the biogeochemistry of submerged soils

299

Upload: guy-kirk

Post on 22-Dec-2016

249 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: The Biogeochemistry of Submerged Soils
Page 2: The Biogeochemistry of Submerged Soils

The Biogeochemistryof Submerged Soils

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 3: The Biogeochemistry of Submerged Soils

The Biogeochemistryof Submerged Soils

Guy KirkNational Soil Resources InstituteCranfield University,UK and formerly International RiceResearch Institute, Philippines

Page 4: The Biogeochemistry of Submerged Soils

Copyright 2004 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester,West Sussex PO19 8SQ, England

Telephone (+44) 1243 779777

Email (for orders and customer service enquiries): [email protected] our Home Page on www.wileyeurope.com or www.wiley.com

All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system ortransmitted in any form or by any means, electronic, mechanical, photocopying, recording,scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 orunder the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham CourtRoad, London W1T 4LP, UK, without the permission in writing of the Publisher. Requests to thePublisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, TheAtrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed [email protected], or faxed to (+44) 1243 770620.

This publication is designed to provide accurate and authoritative information in regard to thesubject matter covered. It is sold on the understanding that the Publisher is not engaged inrendering professional services. If professional advice or other expert assistance is required, theservices of a competent professional should be sought.

Other Wiley Editorial Offices

John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA

Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA

Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany

John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia

John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809

John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1

Wiley also publishes its books in a variety of electronic formats. Some content that appearsin print may not be available in electronic books.

Library of Congress Cataloging-in-Publication Data

Kirk, G. J. D.The biogeochemistry of submerged soils / Guy Kirk.

p. cm.Includes bibliographical references (p. ).

ISBN 0-470-86301-3 (cloth : alk. paper)1. Hydromorphic soils. 2. Soil chemistry. 3. Biogeochemistry. I.

Title.S592.17.H93K57 2004631.4′1—dc22 2003019773

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN 0-470-86301-3

Typeset in 10/12pt Times by Laserwords Private Limited, Chennai, IndiaPrinted and bound in Great Britain by Antony Rowe Ltd, Chippenham, WiltshireThis book is printed on acid-free paper responsibly manufactured from sustainable forestryin which at least two trees are planted for each one used for paper production.

Page 5: The Biogeochemistry of Submerged Soils

Contents

Preface ix

Acknowledgements xi

1 Introduction 11.1 Global Extent of Submerged Soils and Wetlands 11.2 Biogeochemical Characteristics 31.3 Types of Submerged Soil 9

1.3.1 Organic Soils 91.3.2 Mineral Soils 101.3.3 Relation between Soils and Landform 12

2 Transport Processes in Submerged Soils 172.1 Mass Flow 192.2 Diffusion 22

2.2.1 Diffusion Coefficients in Soil 222.2.2 Propagation of pH Changes Through Soil 35

2.3 Ebullition 382.4 Mixing by Soil Animals 39

3 Interchange of Solutes between Solid, Liquid and Gas Phases 45A. WATER 45

3.1 Composition of the Water 453.1.1 Acid and Bases 463.1.2 Speciation 473.1.3 Equilibrium Calculations 50

3.2 pH Buffer Capacity 533.3 Equilibrium with the Gas Phase 54

3.3.1 Floodwater CO2 Dynamics 563.4 Gas Transport Across the Air–Water Interface 58

3.4.1 CO2 Transfer Across the Air–Water Interface 61B. SOIL 65

3.5 The Solid Surfaces in Soils 653.6 The Solid Surfaces in Submerged Soils 69

3.6.1 Organic Matter in Submerged Soils 743.7 Solid–Solution Interactions 76

3.7.1 Adsorption 76

Page 6: The Biogeochemistry of Submerged Soils

vi Contents

3.7.2 Precipitation 793.7.3 Co-Precipitation in Solid Solutions 823.7.4 Inhibition of Precipitation 853.7.5 Equations for Solid—Solution Interactions 87

4 Reduction and Oxidation 934.1 Thermodynamics and Kinetics of Redox Reactions 93

4.1.1 Electron Activities and Free Energy Changes 934.1.2 Redox Potentials 974.1.3 Relation between pe and Concentration of Redox

Couples 974.1.4 pe–pH Diagrams 994.1.5 Energetics of Reactions Mediated by Microbes 102

4.2 Redox Conditions in Soils 1064.2.1 Changes with Depth in the Soil 1074.2.2 Changes with Time 1094.2.3 Calculated Changes in pe, pH and Fe During Soil

Reduction 1134.2.4 Measurement of Redox Potential in Soil 116

4.3 Transformations of Nutrient Elements AccompanyingChanges in Redox 1194.3.1 Transformations of Carbon 1204.3.2 Transformations of Nitrogen 1204.3.3 Transformations of Sulfur 1224.3.4 Transformations of Phosphorus 124

4.4 Oxidation of Reduced Soil 1274.4.1 Kinetics of Fe2+ Oxidation 1284.4.2 Simultaneous Diffusion and Oxidation in Soil 131

5 Biological Processes in the Soil and Floodwater 1355.1 Microbiological Processes 135

5.1.1 Processes Involved in Sequential Reduction 1365.1.2 Nitrate Reduction 1415.1.3 Iron and Manganese Reduction 1425.1.4 Sulfate Reduction 1435.1.5 Methanogenesis 1445.1.6 Aerobic Processes 147

5.2 Macrobiological Processes 1505.2.1 Net Primary Production and Decomposition 1505.2.2 The Floodwater–Soil System 1515.2.3 Floodwater Properties 1525.2.4 Floodwater Flora 1545.2.5 Fauna 159

5.3 Is Biodiversity Important? 163

Page 7: The Biogeochemistry of Submerged Soils

Contents vii

6 Processes in Roots and the Rhizosphere 1656.1 Effects of Anoxia and Anaerobicity on Plant Roots 165

6.1.1 Adaptations to Anoxia 1676.1.2 Armstrong and Beckett’s Model of Root Aeration 170

6.2 Architecture of Wetland Plant Root Systems 1716.2.1 Model of Root Aeration versus Nutrient Absorption 1726.2.2 Root Surface Required for Nutrient Absorption 177

6.3 Nutrient Absorption Properties of Wetland Plant Roots 1806.3.1 Ion Transport in Roots 1806.3.2 Ion Transport in Wetland Roots 184

6.4 Root-Induced Changes in the Soil 1906.4.1 Oxygenation of the Rhizosphere 1916.4.2 The pH Profile Across the Rhizosphere 194

6.5 Consequences of Root-induced Changes 1966.5.1 Nitrification–Denitrification in the Rhizosphere 1966.5.2 Solubilization of Phosphate 1976.5.3 Solubilization of Zinc 2006.5.4 Immobilization of Cations 200

6.6 Conclusions 202

7 Nutrients, Toxins and Pollutants 2037.1 Nutrient and Acidity Balances 203

7.1.1 Nutrient Balances in Ricefields 2037.1.2 Acidity Balances in Ricefields 2087.1.3 Peat Bogs 2107.1.4 Riparian Wetlands 2107.1.5 Tidal Wetlands 211

7.2 Toxins 2127.2.1 Acidity 2127.2.2 Iron Toxicity 2147.2.3 Organic Acids 2157.2.4 Salinity 216

7.3 Trace Elements 2187.3.1 Global Cycling of Trace Elements 2187.3.2 Transport Through Soil and into Plant Roots 2187.3.1 Mobilities of Individual Trace Elements 220

8 Trace Gases 2338.1 Methane 233

8.1.1 Global Budget 2338.1.2 Processes Governing Methane Emissions from Rice 2348.1.3 Modelling Methane Emission 237

Page 8: The Biogeochemistry of Submerged Soils

viii Contents

8.1.4 Estimating Emissions at the Regional Scale 2448.1.5 Possibilities For Decreasing Emissions 246

8.2 Nitrogen Oxides 2478.2.1 Global Budget 2478.2.2 Processes Governing Nitrous and Nitric Oxide

Emissions from Rice 2498.2.3 Differences between Rice Production Systems 250

8.3 Ammonia 2528.3.1 Global Budget 2528.3.2 Processes Governing Ammonia Emissions from Rice 254

8.4 Sulfur Compounds 2568.4.1 Global Budget 2568.4.2 Emissions from Ricefields 256

8.5 Carbon Sequestration 258

References 259

Index 283

Page 9: The Biogeochemistry of Submerged Soils

Preface

This book is about the movements and transformations of energy and matter insoils that are continuously or intermittently submerged with water. Submergedsoils cover a huge area, from 5 to 7 per cent of the Earth’s land surface, and theyare undoubtedly of great practical importance: in local, regional and global ele-ment cycles, as habitats for plants and wildlife, and in food and fibre production.The submerged soils in ricefields, for example, produce the basic food of morethan 2 billion people, a third of the world population. But submerged soils arealso inherently interesting scientifically, and that is the main theme of the book.

When a soil is submerged, air is excluded and the soil quickly becomes anoxic.A submerged soil is therefore an open, anoxic chemical system, surrounded byoxic systems with very different characteristics. Energy enters through photosyn-thesis, and inorganic matter enters with percolating water and by gas exchange.Chemical reactions occur through a complicated interchange between solid, liq-uid and gas phases, largely mediated by biological processes. Further, becauseplants are the main conduits for gas exchange between the soil and overlyingatmosphere, they have a particularly important influence. Submerged soils there-fore provide a unique natural laboratory for studying a great range of physical,chemical and biological processes that are important in environmental systems.They form under a wide range of hydrological, geological and topographicalconditions, but because of the overriding influence of anoxia, the soils and plantsand microbes adapted to them have various characteristics in common.

The book describes the physical, chemical and biological processes operatingin submerged soils and links them to the dynamics of nutrients, toxins, pollutantsand trace gases. Far less research has been done on these topics for submergedsoils than for dryland soils, in spite of their importance. But knowledge and under-standing of them have increased substantially in the past few decades. Much ofthe research has been on rice soils, particularly at the International Rice ResearchInstitute (IRRI) which has been involved in research on submerged soils since itwas founded in 1960. But there is also much in the ecological and environmentalliteratures concerned with natural wetlands. In preparing the book I have aimedto deal with generic principles relevant to both natural and artificial wetlandswith the aim of serving audiences for both.

Page 10: The Biogeochemistry of Submerged Soils

Acknowledgements

I thank the following friends and colleagues for their help in planning the bookand reviewing draft chapters: Dave Bouldin, Roland Buresh, Ralph Conrad,Achim Dobermann, Dennis Greenland, Peter Nye, Bill Patrick, John Sheehy,Siobhan Staunton, Dick Webster and Oswald van Cleemput. I am indebted to theDirector General of IRRI, Ron Cantrell, for the award of a consultancy to writethe book and for his encouragement throughout. Most of the writing was doneduring a sabbatical in the Department of Plant Sciences, University of Cambridge,and I am grateful to the Head of Department, Roger Leigh, and member of theDepartment for their hospitality. The book was completed during my first monthsat the National Soil Resources Institute, Cranfield University, and I am indebtedto the Director, Mark Kibblewhite, for his encouragement and forbearance. Forhelp with the artwork I am grateful to Edwin Javier, Ely Tabaquero and GeneHettel, all of IRRI.

Page 11: The Biogeochemistry of Submerged Soils

1 Introduction

Submerged soils behave and affect the environment in substantially different waysto dryland soils. This chapter discusses the main characteristics and environmentaleffects of submerged soils and the wetlands they support, and their extent acrossthe globe.

1.1 GLOBAL EXTENT OF SUBMERGED SOILS AND WETLANDS

For the purposes of the book I define wetlands as lands that are intermittentlyor permanently inundated with water to a depth of no more than a few metres.Depending on the precise definition applied, estimates of the total global wetlandarea range from 700 to 1000 Mha (Aselmann and Crutzen, 1989; Scharpenseel,1997; Mitsch and Gosselink, 2000). Figure 1.1 shows their approximate distri-bution and Table 1.1 the extents of different types distinguished by hydrology,vegetation and soil characteristics. The largest areas are the bogs and fens inpolar and boreal regions in North America and Russia (34 % of total area); trop-ical swamps, especially in East Africa and South America (14 % of total area);tropical floodplains, especially of the Amazon and the rivers of South East Asia(10 %); and temperate and tropical ricefields (4 and 12 %, respectively). Almosthalf the global wetland area is in the tropics. There has been considerable lossof wetlands in many parts of the world over the past 200 years as a result ofconversion to agricultural and aquacultural uses. In the US for example, it isestimated that the area has declined from 89 Mha in the 1780s to 49 Mha in the1980s (Mitsch and Gosselink, 2000).

A special class of wetland is the lowland ricefield, which accounts for almost afifth of the wetland area worldwide. Much of our knowledge and understandingof submerged soils has been gained from research on rice soils. The successof rice as a food crop stems from its origins as a wetland plant and its abilityto withstand soil submergence with the attendant improvements in water andnutrient supplies. A corollary is that rice is more sensitive to water deficiencythan most other crops, and the critical factors in its productivity are the supplyof water to the soil, from rain, river, reservoir or groundwater, and the abilityof the soil to retain water. Hence most rice is produced and the highest yieldsattained on the alluvial deposits associated with major rivers and their deltas.More than 90 % of the production is in Asia, distributed unevenly over four rice

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 12: The Biogeochemistry of Submerged Soils

2 Introduction

40°

20°

20°

40°

160° 140° 120° 100° 80° 60° 40° 20° 0° 20° 40° 60° 80° 100° 120° 140° 160°

Equator

Major Wetland Area

Area with AbundantWetlands

Figure 1.1 Global distribution of wetlands (Mitsch and Gosselink, 2000). Reproducedby permission of Wiley, New York

Table 1.1 Global extent of wetlands of different types

Area (Mha)

Polar Boreal Temperate Tropical Total

Bogs 21 104 42 20 187Fens 54 62 32 — 148Swamps — 1 10 102 113Marshes — — 17 10 27Floodplains — — 8 74 82Shallow lakes — — 1 11 12Ricefields — — 29 80 109Total 75 167 139 297 678

Definitions of wetland types:Bogs are raised peat-producing wetlands formed in wet climates where organic material has accumulated overlong periods. Because they are raised, water and nutrients are entirely derived from the atmosphere, and they aretherefore nutrient deficient and acid. Sphagnum moss is the main vegetation, though other types of vegetation arealso present in tropical regions.Fens are peat-producing wetlands that receive water and nutrients through inflow from neighbouring land. Theyare generally less acid than bogs and may be alkaline, and tend to be dominated by grasses and sedges. Becauseof their better nutrient status they are generally more prolific than bogs.Swamps are forested, freshwater wetlands on submerged soils in which little peat accumulates. This is the USdefinition; elsewhere the term also includes non-forested wetlands with reeds. Swamps tend to form in warmerclimates.Marshes are herbaceous freshwater, non-peat-producing wetlands dominated by grasses, sedges or reeds. Thedistinction between swamps and marshes may be blurred.Floodplains are periodically inundated areas along rivers or lakes. Their vegetative cover is variable.Shallow lakes are open water bodies a few metres deep. Only considered foe temperate and tropical regions; inpolar and boreal regions it is difficult to separate shallow lakes from bogs and fens.Ricefields exclude upland rice. The physical area is calculated from the sum of irrigated rice (51 Mha of which47 % is double- or triple-cropped with rice and 33 % under rice–wheat rotation), rainfed lowland rice (54 Mha)and deepwater rice (4 Mha). The riceland of China is taken to be all temperate.Sources : Aselmann and Crutzen (1989); Huke and Huke (1997); IRRI (2002).

Page 13: The Biogeochemistry of Submerged Soils

Biogeochemical Characteristics 3

Table 1.2 Rice ecosystems in the main rice-producing countries in Asia

Harvested area (kha)

Irrigated Rainfed lowland Flood-prone Upland Total

WSa DSa 0–30b 30–100b

India 15 537 4123 11 985 4447 1364 5060 42 516China 20 490 9146 1990 0 0 499 32 125Indonesia 2963 2963 2872 1006 2 1209 11 015Bangladesh 351 2267 3271 2873 1220 697 10 679Thailand 274 665 6382 1778 342 203 9644Vietnam 1630 1630 1963 651 177 322 6373Myanmar 1812 1386 2033 478 362 214 6285Philippines 1175 1029 911 341 0 165 3621Pakistan 2125 0 0 0 0 0 2125Cambodia 140 165 1069 349 152 24 1899Nepal 706 24 406 166 118 68 1488Korea, Rep. of 776 0 326 0 0 1 1103Sri Lanka 377 251 213 26 0 0 867Total 49 211 24 003 34 056 12 131 3737 8853 13 1991

a Wet/dry season.b Depth of floodwater (cm).Definitions of ecosystem types:Irrigated. Grown in levelled, bunded fields with good water control. Crop is transplanted or direct seeded inpuddled soil, and a shallow floodwater is maintained on the soil surface so that the soil is predominantly anoxicduring crop growth.Rainfed lowland. Grown in level to gently sloping, bunded fields that are flooded for at least part of the croppingseason. Water depths exceed 100 cm for no more than 10 consecutive days. Crop is transplanted in puddled soil ordirect seeded on puddled or ploughed dry soil. During season soil alternates between oxic and anoxic conditionsof variable duration and frequency.Flood-prone. Distinguished from rainfed lowland rice by extent and duration of flooding. Fields are flooded to atleast 100 cm and often much more for at least 10 consecutive days in the growing season. Crop is transplantedin puddled soil or direct seeded on ploughed dry soil; soil may alternate between oxic and anoxic conditionsduring season.Upland. Grown in level to steeply sloping fields that are rarely flooded. No effort is made to impound water asfor other rice ecosystems. Crop is direct seeded on ploughed dry soil or dibbled in wet, non-puddled soil.Source: IRRI (2002). Reproduced by permission of IRRI.

‘ecosystems’ distinguished by land and water characteristics and adaptations ofthe rice plant to them. These are defined in Table 1.2 together with their extentsin the main rice-producing countries in Asia.

1.2 BIOGEOCHEMICAL CHARACTERISTICS

Wetlands are intermediate between upland systems and true aquatic systems, bothin terms of their hydrologies, being intermittently to permanently flooded, andin terms of their biogeochemistries, being sources, sinks and transformers of

Page 14: The Biogeochemistry of Submerged Soils

4 Introduction

nutrients and carbon, whereas uplands tend to be sources and aquatic systemssinks. Three types of wetland are distinguished based on hydrology (Figure 1.2):

a. fluxial, which receive water wholly or in part from surface flow, such as inrunoff or streams;

b. phreatic, which receive water from groundwater that rises to the soil surfacefor at least part of the year; and

c. pluvial, which receive water entirely from rainfall.

In fluxial wetlands water flowing in from neighbouring upland brings with itsediment and nutrients which are only slowly lost to deepwater areas downslope,and may be supplemented by seasonal inflow from deepwater areas. Because ofthe net inflow of nutrients, the abundance of water, and beneficial changes in thesoil resulting from chemical reduction under anoxia, fluxial wetlands are amongthe most productive ecosystems on Earth. By contrast pluvial wetlands rely onnutrients brought in by rainfall or fixed biologically from the atmosphere, and theytherefore tend to be much less productive. Phreatic wetlands are intermediate.

As sources, sinks and transformers of matter and energy, wetlands have impor-tant roles in element cycles at local, regional and global scales. They contributeto the global stability of carbon dioxide, methane and sulfur in the atmosphereand of available nitrogen and phosphorus in surface waters, and they are impor-tant regionally as sinks for organic and inorganic pollutants released into themaccidentally or otherwise. These topics are introduced in the following sections;all are returned to in greater detail later in the book.

Carbon Balances in Wetlands

Table 1.3 shows the net primary production of different wetlands compared withupland and aquatic ecosystems. The generally greater productivity of wetlands isevident. Net primary production (NPP) is the gross rate of carbon fixation in pho-tosynthesis less the rate of loss in plant respiration. The chief factors governingNPP are radiation, temperature, water, nutrients and toxins. Hence for a given typeof wetland, NPP tends to increase from polar to tropical regions as incident radi-ation and day length increase; correspondingly nutrients and temperature becomeincreasingly limiting. There are of course interactions between these changes. Forexample, the greater productivity of temperate compared with tropical ricefieldson a per crop basis shown in Table 1.3 arises because of interactions betweenradiation and temperature: in temperate rice areas with high radiation during thegrowing season, low night-time temperatures result in lower respiratory lossescompared with tropical areas and hence greater net productivity.

Because of their often high biological productivity and low rates of decompo-sition under anoxia, wetlands are one of the largest terrestrial sinks for carbon.They account for about a third of the soil carbon globally (Table 1.4). Howeverthere are large differences between wetland types. Organic wetland soils tend

Page 15: The Biogeochemistry of Submerged Soils

5

Wet

lan

dA

qu

atic

Up

lan

dW

etla

nd

Wet

lan

d

Inte

rmitt

ently

-per

man

ently

flood

ed

Inte

rmitt

ently

-per

man

ently

flood

ed

Inte

rmitt

ently

-per

man

ently

flood

edD

ryP

erm

anen

tlyde

eply

flood

ed

Net

exp

ort

Net

exp

ort o

rim

port

, and

tran

sfor

mat

ion

Net

impo

rtN

et im

port

and

tran

sfor

mat

ion

Tra

nsfo

rmat

ion

Low

tom

ediu

mG

ener

ally

high

Gen

eral

lylo

w

Gen

eral

lym

ediu

mto

hig

h

Gen

eral

lylo

w

Energyconversion

Nutrientregime

Waterregime

(a)

Flu

xial

(b)

Phr

eatic

(c)

Plu

vial

……

……

……

……

…… …

… ……

……

………

…………

……

…… …

……

……

Up

lan

dU

pla

nd

Up

lan

dU

pla

nd

Fig

ure

1.2

Wet

land

sin

(a)

fluxi

al,

(b)

phre

atic

and

(c)

pluv

ial

land

scap

es

Page 16: The Biogeochemistry of Submerged Soils

6 Introduction

Table 1.3 Net primary production of wetlands compared with other ecosystems

Net primary production (g C m−2 year−1)

Polar Boreal Temperate Tropical

WetlandsBogsa 100–300 300–700 400–800 600–1200Fensa 100–300 400–700 400–1200Swampsa 500–1000 700–1500 1500–3000Marshesa 800–2000 1500–4000Floodplainsa 800–1800 1500–2500Shallow lakesa 400–600 500–800Wetland riceb 850 1050

OthersForestc 430 650 620 (dry), 800 (humid)Grasslandc 320 450Arabled 750 600Desertc <100 <100

Source:a Aselmann and Crutzen (1989).b Irrigated rice, assuming temperate mean grain yield = 8 t ha−1, tropical mean grain yield = 5 t ha−1 ×2 crops year−1, harvest index = 0.5, root mass/above-ground mass = 0.2, mass of C/plant mass = 0.44.c Houghton and Skole (1993).d Based on data of Evans (1993) for wheat and maize yields and harvest indices.

to be confined to cold regions and are rare in tropical regions where decom-position is accelerated by higher temperatures and seasonal wetting and dryingof the soil because of the seasonal rainfall pattern. Decomposition is particu-larly accelerated in wetland rice soils for reasons connected with their fertilityand particular ecology, and they therefore tend not to have large organic mat-ter contents in spite of their very large productivities. Mineral wetland soilscover about 5 % of the Earth’s land surface and account for 18 % of soil carbon;organic wetland soils account for 12 % of soil carbon though less than 3 % ofthe land surface.

Countering their value as a carbon sink, wetlands are also the largest singlesource of atmospheric methane accounting for nearly 50 % of global emissions.Wetlands may also be sources of particulate carbon for aquatic systems down-stream. Salt marshes especially are important carbon sources for adjacent estuar-ies where in situ biological production may be limited by nutrient supplies. Smalldifferences in climate, water and nutrient regimes, and land use can markedlychange the delicate carbon balance in wetlands.

Nitrogen

Nitrogen occurs in several oxidation states under Earth-surface conditions, from+V to −III, and its fixation from and loss to the atmosphere depend on trans-formations between these states. Because wetlands are the main reducing system

Page 17: The Biogeochemistry of Submerged Soils

Biogeochemical Characteristics 7

Table 1.4 Estimates of organic carbon in wetland and other soilscompared with other global carbon pools

Pool Area(Mha)

Carbonstocks(Pg)

Annualchange

(Pg year−1)

SoilWetland minerala,b 670 380Wetland organica,b 350 260Wetland ricec 109 10Forestd 4200 790Grasslanda 3000 500Arablea 1400 140Desert 3750 10

Total 13 500 2090

Otherse

Atmosphere 770 +3Land

Short-lived biota 130Long-lived biota 700 −1 to −2Litter 60

SeaSurface water 725Deep water 37 675

Fossil fuel 5000–10 000 −5Sediments 108

Source:a Scharpenseel (1997).b Armentano and Verhoeven (1990).c IRRI (2002) (2 % C to 30 cm depth).d Dixon et al. (1994).e Bolin and Cook (1983) (atmosphere adjusted to 2003).

in most landscapes and maintain the widest range of redox conditions of anyecosystem, they have a central role in the global nitrogen cycle. Nearly 20 % ofnatural N2 fixation occurs in wetlands (Table 1.5) because of the favourable waterand nutrient status for N2 fixing organisms (Buresh et al., 1980; Bowden, 1987).Wetlands are also important sinks for nitrate which under anoxic conditions isreduced to N2 by microbes in denitrification:

5CH2O + 4NO3− + 4H+ −−−→ 2N2 + 5CO2 + 7H2O

Table 1.5 shows the importance of denitrification in wetlands on a global scale.Further, agricultural wetlands are important sources of NH3 which is emitted byvolatilization of ammoniacal-N in the floodwater:

NH4+ −−−→ NH3 + H+

Page 18: The Biogeochemistry of Submerged Soils

8 Introduction

Table 1.5 Nitrogen fixation and denitrification in wetlands

N fixation Denitrification

Mean rate(g m−2 year−1)

Total(Tg year−1)

Mean rate(g m−2 year−1)

Total(Tg year−1)

TemperateBogs/fens 1.0 3.0 0.4 1.2Floodplains 2.0 6.0 1.0 3.0

TropicalBogs 1.0 0.5 0.4 0.2Swamps 3.5 7.8 1.0 2.2Floodplains 3.5 5.2 1.0 1.5Ricefields 3.5 5.0 7.5 10.8

Total 27.5 18.9Total terrestrial 139 43–390

Source: after Armentano and Verhoeven (1990).

The floodwater often has a high pH as a result of CO2 removal by photosynthe-sizing organisms, favouring NH3 volatilization.

As a result of gaseous losses, and in spite of biological N2 fixation, N is oftenthe most limiting nutrient in wetlands together with P. It is also often one of themost limiting nutrients in coastal waters, so the extent of denitrification of NO3

−in coastal wetlands has a particular significance. Pollution with excess NO3

−causes hypoxia in coastal waters and lakes worldwide (Mitsch and Gosselink,2000). Currently atmospheric N2 is fixed artificially for N fertilizers at morethan double the rate of natural biological N2 fixation, so the return of N to theatmosphere through denitrification in wetlands is an important brake on excessNO3

−.

Sulfur

Like nitrogen, sulfur occurs in several oxidation states in submerged soils andits transformations are microbially mediated. Sulfate washed into wetlands ordeposited from the atmosphere is largely reduced to S2− in reactions mediatedby sulfate reducing bacteria. Subsequent precipitation of S2− with metals, espe-cially Fe2+, results in more or less permanent removal of the S from the globalS cycle. Wetlands are therefore a potentially important sink for excess S releasedin fossil fuel burning. Some of the S2− may be emitted as H2S in organic soils,but in submerged mineral soils the concentration of Fe2+ is usually sufficient toprevent this. Hence measured emissions of H2S and other forms of volatile Sfrom wetlands are generally modest, and in general wetlands are a net sink forS. The concentrations and availability of S in wetland soils rarely limit biologi-cal production.

Page 19: The Biogeochemistry of Submerged Soils

Types of Submerged Soil 9

Phosphorus

Though orthophosphate itself is generally not reduced in submerged soils, reduc-tion of ferric iron compounds and changes in the electrochemical properties ofthe surfaces with which orthophosphate reacts strongly affect its solubility anddynamics. In agricultural wetlands with a history of P fertilization, submergenceoften results in enhanced availability of P to plants, and responses to additionsof further fertilizer in ricefields are often weak. However in natural wetlands, Pconcentrations are much smaller and P washed into the soil becomes stronglysorbed on the surfaces of reduced soil constituents. Phosphorus brought in withsediment may also be effectively filtered out of water passing through wetlands.Phosphorus retention is therefore a highly valuable attribute of wetlands receivingdiffuse pollution. However, the net primary productivity of most natural wetlands,particularly freshwater wetlands, is limited by deficiency of P.

Metals and Other Pollutants

Heavy metals, toxic organics and other pollutants have often frequently beenadded to wetlands both accidentally and on purpose, exploiting their bufferingand storage capacities. The chemistry of submerged soils and sediments is suchthat pollutants may be effectively removed from the percolating water in redox,sorption and precipitation reactions. But the effects of long-term accumulation ofpollutants on nutrient cycles and other wetland functions are not well understood.

1.3 TYPES OF SUBMERGED SOIL

The principal distinguishing feature of wetland soils is that they develop underpredominantly anoxic conditions. Although anoxia is also sometimes found inother ecosystems, it prevails in wetlands and dominates soil properties. Becauseof the very large organic matter content of some wetland soils, a rough separa-tion into organic and mineral types based on organic matter content is a usefuldelineation.

1.3.1 ORGANIC SOILS

The USDA (1999) defines organic wetland soils as having an organic carboncontent of at least 12 % if the mineral fraction has no clay, 18 % if ≥ 60 % clay,or 12–18 % if < 60 % clay. Further differentiation is based on the botanical originof the organic matter–whether mosses, herbaceous plants, or woody plants–andits state of decomposition: fibrists contain predominantly recognizable, little-decomposed plant debris, saprists predominantly well-decomposed plant debris,

Page 20: The Biogeochemistry of Submerged Soils

10 Introduction

and hemists are intermediate. Equivalent terms in other classifications are peats,mucks and mucky-peats, respectively.

By virtue of being unconsolidated and structureless in comparison with mineralsoils, organic soils have much smaller bulk densities, greater porosities and watercontents (>80 %), and smaller load-bearing capacities. These factors make theirartificial management highly problematic.

Organic wetland soils also tend to be poor chemically. Organic soils tendto form under nutrient deficient conditions, which limit organic matter decom-position. This occurs particularly in pluvial wetlands where the only nutrientinputs are from rainfall and biological fixation from the atmosphere. Althoughthe organic material may have a high cation exchange capacity by virtue of thecharged functional groups on its surfaces, the exchange capacity tends to bedominated by H+ ions and the soil is acid. The acidity is an inevitable conse-quence of the circumstances in which organic soils form. By contrast, mineralwetland soils tend to have pHs near neutral as a result of electrochemical changesaccompanying soil reduction.

1.3.2 MINERAL SOILS

The most productive wetlands are on mineral soils, often developed on alluvialdeposits in fluxial wetlands. Nutrients and fertile sediments seasonally flow intothese areas under high rainfall and surface water flow.

Under prolonged submergence, mineral soils develop so-called redoximorphicfeatures associated with anaerobic soil metabolism (Figure 1.3). As oxygen isexcluded by submergence, soil organisms must use other soil constituents as theiroxidizing agents in deriving energy from organic matter. This typically occurs inthe sequence: nitrate ions to nitrogen, manganic manganese to manganous, ferriciron to ferrous, and then sulfate ions to sulfide. Subsequently organic matteris decomposed by methanogenic bacteria to carbon dioxide and methane. Thissequence is predicted by thermodynamics.

The most visible change associated with this process is the reduction of thered and brown compounds of ferric iron to blue-grey compounds of ferrous iron.Subsequent translocation of soluble ferrous iron to zones where oxygen enters thesoil–such as at the soil surface or near plant roots–produces reddish-brown mot-tles of insoluble ferric iron. Likewise there may be movement and re-oxidationof manganous manganese forming black manganic compounds. These changesproduce the characteristic redoximorphic features of submerged mineral soils.

The soil profile that develops under prolonged submergence is sensitive to thenature of the water saturation. In very wet areas where the soil is perennially sat-urated, the profile may be largely reduced throughout with little development ofdistinct pedogenic horizons. Such conditions arise in tidal marshes, lake margins,floodplains or in wet footslope areas. In better-drained areas where the flooding

Page 21: The Biogeochemistry of Submerged Soils

Types of Submerged Soil 11

0

10

20

30

40

Dep

th (

cm)

Ofw

Apox

Apg

B

Layer of standing water occupied bymicro and macro-fauna and -flora

Oxic floodwater−soil interface, a fewmm to a few cm thick depending onfloodwater aeration, soil reducingconditions, mixing by soil animalsand percolation

Anoxic soil layer in which pe+pH isbelow the range at which Fe(III) isreduced, except in the rhizosphere

Subsoil properties vary with the typeof water saturation. In aquicmoisture regimes the whole horizonis largely reduced throughout; inepiaquic regimes, where the watertable is perched, the horizongenerally remains oxic and ismottled along wide pores

Figure 1.3 Schematic profile of a submerged soil with redoximorphic features

is more intermittent, there may be more-distinct layers in the soil with differ-ent redoximorphic features. The transition between these soil types in partiallydrained wetlands may occur in a matter of decades.

At the boundary between uplands and wetlands there is, in some circum-stances, an interaction between organic matter accumulation in sediments and thedevelopment of wetland conditions. Some level of organic matter accumulationis required to drive anaerobic metabolism. But also, because, in general, well-decomposed organic matter improves the water holding capacity of mineralsoils, particularly in medium to coarse textured sediments, and particularly ifthe clay mineralogy is dominated by low activity kaolinitic clays, there is afeedback between organic matter accumulation and the extent and duration ofwater saturation.

Particular modifications of these patterns occur in wetland rice soils. Repeatedworking of the soil for rice often results in permanent changes that mask thesoil’s original character. Gross changes are caused by levelling, terracing andpuddling the soil for rice, which destroys the soil structure. Over time a ‘traffic’pan of compacted soil often develops, 5–10 cm thick at 10–40 cm depth. Thishas a greater bulk density and is less permeable than the overlying surface soil,

Page 22: The Biogeochemistry of Submerged Soils

12 Introduction

but has similar texture. Over time the surface soil often becomes more coarse-textured, possibly because of weathering of clay under alternate flooding anddrainage (Brinkman, 1970; Moormann and van Breemen, 1978). Clay may alsobe lost from the surface during puddling by movement downslope with surfacewater. But equally clay may be added from upslope. Freely drained soils thatare repeatedly flooded and puddled for rice may show downward movement ofreduced Fe and Mn and their subsequent accumulation in oxidized forms at theboundary with oxic subsoil. However this is rarely seen in naturally hydromor-phic, finer textured soils.

Because of rice’s origins as a wetland plant, it is more sensitive to waterdeficiency than most other crops. But provided sufficient water is supplied toperiodically inundate the land and the soil is able to retain the water, rice willthrive on almost any type of soil. The productivity of rice land therefore oftendepends more on position in the landscape and soil physical properties thanon the finer attributes of the soil. Nonetheless, subtle differences in propertiesdistinguish productive and ‘problem’ soils and affect the behaviour of the soil inthe environment.

1.3.3 RELATION BETWEEN SOILS AND LANDFORM

Most of the landforms in which wetlands form can be seen in tracing a river fromits source in hilly or mountainous areas to its outflow in coastal floodplains andthe sea. The main landforms are inland valleys, alluvial fans or fan complexes,meander or lacustrine floodplains, and alluvial terraces (Figure 1.4), and eachof these is associated with particular soils as illustrated for ricelands in Asiain Table 1.6. This section gives a brief description of these associations. Morecomplete descriptions are given in Moormann and van Breemen (1978), Driessenand Moormann (1985) and Richardson and Vepraskas (2001). Following theseauthors I use the USDA (1999) soil classification.

Inland Valleys

Wetlands occur on the valley floors and the lower slopes. The soils vary widelywith parent materials and other factors, but there are some general patterns. On thevalley floors, slopes decrease from the top to the bottom and the age and textureof the deposits vary accordingly. Where deposition is most active, the soils areyoung and have little profile development. These are Entisols. But most soils inthe valley bottoms show at least some profile development and are Inceptisols orAlfisols where there is a pronounced dry season.

Where the valley slopes have been terraced for rice and the soil has remainedin situ for a long time–hundreds of years–there may be inherited clay illuviationleading to man-made Alfisols and Ultisols. Artificial Entisols may occur where

Page 23: The Biogeochemistry of Submerged Soils

Types of Submerged Soil 13

Inlandvalley

Inlandvalley

Inland valleys

Marine sediments

Coastal

Sea

Tidal

land

Alluvial fan

PlainMeander

Terrace

Terrace

Terrace

Floodplain

Riverine sediments

Figure 1.4 Major wetland land forms (Moorman and van Breemen, 1978). Reproducedby permission of IRRI

terracing has completely disturbed the original soil profile and also on valleybottoms that have been perennially irrigated with muddy water.

Alluvial Fans

The soils again vary greatly with the age and origin of sediments, from youngEntisols to well-developed Alfisols and Ultisols. There are some common trends.The deposits are often coarsest and youngest near the apex of a fan and theybecome finer and older towards the more gently sloping base. There are corre-sponding differences in hydrology with seepage of water from the better-drainedupper parts and accumulation in the lower resulting in marshland to developwhere fans meet adjacent floodplains. Entisols and Inceptisols are common inthe upper fans; Alfisols in the lower fans or Ultisols where the surroundinguplands are highly weathered.

Active Floodplains

The main wetland areas are in the river basins. Levee deposits become increas-ingly fine textured with distance downstream and distance away from the river.The soils are mostly Entisols and Inceptisols, and, where levees grade into basins,Alfisols or Ultisols. Soils in the basins are typically fine-textured and wet butmany types occur due to differences in parent materials, rates of deposition,

Page 24: The Biogeochemistry of Submerged Soils

14

Tabl

e1.

6T

hem

ain

wet

land

soils

inri

veri

nean

dco

asta

lla

ndfo

rms

His

toso

lsE

ntis

ols

Ince

ptis

ols

Alfi

sols

Ult

isol

s

Aqu

ents

Fluv

ents

Aqu

epts

Tro

pept

s/O

chre

pts

Aqu

alfs

Ust

alfs

/U

dalf

sA

quul

tsU

stul

ts/

Udu

lts

Coa

stal

plai

ns++

++na

+++

+−

−−

Inla

ndva

lleys

+++

++++

++

++

+A

lluvi

alfa

ns−

+++

++++

+++

++

Floo

dpla

ins

−+

++++

++++

+++

+A

lluvi

alte

rrac

es−

++

+++

++++

++++

USD

A(1

999)

soil

clas

sific

atio

n.na

,no

tap

plic

able

;−,ab

sent

orra

re;+

,co

mm

on;+

+,ab

unda

nt.

Exp

lana

tion

ofso

ilca

tego

ries

(FA

O(1

999)

equi

vale

nts

inpa

rent

hese

s):

His

toso

ls(H

isto

sols

)ha

vehi

ghor

gani

cm

atte

rth

roug

hout

the

profi

le.

Ent

isol

ssh

owno

evid

ence

ofso

il-fo

rmin

gpr

oces

ses

lead

ing

topr

ofile

deve

lopm

ent:

Aqu

ents

(Gle

ysol

s,pt

;F

luvi

sols

,pt

)ar

efo

rmed

inco

ntin

uous

lyor

near

-con

tinuo

usly

wet

envi

ronm

ents

;F

luve

nts

(Flu

viso

ls,

pt)

inre

cent

allu

vium

inar

eas

that

are

freq

uent

lyflo

oded

byri

vers

depo

sitin

gne

wse

dim

ent.

Ince

ptis

ols

show

wea

kpr

ofile

deve

lopm

ent:

Aqu

epts

(Gle

ysol

s,pt

;T

hion

icL

uvis

ols,

pt)

are

wat

er-s

atur

ated

for

atle

ast

part

ofth

eye

aran

dsh

owgr

ayor

rust

ym

ottli

ng;

Trop

epts

(Cam

biso

ls,

pt)

are

wel

l-dr

aine

dan

doc

cur

inw

arm

regi

ons

with

only

slig

htan

nual

tem

pera

ture

chan

ges;

Och

rept

s(C

ambi

sols

,pt

)oc

cur

inre

gion

sw

ithgr

eate

ran

nual

tem

pera

ture

chan

ges.

Alfi

sols

show

mar

ked

clay

tran

sloc

atio

ndo

wn

the

profi

lew

ithou

tex

cess

ive

depl

etio

nof

base

s:A

qual

fs(G

leyi

cL

uvis

ols)

are

wat

er-s

atur

ated

for

part

ofth

eye

ar;

Ust

alfs

(Luv

isol

s,ex

cept

Gle

yic,

pt;

Eut

ric

Nito

sols

,pt

)ar

ese

ason

ally

dry;

Uda

lfs

(Luv

isol

s,ex

cept

Gle

yic,

pt;

Eut

ric

Nito

sols

,pt

)ar

eco

ntin

uous

lym

oist

.U

ltis

ols

show

mar

ked

clay

tran

sloc

atio

nw

ithin

tens

ive

leac

hing

and

depl

etio

nof

base

s:A

quul

ts(G

leyi

cA

cris

ols;

Plin

thic

Acr

isol

s,pt

;D

ystr

icP

lans

ols,

pt),

Ust

ults

(Acr

isol

s,pt

;D

ystr

icN

itoso

ls)

and

Udu

lts

(Acr

isol

s,pt

;D

ystr

icN

itoso

ls)

asfo

rA

lfiso

ls.

Sour

ce:

afte

rD

ries

sen

and

Moo

rman

(198

5).

Rep

rodu

ced

bype

rmis

sion

ofIR

RI.

Page 25: The Biogeochemistry of Submerged Soils

Types of Submerged Soil 15

relic riverbeds, and other factors: Entisols and Inceptisols in young depositsand Alfisols and Ultisols in older. More humic soils, Mollisols, occur locally indepressions in richer floodplains and Vertisols (swelling clay soils that are ‘self-mulching’ as a result of seasonal shrinking and cracking) in basins receivingbase-rich water from adjacent higher areas.

Alluvial Terraces

Terraces vary in age, parent material, height above the base drainage andtopography. Within a given river system, the lower terraces are the youngest andthe main soils are Inceptisols with Alfisols in poorly drained areas. Higher, older,Pleistocene terraces have soils reflecting their age and degree of pedogenesis.The relatively low Pleistocene terraces accompanying many rivers in SoutheastAsia are important rice areas and the soils are Ultisols and to a lesser extentAlfisols. The highest, oldest Pleistocene terraces are often so dissected and highlyweathered that they are not suitable for rice.

Coastal Plains

Rapidly aggrading coastal plains are all wet and young and many of the soils areEntisols with no distinct profile development. Upon ripening, whether throughdrainage or further accretion of sediment, they may develop into Inceptisols.Further inland, soils are increasingly older and profile development more con-spicuous. The groundwater is shallow and the soils are grayish with mottlingreflecting seasonal fluctuations in the water table. Similar soils occur in slowlyaggrading or stationary coastal plains, but the greater extent and duration of man-grove vegetation and more intense tidal influence often cause accumulation ofpyrite leading to potential ‘acid sulfate’ soils. Upon drainage and aeration thesebecome extremely acid and are notoriously difficult to manage (Section 7.2).Where the land is protected from direct intrusion by the sea by beach ridges, andsedimentation is minimal, Histosols may develop. Typically these areas have adomed relief with raised central portions where the peat-forming forest vegetationgrew, grading into Entisols close to the rivers and sea.

Page 26: The Biogeochemistry of Submerged Soils

2 Transport Processesin Submerged Soils

The properties of submerged soils are, to a large extent, determined by transportprocesses controlling the fluxes of solutes and gases through the soil or throughplants growing in it. For example, the reason the soil rapidly becomes anoxicfollowing submergence is the much slower transport of oxygen through the water-filled pores of submerged soil than through the air spaces of well-drained soil.Diffusion coefficients in the liquid phase are four orders of magnitude smallerthan those in the gas phase. It therefore makes sense to start with an account of thevarious transport processes that operate in submerged soils. Transport processesin plants are considered in Chapter 6.

Because of the central importance of transport, and because there is a well-established theory and mathematics of transport processes in soils, submergedsoils lend themselves well to mathematical modelling. Models necessarily giveonly a crude picture, particularly of the biological processes. But some form ofmodelling is essential to unravel the complex interactions taking place. Most mod-els involving transport processes in soils are based on the ‘continuity equation’which relates the change in mass of a substance in a small volume of soil over asmall time to the fluxes of the substance into and out of the volume. I here explainthe basis of the continuity equation and then describe the transport equationsderived from it that are used later in the book. For an introduction to the mathe-matics of transport processes in environmental systems see Crank et al. (1981).

Considering the mass balance of a solute moving in soil between two planesof unit cross-section at distances x and x + δx, the rate of change in mass isequal to the rate of entry across the plane at x less the rate of removal across theplane at x + δx. Hence

δx

(∂C

∂t

)x

≈ (Fx − Fx+δx)t ≈ −δx

(∂F

∂x

)t

(2.1)

where Fx and Fx+δx are the fluxes across x and x + δx and C is the amount ofsolute per unit volume of soil. In the limit δx → 0,(

∂C

∂t

)x

=(

−∂F

∂x

)t

(2.2)

This is the continuity equation in one dimension.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 27: The Biogeochemistry of Submerged Soils

18 Transport Processes in Submerged Soils

If the movement is solely by diffusion, then from Fick’s first law,

F = −D

(∂C

∂x

)t

(2.3)

where D is the solute diffusion coefficient. If the soil solution is also moving,then the solute will also be carried by mass flow, and

F = −D∗ ∂C

∂x+ vCL (2.4)

where v is the water flux in the x direction, CL is the concentration of the solutein the soil solution and D∗ is the dispersion coefficient which differs from thediffusion coefficient because the movement of the solution itself causes somedispersion of the solute.

The continuity equation for combined diffusion and mass flow is obtained bycombining Equations (2.2) and (2.4):

∂C

∂t= ∂

∂x

(D∗ ∂C

∂x− vCL

)t

(2.5)

This is an expression of Fick’s second law.The concentration of the solute may also change as a result of processes occur-

ring within the volume δx. This is allowed for by adding a term R(C, x, t) toEquation (2.2) to give

∂C

∂t= ∂

∂x

(D∗ ∂C

∂x− vCL

)t

+ R(C, x, t) (2.6)

Note that R can be positive or negative. Generally conditions in submergedsoils are strongly affected by the vegetation present, which acts as the mainconduit for gas transfer between the soil and overlying atmosphere. The effectsof vegetation can be allowed for in the R term, suitably modified with depth inthe soil and time. Time-dependent reactions, microbially mediated reactions andother reactions adding or removing the solute can be represented with additionalR terms.

The equivalent equation for movement normal to a cylinder, such as a plantroot, is

∂C

∂t= 1

r

∂rr

(D∗ ∂C

∂r− vCL

)t

+ R(C, x, t) (2.7)

where r is the radial distance from the axis of the cylinder.In simple cases these equations can be solved analytically but more often

numerical solutions are necessary.

Page 28: The Biogeochemistry of Submerged Soils

Mass Flow 19

2.1 MASS FLOW

Submergence greatly alters a soil’s hydraulic properties. Following submergence,air trapped in the pores inside aggregates becomes compressed. Further compres-sion develops as volatile products of respiration accumulate in the pores and as2:1 type clays swell. As a result, large aggregates tend to rupture. Further rup-ture occurs as a result of the dissolution of organic matter and oxides, whichact as cementing agents within aggregates (Greenland, 1981). Hence in the firstfew days following submergence, the permeability of the soil increases as gasesaccumulate in the pores. But as the soil begins to disaggregate, the permeabilitygradually declines. The decline accelerates as pores become clogged with dis-persed clay and other debris. Allison (1947) found the decrease in permeabilitywas less if the soil was sterilized, indicating that the effects of microbes wereimportant, presumably because of increased disruption of aggregates with theaccumulation of respiratory gases and dissolution of cementing agents.

The extent of disaggregation varies greatly between soils and with the qualityof the water. In pure layer silicate systems with high pH or sodium saturation,the disintegration of aggregates and dispersion of clays can be near complete.Whereas in highly structured soils with large contents of organic matter orhydrous oxides, aggregation may be little affected, although it may be easilydisrupted by subsequent application of force.

In wetland rice cultivation, further disaggregation is caused by the process oftilling the soil when wet, which is an integral part of the land preparation priorto transplanting. Wet tillage results in near complete destruction of water-solubleaggregates and dispersion of fine clay particles. The aim is to reduce losses ofwater through percolation, both to conserve water and to control weeds, and tofacilitate transplanting. Some flow through of water should be maintained so thatthe soil does not become entirely anoxic. Also, if the structure is completelydestroyed the soil will dry only very slowly following the rice crop, and thiswill delay the establishment of a following dryland crop. Table 2.1 shows theeffect of puddling on percolation rates in a range of flooded soils measured

Table 2.1 Effect of puddling on percolation rates in a range of flooded Philippine soils

Soil Mineralogy Clay Percolation rate(%) (cm day−1)

Unpuddled Puddled

Psamment Siliceous 9 267 0.45Fluvent Mixed 24 215 0.17Aquept Montmorillonitic 30 183 0.05Aqualf Montmorillonitic 40 268 0.05Ustox Kaolinitic 64 155 0.05Andept Allophanic 46 214 0.31

Source: Sanchez (1976). Reproduced by permission of Wiley, New York.

Page 29: The Biogeochemistry of Submerged Soils

20 Transport Processes in Submerged Soils

Table 2.2 Effect of cultivation at different soil-water states on components of percentageporosity in a Vertic Tropaquept clay soil

Sampling state Total porosity >50 µm 0.5–50 µm <0.5 µm

Before cultivation 66 26 6 31After moist

cultivation61 5 15 41

After saturatedcultivation

62 3 16 43

SE 1 1 1 1

Source: Reprinted from Painuli et al. (1988) with permission from Elsevier Science.

under laboratory conditions. Puddling decreases percolation rates by up to threeorders of magnitude. Among the soils in the table, the sandy Entisol and clayeyAndept were difficult to puddle and consequently had greater percolation rates.However, the low percolation rates in the other soils are comparable to rates inmost rice soils under field conditions. Puddling generally leads to an increasein total porosity because the destruction of aggregates decreases intra-aggregatepores but increases inter-aggregate and inter-domain pores, as shown in Table 2.2.

The percolation rate also depends on the depth of water standing on the soilsurface. Consider the submerged soil shown in Figure 2.1. The soil overlies acompacted traffic layer caused by repeated working for wetland rice cultivation,beneath which the soil may be saturated or unsaturated depending on its propertiesand the depth of the ground water. The flux of water through the soil is relatedto the gradient in water potential by Darcy’s law:

v = −Kdψ

dz(2.8)

where v is the flux in direction z, K the hydraulic conductivity and ψ the waterpotential. If v and K are constant in a given layer in Figure 2.1, as they generally

ψ0

floodwater

puddled layer

compacted layer

z1

z2

z3

Figure 2.1 Changes in water potential with depth in a puddled flooded soil

Page 30: The Biogeochemistry of Submerged Soils

Mass Flow 21

will be, then

v = −K�ψ

�z(2.9)

and for steady-state flow of water through the soil,

v = −Kpψ2 − ψ1

z2 − z1= −Kc

ψ3 − ψ2

z3 − z2(2.10)

where subscripts 1, 2 and 3 refer to the indicated depths in Figure 2.1 and Kp

and Kc are the conductivities of the puddled and compacted layers, respectively.Rearranging Equation (2.10) and substituting Lp and Lc for the depths of thepuddled and compacted layers gives

v = ψ1 − ψ3

Lp/Kp + Lc/Kc(2.11)

Equation (2.11) shows how the flow increases with increasing depth of the flood-water and decreases with increasing impermeability of the compacted layer.

The effect of percolation on transport of solutes through the soil is quantifiedas follows. If there is a concentration gradient of a solute through the soil, fromEquation (2.4) the net flux due to mass flow and diffusion is

F = −DdC

dz+ vC (2.12)

Mass flow and diffusion act together and cannot be separated. However an ideaof their relative contributions to the net flux can be obtained by estimating thedistance the solute would be transported if each process acted independently. Ifin time t mass flow transports the solute a distance

z1 = vt (2.13)

the mean distance moved by diffusion would be

z2 = √Dt (2.14)

and the ratio of the two would be

z1

z2= v

√t

D(2.15)

This equation indicates that, for a constant flow rate and diffusion coefficient, thedistance transported by mass flow will exceed that by diffusion after a certain timehas elapsed, i.e. mass flow eventually becomes more important than diffusion.However note that z2 is only the mean distance moved; some of the solute willhave diffused beyond this.

Equation (2.15) can be used to calculate the relative importance of mass flowand diffusion under conditions in ricefields. From the discussion above, rates of

Page 31: The Biogeochemistry of Submerged Soils

22 Transport Processes in Submerged Soils

percolation in puddled submerged soils are generally less than a few mm day−1,that is ≤ 2 × 10−7 dm s−1. Solute diffusion coefficients in submerged soils areof the order of 5 × 10−8 dm2 s−1 (next section). Therefore the time taken for thedistance moved by mass flow to exceed the mean distance moved by diffusionwould be 14 days.

2.2 DIFFUSION

The rates of many important processes in submerged soils are governed by ratesof diffusion. A comprehensive theory of diffusion in soils exists, allowing thedevelopment of mechanistic models of soil processes involving diffusion. I brieflydescribe this theory in this section; more complete treatments are given in Nye(1979) and Tinker and Nye (2000).

Diffusion results from the random thermal motion of particles. If there is aconcentration gradient of a substance through a medium in which it is mobile,the net amount of substance crossing a unit section in unit time is given by Fick’sfirst law (Equation 2.3):

F = −D

(dC

dx

)(2.16)

where F is the flux of the substance, dC/dx is its concentration gradient acrossthe section, and D is the diffusion coefficient, which is defined by this relation.

2.2.1 DIFFUSION COEFFICIENTS IN SOIL

In submerged soils there is no continuous gas phase through which volatile solutescan diffuse. Hence we are mainly concerned with the liquid and solid phases.If the concentration of a volatile solute in the liquid becomes sufficiently largefor an appreciable amount to come out of solution, the resulting gas bubbleswill rise through the soil by the process of ebullition, partly becoming entrappedbeneath soil particles until they are dislodged by mechanical forces. Ebullition isdiscussed in Section 2.3. For diffusion in the liquid and solid phases, the sameprinciples apply as for non-submerged soils though there are some additionaleffects, which I shall describe.

Most solutes in soils are to some extent adsorbed on the soil solid; only asmall fraction is in the solution in the pores. However some adsorbed solutes,particularly exchangeable cations, can have considerable mobility on soil surfaces(see below), so it is important to consider the solid phase pathway as well asthe solution. Because the diffusing solute passes rapidly between the solid andsolution, the two pathways partly act in series. In such a heterogeneous mediumas soil it is not realistic to account for the mobilities and concentration gradientsof solutes in all the constituent parts. But if the soil volumes and reaction times

Page 32: The Biogeochemistry of Submerged Soils

Diffusion 23

considered are large enough to average microscale variations, the soil can betreated as quasi-homogeneous and Fick’s first law can be applied to the systemas a whole. The term C in Equation (2.16) is then the concentration of thediffusate in the whole soil system; that is, ‘all those ions or molecules that are inor pass through a mobile phase during a period that is short in comparison withthe time of the diffusion process’ (Nye, 1979). Solutes that do not interchangecompletely between the solid and solution within this time frame, i.e. a matterof hours, are treated as having a rate of reaction and are dealt with by adding asource or sink term to the appropriate form of the continuity equation.

Following from this definition, the diffusive flux of a solute through the solutionand solid in the x direction is given by (Tinker and Nye, 2000, Equation 4.17)

F = −DLfLθLdCL

dx− DLfSθS

dCS

dx(2.17)

where DL is the diffusion coefficient of the solute in free solution, θL is thefraction of the soil volume occupied by solution, θS is the fraction of the soilvolume occupied by soil solid, fL and fS are the impedance factors for the liquidand solid phase, respectively, and CL and CS are the amounts of solute per unitvolume of liquid and solid phase, respectively.

The first term on the right-hand side of Equation (2.17) represents diffusionexclusively in solution; the second term represents the additional diffusion thatoccurs as a result of mobility on the soil solid.

The concentration of the solute in the solid is expressed in terms of the amountper unit weight of solid, SS, by θSCS = ρSS where ρ is the weight of dry solidper unit soil volume. By definition, F = −DdC/dx. Substituting for F and θSCS

in Equation (2.17) and rearranging gives

D = DL

(fLθL + fSρ

dSS

dCL

)dCL

dC(2.18)

In the following sections I discuss the components of the diffusion coefficient sodefined in turn. Note all the components of D are altered by flooding the soil.As well as increasing the cross-sectional area for diffusion, represented by θL,flooding affects the geometry and tortuosity of the soil pore network, representedby fL and fS, and solute sorption on the soil solid, represented by dCL/dC.

The Diffusion Coefficient in Free Solution, DL

Table 2.3 gives the self-diffusion coefficients of some important ions in sub-merged soils and Figure 2.2 shows the values for the elemental ions plottedagainst ionic potential (|z|/r where |z| is the absolute ionic charge and r thecrystal ionic radius). As the ionic potential increases the hydration layer of watermolecules around the ion increases, and therefore the mobility tends to decrease.Also, at the same ionic potential, cations diffuse faster than anions. The mobilities

Page 33: The Biogeochemistry of Submerged Soils

24 Transport Processes in Submerged Soils

Table 2.3 Self-diffusion coefficients of ions in aqueous solution at 25 ◦Ca

Cations D0

(dm2 s−1 × 10−7)Anions D0

(dm2 s−1 × 10−7)

H3O+ 9.31 OH− 5.27Li+ 1.03 F− 1.47Na+ 1.33 Cl− 2.03K+ 1.96 Br− 2.08Rb+ 2.06 I− 2.04Cs+ 2.07 HS− 1.73NH4

+ 1.98 SO42− 1.06

Mg2+ 0.70 NO2− 1.91

Ca2+ 0.79 NO3− 1.90

Sr2+ 0.79 HCO3− 1.18

Ba2+ 0.85 CO32− 0.92

Mn2+ 0.69 H2PO4− 0.85

Fe2+ 0.72 HPO42− 0.73

Ni2+ 0.68 H2AsO4− 0.91

Cu2+ 0.73 HCOO−b 1.45Zn2+ 0.72 CH3COO−b 1.09Cd2+ 0.72 CH3CH2COO−b 0.95Pb2+ 0.95

a Calculated with the relation D0 = RT λ0

|z|F 2using values of the limiting equivalent conductivity, λ0,

from Landolt et al. (1960).b Robinson and Stokes (1959).

|z |/r (nm−1)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

D0

(dm

2 s−1

× 1

0−7)

0.0

0.5

1.0

1.5

2.0

2.5

Ba2+

Pb2+

Sr2+

Fe2+, Zn2+

Cd2+

Ca2+

Mn2+Cu2+ Mg2+

Ni2+

Cs+Rb+

K+NH4

+

Na+

Li+

Br−

I−Cl−

F−

Figure 2.2 Self-diffusion coefficients at 25 ◦C plotted against ionic potential (after Liand Gregory, 1974). Reprinted with permission from Elsevier Science

Page 34: The Biogeochemistry of Submerged Soils

Diffusion 25

of H3O+ and OH− are anomalously large because they move by a proton jumpmechanism in which protons are passed between favourably orientated watermolecules (Glasstone et al., 1941).

However, in bulk diffusion, ions cannot move independently of each otherbecause electrical neutrality must be maintained. Consequently there is an electricpotential between diffusing ions such that the faster ions tend to be slowed downby the slower ones and vice versa. The flux of a particular ion is therefore thesum of the diffusion due to its own concentration gradient and that due to thegradient of the diffusion potential arising from differences in the mobilities of theions present. This is expressed by the Nernst-Planck equation along the x-axis:

FLA = −DLA

(dCLA

dx+ ZACLAF ′

RT

dx

)(2.19)

where F ′ is the Faraday, ZA is the charge of ion A and ψ is the potential. If Ais present in only small concentrations, the diffusion potential term is much lessimportant than the concentration gradient term, and can be ignored. However, ifA is a large part of the total ionic strength, and ions are present with differingmobilities, the diffusion potential will be important. Vinograd and McBain (1941)used the condition of no net flux of charge:

�ZiFLi = 0 (2.20)

where subscript i refers to a particular species and Zi is its charge, to expressthe term dψ/dx in terms of the ionic concentrations gradients, giving

FLA = −DLAdCLA

dx+ ZACLADLA

�DLidCLi/dx

�Zi2DLiCLi

(2.21)

Equation (2.21) shows that the greater ZA, CLA and DLA, and the smaller dCLA/dx,the greater will be the effect. But the effect is small for ions with similar mobili-ties and for ions whose concentrations are small compared with the total solutionconcentration.

A further point is that a significant proportion of many of the cations in solu-tion in submerged soils may be complexed with organic ligands (Chapter 3).The diffusion coefficients of the complexed ions will be smaller than the corre-sponding free ions. Table 2.4 compares self-diffusion coefficients of chelated andunchelated Fe3+ and Zn2+. Table 2.5 gives the diffusion coefficients in aqueoussolution of other uncharged solutes important in submerged soils, and diffusioncoefficients in air.

The Soil Moisture Content, θL and Bulk Density, ρ

In submerged soils there tends to be a gradient of bulk density with depth asa result of the settling of disturbed sediment. As a result, the bulk density is

Page 35: The Biogeochemistry of Submerged Soils

26 Transport Processes in Submerged Soils

Table 2.4 Diffusion coefficientsof chelated ions at 25 ◦C

D(dm2 s−1)

Zn2+ 8.8 × 10−8 a

ZnEDTA 6.8 × 10−8 a

Fe3+ 6.2 × 10−8 b

FeEDTA 3.9 × 10−8 b

FeDTPA 4.2 × 10−8 b

FeEDDHA 5.4 × 10−8 b

Source:a Elgawhary et al. (1970).b O’Connor et al. (1971).

Table 2.5 Diffusion coefficients of gasesin air and water at 25 ◦C and 1 atm

D(dm2 s−1)

in air in water

O2 2.05 × 10−3 2.26 × 10−7

CO2 1.55 × 10−3 1.93 × 10−7

CH4 2.20 × 10−3 1.73 × 10−7

H2S 1.66 × 10−3 2.00 × 10−7

N2 2.04 × 10−3 2.02 × 10−7

NH3 2.19 × 10−3 2.49 × 10−7

N2O 1.55 × 10−3 1.98 × 10−7

NO 2.04 × 10−3 2.55 × 10−7

Source: Lerman (1979).

near zero at the boundary between the soil and overlying water and graduallyincreases over the upper 0.5 to 1 cm of soil or deeper. The cross-sectional areafor diffusion and ion interchange with the soil solid are altered correspondingly.

The Impedance Factor for the Liquid Phase, fL

The impedance factor is strictly empirical, accounting primarily for the geometryof the soil pore network but also for ion exclusion by negative adsorption fromnarrow pores, and for the increased viscosity of water near charged surfaces. Itis similar for all simple ions and molecules. It can be measured by following theself diffusion of a nonadsorbed ion, such as Cl−, for which C = θLCL and henceD = DLfL.

Discontinuities in the liquid pathway and the effects of anion exclusion fromnarrow pores and of viscosity near charged surfaces are important at low moisturecontents. The value of fL therefore approaches zero in very dry soil and the

Page 36: The Biogeochemistry of Submerged Soils

Diffusion 27

0.0 0.1 0.2 0.3 0.4 0.5 0.60.0

0.1

0.2

0.3

0.4

0.5

0.6

5

15

30a

2423

5316

37

26

4

19336

38

17b

Volumetric water content, qL

Impe

danc

e fa

ctor

, fL

Figure 2.3 Relation between diffusion impedance factor, fL, and moisture content, θL,in a range of soils. Numbers shown are % clay contents. aMean of six soils; bvolcanicash soil. (After Tinker and Nye, 2000; Olesen et al., 2001). Reproduced by permission ofOxford University Press

relation between fL and θL is curvilinear. Figure 2.3 shows relations between fL

and θL in a range of soils given by Nye (1979); Olesen et al. (2001) give furthervalues. The figure shows that, at a given moisture content, fL is smaller in claysoils than in sandy soils, probably because a greater proportion of the soil wateris in fine pores. But at a given water potential, fL is larger in clayey than sandysoils because they hold more water (So and Nye, 1989).

Effects of Flooding and Redox Conditions on fL. As well as increasing the cross-sectional area for diffusion through the soil pores, flooding affects fL becausethe anoxic conditions that develop result in dissolution and re-precipitation ofthe soil solid and changes in its electrical properties. Kirk et al. (2003) investi-gated these effects in four soils with contrasting properties. Figure 2.4 shows therelation between fL and bulk density in the soils under water-saturated condi-tions. The relation is linear with similar slopes but different intercepts in the foursoils. As bulk density increases, porosity decreases, and the pathway for diffu-sion becomes more tortuous. The dotted line in Figure 2.4 shows the theoreticalrelation between fL and ρ for a mixture of spherical particles of various sizes:fL = θL

0.5 = (1 − ρ/ρP)0.5 where ρP is the particle density, taken as 2.65 g cm−3

(Nye, 1979, Section V.B). The values of fL for the more coarse-textured soil,Iloilo, come closest to this line, but the values are progressively far from it forthe more clayey soils, and they are not parallel to it in any of the soils. There areseveral reasons for this. In soils electrostatic and viscosity interactions betweendiffusing solutes and solid surfaces are important and tend to diminish fL at a

Page 37: The Biogeochemistry of Submerged Soils

28 Transport Processes in Submerged Soils

Bulk density, r (g cm−3)

0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5

Impe

danc

e fa

ctor

, fL

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

IloiloMaahasN. Ecija Tarlac

Figure 2.4 Relation between diffusion impedance factor, fL, and bulk density, ρ, infour water-saturated rice soils. Dotted line is the theoretical relation between fL and ρ fora mixture of different-sized spherical particles (Kirk et al., 2003). Iloilo: Epiaquult; clay21 %; org C 1.04 %; pH 3.93. Maahas: Haplaquoll; clay 54 %; org C 1.83 %; pH 5.89.Nueva Ecija: Epiaquert; clay 35 %; org C 1.57 %; pH 5.25. Tarlac: Tropaquept; clay 33 %;org C 1.06 %; pH 6.02. Reproduced by permission of Blackwell Publishing

given ρ. These interactions increase as ρ increases and an increasing proportionof the pores are fine, and as clay content increases. Hence the regression lines aresteeper than the theoretical line and farther from it with increasing clay content.Also the approximation of spherical shape is increasingly invalid with increasingclay content because small surface irregularities become an increasing proportionof the whole.

Figure 2.5 shows the effects of changes in redox conditions on fL in the foursoils with time following flooding. In three of the soils, fL decreases during thefirst few weeks following flooding but then gradually returns to its initial value.Since ρ and θL were constant over this period, the changes in fL were evidentlydue to changes in chemical and biological conditions in the soils following flood-ing. This is explained as follows.

Following flooding O2 entrapped in the soil is rapidly consumed in aerobicmicrobial respiration, and then other inorganic electron acceptors are used inmicrobial respiration in the sequence NO3

−, Mn(IV), Fe(III), SO42− (Chapter 4).

Concomitantly, organic matter is oxidized, dissolved CO2 accumulates, and thepH of acid soils tends to increase and that of alkaline soils to decrease, stabilizingin the range 6.5 to 7. Figure 2.6 shows the changes in EH, pH, the concentrationof HCO3

− in solution and cation exchange capacity (CEC) in the four soilsfollowing flooding, and Figure 2.7 shows the corresponding changes in soil Fe.In all four soils, Fe(III) is the main inorganic reductant and large concentrationsof Fe2+ develop in the soil solution in the weeks following flooding as Fe(III) is

Page 38: The Biogeochemistry of Submerged Soils

Diffusion 29

0.4

0.5

0.6

0.7

0.4

0.5

0.6

0.7

0.4

0.5

0.6

0.7

0.4

0.5

0.6

0.7

Iloilo (r = 1.21 ± 0.01)

Nueva Ecija (r = 1.00 ± 0.02)

Impe

danc

e fa

ctor

, fL

Maahas (r = 0.80 ± 0.01)

Time after flooding (days)

0 20 40 60 80

Tarlac (r = 0.98 ± 0.02)

Figure 2.5 Changes in fL with time following submergence of the four soils in Figure2.4 (Kirk et al., 2003). Reproduced by permission of Blackwell Publishing

Page 39: The Biogeochemistry of Submerged Soils

30 Transport Processes in Submerged Soils

−300

−200

−100

0

100

200

300

Soi

l pH

4

5

6

7

[HC

O3− ]

in s

olut

ion

(mM

)

0

5

10

15

20

25

Time after flooding (days)

0 20 40 60 80

CE

C (

mm

olc

kg−1

)

0

50

100

150

200

250

Iloilo

Tarlac

MaahasNueva Ecija

Iloilo

Tarlac

Maahas

Nueva Ecija

Iloilo

Maahas

TarlacNueva Ecija

Maahas

Nueva Ecija

Iloilo

TarlacS

oil E

H (

mV

)

Figure 2.6 Changes in EH, pH, HCO3− and CEC with time in the experiment in

Figure 2.5 (Kirk et al., 2003). Reproduced by permission of Blackwell Publishing

Page 40: The Biogeochemistry of Submerged Soils

Diffusion 31

[Fe2+

] in

solu

tion

(mM

)

0.0

0.1

0.2

0.3

0.4

0 80

[Fe2

+ ] /

mM

0

1

2

3

4

Iloilo

Tarlac

Nueva Ecija

Maahas

NH

4OA

c-ex

trac

tabl

e F

e(II)

(mm

ol k

g−1)

0

1

2

3

4

5

6

Iloilo Maahas

Nueva EcijaTarlac

Fer

rozi

ne-e

xtra

ctab

le F

e(II)

(mm

ol k

g−1)

0

2

4

6

8

Iloilo

Maahas

Nueva Ecija

Tarlac

Time after flooding (days)

0 20 40 60 80

HC

l-ext

ract

able

Fe(

II)(m

mol

kg−1

)

0

20

40

60

80

100

Iloilo

MaahasNueva Ecija

Tarlac

Figure 2.7 Changes in soil Fe(II) with time in the experiment in Figure 2.5 (Kirk et al.,2003). Reproduced by permission of Blackwell Publishing

Page 41: The Biogeochemistry of Submerged Soils

32 Transport Processes in Submerged Soils

reduced and dissolves. The increases and subsequent decreases in Fe2+ in solutioncoincide with the increases and decreases in HCO3

−, suggesting that insolubleferrous carbonates are formed. The ion activity products of well-known ferrouscarbonates and hydroxides are exceeded up to 10-fold in the four soils in theearly stages following flooding. Evidently precipitation of these compounds isinhibited, probably as a result of adsorption of foreign solutes, such as dissolvedorganic matter, on nucleation sites. However, once a sufficient supersaturationhas been reached, amorphous solid phases are precipitated, and these later re-order to more crystalline forms. The changes in Fe(II) in the solid phase areconsistent with this, the more soluble pools showing peaks roughly matchingFe2+ in solution but the HCl-extractable Fe(II) continuing to increase over time.

Mixed Fe(III)–Fe(II) compounds form initially upon flooding with a progres-sively greater proportion of Fe(II) as reduction proceeds. In the Maahas, NuevaEcija and Tarlac soils, which contain 2:1 clays, some of the Fe is structural in claylattices. Reduction of this structural Fe causes a net increase in the negative sur-face charge on the clay, resulting in increased CEC and decreased clay swellingand surface area (Stucki et al., 1997). Further, in soils that undergo intermittentreduction and oxidation, as all the soils here do, a large part of the easily reducibleFe is present as coatings of oxyhydroxides on clay surfaces (Brinkman, 1985),and these are dissolved during soil reduction. Where positively charged oxyhy-droxides neutralize negatively charged sites on the clay, dissolution of the coat-ings will cause the net surface negative charge and hence CEC to increase (Rothet al., 1969). The rapid increase in CEC in the early stages of reduction and itssubsequent stabilization (2.7) are consistent with the changes in pH and Fe.

The changes in fL in the four soils following flooding roughly parallel thechanges in CEC and Fe. Increased CEC will cause increased anion exclusionfrom narrow pores, decreasing fL. Decreased clay swelling and interlayer spacewith reduction of structural Fe will also increase anion exclusion and exacerbatethe decrease in fL. There will also be changes in pore geometry with dissolutionof ferric oxyhydroxides coatings, but these will be complicated by subsequent re-precipitation in initially amorphous but later crystalline forms. Poorly crystallinecompounds lining soil pores should increase the tortuosity of the diffusion path-way, but as they revert to more crystalline forms with smaller specific surfaces,fL should increase. Although the changes in diffusion impedance due to changesin redox are smaller than the changes due to water content per se, they will beimportant in some soils.

The Impedance Factor for the Solid Phase, fS

In Equation (2.18) fS is defined in relation to DL: it takes account of all factorsdecreasing the mobility of the sorbed solute from the mobility it would havein free solution. This includes the binding of the solute to the surface and thelimited thickness of the layer of water molecules associated with the surface,

Page 42: The Biogeochemistry of Submerged Soils

Diffusion 33

in which the sorbed solute diffuses. It can be measured by assuming that thesolution pathway is the same as that for a nonabsorbed ion, and then deducingthe movement in the solid from the excess movement not accounted for by thesolution pathway.

While fS is negligible for strongly sorbed anions, such as phosphate, whichare covalently bound to the surface, it is often substantial for exchangeablecations (Staunton and Nye, 1983, 1987; Staunton, 1990; Nye and Staunton, 1994).Although the values of fS are much smaller than those of fL, the flux on the sur-face may exceed that in solution because a much larger proportion of the cationsare held on the surface. Staunton (1990) measured fS for different exchangeablecations (Na+, Ca2+, Rb+ and Cs+) in soils with a range of clay contents andmineralogies, and found that fS decreased as the degree of binding to the sur-face, measured by the ratio SS/CL, increased. Thus the two tend to offset eachother and as a result the contribution of surface diffusion to D is variable. Forthe cations and soils considered by Staunton the contribution of surface diffusionto D ranged from 27 to 97 %. There was no correlation between fS and soilmineralogy or clay content. Staunton and Nye studied alkali and alkaline earthcations sorbed in freely exchangeable forms, over reaction times of a few hours.More strongly sorbed cations, such as those of transition and heavy metals attrace concentrations, are likely to have less surface mobility.

A consequence of high surface mobility is that equilibrium between inter- andintra-aggregate pores is maintained more easily. Thus Pinner and Nye (1982), forCl−, and Staunton and Nye (1983, 1987), for exchangeable cations, found no evi-dence of slow equilibration between inter- and intra-aggregate pores for diffusionin naturally structured soils, and the soil pore network behaved homogeneouslyto the diffusant. However for solutes with little surface mobility, such as stronglysorbed anions, access to exchange sites through narrow access pores is likely tolimit equilibration between inter- and intra-aggregate pores. It is important toconsider this in defining continuity equations and in measuring rates or reaction.

Effects of Flooding and Redox Conditions on fS. I know of no published data onthis. But it is likely that the nature of particle surfaces in intermittently floodedsoils would restrict surface mobility. For ions to diffuse freely on the surfacethere must be a continuous pathway of water molecules over the surface anduniform cation adsorption sites. But in intermittently flooded soils the surfacetypically contains discontinuous coatings of amorphous iron oxides on other clayminerals, and on flooding reduced iron is to a large extent re-precipitated asamorphous hydroxides and carbonates as discussed above, resulting in muchmicroheterogeneity with adsorption sites with disparate cation affinities.

The Derivative dCL/dC

This derivative is the reciprocal of the buffer power and describes the distributionof the diffusing solute between the soil solid and solution. Its value varies by

Page 43: The Biogeochemistry of Submerged Soils

34 Transport Processes in Submerged Soils

several orders of magnitude for different solutes in a given soil and to a lesserextent for the same solute in different soils. For nonadsorbed solutes, such asthe Cl− ion, C = θLCL and therefore θLdCL/dC = 1. Hence in Equation (2.18),D = DLfL. For a strongly sorbed ion, such as H2PO4

−, dCL/dC may be 1000and the value of D correspondingly small. Values of dCL/dC are also sensitiveto the method by which they are measured. This must therefore be as close aspossible to the conditions under which the diffusion coefficient is to be applied.The difficulties in measuring dCL/dC correctly are discussed by Tinker and Nye(2000, pp. 84–88).

Effects of Flooding and Redox Conditions on dCL/dC . Reductive dissolutionreactions of the sort indicated in Figures 2.6 and 2.7 will affect the amount of asolute in diffusible forms in the soil and the distribution of the diffusible formsbetween the soil solid and solution. These processes are discussed in detail inChapter 3. I here exemplify their effects by reference to a study of phosphatediffusion in a soil under different water regimes.

Huguenin-Elie et al. (2003) measured the diffusion of P to a resin sink placedin contact with a soil that was either moist, flooded or flooded then moist, andderived values of the diffusion coefficient of P in the soil by fitting to the resultsthe equation for diffusion from a semi-infinite medium to a planar sink:

Mt = 2C∞

√DPt

π(2.22)

where Mt is the amount of P accumulated in the resin at time t, C∞ is theconcentration of diffusible P in the soil bulk, measured independently, and DP

is the diffusion coefficient of P. The results are shown in Figure 2.8. They then

Time (days)

0 42 6 8 10

P a

ccum

ulat

ed in

resi

n, M

t(µ

mol

dm

−2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Mt = 14.96 × 10−10 × √t

Mt = 4.93 × 10−10 × √t

Mt = 4.60 × 10−10 × √t

Figure 2.8 Amounts of P absorbed by a planar resin sink in contact with columns of soilthat was flooded (circles), moist (triangles) or flooded then moist (squares). Lines are fitsto Equation (2.22) (Huguenin-Elie et al., 2003). Reproduced by permission of BlackwellPublishing

Page 44: The Biogeochemistry of Submerged Soils

Diffusion 35

calculated values of dCL/dC from the values of DP so obtained using measuredvalues of θL and fL and Equation (2.18). Flooding increased θLfL from 0.12 to0.21, C∞ from 199 to 352 µmol dm−3 (whole soil), and dCL/dC from 1/2220to 1/1330. Drying previously flooded soil to the original θLfL value decreasedC∞ to 149 µmol dm−3 and dCL/dC to 1/1430. The changes were consistent withchanges in soil Fe upon flooding and drying.

Equations for Sorption. The following two simple equations often adequatelydescribe the relation between the amount of an anion or a cation sorbed on thesoil solid and its concentration in solution:Freundlich equation

CS = a CLb (2.23)

Tempkin equationCS = a ln CL (2.24)

where CS is the concentration in the solid, CL the concentration in solution,and a and b are coefficients (b < 1). Log-log transformations of the Freundlichequation are linear and linear-log transformations of the Tempkin are linear.These equations correspond to mechanisms by which sorption is progressivelyinhibited by the accumulation of sorbed solute on the solid. However, as appliedto heterogeneous soil systems they are empirical and more precise mechanismsthan this cannot be inferred.

2.2.2 PROPAGATION OF pH CHANGES THROUGH SOIL

An important application of the theory of solute diffusion in soil is Nye’s (1972)theory of how pH changes are propagated by acid–base transfer. Sources of pHchanges in submerged soils are legion and the resulting pH gradients through thesoil have important effects on soil processes. Examples are diurnal changes infloodwater pH caused by algae and pH changes in the rhizosphere induced byroots. I here give a summary of the theory.

Changes in pH are propagated by the movement of protons. But because freeprotons do not exist they must move by transfer between proton donors andacceptors, i.e. Bronsted acids and bases:

acid (proton donor) = base (proton acceptor) + H+ (2.25)

The main acid–base pairs in soils are H3O+ –HCO3− and H2CO3 –HCO3

−. Inparticular cases other pairs, such as NH4

+ –NH3, H2PO4− –HPO4

2− and H2S–HS−,may also be important. When a pH gradient exists in a soil, acids and bases willmove in the soil solution between the solid surfaces: acids from zones of low pHto high and bases in the opposite direction. The acid arriving in a portion of soilwill react with it, and a local acid–base equilibrium will be established. Thus the

Page 45: The Biogeochemistry of Submerged Soils

36 Transport Processes in Submerged Soils

acidity of the solution is buffered. Electrical neutrality is maintained by the co- andcounter-diffusion of the other cations and anions present. Nye derives equations ofthe transfer of protons through the soil as follows.

Consider the balance of acid, HS, between two imaginary planes of unit cross-section at x and x + δx in the soil. The general reaction of the acids and baseswith the soil may be written

HB + MS = HS + M+ + B− (2.26)

where HB is an acid and B− its conjugate base. In unit time the increase in HSwill be the result of all reactions of this type and will equal the loss by reactionof acids HB within the region. This loss must equal the sum of the fluxes ofall acids entering across the plane at x less those leaving across the plane atx + δx. If M+ and H+ do not move in the solid phase, then by analogy withEquations (2.5) and (2.18),

∂[HS]

∂t= ∂

∂x

(θLfL�DLHB

∂[HB]

∂x

)(2.27)

where DLHB is the diffusion coefficient of HB in free solution, [HB] is the con-centration of acid in solution and [HS] is the concentration of acid soil, i.e. ofproton donating groups as measured by the amount of strong base consumed byunit volume of soil in raising the equilibrium pH to a standard pH. (Since onlydifferences in [HS] arise it is not necessary to define this pH.)

It is assumed that the impedance factor, fL, is the same for all mobile acidsand bases. It is also assumed that the solid equilibrates rapidly with the adjacentsolution; in cases where it does not, terms for the rates of reaction can be addedto the equation.

The term ∂[HB]/∂x in Equation (2.27) is expressed in terms of ∂[HS]/∂x asfollows. The pH buffer power of each acid–base pair is defined as:

bHB = d[B−]

d pH(2.28)

However, because in general there is no net flux of component B, d[B−] =−d[HB] and

bHB = −d[HB]

d pH(2.29)

Likewise the pH buffer power of the soil is

bHS = −d[HS]

d pH(2.30)

bHS is often fairly constant over a wide range of pH.

Page 46: The Biogeochemistry of Submerged Soils

Diffusion 37

Since [HB] and [HS] are both functions of pH, and pH is a function of x,

∂[HB]

∂x= ∂[HS]

∂x

d[HB]

d pH

d pH

d[HS]= bHB

bHS

∂[HS]

∂x

Hence substituting in Equation (2.27),

∂[HS]

∂t= ∂

∂x

{(θLfL

bHS�bHBDLHB

)∂[HS]

∂x

}(2.31)

where the term in parentheses is the soil acidity diffusion coefficient, DHS. If bHS

is constant, then by substituting for d[HS] from Equation (2.30), Equation (2.31)may be written

∂pH

∂t= ∂

∂x

(DHS

∂pH

∂x

)(2.32)

For a soil in which the only important acid–base pairs are H3O+ –H2O andH2CO3 –HCO3

−, Nye (1972) shows that:

DHS = 2.303θLfL

bHS(DLH[H3O+] + DLC[HCO3

−]) (2.33)

The relative contribution of the pairs H3O+ –H2O and H2CO3 –HCO3− to the

overall soil acidity diffusion coefficient is given by the term in parentheses inEquation (2.33) and is plotted at different pHs in Figure 2.9(a). The figures shows

pH

4 5 6 7 8

DLq

LfLb

HB

/bH

S (

dm2

s−1 ×

10−9

)

0.0

0.2

0.4

0.6

0.8

1.0H3O+---H2O H2CO3---HCO3

PCO2 =

1 kPa

0.03

0.1

pH

3 4 5 6 7 8

DH

S (

dm2

s−1 ×

10−9

)

0.0

0.2

0.4

0.6

0.8

1.0(a) (b)

PCO2 =

0.05 kPa

Figure 2.9 (a) Contributions of acid–base pairs H3O+ –H2O and H2CO3 –HCO3− to the

soil acidity diffusion coefficient over a range of pH; θLfL = 0.3, bHS = 0.05 mol dm−3 pH−1

(after Nye, 1972). (b) Observed and calculated soil acidity diffusion coefficients (Nye andAmeloko, 1986). Reproduced by permission of Blackwell Publishing

Page 47: The Biogeochemistry of Submerged Soils

38 Transport Processes in Submerged Soils

that the soil acidity diffusion coefficient passes through a minimum in the pHrange in which H3O+ and HCO3

− are both low: in this pH range a flux of acidor base through the soil results in steep pH gradients. This has been corroboratedexperimentally over a wide range of soil pHs by Nye and Ameloko (1986), asshown in Figure 2.9(b). The data in Figure 2.9(b) were obtained from profiles ofpH measured in two blocks of soil of different initial pHs placed in contact.

2.3 EBULLITION

Ebullition is the process by which gas bubbles form from volatile solutes insolution and rise to the surface and atmosphere. Bubbles form spontaneouslywhen a solution becomes supersaturated with a volatile solute. Rates of formationof bubbles and ebullition depend on the volatility of the particular solute as well asits concentration in solution. In a soil producing methane, for example, althoughCH4 and CO2 may be generated in equal proportions (Chapter 5), gas bubbleswill contain a large excess of CH4 over CO2 because CH4 is about 20 timesmore volatile than CO2.

Bubbles form when the sum of the partial pressures of the volatile solutes exceedsthe hydrostatic pressure. For a water column containing dissolved N2, CO2 andCH4, the condition for formation of a bubble is therefore (Morel and Herring, 1993,Equation 142)

PN2 + PCO2 + PCH4 + PH2O > Pz (2.34)

where Pz is the hydrostatic pressure at depth z (= Patm + ρgz). This inequalitymay be met either because of accumulation of dissolved gases or because ofchanges in pressure, as for example when a core of mud is brought up from ananaerobic sediment.

In flooded soil or sediment, bubbles form through heterogeneous nucleationat the surface of solid particles, rather than by homogeneous nucleation in freesolution. Because of this, bubbles form easily and the sum of the partial pressuresof volatile solutes tends to be maintained at or near the hydrostatic pressure.Therefore, for a methanogenic sediment,

PN2 + PCO2 + PCH4 + PH2O = Patm + ρgz (2.35)

This equation can be used to calculate the composition of bubbles and rates ofebullition from rates of gas formation and the volatility of the different species.

Thus for a methanogenic sediment in which rates of CH4 and CO2 genera-tion are balanced by their rates of loss to the atmosphere above by diffusionand ebullition, we have for each volatile solute (cf. Morel and Herring, 1993,Equations 144–146)

D

(CZ − C0

Z

)+ ε

(CZ/KH

Patm + ρgZ

)− R = 0 (2.36)

Page 48: The Biogeochemistry of Submerged Soils

Mixing by Soil Animals 39

0.0 0.3 0.6 0.9

0

10

20

30

40

50

Concentration of dissolved gas in water (mM)0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 0.0 0.3 0.6

CH4 CO2 N2

FE = 2.9 FD = 2.1 FE = 0.3 FD = 4.7 FE = 0.8 FD = −0.8

Dep

th in

ove

rlyin

g w

ater

(cm

)

Figure 2.10 Concentrations and fluxes of CH4, CO2 and N2 in anoxic acidic marsh(after Morel and Herring, 1993). FE and FD are the fluxes by ebullition and diffusion,respectively Reproduced by permission of Wiley, New York

where Z is the depth of overlying water, D is the diffusion coefficient of thesolute in water, CZ and C0 are the concentration of dissolved solute at sedi-ment surface and water surface, respectively, ε is the rate of ebullition of allgases together, KH is Henry’s law constant and R is the rate of generation of thesolute in the sediment.

An equation of this type can be written for N2, CH4 and CO2 and combinedwith Equation (2.35) and the resulting equation solved to obtain the rates ofebullition and the concentrations of each gas at the sediment surface given theambient atmospheric concentrations, the rate of methanogensis and the depth ofthe water.

Figure 2.10 compares the relative contributions of ebullition and diffusion tofluxes of CH4, CO2 and N2 in an anoxic marsh so calculated. The figure showsthat CO2 escapes mainly by diffusion whereas more than half the CH4 escapes byebullition. The bubbles contain 69 % CH4, 19 % N2, 5 % H2O and only 7 % CO2.

In practice gas bubbles may become entrapped under irregularly shaped soilparticles, and so the simple steady state described by Equation (2.36) does nothold. The rate of ebullition is then sensitive to mechanical disturbances, inducedfor example by wading animals or by the action of wind on plants in the sediment.This is discussed further in Chapter 8.

2.4 MIXING BY SOIL ANIMALS

The upper few centimetres of the soil are subject to mixing by invertebrates bur-rowing through the soil and ingesting soil particles. If populations are sufficiently

Page 49: The Biogeochemistry of Submerged Soils

40 Transport Processes in Submerged Soils

dense, this may have a large effect on solute transfer between the soil and overly-ing water. Oligochaete worms are often present in submerged soils in populationsexceeding several thousand per m2 with burrows extending to several centimetres(Chapter 5). Once the burrows are constructed, the worms remain in them feedingwith their heads downward and their posterior ends upward in the overlying water.By waving their posteriors and moving their bodies in a peristaltic motion theycause the water in the burrows to be mixed with the overlying water. Solid particlesalso fall into the burrows and are mixed. Hence solutes diffusing into a burrowwill be rapidly transferred to the surface, and vice versa. The ecology of tubificidsand other organisms in the soil and floodwater are discussed in Chapter 5. I herediscuss approaches to modelling their effects on solute transfer between the soiland floodwater.

Three approaches have been taken to the analogous problem of mixing byinvertebrates in marine sediments (Aller, 1980a; Berner, 1980). The simplestapproach has been to lump together all the processes involved and to assumethat mixing is random and complete to a specified depth. This has been appliedsuccessfully to the long-term mixing of sediments under the combined actionof invertebrates and waves or currents, but is inappropriate for less perturbedsystems and short times. A second approach has been to express the effect ofburrowing as increased effective diffusion coefficients of solutes in the pore water,derived by fitting diffusion equations to empirical data (Aller, 1980a; Berner,1980; van Rees et al., 1996). But the physical basis of this approach is doubtful.

A third approach was developed by Aller (1980a, b) who studied solute fluxesin near-shore marine sediments showing seasonal variation. In this approach, thegeometry of the burrow–sediment system is allowed for explicitly and trans-port in the sediment between the burrows is described with appropriate diffusionequations. It is assumed that the burrows are oriented normal to the sediment sur-face and distributed uniformly or randomly in the horizontal plane (Figure 2.11).Thereby a cylindrical zone of influence is ascribed to each burrow with a radius

zone of influence(radius = r2 = 1/√πN)

burrow(radius = r1)

Figure 2.11 Distribution of worm burrows and cylinders of influence represented byboundary conditions for Equation (2.37)

Page 50: The Biogeochemistry of Submerged Soils

Mixing by Soil Animals 41

such that the whole sediment volume is accounted for. The water in a burrow isassumed to mix instantaneously with the overlying seawater, and solutes diffuseradially between the burrow and the sediment surrounding it as well as verti-cally between the sediment and overlying water. The corresponding continuityequation for transport in the sediment influenced by a particular burrow is

∂C

∂t= ∂

∂z

(D

∂C

∂z

)+ 1

r

∂r

(rD

∂C

∂r

)+ R (2.37)

where z is the distance from the sediment surface, r is the radial distance fromthe centre of the cylinder and R is the rate of production or consumption ofsolute in the sediment.

In Equation (2.37), the first term on the right-hand side accounts for diffusionin the vertical direction; the second term accounts for radial diffusion across thecylinder. The following boundary conditions apply. At the sediment–water andsediment–burrow interfaces, the concentrations are the same as in the overly-ing water:

z = 0 C = C0

r = r1 C = C0

At the boundary between adjacent cylinders, there is effectively no transferof solute:

r = r2 dC/dr = 0

where the radius of the cylinder, r2,= 1/√

πN , where N is the density of wormsper unit sediment surface area. At the bottom of the cylinder, the flux of soluteis constant:

z = L dC/dz = B

The value of B is specified from empirical observations.Aller (1980b) shows that if the mean distance between burrows is small com-

pared with their length, then a steady state (∂C/∂t = 0) will be attained rapidly,and he provides an analytical solution of Equation (2.37) for the steady statesubject to the above boundary conditions. (The solution is complicated, involv-ing Bessel functions, and is not reproduced here.) The mean concentration at aparticular depth is found by integrating the concentration across the cylinder ofsediment at that depth:

Cz =

∫ r2

r1

2πrC.dr

∫ r2

r1

2πr.dr

(2.38)

Aller uses the model to explain seasonally fluctuating profiles of NH4+ concen-

tration in sediments in Long Island Sound. In this system NH4+ is produced

Page 51: The Biogeochemistry of Submerged Soils

42 Transport Processes in Submerged Soils

in anoxic decomposition of organic matter in the sediment at a rate decreasingexponentially with depth,

R = R0 exp(−αz) + R1 (2.39)

and it is removed by nitrification in the overlying water and in worm burrows.The rate of NH4

+ formation and the density of the worms vary with seasonaltemperature changes. Figure 2.12 shows concentration profiles of NH4

+ in thesediments measured over 2 years and the corresponding profiles predicted by themodel using independently measured parameter values. It shows that the mainfeatures of the profiles and their seasonal dynamics are satisfactorily predicted.By comparison, a model using the same parameter values but only allowing fordiffusion in the vertical direction over-predicted the concentrations several fold.Aller found similar good agreement between observed and predicted profiles andfluxes of SO4

2− and Si in the sediments. He concluded that the model accountedsatisfactorily for the important processes operating.

The application of this approach is illustrated in Figures 2.13 and 2.14, whichshow the effects of tubificid worms on the movement of P between a submerged

July 1974

Dep

th (

cm)

0

5

10

15

November 1974

0

5

10

15

March 1975

0

5

10

15

July 1975

0 100 200 300 400 0 100 200 300 400 0 100 200 300 400

0 100 200 300 400 0 100 200 300 400 0 100 200 300 400

0

5

10

15

October 1975

0

5

10

15

March 1976

0

5

10

15

Concentration of NH4+ in solution (µM)

Figure 2.12 Concentration profiles of NH4+ at different times in sediments in Long

Island Sound. Points are measured data; lines are predicted with Equations (2.37) and(2.39) using independently measured parameter values (after Aller, 1980a). Reprintedwith permission from Elsevier

Page 52: The Biogeochemistry of Submerged Soils

Mixing by Soil Animals 43

Concentration of P in solution (µM)

0 100 200 300 400 500 600

Dep

th (c

m)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1000

5000

15 000

N =30 000

Figure 2.13 Effect of mixing of pore water by tubificid worms on profiles of P concen-tration in submerged soil calculated with Equations (2.37) and (2.40). Numbers on curvesare densities of tubificids

N (m−2)

0 10 000 20 000 30 000

P fl

ux (

mm

ol m

−2 d

ay− 1

)

0.0

0.2

0.4

0.6

0.8

L = 5 cm

1.5

3

N (m−2)

0 10 000 20 000 30 0000.0

0.2

0.4

0.6

0.8

1.0

5

1.5

3

Rat

io o

f soi

l sur

face

flux

to to

tal

Figure 2.14 Effect of mixing by tubificids on flux of P between soil and floodwatercalculated with Equations (2.37) and (2.40). Numbers on curves are depths of mixing

soil and overlying floodwater. The primary productivity of the floodwater, includ-ing the fixation of N by photosynthetic aquatic organisms, is often limited by thesupply of P from the soil. So enhanced P transport resulting from tubificid activ-ities can be important. Figure 2.13 shows calculated concentration profiles of Pin the soil near the floodwater for realistic densities of tubificids (see Chapter 5)and other parameters, and Figure 2.14 shows the corresponding fluxes from thesoil into the floodwater. Following Aller (1980b) for Si desorption in marinesediments, the rate of P desorption from the soil is calculated with the formula

R = k(Ceq − C) (2.40)

Page 53: The Biogeochemistry of Submerged Soils

44 Transport Processes in Submerged Soils

where k is a rate constant and Ceq an apparent equilibrium P concentration.Values of k = 10−6 s−1 and Ceq = 0.5 mm were used for the calculations inFigures 2.13 and 2.14. The values of the other parameters used were θL = 0.6,fL = 0.4, b = 100, B = 0, r1 = 0.5 mm, r2 = 1/

√πN , where the values of N

are given in the figures and L = 3 cm. It will be seen that the tubificids have alarge effect and the flux of P to the floodwater increased several fold for realisticnumbers and dimensions. For comparison, the fluxes of P from the soil requiredto sustain typical rates of primary production in the floodwater in ricefields arein the range 0.05–0.25 mmol m−2 day−1, calculated from measured primary pro-duction and the P contents of likely photosynthetic organisms given by Roger(1996). Because the tubificids depend upon the photosynthetic organisms for theircarbon, there will be a positive feedback between mixing by tubificids and netprimary production in the floodwater.

Note that the sensitivity of the net flux between the soil and water to theworms’ activities depends on the relation between the rate R and the solute con-centration. For the calculations in Figures 2.13 and 2.14, R varies linearly withconcentration as specified in Equation (2.40), and the flux is sensitive to wormactivity. But where the rate is independent of concentration, as for NH4

+ forma-tion in Equation (2.39), the net flux, which in this case is roughly R0/α + LR1,is necessarily independent of worm activity, though the distribution of the fluxbetween burrows and the sediment surface and the concentration profile are not.In practice the rate will always depend to some extent on concentration. But thepredictions here for the idealized steady state indicate the expected sensitivities.

Page 54: The Biogeochemistry of Submerged Soils

3 Interchange of Solutes betweenSolid, Liquid and Gas Phases

This chapter is concerned with how ions and uncharged solutes in the waterand soil solution in submerged soils interchange between the solid, liquid andgas phases present. This is a large topic. I give here the bare essentials neededto understand the transport and transformation processes discussed elsewhere inthe book, and I give references to more detailed treatments where appropriate.The water and atmosphere overlying the soil are dealt with first and then theadditional complexities in the soil.

A. WATER

3.1 COMPOSITION OF THE WATER

The water contains:

• dissolved matter– free ions;– inorganic and organic complexes;– uncharged molecules.

• particulate matter– large organic and inorganic polymers;– oxides;– clay minerals;– organic matter.

Because of their large surface areas, charged particles are very efficient scav-engers of ions from solution, and where the sediment load is large the concentra-tion of adsorbed ions may greatly exceed the concentration in solution. Similarlyfor ions that form complexes with organic or inorganic ligands, their total con-centration in solution may be far greater than the concentration of the free ion.Complexation and sorption are especially important in regulating the concentra-tions of trace metals in natural water systems. The interactions between ions andcharged particles are discussed in the sections on soil.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 55: The Biogeochemistry of Submerged Soils

46 Interchange of Solutes between Solid, Liquid and Gas Phases

3.1.1 ACIDS AND BASES

The concentrations of dissolved species in natural waters depend ultimately on thedissolution of basic rocks–carbonates, silicates and aluminosilicates–induced bythe action of weak acids in the water derived from dissolved gases–e.g. H2CO3

derived from CO2. Anions produced in acid–base reactions balance cations pro-duced in dissolution reactions. The charge balance is:

�m[cationm+] = �n[anionn−] (3.1)

Table 3.1 shows the main weak acids present in natural waters and typicalconcentration ranges. Table 3.2 shows the corresponding equilibrium constants.Table 3.1 shows that carbonic acid is by far the dominant acid with concentrationstypically of the order of several mM. It arises from the dissolution of carbonaterocks and atmospheric CO2, and from the respiration of aquatic and soil organ-isms. The concentrations of dissolved silica are 5–10 times smaller, and thoseof ammonium and orthophosphate smaller again, although NH4

+ concentrationsin the mM range may arise in the water in ricefields following fertilizer appli-cations. The hydrolysis products of certain metals, such as Fe(III) and Al(III),also behave as weak acids and may be important under particular circumstances.Dissolved amino, organic and humic acids are rarely a large part of the chargebalance in solution but may be important as metal ligands.

The distributions of different acid–base pairs with pH are shown in Figure 3.1.Bicarbonate (HCO3

−) is the dominant carbonate species at near neutral pH;silicic acid (H4SiO4) is essentially undissociated at all pHs of interest; and theammonium ion (NH4

+) is the dominant form of ammoniacal-N at pHs belowabout 8. Orthophosphate and sulfide have acidity constants near neutral pH.

For a given concentration of a particular dissolved acid, the proportions of thecomponent species in the equilibrium solution will depend on the alkalinity ofthe solution; that is, the balance of cations and non-dissociating anions present.This can be calculated as shown in Table 3.3 for the aqueous carbonate equilibria

Table 3.1 Concentrations of weak acids and bases in natural waters

Global meanfor freshwatera

Range in water Range insubmerged soil

solutions

Carbonate 0.97 mM 0.01–10 mM 5–100 mMSilicate 0.22 mM 0.1–0.5 mM 0.1–1.5 mMAmmonium 0–10 µM 0.001–1 mM 0.001–1 mMPhosphate 0.7 µM 0.5–25 µM 0.5–100 µMSulfide — Trace 0.01–10 µMAmino acids — Trace 0.1–10 µMOrganic acids — 0.001–1 mM 0.1–10 mM

Source:a Morel and Herring (1993).

Page 56: The Biogeochemistry of Submerged Soils

Composition of the Water 47

Table 3.2 Equilibrium constants for acid–base equilib-ria at 25 ◦C, I = 0

Equilibrium − log K

H2O = H+ + OH− 14.0CO2(g) + H2O = H2CO3

∗ 1.46

H2CO3∗ = H+ + HCO3

− 6.35

HCO3− = H+ + CO3

2− 10.33

H4SiO4 = H+ + H3SiO4− 9.86

H3SiO4− = H+ + H2SiO4

2− 13.1NH3(g) = NH3(aq) −1.76

NH4+ = H+ + NH3(aq) 9.24

H3PO4 = H+ + H2PO4− 2.15

H2PO4− = H+ + HPO4

2− 7.20

HPO42− = H+ + PO4

3− 12.35H2S(g) = H2S(aq) 0.99

H2S(aq) = H+ + HS− 7.02

HS− = H+ + S2− 13.9

CH2 NH2COOH = H+ + CH2 NH2COO− 9.78

CH3COOH = H+ + CH3COO− 4.76

with the equilibrium constants in Table 3.4. Similar calculations can be made forthe other dissolved acids.

Table 3.3 gives the equilibria in a closed system in which the total carbonateconcentration, CT, is fixed. In an open system, such as the water on the surfaceof a submerged soil, CT is variable and the resulting changes in pH dependon the balance of charge between the non-carbonate anions and cations present.Likewise if a quantity of strong acid, HX, or base, MOH, is added to the solution,the equilibria will adjust so as to neutralize part of the H+ or OH− added and sobuffer the change in pH. The changes in [H+] with alkalinity or dissolved CO2

can be found from (see Equation 9, Table 3.3):

CB − CA = [HCO3−] + 2[CO3

2−] + [OH−] − [H+] (3.2)

where CA is the concentration of non-carbonate anions after the addition ofacid HX and CB is the concentration of cations after addition of base MOH. IfCB > CA, the difference CB − CA is the alkalinity of the solution; if CA > CB,the difference CA − CB is the mineral acidity.

3.1.2 SPECIATION

Many ions and uncharged molecules are present in solution as more than onespecies, depending on the concentrations of ligand ions and molecules and the

Page 57: The Biogeochemistry of Submerged Soils

48 Interchange of Solutes between Solid, Liquid and Gas Phases

H2CO3*

H2CO3

CO32−HCO3

H2PO4− HPO4

2−

H3PO4PO4

2−

H4SiO4

H3SiO4−

H2SiO42−

log

activ

ity (

arbi

trar

y un

its)

pH

4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11

NH4+ NH3

H2S HS−

S2− CH3COOH

CH3COO−

(a) (b)

(c) (d)

(e) (f)

Figure 3.1 Distribution of dissolved acid–base species at constant total concentration insolution (after Morel and Herring, 1993). Reproduced by permission of Wiley, New York

solution pH and ionic strength. Complexation of metals with ligands can resultin the total concentration of the metal being far greater than the concentration ofthe free ion. This topic is covered in detail by Morel and Herring (1993), Stummand Morgan (1996) and, for chelation by humic substances, by Tipping (2002).

A complex is a species in which a metal atom or ion is attached by coordinatebonds to one or more ligand ions or uncharged molecules. The complex mayitself be positive, negative or uncharged. In forming a coordinate bond the liganddonates a pair of electrons to the metal. In so doing the ligand is acting as a

Page 58: The Biogeochemistry of Submerged Soils

Composition of the Water 49

Table 3.3 Equilibria in aqueous carbonate solutions

Species

CO2(g), CO2(aq), H2CO3, HCO3−, CO3

2−, H+, OH−, M+, X−

[H2CO3∗]a = [CO2(aq)] + [H2CO3]

Equilibriab

[CO2(aq)]/[CO2(g)] = H c (0)

[H2CO3∗]/pCO2 = KH (0a)

[CO2(aq)]/[H2CO3] = K (1)

[H+][HCO3−]/[H2CO3] = KH2CO3 (2)

[H+][HCO3−]/[H2CO3

∗] = K1 (2a)

[H+][CO32−]/[HCO3

−] = K2 (3)

[H+][OH−] = KW (4)

Ionization fractions for constant total carbonate concentration, CT

CT = [H2CO3∗] + [HCO3

−] + [CO32−] (5)

[H2CO3∗] = α0CT[HCO3

−] = α1CT[CO32−] = α2CT

α0 =(

1 + K1

[H+]+ K1K2

[H+]2

)−1

(6)

α1 =(

[H+]

K1+ 1 + K2

[H+]

)−1

(7)

α2 =(

[H+]2

K1K2+ [H+]

K2+ 1

)−1

(8)

Electrical neutrality condition

[H+] + [M+] = [HCO3−] + 2[CO3

2−] + [OH−] + [X−] (9)

a The ‘apparent’ concentration of H2CO3 since [CO2(aq)] � [H2CO3].b Equilibrium constants are defined at constant ionic strength.c Dimensionless Henry’s law constant, in which [CO2(g)] = PCO2 /RT .Source: Stumm and Morgan (1996). Reproduced by permission of Wiley, New York.

Table 3.4 Equilibrium constants for carbonate equilib-ria at 25 ◦C, I = 0

Equilibrium Constant − log K

CO2(g) + H2O = H2CO3∗ KH

a 3.47

H2CO3∗ = H+ + HCO3

− K1 6.35

HCO3− = H+ + CO3

2− K2 10.33

H2O = H+ + OH− KW 14.0

a PCO2 in kPa.

Page 59: The Biogeochemistry of Submerged Soils

50 Interchange of Solutes between Solid, Liquid and Gas Phases

Lewis base and the metal as a Lewis acid. A characteristic of ligands is that theyhave a lone pair of electrons which they can donate to empty electron orbitalson the metal. Some ligands also have empty p- or d-orbitals and can producecomplexes in which a double bond is formed with the metal: a sigma bond bydonation of the lone pair from the ligand to the metal and a pi bond by backdonation of electrons on the metal to empty d-orbitals on the ligand. The termchelate is reserved for species involving polydentate ligands that form a ring ofatoms including the metal.

Inorganic and organic ligands contain the following electron donor atoms fromGroups IVB to VIIB of the Periodic Table:

C N O FP S ClAs Se Br

Te I

Formation of coordination complexes is typical of transition metals, but othermetals also form complexes. The tendency to form complexes is a function ofthe metal’s electron configuration and the nature of its outer electron orbitals.Metal cations can be classified into types A and B based on their coordinationcharacteristics, as shown in Table 3.5. A-type cations, which tend to be from theleft side of the Periodic Table, have the inert-gas type electron configuration withlargely empty d-orbitals. They can be imagined as having electron sheaths noteasily deformed under the influence of the electric fields around neighbouringions. B-type cations have a more readily deformable electron sheath.

In consequence, A-type cations form complexes preferentially with the fluorideion and ligands having oxygen as their electron donor atom. They are attracted toH2O more strongly than to NH3 or CN−, and they do not form sulfides becauseOH− ions readily displace HS− or S2− ions. They tend to form sparingly solubleprecipitates with OH−, CO3

2− and PO43−. By contrast, B-type cations coordinate

preferentially with ligands containing I, S or N as electron donors. They may bindNH3 more strongly than H2O and CN− more strongly than OH−, and they tendnot to form complexes with the main functional groups in organic matter, whichhave O as electron donor. They form insoluble sulfides and soluble complexeswith S2− and HS−.

Table 3.6 shows the major inorganic species expected in a solution with acomposition typical of natural fresh water. Some calculations for organic ligandsin submerged soil solutions are given in Section 3.7.

3.1.3 EQUILIBRIUM CALCULATIONS

Complete calculations of chemical equilibria in natural waters and soil solutionsare complicated because such a large number of solutes, solids and gases are

Page 60: The Biogeochemistry of Submerged Soils

Composition of the Water 51

Table 3.5 Classification of metal ions

A-type metal cations Transition-metal cations B-type metal cations

Electron configuration ofinert gas, lowpolarizability, ‘hardspheres’

One to nine outer shellelectrons, notsphericallysymmetric

Electron number correspondsto Ni0, Pd0 and Pt0 (10 or 12outer shell electrons), lowelectronegativity, highpolarizability, ‘soft spheres’

(H+), Li+, Na+, K+, Be2+,Mg2+, Ca2+, Sr2+, Al3+,Sc3+, La3+, Si4+, Ti4+,Zr4+, Th4+

V2+, Cr2+, Mn2+,Fe2+, Co2+, Ni2+,Cu2+, Ti3+, V3+,Cr3+, Mn3+, Fe3+,Co3+

Cu+, Ag+, Au+, Tl+, Ga+,Zn2+, Cd2+, Hg2+, Pb2+,Sn2+, Tl3+, Au3+, In3+, Bi3+

Ligands Ligands

F > O > N = Cl > Br >I > S

S > I > Br > Cl = N > O > F

OH− > RO− > RCOO−

CO32− � NO3

PO43− � SO4

2− � ClO4−

Source: Stumm and Morgan (1996). Reproduced by permission of Wiley, New York.

involved. However general computer programs are available to perform such cal-culations using successive approximation (Melchior and Bassett, 1990; Mangoldand Tsang, 1991; Sposito, 1994). WHAM (Tipping, 1994, 2002) gives particularattention to reactions involving humic substances.

An important component of equilibrium calculations is the conversion betweenion activities, which equilibrium constants refer to, and ion concentrations, whichmass balance and electrical neutrality equations refer to. The conversion is madewith activity coefficients defined by the relation:

ai = γiCi (3.3)

Various empirical relations are available for calculating individual ion activitycoefficients [discussed by Stumm and Morgan (1996) for natural waters andSposito (1984a, b), for soil solutions]. In the calculations in this book I usedthe Davies equation:

log γ = −AZ2

( √I

1 + √I

− 0.3I

)(3.4)

where I is ionic strength (= 12�CiZi

2), Z is ionic charge and A = 1.82 ×106(εT )−1.5, where ε is the dielectric constant (A ≈ 0.5 for water at 25 ◦C).This relation is valid for I < 0.5 m.

Page 61: The Biogeochemistry of Submerged Soils

52 Interchange of Solutes between Solid, Liquid and Gas Phases

Table 3.6 Major inorganic species in representative natural water

Condition Element Major species Fresh water[Mn+/MT]

B(III) H3BO3, B(OH)4−

V(V) HVO42−, H2VO4

Hydrolysed, Cr(VI) CrO42−

anionic As(V) HASO42−

Se(VI) SeO42−

Mo(VI) MoO42−

Si(IV) Si(OH)4

Li Li+ 1.00

Na Na+ 1.00

Mg Mg2+ 0.94

Predominantly K K+ 1.00

free aquo ions Ca Ca2+ 0.94

Sr Sr2+ 0.94

Cs Cs+ 1.00

Ba Ba2+ 0.95

Be(II) BeOH+, Be(OH)20 1.5 × 10−3

Al(III) Al(OH)3(s), Al(OH)2+, Al(OH)4

− 1 × 10−9

Ti(IV) TiO2(s), Ti(OH)40

Mn(IV) MnO2(s)

Fe(III) Fe(OH)3(s), Fe(OH)2+, Fe(OH)4

− 2 × 10−11

Co(II) Co2+, CoCO30 0.5

Ni(II) Ni2+, NiCO30(Ni2+, NiCl) 0.4

Complexation with Cu(II) CuCO30, Cu(OH)2

0 0.01

OH−, CO32−, Zn(II) Zn2+, ZnCO3

0(Zn2+, ZnCl) 0.4

HCO3−, Cl− Ag(I) Ag+, AgCl0(AgCl2

−, AgCl) 0.6

Cd(II) Cd2+, CdCO30(CdCl2) 0.5

La(III) LaCO3+, La(CO3)2

− 8 × 10−3

Tl(I), Tl+, Tl(OH)30, Tl(OH)4

− 2 × 10−21

Tl(III)

Hg(II) Hg(OH)20(HgCl4

2−) 1 × 10−10

Pb(II) PbCO30(PbCl+, PbCO3) 5 × 10−2

Bi(III) Bi(OH)3 7 × 10−16

Fresh water conditions: pH = 8, Alk = 2 mM, [SO42−]T = 0.3 mM, [Cl−] = 0.25 mM, [Ca2+]T = 1 mM,

[Mg2+]T = 0.3 mM, [Na+]T = 0.25 mM, O2 at saturation with air, I = 5 mM.Source: Stumm and Morgan (1996). Reproduced by permission of Wiley, New York.

Page 62: The Biogeochemistry of Submerged Soils

pH Buffer Capacity 53

3.2 pH BUFFER CAPACITY

The extent to which the pH of a solution is buffered against additions or removalsof protons is measured by the solution’s pH buffer capacity. This is defined asthe amount of strong acid or base required to produce unit change in pH. Thebuffering depends on the transfer of protons between donors and acceptors, i.e.Bronsted acids and bases, which form conjugate acid–base pairs. The pH buffercapacity of a solution is calculated from the buffer capacities of the individualacid–base pairs present.

Consider a generic acid–base pair HX–X− representing the various acid–basepairs in a solution. The pH buffer capacity of the HX–X− pair is defined as

bHX = d[X−]

d pH(3.5)

The total concentration of the pair is [HX] + [X−] = [Xtotal] and from the acidityconstant, K, [HX] = [H+][A]/K , hence

[X−] = [Xtotal]

[H+]/K + 1(3.6)

Substituting in Equation (3.5) for [X−] from Equation (3.6)

bHX = [Xtotal]d 1/

([H+]/K + 1

)d pH

(3.7)

Hence

bHX = [Xtotal]d[H+]

d pH

d 1/([H+]/K + 1

)d[H+]

= 2.303[Xtotal]K[H+](

K + [H+])2

or

bHX = 2.303[HX][X−]

[HX] + [X−](3.8)

The total buffer capacity of the solution is equal to the sum of the buffercapacities of the individual acid–base pairs present, each given by an equationlike Equation (3.8). In an aqueous solution of acid HA, three acid–base pairs arepresent: HA–A−, H3O+ –H2O and H2O–OH−. Because [H3O+] and [OH−] areboth negligible compared with [H2O], in Equation (3.8) [X−] = ([HX] + [X−])for H3O+ –H2O and [HX] = ([HX] + [X−]) for H2O–OH−. Hence

bH3O+ = 2.303[H3O+] (3.9)

bOH− = 2.303[OH−] (3.10)

Therefore

bsolution = 2.303

([H3O+] + [OH−] + [HA][A−]

[HA] + [A−]

)(3.11)

Page 63: The Biogeochemistry of Submerged Soils

54 Interchange of Solutes between Solid, Liquid and Gas Phases

If other acid–base pairs are present the buffer capacity is

bsolution = 2.303

([H3O+] + [OH−] + [HA][A−]

[HA] + [A−]+ [HB][B−]

[HB] + [B−]+ . . . .

)

(3.12)

Polyprotic acids can be treated as a mixture of monoprotic acids. For example,consider the diprotic acid H2C which forms the acid–base pairs H2C = HC− +H+ and HC− = C2− + H+. The acidity constants are K1 = [H+][HC−]/[H2C]and K2 = [H+][C2−]/[HC−], respectively. The total buffer capacity of the solu-tion is therefore

bsolution = 2.303

([H3O+] + [OH−] + [H2C][HC−]

[H2C] + [HC−]+ [HC−][C2−]

[HC−] + [C2−]

)

(3.13)

For example, for a solution buffered by the CO2 –H2O–H2CO3 –HCO3− sys-

tem, application of Equation (3.13) gives

bsolution = 2.303{[H3O+] + [OH−] + [α1(α0 + α2) + 4α2α0] CT} (3.14)

where α0, α1 and α2 are ionization fractions defined in Table 3.3. This equationpredicts that the buffer capacity will pass through minima at pH 4–4.5 where[H3O+] and [HCO3

−] are both low, at pH 8.3 where [H2CO3] and [CO32−] are

low, and at pH 10.5–11 where [HCO3−] and [OH−] are low. At these pH ranges,

changes in the concentrations of acids or bases in the solution will cause largepH changes.

3.3 EQUILIBRIUM WITH THE GAS PHASE

The equilibrium distribution of a volatile solute between gas and liquid phases isdescribed by Henry’s law. For the equilibrium A(l) = A(g) in a dilute solutionat low gas pressure,

[A(l)] = KHpA (3.15)

where [A(l)] is the concentration of the dissolved gas in solution, pA is the partialpressure in the gas phase and KH is the Henry’s law constant. (At high concentra-tions or gas pressures, [A(l)] and pA are replaced by the corresponding activitiesand fugacities.) The constant is also sometimes expressed in dimensionless form,H, such that

[A(l)] = H [A(g)] (3.16)

where [A(g)] is the concentration in the gas phase. Hence

H = RT KH (3.17)

Page 64: The Biogeochemistry of Submerged Soils

Equilibrium with the Gas Phase 55

Table 3.7 Henry’s law constants for importantgases in submerged soils at 25 ◦C and typical par-tial pressures in the atmosphere

KH Typical partialpressure

(M kPa−1) (kPa)

N2 6.52 × 10−6 78O2 1.24 × 10−5 21CO2 3.35 × 10−4 3.5 × 10−2

CH4 1.27 × 10−5 1.7 × 10−4

NH3 5.63 × 10−1 0.1–5 × 10−7

H2S 1.04 × 10−3 < 2 × 10−8

NO2 9.87 × 10−5 1–5 × 10−7

NO 1.88 × 10−5 1–5 × 10−8

N2O 2.54 × 10−5 3 × 10−5

where R is the gas constant (= 8.314 kPa L mol−1 K−1) and T is the temper-ature (K). Values of KH for important gases in submerged soils are given inTable 3.7.

For some volatile solutes, slow reactions influence the rate of equilibrationbetween the gas and liquid phases. Generally the rate of gas transfer across theliquid–gas interface is the rate-limiting step, as discussed in Section 3.4. Butthere may also be slow hydration or other reactions in solution that must beallowed for. An important example is the hydration of CO2, whose half-life maybe comparable to rates of transfer of CO2 across the air–water interface.

Kinetics of CO2 Hydration

The kinetics of the hydration and dehydration reactions are slow in comparisonwith some processes in the water. The reactions are

CO2 + H2Okf 1−−−→←−−−kb1

H2CO3 (3.18)

and

CO2 + H2Okf 2−−−→←−−−kb2

HCO3− + H+ (3.19)

and the corresponding rate law is

−d[CO2]

dt= (

kf 1[CO2] − kb1[H2CO3]) + (

kf 2[CO2][OH−] − kb2[HCO3−]

)(3.20)

Page 65: The Biogeochemistry of Submerged Soils

56 Interchange of Solutes between Solid, Liquid and Gas Phases

Substituting from Table 3.3 for the equilibrium constant for dissociation ofH2CO3, which is fast,

−d[CO2]

dt= (

kf 1 + kf 2)

[CO2] − (kb1 + kb2KH2CO3

)[H2CO3] (3.21)

or

−d[CO2]

dt= kCO2 [CO2] − kH2CO3

KH2CO3

[H+][HCO3−] (3.22)

where kCO2 = kf1 + kf2 and kH2CO3 = kb1 + kb2KH2CO3 . Equation (3.22) corres-ponds to the simplified scheme

CO2 + H2OkCO2−−−→←−−−

kH2CO3

H2CO3

fast−−−→←−−− H+ + HCO3− (3.23)

That is, the hydration reaction is first order with respect to dissolved CO2. The rateconstant kCO2 = 0.025–0.04 s−1(25 ◦C) and activation energy ≈63 kJ mol−1. Forthe dehydration reaction, kH2CO3 = 10–20 s−1(20–25 ◦C) and activation energy≈67 kJ mol−1.

3.3.1 FLOODWATER CO2 DYNAMICS

The pH of the water on the surface of a submerged soil often depends on theactivity of photosynthetic organisms. Photosynthesis by aquatic plants and algaeremoves dissolved CO2 during the day, but at night the net respiratory activity ofthe organisms returns CO2 to the water and the concentration of dissolved CO2

and acidity increase:

CO2 + H2Ophotosynthesis−−−−−−−→←−−−−−−−

respirationCH2O + O2 (3.24)

where CH2O is organic matter produced in photosynthesis or consumed in respi-ration. As a result the pH may rise as high as 10 during the day but fall by twoor three pH units at night. Figure 3.2 shows measured diurnal changes in pH andcarbonate species in the floodwater of a ricefield. The relations between pH, alka-linity and carbonate equilibria are described by Equation (3.2). Equation (3.24)shows that photosynthesis and respiration do not affect the alkalinity of the waterper se. The pH increases or decreases with the change in CT at constant alkalin-ity. The change in pH depends on the alkalinity as it affects the initial pH and theconsequent acid–base system operating. At pHs below pK1(= 6.3), CO2(aq) isthe dominant species and there is little change in pH with CT. Between pK1 andpK2(= 10.3)HCO−

3 is the dominant species and roughly 1 mol of H+ is consumedper C fixed in photosynthesis (HCO3

− + H+ → CH2O + O2), with a correspond-ingly greater pH change. At pHs above pK2, CO3

2− is the dominant species androughly 2 mol of H+ are consumed per C fixed (CO3

2− + 2H+ → CH2O + O2),

Page 66: The Biogeochemistry of Submerged Soils

Equilibrium with the Gas Phase 57

Time

600 800 1000 1200 1400 1600 1800 2000 2200

Per

cent

mol

e fr

actio

n of

H2C

O3,

HC

O3−

and

CO

32−

0

5

10

15

80

90

100

pH

6

7

8

9

10

Fre

e C

O2

cont

ent (

mg

L−1)

0

10

20

30

40

50

60HCO3

pH

Free CO2

CO32−

H2CO3

Figure 3.2 Diurnal changes in pH and concentrations of carbonate species in the flood-water in a ricefield (Mikkelsen et al., 1978). Reproduced by permission of Soil Sci.Soc. Am.

and the pH change is correspondingly larger again. Figure 3.3 shows calculatedchanges in pH for a sinusoidally varying floodwater [H2CO3

∗] over the dayfor two different alkalinities. The dissolved CO2 concentrations are the same inFigure 3.3(a) and (b); only the alkalinities differ.

In principle, the alkalinity of the water will also be affected by the balanceof nutrient ions consumed and released by organisms in the water. But in prac-tice these have a minor affect compared with CO2. The average composition ofthe algal biomass in natural waters is given by the Redfield formula (Redfield,1934) as C106H263O110N16P. Therefore for the complete stoichiometry of algalphotosynthesis and respiration, we have with NO3

− as the source of N

106CO2 + 16NO3− + H2PO4

− + 122H2O + 17H+

= C106H263O110N16P + 133O2 (3.25)

and with NH4+

106CO2 + 16NH4+ + H2PO4

− + 106H2O

= C106H263O110N16P + 106O2 + 15H+ (3.26)

The corresponding changes in alkalinity are +17/106 = +0.16 molc per mol Cfixed for NO3

− nutrition and −15/106 = −0.14 molc per mol C fixed for NH4+

nutrition. More significant changes in the alkalinity of ricefield floodwater are

Page 67: The Biogeochemistry of Submerged Soils

58 Interchange of Solutes between Solid, Liquid and Gas Phases

6.0

6.5

7.0

7.5

8.0

8.5

9.0

0.0

0.2

0.4

0.6

0.8

1.0

0

10

20

30

40

Time (h past midnight)

6 8 10 12 14 16 18 20 22pH

6.0

6.5

7.0

7.5

8.0

8.5

9.0

Rat

io o

f H

2CO

3*,

HC

O3−

or C

O32−

to

CT

0.0

0.2

0.4

0.6

0.8

1.0

Fre

e C

O2

(mg

L−1)

0

10

20

30

40

HCO3−/CT

pHFree CO2

H2CO3*/CT CO32−/CT

H2CO3*/CT

Free CO2 HCO3−/CT

pH

(a) [Alk] = 10 mM

(b) [Alk] = 0.5 mM

Figure 3.3 Calculated diurnal changes in the pH and concentrations of carbonatespecies in ricefield floodwater for sinusoidally varying [H2CO3

∗] with (a) [Alk] = 10 mM,(b) [Alk] = 0.5 mM. The free CO2 concentrations are in mg L−1 to be consistent withFigure 3.2

caused by additions of nitrogenous fertilizers. Effects on pH again depend oninitial pH and corresponding buffer systems operating.

3.4 GAS TRANSPORT ACROSS THE AIR–WATER INTERFACE

The floodwater is for the most part not in equilibrium with the atmospherebecause rates of production of volatile solutes in the water exceed rates of gasexchange across the air–water interface. In particular, during the day, rates ofCO2 consumption and O2 production by photosynthesizing organisms are gen-erally sufficient to cause undersaturation of CO2 and supersaturation of O2.Conversely, at night, respiration causes depletion of O2 and supersaturation ofCO2. The underlying soil is also a large sink for O2 and source of CO2. Theresulting diurnal changes in dissolved CO2 can cause large changes in floodwaterpH, often from near neutral at night to pH 10 during the day.

Page 68: The Biogeochemistry of Submerged Soils

Gas Transport Across the Air–Water Interface 59

still air layer

still water layer

turbulent bulk water

turbulent bulk air

dzG

dzL

Figure 3.4 The air–water interface

Two main approaches have been taken to modelling the air–water interfacein natural systems so as to calculate rates of volatilization and dissolution (Lissand Slater, 1974; Frost and Upstill-Goddard, 1999; McGillis et al., 2001). In thesimpler the interface is viewed as two thin still layers, one in the air and one inthe water, separating well-mixed bulk phases (Figure 3.4). Transport across thestill layers is by diffusion. The still layers arise because of the increased viscosityof the air and water near the interface. Their thicknesses depend on such factorsas wind speed and surface roughness. Under turbulent conditions, the thickness ofthe still layers is reduced and rates of gas transport correspondingly increased. Atsteady state the fluxes across the layers are equal. Therefore, if the gas undergoesno reactions, we have from Fick’s first law

F = − DG

δzG(CG0 − CG) = − DL

δzL(CL − CL0) (3.27)

where subscripts G and L indicate the gas and liquid phases, respectively, andsubscript 0 indicates the interface.

The alternative approach considers that turbulent eddies periodically mix thesurface layers with the bulk fluids. The flux across the interface is related tothe concentration difference by a transfer coefficient equal to the square rootof the diffusion coefficient divided by a characteristic time, τ , representing thefrequency of mixing. Thus

F = −√

DG

τG(CG0 − CG) = −

√DL

τL(CL − CL0) (3.28)

Neither model accounts completely for the processes operating in the interface,and they provide similar fits to empirical data (Frost and Upstill-Goddard, 1999).However the first model has the advantage of conceptual simplicity and I use itin the following sections.

Page 69: The Biogeochemistry of Submerged Soils

60 Interchange of Solutes between Solid, Liquid and Gas Phases

If the gas obeys Henry’s law, then

CL0 = HCG0 (3.29)

where H is the dimensionless Henry’s law constant. Eliminating CG0 betweenEquations (3.27) and (3.29) gives

F = 1

kG(CG − CL0/H) = 1

kL(CL0 − CL) (3.30)

where kG(= DG/δzG) and kL(= DL/δzL) are transfer coefficients for the gas andliquid phases, respectively. Hence CL0 can be eliminated to obtain the followingequation for the flux through the water film

F = 1

1/kL + H/kG(CG − HCL) (3.31)

In Equation (3.31), 1/kL is the resistance to transfer through the liquid film andH/kG is the resistance to transfer through the gas film. The relative importanceof these resistances is given by the ratio

resistance in gas phase

resistance in liquid phase= H

kL

kG(3.32)

Table 3.8 gives values of this ratio for important gases in submerged soils. Mostof the gases are sparingly soluble, and the resistance in the liquid phase is muchgreater than that in the gas. This is because diffusion coefficients in water aretwo orders of magnitude smaller than those in air and because, for these gases,H is small. But for very soluble gases, such as NH3, resistance in the gas phasemay be limiting. Solubility varies much more between different gases than thediffusion coefficients and is therefore the main determinant of whether gas orliquid phase resistance is limiting. If H is less than about 5, transport in the

Table 3.8 Relative importance of resistances in air (rG) and in water (rL) togas transfer across an air–water interface at 25 ◦C and 1 atm (Equation 3.32)

DG(dm2 s−1)

DL(dm2 s−1)

H rG/rLa

O2 2.05 × 10−3 2.26 × 10−7 3.08 × 10−2 3.40 × 10−6

CO2 1.55 × 10−3 1.93 × 10−7 8.29 × 10−1 1.03 × 10−4

CH4 2.20 × 10−3 1.73 × 10−7 3.16 × 10−2 2.48 × 10−6

H2S 1.66 × 10−3 2.00 × 10−7 2.57 3.09 × 10−4

N2 2.04 × 10−3 2.02 × 10−7 1.62 × 10−2 1.59 × 10−6

NH3 2.19 × 10−3 2.49 × 10−7 1.39 × 103 0.159N2O 1.55 × 10−3 1.98 × 10−7 6.29 × 10−2 8.03 × 10−6

NO 2.04 × 10−3 2.55 × 10−7 4.65 × 10−2 5.81 × 10−6

a For δzL = δzG and assuming gases do not react with water.

Page 70: The Biogeochemistry of Submerged Soils

Gas Transport Across the Air–Water Interface 61

liquid phase is limiting; if it is greater than about 500, transport in the gas phaseis limiting.

However, this simple picture only applies to gases that do not undergo reactionsin the boundary layers. For gases that do react, for example through hydrationand acid–base reactions, the net flux depends on the simultaneous movementof all the solutes involved, and the flux will not be the simple function of con-centration expressed in Equation (3.25). An example is CO2, which reacts withwater to form carbonic acid and carbonate species–H2CO3, HCO3

− and CO32−.

The situation is complicated because the exchange of H+ ions in the carbonateequilibria results in a pH gradient across the still layer, and it is therefore nec-essary to account for the movement of H+ ions across the still layer as well asthe movement of carbonate species. The situation is further complicated in thecase of CO2 by the kinetics of hydration and dehydration, which may be slow incomparison with transport.

3.4.1 CO2 TRANSFER ACROSS THE AIR–WATER INTERFACE

Under equilibrium conditions, the bulk of the dissolved CO2 is present as HCO3−

or CO32− or both if the pH is greater than about 6. Therefore, where a gradi-

ent of CO2 concentration exists across a solution, the net flux of CO2 will begreatly increased if there is rapid equilibration between the dissolved CO2 andcarbonate species. Consequently, most plants and animals have evolved enzymesystems to catalyse the hydration–dehydration equilibria and the enzyme respon-sible—carbonic anhydrase—is present in most plant and animal cells. It islikely that this enzyme will often be present extracellularly in natural waters.This is because many aquatic plants use HCO3

− for photosynthesis under lowCO2 conditions by catalysing the conversion of HCO3

− to CO2 outside theplasma membranes of leaf cells. The mechanism involves catalysis by extra-cellular carbonic anhydrase in conjunction with H+ extrusion across the plasmamembrane (Graham et al., 1984; Tsuzuki and Miyachi, 1989). Since at least someforms of the enzyme are soluble, appreciable concentrations should arise in thewater under intense algal growth, though the stability of the enzyme under highlight and O2 conditions is unknown. The presence of carbonic anhydrase orsimilar enzymes catalysing CO2 hydration has been demonstrated in seawaterwith corresponding differences in rates of CO2 exchange (Berger and Libby,1969).

The following calculations show the range of effects from infinitely slow hydra-tion–dehydration to infinitely fast. Emerson (1975) and Kirk and Rachhpal-Singh(1992) and have made calculations allowing for the kinetics of the uncatalysedhydration–dehydration reactions, giving intermediate results.

We have for the flux of CO2 across the still air layer an equation of the type

FG = − DG

δzG(CG − CL0/H) (3.33)

Page 71: The Biogeochemistry of Submerged Soils

62 Interchange of Solutes between Solid, Liquid and Gas Phases

At steady state the flux of CO2 gas must equal the net flux of dissolved CO2 andcarbonate species, therefore

FGC = FLC = FLH2CO3∗ + FLHCO3

− + FLCO32− (3.34)

where H2CO3∗ represents CO2(aq) + H2CO3.

The fluxes of the uncharged solutes, CO2 and H2CO3, are given by equationsof the type

FLA = −DLA

δzL(CLA − CLA0) (3.35)

The fluxes of charged solutes depend on the diffusion potential arising fromdifferences in the mobilities of individual ions, as well as on an ion’s ownconcentration gradient (Equation 2.21). The effect of diffusion potentials will beimportant if the carbonate species are a large part of the total ion concentration,as they often will be. Therefore we have for the net flux of ion B

FLB = −DLBdCLB

dz+ ZBCLBDLB

�DLidCLi/dz

�Zi2DLiCLi

(3.36)

where subscript i refers to all the co- and counter-ions in solution and ZB andZi are the ionic charges. This gives

FLB ≈ −DLB

δzL(CLB0 − CLB) + ZB

2(CLB + CLB0)DLBΦ (3.37)

where

Φ = �DLidCLi/dz

�Zi2DLiCLi

There is an equation of this type for each of the ions present.The principal cations and anions in floodwaters are generally Ca2+, Cl−,

HCO−3 , CO2−

3 and OH−. Therefore we have five equations of type (3.37) forthe fluxes of the five charged species, Equation (3.33) for CO2 gas and Equa-tion (3.35) for H2CO3

∗. These seven equations contain six unknowns—the con-centrations of H2CO3

∗ and the five ions in solution at the interface—and theseare found with the following six equations: Equation (3.34), Equations (1)–(3)in Table 3.3, and FLCa2+ = 0 and FLCl− = 0—i.e. no net flux of Ca2+ and Cl−across the interface, their concentration gradients being balanced by their diffu-sion potential gradients.

Note that charge balance between the diffusing ions is inherent in Equa-tion (3.37). Note also that the movement of H+ ions formed in the carbonateequilibria is allowed for in the movements of the various conjugate acid–basepairs present: H2CO3 –HCO3

−, HCO3− –CO3

2− and H2O–OH−. For each molof CO2 entering or leaving the water, 1 mol of H+ is added or removed at the

Page 72: The Biogeochemistry of Submerged Soils

Gas Transport Across the Air–Water Interface 63

interface and transferred to or from the bulk solution by the diffusion of conjugateacid–base pairs. Thus

FGC = FLH2CO3∗ − FLCO3

2− − FLOH− (3.38)

which is inherent in the mass and charge balances.Figure 3.5 shows calculated concentration profiles in the still water layer for

realistic conditions in ricefields. In figure 3.5(a) the CO2 pressure is large, the pHin the bulk solution correspondingly low (pH 6.7), and the movement of dissolvedCO2 to the interface primarily as H2CO3

∗. The loss of CO2 raises the pH at theinterface (to pH 8.2), tending to offset the depletion of HCO3

− and the gradientof HCO3

− concentration is small. In figure 3.5(b) the CO2 pressure is small, thepH in the bulk solution correspondingly high (pH 10.6), and the movement ofdissolved CO2 away from the interface is primarily as HCO3

−. Dissolution ofCO2 lowers the pH at the interface (to pH 8.3) and there is therefore a gradient ofdecreasing OH− towards the interface. The gradient of CO3

2− is also negative.Since the mobility of OH− is about five times that of HCO3

− and CO32−, there

is therefore an excess negative potential at the interface and as a result Ca2+diffuses to the interface and Cl− away.

Concentration (mM)(a)

(b)

0.00 0.25 0.50 0.75 1.00 1.250.0

0.1

0.0

Dep

th (

mm

)D

epth

(m

m)

0.1

HCO3−

HCO3−

Ca2+Cl−

Ca2+Cl−

H2CO3*OH−

H2CO3*

OH−

CO32−

Figure 3.5 Profiles of CO2, HCO3−, etc. across still water layer. Still layer

thickness both = 1000 µm, [Ca2+]L∞ = 0.5 mM, [Cl−]L∞ = 0.15 mM, PCO2L∞ = 1 kPa(a), 2.5 × 10−5 kPa (b)

Page 73: The Biogeochemistry of Submerged Soils

64 Interchange of Solutes between Solid, Liquid and Gas Phases

CO2 pressure (Pa)

0.01 0.1 1 10 100 1000

CO

2 flu

x (k

g C

ha−1

h−1

)

−10

−5

0

5

10

15

with CO2 hydration

without

Figure 3.6 Flux of CO2 as a function of CO2 pressure with and without carbonateequilibria

Figure 3.6 shows how the flux of CO2 across the interface varies with CO2

pressure in the bulk solution, with and without equilibration between CO2 andcarbonate species in the boundary layer. A positive flux indicates dissolution anda negative flux volatilization. The figure shows that the effect of the carbonateequilibria is very marked at small CO2 pressures, but insignificant at large pres-sures where transport across the boundary layer is primarily as H2CO3

∗. At smallCO2 pressures the rate of dissolution is enhanced many fold by the carbonateequilibria, the effect increasing as the CO2 pressure decreases and the pH of thebulk solution correspondingly increases.

An important practical problem in ricefields is the loss of N fertilizer throughvolatilization of NH3 from the floodwater. Loss of NH3 is sensitive to the pHof the floodwater, and hence is intimately linked to the dynamics of dissolvedCO2 (Bouldin and Alimago, 1976). To quantify this it is necessary to considerthe simultaneous transfers of CO2 and NH3 across the air–water interface andtheir coupling through acid–base reactions. There is an equation of type (3.33)for the flux of NH3 across the still air layer and, as for the dissolved CO2 andcarbonate species, the flux across the still water layer is

FGN = FLN = FLNH4+ + FLNH3 + FLNH4OH (3.39)

The acid–base pairs involved are NH4+ –NH3 and NH4

+ –NH4OH, in additionto those listed above, and we have

FGC − FGN = FLH2CO3 − FLCO32− − FLNH3 − FLNH4OH − FLOH (3.40)

Equation (3.40) is inherent in the mass and charge balances. These equations canbe solved as before to calculate the simultaneous fluxes of CO2 and NH3 acrossthe air–water interface.

Page 74: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Soils 65

B. SOIL

In addition to the factors considered for water, we need to consider for soil: (a) thefar greater importance of interactions with solid surfaces and the buffering of ionsin solution by ions adsorbed on the surfaces; and (b) the more-strongly reducingconditions that develop in soil because of the greater sink for O2, resulting intransformations of soil surfaces as well as of species in solution.

Figure 3.7 shows the concentrations of cations in solution following submer-gence of four representative rice soils; the corresponding changes in EH, pH,HCO3

−, CEC and soil Fe are shown in Figures 2.6 and 2.7. The main anionin solution is HCO3

− derived from CO2 together with Cl−: any NO3− ions

present before submergence are rapidly consumed in reduction and consequentlytheir concentration is generally negligible after the initial stages, and SO4

2− ionsare also reduced though more slowly. These unadsorbed anions determine theoverall strength of the soil solution and balance cations derived from exchange,dissolution and redox reactions involving the soil solid.

Redox processes are discussed in detail in Chapter 4. The rest of this chapterdeals with solid–solution interactions, firstly for soils in general and then forsubmerged soils. Recent reviews of solid–solution interactions in soils includeSposito (1994), Sparks (2003) and the relevant chapters of Sumner (2000).

3.5 THE SOLID SURFACES IN SOILS

The main surfaces with which ions interact are clay-sized particles of layer sili-cates and Al, Fe and Mn oxides, and organic matter bound to clay particles. Theirinteractions with ions depend on their functional groups. These are analogous tothe functional groups on molecules in solution but differ in that they are held afixed distance apart and their charge characteristics may be strongly influencedby the neighbouring functional groups.

Layer Silicates

The layer silicates comprise tetrahedral sheets of silica and octahedral sheets ofaluminium and magnesium hydroxide, with varying amounts of the Si, Al and Mgreplaced by cations of lower valence giving the lattice a net negative charge. Twobasic combinations occur: 1 tetrahedral sheet with 1 octahedral (e.g. kaolinite,halloysite), and 2 tetrahedral with 1 octahedral (e.g. smectite, vermiculite, illite).

1:1 Type. The tetrahedral and octahedral sheets are bound together because theyshare an oxygen atom, and the resulting layers are bound together by hydrogenbonds between the oxygens of the tetrahedral sheet and the hydroxyls of theadjacent octahedral sheet. The structure is therefore rigid. Pure kaolinite, formula

Page 75: The Biogeochemistry of Submerged Soils

66 Interchange of Solutes between Solid, Liquid and Gas Phases

0

1

2

3

4

5

6

7Iloilo

Con

cent

ratio

n in

soi

l sol

utio

n (m

mol

c L−1

)

0

2

4

6

8

10

12

14

16

0

5

10

15

20

Time after flooding (days)

0 20 40 60 800

1

2

3

4

5

6

Tarlac

Ca2+

Fe2+

NH4+

Mg2+

Na+

K+

Ca2+

Mg2+

Fe2+

Ca2+

Mg2+

Fe2+

Mn2+

Mg2+

Ca2+

Fe2+

Na+

Na+

Na+

Maahas

Nueva Ecija

Figure 3.7 Changes in the concentrations of cations in solution following flooding offour rice soils (Kirk et al., 2003). The corresponding changes in EH, pH, HCO3

−, CECand soil Fe are shown in Figures 2.6 and 2.7. Reproduced by permission of BlackwellPublishing

Page 76: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Soils 67

Table 3.9 Structural charge and surface area of layer silicates

Group Layerstructure

Typical formulaa Structuralnegativecharge

(molc kg−1)

Specificsurface

area(m2 kg−1)

Kaolinite 1:1 [Si4]Al4O10(OH)8 0–0.01 0.5–3

Halloysite 1:1 [Si4]Al4O10(OH)8·4H2O 0–0.01 1–4.5

Illite 2:1 M1.4 – 2[Si6.8Al1.2]Al3Fe0.25 Mg0.75O20(OH)4 1.9–2.8 8–15

Vermiculite 2:1 M1.2 – 1.8[Si7Al]Al3Fe0.5 Mg0.5O20(OH)4 1.6–2.5 30–50

Smectite 2:1 M0.5 – 1.2[Si8]Al3.2Fe0.2 Mg0.6O 20(OH)4 0.7–1.7 60–80

Chlorite 2:1:1 (Al(OH)2.55)4[Si6.8Al1.2]Al3.4 Mg0.6O20(OH)4 Variable 2.5–15

a M = monovalent cation.

Si4Al4O10(OH)8, has no structural charge, but moderate structural charge (< −10 mmolc kg−1) arises in most natural kaolinites as a result of substitution ofMg2+ for Al3+ or Al3+ for Si4+ (Table 3.9). More important is the pH-dependentcharge that arises as a result of adsorption and desorption of protons by the –OHand –O groups at the lattice edges. This is in the range +10 to −50 mmolc kg−1

over the usual range of soil pHs, with zero charge at about pH 4.6.

2:1 Type. The tetrahedral sheets of adjacent layers cannot form hydrogen bondswith each other, lacking –O groups, but are held together by electrostatic forcesarising from their charge. In illites the interaction is strong because the latticecharge is neutralized by K+ ions, which effectively glue the sheets together. Insmectites and vermiculites water molecules and solutes can enter between thelayers, increasing the layer spacing.

Swelling clay minerals have a moderate proportion of substituted atoms andhence a relatively low charge density (Table 3.9). The cations that compensatethe permanent charge are therefore bound to the surface by relatively weak purelyelectrostatic forces. They form a broad diffuse layer in the solution near the sur-face. Non-swelling 2:1 clay minerals have a much larger proportion of substitutedatoms and therefore greater charge density (Table 3.9). As a result the charge-compensating cations may form covalent or ionic bonds with the surface, as wellas being held electrostatically. Certain ions, e.g. K+ and NH4

+, fit neatly in holesin the clay structure and produce a rigid stack of clay layers. Only the cationsheld on outer surfaces and, to a variable extent, on imperfectly fitting layerswithin the clay structure are then exchangeable with cations in the soil solution.

2:1:1 Type. The interlayer spaces of 2:1 silicates may be blocked by poorlyordered sheets of Al hydroxy polymers, such as [Al(OH)2.5

0.5+]n(n ≥ 6). SuchAl interlayers neutralize a considerable part of the surface charge and restrictswelling, and effectively convert 2:1 clays into materials similar to kaolinite.

Page 77: The Biogeochemistry of Submerged Soils

68 Interchange of Solutes between Solid, Liquid and Gas Phases

Even moderate Al interlayering greatly affects clay properties. Such materialsare important in submerged soils that have become ferrolysed (Section 7.1).

Amorphous Aluminosilicates. These occur in soils influenced by volcanic activ-ity and are associated with very high moisture retention and anion fixation, andlow to very high pH-dependent CEC. They may also bind organic matter tightly,protecting it against decomposition. Examples are allophane and imogolite.

Oxides

Being widespread in the lithosphere and insoluble in the usual range of soilpH, oxides and hydroxides of Al, Fe and Mn are common in soil clays. Redor yellow coloration of soils is apparent at Fe oxide contents of only 0.1 % orless, especially if the Fe is amorphous and coats other minerals. The most visiblechange occurring when soils are submerged is the conversion of the red andyellow compounds of Fe(III) to the bluish-grey compounds of Fe(II).

Metal oxides and hydroxides have little or no structural charge but developpH-dependent charge as the hydroxyl groups at the lattice edges gain or loseprotons. The surface charge is a function of both pH and the concentration ofsalts in the solution, as these affect the dissociation of the –OH groups. How-ever the pH at which the surface negative charge is equal to the surface positivecharge—the point of zero charge (pzc)—is independent of the salt concentrationif the salt does not react with the surface. The point of zero charge is an impor-tant characteristic of the surface. Table 3.10 gives pzc values for common soilmaterials. Values are large for metal oxides and hydroxides but small for silicaand soil organic matter.

In real soils where oxides, layer silicates, organic matter and other materialsare present in intimate mixtures, with the oxides and organic matter often coatingthe surfaces of the other materials, the different functional groups interact with

Table 3.10 Points of zero charge (pzc)of oxides and aluminosilicates

Material pzc

α-Al(OH)3 5.0γ -AlOOH 8.2α-FeOOH 7.8γ -Fe2O3 6.7Amorphous Fe(OH)3 8.5MgO 12.4δ-MnO2 7.2SiO2 2.0Feldspars 2–2.4Kaolinite 4.6Montmorillonite 2.5

Page 78: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Submerged Soils 69

each other and so the distinction between permanent and pH-dependent chargeis blurred.

Organic Matter

Soil organic matter is a weak acid and becomes negatively charged by losing pro-tons. The main functional groups are carboxylates and, to a lesser extent, phenols,which are weaker acids. Their acid–base behaviour is complicated because oftheir heterogeneity and because of the effects of neighbouring functional groupson soil surfaces. With increasing dissociation, the build up of negative chargeon the surface tends to inhibit further dissociation. Thus a plot of the extent ofdissociation of organic functional groups versus pH tends to be steeper than theequivalent plot for simple monoprotic acids, and it approaches a straight line overthe usual pH range in soils. This leads to the following rough empirical relationfor the negative charge on soil organic matter as a function of pH (McBride,1994):

SOM charge(mmolc g−1organic C) = −0.6 + 0.5 pH (3.41)

As a rule of thumb, at near neutral pH, each g of organic C per kg of soilincreases the surface negative charge by about 3 cmolc kg−1 soil.

A further complication is that soil organic matter becomes more soluble athigher pH as dissociation increases the surface negative charge. Also, organicmatter may form coordination complexes with some metals involving cova-lent bonds.

3.6 THE SOLID SURFACES IN SUBMERGED SOILS

Many submerged soils are developed in recent in alluvium and are often youngor only weakly weathered (Section 1.3). The overall composition of the clayfraction is therefore often close to that of the parent sediment. Hence the followinggeneralizations can be made for rice soils in the humid tropical lowlands (Kyuma,1978; Binkman, 1985)

• Soils derived from marine alluvial sediments tend to be dominated by mont-morillonitic 2:1 clays whereas those from riverine sediments have vermiculitic2:1 clays with mixtures of 1:1 clays and metal oxides, the sediment beingdeveloped under more strongly weathered conditions.

• Soils developed in positions higher in the landscape tend to be dominated bymore-weathered material.

• Soils derived from basic volcanic ejecta, metamorphic rocks and granitic rockshave corresponding mineralogies.

Various changes in mineralogy are induced by seasonal flooding. The firstfactor in this is the change in base status of the soil due to the flow of water

Page 79: The Biogeochemistry of Submerged Soils

70 Interchange of Solutes between Solid, Liquid and Gas Phases

through and across it. Depending on the alkalinity of the water entering andleaving, the soil may be enriched with bases or depleted. Large quantities ofbases are liberated in soil reduction following flooding and may be leached.The balance will depend on the topological and hydrological situation of thesoil, and in general soils low in the landscape will accumulate bases and thosehigher will be depleted. There may be biological fixation of bases, for exampleby aquatic snails in fields receiving base-rich interflow water or irrigation. Forexample, after 15 years of intensive irrigation of ricefields at the International RiceResearch Institute (Laguna, Philippines) with base-rich water (4 mmolc L−1Na+and 1 mmolc L−1 Mg2+), there was a marked accumulation of CaCO3 in snailshells and the aerobic soil pH increased from 5.6–6.0 to 6.5–7.0 (Moormann andvan Breemen, 1978). The reverse process—decalcification—may also occur. Forexample, in the calcareous soils of the Ganges and Megha sediments, Bangladesh,where the ricefields are rainfed and the rainwater tends to be acid, Brammer(1971) reported losses of 1 % of the CaCO3 in 25 years from sediments thatoriginally contained 5–10 % CaCO3. The calcite is dissolved by acids in the rainand CO2 formed during soil reduction, and Ca(HCO3)2 is leached out of the soil.

The second factor is the changing redox state of the soil. Generally iron isthe most abundant redox species present. Table 3.11 shows total iron contentsof a range of rice soils across Asia. From 20 to 80 % of the iron is presentas free Fe(III) oxides and often from 1 to 20 %—and sometimes as much as90 %—of the free Fe(III) is reduced to soluble and exchangeable Fe(II) fol-lowing submergence (see for example Figure 2.8). Some of the exchangeableFe(II) is subsequently reprecipitated as mixed Fe(II)Fe(III) compounds of uncer-tain composition. There may also be reduction of structural Fe(III) to Fe(II)within clay minerals. These changes take place over a matter of weeks. Uponsubsequent drying and re-oxidation, the exchangeable and amorphous Fe(II) arerapidly converted to ferric hydroxides, initially in amorphous forms that recrys-tallize only very slowly (Figure 3.8). As a result, amorphous ferric hydroxidesand similar materials tend to accumulate in the soil at the expense of more stable

Table 3.11 Total iron contents (mg Fe g−1) of rice soils

Country Mean Min. Max. n

Bangladesh 40.1 8.0 66.1 53Burma 39.7 5.9 65.5 16Cambodia 32.2 0.0 80.1 16India 72.3 9.0 117.6 73Indonesia 20.8 1.7 50.3 44Malaysia 20.8 1.9 50.3 41Philippines 54.1 28.8 86.7 54Sri Lanka 37.4 5.0 149.6 33Thailand 25.2 0.0 90.0 80Japan 42.0 — — 155

Source: Kyuma (1978).

Page 80: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Submerged Soils 71

Amorphous Fe(III) oxidesCrystalline Fe(III) oxides Fe(II)

Slow reduction

Fast reduction

Fast oxidation

Slow crystallization

Figure 3.8 Accumulation of amorphous Fe(III) compounds under alternating reductionand oxidation (after Moormann and van Breemen, 1978). Reproduced by permissionof IRRI

ferric oxides. In turn, the amorphous materials are more easily reduced duringsoil flooding and over time the iron compounds reach a steady state in which eas-ily reducible amorphous materials are combined with more recalcitrant minerals.The proportion of amorphous and crystalline materials will be influenced by thehydrological regime and climate. Concomitantly there are short- and long-termchanges in soil organic matter.

Changes in Surface Properties Following Submergence

For the most part, the overall composition of the clay fraction in soils is deter-mined by long-term processes and does not change rapidly with changes inconditions. However the properties of the clay surface can change rapidly andthe surface is often not in equilibrium with the rest of the solid phase. Thusalternating reduction and oxidation under variable water regimes cause rapid buttransient changes in surface properties, as well as more persistent changes inthe overall composition of the solid phase. The changes in surface propertiesfollowing soil submergence are as follows.

Dissolution of Oxide Coatings. The net negative charge on the soil solid mayincrease following reduction as a result of dissolution of oxide and oxyhydroxidecoatings on clay surfaces. The pzc of oxides and oxyhydroxide are at or aboveneutral pH (Table 3.10), and so the coatings on clays are positively charged inneutral and acid soils and neutralize some of the clay’s negative charge. Theirdissolution therefore results in an increase in the net negative charge on thesurface. Hence, for example, an oxide with composition Fe(OH)2.5

0.5+, is reducedaccording to the half reaction

[soil—2Fe(OH)2.5] + 5H+ + 2e− −−−→ [soil—]− + 2Fe2+ + 5H2O (3.42)

where e− represents an electron transferred in the reduction. If the correspondingoxidation of soil organic matter is (Chapter 4)

CH2O + H2O −−−→ CO2 + 4H+ + 4e− (3.43)

Page 81: The Biogeochemistry of Submerged Soils

72 Interchange of Solutes between Solid, Liquid and Gas Phases

then the overall reaction is

2[soil—2Fe(OH)2.5] + CH2O + 6H+ −−−→ 2[soil—]− + 4Fe2+ + CO2 + 9H2O(3.44)

In Reaction (3.44), for each mol of Fe reduced the surface negative chargeincreases by 0.5 molc and 1.5 mol of H+ are consumed.

Roth et al. (1969) found increases in surface negative charge equivalent to10–60 % of the initial charge for a range of soils and soil clays. The changecould be attributed quantitatively to the removal of the positively charged oxidecoatings and was reversed by re-oxidizing the samples. Changes in charge withreductive dissolution of oxides have been demonstrated using chemical reduc-ing agents (Roth et al., 1969) and microbial reducing agents (Bloomfield, 1951;Ottow, 1973; Lovley, 1991), and under field conditions (Favre et al., 2002).

Dissolution and reduction of crystalline Fe(III) minerals is accelerated by chela-tion with carboxylate ligands in the presence of Fe(II) (Zinder et al., 1986; Blesaet al., 1987; Phillips et al., 1993; Kostka and Luther, 1994). Therefore as soilreduction proceeds and carboxylates formed in oxidation of organic matter accu-mulate in solution together with Fe2+, dissolution and reduction of crystallineFe(III) will commence. Dissolution of oxyhydroxide coatings will therefore lagbehind the initial reduction of Fe(III).

Reduction of Structural Fe. There may also be changes in charge due to reduc-tion of structural Fe(III). Virtually all soil clay minerals contain some iron intheir crystal structures and reduction of this structural Fe by chemical or micro-bial reducing agents, with the iron remaining octahedrally coordinated in the claystructure, is well documented (Stucki, 1988; Stucki et al., 1997). The extent ofreduction, whether by microbes or chemical reducting agents, can be as muchas 90 % of the octahedral Fe(III) in a few days (Kostka et al., 1999). The rateis enhanced by the presence of organic chelating agents that commonly occur insediments and flooded soil solutions, and under such conditions Fe(III) reductionmay lead to partial dissolution of the clay (Kostka et al., 1999).

As structural Fe(III) is reduced, the negative charge on the clay will increase.It is found experimentally that the increase in negative charge is not directlyequivalent to the amount of Fe(III) reduced, and the more reduced the clay is thesmaller is the change in charge. An example is shown in Figure 3.9. The mech-anism behind this is uncertain but involves dehydroxylation of the clay structureduring reduction and sorption of metal cations from the solution (Stucki et al.,1997; Drits and Manceau, 2000). The extent of dehydroxylation and sorptionvaries with the extent of reduction, hence the change is nonlinearly related to theamount of Fe reduced. For example, for a nontronite:

M[Si7Al]Fe(III)4O20(OH)4 + mM+ + nH+ + pe− −−−→M1+m[Si7Al]Fe(III)4−pFe(III)pO20(OH)4−n + nH2O

(3.45)

Page 82: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Submerged Soils 73

Amount of Fe(III) reduced (mmol g−1)

0.0 0.5 1.0 1.5 2.0

Sur

face

cha

rge

(mm

olc

g−1)

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

Figure 3.9 Relation between surface charge and reduction of structural Fe in a diocta-hedral smectite. Points are experimental data; lines are theoretical relations discussed inthe text (Drits and Manceau, 2000). Reproduced by permission of Clay Minerals Society

where M is a sorbed cation and m, n and p are coefficients. The solid line inFigure 3.9 shows the fit to Equation (3.45) and the dotted line shows the expectedrelationship if only dehydroxylation occurs. If the generic reaction is simplified to

[clay—Fe(III)OH] + nH+ + e− −−−→ [clay—Fe(II)OH1−n](1−n)− + nH2O(3.46)

then using the upper value n = 0.75 (Figure 3.9), the complete reaction withsimultaneous oxidation of organic matter (as for Reaction 3.44) is:

4[clay—Fe(III)OH] + CH2O −−−→ 4[clay–Fe(II)OH0.25]0.25−

+ H+ + CO2 + H2O (3.47)

In Reaction (3.47), for each mol of Fe reduced the surface negative chargeincreases by 0.25 molc and 0.25 mol of H+ are released.

For moderate reduction the changes are completely reversible but they areprogressively less so with more thorough reduction (Stucki et al., 1984; Komadelet al., 1995; Gates et al., 1996). There are concomitant changes in the clay’sphysical and chemical properties, including its surface area, swelling behaviour,and capacity to sorb cations.

Changes in pH-dependent Charge. Changes in pH with soil reduction will causechanges in the charges on inorganic –OH functional groups and organic mat-ter. From Equation (3.41), the increase due to organic functional groups willbe approximately 0.5 mmolc g−1 organic C per unit pH increase, or, for a soilwith 1 % organic C, 5 mmolc kg−1 soil. This is small compared with the changesdue to dissolution of oxide coatings and reduction of structural Fe, which are ofthe order of several tens of mmolc kg−1 soil. But it may be important in highly

Page 83: The Biogeochemistry of Submerged Soils

74 Interchange of Solutes between Solid, Liquid and Gas Phases

weathered soils with low cation exchange capacity. The changes due to edge –OHgroups on kaolinites will be of a similar magnitude. The increase in oxide chargeper unit increase in pH will be of the order of 5 mmolc kg−1 oxide at the ionicstrength and pH of typical flooded soil solutions. However, if the surface oxidecoatings dissolve in the process of reduction, this will be of no consequence.

Formation of New Solid Phases. Once a sufficient concentration of dissolvedconstituents has been reached following submergence and soil reduction, newsolid phases will precipitate. The nature of these compounds is discussed indetail in Chapter 4.

In neutral soils with smectite or vermiculite in the clay fractions, the changesin redox and bases status following soil flooding may cause synthesis of materialssimilar to smectite with Fe2+ in the octahedral sheet. Other cations, e.g. Zn2+,may also become entrapped.

In acids soils, particularly those with kaolinite clay minerals, soluble Fe2+concentrations tend to rise to high levels because of low CEC and because con-ditions do not favour precipitation of Fe(II) oxides or carbonates or synthesisof silicates.

When a reduced soil is re-oxidized, Fe2+ changes into Fe(OH)3. The originalFe oxides are thus distributed differently, generally with a higher specific surfaceand activity. In high-activity clay soils, this may increase the stability of thestructure established just before flooding. In low-activity clay soils the effectsof alternate reduction and oxidation are less clearly beneficial, partly because ofleaching of nutrients.

3.6.1 ORGANIC MATTER IN SUBMERGED SOILS

In general the organic matter in soils tends towards a steady state in whichadditions from net primary production balance the various processes removingit, and the organic matter attains a level and composition characteristic of theprevailing conditions. As discussed in Chapter 1, net primary productivity ofwetlands is in general far greater than that of uplands and rates of organic matterloss due to decomposition, run-off, leaching and erosion tend to be less. Ingeneral, the yield of energy per unit of carbon oxidized is smaller in anaerobicfermentation than in aerobic respiration, and so, other things being equal, ratesof decomposition are less. Hence in the temperate zone wetland soils often havelarge organic matter contents or are peaty. However, in the tropics, peat bogsand fens are rare (Table 1.1) and wetland rice soils tend not to have particularlylarge organic matter contents.

In data assembled by Greenland (1997), the mean level of organic carbon inthe topsoils of wetland rice soils from across tropical Asia was 2 %, and afterexcluding acid peaty soils the mean was 1 %. This compares with a range of1.27–1.81 % for Oxisols and Ultisols of the Cerrado region of Brazil (Sanchez,

Page 84: The Biogeochemistry of Submerged Soils

The Solid Surfaces in Submerged Soils 75

1981) and 2.78–4.80 for Oxisols and Ultisols of the humid tropical forest zoneof Sumatra (van Noordwijk et al., 1997). Under intensive multiple-rice cropping,the soil organic mater may increase until a new steady state level is reached aftersome years. Table 3.12 gives an example for a rice–rice system over five cropsin 2 years compared with a maize–rice system. However, very large soil organicmatter contents do not develop.

Evidently rates of decomposition are greater than expected for simple anaerobicdecomposition. Figure 3.10 shows comparable rates of organic matter decompo-sition in soils that were kept continuously flooded or well-drained under otherwisesimilar, tropical conditions for 3–4 years (Neue and Scharpenseel, 1987). Clearly,

Table 3.12 Carbon balances for rice–rice and maize–rice cropping systems over fiveconsecutive crops in 3 years at IRRI, Laguna, Philippines. Except for short stubble allstraw was removed from the fields and no organic manures were applied. Values are t Cha−1± SEs

Cropping system Rice–rice Maize–rice

N fertilizer (kg ha−1) 0–0 190–100 0–0 190–100

Initial SOCa 19.13 ± 0.83 19.41 ± 0.27 19.22 ± 0.79 19.38 ± 0.97Final SOC 20.97 ± 0.49 22.15 ± 0.54 19.01 ± 0.40 19.83 ± 0.43Change in SOC +1.84 ± 0.44 +2.74 ± 0.67 −0.22 ± 0.50 +0.46 ± 0.96(% change) (+10) (+14) (−1) (+2)C from crop residues 4.93 ± 0.18 7.72 ± 0.21 4.15 ± 0.28 7.09 ± 0.18C mineralized 3.09 ± 0.56 4.98 ± 0.47 4.37 ± 0.56 6.63 ± 0.87(% of crop residues) (63) (64) (105) (94)

a SOC, soil organic carbon.Source: Witt et al. (2000). Reproduced with kind permission of Kluwer Academic Publishers.

Time (months after straw incorporation)

0 10 20 30 40 50 0 10 20 30 40 50

14C

rem

aini

ng (

%)

10

100

TropeptTropeptHumult

10

100

AqueptAquollAquult

(a) well drained (b) submerged

30

3

30

3

Figure 3.10 Decomposition of 14C-labelled rice straw in (a) well-drained uplandsoils and (b) continuously submerged lowland soils under tropical conditions (adaptedfrom Neue and Scharpenseel, 1987). Reprinted with permission from Elsevier

Page 85: The Biogeochemistry of Submerged Soils

76 Interchange of Solutes between Solid, Liquid and Gas Phases

other factors are at work. In addition to temperature and aeration, and the quan-tity and nature of organic matter inputs, other factors influencing decompositioninclude the soil pH, which may be more favourable following submergence; thelevel and balance of nutrients, which may also be more favourable; and thecommunities of micro- and macro-fauna that together bring about the decom-position. Submerged soils are never wholly anoxic and contain aerated zonesat the interfaces between the soil and floodwater and the soil and plant roots.Burrowing oligochaete worms transport fresh and partially decomposed organicmatter between the soil and floodwater, and the activities of organisms in the soiland floodwater are thereby linked. Hence the soil–floodwater system as a wholebehaves quite differently from its component parts in isolation. Decompositionprocesses are discussed further in Chapter 5.

Studies by Olk and others of long-term (≤ 30 years) changes in organic matterin soils under intensive wetland rice cultivation have shown a gradual accumula-tion of less humified and more phenolic material (Olk et al., 1996, 1998, 1999;Mahieu et al., 2000a, b, 2002). These authors compared the chemical composi-tion of organic matter from four soils with different histories of cropping andsubmergence: (a) one crop of upland rice annually without soil submergence;(b) one wetland rice crop and one soybean crop annually; (c) two wetland ricecrops annually; and (d) three wetland rice crops annually. With increasing inten-sity of wetland rice cropping there were large increases in the proportions ofless humified material in the soil organic matter, measured as diester P, amideN and phenolic C in nuclear magnetic resonance spectra. There were also posi-tive correlations with visible light absorption and concentrations of free radicals,both of which are indices of humification, and negative correlations with theconcentration of H, a negative index of humification. The authors speculate thatslower lignin decomposition under restricted O2 supply in submerged soil leadsto incorporation of phenolic compounds into young soil organic matter as it isturned over. Since phenolic compounds can react strongly with nitrogenous com-pounds, they further speculate that the mineralization and immobilization of Nin intensively cropped rice soils may be adversely affected by accumulation ofphenolic material.

3.7 SOLID–SOLUTION INTERACTIONS

3.7.1 ADSORPTION

Adsorption depends on the interaction of ions and uncharged solutes with thefunctional groups on soil surfaces. It is in some ways analogous to complexationof ions with ligands in solution (Section 3.1), with the difference that the surfacefunctional groups are stationary and their properties depend to a greater extenton interactions with neighbouring groups. Two types of surface complex aredistinguished:

Page 86: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 77

(a) inner-sphere complexes in which the adsorbed species is bound directly tothe surface functional group, with no intervening water molecules; and

(b) outer-sphere complexes in which at least one water molecule remains betweenthe absorbed species and the surface.

In inner-sphere complexes the bonding is covalent or ionic and the reactivity ofthe surface is altered by the interaction; in outer-sphere complexes the bondingis largely electrostatic and the reactivity of the surface is largely unaltered. Inner-sphere complexation is usually slower than outer-sphere and is often not readilyreversible. It can occur regardless of the net surface charge and is little influencedby the ionic strength of the external solution. If an ion is adsorbed without forminga surface complex, neutralizing surface charge in only a delocalized way, thenit is said to be part of the diffuse ion swarm. Such ions are free to move aboutin the soil solution near the surface. Figure 3.11 shows surface complexes on aninorganic hydroxyl surface. It shows the distinction between inner- and outer-sphere complexes, depending on the presence of water molecules between thesurface and complexed species. The region of the diffuse ion swarm begins atthe outer edge of the water molecules solvating ions in outer-sphere complexes.

There are a number of more loosely defined terms for different types of adsorp-tion that are related to the form of surface complexation. Specifically adsorbedions are held in inner-sphere complexes whereas non-specifically adsorbed ionsare in outer-sphere complexes or the diffuse-ion swarm. Readily exchangeable

Watermolecules

Oxygen

Metal

H+

X H

M

H

H

Inner-spherecomplexes

e.g. M = Zn, Pb, Cd X = P, As, Si

Outer-spherecomplexes

e.g. M = K, Ca, Mg, Fe X = Cl, NO3, HCO3

X−

M+

Solid−waterinterface

Figure 3.11 Complexes formed between solutes and hydroxyl groups on oxides andlayer silicate edges (adapted from Sposito, 1984b). Reproduced by permission of OxfordUniversity Press

Page 87: The Biogeochemistry of Submerged Soils

78 Interchange of Solutes between Solid, Liquid and Gas Phases

ions are those that can be replaced easily by leaching with an electrolyte solu-tion. This is an empirical definition, but only fully solvated ions can be readilyexchangeable and therefore must be either in the diffuse-ion swarm or in outer-sphere complexes.

Adsorption interacts strongly with complexation in solution. Table 3.13 indi-cates the range of complexes between metal ions and inorganic and organicligands in soil solutions. In a submerged soil the organic ligands present includeacetate, formate and propionate at concentrations of 10–40 mM in the early stagesfollowing submergence though less than 1 mM after 3–4 weeks. In addition con-centrations of amino acids, phenolic acids and larger molecular weight humicacids may reach a few hundred µM, though transiently. Figure 3.12 shows thecalculated effects of realistic concentrations of acetate, formate, propionate, glu-tamate, glycine, benzoate and phenylacetate on Fe(II), Mn(II) and Zn(II) species.The figure shows that for Fe(II) and Mn(II) the free ion dominates at all pHs,except for Fe above pH 9 where hydroxy complexes are important. Complexeswith acetate are also significant at pHs above about 5, and FeHCO3

+ abovepH 6 and MnGlu above pH 5. Complexes with formate, propionate or eitherof the phenolic acids are unimportant at all pHs. The picture is more compli-cated for Zn(II) with many more significant species. The free ion dominates atpH ≤ 7.5 but complexes with acetate, HCO3

−, glutamate and especially CO32−

are important at various pHs. Hydroxy complexes are only important at pH>9.Figure 3.13 shows the solubility of Zn2+ in soil at four Zn levels and different

pHs. The figure shows that the soil solution is under-saturated with respect tolikely pure Zn precipitates up to high pHs, and there is a marked minimumin solubility at near neutral pH. The explanation involves cation exchange andspecific adsorption reactions, trace amounts of Zn2+ being sorbed preferentiallyover the main exchanging cations, and complexation reactions between Zn2+ andorganic ligands in solution. The negative charge on soil surfaces increases as thepH increases, tending to increase sorption of Zn2+ on variable-charge surfaces.But at near neutral pH the concentration of dissolved organic matter in solution

Table 3.13 The main species of trace metals in soil solutions

Metal Acid soils Alkaline soils

Mn(II) Mn2+, MnSO40, Orga Mn2+, MnSO4

0, MnCO30, MnHCO3

+

Fe(II) Fe2+, FeSO40, FeH2PO4

+ FeCO30, Fe2+, FeHCO3

+, FeSO40

Ni(II) Ni2+, NiSO40, NiHCO3

+, Org NiHCO30, NiHCO3

+, Ni2+

Cu(II) Org, Cu2+ CuCO30, Org

Zn(II) Zn2+, ZnSO4 ZnHCO3+, ZnCO3

0, Zn2+, ZnSO4

Cd(II) Cd2+, CdSO40, CdCl+ Cd2+, CdCl+, CdSO4

0, CdHCO3+

Pb(II) Pb2+, Org, PbSO40, PbHCO3

+ PbCO30, PbHCO3

+, Pb(CO3)22−, PbOH+

a Org, organic complexes, e.g. with fulvic acids.Source: adapted from Sposito (1983). Reproduced by permission of Elsevier.

Page 88: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 79

pH

4 5 76 8 9 1010−11

10−10

10−9

10−8

10−7

Zn2+ZnHCO3

+ ZnCO3

ZnGlu

Zn(OH)2 Zn(OH)3−ZnGly

ZnOH+

ZnForm

ZnAc

ZnProp

Con

cent

ratio

n (M

)

10−7

10−6

10−5

10−4

10−3

Mn2+

MnOH+

MnAcMnGlu

10−6

10−5

10−4

10−3

10−2

Fe2+

FeAc

FeGly

FeHCO3+

Fe(OH)3−

Fe(OH)2

FeOH+

MnHCO3+

Figure 3.12 Distributions of Fe(II), Mn(II) and Zn(II) species in a simulated flooded soilsolution. Total concentrations in mM are Fe(II) = 2.5, Mn(II) = 0.25, Zn(II) = 0.0001,CT = 20, acetate = 10, formate = 1, propionate = 1, glutamate = 0.1, glycine = 0.1,benzoate = 0.1, phenylacetate = 0.1. Species accounting for <1 % of total not shown

increases appreciably and so complexation of Zn2+ in solution increases. Thereis also the possibility of co-precipitation in solid solutions that are much lesssoluble than pure precipitates (see below).

3.7.2 PRECIPITATION

Adsorption is the process by which a net accumulation of a substance occursat the boundary between two phases (e.g. gas—solid as well as liquid—solid),

Page 89: The Biogeochemistry of Submerged Soils

80 Interchange of Solutes between Solid, Liquid and Gas Phases

pH

3 54 6 7 8 9 10

pZn

3

4

5

6

7

Zn (O

H)2 (am

orphous)

ZnC

O3

10 ppm Zn 20 ppm Zn40 ppm Zn

70 ppm Zn

Zn5 (O

H)6 (C

O3 )2

Zn(OH

)2

Figure 3.13 Solubility of Zn2+ versus pH in soil at four Zn levels. The straight linesindicate the solubilities of possible precipitates at atmospheric CO2 pressure (McBride,1994). Reproduced by permission of Oxford University Press

whereas precipitation is the process by which a substance accumulates to forma new solid phase. Both imply a net removal of solute from solution, but oneis inherently two-dimensional and the other three-dimensional. However the dis-tinction is blurred because similar types of chemical bond are often involved.Figure 3.14 shows the formation of a precipitate on the surface of an oxide orlayer silicate in an analogous way to the formation of inner-sphere complexesin Figure 3.11. In practice there is a continuum between the two ranging fromextremely insoluble inner-sphere complexes to precipitates that are much moresoluble. In general when the concentration of a sorbed species is small sur-face complexation dominates and when it is large precipitation dominates. Insubmerged soils very large concentrations of dissolved ions and gases developfollowing anaerobic metabolism and reductive dissolution of solid phases, andso precipitation reactions often dominate.

Precipitation is generally much slower than adsorption. Table 3.14 comparesrates of precipitation with rates of adsorption and other surface phenomena insoil systems. Rates vary greatly between precipitating compounds. They are alsooften subject to inhibition and catalysis by other solutes and solid phases present.

Page 90: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 81

Oxygen

Metal

H

Solid−waterinterface

H

H H

Precipitatedmetal

Figure 3.14 Precipitation on the surface of an oxide or edge of a layer silicate (cf.Fig. 3.11)

Table 3.14 Rates of solid-solution interactions in soils

Process Time scale(h)

Complexation in solution 10−8 –10Adsorption 10−8 –10Desorption 10−6 –104

Dissolution 1–104

Redox reactions in solution 10−6 –108

Redox reactions on surfaces 0.1–104

Solid solution formation 0.1–10Cluster formation 10−2 –1Heterogeneous nucleation and precipitation 1–102

Homogeneous nucleation and precipitation 0.1–104

Recrystallization into pure phases 104 –108

Diffusion on surfaces 10–108

Diffusion in crystals 105 –108

Thus rate laws for precipitation reactions tend to be complicated, even in puresolutions. Mixed precipitates can be inhomogeneous solids with one componentrestricted to a thin outer layer because of slow diffusion. New solid phases canprecipitate homogeneously onto the surfaces of existing solid phases. Weatheringsolids may provide host surfaces onto which more stable phases may precipitate.

At least three potentially rate-limiting steps can be distinguished: the diffusionof the reacting solutes to the site of precipitation; their reaction to form theinsoluble compound; and accumulation of the compound as a solid phase. In

Page 91: The Biogeochemistry of Submerged Soils

82 Interchange of Solutes between Solid, Liquid and Gas Phases

general, in water systems rates of diffusion are much greater than the rates ofreaction and are rarely rate limiting. However in soils diffusion may be muchslower, for example where a reaction is catalysed by sorption on inaccessiblesurfaces, and so diffusion is often rate limiting. The formation of the solid phasethen has three stages. First a critical cluster of the constituent solutes must formfrom a supersaturated solution. Spontaneous crystal growth may then occur. Ineffect the solubility of the nuclei that are initially formed is greater than thatof the larger crystallites that grow from them. The difference arises from themuch greater surface energy of the nuclei compared with the crystallites. Inthe third stage large crystals form slowly from crystallites through the processknown as ripening. Nucleation may occur homogeneously from a pure solutionthat is sufficiently supersaturated. But in soil systems it is more likely to occurheterogeneously with the surfaces of soil particles acting as catalysts. If thesurface matches the precipitating phase well, the interfacial energy between thetwo solids is less than that between the precipitate and the solution, and theenergy barrier for nucleation is therefore decreased. Consequently the degreeof supersaturation necessary for precipitation to start is smaller. For example,soil solutions are often highly supersaturated with respect to gibbsite, but it willprecipitate from similar solutions containing smectite.

The particular combinations of ions and molecules that will form precipitatesin a given solution can be predicted from equilibrium thermodynamics. However,this often gives a misleading picture because there are kinetic limitations or thereis inhibition, particularly in soil solutions. There may also be isomorphous sub-stitution of one cation for another in the precipitate, resulting in a solid solutionwith a different solubility to the pure compound.

3.7.3 CO-PRECIPITATION IN SOLID SOLUTIONS

In general the solid phases formed under natural conditions are not simple purecompounds but contain foreign ions isomorphously substituted in crystal lattices.The activity of the solid phase is thereby decreased. This may have little effecton the solubility of the major component, but it may greatly decrease the solu-bility of the minor component compared with its pure form. This process differsfrom adsorption or occlusion in that it represents the equilibrium state of the sys-tem. A requirement is that the foreign ion can diffuse freely during precipitationand that there is a high structural compatibility between the two pure phases.Hence the radii of the ions involved must be similar and the minor compo-nent should become uniformly distributed through the solid. As discussed earlierrates of precipitation tend to be slower than rates of adsorption; rates of solidsolution formation are intermediate between rates of adsorption and precipita-tion (Table 3.14). Solid solutions seem likely to form in submerged soils becausereductive dissolution reactions generate large concentrations of dissolved ionsover short times, and so there are opportunities for mixed precipitates to form.

Page 92: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 83

Consider a two-phase system with components AY and BY in which some ofthe BY(s) becomes dissolved in AY(s): BY is the solute and AY the solvent. Ifthe resulting solid solution is homogeneous, that is, it contains no concentrationgradient, the equilibrium distribution of A and B between the liquid and solid is

AY(s) + B+ = BY(s) + A+ (3.48)

and the equilibrium constant is

(BY(s))(A+)

(AY(s))(B+)= XBYfBY(A+)

XAYfAY(B+)= KSP(AY)

KSP(BY)

= K (3.49)

where the ratio of the activities of the solids is replaced by the ratio of the molefractions [XAY = nAY/(nAY + nBY) and XBY = nBY/(nAY + nBY)] multiplied byactivity coefficients, and KSP(AY) and KSP(BY) are the solubility products of pureAY and BY. Pure AY and BY are the ‘end members’ of the solid solution series.This equation shows that the dissolution of BY in solid AY is a function of:

(a) the ratio of the solubility products of AY and BY (K);(b) the solution composition with respect to A+ and B+;(c) a solid solution factor equal to the ratio of the activity coefficients of the

solid solution components (fAY/fBY).

Assuming as a first approximation fAY/fBY = 1 (an ideal solid solution) andthat the activity ratio of the species in solution can be replaced by the concen-tration ratio, then

XBY

XAY= K

[B+]

[A+](3.50)

As an example, consider a solid solution of 5 % ZnCO3 in MgCO3 (95 %) inequilibrium with Mg2+, Zn2+ and [CO3

2−] = 10−5 M (realistic conditions formany submerged soils). We have

K = KSP(MgCO3)

KSP(ZnCO3)

= 10−7.46

10−10.0= 347

Assuming the solubility of MgCO3 is little altered, then [Mg2+] = 3.47 × 10−3 Mand from Equation (3.49)

[Zn2+] = [Mg2+]1

K

XZnCO3

XMgCO3

= 5.26 × 10−7 M

This compares with [Zn2+] = 10−5 M for equilibrium with pure ZnCO3.If K is very large or small, the solid solution cannot be homogeneous because

A or B will be selectively scavenged from the liquid as the solid precipitates. Inthis case A or B will be occluded within the solid and will only be desorbed backinto solution very slowly. This mechanism provides a means by which toxic or

Page 93: The Biogeochemistry of Submerged Soils

84 Interchange of Solutes between Solid, Liquid and Gas Phases

essential metals may become buried in forms largely inaccessible to desorption,so they may accumulate over time.

Examples in Soil Systems

Solid solutions will only form between ions with similar radii (±15 %).Table 3.15 shows the radii in crystal lattices of divalent cations that might formsolid solutions in soils. Hence, for example Mn2+, Fe2+ and Cd2+ might beexpected to form solid solutions in CaCO3, but Cu2+ and Zn2+ would not.However, soils do not necessarily behave the same as pure systems. Thus there islittle evidence for strong association of Cd2+ or Pb2+ with calcite (CaCO3) in soilsystems, despite having similar radii to Ca2+ (McBride, 1994). However Cd2+and Pb2+ are both commonly associated with hydroxyapatite (Ca10(PO4)6(OH)2),and hence may contaminate phosphate fertilizers. Also Cu2+, Zn2+, Ni2+ andCo2+ are excluded from haematite during its crystallization but substitute intomagnetite (Fe(III)2Fe(II)O4). Small divalent cations such as Zn2+, Cu2+ andMg2+ often react with Al hydroxide to form solid solutions called hydrotalcitesof general formula [M2+

1−xAlx(OH)2]x+Xn−x/n, where Xn− is an anion. These

are noted for their permanent positive charge and hence anion exchange capacity.Features of metal sorption in soils that indicate—but do not prove—solidsolution formation include (McBride, 1994): sorption capacities that increaseover time; decreasing reversibility of sorption with time; selectivity for metalsbased largely on ionic radii; and solubilities well below those predicted from thesolubility products of pure systems.

The cycles of reduction and oxidation of Fe and Mn oxides in intermittentlysubmerged soils provide opportunities for co-precipitation with trace metals. Inmost natural systems it is the rate of dissolution of the solid phase that limitssolid solution formation rather than thermodynamics, so conditions in submergedsoils are highly conducive to formation of solid solutions.

Table 3.15 Radii of divalent cationsin crystal lattices

Cationa Ionic radius (nm)

Mg2+ 0.072Ca2+ 0.100Mn2+ 0.083Fe2+ 0.078Co2+ 0.075Ni2+ 0.069Cu2+ 0.073Zn2+ 0.074Cd2+ 0.075Pb2+ 0.119

a In six-fold coordination.

Page 94: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 85

3.7.4 INHIBITION OF PRECIPITATION

Rates of precipitation in soil are sensitive to inhibition by organic and inorganicligands that may be sorbed on soil surfaces and thereby interfere with nucleationand crystal growth. This is particularly important in submerged soils becauselarge concentrations of organic and inorganic ligands develop in the soil solutionfollowing submergence.

Inskeep and Bloom (1986) measured inhibition of calcite precipitation byorganic ligands in simulated soil solutions prepared from CaCl2, KHCO3 andseeds of CaCO3, and maintained at constant pH and CO2 pressure. The datafitted the rate equation:

R = ks[(Ca2+)(CO32−) − KSP] (3.51)

where R is the rate of CaCO3 precipitation, k is the rate constant, s is the surfacearea of CaCO3 seeds, and KSP is the solubility product of pure CaCO3. Thevalue of k was 117 dm6 mol−1 m−2 s−1 and it decreased to zero in the presenceof 0.15 mM water-soluble soil organic matter or 0.028 mM fulvic acid.

However, rates of precipitation in soil systems may be quite different fromthose in solutions because: precipitation is catalysed by adsorption of the reactingsolutes onto soil surfaces; the nature of the solid phases formed may be different;and sorption may also alter the effects of inhibitors. There are very few datain the literature on these effects actually measured in soils. Figure 3.15 showsdata of Huang (1990) for calcite precipitation in three soils incubated with urea.Precipitation was induced as the pH increased during urea hydrolysis:

CO(NH2)2 + 3H2O −−−→ 2 NH4+ + HCO3

− + OH− (3.52a)

Soil2 —Ca + 2 NH4+ + HCO3

− + OH− −−−→ 2Soil—NH4 + CaCO3(s) + H2O

(3.52b)

Simultaneously, in all the soils the concentrations of P and dissolved organiccarbon (DOC) in the soil solution increased as the pH increased, and this willhave inhibited CaCO3 precipitation. By analysing the combined data by multipleregression, Huang (1990) developed a rate equation for the three soils allowingfor inhibition by P and DOC. The rate equation was developed from an earlierequation developed from studies in solution systems:

R = kw0.379[(Ca2+)(CO32−)/KSP] (3.53)

where R is the rate of precipitation in mol dm−3 soil solution s−1, k is the rateconstant and w is the weight of newly formed CaCO3 in kg dm−3 soil solution,which is a surrogate for the surface area term in Equation (3.51), surface areabeing unmeasurable in soil. The rates calculated with this equation are comparableto those with Equation (3.51). The modified equation for soil systems is:

R = kw0.379ea[PL]eb[CL]ec[PL][CL][(Ca2+)(CO32−)/KSP] (3.54)

Page 95: The Biogeochemistry of Submerged Soils

86 Interchange of Solutes between Solid, Liquid and Gas Phases

CaC

O3

prec

ipita

ted

(mm

ol k

g−1 s

oil)

0

20

40

60

80

100

120

A 0.5A 0.7B 0.3B 0.5C 0.3C 0.5

pH

7.0

7.5

8.0

8.5

9.0

9.5

P c

once

ntra

tion

(mm

ol d

m−3

sol

utio

n)

0

1

2

3

4

5

Time (h)

0 20 40 60 80 100 120

DO

C c

once

ntra

tion

(mm

ol d

m−3

sol

utio

n)

0

100

200

300

400

500

Figure 3.15 Precipitation of calcite in three soils (A, B, C) as pH increases duringhydrolysis of urea, and simultaneous changes in pH, P and dissolved organic C (DOC) inthe soil solution. Moist soil was incubated with CaCl2 and urea at constant CO2 pressure.Numbers with symbols are initial concentrations of urea (M). Data of Huang (1990).Reproduced by permission

Page 96: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 87

where [PL] is the concentration of P in the soil solution, [CL] is the concentrationof DOC in the soil solution, a, b, c are coefficients and k is the rate constant,which is soil specific.

The values of k were 0.18, 1.98 and 1.16 × 10−6 (mol dm−3 s−1 basis) forSoils A, B and C in Figure 3.15, respectively. These values are more than fourtimes k for the solution system. The values of the inhibition coefficients–a =−1686, b = 6.13, c = 3854—were smaller than in the solution system. As aresult the concentrations of P and DOC required to halve the rate of precipitationwere 10 times those in the solution system. Also the interaction between [PL] and[CL] was negligible in the solution system but important in the soils. Figure 3.16shows plots of Equation (3.54) for different values of [PL] and [CL] and w =0.75 g dm−3. For the values used, which are realistic for submerged-soil solutions,the combined inhibitory effect of P and DOC was such that an order of magnitudegreater degree of supersaturation [(Ca2+)(CO3

2−)/KSP] is necessary to producethe same rate of precipitation as in the absence of inhibitors.

This sensitivity of precipitation in soil to organic ligands and other inhibitorsexplains why the soil solutions of submerged soils may be as much as a hundred-fold over-saturated with respect to solid phases in the first few weeks followingsubmergence (Chapter 4).

3.7.5 EQUATIONS FOR SOLID—SOLUTION INTERACTIONS

In the initial few weeks following submergence, the properties of the soil surfacechange markedly as a result of reductive dissolution and precipitation reac-tions. But in time, a steady- or quasi-steady-state is reached, and then the samerules govern the distributions of exchangeable ions between the soil solid and

Saturation index

20 40 60 80 100

Rat

e of

CaC

O3

prec

ipita

tion

(µm

ol d

m−3

s−1

)

0

1

2

3

4

5

6

7

8

9

0.5, 0

0, 2000.5, 200

0, 0

Figure 3.16 Inhibition of calcite precipitation by P and water-soluble organic mattercalculated with Equation (3.54) derived using the data in Figure 3.15. Numbers on curvesare concentrations of P and DOC in the soil solution (mM)

Page 97: The Biogeochemistry of Submerged Soils

88 Interchange of Solutes between Solid, Liquid and Gas Phases

solution as in non-submerged soils. The first consideration is the concentrationof non-adsorbed anions in solution because this determines the total strength ofthe solution. In submerged soils the principle anion is generally HCO3

−. Thenext consideration is what proportions of the exchangeable cations balance theanions in solution, and for this some form of empirical relation is necessary.There have not been many attempts to apply ion exchange equations to sub-merged soils (Pasricha and Ponnamperuma, 1976, 1977). But in principle thesame equations should apply as for non-submerged soils.

Considering the monovalent–divalent exchange reaction

soil—B + 2A+ = soil—2A + B2+ (3.55)

and applying the law of mass action:

(soil—2A)(B2+)

(soil—B)(A+)2= KE (3.56)

where KE is the equilibrium constant for the reaction and the terms in parenthesesare activities. Rearranging Equation (3.56) gives

(soil—2A)

(soil—B)= KE

(A+)2

(B2+)(3.57)

Because there is generally a large reserve of exchangeable cations on the solid,a small change in A+ results in little change in the ratio on the left-hand side.Hence the ‘reduced activity ratio’ (A+)/

√(B2+) tends to remain constant. The

activity coefficients for the ions in solution can be evaluated with Equation (3.3).Because of the complexity of soils, there are no general relations between the

proportions of two cations on the exchange complex and their reduced activityratio in solution. But two equations are commonly used:the Gaines and Thomas

NA√NB

= KGT(A+)√(B2+)

(3.58)

and the GaponNA

NB= KG

(A+)√(B2+)

(3.59)

where NA and NB are equivalent fractions of the total exchange capacity and KGT

and KG are exchange constants. When there are three or more competing ions, asthere generally will be, it is not practical to determine the exchange isotherms forall possible combinations of the ions. However Bond and Verburg (1997) haveshown that ternary and higher order exchanges can be predicted from the binaryexchange isotherms of the component ions.

Page 98: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 89

Calculated Changes in Exchangeable Cations Following Soil Reduction. Wecan use these equations to calculate how the exchangeable cations and the com-position of the soil solution will change following soil reduction. As we haveseen, precipitation of insoluble reduced compounds is often inhibited until a largesupersaturation is reached. Therefore for simplicity I assume no precipitation; theeffects of precipitation are considered in Chapter 4.

The major changes in the soil solid affecting exchangeable cations are: reduc-tion and dissolution of Fe oxyhydroxide coatings (cf. Equation 3.44):

2[soil—2Fe(OH)3−m] + CH2O + (8 − 4 m)H+ −−−→2[soil—]2m− + 4Fe2+ + CO2 + (11 − 4 m)H2O

(3.60)

in which, for each mol of Fe reduced, the change in surface negative charge is0.5 m molc and 2 − m mol of H+ are consumed; reduction of structural Fe inclay lattices (cf. Equation 3.47):

4[clay—Fe(III)OH] + CH2O −−−→4[clay—Fe(II)OH1−n](1−n)− + 4(1 − n)H+ + CO2 + (4n − 1)H2O

(3.61)

in which, for each mol of Fe reduced, the surface negative charge increases by1 − n molc and 1 − n mol of H+ are released; and changes in the charge onorganic matter and variable-charge clays due to the changes in pH. If ψ is theratio of structural Fe reduced to total Fe reduced, the total changes in surfacenegative charge and acidity are

[Z] = �[cations]S + [HS]S = {(1 − ψ)0.5 m + ψ(1 − n)}[Fe(III)](3.62)

and[HS] = −{(1 − ψ)(2 − m) − ψ(1 − n)}[Fe(III)] (3.63)

where �[cations]S and [HS]S are the changes in exchangeable cations andacidity in the soil solid. The latter is related to [HS] by

[HS]S = [HS] − θ/ρ([H+]L − [HCO3−]L) (3.64)

It is often found empirically that the change in soil pH for a given addition ofacid or base is constant over a wide range of pH and this relation is not greatlyaltered by soil reduction. Hence the pH buffer power, bHS, is constant and

pH = −[HS]/bHS (3.65)

Hence [H+]L and [HCO3−]L in Equation (3.64) can be found from [HS].

We therefore have the basis for the calculation.Consider the effect of Reactions (3.60) and (3.61) on the composition of a soil

solution containing exchangeable cations A+ and B2+ balanced by the anion X−.As Fe2+ and CO2 are formed and H+ consumed:

Page 99: The Biogeochemistry of Submerged Soils

90 Interchange of Solutes between Solid, Liquid and Gas Phases

(a) the strength of the soil solution increases, with the additional cations balancedby HCO3

−;(b) the CEC of the soil solid increases;(c) some A+ and B2+ are displaced from the exchange complex by Fe2+.

Given the values of [Fe(III)],CT,[Z] and [HS], we have nine unknowns:the concentrations of A+, B2+, Fe2+, H+ and HCO3

− in the soil solution, and theconcentrations of A+, B2+, Fe2+ and HS in the soil solid. These may be foundfrom the following nine equations:

(1) from Equation (3.65)

[H+]L = 10−(pH−[HS]/bHS) (3.66)

(2) from the carbonate equilibria,

[HCO3−]L = CT

[H+]L/KC1 + 1(3.67)

(3) from the requirement of electrical neutrality in the solution,

[A+]L + 2[B2+]L + 2[Fe2+]L + [H+]L = [HCO3−]L + [X−]L + [OH−]L

(3.68)

(4) from the requirement of electrical neutrality in the solid,

[A+]S + 2[B2+]S + 2[Fe2+]S =[A+]S0 + 2[B2+]S0 + [Z] − {[HS] − θ/ρ([H+]L − [HCO3

−]L)}(3.69)

where subscript 0 indicates initial values;(5) from monovalent–divalent cation exchange equilibria,

[A+]L√[B2+]L + [Fe2+]L

= KE1[A+]S

[B2+]S + [Fe2+]S(3.70)

(6) from divalent–divalent cation exchange equilibria,

[B2+]L

[Fe2+]L= KE2

[B2+]S

[Fe2+]S(3.71)

and (7), (8), (9) from conservation of mass there are three equations of the type

θ/ρ[A+]L + [A+]S = [A+] (3.72)

These equations can be solved simultaneously to obtain the new composition ofthe soil solution. Assume KE1 and KE2 constant in spite of reductive dissolutionreactions.

Page 100: The Biogeochemistry of Submerged Soils

Solid–Solution Interactions 91

Fe(III) reduced (mmol kg−1)

0 20 40 60 80 100

Con

cent

ratio

n in

soi

l sol

utio

n (m

mol

c L−1

)

0

5

10

15

20

25

30

pH

6.00

6.25

6.50

6.75

7.00

CO

2 pr

essu

re (

kPa)

0

5

10

15

20

25

30

CO2pressure

pH

HCO3−

Fe2+

A+

B2+

Figure 3.17 Calculated changes in a soil solution upon reduction of Fe(III)oxide coatings on soil surfaces and structural Fe(III) in clay lattices with-out re-precipitation of Fe(II). A+ and B2+ are exchangeable cations. Param-eter values in Equations (3.60)–(3.72): [A+] = 0.1[B2+], CEC0 = 100 mmol kg−1,[X]L = 5 mmolc L−1, pH0 = 6, bHS = 50 mmol pH−1 kg−1, KE1 = 1, KE2 = 1, ψ = 0.5,m = 0.5, n = 0.75, θ/ρ = 0.7

Figure 3.17 shows changes in the composition of a simulated soil solution socalculated. The proportions of Fe reduced in oxide coatings and structural Feare equal. The figure shows that as Fe2+ accumulates and the CEC, pH and[HCO3

−]L increase, the concentrations of Fe2+ and B2+ in solution increase, butthe concentration of the monovalent A+ in solution changes less because it ismore poorly buffered by the exchange complex. With a greater proportion ofstructural Fe reduced, less Fe2+ accumulates and the increase in pH is smallerbut nonetheless the CEC increases and so the increases in B2+ in solution withoxide dissolution and accumulation of HCO3

− is smaller.

Page 101: The Biogeochemistry of Submerged Soils

4 Reduction and Oxidation

The characteristics of submerged soils depend above all on the reduction and oxi-dation (redox) reactions that take place as a result of oxygen being excluded fromthe soil. With few exceptions, redox reactions on the Earth’s surface are drivenby the redox disequilibrium caused by photosynthesis. In photosynthesis greenplants use solar energy to reduce inorganic carbon to strongly reduced organiccompounds, and simultaneously water is oxidized to O2. Non-photosyntheticorganisms tend to restore equilibrium by catalysing the oxidation of the organiccompounds back to inorganic compounds in energy-yielding reactions. The bulkof this oxidation takes place in soil. In aerated soils, the preferred oxidizing agentis O2 itself. However where O2 is not available, as for example in submergedsoils, alternative oxidants must be used. These may be organic, in which case theprocess is fermentation, or inorganic, in which case it is anaerobic respiration,though this term is often used to cover fermentation as well. In this chapter Igive an overview of the thermodynamics of redox reactions and their kinetics innatural systems, and I then discuss the particular redox processes that occur insubmerged soils.

4.1 THERMODYNAMICS AND KINETICS OF REDOX REACTIONS

4.1.1 ELECTRON ACTIVITIES AND FREE ENERGY CHANGES

In the same way that acid–base reactions involve the transfer of protons betweenproton donors and proton acceptors, redox reactions involve the transfer of elec-trons between electron donors, called reducing agents or reductants, and electronacceptors, called oxidizing agents or oxidants. Thus when a redox reaction takesplace, a reductant loses electrons and is oxidized to its conjugate oxidant:

Red1 −−−→ Ox1 + ne− (4.1a)

and simultaneously an oxidant gains electrons and is reduced to its conjugatereductant:

Ox2 + ne− −−−→ Red2 (4.1b)

where e− represents the electron. Equations (4.1a) and (4.1b) are redox halfreactions or couples, and together they constitute a complete redox reaction.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 102: The Biogeochemistry of Submerged Soils

94 Reduction and Oxidation

Which redox couple in a redox reaction has the oxidizing role and which thereducing role depends on the relative abilities of the two couples to accept ordonate electrons. For example O2 has a greater affinity for electrons than otherpotential oxidants in natural systems, and is therefore reduced preferentially. Themeans of quantifying the relative abilities of redox couples to accept or donateelectrons and the corresponding free energy changes is as follows.

Just as free protons do not exist in solution in acid–base reactions, there are nofree electrons in redox reactions. However it is possible to define the activity ofelectrons relative to a specified standard state and thereby treat electrons as dis-crete species in equilibrium calculations in the same way as ions and molecules.The standard state of electron activity for this purpose is by convention definedwith respect to the redox couple made by hydrogen ions and hydrogen gas:

H+ + e− = 12 H2(g) (4.2)

By convention the standard free energy change, �Go, for this reaction is set atzero, i.e. since �Go = −RT ln K :

K = (H2(g))1/2

(H+)(e−)= 1 (4.2a)

The thermodynamic relations of any particular redox couple can therefore becalculated from the values for the reaction between the couple of interest and thehydrogen ion—hydrogen gas couple. Thus for the reduction of oxidant Ox to itsconjugate reductant Red, we have:

Ox + ne− = Red (4.3a)

and12nH2(g) = nH+ + ne− (4.3b)

and the full reaction is

Ox + 12nH2(g) = Red + nH+ (4.3c)

If K1, K2 and K are the equilibrium constants for Reactions (4.3a), (4.3b) and(4.3c), then

K = K1K2 = K1 = (Red)

(Ox)(e−)n(4.4)

i.e.

(e−) =[

1

K

(Red)

(Ox)

]1/n

(4.5)

or, on a logarithmic scale,

pe = − log(e−) = 1

n

[log K − log

(Red)

(Ox)

](4.6)

Page 103: The Biogeochemistry of Submerged Soils

Thermodynamics and Kinetics of Redox Reactions 95

Writing (1/n) log K = peo,

pe = pe0 − 1

nlog

(Red)

(Ox)(4.7)

and since �G = −2.303RT log K ,

peo = − �Go

2.303nRT(4.8)

peo is the electron activity with all the components of the redox couple at unitactivity, with unit activity of gases taken as a partial pressure of 1 atm. Table 4.1gives values of peo for important redox couples in natural systems expressed interms of unit electron transfer (i.e. n = 1 and peo = log K).

The pe of a solution indicates its relative tendency to accept or donate electronsin the same way that pH indicates the tendency to accept or donate protons. In astrongly reducing solution the tendency to donate protons and the correspondingelectron activity are large, and the pe is low. Likewise an acid solution has alow pH. Table 4.1 shows that the pe values and hence equilibrium constants ofmany redox reactions are very large or very small, which means that the reactionsproceed to completion in one direction or the other and the free energy changesinvolved are large.

The values of peo and �Go for any particular redox half reaction can becalculated from the values for complete redox reactions or combinations of reac-tions having the half reaction in common. For example, the oxidation of glucoseby oxygen,

16 glucose + O2(g) = CO2(g) + H2O (4.9)

is equivalent to the two half reactions

12 O2(g) + 2H+ + 2e− = H2O (4.9a)

and16 glucose + H2O = CO2(g) + 4H+ + 4e− (4.9b)

The change in free energy for the glucose—CO2 half reaction can be found from�Go for the full reaction less that for the O2–H2O half reaction = −477.7 − 2 ×(−236.6) = −4.5 kJ mol−1.

The values of peo and �Go can also be calculated from the standard freeenergies of formation, �Go

f , of each of the reactants and products in a redox halfreaction. For example, for the reduction of ferrihydrite [amorphous Fe(OH)3] thehalf reaction is

Fe(OH)3(s, amorph) + 3H+ + 2e− = Fe2+ + 3H2O (4.10)

and the values of �Gof are between −699 and −712 kJ mol−1 for amorphous

Fe(OH)3, −78.87 kJ mol−1 for Fe2+ and −711.54 kJ mol−1 for H2O. Therefore�Go

r = ��Gof products −��Go

f reactants = −91 to −78 kJ mol−1, and peo =(log K)/n = −�Go

r /2.303nRT = 16.0 to 13.7. Values of �Gof and other

Page 104: The Biogeochemistry of Submerged Soils

96 Reduction and Oxidation

Table 4.1 Equilibrium constants of important reduction half-reactions innatural systems at 25 ◦C

peo = log K

HH+ + e− = 1

2 H2(g) 0.00

O14 O2(aq) + H+ + e− = 1

2 H2O 20.75

N16 N2 + 4

3 H+ + e− = 13 NH4

+ 4.6312 NO3

− + H+ + e− = 12 NO2

− + 12 H2O 14.15

16 NO2

− + 43 H+ + e− = 1

6 NH4+ + 1

3 H2O 15.1414 NO3

− + 54 H+ + e− = 1

8 N2O(g) + 58 H2O 18.81

15 NO3

− + 65 H+ + e− = 1

10 N2(g) + 35 H2O 21.05

S12 SO4

2− + H+ + e− = 12 SO3

2− + 12 H2O −1.65

18 SO4

2− + 98 H+ + e− = 1

8 HS− + 12 H2O 4.25

14 SO4

2− + 54 H+ + e− = 1

8 S2O32− + 5

8 H2O 4.8518 SO4

2− + 54 H+ + e− = 1

8 H2S(g) + 12 H2O 5.25

Mn12 MnO2(s) + 2H+ + e− = 1

2 Mn2+ + 2H2O 21.82

MnOOH(s) + 3H+ + e− = Mn2+ + 2H2O 25.3312 Mn3O4(s) + 4H+ + e− = 3

2 Mn2+ + 2H2O 30.79

Fe

Fe3+ + e− = Fe2+ 13.00

α-FeOOH(s) + 3H+ + e− = Fe2+ + 2H2O 11.31

Fe(OH)3(s) + 3H+ + e− = Fe2+ + 3H2O 16.54

C12 CH2O + H+ + e− = 1

2 CH3OH(aq) 3.9914 CH2O + H+ + e− = 1

4 CH4(g) + 14 H2O 6.94

12 CO2(g) + 1

2 H+ + e− = 12 HCOO− −5.22

14 CO2(g) + H+ + e− = 1

4 HCHO(aq) + 12 H2O −1.20

16 CO2 + H+ + e− = 1

6 CH3OH(aq) + 16 H2O 0.50

18 CO2(g) + H+ + e− = 1

8 CH4(g) + 14 H2O 2.87

14 CO2(g) + 7

8 H+ + e− = 18 CH3COO− + 1

4 H2O 1.2715 CO2(g) + H+ + e− = 1

10 CH3CHO(g) + 310 H2O 0.99

16 CO2(g) + H+ + e− = 1

12 CH3CH2OH(aq) + 14 H2O 1.52

16 CO2(g) + H+ + e− = 1

12 C2H4(g) + 14 H2O 1.34

14 CO2(g) + 11

12 H+ + e− = 112 (lactate) + 1

4 H2O 0.68310 CO2(g) + 9

10 H+ + e− = 110 (pyruvate) + 3

10 H2O 0.0514 CO2(g) + H+ + e− = 1

24 (glucose) + 14 H2O −0.20

Source: Morel and Herring (1993) and Stumm and Morgan (1996).

Page 105: The Biogeochemistry of Submerged Soils

Thermodynamics and Kinetics of Redox Reactions 97

thermodynamic properties of common chemical species in natural systems aregiven in Stumm and Morgan, 1996 (Appendix 3).

4.1.2 REDOX POTENTIALS

An alternative approach to quantifying redox equilibria is to treat them as elec-trode reactions and to calculate the electric potential that would exist if the coupleof interest formed a half cell with an inert electrode. The standard reference pointfor this is again taken as the hydrogen half reaction, the electrode potential forwhich is set at zero, which is equivalent to setting �Go equal to zero. Thisapproach lacks the simplicity of the system based on electron activities and isnot so well suited to equilibrium calculations. However it has historical prece-dence and is widely used so must be mentioned. Also, the electrode potentials ofmany redox couples can be measured directly. An inert electrode, such as plat-inum metal, is placed in a solution containing the redox couple of interest andlinked via an external circuit to a hydrogen half cell comprising an inert electrodein a solution in equilibrium with H2 gas at 1 atm and having unit activity of H+ions. The potential is measured with a voltmeter.

The standard electrode potential, EoH, in which the suffix H indicates that the

potential is on the H2–H+ scale, is derived as follows. We have

�G = −nFEH (4.11)

where F is the Faraday. Combining Equation (4.11) with Equation (4.8) gives

EH = 2.303RT

Fpe (4.12)

Substituting for pe from Equation (4.6) in Equation (4.12) gives

EH = 2.303RT

F

1

n

[log K − log

(Red)

(Ox)

](4.13)

or

EH = EoH − 2.303RT

nFlog

(Red)

(Ox)(4.14)

where EoH = (2.303RT/nF) log K . This is the Nernst equation.

At 25 ◦C, 2.303RT/F = 0.059 V, therefore

EH = 0.059 pe (4.15)

4.1.3 RELATION BETWEEN pe AND CONCENTRATIONOF REDOX COUPLES

If one redox couple in a redox reaction is present at a much greater concentrationthan the other, then the concentration of the reduced and oxidized species in this

Page 106: The Biogeochemistry of Submerged Soils

98 Reduction and Oxidation

couple are little influenced by the advancement of the reaction towards equi-librium and the equilibrium electron activity is effectively that of the dominantredox couple. This is given by the appropriate form of Equation (4.7) for thedominant couple:

pe = peo − 1

nlog

(Red)

(Ox)(4.16)

(Note that this equation can be expressed in terms of the concentrations of Redand Ox by dividing the activities by the appropriate activity coefficients.) Thepe of a solution will therefore be ‘poised’ at the pe determined by the dominantcouple until that couple is exhausted. The pe of all other redox couples operatingwill tend to adjust to this electron activity.

For example, in oxic natural waters the principal oxidant is O2 and in agreementwith expectations the pe of such waters is generally poised in the range expectedfor the O2–H2O couple (Morel and Herring, 1993). Thus for water at pH = 7 inequilibrium with atmospheric PO2(= 10−0.7 atm), the half reaction is

14 O2(g) + H+ + e− = 1

2 H2O peo = 20.75

therefore

pe = peo − log1

P1/4O2

(H+)= 13.58, (4.17)

which agrees well with the typical range of pe in oxic natural waters; pe = 12to 14. Equation (4.17) indicates the pe varies as log (H+) but only as 1

4 log PO2 .Thus the pH has a large effect on pe but the concentration of O2 has only a minoreffect and small concentrations of O2 maintain water in an oxidized state.

A further informative example is the organic matter—CO2 couple, which is theprincipal reductant in natural systems. Consider a solution in equilibrium withatmospheric CO2 at neutral pH and containing 10 µM ‘CH2O’, where CH2Orepresents average organic C in natural systems, whose composition is similarstoichiometrically to that of carbohydrates. The half reaction is

14 CO2(g) + H+ + e− = 1

4 CH2O + 14 H2O peo = −1.20

therefore

pe = peo − log(CH2O)1/4

P1/4CO2

(H+)= −7.83 (4.18)

Such very low pe values do not generally occur in natural systems because oxidiz-ing couples such as O2–H2O or Fe(OH)3–Fe2+ are usually present in much greaterconcentrations. However pe values this low may occur in bacterial cells whereorganic matter is being oxidized in the absence of large concentrations of inor-ganic oxidizing couples, providing strongly reducing microenvironments whichmay be linked to redox couples in the external environment via intermediariespassing across the cell wall.

Page 107: The Biogeochemistry of Submerged Soils

Thermodynamics and Kinetics of Redox Reactions 99

Notice that, as in Equation (4.17), pe in Equation (4.18) is sensitive to pH butnot to the concentrations of the redox species. The sensitivity to the concentrationof the redox species depends on the reaction stoichiometry. For the Fe(OH)3–Fe2+couple, for example, the half reaction is

Fe(OH)3(s) + 3H+ + e− = Fe2+ + 3H2O peo = 16.54

giving

pe = peo − log(Fe2+)

(H+)3(4.19)

and thus pe is sensitive to both the Fe2+ concentration and pH, and pe falls asthe reduction of Fe(OH)3 proceeds and Fe2+ accumulates.

4.1.4 pe–pH DIAGRAMS

Most redox reactions consume or produce protons and the stoichiometry is oftensuch that pe is very sensitive to pH, as the examples in the previous sectionshow. A simple method for determining which species will predominate underparticular conditions of pe and pH in an unknown redox system is to construct‘pe–pH diagrams’. This is done as follows. Consider the following redox halfreaction involving H+:

Ox + mH+ + ne− = Red

We have

pe = peo − 1

nlog

(Red)

(Ox)− m

npH (4.20)

i.e.

n(pe − peo) + mpH = log(Ox)

(Red)(4.21)

If n(pe − peo) + mpH > 0 then (Ox) > (Red) and the oxidant is the dominantspecies, and vice versa. Hence a plot of pe versus pH with (Ox) = (Red) has slopem/n and intercept pe, and for points above the line the oxidant is dominant andfor points below the reductant is dominant. pe is taken as the dependent variable,plotted on the ordinate, because pH is often controlled by processes in additionto redox reactions and is therefore more properly the independent variable.

Figure 4.1 gives examples for biological redox couples important in naturalsystems. Figure 4.1(a) shows the diagram for the H2O–O2 and H2O–H2 cou-ples. The respective lines are pe = 20.75 − 1

4 log PO2 − pH with PO2 = 1 atm,and pe = − 1

4 log PH2 − pH with PH2 = 1 atm. These are upper limits for the par-tial pressures of O2 and H2 in natural waters. For points above the upper line, H2Ois an effective reductant, producing O2; for points below the lower line H2O is aneffective oxidant, producing H2. The region between the lines, where O2 acts asan oxidant and H2 as a reductant, covers most circumstances in natural systems.

Page 108: The Biogeochemistry of Submerged Soils

100

45

67

89

104

56

78

910

45

67

89

104

56

78

910

114

56

78

910

11

pe

−10−505101520

PO

2 > 1

atm

PH

2 > 1

atm

pH

H2O

N2

NO

3−

NH

4+N

H3

NO

3−

NH

4+NH

3

NO

2−S

O42−

S(s

)

H2S

HS

−C

H4

CO

2H

CO

3−

CO

32−

(a)

(b)

(c)

(d)

(e)

Fig

ure

4.1

pe–p

Hdi

agra

ms

for

impo

rtan

tbi

olog

ical

redo

xco

uple

sin

natu

ral

syst

ems.

(a)

H2O

–O2.

(b)

The

nitr

ogen

syst

emco

nsid

er-

ing

only

stab

leeq

uilib

ria:

the

only

oxid

atio

nst

ates

invo

lved

are

(−II

I),

the

elem

enta

lst

ate

and

(+V

).(c

)T

heni

trog

ensy

stem

trea

ting

NH

4+ ,

NH

3,

NO

3−

and

NO

2−

asm

etas

tabl

ew

ithre

spec

tto

N2

whi

chis

trea

ted

asre

dox-

iner

t.(d

)T

heSO

42−

–S(s

)–H

2S(

aq)

syst

em,

[tot

also

lubl

eS]

=10

−2M

.(e

)T

heca

rbon

syst

emig

nori

ngel

emen

tal

C.

Aft

erSt

umm

and

Mor

gan

(199

6).

Rep

rodu

ced

bype

rmis

sion

ofW

iley,

New

Yor

k

Page 109: The Biogeochemistry of Submerged Soils

Thermodynamics and Kinetics of Redox Reactions 101

Figure 4.1(b) shows the diagram for the stable equilibria in the nitrogen system.According to this diagram N2 should be largely oxidized to NO3

− in most naturalwaters. The fact that it is not and N2 is known to persist in oxic waters indicatesthat a complete redox equilibrium does not exist; only a partial equilibrium isattained under the mediation of microbes. Figure 4.1(c) shows the diagram forthe nitrogen system with N2 treated as redox-inert and NH4

+, NH3, NO3− and

NO2− as metastable with respect to N2. This diagram more correctly represents

conditions in natural systems, with NH4+ as the stable species under mildly

reducing conditions and NO3− under oxic conditions. This example illustrates

the difficulty in choosing the correct redox couples to represent real systems inpe–pH diagrams. Some independent insight into the system is generally requiredto choose the correct couples. Figure 4.1(d) and (e) show diagrams for sulfur andcarbon systems.

The situation is further complicated for redox reactions involving several solidphases. An example is the Fe–CO2 –H2O system shown in Figure 4.2. Thisshows that ferrihydrite, Fe(OH)3, can be formed over a wide range of pe and pH,though the pe range is increasingly restricted under increasingly acid conditions,and Fe2+ is then the stable form. Siderite, FeCO3, and the hypothetical Fe(II)hydrous oxide Fe(OH)2 may be formed under moderately reducing conditionsbut only at pH > 7. Elemental Fe is only stable under very strongly reducingconditions, outside the range in which water is stable. In real systems the situation

pH

0 2 4 6 8 10 12 14

pe

−15

−10

−5

0

5

10

15

20

25

PO2 > 1 atm

PH2 > 1

Fe(OH)3(amorph, s)

Fe3+

FeOH2+

Fe(OH)4−

Fe2+

FeCO3(s)

Fe(OH)2(s)Fe(s)

Figure 4.2 pe–pH diagram for the Fe–CO2 –H2O system. [Fe(II)] = 1 mM, [Fe(III)] =0.01 mM, CT = 5 mM. Amorphous Fe(OH)3 is ferrihydrite, FeCO3 siderite, and Fe(OH)2a hypothetical Fe(II) hydrous oxide. The details of the construction of this diagram areexplained in Stumm and Morgan (1996, Chapter 8). Reproduced by permission of Wiley,New York

Page 110: The Biogeochemistry of Submerged Soils

102 Reduction and Oxidation

may be complicated by the presence of other redox species with which Fe reacts,such as sulfide, and by the slow kinetics of redox and precipitation reactions andthe need for microbial mediation. Thus for example siderite is rarely found insoils though in many cases it is the thermodynamically favoured phase as shownin Figure 4.2. These points are discussed further in later sections.

Because of the sensitivity of pe to pH it is often convenient to compare peo

values ‘corrected’ to pH 7 and termed peo∗, where:

peo∗ = peo − 1

nlog

(Red)

(Ox)− 7

m

n(4.22)

As discussed earlier, the concentration-dependent term in Equation (4.22) willoften be small in comparison with the pH term and can be ignored. For couplesin which the concentration term is more important, such as Fe(OH)3–Fe2+, peo∗values can be calculated for representative concentrations.

Table 4.2 gives peo∗ values for important redox couples in natural systemsarranged in order of decreasing peo∗ with strong oxidants at the top and strongreductants at the bottom. From such a table it is possible to infer which coupleswill react when present together and which will have the oxidizing role and whichthe reducing role. The table shows for example that Fe(OH)3 can readily oxidizeorganic matter ‘CH2O’ to form CO2 and Fe2+ but it cannot oxidize N2 to NO3

−.However note that the peo∗ value for the Fe(OH)3 –Fe2+ couple is sensitive tothe value of (Fe2+).

4.1.5 ENERGETICS OF REACTIONS MEDIATED BY MICROBES

Most redox reactions in vitro reach equilibrium only extremely slowly with halftimes of the order of months or years, even though they may be highly favouredthermodynamically. This is illustrated by the persistence of N2 in oxic systemseven though its oxidation to NO3

− is strongly favoured (Table 4.1). However,microbes in soil and water are capable of catalysing particular reactions fromwhich they obtain energy for metabolism. The half times of such microbiallycatalysed reactions are of the order of hours or days.

The amounts of energy consumed or produced in redox reactions, and hencethe efficiency with which they can be exploited by microbes, can be calculatedfrom thermodynamic data. This gives surprisingly good insights into the dynam-ics of microbial communities in natural systems without detailed knowledge ofthe biochemical and physiological pathways involved. For example, the sequenceof reduction reactions that occur in submerged soils following exclusion of O2

matches the order of decreasing free energy change for the corresponding redoxreactions. Note that organisms cannot carry out gross reactions that are thermo-dynamically impossible: they do not oxidize substrates or reduce oxidants per se,but merely catalyse the process by mediating the electron transfers occurring. Theenergy produced or consumed in a given redox reaction is calculated as follows.

Page 111: The Biogeochemistry of Submerged Soils

Thermodynamics and Kinetics of Redox Reactions 103

Table 4.2 Equilibrium constants of reduction half-reactions at pH 7 and 25 ◦C

peo peo∗

14 O2(aq) + H+ + e− = 1

2 H2O 20.75 13.7515 NO3

− + 65 H+ + e− = 1

10 N2(g) + 35 H2O 21.05 12.65

14 NO3

− + 54 H+ + e− = 1

8 N2O(g) + 58 H2O 18.81 10.06

12 MnO2(s) + 2H+ + e− = 1

2 Mn2+ + 2H2O 21.82 9.67a

12 Mn3O4(s) + 4H+ + e− = 3

2 Mn2+ + 2H2O 30.79 8.33

MnOOH(s) + 3H+ + e− = Mn2+ + 2H2O 25.33 8.02a

12 NO3

− + H+ + e− = 12 NO2

− + 12 H2O 14.15 7.15

18 NO3

− + 54 H+ + e− = 1

8 NH4+ + 3

8 H2O 14.90 6.1516 NO2

− + 43 H+ + e− = 1

6 NH4+ + 1

3 H2O 15.14 5.8214 CH2O + H+ + e− = 1

4 CH4(g) + 14 H2O 6.94 −0.06

Fe(OH)3(s) + 3H+ + e− = Fe2+ + 3H2O 16.54 −1.46a

12 CH2O + H+ + e− = 1

2 CH3OH 3.99 −3.0118 SO4

2− + 54 H+ + e− = 1

8 H2S(g) + 12 H2O 5.25 −3.50

18 SO4

2− + 98 H+ + e− = 1

8 HS− + 12 H2O 4.25 −3.63

18 CO2(g) + H+ + e− = 1

8 CH4(g) + 14 H2O 2.87 −4.13

16 N2 + 4

3 H+ + e− = 13 NH4

+ 4.63 −4.70

α-FeOOH(s) + 3H+ + e− = Fe2+ + 2H2O 11.31 −6.69a

H+ + e− = 12 H2(g) 0.00 −7.00

14 CO2(g) + H+ + e− = 1

24 (glucose) + 14 H2O −0.20 −7.20

14 CO2(g) + H+ + e− = 1

4 CH2O + 12 H2O −1.20 −8.20

a For reductive dissolution of Mn and Fe oxides peo∗ values are calculated with (Mn2+) = 0.2 mM and(Fe2+) = 1 mM to represent conditions in submerged soil solutions; in other couples reactants are givenunit activities.

Consider the reaction

Ox1 + Red2 = Red1 + Ox2

for which the reduction half reactions are:

Ox1 + ne− = Red1

Ox2 + ne− = Red2

with equilibrium constants K1 and K2, i.e. pe1 = 1/n log K1 and pe2 =1/n log K2. Therefore

�Go = −2.303RT log K = −2.303RT logK1

K2= −2.303RT n(peo

1 − peo2)

(4.23)

Page 112: The Biogeochemistry of Submerged Soils

104 Reduction and Oxidation

The free energy change for the reaction is

�G = �Go + 2.303RT log(Ox2)(Red1)

(Ox1)(Red2)(4.24)

Combining Equations (4.23) and (4.24) gives

�G = −2.303RT

[n(peo

1 − peo2) − log

(Ox2)(Red1)

(Ox1)(Red2)

](4.25)

As discussed earlier, the dependence of pe on the concentrations of reductantsand oxidants is often small in comparison with its dependence on pH. The termin the square brackets in Equation (4.25) can therefore be replaced by peo∗ termsgiving for the approximate standard free energy change:

�Go∗ ≈ −2.303RT n(peo∗1 − peo∗

2 ) (4.26)

Figure 4.3 shows oxidation and reduction reactions used by microbes as energysources on a scale of peo∗ (data from Table 4.1). The free energy changes forthe different complete reactions can be read from the �Go∗ scale, in accordancewith Equation (4.26). The energy expended by microbes in elaborating carbon,for example through fixation of CO2, and other elements, for example nitrogenthrough fixation of atmospheric N2, can be calculated in a similar way.

Such calculations indicate the maximum energy available from a reaction or theminimum required to carry it out. The true gain to a microbe is smaller, or the costlarger, because of the energy required for cell maintenance and reproduction andother processes. The energetic efficiencies of biochemical processes are typicallyof the order of 30–40 %. Nonetheless the ecological succession of microbes inresponse to the stepwise oxidation of reduced compounds and exhaustion ofoxidants can be predicted from such calculations. Thus the succession of aerobicorganisms, denitrifiers, manganese reducers, iron reducers, sulfate reducers andmethanogenic bacteria following submergence of a soil directly matches the orderof decreasing peo∗ for the corresponding redox couples in Figure 4.3(b): O2–H2O,NO3

−–N2, MnO2(s)–Mn2+, Fe(OH)3(s)–Fe2+, SO42−–HS− and CH2O–CH4.

Microorganisms and organisms in general can be classified according to theprincipal sources of their energy, carbon and electrons. A hierarchical classifica-tion is not possible because all combinations of these three occur. Thus all threecan be separate, as for green plants which obtain their energy from sunlight,carbon from CO2 and electrons by oxidizing water to O2; and all three can bethe same, as for the majority of bacteria which use organic compounds as theirsources of energy, carbon and electrons.

Organisms that obtain their carbon from inorganic compounds, mainly CO2, arecalled autotrophs. They are subdivided into photoautotrophs which obtain energyfrom sunlight—for example, green plants and photosynthetic bacteria—and che-moautotrophs which obtain energy from chemical processes—for example, inFigure 4.3(a), nitrifying bacteria, which oxidize NH4

+ to NO3−, sulfur oxidizing

Page 113: The Biogeochemistry of Submerged Soils

105

−∆Go* (kJ mol−1)

102030405060708090100

110

120

130

peo*

−10−9−8−7−6−5−4−3−2−10123456789101112131415

peo*

−10

−9−8−7−6−5−4−3−2−10123456789101112131415

O2

H2O

O

2 H

2O

CH

2O C

O2

H+

H2

CO

2 C

H4

CH

2O C

H3O

H

SO

42− H

S−

Fe(

OH

) 3

Fe2+

NO

3− N

2

NO

3− N

2O

MnO

2 M

n2+

Red

uctio

nsO

xida

tions

NO

3− N

O2−

(a)

(b)

Oxi

datio

nsR

educ

tions

−∆Go* (kJ mol−1)10 20 30 40 50 60 70 80 90 10

0

110

120

130

Fe2+

Fe(

OH

) 3

NO

2− N

O3−

NH

4+ N

O2−

H2

H+

H2S

SO

42−

NO

3− N

H4+

Fig

ure

4.3

Free

ener

gych

ange

sin

redo

xre

actio

nsm

edia

ted

bym

icro

bes.

(a)

Oxi

datio

nof

redu

ced

inor

gani

cco

mpo

unds

linke

dto

redu

ctio

nof

O2.

(b)

Oxi

datio

nof

orga

nic

mat

ter

‘CH

2O

’lin

ked

tore

duct

ion

ofva

riou

sor

gani

can

din

orga

nic

oxid

ants

.pH

=7

and

unit

oxid

ant

and

redu

ctan

tac

tiviti

esex

cept

(Mn2+

)=

0.2

mM

and

(Fe2+

)=

1m

M

Page 114: The Biogeochemistry of Submerged Soils

106 Reduction and Oxidation

bacteria, which oxidize reduced S compounds to SO42−, and Mn(II) and Fe(II)

oxidizing bacteria, which produce insoluble Mn and Fe oxides, though it is notcertain that useful energy is derived from this. Examples of autotrophs usingdifferent electron sources in fixing CO2 are green plants which derive electronsfrom the oxidation of water, sulfide oxidizers which oxidize H2S(g) to colloidalS, and ammonium oxidizers which oxidize NH4

+ to NO2−.

Organisms that obtain carbon from ingested organic compounds are calledheterotrophs, and most also derive their energy and electrons from these organiccompounds. Examples are fungi, protozoa and most bacteria. A wide range oforganic and inorganic oxidants are used as end electron acceptors in oxidizing theorganic compounds, as in the reactions shown in Figure 4.3(b). Also a wide rangeof organic compounds are oxidized. The resulting free energy changes may differsubstantially from those in Figure 4.3(b) for oxidation of the average compound‘CH2O’. For example the oxidation of glucose yields about 54 kJ more energyper mole of C than oxidation of acetate. This makes an increasingly significantdifference the lower the peo∗ of the oxidizing couple. Thus it makes only a smalldifference for O2 or NO3

− reduction (12 and 15 %, respectively), but a largedifference for SO4

2− reduction (69 %).An important component of the overall efficiency of energy production by

microbes is the location of the linked couples and the resulting need to trans-port reactants and products across cell membranes. In denitrification and SO4

2−reduction, because all of the NO3

− and to a lesser extent the SO42− are dissolved

in the soil solution, they are readily imported into the cell and their reductionlinked directly to the oxidation of organic compounds via electron transfer sys-tems. But in Mn and Fe reduction, the oxides are only sparingly soluble, and sothe concentrations of Mn(III, IV) and Fe(III) in solution are small, even whenlarge concentrations of the solid oxides are present. This presented a problem inestablishing that Mn and Fe reduction was directly linked to microbial respirationin natural systems, rather than being an indirect effect through abiotic reactionsinvolving side products of respiration. The evidence for the direct involvementof microbes is discussed in Chapter 5.

4.2 REDOX CONDITIONS IN SOILS

This topic has a long history of research (Harrison and Aiyer, 1920; Sturgis,1936; Pearsall and Mortimer, 1939; Shioiri, 1943; De Gee, 1950; Takai, 1952;Ponnamperuma, 1955; Baas-Becking et al., 1960; Jeffrey, 1961; Patrick, 1966;Ponnamperuma, 1972; Yu, 1985; Kyuma, 2003). The following factors result inconditions differing from those in simple aquatic systems:

• The soil has a structure and contains a network of pores filled to a varyingextent with water, and the soil is overlain by a layer of standing water of vary-ing depth and degree of oxygenation. The filling and emptying of the poresis often very dynamic changing from complete saturation to near emptiness

Page 115: The Biogeochemistry of Submerged Soils

Redox Conditions in Soils 107

and vice versa within a matter of days. Redox conditions are correspondinglydynamic.

• Transport of solutes and gases through the soil is much slower than through soil-free water because of the restricted cross-sectional area for transport throughthe soil pore network and because of adsorption and reaction on soil sur-faces (Chapter 2). Redox conditions are therefore closely linked to transportprocesses.

• Mineral surfaces have a much greater influence through sorption and precipi-tation of solutes and direct mediation of redox reactions.

• The soil contains organic matter which is humified to a varying extent andinputs of fresh organic matter are often much larger than in aquatic sys-tems because of greater net primary productivity. The organic matter bothprovides substrates for microbial processes and participates in sorption andother reactions.

• The micro- and macrobiological ecologies are different.

In this section the redox conditions developing in soils following submergenceare described and the processes governing these conditions are analysed in termsof the soil chemistry and microbiology discussed so far.

4.2.1 CHANGES WITH DEPTH IN THE SOIL

The floodwater standing on the soil surface is usually sufficiently shallow, wellmixed by wind and thermal gradients, and oxygenated by photosynthetic organ-isms that it is essentially aerobic. However transport of O2 into the underlyingsoil is too slow for more than a thin layer to be aerobic. In this layer the con-centrations of Mn2+, Fe2+ and other reduced species are negligible, and CO2 isthe main end product of microbial respiration. In the underlying anaerobic soil,only a few millimetres away, the concentrations of Mn2+, Fe2+ and the vari-ous organic products of anaerobic respiration can be very large. Thus conditionschange dramatically over a very short distance.

The distribution of reduced species with depth follows a characteristic patternreflecting the succession of terminal electron acceptors used by microbes—O2,NO3

−, Mn(III, IV), Fe(III), SO42− and organic C in fermentation reactions.

Sorption, precipitation and dissolution reactions also influence the distribution.Figure 4.4 shows profiles of EH and extractable Mn2+, Fe2+ and S2− in soilcolumns following flooding, illustrating some of these effects. A steady statedevelops over time and the profiles of the reduced species in the soil then reflectthe profile of EH being progressively deeper for the less-easily reduced species.Thus the profile of Mn2+ in the figure extends closest to the surface, Mn reduc-tion taking place at the highest EH, and the S2− profile extends least close to thesurface. The depth to which O2 penetrates the soil, as indicated by the depth atwhich EH begins to decrease, is about 10 mm under the conditions of the figure.This depth depends on such factors as the oxygenation of the floodwater, the

Page 116: The Biogeochemistry of Submerged Soils

108

Depth in soil (mm)

−100

010

020

0E

H (

mV

)300

400

500

0 5 10 15 20 25

1 da

y

13 w

k

2 w

k

4 w

k

10 w

k

[Fe2+

] (µ

mol

g−1

)0

1020

3040

1 da

y1 w

k

2 w

k 4 w

k

6 w

k

8 w

k

13 w

k

7 w

k

10 w

k

[Mn2+

] (µ

mol

g−1

)0

26

48

10

2 w

k

1 w

k

6 w

k

8 w

k

13 w

k

10 w

k

1 da

y

025

050

075

010

0012

5015

00[S

2−]

(cpm

g−1

)

2 w

k3

wk

5 w

k

13 w

k

8 w

k11

wk

Fig

ure

4.4

Dep

thdi

stri

buti

onof

EH

,M

n2+,

Fe2+

,S

2−at

diff

eren

ttim

esfo

llow

ing

flood

ing

(Pat

rick

and

DeL

aune

,19

72).

Rep

rodu

ced

bype

rmis

sion

ofSo

ilSc

i.So

c.A

m.

Page 117: The Biogeochemistry of Submerged Soils

Redox Conditions in Soils 109

amount and quality of organic matter present, and the concentrations of easilyreducible Fe(III) and other reductants.

I now summarize the changes in electrochemical conditions that occur in thesoil with time following submergence as they affect the profiles and dynamicsof redox species.

4.2.2 CHANGES WITH TIME

Reduction of a submerged soil proceeds roughly in the sequence predicted bythermodynamics:

O2 + CH2O −−−→ CO2 + H2O (4.27)

4NO3− + 5CH2O + 4H+ −−−→ 2N2 + 5CO2 + 7H2O (4.28)

2MnO2 + CH2O + 4H+ −−−→ 2Mn2+ + CO2 + 3H2O (4.29)

4Fe(OH)3 + CH2O + 8H+ −−−→ 4Fe2+ + CO2 + 11H2O (4.30)

SO42− + 2CH2O + 2H+ −−−→ H2S + 2CO2 + 2H2O (4.31)

and2CH2O −−−→ CH4 + CO2 (4.32)

Typically O2 becomes undetectable within a day of submergence and then NO3−

is reduced. Reduction of NO3− will not occur until the O2 concentration reaches

a very small value. Likewise, whilst NO3− is being reduced, the pe is poised in

the range 3–6 and reduction of Mn and Fe are prevented. However NO3− will be

exhausted within a matter of days and then reduction of Mn and Fe may proceed.In the absence of O2 Fe(III) is generally the main oxidant in the soil, its

concentration typically exceeding concentrations of NO3−, Mn(III, IV) or SO4

2−by at least an order of magnitude (Chapter 3). Between 1 and 20 % and sometimesas much as 90 % of the free Fe(III) in the soil is reduced to Fe(II) over 1–2months of submergence (Ponnamperuma, 1972; van Breemen, 1988). Some ofthe structural Fe(III) in soil clays is also reduced (Stucki et al., 1997). The courseof soil reduction and the changes in pe and pH are therefore generally dominatedby the reduction of Fe(III).

Changes in pe, pH and Alkalinity

It is difficult to obtain reliable measurements of EH and hence pe in soils. Strictly,only measurements made with the electrodes in soil solution extracts rather thandirectly in soil are thermodynamically meaningful, and these are also subject tovarious errors, particularly due to the presence of mixed redox systems. Nonethe-less it is a useful parameter and is the only single electrochemical property thatcan distinguish submerged soils from well-drained ones.

Figure 4.5 shows changes in pe, pH and Fe2+ in the soil solution of fourrepresentative soils following flooding (IRRI, 1964). The figure shows that inall the soils there is a minimum in pe after a few days followed by an increase,

Page 118: The Biogeochemistry of Submerged Soils

110

02

46

810

1214

16024681012

02

46

810

1214

16−101234567

10

1

.5

1.6

7 2

1

4.1

2

.78

26

1

.5

0.3

0 3

0

5.1

1

.2

Tim

e (w

eeks

afte

r flo

odin

g)

[Fe2+] in solution (mM)

pH

pe

02

46

810

1214

165.

2

5.4

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

7.2

Soi

l O

rg C

Act

ive

(

%)

Fe

(%)

Fig

ure

4.5

Cha

nges

inpe

,pH

and

Fe2+

inso

ilso

luti

ons

ofva

riou

sso

ils

foll

owin

gsu

bmer

genc

eat

25◦ C

.pe

valu

esar

eca

lcul

ated

from

mea

sure

dE

H(V

)va

lues

inso

ilso

lutio

ns.

Dat

afr

omIR

RI

(196

4).

Rep

rodu

ced

bype

rmis

sion

ofIR

RI

Page 119: The Biogeochemistry of Submerged Soils

Redox Conditions in Soils 111

and this is characteristic of most soils following flooding (Ponnamperuma, 1972).The minimum can be less than zero and can be accompanied by evolution of H2

gas. It is due to fermentation reactions starting as soon as O2 and NO3− are used

up but before populations of Mn and Fe reducing bacteria are established. Asdiscussed in Section 5.3, the low solubility of Mn(III, IV) and Fe(III) oxides mayinitially limit the rate of their reduction. Organic acids produced in fermentationreactions will help dissolve Mn(III, IV) and Fe(III) from oxide particles andthereby facilitate the establishment of the Mn and Fe reducers. As Mn and Fereduction then proceed, the pe will increase to values corresponding to the Mnand Fe couples involved, and then gradually decline.

Simultaneously H+ ions are consumed in Reactions (4.28)–(4.31) and thepH tends to increase. Initially the pH of aerobic soils may decrease followingsubmergence because CO2 formed in aerobic respiration escapes from the soilonly very slowly, and it therefore accumulates. As CO2 continues to accumulateduring anaerobic respiration and fermentation, large partial pressures develop,typically in the range 5 to 20 kPa. The accumulation of CO2 lowers the pH ofalkaline soils and curbs the increase in pH of acid soils. As a result the pHs ofall soils tend to converge following submergence in the range 6.5–7.

As the partial pressure of CO2 increases, the concentration of HCO3− in the

soil solution increases and therefore the concentrations of balancing cations insolution increase. Changes in alkalinity and concentrations of cations in solutionfollowing submergence are shown in Figure 4.6. The NH4

+, Mn2+ and especiallyFe2+ ions formed in soil reduction displace exchangeable cations into solution.

Time (weeks after submergence)

Con

cent

ratio

n in

soi

l sol

utio

n (m

mol

c L−1

)

0 2 4 6 8 10 12 14 160

5

10

15

20

25

Ca2+ + Mg2+ +

NH4+ + Na+ + K+

Fe2+ + Mn2+

Alkalinity

Figure 4.6 Changes in alkalinity and concentrations of cations in the soil solution fol-lowing submergence (Ponnamperuma, 1972)

Page 120: The Biogeochemistry of Submerged Soils

112 Reduction and Oxidation

Also, the changes in pH will cause changes in the charges of variable-chargeclays and organic matter, thus the cation exchange capacity of acid soils willtend to increase and that of alkaline soils decrease.

Changes in Fe

Large concentrations of Fe2+ develop in the soil solution in the weeks follow-ing flooding, often several mM or tens of mM (Figure 4.5). Calculations withchemical equilibrium models show that the ion activity products of pure ferroushydroxides, carbonates and other minerals are often exceeded 100-fold (Neueand Bloom, 1989). Evidently precipitation of these minerals is inhibited, proba-bly as a result of adsorption of foreign solutes, such as dissolved organic matterand phosphate ions, onto nucleation sites (Section 3.7). However, once a suffi-cient supersaturation has been reached there is a rapid precipitation of amorphoussolid phases, which may later re-order to more crystalline forms. Only a smallpart of the Fe(II) formed in reduction remains in solution; the bulk is sorbed inexchangeable forms or, eventually, precipitated.

The identities of the solid phases that form remain a mystery. Direct identi-fication is difficult because Fe(II) and Mn(II) solid phases are readily oxidizedby O2 and it is therefore necessary to maintain scrupulously anoxic conditionsto ensure that the material examined actually represents that in anoxic soil. Analternative is to make indirect assessments through measurements of pe, pH and[Fe2+] in solution, but these too are difficult (see section on measurement ofredox potential in soil).

Some of the well-known solid phases that might form are shown in Table 4.3.None of these appears to be quantitatively important, at least in the first few

Table 4.3 Some possible mineral phases in reduced soils and their equilibrium constantsat 25 ◦C

Compound Equilibrium log K

Mn(II) hydroxide Mn(OH)2(s) + 2H+ = Mn2+ + 2H2O 15.13a

Rhodocrosite MnCO3(s) = Mn2+ + CO32− −10.39b

Hauerite MnS2(s) = Mn2+ + S22− −14.79c

Fe(II) hydroxide Fe(OH)2(s) + 2H+ = Fe2+ + 2H2O 11.67a

Fe(II)Fe(III) hydroxide Fe3(OH)8(s) + 2H+ = Fe2+ + 2Fe(OH)3 + 2H2O −10.60d

Siderite FeCO3(s) = Fe2+ + CO32− −10.45b

Vivianite Fe3(PO4)2·8H2O(s) = 3Fe2+ + 2H2PO4− + 8H2O 3.11c

Pyrite FeS2(s) = Fe2+ + S22− −26.93c

Source:a Calculated from �Go

f values.b Stumm and Morgan (1996).c Lindsay (1979).d Arden (1950).

Page 121: The Biogeochemistry of Submerged Soils

Redox Conditions in Soils 113

weeks or months following submergence. The large increases in CO2 pressureas dissolved Mn(II) and Fe(II) accumulate would suggest Mn and Fe carbon-ates should be precipitated. However it is unlikely that simple Mn and Fecarbonates are formed because Mn2+ and Fe2+ ions have similar radii (0.083and 0.078 nm, respectively) and can readily substitute for each other in crystallattices. Rhodocrosite (MnCO3) and siderite (FeCO3) are end members of a con-tinuous series of solid solutions of Fe(II)–Mn(II) carbonates (Deer et al., 1992).Iron–manganese minerals also readily incorporate Mg2+ (radius 0.072 nm) andto a lesser extent Ca2+ (0.1 nm) and other divalent cations. It is therefore likelythat various solid solutions are formed.

There is evidence that mixed Fe(II)–Fe(III) hydroxides are formed. These canbe produced easily in vitro by partial oxidation of pure Fe(II) hydroxy saltsand they have some of the observed properties of the solid phase Fe(II) foundin reduced soils, including the grayish-green colours characteristic of reducingconditions in soils. This material is ‘green rust’ and has the general formulaFe(II)6Fe(III)2(OH)18 with Al3+ partly substituted for Fe3+ and Cl−, SO4

2− andCO3

2− substituted for OH−.Once precipitation begins, a quasi-steady state will eventually be attained in

which the soil pe and pH are poised by the redox and precipitation equilibriaoperating. In the transition to the steady state, protons will be provided by dis-sociation of acids in the soil solution—e.g. H2CO3 derived from CO2–and byreactions with the soil exchange complex. The course of reduction and the even-tual steady state will depend on these reactions and it is therefore necessary toallow for them in predicting what the steady state conditions will be.

In the following section I describe a simple model for calculating the changesin pe, pH and concentrations of inorganic reductants during soil reduction, allow-ing for the effects of pH buffering and cation exchange, and the characteristicsof the mineral phases formed. The approach is based on that of van Breemen(1988) for partial redox equilibrium in soil without pH buffering and cationexchange.

4.2.3 CALCULATED CHANGES IN pe, pH AND Fe DURINGSOIL REDUCTION

Consider an idealized soil containing ferric hydroxide and readily decomposableorganic matter. The following conditions hold:

• the soil is initially saturated with the atmospheric partial pressure of O2 butotherwise closed to exchange of O2;

• the partial pressure of CO2 is constant;• the soil exchange complex is initially saturated with divalent cations M2+, i.e.

H+ is treated as non-exchangeable and there are no other monovalent cations;• the soil reaches a steady state following reduction in which the soil solution

is in equilibrium with Fe(OH)3 and Fe3(OH)8.

Page 122: The Biogeochemistry of Submerged Soils

114 Reduction and Oxidation

Following flooding, O2 dissolved in the soil solution is consumed according toReaction (4.27). There is no pH change, the CO2 pressure being constant, andpe is poised by the O2–H2O couple:

O2(aq) + 4H+ + 4e− = H2O

i.e.pe = 21.45 + 1

4 log[O2]L − pH (4.33)

Once all the O2 has been used up Fe(OH)3 is reduced according to Reaction(4.30) and the pe is poised by the Fe(OH)3–Fe2+ couple:

Fe(OH)3(s) + 3H+ + e− = Fe2+ + 3H2O

i.e.pe = 16.54 − log[Fe2+]L − 3pH (4.34)

Once pe falls sufficiently, precipitation of Fe3(OH)8 commences and Fe3(OH)8

is formed at the expense of Fe(OH)3 according to the reaction

3Fe(OH)3 + CH2O + 8H+ → Fe3(OH)8 + CO2 + 11H2O

The pe is now poised by the Fe(OH)3–Fe3(OH)8 half reaction:

3Fe(OH)3(s) + H+ + e− = Fe3(OH)8 + H2O

i.e.pe = 1.46 − pH (4.35)

Hence for a given generation of Fe2+ in Fe(OH)3 reduction, and for a specifiedinitial soil CEC and concentration of non-carbonate anions in the soil solution([X−]L), we have five unknowns: the soil pH and the concentrations of Fe2+ andM2+ in the soil solid and solution; and these can be found from the followingfive equations:

(1) Equation (3.69) for the electrical neutrality of the solution, with [HCO3−]

found from pCO2 and pH;(2) Equation (3.70) for the electrical neutrality of the solid, with changes in

acidity in the solid related to changes in pH with the soil pH buffer capacity;(3) Equation (3.72) for divalent–divalent cation exchange; and(4) and (5) two equations like Equation (3.73) for conservation of M2+ and Fe2+.

These equations can be solved simultaneously with Equations (4.33)–(4.35) toobtain values of pe, pH, [O2] and [Fe2+] over the course of reduction. Figure 4.7shows results for realistic flooded soil conditions, expressed in terms of theamounts of CH2O oxidized in the different reactions. Figure 4.7(a) gives resultsin the absence of pH and cation buffering by the soil; Figure 4.7(b)–(d) givesresults for different values of bHS, CEC and [X−]L.

Page 123: The Biogeochemistry of Submerged Soils

115

CH

2O o

xidi

zed

(mm

ol k

g−1 )

pe

08

64

210

1214

160

86

42

1012

1416

08

64

210

1214

160

86

42

1012

1416

[Fe2+], [O2] in solution (mM)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

peFe2+

pH

O2

O2

(b)

pH Fe2+

pe

−2024681012141618

pH

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

(d)

(c)

Fe2+

pH pe

Fe2+

pH pe

O2

O2

(a)

b HS =

0, C

EC

= 0

,[X

− ] L =

0b H

S =

15,

CE

C =

35,

[X− ] L

= 2

0b H

S =

50,

CE

C =

35,

[X− ] L

= 2

0b H

S =

50,

CE

C =

10,

[X− ] L

= 1

0

Fig

ure

4.7

Cal

cula

ted

chan

ges

inpe

,pH

,[O

2]

and

[Fe2+

]in

anid

ealiz

edso

ildu

ring

redu

ctio

n.M

iner

alph

ases

Fe(O

H) 3

and

Fe3(O

H) 8

.Pa

ram

eter

sfo

rpH

buff

erin

gan

dca

tion

exch

ange

diff

erbe

twee

n(a

)–(d

)as

indi

cate

d.U

nits

ofb

HS

(soi

lpH

buff

erpo

wer

),m

mol

kg−1

pH−1

;C

EC

(ini

tial

catio

nex

chan

geca

paci

ty),

cmol

ckg

−1;

and

[X− ] L

(con

cent

ratio

nof

non-

carb

onat

ean

ions

),m

M.

CO

2pr

essu

re=

10kP

a,θ

=0.

6,ρ

=1.

0,in

itia

lpH

=4.

5

Page 124: The Biogeochemistry of Submerged Soils

116 Reduction and Oxidation

It will be seen that pe is initially buffered at about 15 until the O2 is exhausted,and it then falls rapidly to the point where it is buffered by the Fe(OH)3 –Fe2+couple. Ferrous ions are now released into the solution and protons removedfrom it, and the changes in pe and pH now depend on the buffering of Fe2+ andH+ by the soil solid. As a result of buffering, the increases in Fe2+ and pH asFe(OH)3 is reduced are more gradual.

Comparing Figure 4.7(b) and (c), for which the CEC and [X−]L are the samebut bHS different, the effect of increasing bHS is to further slow the increase inpH, and the pH at the steady state when Fe3(OH)8 is formed is smaller and thepe and concentration of Fe2+ in solution correspondingly larger. Also a muchlarger quantity of CH2O is consumed in reaching the steady state. From thestoichiometry of Reaction (4.30), the amount of exchangeable Fe2+ formed inmmolc kg−1 is roughly four times the amount of CH2O oxidized.

The effect of varying CEC can be seen by comparing Figure 4.7(c) and (d).With decreasing CEC at constant bHS, [X−]L and PCO2 , the concentration ofFe2+ at steady state is increased and the pH correspondingly decreased andpe increased.

In summary, the calculations predict that:

(1) O2 will disappear rapidly after flooding;(2) dissolved Fe2+ will appear in the soil solution and its concentration increases

to a constant steady-state level;(3) an initially low pH will increase to between 6.5 and 7 at the steady state;(4) pe will decrease from about 15 to near 0;(5) the rates of change in Fe2+, pH and pe and their steady-state values, and the

amounts of organic matter oxidized in reaching the steady state, depend onpH buffering and cation exchange by the soil.

These predictions can be compared with the results for real soils shown inFigure 4.7.

In the real soils the ranges of pe and pH are similar and a steady state is attainedin which the concentrations of Fe2+ in solution are similar to those predicted withthe model. However the large peak in Fe2+ concentrations in some soils beforethe steady state is reached is not predicted. The peak occurs because precipitationof ferrous carbonate is slow and may be inhibited by interfering solutes in thesoil, resulting in supersaturation with respect to the expected solid phases.

Note that although the pe–pH–[Fe2+] relationships shown in Figure 4.7 areconsistent with control by the Fe(OH)3–Fe3(OH)8 system, in fact various otherreduced Fe solid phases are possible and as discussed above it is difficult toestablish unequivocally which phase controls Fe2+ solubility in reduced soils.

4.2.4 MEASUREMENT OF REDOX POTENTIAL IN SOIL

In principle the redox potential provides a simple means of gauging a soil’sredox status. However in practice it is difficult to make reliable measurements.

Page 125: The Biogeochemistry of Submerged Soils

Redox Conditions in Soils 117

Stumm and Morgan (1996) discuss the problems for simple aquatic systemsand van Breemen (1969), Ponnamperuma (1972), McBride (1994) and Patricket al. (1996) discuss the additional problems for soil systems. I here give themain points.

Measurements of EH are usually made with a platinum electrode placed in thesoil solution together with a reference half cell electrode of known potential. Theplatinum electrode transfers electrons to and from the soil solution without react-ing with it. Reducing half reactions in the soil tend to transfer electrons to theplatinum electrode and oxidizing half reactions to remove them. At equilibriumno electrons flow and the electric potential difference between the half cell com-prising the platinum electrode and the soil solution and the half cell comprisingthe reference electrode is recorded.

The first problem to mention is that thermodynamically meaningful measure-ments of EH must be made on soil solution extracts and not directly on the soilitself. Although EH values measured in soil following reduction may show theexpected qualitative trends and expected differences between soils, they are notsatisfactory for quantitative interpretation. Hence duplicate measurements of EH

in soil can vary by as much as 100 mV and values are generally far too lowin terms of the Fe2+ and Mn2+ concentrations measured (IRRI, 1964). This isfirstly because the measurement indicates the potential in the immediate vicinityof the electrode and not that of the whole soil, and there may be large microscalevariations in EH especially near the surfaces of bacterial cells. Secondly theyare subject to liquid junction potential errors. It is therefore necessary to makemeasurements in solution withdrawn from the soil ensuring the minimum of gasexchange during sampling. It is particularly important that no O2 is allowed toenter the solution and that no CO2 is lost. This may be achieved by withdrawingthe solution through porous tubing into previously evacuated tubes.

Apart from these sampling errors there are a number of intrinsic errors in themeasurement of EH in soil solutions. The measurement depends on there beingno net flow of current through the circuit made by the platinum electrode andreference electrode. However the current in one direction, called the exchangecurrent, i0, is not zero. Its value for each half-reaction varies with the electrodepotential and with the concentrations of the oxidant and reductant. Figure 4.8shows this schematically for the Fe2+–Fe3+ couple. As can be seen from thefigure, an infinitesimal shift in the electrode potential from its equilibrium valuewill make the half reactions proceed in one direction or the other and a net currentwill flow through the circuit. The equilibrium potential of the system can be foundfrom the potential at which no net current flows. How precisely and reproduciblythis measurement indicates the equilibrium potential depends on how steeply thenet current deviates from zero near the equilibrium potential. The greater theexchange current, i0, the more steeply the net current varies with the potential.This in turn depends on the redox couple operating and its concentration.

Modern instruments will give reliable measurements for i0 values greater thanabout 0.1 µA. Figure 4.8 shows that for the Fe2+–Fe3+ couple, i0 ≈ 100 µA

Page 126: The Biogeochemistry of Submerged Soils

118 Reduction and Oxidation

−100

525

50

100

475

−50

net current

+i 0

+i 0

−i 0

475

100

50

−50

525

net current

Fe3+→ Fe2+

Fe2+→ Fe3+

Fe3+→Fe2+

Fe2+→ Fe3+

−100

0 0←Potential (mV)

Cur

rent

(µA

)(a) (b)

525

50

500 475

−50 net current

−i 0

Fe3+→ Fe2+

0

500

450 425

Fe2+→ Fe3+

(c)

+i 0

Figure 4.8 Electrode current versus electrode potential curves for the Fe2+–Fe3+ couplein water at pH 2 with (a) [Fe3+] = [Fe2+] = 1 mM; (b) [Fe3+] = [Fe2+] = 0.1 mM;(c) [Fe3+] = 0.1 mM, [Fe2+] = 1 mM. Electrode area = 1 cm2 (Stumm and Morgan,1996). Reproduced by permission of Wiley, New York

for [Fe3+] = [Fe3+] = 10−3 M (Figure 4.8a). If the concentration of both ionsis 10-fold smaller, i0 and the slope are 10-fold smaller (Figure 4.8b). How-ever if the concentration of only one of the ions is decreased the drop in i0

is not as great (Figure 4.8c); note also that the equilibrium potential is shifted.If [Fe3+] = [Fe3+] = 10−7 M, i0 ≈ 0.1 µA and measurements are no longer reli-able. In practice the limiting concentration is nearer 10−5 M because of the effectsof trace impurities. The value of i0 will increase with the surface area of theelectrode. However the benefit of this tends to be offset by greater effects ofimpurities. In the case of the O2–H2O couple, the net current is virtually zeroover a wide range of electrode potentials as shown in Figure 4.9(a). This makes itextremely difficult to determine the equilibrium potential for the O2–H2O couple,and so EH measurements in aerated soils are not reliable.

A further problem, particularly in soil systems, is that several redox systemsmay be present, in which case the apparent equilibrium potential may be the result

Page 127: The Biogeochemistry of Submerged Soils

Transformations of Nutrient Elements Accompanying Changes in Redox 119

+1

1

−1

−1

Potential (V)

Fe3+→Fe2+

Fe2+→Fe3+

H2O→O2

0 0

Cur

rent

(m

A)

(a) (b)

0

O2→H2O

H2O→H2

O2→H2O

Potential (V)

Em Eeq

Figure 4.9 Electrode current versus electrode potential curves for solutions containingO2: (a) in otherwise pure water; (b) in the presence of Fe2+. In (a) the net current isclose to zero over a wide range of potential, so it is difficult to locate the equilibriumpotential. In (b) the measured equilibrium potential is a mixed potential, Em, obscuringthe true equilibrium potential of the system, Eeq (Stumm and Morgan, 1996). Reproducedby permission of Wiley, New York

of the combined exchange currents of two or more redox couples. Figure 4.9(b)illustrates this for the Fe2+–Fe3+ system in the presence of trace concentrationsof dissolved O2. The measured equilibrium potential, Em, at which the net currentis zero may be the potential at which the rate of reduction of O2 at the electrodeequals the rate of Fe2+ oxidation. This would be likely if the concentration ofFe2+ greatly exceeded that of Fe3+, as in general it will in submerged soils. Thetwo couples are not in equilibrium with each other and the measured potentialis termed a mixed potential. The mixed potential does not represent either of theindividual couples operating and is therefore difficult to interpret. Many redoxcouples do not react reversibly at electrode surfaces. Examples are CO2–CH4

and NO3−–N2. This too complicates interpretation.

These factors rather constrain the usefulness of EH measurements in soil solu-tions. Inferences about the thermodynamics of redox processes in soils that relyheavily on measurements of redox potential should be treated with caution.Nonetheless soil EH measurements provide a ready measure of redox status,for example in experiments in which constant EH and pH are required (Patricket al., 1973).

4.3 TRANSFORMATIONS OF NUTRIENT ELEMENTSACCOMPANYING CHANGES IN REDOX

These are briefly discussed here in the context of redox chemistry. More completediscussions are given in Chapters 5–8.

Page 128: The Biogeochemistry of Submerged Soils

120 Reduction and Oxidation

4.3.1 TRANSFORMATIONS OF CARBON

In broad terms the decomposition of organic matter under anaerobic conditionsis expected to be slower than under aerobic conditions because the free energychanges for the reactions involved are much smaller (Table 4.1 and Figure 4.3).For example, for the aerobic decomposition of ‘CH2O’,

14 ‘CH2O’ + 1

4 O2 = 14 CO2(g) + 1

4 H2O

�Go = −119 kJ mol−1 at pH 7, whereas for its anaerobic decomposition inmethanogenesis,

14 ‘CH2O’ = 1

4 CO2(g) + 14 CH4(g)

�Go = −17.7 kJ mol−1 at pH 7. Consequently the microbes mediating the de-composition derive less energy and produce fewer cells per unit of carbonmetabolized. The accumulation of organic matter in marshes and peat bogs illus-trates this point. (But note the rarity of tropical wetland soils with large organicmatter contents, discussed in Section 3.7.)

The most striking difference between anaerobic and aerobic decompositionis in the nature of the end products. In aerobic decomposition the main prod-ucts are CO2, NO3

−, SO42− and resistant residues; in anaerobic decomposition

they are CO2, H2, CH4, N2, NH4+, H2S and various partially decomposed and

humified residues.The decomposition proceeds in two stages. The first involves formation of

organic acids, particularly acetic, propionic and butyric, plus various aliphaticsand phenolics, some of which are toxic to plants. The second involves conversionof organic acids to gaseous products and follows a characteristic pattern. In thefirst few days, H2 formed in fermentation reactions may be evolved together withCO2. Nitrogen gas is also evolved, formed in denitrification of NO3

−. As inor-ganic redox couples then begin to buffer the redox potential, H2 evolution ceasesand CO2 is the main end product of carbohydrate metabolism. This continuesuntil the pe and pH reach values at which methanogenesis is possible, typically1 or 2 weeks after submergence. The concentration of CH4 in the soil solutionand in gas bubbles then exceeds the concentration of CO2 several-fold as a resultof solubility and precipitation effects. Although there is wide variation in thecomposition of gases formed between soils, this general pattern is always seen.At higher temperatures, CO2 and CH4 are formed sooner and at greater rates.Also, at higher temperature and pH, the ratio of CH4 to CO2 in the soil gaseschanges in favour of CH4 because of solubility and precipitation effects and thehigher optimal temperatures for methanogens.

4.3.2 TRANSFORMATIONS OF NITROGEN

The main transformations of N are summarized in Figure 4.10. In the absence ofoxygen, mineralization of organic N proceeds only as far as NH4

+, and NH4+

Page 129: The Biogeochemistry of Submerged Soils

Transformations of Nutrient Elements Accompanying Changes in Redox 121

denitrification

NO3−

NO2

NO2−

NO

N2O

N2

N fixation

nitr

ifica

tion

+5

+4

+3

+2

+1

0

−3

12

−4

Organic NNH4+

Oxi

datio

n st

ate

of N

Red

uctio

n

Oxi

datio

n

pe

mineralization

immobilization

Figure 4.10 Nitrogen transformations in submerged soils on a redox scale (McBride,1994). Reproduced by permission of Oxford University Press

accumulates in the soil solution and exchange complex. Because of the low Nrequirement of anaerobic metabolism, subsequent immobilization by microbestends not to be important, or, if it occurs—as when organic matter with a wideC:N ratio is present—the immobilization is temporary.

Further transformations of N take place at the oxic interfaces between the soiland floodwater and root and soil where NH4

+ diffusing in from the neighbouringanoxic soil may be nitrified to NO3

−. Subsequently, NO3− diffusing out into the

anoxic soil may be denitrified to N2. This process results in significant losses ofN from wet soils but its importance in submerged soils is unclear (Section 5.3).

Under strongly reducing conditions (pe < −4) reduction of N2 to NH4+ is

thermodynamically possible. The net reaction is

16 N2(g) + 1

3 H+ + 14 ‘CH2O’ = 1

3 NH4+ + 1

4 CO2(g)

�Go = −14.3 kJ mol−1 at pH 7. However this reaction has a very large acti-vation energy because of the energy required to break the N≡N triple bond(942 kJ mol−1). Therefore only highly specialized ‘nitrogen fixing’ organisms arecapable of maintaining sufficiently reducing conditions in their cells to mediatethe reaction. The niches in submerged soils in which nitrogen fixers may operateare discussed in Chapter 5.

Most of the mineralizable N in the soil is converted to NH4+ within a few

weeks of submergence if the temperature is favourable and the soil not stronglyacid or deficient in other nutrients. The concentration of NH4

+ in the soil solutiontypically reaches 0.1 to 5 mM buffered by from 5 to 20 times this concentration

Page 130: The Biogeochemistry of Submerged Soils

122 Reduction and Oxidation

0 2 4 6 8 10 12 14 160

2

4

6

8

0

Soil pHOrg C(%)

ActiveFe (%)

2 4 6 8 10 12 14

14

23

26

21

1

392518

23

21

2928

27

1

16

1

25

262714

0

2

4

6

8

10

12

16.5 mMat 0.5 wk

Time (weeks after flooding)

Con

cent

ratio

n of

NH

4+ in

soi

l sol

utio

n (m

M)

Con

cent

ratio

n of

SO

42− in

soi

l sol

utio

n (m

M)

Con

cent

ratio

n of

P in

soi

l sol

utio

n (µ

M)

0 2 4 6 8 10 12 14 160

20

40

60

80

100

120

140(c)

(a) (b)

1 7.6 2.3 0.1814 4.8 2.8 2.1318 5.6 6.0 0.2721 4.6 4.1 2.7823 5.7 8.0 0.4725 4.8 4.4 0.1826 7.6 1.5 0.3027 6.6 2.0 1.6028 4.9 2.9 4.7029 5.8 7.7 1.8039 8.1 2.0 -

Figure 4.11 Changes in (a) NH4+, (b) SO4

2− and (c) P in the soil solution of varioussoils following flooding (modified from IRRI, 1964, 1965). Reproduced by permission ofIRRI

of NH4+ on the soil exchange complex. Figure 4.11 shows changes in NH4

+ insolution following submergence of a range of soils.

4.3.3 TRANSFORMATIONS OF SULFUR

The stable form of sulfur under moderately strong reducing conditions (pe < −3)

is hydrogen sulfide, H2S, which is readily soluble and under non-acid conditions

Page 131: The Biogeochemistry of Submerged Soils

Transformations of Nutrient Elements Accompanying Changes in Redox 123

dissociates to HS−(pK = 7.02). For the reduction of SO42− the net reaction is

14 ‘CH2O’ + 1

8 SO42− + 1

8 H+ = 14 CO2(g) + 1

8 HS−

and �Go = −20.5 kJ mol−1 at pH 7. H2S and HS− are also produced in thehydrolysis of the S-containing amino acids.

The HS− formed further dissociates to S2−(pK = 13.9). However in mostsubmerged soils the concentration of Fe2+ in the soil solution is sufficient thatvirtually all S2− is precipitated as amorphous ferrous sulfide and very smallconcentrations of H2S and HS− remain in solution. The relations between theSO4

2−–HS− and Fe(OH)3–Fe2+ systems at neutral pH are shown in Figure 4.12.Amorphous ferrous sulfide may gradually crystallize as mackinawite (FeS). Undersome circumstances pyrite is then formed, e.g. FeS(s) + S(s) → FeS2(s), leadingto potential acid sulfate soils (Section 7.3).

There may be a cycling of S compounds of different oxidation state betweenanaerobic and aerobic zones in the soil, such as at the soil—floodwater interface.In reduced lake and marine sediments this leads to accumulation of insolublesulfides as SO4

2− carried into the sediment from the water above is immobilized.Such deposits function as sinks for heavy metals. Plants absorb S through theirroots as SO4

2−; H2S is toxic to them. Therefore HS− must be oxidized to SO42−

in the rhizosphere before it is absorbed.Figure 4.12 shows changes in the concentration of SO4

2− in the soil solutionfollowing submergence of a range of soils. In neutral and alkaline soils con-centrations of SO4

2− greater than 10 mM may decrease to 0 within 6 weeks ofsubmergence. In acid soils the concentration of SO4

2− in solution may initiallyincrease following submergence and then slowly decline over several months.

pe−8 −6 −4 −2 0 2 4 6 8 10

−12

−10

−8

−6

−4

−2

0

H2S + HS− SO42−

Fe2+

FeS

log

conc

entr

atio

n (M

)

FeCO3 Fe(OH)3

Figure 4.12 Concentration–pe diagram for FeS, FeCO3 and Fe(OH)3 at pH = 7, CT

(total carbonate carbon) = 5 mM and [SO42−] + [H2S(aq)] + [HS−] = 1 mM (modified

from Stumm and Morgan, 1996). Reproduced by permission of Wiley, New York

Page 132: The Biogeochemistry of Submerged Soils

124 Reduction and Oxidation

The initial increase occurs because SO42− sorbed on variable charge clays and

oxides is desorbed as the pH increases. The rate of subsequent reduction will below if the pH remains below 5.5, the optimal range of pH for SO4

2− reducingbacteria being greater than this.

4.3.4 TRANSFORMATIONS OF PHOSPHORUS

Phosphorus is often the most limiting nutrient in natural wetlands. Because ofits association with soil Fe, its solubility changes markedly during reduction andoxidation. In general it is not itself reduced and remains in the +5 oxidationstate, though production of phosphine gas (PH3; +3 oxidation state) at rates ≤6.5 ng m−2 h−1 has been reported in laboratory experiments with brackish andsaline marsh soils (Devai and Delaune, 1995). Review articles on transformationsof P in submerged soil include Patrick and Mahapatra (1968), Kirk et al. (1990a)and Willett (1991).

Typically when a soil is submerged the concentrations of water- and acid-soluble P increase, reach a peak or plateau, and then decrease (Figures 4.11c and4.13). For the soils shown in the figures, the peak P concentrations in solutionwere smallest for acid soils high in active Fe and greatest for a sandy soil low inFe. The increases in acid-soluble P were greatest in an alkali soil low in active

Time (weeks after submergence)

Con

cent

ratio

n of

aci

d-so

lubl

e P

in s

oil

(mm

ol k

g−1)

0 2 4 6 8 10 120.0

0.2

0.4

0.6

0.8

1.0

1.2

26

27

18

21

28

14

Figure 4.13 Changes following flooding in the concentration of P soluble in an acetatebuffer at pH 2.7. Numbers next to curves identify soils; properties given in table inFigure 4.11 (modified from Ponnamperuma, 1985). Reproduced by permission of IRRI

Page 133: The Biogeochemistry of Submerged Soils

Transformations of Nutrient Elements Accompanying Changes in Redox 125

Fe, intermediate in sandy loams high in organic C and low in active Fe, and leastin acid clays high in active Fe. The increases in soluble P are particularly linkedto the transformations of Fe and changes in pH. The main processes are:

• reduction of Fe(III) compounds holding P on their surfaces and within theircrystal lattices;

• dissolution of Ca-P compounds in alkaline soils as the pH decreases and des-orption of P held on variable-charge surfaces in acid soils as the pH increases;

• displacement of sorbed P by organic anions and chelation of metal ions thatwould otherwise immobilize P; and

• mineralization of organic P.

Subsequent decreases in solubility may be due to re-sorption or precipitation onclays and oxides as soil conditions continue to change, and decomposition oforganic anions chelating P or chelating Al and Fe with which it would other-wise react.

Following submergence soils often release more P to solutions low in P butadsorb more P from solutions high in P. This apparent paradox can be explainedby the reduction of Fe(III) oxides to poorly ordered gel-like Fe(II) compoundswith large surface areas. Phosphorus solubilized in soil reduction is sorbed onthe amorphous surfaces and desorbed when P is removed from the soil solution;but fresh P added to the soil is removed from solution by sorption onto theFe(II) surfaces. Consequently many soils do not show significant increases in Psolubility during flooding (Willett, 1991), and with prolonged flooding the P maybecome re-immobilized in less soluble forms.

Gradual immobilization of P with prolonged anaerobicity is shown inFigure 4.14, which gives change in labile P over 3 years of double rice croppingof a perennially wet soil (Bucher, 2001). The labile P declines even in plotsthat received sufficient P fertilizer to more than off-set crop removals. Periodicdrying of the soil during the fallow periods tended to increase labile P in the soil,but not in years when the soil remained anaerobic during the fallow (followingthe 1998 and 1999 wet season crops). The effect was greatest when tillage wasdelayed until the end of the fallow, resulting in more-reducing conditions in thesoil, and it carried through to the succeeding rice crop. Supporting laboratory andgreenhouse studies showed that changes in soil Fe with reduction and oxidationwere responsible for the changes in P.

Rapid drying and oxidation of the soil can also result in the P becomingvery insoluble (Brandon and Mikkelsen, 1979; Willett, 1979; Sah et al., 1989;Huguenin-Elie et al., 2003). Re-oxidized Fe(II) compounds may be precipitatedin poorly crystalline forms with large specific surface areas, on and in which Pbecomes immobilized. Hence upland crops grown in rotation with rice frequentlysuffer P deficiency even though crops on similar soils not used for rice growhealthily. The problem is in part also due to disruption of mycorrhizal networksduring flooding (Ilag et al., 1987; Ellis, 1998; Miller, 2000).

Page 134: The Biogeochemistry of Submerged Soils

126

Tim

e

4 Dat

21 Dat

42 Dat

63 Dat

2 Dat

21 Dat

42 Dat

63 Dat

Harvest

Mid-fallow

End of fallow

Harvest

Mid-fallow

End of fallow

2 Dat

21 Dat

42 Dat

63 DatHarvest

Mid-fallow

End of fallow2 Dat

21 Dat

42 Dat

63 DatHarvest

2 Dat

21 Dat

42 Dat

63 DatHarvest

Mid-fallow

End of fallow

051015N

K p

lots

0510152025N

PK

plo

ts

Late

tilla

ge,

with

str

awLa

te ti

llage

,no

str

aw

Ear

ly ti

llage

,w

ith s

traw

Ear

ly ti

llage

,no

str

aw

Fal

low

1998

DS

Fal

low

1998

WS

Fal

low

1999

DS

Fal

low

1999

WS

Fal

low

2000

DS

Resin-extractable P (mg kg−1 dry soil)

Fig

ure

4.14

Cha

nges

inla

bile

soil

P(e

xtra

ctab

lew

ithH

CO

3− -f

orm

anio

nex

chan

gere

sin)

duri

ng3

year

sof

wet

land

rice

crop

ping

asaf

fect

edby

timin

gof

tilla

ge(e

arly

,la

te=

star

t,en

dof

fallo

w),

inco

rpor

atio

nof

prev

ious

crop

’sst

raw

,an

dap

plic

atio

nof

P(2

0kg

ha−1

inN

PKpl

ots)

.T

heov

eral

lP

bala

nces

over

3ye

ars

wer

e+3

7an

d+7

kgP

ha−1

inth

eN

PKpl

ots

with

and

with

out

stra

w,

and

−90

and

−115

kgP

ha−1

inth

ePK

plot

s.D

S,W

S,dr

y,w

etse

ason

;D

AT

,da

ysaf

ter

tran

spla

ntin

g(B

uche

r,20

01).

Rep

rodu

ced

bype

rmis

sion

Page 135: The Biogeochemistry of Submerged Soils

Oxidation of Reduced Soil 127

4.4 OXIDATION OF REDUCED SOIL

When a spadeful of wet, anaerobic soil is brought to the surface and allowedto dry, air enters through drying cracks and the soil tends to become uniformlyoxidized and turn a uniform brown. Whereas when oxidation occurs withoutdrying—as, for example, near a root releasing O2 into wet soil—it is far lessuniform and reddish-brown ferric oxide deposits form on and near the oxidizingsource. The difference depends on the relative rates of movement of O2 into thesoil and of ferrous iron and other reductants in the opposite direction, and therates of reaction.

Figure 4.15 indicates the range of rates of O2 consumption in different soils.Oxygen is consumed in oxidation of inorganic reductants, such as Fe(II), aswell as in oxidation of organic matter by microbes. Bouldin (1968) and Howelerand Bouldin (1971) compared measured rates of O2 movement into anaerobicsoil cores with the predictions of various models, and obtained the best fitswith a model allowing for both microbial respiration and abiotic oxidation ofmobile and immobile reductants; abiotic oxidation accounted for about half theO2 consumed. The kinetics of the abiotic reactions are complicated. They oftendepend on the adsorption of the reductant on solid surfaces as, for example, in

Time (h)

0 20 40 60 80

[O2]

/[O2]

initi

al

0.01

0.1

1

6.2 1.60 42.3

6.6 2.30 39.6

5.9 0.82 33.6

6.8 1.71 18.9

5.6 0.72 26.9

5.6 1.01 5.3

7.6 0.54 16.5

pH Org C [Fe2+]

(%) (µmol g−1)

Figure 4.15 Rates of oxygen consumption by shaken suspensions of anaerobic soils.Points are measured data, lines are fits to two first-order rate equations. The apparentrate constant for the initial reaction is common to all soils; that for the main reactionvaries 30-fold between the soils and is well correlated with [Fe2+] (Reddy et al., 1980).Reproduced by permission of Soil Sci. Soc. Am.

Page 136: The Biogeochemistry of Submerged Soils

128 Reduction and Oxidation

the autocatalysis of Fe2+ oxidation by adsorption of Fe2+ on ferric hydroxideformed in the reaction. The adsorption is likely to be pH-dependent, a decreasein pH tending to decrease sorption and increase the concentration of Fe2+ insolution. Hence there may be complex interactions between the mobility of Fe2+,the rate of oxidation, and pH changes caused by the reaction. Such interactionscan produce banded distributions of iron around an O2 source, as found, forexample, by Saleque & Kirk (1995) for the distribution of iron near rice rootsand calculated by Kirk et al. (1990) with a model of the coupled diffusion andreaction of O2, Fe2+ and acidity in soil. This is an example of the Liesegangphenomenon (Stern, 1954; Keller, 1980).

4.4.1 KINETICS OF Fe2+ OXIDATION

Aqueous Solution

The reaction between Fe2+ and O2 to form insoluble ferric hydroxide can bewritten

4Fe2+ + O2 + 10H2O = 4Fe(OH)3 + 8H+ (4.36)

Equation (4.36) shows that two H+ ions are produced for each mole of Fe2+oxidized, i.e. the reaction is accompanied by acidification. In aqueous solution,the rate is found to be very sensitive to pH and at near neutral pH the reactionis accelerated 100-fold if the pH is raised by one unit. The following empiricalrate law applies in the pH range 5–8 (Stumm and Lee, 1961; Wehrli, 1990)

−d[Fe(II)]/dt = k[O2][OH−]2[Fe(II)] (4.37)

where k ≈ 2 × 1014 mol3 dm−9 s−1 at 25 ◦C and [Fe(II)] is the sum of theconcentrations of Fe(II) species present—Fe2+ and its hydroxy complexes,FeOH+ and Fe(OH)2, for which the formation constants are 10−4.5 mol−1 dm3

and 10−7.4 mol−2 dm6, respectively. Therefore [Fe(II)] ≈ [Fe2+], but the pHdependence of the rate is due to the parallel oxidation of the three species.At [O2] = 0.28 mM (i.e. in equilibrium with atmospheric PO2 ), the half time forthe reaction is 0.34 h at pH 7 and 143 days at pH 5.

As discussed in Section 4.1, most redox reactions reach equilibrium onlyslowly if they are not catalysed. Oxidation of Fe2+ is catalysed by adsorption ofFe2+ onto Fe(OH)3 formed in the reaction, so Equation (4.36) only holds for theinitial rates of reaction. Tamura et al. (1976) studied the oxidation of a solutionof Fe2+ at different controlled pHs near neutral and with varying additions ofFe(OH)3. The reaction obeyed the rate law

−d[Fe2+]/dt = k[O2][OH−]2[Fe2+] + kS[O2][Fe2+]ad (4.38)

where [Fe2+] is the concentration in solution, [Fe2+]ad the concentration adsorbedon Fe(OH)3 and kS the rate constant for oxidation of adsorbed Fe2+ (= 73 mol−1

Page 137: The Biogeochemistry of Submerged Soils

Oxidation of Reduced Soil 129

dm3 s−1, with all concentrations in mol dm−3 suspension). Adsorption isdescribed by

[Fe2+]ad/[Fe2+] = K[Fe(III)]/[H+] (4.38a)

where [Fe(III)] is the concentration of Fe(OH)3 and K = 10−14.3. Other metaloxidation reactions catalysed by sorption onto oxide surfaces are described inSection 7.3.

Soil

A similar catalysis occurs on soil surfaces. Ahmad and Nye (1990) and Kirkand Solivas (1994) studied the kinetics of Fe2+ oxidation in soil suspensions bymeasuring changes in extractable Fe2+ in the whole soil and in solution duringoxidation at constant [O2]. They found that 75 % of the initial Fe2+ was oxidizedrapidly (t1/2 ≈ 2 h) and the remainder only very slowly (t1/2 ≈ 8 days). In thesoils studied, the pH fell from near neutral to less than 5 over the course of the fastreaction. Measurements of the fast reaction at constant pH (Figure 4.16) showedthat the oxidation of adsorbed Fe2+ was much faster than solution Fe2+, andthat the adsorbed Fe2+ was oxidized at a rate that was nearly independent of pH.Figure 4.16 shows that the overall rate of oxidation is more dependent on the con-centration of sorbed Fe2+([Fe2+]S) than the concentration in solution ([Fe2+]L).Thus, although at pH 6.5 [Fe2+]L drops to one-tenth of its initial value within 1 h,d[Fe2+]/dt does not decrease to nearly the same extent. The figure also shows thatoxidation of sorbed Fe2+, indicated by the slopes of the lines in Figure 4.16(c),is surprisingly little influenced by pH and roughly follows first-order kinetics.

The overall rate equation at constant pH is therefore

−d[Fe2+]/dt = RkL[O2]L[Fe2+]L + kS[O2]L[Fe2+]S (4.39)

where kL and kS are the rate constants for the reactions in the solution andsolid and R the solution to solid ratio. Values of kS, calculated from the data inFigure 4.16(c) range from 0.19 mol−1 dm3 s−1 at pH 6.5 to 0.15 mol−1 dm3 s−1

at pH 5.Initially, most of the readily oxidizable Fe(II) is sorbed on the soil exchange

complex. As the soil is oxidized, the Fe(OH)3 formed provides fresh sorptionsites, as well as possibly blocking some of the original sites. Ahmad and Nye(1990) estimated the importance of the freshly formed Fe(OH)3 in sorbing Fe2+compared with the original soil exchange complex, and found that the importanceof the freshly formed Fe(OH)3 was much greater at higher pH, consistent withthe expected greater pH-dependence of sorption on Fe(OH)3 surfaces.

They also found that the oxidation of Fe2+ when sorbed on a mixture of soilexchange and Fe(OH)3 sites was much slower than on Fe(OH)3 in the absenceof soil, described by Equations (4.38) and (4.38a).

Page 138: The Biogeochemistry of Submerged Soils

130 Reduction and Oxidation

Time (min)[F

e2+] L

(m

M)

[Fe2+

] (m

mol

kg−1

)

Time (min)

Time (min)

0 50 100 150 200 250 0 50 100 150 200 25030

45

60

75

90

100

4.5

5.0

6.5

6.0

0.1

1

10

100

6.5

6.0

5.5

5.0

4.5

0.3

3

30

(a)

[Fe2+

] s (

mm

ol k

g−1)

(b)

0 50 100 150 200 250

15

30

45

60

7590

6.0

4.5

5.0

(c)

Figure 4.16 Changes in concentrations of Fe2+ in (a) whole soil, (b) soil solution and(c) soil solid during oxygenation of reduced soil suspensions at different pHs. [Fe2+]S wascalculated from [Fe2+]–R[Fe2+]L (Kirk and Solivas, 1994). Reproduced by permissionof Blackwell Publishing

A possible explanation is that access of O2 to the exchange or Fe(OH)3 siteswhere the Fe2+ is adsorbed is restricted. Possibly Fe(OH)3 is precipitated betweenclay lamellae at the oxidation sites and it partially blocks the original exchangesites. This mechanism would also imply a wide range of reaction rates betweensoils, with kS values ranging by perhaps an order of magnitude, as in Figure 4.15.

In summary, the reaction can be represented by the following simplified scheme:

Fe2+L ⇀↽ Fe2+

SkS−−−→ Fe(OH)3

Page 139: The Biogeochemistry of Submerged Soils

Oxidation of Reduced Soil 131

in which the exchange of Fe2+ between the solid and liquid is rapid and theoverall rate depends only on the approximately first-order oxidation of sorbedFe2+. Although this rate is independent of pH, the distribution of Fe2+ betweenthe solid and liquid is not and it is therefore necessary to allow for pH changesin calculating the rate.

4.4.2 SIMULTANEOUS DIFFUSION AND OXIDATION IN SOIL

Kirk et al. (1990b) and Kirk and Solivas (1994) used the above understandingof oxidation kinetics to develop a model of soil oxygenation. The model allowsfor the diffusion of O2 into the soil, the diffusion of Fe2+ towards the oxidizingsurface, the rate of formation and concentration profile of the Fe(OH)3 formed,and the diffusion by acid–base transfer of the acidity formed: H3O+ diffusingaway from the zone of acidification and HCO3

− (derived from CO2) towards it.The principal equations are as follows, expressed in planar geometry so as to beable to test the predictions against experimentally measured reactant profiles.

(1) For the diffusion and reaction of O2:

∂[O2]

∂t= ∂

∂x

(DLOθf

∂[O2]L

∂x

)− 1

4S1 − S2 (4.40)

where [O2] and [O2]L are the concentrations of O2 in the whole soil andsolution, respectively, S1 is the rate of Fe2+ oxygenation, S2 is the rate of O2

consumption in microbial respiration, and the other parameters are as definedin Chapter 2.

(2) For the diffusion and reaction of Fe2+:

∂[Fe2+]

∂t= ∂

∂x

(DLIθf

∂[Fe2+]L

∂x

)− S1 (4.41)

where [Fe2+] and [Fe2+]L are the concentrations of mobile Fe2+ in the wholesoil and solution, respectively.

(3) For the diffusion and reaction of soil acidity (Section 2.2)

∂[HS]

∂t= − ∂

∂x

{2.303θf (DLH[H3O+]L + DLC[HCO3

−]L)∂pH

∂x

}+ 2S1

(4.42)

where [HS] is the concentration of titratable soil acid.

Applying Equation (4.39) to the structured-soil system, and ignoring the slowoxidation of Fe2+ in solution, gives

S1 = ρkS[O2]L[Fe2+]S (4.43)

Page 140: The Biogeochemistry of Submerged Soils

132

02

46

810

01020304050607080

3 da

ys6

days

9 da

ys

Dis

tanc

e fr

om s

urfa

ce e

xpos

ed to

O2

(mm

)

[Fe(II)] (mmol kg−1)

[Fe(III)] (mmol kg−1)

pH

02

46

810

0255075100

125

02

46

810

4.0

4.5

5.0

5.5

6.0

6.5

7.0

(a)

(b)

(c)

mm

02

46

810

020406080100

120

9 da

ys

3 da

ys

[Fetotal] (mmol kg−1)F

igur

e4.

17Pr

ofile

sof

(a)

Fe(I

I),

(b)

Fe(I

II)

and

(c)

pHin

colu

mns

ofre

duce

dso

ilex

pose

dto

O2

aton

een

dfo

rdi

ffer

ent

times

.Po

ints

are

expe

rim

enta

llym

easu

red;

lines

are

pred

icte

dus

ing

the

mod

elde

scri

bed

inth

ete

xtw

ithin

depe

nden

tlyes

timat

edpa

ram

eter

valu

es(K

irk

and

Soliv

as,

1994

).R

epro

duce

dby

perm

issi

onof

Bla

ckw

ell

Publ

ishi

ng

Page 141: The Biogeochemistry of Submerged Soils

Oxidation of Reduced Soil 133

where ρ is the soil bulk density. The rate of microbial O2 consumption isdescribed by a Michaelis–Menten type equation:

S2 = ρvmax[O2]L/(KM + [O2]L) (4.44)

Kirk and Solivas (1994) measured profiles of Fe(II) and Fe(III) concentrationsand pH in columns of reduced soil exposed to O2 at one end and compared theresults with the predictions of the model using independently measured parametervalues. The agreement between the observed and calculated results, shown inFigure 4.17, is good. The measured profiles of [Fe(II)] (Figure 4.17a) are scat-tered, probably because of the spatial variability inherent in soil reduction andthe clustering of microbes around favourable microsites. There was much lessscatter in the Fe(OH)3 and pH profiles which are the result of abiotic reactions.The zone of Fe(II) depletion extends further than the zone of Fe(OH)3 accu-mulation, as expected because Fe2+ is mobile but Fe(OH)3 is not. As a result,Fe(OH)3 accumulated in the oxidation zone close to the source of O2, as shownin the inset in Figure 4.17(b).

The good agreement between observed and calculated results and the fact thatthe model contains no arbitrary fitting parameters show that the important pro-cesses are well understood and that the model provides a satisfactory descriptionof the system. It can therefore be used to explore other conditions through a sen-sitivity analysis (Figure 4.18). The figure shows that over the range of parametervalues expected for submerged soils, substantial amounts of iron are transferredtowards the O2-exposed surface leading to a well-defined zone of Fe(OH)3 accu-mulation. For a given soil Fe(II) content, the accumulation is sensitive to the soilFe2+ buffer power, the oxidation rate constant and the soil bulk density. The fallin pH in the oxidation zone is sensitive to the initial soil pH, the soil pH bufferpower, and the partial pressure of CO2.

By contrast if the soil dries to any extent resulting in partially air-filled pores,the penetration of O2 increases dramatically: for an air-space of just 1 % of total

r

0.14

0.12

0.10

Spr

ead

oxid

atio

n fr

ont (

cm)

0.08

0.06

0.04

0.020.01 0.1 1 10 100

[Fe]

ks

6

5

4

3r

2

1

00.01 0.1 1 10 100

Fe

tran

sfer

red

(mol

cm

−2)

[Fe]

ks

bFe

Multiple of standard parameter value

r

pH a

t soi

l sur

face

0.013.0

3.5

4.5

5.5

4.0

5.0

0.1 1 10 100

[Fe]

[H+]

ks

PCO2

bHs

Figure 4.18 Sensitivity of the model used for the calculations in Figure 4.17 to itsparameters: [Fe] is the initial concentration of mobile Fe2+, bFe is the soil Fe2+ bufferpower, bHS is the soil pH buffer power, kS is the Fe2+ oxidation rate constant and ρ isthe soil bulk density. Standard values as for calculations in Figure 4.17

Page 142: The Biogeochemistry of Submerged Soils

134 Reduction and Oxidation

porosity, Kirk et al. (1990b) calculated a four-fold increase in the penetration ofO2. There is then little accumulation of iron in the oxidation zone because the O2

diffuses so much faster than the Fe2+ that almost no Fe2+ can move towards theoxidation front before it is oxidized. The generation of acidity is correspondinglydispersed through the soil. These conclusions are discussed further in Chapter 6in relation to the rhizospheres of wetland plants.

Page 143: The Biogeochemistry of Submerged Soils

5 Biological Processes in the Soiland Floodwater

The soil and floodwater in wetlands are busy with life, and this drives the biogeo-chemistry. The remarkable long-term productivity of wetland rice systems dependson the fixation of carbon and nitrogen from the atmosphere by organisms in the soiland water, for which conditions are optimal. For example, in a long-term experi-ment at the International Rice Research Institute in the Philippines in which threecrops of rice have been grown each year for 30 years without additions of fertil-izers or manures and with complete removal of the rice straw, grain yields haveremained nearly constant at 3 to 3.5 t ha−1 per crop or a total of 9 to 10 t ha−1

per year (Dobermann et al., 2000). No other intensive agricultural system withoutartificial inputs of nutrients comes close to this level of productivity. The accumula-tion of nitrogen by crops in this experiment has remained constant at about 50 kg Nha−1 per crop, largely due to additions from biological fixation in the floodwaterand floodwater–soil interface (Ladha et al., 2000). Comparable rates of nitrogenfixation are attained in other fluxial wetland systems (Table 1.5).

This chapter describes the important micro- and macrobiological processes insubmerged soil and the overlying floodwater. Processes in plants and their rhizo-spheres are discussed in Chapter 6. The microbiological processes are discussedfirst and then the additional complexities caused by macrobiological processesand the particular ecology of the floodwater–soil system.

5.1 MICROBIOLOGICAL PROCESSES

Descending through the soil from the floodwater there is a gradient of redoxpotential and a sequence of zones characterized by progressively more-reducedelectron acceptors. Figure 5.1 shows hypothetical concentration profiles of redoxspecies with depth. At sufficient depth the only electron acceptors are CO2 andH+, and this zone is dominated by fermentation and methanogenesis. At inter-mediate depths there are successive zones of sulfate reduction, iron reduction,manganese reduction and denitrification. The microbes mediating these pro-cesses are largely prokaryotic; populations of fungi and other eukaryotes thatare important in digesting organic matter under aerobic conditions are much lesssignificant in anaerobic soil.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 144: The Biogeochemistry of Submerged Soils

136 Biological Processes in the Soil and Floodwater

Concentration

NH4+

CH4

Fe2+

Mn2+

O2

NO3−

Dep

th

Figure 5.1 Indicative concentration profiles of redox species with depth in sub-merged soil

5.1.1 PROCESSES INVOLVED IN SEQUENTIAL REDUCTION

The sequence of reactions by which organic matter is oxidized following sub-mergence loosely follows the predictions of thermodynamics—i.e. in the orderof decreasing free energy change—as described in Chapter 4. However, ratesof reduction vary greatly between soils and there are complicated interactionsbetween the microbial processes involved. Hence it is difficult to predict a priori,for example, how long after submergence a given soil will become methanogenicand what the rate of methane production will be.

The free energy change for a particular redox reaction varies with pe, pH, andthe concentrations of reductants and oxidants according to Equation (4.26):

�G = −2.303RT

[n(pe0

1 − pe02) − log

(Ox2)(Red1)

(Ox1)(Red2)

]

In this equation, the value of (Red) is a function of the nature of the reductant,its solubility, the crystallinity of solid phases containing it, effects of solubilizingagents, transport limitations, and other factors. Likewise the value of (Ox) is afunction of various factors. As discussed in the previous chapter, most redoxreactions are very slow and the prevailing conditions are therefore sensitive tocatalysis. Three types of catalysis are involved:

• Abiotic, for example by adsorption of reactants onto mineral surfaces, dis-tinguished from biotic catalysis by the absence of a temperature optimum.

Page 145: The Biogeochemistry of Submerged Soils

Microbiological Processes 137

Abiotic catalysis is generally less important than biotic but may be important.Examples are Mn(III,IV) and Fe(III) reduction by microbial metabolites, andFe(II) oxidation which is catalysed by sorption onto soil particles.

• Abiontic, involving free extracellular enzymes or solubilizing agents, enzymesbound to soil surfaces, enzymes within dead or non-proliferating cells, orenzymes associated with dead cell fragments. Extracellular enzymes are impor-tant in the initial stages of organic matter oxidation, in which polysaccharidesand proteins are hydrolysed to soluble compounds that can be absorbed bymicrobial cells and further oxidized in biotic processes.

• Above all, biotic catalysis by microbes is important.

Biotic catalysis is complicated. Different communities of microbes deal with dif-ferent parts of the sequence of processes degrading organic matter. Anaerobicdecomposition involving organic electron acceptors (i.e. fermentation) generallyoccurs concurrently with respiration involving inorganic electron acceptors, andboth produce intermediates that act as both oxidants and reductants. There areoften syntrophic relationships between microbes in which the metabolisms of twoor more organisms are linked and mutually beneficial. For example, in methano-genesis, oxidation of fatty and amino acids to H2, CO2 and acetate is endergonicunder standard conditions (i.e. PH2 = 1 atm), but a sufficiently small concentra-tion of H2 is maintained locally by methanogens that utilize H2 (Conrad et al.,1986; Zehnder and Stumm, 1988; Krylova and Conrad, 1998). Likewise there areantagonisms between microbes, for example where one microbe maintains theconcentration of a substrate below the threshold of a competitor, such as in theinhibition of methanogens by SO4

2− reducers competing for H2 (Achtnich et al.,1995). There are also specific inhibitory effects through particular metabolites,such as in the inhibition of methanogens by denitrifiers (Roy and Conrad, 1999).Hence the initial microbial populations, growth rates and community structuresmay all be important in the overall course of reduction.

The main pathways of organic matter oxidation in anaerobic soil are as fol-lows. In the initial stages, fermenting bacteria excrete extracellular enzymes thathydrolyse polysaccharides and proteins to soluble compounds. These may then beabsorbed by microbial cells and converted to alcohols, fatty acids and H2. If inor-ganic electron acceptors are available, the alcohols and fatty acids are completelyoxidized to CO2 in sequential reduction reactions. If inorganic electron acceptorsare not available—whether because they have been exhausted or because theyare otherwise inaccessible—communities of fermenting bacteria decompose thealcohols and fatty acids to acetate, H2 and CO2. These then serve as substratesfor methanogenic archaea. Sugar monomers may also be directly converted toacetate by homacetogenic bacteria. Likewise proteins are hydrolysed to aminoacids by extracellular enzymes, and the amino acids then ultimately oxidized toacetate, H2, NH4

+ and CO2.Figure 5.2 shows the sequential reduction of inorganic electron acceptors and

production of CO2, CH4 and intermediaries in two representative soils from a

Page 146: The Biogeochemistry of Submerged Soils

138S

oil N

o. 8

(hig

h la

bile

C,

low

Fe)

CH4 or CO2 (kPa)

CH4 or CO2 (kPa)

0246810121416

H2 (kPa)

H2 (kPa)

01020304050

H2

H2

CH

4

CH

4

CO

2

CO

2

Acetate (µmol g−1)

Acetate (µmol g−1)

0

200

400

600

800

1000

1200

−100

0100

200

300

400

500

pH

5.0

5.5

6.0

6.5

7.0

7.5

Acid-soluble Fe(II) (µmol g−1)

Acid-soluble Fe(II) (µmol g−1)

0255075100

125

150

175

200

SO42− (µmol g−1

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

NO3− (µmol g−1

)

012345

pH

Fe(

II)

NO

3−

NO

3−

SO

42−

0246810121416

012345

0

200

400

600

800

−100

0100

200

300

400

500

600

pH

5.5

6.0

6.5

7.0

7.5

Tim

e af

ter

subm

erge

nce

(day

s)

020

4060

8010

012

0

Tim

e af

ter

subm

erge

nce

(day

s)

020

4060

8010

012

0

0255075100

125

150

175

200

225

250

275

SO42− (µmol g−1

)

0.0

0.5

1.0

1.5

2.0

NO3− (µmol g−1

)

012345

pH

Fe(

II)

SO

42−

Soi

l No.

14

(low

labi

le C

, hi

gh F

e)

EH (mV)

EH (mV)

Ace

tate

EH

EH

Ace

tate

Fig

ure

5.2

Sequ

entia

lre

duct

ion

ofel

ectr

onac

cept

ors

and

accu

mul

atio

nof

CO

2an

dC

H4

intw

ori

ceso

ils.

The

soils

wer

esu

bmer

ged

and

incu

bate

dat

30◦ C

inse

aled

bottl

es(Y

aoet

al.,

1999

).R

epro

duce

dw

ithki

ndpe

rmis

sion

ofK

luw

erA

cade

mic

Publ

ishe

rs

Page 147: The Biogeochemistry of Submerged Soils

Microbiological Processes 139

sample of 16 rice soils studied by Yao et al. (1999). Three distinct phases canbe distinguished:

(1) an initial reduction phase lasting 19–75 days in the 16 soils, during whichmost of the inorganic electron acceptors are depleted and the rate of CO2

production, given by the slope of the CO2 accumulation line in the figure,is maximal;

(2) a methanogenic phase starting after 2–87 days and lasting 38–68 days, duringwhich the rate of CH4 production is maximal; and

(3) a pseudo steady-state phase during which rates of CH4 and CO2 productionand concentrations of H2 and acetate are roughly constant.

The line of H2 accumulation in the figure is informative because H2 is turnedover rapidly as it is produced in fermentation and consumed in Fe(III) and SO4

2−reduction and methanogenesis. Hence there are peaks in H2 pressure in the earlystages of Fe(III) and SO4

2− reduction and again at the transition from Fe(III)and SO4

2− reduction to methanogenesis. Because consumption tends to increasewith the concentration of H2 but production is independent of it, there is apoint at which consumption equals production, characterized by H2 concentra-tions in the nM range. Acetate is also produced in fermentation and consumedin methanogenesis, but its turnover is slower and larger concentrations build up.

Figure 5.3 compares the quantities of electrons consumed in reduction of inor-ganic electron acceptors and methanogenesis in the 16 soils with those donated inthe oxidation of organic matter to CO2. At the end of the initial reduction phase,the former exceeded the latter in nine of the soils, probably in part because CO2

was precipitated in carbonates and in part because some of the organic carbon wasconverted to forms more oxidized than that in CO2. However by the end of theincubation the electron balance was zero in all but three of the soils. At the endof the incubation, only 6–17 % of the organic carbon in the soils was released asgases: 61–100 % as CO2, <0.1–35 % as CH4 and <5 % as non-methane hydro-carbons. Most of the CO2 was produced in Fe(III) reduction during the initialreduction phase.

Yao and Conrad (1999) calculated the free energy changes in methanogenesisduring the three phases above. During the initial reduction and while redox poten-tials were still positive (360–510 mV), acetate and H2 concentrations allowedexergonic methanogenesis with �G < −30 kJ mol−1CH4. After about 4 daysCH4 accumulation slowed and ceased in most soils. At this time CH4 partialpressures were still small (about 10–100 Pa), but H2 depleted by Fe or S reduc-tion and �G increased to −10 kJ mol−1 CH4, indicating that methanogenesis wasnot possible. At the end of Fe and S reduction, �G decreased to < − 25 kJ mol−1

CH4 and CH4 production resumed. Vigorous CH4 production continued until thepseudo steady state was reached. In a few soils the initial CH4 production was notinterrupted by an intermediate increase of �G so CH4 was released throughoutthe experiment, resulting in the highest maximum CH4 production rates.

Page 148: The Biogeochemistry of Submerged Soils

140 Biological Processes in the Soil and Floodwater

0

100

200

300

400

500CO2

FeCH4

MnNO3

SO4

(a)

Soil No.

Ele

ctro

ns t

rans

ferr

ed (

µmol

g−1

)

0

200

400

600

800

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

(b)

No. pH Org C(mg g−1)

Labile C(µmol g−1)

Free Fe(µmol g−1)

1 7.7 10.4 79.5 12.12 6.0 16.8 582.7 13.83 5.1 18.5 713.9 3.64 7.4 13.4 376.5 8.25 6.3 11.5 424.3 13.66 6.7 13.5 477.3 10.27 7.6 9.5 265.4 7.38 5.9 19.7 818.7 4.99 5.1 16.5 575.2 22.410 6.2 21.4 331.1 22.311 6.8 26.2 319.0 24.912 6.0 15.1 200.4 13.013 6.7 10.7 118.0 24.514 5.8 13.9 272.3 33.115 6.1 8.1 405.9 4.316 6.0 15.5 422.0 10.8

Figure 5.3 Calculated electron balance between CO2 produced and electron acceptorsreduced during anaerobic incubation of 16 rice soils (a) after the initial reduction phaseand (b) after 120 days (Yao et al., 1999). Reproduced with kind permission of KluwerAcademic Publishers

Page 149: The Biogeochemistry of Submerged Soils

Microbiological Processes 141

5.1.2 NITRATE REDUCTION

Figure 4.1 shows that NO3− is the stable form of nitrogen over the usual range of

pe + pH in aerobic environments. The fact that most of the N2 in the atmospherehas not been converted to NO3

− therefore indicates that the biological mediationof this conversion in both directions is inefficient. Hence NO3

− reduction to N2

occurs by indirect mechanisms involving intermediaries. Dissimilatory reductionof NO3

− (i.e. where the nitrogen oxide serves as an electron acceptor for the cell’smetabolism but the N reduced is not used by the microbes involved) potentiallyoccurs by two processes:denitrification,

NO3− −−−→ NO2

− −−−→ NO −−−→ N2O −−−→ N2

and reduction to NH4+,

NO3− −−−→ NO2

− −−−→ NH4+

Assimilatory NO3− reduction might also occur. But because concentrations of

NH4+ and organic N are in general large in anaerobic environments, it is sup-

pressed and insignificant. The literature on NO3− reduction is reviewed by

Tiedje (1988).In most submerged soils dissimilatory reduction to NH4

+ is much less impor-tant than denitrification because reduction to NH4

+ is a strictly anaerobic processand any NO3

− entering the soil or formed in oxic zones is denitrified beforeit reaches a sufficiently reduced environment (Buresh and Patrick, 1981). Theimportance of dissimilatory reduction depends on the ratio of available carbon toelectron acceptors. Reduction to NH4

+ produces more electrons per unit NO3−

reduced (8 compared with 5), but less energy. It therefore dominates in contin-uously anaerobic environments with a high ratio of available carbon to electronacceptors, such as the rumen, whereas denitrification dominates in environmentswith a low ratio of available carbon to electron acceptors, such as in most sub-merged soils and in anaerobic microsites in otherwise aerobic soils. Buresh andPatrick (1981) found 15 % of NO3

− reduction was to NH4+ in unplanted sub-

merged sediment, and Buresh et al. (1989) found the equivalent figure for ricesoils was less than 5 %.

Carbon acts as the electron donor for denitrification. The availability of carbonoften limits denitrification in anaerobic microsites in non-submerged soils. As aresult, the reaction does not go to completion and the intermediaries NO2

− andN2O accumulate. Completion of the reaction may also be hindered by low pH. Butunder uniformly anaerobic conditions NO3

− as electron acceptor is more likelyto be limiting than carbon as electron donor because NO3

− is not regenerated.Therefore the rate of denitrification is limited by the supply of NO3

− ratherthan carbon, and proceeds almost completely to N2.

Page 150: The Biogeochemistry of Submerged Soils

142 Biological Processes in the Soil and Floodwater

These general features of NO3− reduction in submerged rice soils are born

out by field observations. Buresh et al. (1993b) found that from 60 to 75 %of 15N-labelled NO3

− applied on the surface of flooded ricefields was lost bydenitrification over 2–3 weeks, as measured by the 15N not recovered in thesoil, floodwater and plants. The recovery of (N2 + N2O)-15N in chambers placedover the floodwater was less than the estimated denitrification loss because gasbubbles became entrapped in the soil. More N2 + N2O was recovered when thechambers were placed over the rice plants showing that some of the gas escapedthrough the plants. The 15NO3

− not lost by denitrification was presumed to havebeen immobilized following absorption by the plants and algae; denitrificationlosses increased when algicides were applied.

5.1.3 IRON AND MANGANESE REDUCTION

Iron reduction has been studied more intensively than manganese because of thegreater abundance of iron in the natural environment. However, because of theirbroadly similar chemistries, the processes involved are probably similar.

Until recently it was thought that microbes were not directly involved in Feand Mn reduction but only indirectly through abiotic reactions involving endproducts of their metabolism. Common metabolites such as H2S and variousorganic acids can reduce Mn and Fe chemically, especially at low pH. How-ever, it is now clear, at least for Fe, that the reduction is directly linked tomicrobial metabolism (Lovley, 1997; Straub et al., 2001). The main obstacle inestablishing this was the insolubility of Fe and Mn oxides, which prevents easyabsorption of Fe and Mn into microbial cells so that their reduction can belinked directly to oxidation of organic compounds via cellular electron transfersystems. It is now established that electrons are transferred out of microbial cellsto extracellular Fe(III) using cytochromes and quinones or similar compounds,and that some species of Fe reducers can solubilize Fe(III) in oxides by excretingchelating agents.

Nevin and Lovley (2002) have shown that microbes can reduce Fe(III) in ironoxide without being in direct contact with the oxide. They incorporated poorlycrystalline iron oxide into porous alginate beads and incubated the beads withcultures of the Fe reducing bacterium Geothrix fermentans, which is found inthe Fe(III) reduction zone of anoxic sediments. Ferric iron from the solid oxidewas reduced and the concentration of dissolved Fe(III) in the solution increasedto as much as 250 mM. Since the pores in the beads were too fine for entry ofthe bacteria, and the amount of Fe(III) reduced was far greater than the amountin oxides on exposed surfaces of the beads, this demonstrates that extracellularexcretions from the bacteria both solubilized Fe(III) and shuttled electrons toit. Further experiments in which Fe(III) in iron oxides was reduced by filteredsuspensions of cultures of G. fermentans confirmed that the electron shuttlingwas extracellular. In contrast, the Fe reducer Geobacter metallireductans is not

Page 151: The Biogeochemistry of Submerged Soils

Microbiological Processes 143

capable or reducing Fe(III) without direct contact with the oxide. Ecologicalinteractions between these and other species of Fe reducers in natural anoxicsediments reflect these and other mechanisms (Snoeyenbos-West et al., 2000;Stein et al., 2001).

Demonstrating that Mn(III,IV) is reduced microbially is complicated by therapid abiotic reduction of Mn(III,IV) by Fe2+ and other reductants. Lovley andGoodwin (1987) obtained indirect evidence for microbial mediation of Mn(III,IV)reduction in experiments in which they follows the consumption of H2 by anoxicsediment. Addition of MnO2 caused H2 to decrease to smaller concentrationsthan possible under Fe reducing conditions, suggesting that Mn reduction wasout competing Fe reduction for H2 in the same way that Fe reduction out competesSO4

2− reduction and methanogenesis.

5.1.4 SULFATE REDUCTION

Widdel (1988) gives a comprehensive review of the microbiology and ecologyof sulfate reduction in natural environments. Dissimilatory reduction of SO4

2−is carried out by certain heterotrophic bacteria, which use SO4

2− as the terminalelectron acceptor in their respiration. The main genera are Desulfovibrio, Desul-fomaculum and Desulfobacter. The bacteria are obligate anaerobes, and beingheterotrophs their activity is sensitive to the supply of carbon. Various organicsubstrates are used with some preferences among species: lactate is the preferredsubstrate for many species but there are also acetate oxidizing sulfate reducers.Lactate oxidizers in particular will also grow well on H2. Competition for H2

and acetate results in inhibition of methanogens by sulfate reducers.Figure 5.2 shows that SO4

2− reduction commences well before Fe(III) reduc-tion is complete, in spite of the lower redox potential required. The onset of SO4

2−reduction coincides with a rapid decline in concentrations of H2 and acetate, forwhich the sulfate reducers compete with Fe(III) reducers. The overlap between Feand SO4

2− reduction is explained by clustering of sulfate reducers in micrositeswithin which they generate more strongly reducing conditions than in the sur-rounding soil. They are able to do this because the SO4

2− and organic substrateson which they subsist are mobile in the soil solution and can therefore diffuseto the microsites where the colonies of sulfate reducers develop. Iron reducerscannot do this because they depend on access to immobile Fe(III) in the soilsolid. Hence the horizontal distribution of sulfate reducers in submerged soilsis generally found to be contagious (Watanabe and Furusaka, 1980). The degreeof clustering increases as the mean number of cells present increases, confirm-ing that the clustering is self-induced. The distribution with depth follows theprofile of redox potential with a peak at an intermediate depth, below the zonedominated by Fe reduction and above the zone of methanogenesis.

Except in some coastal soils, histosols, acid sulfate soils, and soils artificiallyamended with sulfate, the total amount of sulfate present is usually small in

Page 152: The Biogeochemistry of Submerged Soils

144 Biological Processes in the Soil and Floodwater

comparison with the amount of reducible Fe. Hence SO42− reduction generally

does not exert a dominant influence on the soil ecology. Artificial amendmentwith sulfate has been proposed as a means of ameliorating methane emissions(Chapter 8).

5.1.5 METHANOGENESIS

While sufficient inorganic oxidants are present, CO2 is the main end productof organic matter decomposition. But after the inorganic oxidants are used up,methanogenesis is obligatory and the proportion of CH4 in the respiratory gasesincreases. Methane is produced mainly by disproportionation of acetate to CO2

and CH4 or by reduction of CO2 with H2 (see review articles by Conrad, 1989;Kiene, 1991; Zinder, 1993). The relative proportions of the two pathways andthe resulting ratio of CH4 to CO2 produced depend on the balance of elec-trons among the reactants and products. Hence for organic matter whose averageoxidation state is zero—for example carbohydrates, which have an average com-position CH2O—complete oxidation produces equal quantities of CH4 and CO2.But if the organic products are more oxidized or more reduced than the origi-nal compounds, the ratio of CH4 to CO2 produced will be less or greater thanone. Yao and Conrad (2000) have used this principle to analyse organic matterturnover in methanogenic soils. Because the rate of turnover is slow, and the cor-responding changes in the organic matter small and therefore difficult to measureaccurately, this approach is potentially very useful. Yao and Conrad’s method isnow outlined.

Once the inorganic oxidants have been used up and methanogenesis estab-lished, the soil enters a pseudo steady state in which the gross composition ofthe organic matter is little altered by decomposition and the rates of CO2 andCH4 production are roughly constant (cf. Figure 5.2). A simple model of theelectron balance during the pseudo steady state is as follows. Decompositionof soil organic matter (SOM) from SOM0 to SOM1 plus CO2 and CH4 occursthrough the following reactions:

SOM0 + aH2O −−−→ SOM1 + bCH3COOH + cH2 + dCO2 (5.1)

CH3COOH −−−→ CH4 + CO2 (5.1a)

and4H2 + CO2 −−−→ CH4 + 2H2O (5.1b)

where a, b, c and d are coefficients, normalized for the flux of C (i.e. �C =2b + d = 1). In the pseudo steady state, Reaction (5.1) is rate limiting and CO2

abundant. It follows that the net rates of CH4 and CO2 production are

VCH4 = (b + c/4)V and VCO2 = (b − c/4 + d)V

Page 153: The Biogeochemistry of Submerged Soils

Microbiological Processes 145

where V is the rate of Reaction (5.1), and the ratio of these rates is

φ = VCH4

VCO2

= 4b + c

4b − c + 4d(5.2)

Similarly the fractions of CH4 produced in Reactions (5.1b) and (5.1a) are

RH2 = c

4b + cand RAc = 1 − RH2 = 4b

4b + c

and the ratio of these fractions is

ψ = RH2

RAc= c

4b(5.3)

If the composition of SOM0 is

SOM0 = xC + yH + zO

then

SOM1 = (x − 2b − d)C + (y + 2a − 4b − 2c)H + (z + a − 2b − 2d)O

Taking O to be in oxidation state −2, H in state +1 and ignoring all other SOMelements, the charges on SOM0 C and SOM1 C are therefore

Z0 = 2z − y and Z1 = 2z − y + 2c − 4d

and the change in total SOM C charge per mole of C consumed, �Z, is

�Z = Z1 − Z0 = 2c − 4d (5.4)

These equations can be combined to give �Z in terms of φ(= VCH4/VCO2):

�Z = 4φ − 1

φ + 1(5.5)

This relation is plotted in Figure 5.4(a). Negative values of �Z indicate a deficitof electrons in the gaseous products of SOM decomposition and that SOM1 ismore reduced than SOM0; positive values indicate a surplus of electrons in thegaseous products and that SOM1 is more oxidized than SOM0.

Substituting for c from Equation (5.3) and for d from d = 1 − 2b inEquation (5.2) gives the following expression for b in terms of φ and ψ :

b = φ

(1 + φ)(1 + ψ)(5.6)

Here b is the number of moles of acetate produced per mole of SOM carbondecomposed. From Equation (5.3), the number moles of H2 produced is

c = 4bψ = 4φψ

(1 + φ)(1 + ψ)(5.7)

Page 154: The Biogeochemistry of Submerged Soils

146 Biological Processes in the Soil and Floodwater

CH4 produced/CO2 produced (f)

0.0 0.4 0.8 1.2−4

−3

−2

−1

0

1(a)

Cha

rge

tran

sfer

red

per

mol

of

C(∆

Z)

1.6 0.0 0.4 0.8 1.2 1.6

mol

H2

per

mol

C (

c)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

(c)

0.8

0.60.4

0.2

y = 1

0.0 0.4 0.80.0

0.1

0.2

0.3

0.4

0.5(b)

0.2 0.40.6

0.81

mol

ace

tate

per

mol

C (

b)

1.2 1.6

y = 0

Figure 5.4 Calculated (a) electron balance (Equation 5.5), (b) production of acetate(Equation 5.6) and (c) production of H2 (Equation 5.7) as functions of the ratio of CH4to CO2 produced (φ) during anaerobic decomposition of soil organic matter. Numbers oncurves are ratios of CH4 produced from H2 + CO2 to CH4 produced from acetate (ψ)

These relations are plotted in Figure 5.4(b) and (c), showing how the amountsof H2 and acetate produced per mole of C consumed vary with the ratio of CH4

to CO2 produced.Finally, from Equation (5.4) the rate of change in Z is

dZ/dt = (2c − 4d)V1 = 2VH2 − 4(VCO2 − VAc + VH2/4) (5.8)

where VAc = (1 − RH2)VCH4 and VH2 = 4RH2VCH4 , giving

dZ/dt = 4(VCH4 − VCO2) (5.9)

Yao and Conrad (2000) measured rates of CO2 and CH4 production (VCH4 andVCO2 ) and the proportion of CH4 produced from H2 + CO2(RH2) in a range ofsubmerged rice soils under pseudo steady-state conditions, and calculated theelectron balance using Equation (5.9). The results are shown in Figure 5.5. Thefigure shows that in the majority of the soils there was a net deficit of electrons,indicating that the SOM became more reduced during decomposition. The ratioφ = VCH4/VCO2 varied from 0.39 to 0.96 in six of the soils, and from 1.24 to 1.36in the remaining two. Also a fairly small proportion of the CH4 was producedfrom reduction of CO2 with H2; most was produced from disproportionation ofacetate. The ratio ψ = RH2/RAc varied from 0.27 to 0.54.

There are various possible explanations for the decrease in oxidation state ofthe SOM carbon in most of the soils. One is that the SOM comprises differentpools of organic matter with carbon in different oxidation states, and Z decreasesas a result of preferential oxidation of more-oxidized pools of SOM, leaving agreater proportion of the more-reduced forms in the residue. The more oxidizedSOM would include compounds in the original SOM and also compounds gen-erated in the course of reduction, for example as a result of chemical oxidationby Fe(III) and other metal oxides. Alternatively, part of the organic matter could

Page 155: The Biogeochemistry of Submerged Soils

Microbiological Processes 147

e− deficit (µmol g−1 day−1) e− surplus (µmol g−1 day−1)

−3 −2 −1 0 0 1 2 3

Net surplusor deficit of e−

16

15

12

8

6

5

3

2

CO2 consumedby H2

CH4 producedfrom H2 + CO2

CH4 producedfrom acetate

CO2 producedfrom acetate

CO2 producedfrom SOM

SoilNo.

Figure 5.5 Electron balances during anoxic decomposition of soil organic matter to CH4and CO2 in eight rice soils. Soil properties given in Figure 5.3 (Yao and Conrad, 2000).Reproduced by permission of Blackwell publishing

have been acting as an electron acceptor, itself becoming reduced and allowingmore of the SOM to be oxidized to CO2 rather than reduced to CH4. The lattermechanism is consistent with the observed small proportion of CH4 producedfrom H2 and CO2 because an electron sink in addition to CO2 would suppressthe concentration of H2.

The results imply that the average oxidation state of SOM carbon shoulddecrease under continuous reducing conditions. This agrees with the observedlong-term changes in the composition of SOM and accumulation of phenoliccompounds with prolonged flooding of rice soils (Chapter 3). However the fieldsituation differs from Yao and Conrad’s experiments in that the soil receives con-tinuing inputs of living organic matter from growing plants or other sources, withmean oxidation state zero, and the fields are periodically drained and oxidizedfor some part of the year. Therefore general conclusions cannot be drawn.

5.1.6 AEROBIC PROCESSES

The floodwater and uppermost part of the soil are oxygenated by photosyntheticorganisms, and the rhizosphere is oxygenated by leakage of O2 from plant roots.

Page 156: The Biogeochemistry of Submerged Soils

148 Biological Processes in the Soil and Floodwater

The penetration of the O2 into the soil depends on its rate of consumption inaerobic processes and its rate of transport by mass flow and diffusion, and in thefloodwater–soil interface, mixing by burrowing invertebrates. Various aerobicprocesses take place in these oxygenated zones.

Nitrification

Depending on the population and growth rates of nitrifying bacteria, the meet-ing of O2 from the floodwater or roots and NH4

+ from the anaerobic bulk soilwill lead to production of NO3

−. Subsequent movement of the NO3− into the

anaerobic soil will lead to rapid loss by denitrification. The importance of thisprocess will be sensitive to the various factors affecting oxygenation of the inter-face and transport through the soil. Measurements of denitrification in ricefieldshave in fact failed to find very high rates of loss (De Datta and Buresh, 1989).An important point is that conditions favouring high rates of oxygenation ofthe floodwater–soil interface through algal activity during the day will alsofavour volatilization of NH3 because of the concomitant increase in floodwa-ter pH (Section 3.2). Therefore NH3 volatilization may out-compete nitrifiers forNH4

+ in the soil surface.Figure 5.6 shows profiles of NO3

− concentration measured with microsensorsin soil cores taken from ricefields by Revsbech and co-workers (Liesack et al.,2000). During illumination of the cores, O2 generated in the floodwater penetratedto a depth of 2–3 mm and a clear peak of NO3

− was apparent, produced innitrification. However in the dark, entry of O2 from the floodwater diminished,nitrification ceased, and NO3

− moving into the soil from the floodwater wasrapidly consumed in denitrification. The rates of nitrification were slow and allfields investigated had similar rates. Nitrification in the rice rhizosphere may bemore important (Chapter 6).

Concentration of NO3− (µM)

0 1 2 3 4

Dep

th (

mm

)

−3

−2

−1

0

1

2

DarkLight

Figure 5.6 Concentration profiles of NO3− in soil cores from a ricefield, illuminated and

not illuminated (Liesack et al., 2000). Reproduced with permission from Elsevier Science

Page 157: The Biogeochemistry of Submerged Soils

Microbiological Processes 149

CH4 Oxidation

Again depending on the population and growth rates of methane oxidizing bacte-ria, methane diffusing in from the anaerobic soil may be oxidized in the floodwa-ter–soil interface or oxygenated rhizosphere. Figure 5.7 shows profiles of O2 andCH4 measured with microelectrodes in the same way as for Figure 5.6. The rateof oxidation is again sensitive to diurnal changes in oxygenation by algae in thefloodwater. But nonetheless, averaged over the day, of the order of 80 % of CH4

diffusing towards the floodwater will be oxidized (Conrad and Frenzel, 2002).Oxidation of CH4 in the rhizosphere is rather less efficient (10–30 %) becauseof the greater competition with alternative O2 sinks (Section 8.1.3). Because dif-fusion of CH4 through the soil to the floodwater is slow, and oxidation in the

−10

−8

−6

−4

−2

0

2

4

oxygen

(a) illuminated

(b) not illuminated

methane

Concentration of O2 (µM)

0 200 400 600 800 1000 1200 1400

Dep

th (

mm

)

−10

−8

−6

−4

−2

0

2

4

Concentration of CH4 (µM)

0 50 100 150 200 250 300

Figure 5.7 Concentration profiles of O2 and CH4 in soil cores from a ricefield (a)illuminated and (b) not illuminated (Damgaard et al., 1998). Reproduced by permissionof American Society for Microbiology

Page 158: The Biogeochemistry of Submerged Soils

150 Biological Processes in the Soil and Floodwater

floodwater-soil interface efficient, diffusion through the floodwater is a much lessimportant conduit for CH4 emission from ricefields than escape by ebullition orby passage through the plant (Section 8.1).

5.2 MACROBIOLOGICAL PROCESSES

Superimposed on the microbiological processes are processes driven by themacroflora and -fauna in the soil and floodwater. These are responsible for the netprimary production of the system, which ultimately drives the biogeochemistry.Organic matter produced by photosynthetic organisms is the source of energy andnutrients for grazing organisms, populations of which can be very large. Many ofthe invertebrate species found in wetlands create burrows through the soil whichprovide conduits for the movement of oxygen, nutrients and carbon. Hence ratesof interchange between aerobic and anaerobic zones can be much greater thanexpected from simple physical transport processes.

For discussion of the ecology of wetland soils and water see Mitsch andGosselink (2000) for natural wetlands, Roger (1996) for wetland ricefields, andCatling (1992) for the additional niceties of deepwater ricefields.

5.2.1 NET PRIMARY PRODUCTION AND DECOMPOSITION

The net primary production is often far greater in wetlands than in drylands insimilar climate zones (Chapter 1). In a given climate, NPP depends on hydro-logical conditions—the frequency and duration of submergence and the rate ofwater flow—and on the concentrations of nutrients and toxins. Hydrologicalconditions regulate primary producers and decomposers in the soil by limitingthe availability of oxygen for aerobic respiration and by affecting supplies ofnutrients and toxins. In general the frequency of inundation is more impor-tant than the duration, and the more open the system the greater the NPPbecause periodic inundation brings in oxygen and nutrient-rich sediment andflushes out toxins. Likewise decomposition is faster under a fluctuating waterregime and the accumulation of organic matter is greatest in wetlands withprolonged inundation and stagnant water (Moore and Bellamy, 1974; Mitschand Gosselink, 2000). Maximum productivity occurs under intermediate periodsof inundation.

In many wetlands NPP and decomposition are most limited by the availabilityof nutrients, especially N and P. For example, in a review of published data onnutrient limitations in North American bogs, fens, marshes and swamps, Bedfordet al. (1999) found that a large proportion of the wetlands were either P limitedor limited by both N and P, especially those occurring on organic soils. Onlymarshes had N:P ratios in both live tissues and soils that consistently indicated Nlimitation, though the soil data suggested that the majority of swamps were also

Page 159: The Biogeochemistry of Submerged Soils

Macrobiological Processes 151

N limited. Nutrient availability also affects the composition of plant communities:species richness and the presence of rare species often decline as nutrient avail-ability increases beyond some threshold (Bedford et al., 1999). In addition to theinflow of nutrients with the water and sediment, there are complicated interactionsbetween hydrology and nutrient availability that affect productivity and decom-position. Transformations of N and P under anaerobic conditions are discussedin Section 4.3.

5.2.2 THE FLOODWATER–SOIL SYSTEM

Five zones can be distinguished: the floodwater standing on the soil per se, thefloodwater–soil interface, the anaerobic bulk soil, the rhizosphere, and the sub-soil. These are to some extent continuous with each other, and they are certainlylinked so that the function of the system as a whole is greater than the sum ofits parts. But they provide convenient boundaries for discussion.

The floodwater is photic and aerobic. It contains photosynthetic and chemosyn-thetic producers of fixed carbon–bacteria, algae and aquatic weeds—and inver-tebrate and vertebrate consumers that graze on the producers. The communityof producers and consumers provides organic matter to the underlying soil andrecycles inorganic nutrients.

The floodwater–soil interface is also photic and aerobic. The boundary withthe overlying water is diffuse and the bulk density increases from near zero to1 g cm−3 or more in the underlying anaerobic soil. The depth to the underly-ing soil varies from a few mm to a few cm, depending on the aeration of thefloodwater, reducing conditions in the soil, rates of percolation, and mixing byinvertebrates. Nitrate, Mn(III,IV), Fe(III), SO4

2− and CO2 are stable and algaeand aerobic bacteria predominate. In the early stages of land preparation for rice,algae develop on the wet soil surface and support populations of grazers. Asorganic matter accumulates during the crop, populations of benthic filters anddeposit feeders develop. The activities of the invertebrates affect nutrient cyclingboth directly through their excretions and indirectly by moving soil particles andorganic matter.

The anaerobic soil is non-photic and reduction processes predominate. Thevalue of pe + pH is generally below the range at which Fe(III) is reduced, unlessorganic substrates are limiting or there are large concentrations of more oxidizedreductants such as Mn(III,IV). Microbial activity is concentrated within water-stable aggregates containing organic matter, and produces NH4

+, S2−, organicacids and CH4. Decomposing organic matter in the anaerobic soil sustains pop-ulations of aquatic oligochaete worms and chironomid larvae.

The subsoil at greater depths may be aerobic in well-drained soils with aperched water table owing to an impermeable layer—such as the traffic pan inricefields; or anaerobic in soils that are poorly drained throughout. It may providesignificant quantities of nutrients to plants growing in the soil if their roots canreach them.

Page 160: The Biogeochemistry of Submerged Soils

152 Biological Processes in the Soil and Floodwater

Plants shade the floodwater and soil surface, and lower the temperature andconcentration of CO2 above the floodwater during the day. These effects increaseas the plant canopy develops during a growing season, with correspondingchanges in the ecology of the soil and water. Plants also act as substrate forepiphytic growths—for example for nitrogen-fixing cyanobacteria—and theyprovide mechanical support for animals, for example allowing animals to escapehigh temperatures in the floodwater during the day. Plant roots support a partiallyoxic rhizosphere. The rhizosphere is non-photic, but root-derived carbon providessubstrate for microbial growth. The thickness of the oxic rhizosphere is only afew tenths of a millimetre, but because the roots can occupy a large part of theanaerobic soil layer, a significant proportion of it may be oxic (Chapter 6).

5.2.3 FLOODWATER PROPERTIES

Temperature and Radiation

Under submerged conditions, temperatures in the soil and water depend on thedepth of the water and on the density of the plant canopy, as well as on mete-orological conditions. The water transmits incident short-wave radiation to thesoil but it also insulates the soil against emission of long wave radiation. Thefull plant canopy transmits 90 % of the short-wave infrared radiation (i.e. half thetotal short-wave). Hence there is a ‘greenhouse’ effect and consequently the soiland water temperatures tend to be higher than the air temperature. Evaporativecooling reduces the surface water temperature and drives convection currents, sothe water tends to be well mixed.

Figure 5.8 shows temperatures in the air, water and soil in a tropical ricefieldover a year. For the conditions in Figure 5.8, the water and the top 2 cm of soilwere 2–5 ◦C warmer than the maximum air temperatures and they continued torise for some weeks after the annual peak air temperature. The 2–10 cm layerof soil is 2–3 ◦C cooler than the top 0–2 cm. Because of the greenhouse effect,diurnal changes in soil and floodwater temperatures tend to be smaller than thechanges in air temperature. However the effects depend on the depth of thefloodwater and its source. Under continuous flowing irrigation, water tempera-tures tend to be lower than air temperatures in hot areas and vice versa in coolareas. Sediment load and algal cover also have effects.

Incident radiation varies with season, cloud cover and latitude, and the fractionreaching the floodwater varies with the density of the plant canopy and with thepresence of floating macrophytes and plankton and floodwater turbidity. Algaeuse different wavelengths of light to green plants, and this may in part com-pensate for shading by plants. Nonetheless the intensity of photosyntheticallyactive radiation at the floodwater surface becomes deficient at some point dur-ing the season. Figure 5.9 shows the decline in light intensity at the floodwatersurface as the canopy of a rice crop develops. Within 6 weeks of transplanting,

Page 161: The Biogeochemistry of Submerged Soils

Macrobiological Processes 153

Day of year

0 100 200 300

Tem

pera

ture

(°C

)

25

30

35

40Air WaterSoil 0-2 cmSoil 2-10 cm

Rice crop Rice crop

Figure 5.8 Temperatures in ricefields over a year, IRRI, Laguna, Philippines. Valuesare means over a month; air temperature is maximum; water and soil temperatures weretaken at 1400 h (Roger, 1996). Reproduced by permission of IRRI

Time after transplanting (days)

0 20 40 60 80 100

Fra

ctio

nal l

ight

pen

etra

tion

0.0

0.2

0.4

0.6

0.8

1.0

Figure 5.9 Decrease in light penetration to the floodwater as a typical rice crop develops(J. Sheehy, IRRI, unpublished data)

only 10 % of the incident photosynthetically active radiation reaches the flood-water. Similar changes occur in natural wetlands as plants become establishedfollowing inundation.

Dissolved O2 and CO2 and pH

Oxygen is produced by photosynthetic organisms in the water and consumedin respiration and other oxidative processes. Often the concentration varies fromsuper-saturation during the day to near zero at night. Because the partial pressures

Page 162: The Biogeochemistry of Submerged Soils

154 Biological Processes in the Soil and Floodwater

of CO2 and O2 are inversely related, the pH and O2 concentration in the flood-water tend to be positively correlated. Diurnal variations of floodwater CO2 andpH are discussed in Section 3.3.

Inorganic Nutrients

The supply of nutrients in the floodwater depends on the composition of thewater arriving in the field and on the nutrient status and transport propertiesof the underlying soil. As discussed above, P is most often the most limitingnutrient in natural wetlands and also in ricefields (Roger, 1996). Concentrationsin irrigation and flood waters are generally small (≤10 µM) and diffusion fromthe soil into the floodwater is slow, particularly if P is adsorbed to a large extenton iron oxides in the oxic floodwater–soil interface. The effect of turbation byanimals burrowing into the soil may therefore be important. The calculations inChapter 2 show that the flux of P into the floodwater increases several fold withrealistic tubificid populations and soil parameters.

5.2.4 FLOODWATER FLORA

The photosynthetic aquatic biomass comprises cyanobacteria (formerly calledblue-green algae), planktonic, filamentous and macrophytic algae, and vascu-lar macrophytes. The net productivity of the floodwater depends on the levelof primary production by the photosynthetic biomass versus its consumptionby grazing animals, particularly cladocerans, copepods, ostracods, insect larvaeand molluscs. Their role will change as the canopy develops and at a leaf areaindex of about 6–7 there will be no more photosynthetically active radiationavailable to them.

The cyanobacteria and algae are confined to the water and upper few mm ofsoil, whereas the macrophytes may be either floating—for example, in ricefieldsthe water fern Azolla –or rooted in the soil to depth. Planktonic and filamentousalgae move up and down in the water column over the day as their buoyancychanges with photosynthetic O2 production. Table 5.1 compares the compositionsof cyanobacteria and aquatic macrophytes obtained from ricefields. The data showthat cyanobacteria have lower dry matter contents and higher N contents thanthe macrophytes. Aquatic macrophytes notably have much smaller dry mattercontents than terrestrial plants and greater ash contents as % of dry matter. TheP contents of the cyanobacteria and aquatic macrophytes in Table 5.1 are in bothcases smaller than critical values for deficiency (about 1 %), which indicates thatin these ricefields P was the limiting nutrient.

Information on the productivity of the floodwater is scarce. Data compiled forricefields by Roger (1996) give an average of 0.35 t dry wt ha−1 of cyanobacteria,

Page 163: The Biogeochemistry of Submerged Soils

Macrobiological Processes 155

Table 5.1 Compositions of N2-fixing cyanobacteria and aquaticmacrophytes in ricefields. Data are means with ranges in parentheses

Cyanobacteria Aquaticmacrophytes

Dry matter (% fresh wt) 4.0 (0.9–7.0) 8.0 (4.5–12.0)Ash (% dry wt) 45.0 (27–71) 20.0 (12–50)C (% dry wt ash-free) 40.0 (37–47) –N (% dry wt ash-free) 5.0 (3.8–7.4) 2.1 (1.3–2.9)C:N ratio 8 (5–12) 24 (18–47)P (% dry wt ash-free) 0.2 (0.05–0.39) 0.3 (0.1–0.6)

Source: Roger (1996). Reproduced by permission of IRRI.

algae and macrophytes in the field during a cropping season (range 0.1–0.6 tha−1). For freshwater bodies and irrigation canals the average was 4.5 t ha−1

(range 1–13 t ha−1). In experimental plots of rice under a range of fertilizertreatments, values for cyanobacteria were 177 kg dry wt ha−1, 28 kg C ha−1 and4 kg N ha−1. Net primary production in ricefield floodwater may reach 1 to2 g C m−2day−1 but more usually it ranges between 0.2 and 1 g C m−2day−1.The total production over a season is equivalent to 10–15 % of the carbon accu-mulated by the rice crop.

Factors Affecting Primary Production

Climate. The principal variable is light intensity, which affects both the totalbiomass and its composition. Generally the algal biomass increases followingflooding until shading by the rice canopy becomes limiting. In general greenalgae prefer high light intensity, cyanobacteria prefer low light, though not in allcases, and diatoms are indifferent. This results in a characteristic succession ofspecies over the growing season. The effect of light is moderated by temperature,and both low and high temperatures can be inhibitory. Dessication and re-wettingof the soil are also inhibitory and affect the proportions of species present.Enhanced mineralization of soil N following drying and re-wetting favours algaeover cyanobacteria.

Soil. Soil pH and the content of available P have the most consistenteffect (Roger, 1996). Overall algal growth increases as pH increases and higherpH favours N2-fixing cyanobacteria over eukaryotic algae. Liming of acidic soilsincreases the proportion of cyanobacteria in the biomass, and the proportionof cyanobacteria may increase as the soil pH increases during reduction.Applications of mineral fertilizers and organic manures have large effects onalgal growth and succession. Effects of changes in acidity and availabilities of N

Page 164: The Biogeochemistry of Submerged Soils

156 Biological Processes in the Soil and Floodwater

and P are particularly important. Herbicides have direct effects and insecticideshave indirect effects through grazers.

Biotic Factors. Cyanobacteria appear to be ubiquitous in ricefields and attemptsto increase biological N2 fixation by inoculating with improved strains developedin laboratory cultures have not been successful (Roger, 1996). Certain bacteria,fungi and viruses have been shown to be pathogenic to cyanobacteria and algaeunder laboratory conditions, but this has yet to be confirmed under field condi-tions. Likewise antagonistic effects have been observed between different speciesof cyanobacteria and algae, and between algae and macrophytes, and vice versa,but there is not much information on the importance of these effects under fieldconditions. Invertebrates such as cladocerans, copepods, ostracods, insect lar-vae and snails are common grazers of algae in ricefields and their populationdynamics often mirror those of algae with a lag of a week or two (below).

Dynamics Over the Crop Cycle

The plethora of variables affecting algal growth and succession mean that com-plicated dynamics are expected. Figure 5.10 shows a generalized succession forthe dry season flora and fauna in unfertilized ricefields (Grant et al., 1986). Themain points are (Roger, 1996):

• Eukaryotic algae develop first but are quickly succeeded by non-colonial, het-erocystous cyanobacteria, shown by the increase in chlorophyll a in the figure.The bloom of cyanobacteria produces a bloom of ostracods and molluscs,which graze on the cyanobacteria. The resulting collapse of the cyanobacteriapopulation is followed by collapses of the ostracods and molluscs.

• About 4 weeks after transplanting a population of slow-growing, mucilaginous,colonial cyanobacteria develops, which is resistant to grazing. This continuesto grow until the soil is drained for the harvest of the rice crop.

• Primary production typically exceeds net respiration (P:R > 1) over the firstmonth, leading to accumulation of organic matter and hence a decrease in theratio to low values.

• There is a peak of N2 fixation coincident with the early cyanobacterial bloom,shown by the increase in acetylene reducing activity in the figure. Nitrogenfixation by the colonial cyanobacteria is slower but lasts longer.

Biological Nitrogen Fixation in the Floodwater–Soil System

The water column and soil surface are often the main sites of biological nitrogenfixation in wetland systems (Buresh et al., 1980; Roger 1996). Biological nitro-gen fixation is the process by which atmospheric N2 is reduced to NH4

+ and the

Page 165: The Biogeochemistry of Submerged Soils

Macrobiological Processes 157

Days after transplanting

0 10 20 30 40 50 60 70 80 90 1000

2

4

6

8

10

12

14

0

500

1000

1500

2000

0

50

100

150

200

AR

A (

µmol

C2H

2 m

−2 h

−1)

1000

500

100

50

0 Mol

lusc

s (n

o. m

−2)

Colonial algae (g m−2)

Chlorophyll a (mg m−2) andOstracods (no. m−2 × 102)

Chlorophyll a

ARA OstracodsColonialcyanobacteria

HarvestMolluscs

DAT

10 20 30 40 50 60

P:R

0123

Figure 5.10 Generalized fluctuations of algae, acetylene-reducing activity (ARA), andgrazers in floodwater of unfertilized ricefields. P:R is the ratio of primary production to netrespiration (Grant et al., 1986). Reproduced by permission of AB Academic Publishers

NH4+ incorporated into organic compounds. The reduction requires a pe less than

about −4.5 at pH 7 (Section 4.3). This is less negative than the pe required forreduction of CO2 to CH2O, hence cyanobacteria (blue-green algae) and certainother photosynthetic bacteria are able to mediate the reduction at the negative pelevels generated in photosynthesis. Non-photosynthetic N2 fixers require anoxicconditions or must exclude O2 from site of fixation. Because of the large acti-vation energy required to break the N≡N triple bond, there is a kinetic barrierto be overcome, and this is achieved in N2 fixing organisms through the enzymenitrogenase.

In wetlands N2 fixation can occur in the water column, in the aerobic water–soilinterface, in the anaerobic soil bulk, in the rhizosphere, and on the leaves andstems of plants. Phototrophic bacteria in the water and at the water–soil interfaceare generally more important than non-photosynthetic, heterotrophic bacteria inthe soil and on plant roots (Buresh et al., 1980; Roger 1996). The phototrophscomprise bacteria that are epiphytic on plants and cyanobacteria that are bothfree-living and epiphytic. A particularly favourable site for cyanobacteria is belowthe leaf surface of the water fern Azolla, which forms a very efficient symbiosiswith the cyanobacterium Anabaena azollae. This symbiosis and those in variousleguminous plants have been exploited in traditional rice production systems tosustain yields of 2 to 4 t ha−1 of grain without fertilizer for hundreds of years.

Page 166: The Biogeochemistry of Submerged Soils

158 Biological Processes in the Soil and Floodwater

There has been much research aimed at developing N2 fixing systems for greateryield levels. For practical reasons, such systems cannot entirely substitute formineral N fertilizers, and there is the difficult dilemma that N2 fixation withinthe ricefield is inhibited by additions of mineral N.

Table 5.2 gives estimates of N2 fixed by various agents in wetland ricefieldsand Table 5.3 gives estimates of total fixation in ricefields from a review of211 N balances in field and pot experiments by Roger and Ladha (1992). Thevalues range from 0 to 100 kg N ha−1 per crop with averages of 30 kg N ha−1 inplots without N fertilizer, 8 kg N ha−1 in plots with N fertilizer broadcast in thefloodwater, and 12 kg N ha−1 in plots where the fertilizer was placed at depthin the soil. The beneficial effect of the presence of plants on BNF is evident,and also the effect of illumination, especially without inorganic N, indicatingthat the N2 fixing agents are phototrophic. Rates of comparable magnitude andvariability are found in natural freshwater and coastal wetlands (Buresh et al.,1980; Bowden, 1987).

Not all of the nitrogen fixed finds its way into the vegetation. The recovery willdepend on rates of decomposition of the material containing the fixed N, rates of

Table 5.2 Estimates of biological nitrogen fixation (BNF) by various agents in wetlandricefields

BNF (kg N ha−1crop−1)

Heterotrophic bacteria in the rhizosphere 1–7Heterotrophic bacteria in the rhizosphere and bulk soil 1–31Heterotrophic and phototrophic bacteria on added straw 20–40 for 10 t strawFree-living cyanobacteria 0–80Azolla/Anaebena azollae in experimental plots 20–150Azolla/Anaebena azollae in ricefields 10–50Legumes as green manures 50–100 in 50 days

Source: adapted from Roger and Ladha (1992). Reproduced with kind permission of Kluwer Academic Publishers.

Table 5.3 Estimates of nitrogen fixed in rice systems from N balancestudies

Factor No.observations

Mean ± SD

(kg N ha−1crop−1)

No inorganic N 166 29.7 ± 25.4With inorganic N 45 4.0 ± 47.6Planted 193 26.5 ± 30.7Unplanted 18 −0.5 ± 46.2Soil and water light 197 25.0 ± 33.9Soil and water dark 14 13.2 ± 13.8No N, soil and water light 152 31.2 ± 25.7No N, soil and water dark 14 13.2 ± 13.8

Source: adapted from Roger and Ladha (1992). Reproduced with kind permission ofKluwer Academic Publishers.

Page 167: The Biogeochemistry of Submerged Soils

Macrobiological Processes 159

loss, especially by volatilization of NH3 from the floodwater, and how well thedecomposition matches the dynamics of plant growth. Recoveries seldom exceed60 % of the N and may be less than 25 %.

5.2.5 FAUNA

Table 5.4 lists invertebrates commonly found in the soil and floodwater of differ-ent wetland types. Insect larvae, molluscs and, particularly, oligochaete wormsof the order Tubificidae are abundant on and in the soil surface. Table 5.5 listsinvertebrates common in ricefields. In a survey of 32 ricefields in the Philippines,Simpson et al. (1993a) found that oligochaetes were the only macro-invertebratespresent in significant numbers in all the fields. The dominant species wereLimnodrilus hoffmeisteri (81 % of the population on average) and Branchiurasowerbyi (13 %). Their distribution within fields was contagious and ranged from0 to 35 000 m−2, averages 5000 to 10 000 m−2. The fresh weight ranged from 0 to630 kg ha−1. The average population at a site was positively correlated with soil

Table 5.4 Invertebrates commonly found in the soil and floodwater of different typesof wetland

Meiofauna(63–500 µm diameter)

Macrofauna(>500 µm diameter)

Freshwater

Marsh/fen Amphipods, copepods Chironomid insect larvae,gastropods, isopods

Swamp forestDeepwater Nematodes, amphipods,

copepods, oligochaetesCrayfish, clams, oligochaetes,

gastropods, isopods, midgelarvae

Alluvial Nematodes, oligochaetes,‘terrestrial’ invertebrates(mitesacari, springtails-collembola)

Oligochaetes includingearthworms, crayfish

Estuarine

Tidal marshFreshwater Nematodes, amphipods,

oligochaetesOligochaetes, polychaetes, midge

larvaeBrackish Nematodes, amphipods,

oligochaetesOligochaetes, fiddler crabs, snails,

musselsSalt Nematodes, amphipods,

oligochaetesOligochaetes, fiddler crabs, mud

crabs, periwinkles, snails,mussels, oysters, clams

Mangrove forest Nematodes, amphipods,copepods, oligochaetes,polychaetes

Fiddler crabs, oysters, barnacles

Source: adapted from Craft (2001). Reproduced by permission of Lewis Publishers.

Page 168: The Biogeochemistry of Submerged Soils

160 Biological Processes in the Soil and Floodwater

Table 5.5 Invertebrates in ricefield soil and floodwater

Densities (number m−2) Comments

Min. Max. Mean

MicrocrustaceaOstracods 0 98 000 6000 Stimulated by factors that

increase primary production,such as N and P fertilizer.

Filter bacteria and algaefrom water

Copepods 0 40 000 33 000Cladocerans 0 33 000 900

Insect larvaeChironomids 0 10 000 600 Feed on epipelic algae at soil

surface and on floating algae

Mosquitoes 0 7000 170MolluscsSnails 0 1000 200 Inhibited by high acidity, N

fertilizer and pesticides;stimulated by high organicmatter. Graze on epipelicand floating algae, and onalgae epiphytic on plantstems

OligochaetesTubificids 0 40 000 10 000 Stimulated by factors that

increase primary productionand bacterial decomposers;inhibited by high soil bulkdensity

Source: Roger (1996), Simpson et al. (1993a, 1994a,b).

organic matter and applications of N fertilizer, presumably through the effects ofthese on primary production and bacterial decomposers on which oligochaetesfeed. Populations of insect larvae and molluscs are successional over the seasonfollowing the cycles of algal populations. But no general trends have been estab-lished for the dynamics of oligochaetes, though this may reflect the paucity ofdata (Roger, 1996).

The effects of macrofauna on the soil biogeochemistry can be summarized(Aller, 1994):

• manipulation of particles: exposure of substrate resulting in increased decom-position;

• grazing: consumption of microbes, stimulation of microbial growth, increasedmineralization;

• excretion of substrate and nutrients: stimulation of microbial growth, increasedmineralization;

Page 169: The Biogeochemistry of Submerged Soils

Macrobiological Processes 161

• construction of burrows: synthesis of refractory or inhibitory structuralproducts;

• irrigation of burrows: increased transfer of soluble oxidants and nutrients,increased re-oxidation and mineralization;

• transport of particles: transfer between major redox zones, increased re-oxida-tion and mineralization.

Bioturbation

Oligochaetes feed with their heads downward in the burrow and posterior endsupward in the water. They ingest fine soil particles, extract carbon and minerals,and deposit residues in faeces on the soil surface. The faeces may subsequentlyfall into the burrow and be mixed. The net effect is a loosening and mixing ofthe soil to depth. The burrows of the species found in ricefields may be severalcentimetres deep and a millimetre or so in diameter. Deeper and wider burrowsare formed by species found in other wetlands (Table 5.4). Once the burrowsare constructed, the worms tend to remain in them and maintain a supply ofoxygen from the overlying water by waving their posteriors in the water andmoving their bodies in a peristaltic motion. Thereby the water in the burrows ismixed with the overlying water and solutes diffusing into a burrow are rapidlytransferred to the surface and vice versa. The calculations in Section 2.4 showthe great sensitivity of solute transport and mixing to the geometry, density andactivity of the burrowing animals.

The effects of oligochaetes on the soil or sediment depend on the particular cir-cumstances. Limnologists and oceanographers consider oligochaetes to be agentsof aeration, increasing the depth of the oxidized layer and stimulating mineraliza-tion and nitrification–denitrification (Fry, 1982; Aller, 1994). For example, Davis(1974) found that the oxidized layer (EH > 200 mV) in profundal lake sedimentswas increased by 0.3–1.6 cm by tubificid populations of 800 m−2.

By contrast in ricefields, where primary production and the amounts of organicmatter in the floodwater may be much greater, the effect can be to enhance theincorporation of organic matter into the soil and so to make the surface soil on aver-age more reduced, in spite of oxygenation of the solution in the burrows. Kikuchiand colleagues (Kikuchi et al., 1975; Kikuchi and Kurihara, 1977, 1982) found thatwith realistic densities of tubificids and organic matter in ricefields, the oxidizedlayer at the soil surface disappeared altogether. They found that weed growth wasdiminished because seeds were moved to a depth at which the O2 concentration wastoo low for germination, and as a consequence oxygenation of the soil by weedsdecreased and populations of aerobes in the soil decreased and anaerobes increased.The concentrations of NH4

+, ortho-P and acid soluble Fe in the floodwater increasedand the concentration of NO2

− + NO3− decreased (Figure 5.11).

In practice redox conditions in the burrows will oscillate as the oxygenationof the floodwater varies over the diurnal cycle. Aller (1994) found in a widerange of organic matter-rich sediments containing burrowing invertebrates that

Page 170: The Biogeochemistry of Submerged Soils

162 Biological Processes in the Soil and Floodwater

NH

4+ (µ

M)

0

200

400

600

with tubificidswithout

NO

3− +

NO

2− (µ

M)

0

100

200

300

400

500

orth

o P

(µM

)

0

2

4

6

8

10

Day of year

150 175 200 225 250

acid

sol

uble

Fe

(µM

)

0

150

300

450

600

750

Figure 5.11 Effects of tubificids on concentrations of N species, P and Fe in the floodwa-ter of unplanted microplots in ricefields. Species B. sowerbyi, density 1000 m−2 (Kikuchiand Kurihara, 1982). Reproduced with kind permission of Kluwer Academic Publishers

solid particles constantly cycled between oxic and anoxic zones but typicallyspent 10- to 100-times longer under anoxic than oxic conditions. Cyclic redoxpatterns were also common within individual burrows and were accompaniedby rapid switching of metabolic processes. Even brief, periodic re-exposure oforganic matter to O2 resulted in more complete decomposition than under con-stant conditions or unidirectional redox change. Redox oscillation apparentlyresults initially in net remineralization of existing microbial biomass followedby stimulated renewed synthesis. Aller (1994) found that some properties, suchas the accumulation of P in the sediment, were comparable under fully oxic andoscillating redox conditions but differed under continuously anoxic conditions.This is another mechanism by which the operation of the floodwater–soil systemas a whole is not a simple sum of its component parts.

Page 171: The Biogeochemistry of Submerged Soils

Is Biodiversity Important? 163

5.3 IS BIODIVERSITY IMPORTANT?

Submerged and non-submerged soils have huge biological diversity and containorders of magnitude more biological species than above ground habitats andaquatic systems (Liesack et al., 2000; Conrad and Frenzel, 2002; Usher et al.,2004). The origins of this diversity are in the physical and chemical heterogeneityof soils at micro- and macro-scales, and intense competition for substrate. Insubmerged soils there are extreme gradients of redox potential from oxic toanoxic zones due to the slow transport of O2 through the soil, and gradients ofsubstrate from rich to poor zones due to the non-uniform distribution of plantdebris and root exudates. Heterogeneity also arises from soil physical structureand the labyrinthine network of soil pores which constrain the movements of bothorganisms and substrates. Although submerged soils have weak macro-structure,especially rice soils that have been deliberately puddled, they retain considerablemicro-structure in water-stable micro-aggregates and a corresponding networkof pores. These factors in combination result in a near infinite combination ofopportunities and constraints for different organisms.

But does all this biodiversity have any consequences for soil processes at themacro-scale? This is a seemingly straightforward question, but the answer hasbeen surprisingly elusive. Progress has been hampered by the absence of suitableexperimental methods for analysing biological diversity and its relation to soilfunctions. Three types of method are used (Ritz, 2004):

(1) genotypic analysis, which assesses the basic genetic information about thecommunity of microbes present;

(2) phenotypic analysis, which assesses the expression of the genetic information,i.e. the living form of the microbial community; and

(3) functional analysis, which assesses the processes that the microbial commu-nity is actively or potentially engaged in.

The greatest diversity is revealed in genotypic analysis, but there is a corre-sponding lack of discrimination. Analyses of soil DNA often do not show cleardifferences between soils from widely differing environments, including in sub-merged soils (Liesack et al., 2000; Reichardt et al., 2000). Phenotypic analysis,such as by assaying membrane-bound phospholipids from living microbes, ismore discriminatory, and there is now good evidence for phenotypic ‘signatures’in soil microbial communities, modified by the environment the microbes areoperating in, including for submerged soils (Reichardt et al., 1996). In func-tional analysis, the actual or potential activities of the microbial community aremeasured. Techniques for this have been developed based on the ability of soilcommunities to utilize different C-containing compounds using Biolog plates.The results often match phenotypic analysis and the expected effects of soil andenvironmental differences. However the method is biased towards those microbesthat thrive under the particular conditions of the assay in vitro (Preston-Mafham

Page 172: The Biogeochemistry of Submerged Soils

164 Biological Processes in the Soil and Floodwater

et al., 2002). A solution is to measure the utilization of substrates added directlyto soil, and practical methods for doing this are being developed (Degens andHarris, 1997).

Through such techniques evidence is emerging for the importance of biodiver-sity for macro-scale soil processes (Usher et al., 2004). Though there is evidentlya great deal of redundancy in most microbial populations, there are thresholdlevels of biodiversity below which important soil functions are impaired. Forexample, the decomposition of recalcitrant organic matter can only be achievedby consortia of organisms operating together, such as in the anaerobic decom-position of organic matter in submerged soils in sequential reduction reactionsmediated by microbes (Section 5.1.1). Also the growth and activity of individualorganisms is necessarily constrained by the nature of the prevailing community,which is important, for example in the persistence of rare organisms in the soiland management of soil-borne diseases.

Given that biodiversity is important, it is important to understand how soilmanagement affects it. But as yet there is not much information on this. It hasbeen assumed that intensification of rice production and more widespread useof fertilizers and pesticides in the past few decades will have diminished thediversity of microbes and invertebrates in ricefields compared with those undertraditional practices. Roger et al. (1991) compared the diversity of arthropods infarmers’ fields in the Philippines and at the International Rice Research Institute,and found the greatest diversity in fields at the Institute, where there has beenheavy use of fertilizers and pesticides for many years, and the least in fields inthe Ifugao rice terraces at Banaue, where there has been little use of fertilizersand pesticides. This goes against the often hypothesized trend of intensificationreducing biodiversity. Simpson et al. (1993b, 1994a, b) measured the effects offertilizer and pesticides on populations of algae and invertebrates in ricefields,and found complicated interactions. Whereas N fertilizer inhibits N2 fixation bycyanobacteria, P fertilizer stimulates it, and the overall productivity of the flood-water is generally increased by fertilization. Likewise pesticides have variouseffects. Part of the community of organisms responsible for mineralizing organicmatter may be killed by pesticide, but the subsequent collapse of predators mayallow other mineralizing organisms to bloom. Several insecticides reduce thenumbers of ostracods that graze on N2-fixing cyanobacteria, and so N2 fixationis enhanced.

Page 173: The Biogeochemistry of Submerged Soils

6 Processes in Roots and theRhizosphere

Though wetland plants have the advantage of an assured water supply, they mustcontend with various difficulties that their dryland counterparts are largely spared.First, because of the very slow diffusion of respiratory gases through water andsubmerged soil compared with dryland soil, the roots must aerate themselvesinternally by forming internal gas channels to the air above. Second, the rootsmust exclude toxic products of anaerobic metabolism in the soil, such as organicacids and ferrous iron, or tolerate large concentrations of these toxins internally.Third, they must contend with the altered forms and solubilities of nutrients inthe soil under anaerobic conditions, for example the predominance of ammoniumrather than nitrate as the plant-available form of nitrogen. That wetland plantsare capable of surmounting these difficulties is shown by the great productivityand biodiversity of wetland systems. This chapter discusses the various processesand mechanisms involved in this.

6.1 EFFECTS OF ANOXIA AND ANAEROBICITY ON PLANT ROOTS

Generally, in plant cells well supplied with O2, energy is provided for growth andmetabolism by the oxidation of glucose in the three stages shown in Figure 6.1:

(1) glycolysis, in which 1 mol of glucose is converted to 2 mol of pyruvate yield-ing 2 mol of ATP (the main form in which energy is transported and utilizedin plants) and 2 mol of NADH2 (reduced NAD which acts as a universalreducing agent in non-green plant tissues);

(2) the Krebs cycle, in which 1 mol of pyruvate is completely oxidized to CO2

yielding 1 mol of ATP and 5 mol of NADH2; and(3) the mitochondrial cytochrome chain, in which 1 mol of NADH2 generates

3 mol of ATP.

The net result is that complete aerobic respiration of 1 mol of glucose yields38 mol of ATP.

However in the absence of O2, anaerobic glycolysis—fermentation—producesonly 2 mol of ATP per mol of glucose consumed. In the absence of O2 the mito-chondrial cytochrome chain ceases to operate and as a result NADH2 accumulates

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 174: The Biogeochemistry of Submerged Soils

166 Processes in Roots and the Rhizosphere

Glucose 2 PGA

3

1

2

2 (1,3-Diphospho- glycerate)

2 ATP 2 ADP

2 ADP

2 ATP2 PEP

2 ADP

2 ATP2 Pyruvate

Krebs cycle

2 NAD 2NADH2

2 Lactate

2 Ethanol

2 Acetaldehyde

2 CO2

(a)

NAD

NADH2

Pyruvate

Acetyl-CoA

CO2

Citrate

Isocitrate

Oxaloacetate

Fumarate

a-Ketoglutarate

Succinate

FPH2

FP

NAD, ADP

NADH2, ATP

NADH2

NAD Malate

(b)

NADH2

NAD

CO2

CO2

ADP ATP ADP ATP ADP ATP(c)

NADH2 cyta3cytacytccytbFP

O2

OH−e e e ee

Figure 6.1 Pathways involved in glucose oxidation by plant cells: (a) glycolysis,(b) Krebs cycle, (c) mitochondrial cytochrome chain. Under anoxic conditions, Reactions1, 2 and 3 of glycolysis are catalysed by lactate dehydrogenase, pyruvate decarboxylaseand alcohol dehydrogenase, respectively. ATP and ADP, adenosine tri- and diphosphate;NAD and NADH2, oxidized and reduced forms of nicotinamide adenine dinucleotide;PGA, phosphoglyceraldehyde; PEP, phosphoenolpyruvate; Acetyl-CoA, acetyl coenzymeA; FP, flavoprotein; cyt, cytochrome; ε, electron. (Modified from Fitter and Hay, 2002).Reprinted with permission from Elsevier

and the Krebs cycle is suppressed. This leads to an accumulation of acetalde-hyde—the first end-product of fermentation (Figure 6.1a); synthesis of alcoholdehydrogenase catalysing the conversion of acetaldehyde to ethanol; and con-sumption of NADH2 as acetaldehyde is reduced to ethanol and hence continuingproduction of ATP and pyruvate. So fermentation can continue to generate ATP

Page 175: The Biogeochemistry of Submerged Soils

Effects of Anoxia and Anaerobicity on Plant Roots 167

for as long as carbohydrate reserves last. However, because the efficiency of thisprocess is much less than the efficiency of aerobic respiration—only 2 mol ofATP are produced per mol of glucose consumed compared with 38—the rate offermentation must increase sharply under anoxia if the cell energy supply is to bemaintained. This can lead to rapid exhaustion of plant reserves under prolongedanoxia. In addition anoxic cells must contend with toxic products of fermenta-tion, particularly ethanol and lactate. Hence the plant will go to some lengths toavoid anoxia in its active tissues.

6.1.1 ADAPTATIONS TO ANOXIA

The most important adaptation plants make to anoxic soil conditions is the devel-opment of highly porous tissue in the root cortex called aerenchyma (Figure 6.2)(Jackson and Armstrong, 1999). The development of aerenchyma may occurboth through closely regulated separation and expansion of cells, or, more usu-ally, through programmed cell death, also under tight regulation in response toexternal stimuli. The result is a continuous pathway of gas channels between thebase of the root and the tip. This both permits gas transport between the plant’saerial parts and respiring root tissues, and lessens the amount of respiring tissueper unit root volume. In addition, the root wall layers become partially suberizedalong part of the root length, resulting in decreased permeability to gases andhence less loss of O2 to the anaerobic soil outside.

The mechanisms by which the aerenchyma remains gas-filled rather than water-filled are not fully understood but appear to involve metabolic control (Raven,1996). The gas-filled state is favoured by inward gradients of water potentialcreated by evapo-transpiration and by barriers to water movement in the apoplasmsuch as exodermis (van Noordwijk and Brouwer, 1993). Thus the root acts as amoderately gas-tight pipe conveying O2 down from the shoots to the elongatingand actively respiring tip and venting CO2 and other respiratory gases in theopposite direction. Figure 6.3 shows changes in root porosity and respiration ratealong the length of maize roots grown in anoxic media.

Metabolic adaptations in the root to provide alternative respiratory pathwaysare far less important. Where these do occur, they are only of short-term use.Indeed, in plants that tolerate prolonged soil submergence, root tissues are oftenparticularly sensitive to anoxia (Vartapetian and Jackson, 1997). Without mor-phological adaptations and a continuous supply of O2 to the root tip, survivalis limited. That said, some rice genotypes will survive several days of anoxiaresulting from complete submergence of the plant following flash flooding, whichis a widespread phenomenon in rainfed rice systems. The tolerance appears todepend on (a) the water being sufficiently clear and with a sufficient dissolvedCO2 content that the plants can continue to photosynthesize and produce carbo-hydrates; (b) cessation of growth so as to preserve carbohydrates for maintenanceprocesses; and (c) increased alcoholic fermentation to maintain glycolysis, NAD

Page 176: The Biogeochemistry of Submerged Soils

168 Processes in Roots and the Rhizosphere

0.1 mm

0.1 mm

0.1 mm

crack

aerenchyma

0.1 mm

0.1 mm

branch roots

bran

ch ro

ots

CC

CC

SC

AE

CC

P

P E

P

E

SC

RH

(a)

(c)

(b)

(d)

(e)

central cylinder

aerenchyma

Figure 6.2 Cross-sections of primary rice roots. (a) Radial section close to tip showingintercellular spaces (I), central cylinder (CC), and rhizodermis (RH). (b) and (c) Radialsections of younger (39 days) and older (72 days) basal parts showing exodermis (E),schlerenchymatous cylinder (SC), parenchymatous or cortical cells (P) and aerenchyma(AE). (d) and (e) Axial sections of mature root (72 days) showing break through of lateralroots (Butterbach-Bahl et al., 2000). Reproduced by permission of verlag

recycling and ATP synthesis (Setter et al., 1997). However these adaptationswould not serve a vigorously growing root system in normal circumstances.

Transport of gases through the aerenchyma may occur by diffusion and,where pressure gradients develop, by convection. Pressurized flow is importantin wetland plants with root systems permitting a throughflow of gases, butis insignificant in other plants (Beckett et al., 1988; Skelton and Alloway,

Page 177: The Biogeochemistry of Submerged Soils

Effects of Anoxia and Anaerobicity on Plant Roots 169

Distance from apex (mm)

Res

pira

tion

rate

(ng

O2

mL−1

s−1

)

Aer

ench

yma

(%)

0 100 200 300 400 5000

50

100

150

200

250

300

0

10

20

30

40

50

60

Wall(estimated)

SteleTotal

Cortex Cortex wall

Aerenchyma

Figure 6.3 Aerenchyma development and changes in respiration rate along the length ofmaize roots grown in anoxic media (adapted from Armstrong et al., 1991a). Reproducedby permission of Backhuys publishers

1996). In throughflow systems atmospheric gases are driven or sucked intothe above-ground parts of the plant and then vented from some other point onthe above-ground parts as an O2-depleted and CO2-enriched exhaust. There arevarious possible sources of positive pressure—e.g. humidity-induced diffusionand thermal transpiration—and of negative pressure—e.g. wind (Venturi forces),the greater solubility of CO2 than O2 (140-fold at 25 ◦C and pH 7), differences ingas velocities, and thermo-osmosis (references in Jackson and Armstrong, 1999).Resistance to pressure flow is inversely proportional to the fourth power of theradius of the conducting vessel, and so large pore-diameters in the diaphragmpartitions of leaf sheath, stem and rhizome are an essential prerequisite forefficient pressurized flow. A well known example of a pressurized flow systemis the water lily (Dacey, 1980, 1981).

Pressurized flow could in principle occur in a non-throughflow root system,such as that of rice, driven by dissolution of respiratory CO2 produced fromgaseous O2. However, Beckett et al. (1988) have shown that convection by thismeans will always be subordinate to diffusion in non-throughflow systems andwill only ever have a minor effect. Hence diffusion is the principle means ofgas transport.

The effectiveness of the internal O2 transport by diffusion or convectiondepends on the physical resistance to movement and on the O2 demand. Thephysical resistance is a function of the cross-sectional area for transport, thetortuosity of the pore space, and the path length. The O2 demand is a functionof rates of respiration in root tissues and rates of loss of O2 to the soil whereit is consumed in chemical and microbial reactions. The O2 budget of the roottherefore depends on the simultaneous operation of several linked processes andthese have been analysed by mathematical modelling (reviewed by Armstrong

Page 178: The Biogeochemistry of Submerged Soils

170 Processes in Roots and the Rhizosphere

et al., 1991b, 2000). In the following section I describe the model developed byArmstrong and Beckett (1987). This accounts for the most important processeswithin the root and has been corroborated for various wetland species usingmeasurements of O2 gradients within roots with microelectrodes.

6.1.2 ARMSTRONG AND BECKETT’S MODEL OF ROOT AERATION

To summarize, the main factors influencing the O2 budget of a non-throughflowroot in anoxic soil are as follows.

(1) The extent of aerenchyma development by the degradation of the primaryroot cortex.

(2) Rates of respiration in different root tissues. The formation of aerenchymadecreases the respiratory O2 demand per unit root volume because there isless respiring root tissue. Also, some plants can tolerate a degree of anoxiain parts of the root, which substantially reduces the O2 demand per unitroot volume.

(3) The permeability of the root wall to gases. Sub-apical parts of the root canhave permeabilities several orders of magnitude smaller than those in theregion of the tip.

(4) The proportion of fine lateral roots branching off the primary root. Havinghigh surface area to volume ratios, laterals tend to be O2-leaky.

For simplicity, the effects of lateral roots are not dealt with explicitly in Arm-strong and Beckett’s model, but they are dealt with in Section 6.2.

In the model, the internal structure of the root is described as three concentriccylinders corresponding to the central stele, the cortex and the wall layers. Diffu-sivities and respiration rates differ in the different tissues. The model allows forthe axial diffusion of O2 through the cortical gas spaces, radial diffusion into theroot tissues, and simultaneous consumption in respiration and loss to the soil. Asteady state is assumed, in which the flux of O2 across the root base equals the netconsumption in root respiration and loss to the soil. This is realistic because rootelongation is in general slow compared with gas transport. The basic equation is

d

dz

(DGθGfG

d[O2]G

dz

)+ 1

r

d

dr

(rDL

d[O2]L

dr

)− Rroot − Rsoil = 0 (6.1)

where the first term represents axial diffusion through the cortical gas spaces, thesecond term radial diffusion through root tissues, and the third and fourth termsthe rates of O2 consumption in tissue respiration and loss to the soil, respectively.Here DG and DL are the diffusion coefficients of O2 in air and water, respectively,θG is the gas content of root by volume, fG is the impedance factor for diffusionin the cortical gas spaces, r is the radial distance, z is the axial distance and [O2]G

Page 179: The Biogeochemistry of Submerged Soils

Architecture of Wetland Plant Root Systems 171

and [O2]L are the concentrations of O2 in the gas space (mol per unit volumegas space) and in root tissue (mol per unit volume root), respectively.

The boundary conditions for solving Equation (6.1) are: (a) at the root base,[O2]G is the ambient value in the atmosphere; and (b) at the root apex, [O2]G isthe minimum value required for root respiration [≈ 30 µmol dm−3 (gas space)].The equations are solved numerically.

6.2 ARCHITECTURE OF WETLAND PLANT ROOT SYSTEMS

In dryland plants the size of the root system compared with the shoot system isgenerally governed by the plant’s water requirements except under quite severenutrient deficiency (Tinker and Nye, 2000). However, in wetland plants in sub-merged soil, the free availability of water means that the size of the root systemis more often likely to be governed by nutrient requirements. The length densitiesof wetland root systems may be comparable to those of dryland plants: lengthdensities of rice roots are typically 20–30 cm cm−3 in the topsoil (Matsuo andHoshikawa, 1993). A large proportion of the length may be as fine roots. In ricein submerged soil short fine laterals, 1–2 cm long and 0.1–0.2 mm in diameter,develop as branches along the primary roots once the primary roots are a fewcm long. These are much less aerenchymatous than the primary roots (porositiesof 1–2 % compared with ≤50 %) and they do not develop secondary thicken-ings in their walls to the same extent (Matsuo and Hoshikawa, 1993). They maythemselves be branched producing up to sixth order laterals. They account fora small part of the root mass but the bulk of the external surface, and they areplumbed directly into the main water and solute transport vessels in the stele ofthe primary root (as can be seen in Figure 6.2).

The structure of the rice root is therefore apparently dominated by the need forinternal gas transport. On the face of it, this structure may conflict with the needsfor efficient nutrient absorption (Kirk and Bouldin, 1991). The development ofgas-impermeable layers in the root wall seems likely to impair the ability ofthose parts of the root to absorb nutrients, and the disintegration of the cortexmight impair transport from the apoplasm to the main solute transport vessels inthe stele, though these points are uncertain (Drew and Saker, 1986; Kronzuckeret al., 1998a). It seems likely that the short fine lateral roots are responsible forthe bulk of the nutrient absorption by the root system and compensate for anyimpairment of nutrient absorption by the primary roots as a result of adaptationsfor internal aeration.

The question arises: what combination of fine laterals and aerenchymatous pri-mary roots provides the greatest absorbing surface for a given root mass? Nothaving impermeable wall layers and having a large surface area to volume ratio,the laterals will leak O2 more rapidly than the adjacent primary root. A relatedquestion is therefore how the O2 budget of the root system is affected by thecombination of primary roots and laterals. Armstrong et al. (1990, 1996) mod-elled O2 release from adventitious and lateral roots of the rhizomatous wetland

Page 180: The Biogeochemistry of Submerged Soils

172 Processes in Roots and the Rhizosphere

species Phragmites australis, and found that for the appropriate combination ofroot types, properties and dimensions, and a large but realistic soil O2 demand,the ratio of O2 consumption in root respiration to that in loss to the soil was13:1 for adventitious roots but 0.15:1, i.e. reversed, for laterals. Evidence forpreferential loss of O2 from laterals in rice includes measurements of Fe oxidecoatings on roots placed in deoxygenated agar containing Fe(II) (Trolldenier,1988); changes in redox potential as roots grew across rows of Pt electrodes inanaerobic soil (Flessa and Fischer, 1993); and the abundance of methane oxidiz-ing bacteria, which are obligate aerobes, along rice lateral roots in anaerobic soil(Gilbert et al., 1998).

Although O2 leakage compromises the root’s internal aeration, some leakage isdesirable for a number of purposes. These include oxidation of toxic products ofanaerobic metabolism in submerged soil such as ferrous iron (van Raalte, 1944;Bouldin, 1966; van Mensvoort et al., 1985); nitrification of ammonium to nitrate,there being benefits in mixed nitrate–ammonium nutrition (Kronzucker et al.,1999, 2000); and mobilization of sparingly soluble nutrients such as P (Salequeand Kirk, 1995) and Zn (Kirk and Bajita, 1995) as a result of acidification dueto iron oxidation and cation–anion intake imbalance.

6.2.1 MODEL OF ROOT AERATION VERSUS NUTRIENT ABSORPTION

Kirk (2003) has developed a simple model to compare root requirements foraeration with those for efficient nutrient acquisition in rice. The main featuresof the rice root system are summarized in Figure 6.4. The model considers rootsin the anoxic soil beneath the floodwater—soil interface, receiving their oxygensolely from the aerial parts of the plant.

Structure of the Root System

The distribution of primary roots beneath a hill of plants is approximately hemi-spherical with the individual roots randomly distributed with respect to thevertical and horizontal directions. Thus if there are N primary roots per hill,the length of primary roots per unit soil volume, LVP, at any distance r from thecentre of the hill is

LVP(r) = dN/dr

dV/dr= N

2πr2(6.2)

About each primary root there is a cylinder of laterals, increasing in densitywith distance from the root base (Figure 6.5). The laterals may develop up tosixth-order branches. A simple equation to describe this is:

LVL(r) = LVL maxr2

(k + r)2(6.3)

Page 181: The Biogeochemistry of Submerged Soils

Architecture of Wetland Plant Root Systems 173

Oxicsubsoil

Plough pan

Fine rootspenetratingplough pan

Anoxic soil

Primary roots(with laterals)in anoxic soil

Oxic surfacesoil

Floodwater

Superficial rootsin floodwaterand oxic soil

Figure 6.4 Root system of the rice plant (Kirk, 2003). Reproduced by permission ofBlackwell Publishing

where LVL is the length density of laterals in the cylinder of soil occupied bythem, k is a coefficient, equivalent to the distance at which LVL(r) = 0.25LVL max,and r0 < r ≤ rlat. If the cylinder has outer radius x and inner radius aP (i.e. theradius of the primary root), and x and aP are constant along the root length, thenthe total length density of primary and lateral roots at distance r from the centreof the hill is

LV(r) = N

2πr2

[1 + π(x2 − a2

P)LVL maxr2

(k + r)2

](6.4)

Equation (6.4) gives reasonable fits to measured profiles of LV with depth inthe field.

Structure of an Individual Root and its Laterals

The porosity of the cortex, permeability of the root wall and the coverage ofthe root with laterals vary along the root length, with a much smaller porosity,

Page 182: The Biogeochemistry of Submerged Soils

174 Processes in Roots and the Rhizosphere

radiusof hill

zone oflaterals

zone ofdecreasingporosity

zone ofroot tip

rmaxr2r1rlat

r0

Figure 6.5 Idealized primary root and its cylinder of laterals. The parallel lines indicatethe increasing length density of laterals along the primary root. The branching of the lat-erals is not represented (Kirk, 2003). Reproduced by permission of Blackwell Publishing

more-permeable wall and no laterals in the region of the tip. Where the later-als emerge from the primary root, there are generally cracks in the epidermis afew µm wide and apparently directly connected to the primary root aerenchyma(Butterbach-Bahl et al., 2000). It seems likely these will be important in gastransfer, though there are no direct measurements showing this. In practice leak-age of O2 from the cracks and axial gradients of O2 within laterals will lead togradients of O2 release along laterals. However, for the intended purpose of themodel an elaborate treatment of these effects is not necessary; it is sufficient thatthe loss of O2 increases with the density of laterals and a constant leakage alongthe length of laterals is assumed.

Figure 6.5 defines for the purposes of the model the distances at which theporosity and coverage with laterals change. It is assumed that, because of thechanges in wall permeability along the root, nutrients are only absorbed by theprimary root in the zones beyond the laterals (rlat < r < rmax) and by the laterals.This is also the surface across which O2 leaks.

Page 183: The Biogeochemistry of Submerged Soils

Architecture of Wetland Plant Root Systems 175

Transport and Consumption of O2 in the Roots and Losses to the Soil

To avoid unduly complicating the model, radial diffusion within the root is notallowed for. Equation (6.1) therefore reduces to:

d

dr

(DGθGfG

d[O2]G

dr

)− Rroot − Rsoil = 0 (6.5)

where r is the distance from the root bases (not the radial distance across theroot as in Equation 6.1). Rroot at a particular distance along the root is the sum ofthe respiration in the primary root and in any laterals emerging from it. Hence,if the rate of respiration per unit root mass is Q,

Rroot = ρ(1 − θG)Q + ρ(1 − θG2)Qπa2Lπ(x2 − a2

P)LVL

πa2P

(6.6)

Likewise Rsoil at a particular distance is the sum of the rates of loss from the pri-mary root and from the laterals. Hence, if FO2 is the flux across unit root surface,

Rsoil = 2πaPFO2

πa2P

+ 2πaLFO2π(x2 − a2P)LVL

πa2P

(6.7)

It is assumed that the primary root wall is completely impermeable to O2 in thezone covered with laterals. In fact the root wall is not completely impermeablein this zone but the resulting flux is small compared with that from the rest ofthe root system and no serious error arises from ignoring it.

It is also assumed that the flux from the laterals and the primary root in thezone beyond the laterals is constant. In fact the sink for O2 in the surroundingsoil will vary in a complicated way with soil conditions and time, and there willbe differences along the root length. However to some extent these differencescancel each other (Kirk, 2003) and the additional complexity involved in allowingfor them is unjustified.

The same boundary conditions apply as for Armstrong and Beckett’s model,and the equations are solved numerically.

Model Calculations

Figure 6.6 shows results for a realistic set of standard parameter values. Themaximum primary root length is 27.3 cm declining to 17.7 dm as the coveragewith laterals increases from <5 to 80 % of the root length. Although the maximumroot length decreases as the coverage with laterals increases, the absorbing rootsurface per unit root mass increases more than two-fold as the coverage with

Page 184: The Biogeochemistry of Submerged Soils

176 Processes in Roots and the Rhizosphere

Fraction of rmax with laterals

0.0 0.2 0.4 0.6 0.8

Abs

orbi

ng r

oot

surf

ace/

mas

s( A

R/W

R, c

m2 g

−1)

80

120

160

200

240

280

Fraction of rmax with laterals0.0 0.2 0.4 0.6 0.8

Max

imum

prim

ary

root

leng

th( r

max

, cm

)

14

16

18

20

22

24

26

28

30

0.60.70.8

Porosityof cortex

(a)

(b)

(c)

(d)

O2 respired and lost (nmol s−1)

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

Abs

orbi

ng r

oot

surf

ace

( AR, c

m2 )

0

20

40

60

80

100

120

Net root respiration (nmol s−1)

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

Loss

to

soil

( FO

2AR, n

mol

s−1

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Figure 6.6 Effect of cortical porosity of primary root and fraction of root covered withlaterals on (a) maximum primary root length, (b) absorbing root surface per unit root mass,and (c) absorbing root surface per primary root as a function of net O2 consumption,and (d) O2 consumed in root respiration and loss to the soil. Numbers on curves areporosities; other parameters have standard values (Kirk, 2003). Reproduced by permissionof Blackwell Publishing

laterals increases from <5 to 80 %. The net O2 consumption in root respirationand loss to the soil decreases as the coverage with laterals increases above about50 %, in spite of the larger surface releasing oxygen.

Figure 6.6(d) shows that root respiration is the main sink for O2, accountingfor more than 30 times the O2 loss to the soil at the minimum coverage withlaterals, though less than five times the loss to the soil at the maximum coveragewith laterals. Respiration in the lateral roots exceeds that in the primary rootby four-fold at the maximum coverage with laterals. These values compare withratios of respiration to loss of 13:1 in adventitious roots of Phragmites australisand 0.15:1 in laterals estimated by Armstrong et al. (1990) with a somewhatlarger FO2 than here.

The results broadly tally with experimental findings for rice. The maximumlength of primary root required to sustain a plant depends on soil conditions andplanting density. Typically the depth to the plough pan in a puddled ricefieldis less than 2 dm, and a typical spacing between plant hills is 25 cm × 25 cm

Page 185: The Biogeochemistry of Submerged Soils

Architecture of Wetland Plant Root Systems 177

though this varies with rice variety, soil fertility and other factors. The maximumprimary root length required to explore this volume of soil would be 26.7 cm,which is within the range calculated with the model. Note that some roots witha greater coverage of laterals could be shorter than this, exploring the soil atshallower depths.

In summary, the model shows that a system of coarse, aerenchmymatousprimary roots with gas-impermeable walls conducting O2 down to short, fine,gas-permeable laterals provides the best compromise between the need for inter-nal aeration and the need for the largest possible absorbing surface per unit rootmass. Though the model assumes a fairly simple picture of the root architectureand the changes in gas-permeability across the roots, this is the basic system inmost current rice genotypes. The significance of rates of loss of O2 to the soil ofthe magnitude calculated is considered in Sections 6.4 and 6.5.

6.2.2 ROOT SURFACE REQUIRED FOR NUTRIENT ABSORPTION

Having explored in the last section the limits that the need for internal aerationplaces on the size of the root system and its optimal architecture, we may nowconsider what root surface is required for nutrient absorption in submerged soil.I consider the case of nitrogen because it is the nutrient required in the greatestamounts. In fertile moist soil, the main plant-available form of N is usually theNO3

− ion, and because NO3− is largely not adsorbed on soil surfaces and is all

in the soil solution, its rate of delivery to absorbing root surfaces does not limitthe rate of uptake. In submerged soil, however, the principle form of N is theNH4

+ ion which is adsorbed and therefore diffuses through the soil more slowly.In quantifying the rate-limiting step in uptake and the root surface required wetherefore need to allow for the rate of transport through the soil and the rate oftransfer across the root surface.

The main transport processes involved are shown in Figure 6.7. In essencethese are the same as in a non-flooded soil: there is a dynamic equilibriumbetween solutes in the soil solution and those sorbed on the immediately adjacent

Liquid

Solid

Soi

l

Roo

t

diffusion and mass flow

very slow diffusion

rapidexchange

Figure 6.7 Solute transport processes near an absorbing root (Tinker and Nye, 2000).Reproduced by permission of Oxford University Press

Page 186: The Biogeochemistry of Submerged Soils

178 Processes in Roots and the Rhizosphere

soil surfaces. As solutes are removed from the soil solution by root uptake, thesorbed solutes tend to buffer the soil solution against the resulting changes inconcentration. The influx across the root surface, F in moles per unit area perunit time, is related to the concentration in solution at the root surface with a‘root absorbing power’, α, such that

F = αCLa (6.8)

To calculate the inflow, CLa must be found from the concentration in the soilbulk taking into account rates of transport through the soil. Kirk and Solivas(1997) have done this for N uptake by rice growing in flooded soil and used theresulting model to assess the relative importance of root uptake properties andtransport through the soil. Their results are summarized in the following.

Kirk and Solivas measured the time course of N uptake by soil-grown riceplants and the simultaneous changes in soil solution NH4

+ and root length den-sity, and then compared the results with the calculated minimum root lengthdensities required to explain the uptake. The calculation was based on the fol-lowing picture of events.

(1) All the N is absorbed as NH4+.

(2) The rate of uptake per unit root length for a given concentration of NH4+

at the root surface is maximal, as indicated by a Michaelis–Menten rela-tion derived from measurements with plants grown hydroponically undermoderate N-deficiency (Section 6.3).

(3) The concentration of NH4+ in the soil solution at root surfaces is related to

the mean concentration in the bulk soil solution by an equation for steady-state diffusion through the soil. The diffusion coefficient of NH4

+ in the soilwas measured.

The corresponding equations are as follows. For roots uniformly or randomlydistributed in volume V of soil at density LV (length per unit volume), the rateof uptake is

dU/dt = 2πaFLVV = 2πaαCLaLVV (6.9)

where a is the mean root radius. For steady-state diffusion across a cylinderof depletion around a root of radius x, the concentration maintained at the rootsurface is (from Tinker and Nye, 2000, Equation 10.24)

CLa = CL[1 − 1

2(αa/Db) + x2(αa/Db)

(x2 − a2)ln

x

a

] (6.10)

where CL is the mean concentration in solution in the soil around the root, Dthe diffusion coefficient of NH4

+ in the soil and b the buffer power for NH4+.

The assumptions inherent in this equation are discussed by Kirk and Solivas. The

Page 187: The Biogeochemistry of Submerged Soils

Architecture of Wetland Plant Root Systems 179

value of x increases as x = 2√

Dt + a until it coincides with the boundary withthe equivalent cylinders around adjacent roots, at which point, x = 1/

√πLV. In

the experiments, N as urea was mixed into the soil so that the concentration inthe floodwater was negligible. The urea N was all hydrolysed to NH4

+ within afew days.

The results are shown in Figure 6.8. They indicate that it would have been nec-essary for the whole of the measured root length to have been active in uptaketo achieve the measured uptake rates, even though very large root length densi-ties developed—up to 30 cm cm−3, which is near maximal values for the upper10 cm of soil under field conditions (Matsuo and Hoshikawa, 1993, Chapter 2,Section 2), and uptake per unit root length was near maximal.

This is somewhat surprising: in dryland crops the total root length is generallyfar larger than necessary to account for N uptake. An important difference isthat, as a result of NH4

+ adsorption on the soil solid, unlike for NO3−, the

concentration of NH4+ in the soil solution is less than Km for high affinity NH4

+transporters in the root and so V < Vmax and a larger root length is required.

N u

ptak

e (m

mol

pla

nt−1

)

0

2

4

6

8

10

Sol

utio

n N

H4+

(mM

)

0.0

0.5

1.0

1.5

2.0

20 30 40 50

Roo

t den

sity

(dm

dm

−3)

100

1000

Time (days after germination)

20 30 40 50

Roo

t de

nsity

(dm

dm

−3)

100

1000

5000

500

(a) (b)

(c) (d)

+N−N 5000

500

Figure 6.8 The time-course of (a) N uptake, (b) soil solution NH4+, and (c) and (d) root

length density in pots of flooded soil planted with rice with (Ž) and without (ž)added N. In (a), lines are fitted logistic curves, slopes of which give values of dU /dtin Equation (6.10). In (b), solid horizontal lines are CL; broken lines CLa calculatedwith Equation (6.11). In (c) and (d), the lines indicate the minimum root length densitiesrequired to explain uptake calculated with the measured CL values ± SE (full lines) andCL derived from exchangeable NH4

+ values ± SE (Kirk and Solivas, 1997). Reproducedby permission of Blackwell Publishing

Page 188: The Biogeochemistry of Submerged Soils

180 Processes in Roots and the Rhizosphere

However the calculations indicate an unlikely lack of margin for error in the rootlength. A possible explanation is that N species other than NH4

+ are also beingabsorbed—such as NO3

− or amino acids—as discussed in Section 6.3.Transport of NH4

+ to the roots in Kirk and Solivas’ experiment was mainlyby diffusion. The additional transport resulting from mass flow of soil solutionin the transpiration stream would have increased the influx across the roots byabout 100ava/0.5bD % where va is the water flux (Tinker and Nye, 2000, pp.146–148), or about 4 % in Kirk and Solivas’ experiment. A sensitivity analysisshowed that rates of diffusion will generally not limit uptake in well-puddledsoils, but they may greatly limit uptake in puddled soils that have been drainedand re-flooded and in unpuddled flooded soils.

Note that the above conclusions refer to uptake of soil N by the main bodyof the rice root system in the anoxic soil beneath the soil–floodwater interface.Uptake of fertilizer N broadcast into ricefield floodwater and absorbed by rootsin the floodwater or soil near the floodwater is not likely to be limited by rootuptake properties or transport (Kirk and Solivas, 1997).

6.3 NUTRIENT ABSORPTION PROPERTIES OF WETLANDPLANT ROOTS

6.3.1 ION TRANSPORT IN ROOTS

Taiz and Zeiger (2002) give a full account of this topic. Mineral ions absorbedfrom solution outside the root surface must be transported across the root tothe main long-distance transport vessels in the xylem, through which they reachthe shoot. This process is highly specific for different ions and molecules andis closely regulated. The regulation is in part a function of the anatomy of thevarious root tissues and in part a function of active transport processes in rootcells. The pathways and transport processes are affected by root adaptationsto anoxia.

Molecules and ions move across the root through both extracellular and intra-cellular pathways (Figure 6.9). The extracellular route exists because all cellshave walls containing solution separated from the cytosol by plasma membranes,and there is therefore a continuous ‘apoplastic’ pathway through which solutescan diffuse from one cell wall to the next without crossing a plasma membrane.There is also a continuous intracellular ‘symplastic’ pathway because the cytosolsof neighbouring cells are connected by cylindrical pores called plasmodesmata,20–60 nm in diameter, through which ions and molecules that have been takenup into the cytosol may diffuse. In tissues where significant intracellular transportoccurs there may be up to 15 plasmodesmata per square µm of cross-section.

An ion entering a root may immediately enter the symplast by crossing theplasma membrane of an epidermal cell, or it may remain in the apoplasm anddiffuse through cell walls. It may subsequently enter the symplasm by crossing

Page 189: The Biogeochemistry of Submerged Soils

Nutrient Absorption Properties of Wetland Plant Roots 181

EPIDERMISCORTEX

ENDODERMIS

XYLEM

PAREN-CHYMA

TRACHEARY

Casparianstrip

CytoplasmVacuole

Nucleus Plasmamembrane

Tonoplast

EXTERNALSOLUTION

Plasmo-desmata

Apoplasm

Figure 6.9 Idealized structure of a root showing apoplastic and symplastic pathways forsolute transport to the xylem

the plasma membrane of a cortical cell, and thence continue to the xylem. Butit cannot reach the xylem entirely through the apoplasm because its passageis blocked by a suberized layer of endodermal cells called the Casparian strip.These effectively block entry of water and mineral ions into the stele via theapoplasm. They also prevent back diffusion of solutes from the xylem apoplastto the cortical apoplast.

Finally the ion must leave the symplast of the xylem and be loaded intothe xylem’s long-distance conducting vessels. The mechanism of xylem loadingapparently involves both passive and active transfer from the xylem parenchy-mal cells.

The immediate means of regulating ion transport and hence absorption into theroot is through the control of active uptake across plasma membranes. Changesin root anatomy in response to changes in nutritional or other external conditionsare necessarily slower.

Membrane Transport Processes

The cell plasma membrane separates the cell cytoplasm from the external medium.The composition of the cytoplasm must be tightly controlled to optimize cellularprocesses, but the composition of the external medium is highly variable. Themembrane is hydrophobic and impedes solute diffusion. But it also facilitatesand regulates solute transfers as the cell absorbs nutrients, expels wastes andmaintains turgour.

Page 190: The Biogeochemistry of Submerged Soils

182 Processes in Roots and the Rhizosphere

When concentration gradients of solutes exist across a membrane the soluteswill diffuse according to their individual concentration gradients. Because ofdiffering mobilities, an electric potential exists between diffusing ions—the dif-fusion potential (Section 2.2)—and as a result the faster ions speed up the slowerones and vice versa, so that electrical neutrality is maintained everywhere inthe solution. Thus the rates of transfer of negative and positive charges areequal. However an electric potential difference across the membrane persists andis measurable.

Plant cells have several internal compartments separated by plasma membranes.The main compartments controlling the ionic relations of the cell are the cytosoland vacuole. The vacuole occupies more than 90 % of the cell volume and socontains the bulk of the ions. Many different ions permeate the membranes, butin general K+, Na+ and Cl− have the greatest concentrations and permeabilities.Studies of ion relations in plant cells have led to the following conclusions:

• K+ is accumulated in both the cytosol and vacuole by passive diffusion, exceptwhen the external concentration is small in which case its is actively taken up;

• Na+ is actively pumped out of the cytosol into the apoplasm and vacuole;• H+ generated in metabolic processes in the cell is also actively pumped out of

the cytosol to maintain a neutral pH; as a result the pH in the apoplasm andvacuole may be two units lower;

• Cl− and all other anions are actively taken up into the cytosol.

All the ions also diffuse passively according to their electrochemical gradients.It is possible to calculate the resulting diffusion potential across the membranes.Typically the diffusion potential expected from the movements of K+, Na+ andCl− is in the range −80 to −50 mV. But measured membrane potentials are gener-ally much more negative, often −200 to −100 mV, indicating that the membranepotential has a second component. The excess potential is generated by elec-trogenic H+-ATPases in the plasma membrane which pump H+ ions out. Theenergy provided by hydrolysis of ATP is used to pump out H+ against its elec-trochemical gradient. This drives the movements of other ions and moleculesacross the membrane via various transporters which both enhance and regulatethe transfers.

Three types of membrane transporter are found: channels, carriers and pumps(Figure 6.10). Channels are transmembrane proteins that function as selectivepores through which ions or uncharged molecules can diffuse passively. Theirselectivity for solutes depends on the size of the pore and the density of surfacecharges lining it. These are altered in response to external and internal stimuli inthe plant, so regulating the transport.

Carriers consist of proteins that extend completely across the membrane. Asolute being transported is initially bound to external sites on the carrier protein;subsequent changes in the conformation of the protein result in the transfer ofthe bound solute across to the other side of the membrane where it dissociates.

Page 191: The Biogeochemistry of Submerged Soils

Nutrient Absorption Properties of Wetland Plant Roots 183

Ca2+

Zn2+

Fe2+

Mn2+ H+

H2PO4−

H+

NH4+

H+

Na+H+

H+

H+

H+

NO3−

K+

sucroseaminoacid

Symporters Energy dissipated by movement of H+ backinto cell coupled to transport of one moleculeof substrate in

Na+H+

AntiporterEnergy dissipated bymovement of H+ back intocell coupled to activetransport of substrate out

Cl−K+

K+

Inwardrectifying

Outwardrectifying

Outwardrectifying

Inwardrectifying

Transmembrane pores: selectivity for solutesdepends on biophysical properties that areactively regulated

ADP + Pi

ATPH+

ADP + Pi

ATP

ADP+ Pi

ATP

H+

H+

H+ pumps

Energyfrom ATPhydrolysisused topump H+

out againstelectro-chemicalgradient ADP

+ Pi

ATP

Ca2+

Ca2+

pump

Effluxcarrier

sucrose

H+

H+H+

H+H+

H+

Ca2+

Na+

Cd2+ Mg2+

hexose

Antiporters

Ca2+

anions,cations

anions(malate2−,Cl−, NO3

−)

Channels

VACUOLEADP+ Pi

ATP

H+

CYTOSOL

H+

2Pi

PPi

H+

pumps

ADP+ Pi

ATPPC-Cd2+

ABCtransporter

Plasmamembrane

Tonoplast

sucrose

Channels

Figure 6.10 The main transport processes across the plasma membrane and tonoplast ofplant cells (adapted from Taiz and Zeiger, 2002). Reproduced by permission of SinauerAssociates

The binding is highly selective and allows transport of a wide range of bothorganic and inorganic solutes. The transport may be either passive, down anelectrochemical gradient, or, unlike for channels, active. If active, the carrier mustcouple the energetically uphill transfer of solute to a separate energy-yieldingprocess. The coupling may be direct, for example to ATP hydrolysis as in H+

Page 192: The Biogeochemistry of Submerged Soils

184 Processes in Roots and the Rhizosphere

and Ca2+ pumps; or it may be secondary with the membrane potential generatedby H+ pumps dissipated by reuptake of one or more H+ ions coupled to thetransport of a different ion or molecule. The solute may be transported into thecell—symport—or out of it—antiport. Thereby H+ ions circulate across theplasma membrane, outward through primary active transport proteins and inwardthrough secondary transport proteins.

A particular ion or uncharged molecule can be transported by different trans-porters depending on its concentration. For example NH4

+ may be absorbedby a passive low-affinity uptake system when its external concentration is largeand by an active high-affinity system when its external concentration is small.Figure 6.10 summarizes the main transport processes on the plasma membraneand tonoplast of plant cells.

6.3.2 ION TRANSPORT IN WETLAND ROOTS

Of wetland plants, rice has been studied the most extensively, and nitrogen hasbeen the most extensively studied element. In this section the rates at whichrice roots can absorb nitrogen are discussed and whether this is affected by themorphological and physiological adaptations to anoxic soil conditions.

Experimental Systems for Measuring Absorption Kinetics

The aim is to measure the influx of the nutrient into a root for a given concen-tration of the nutrient in the soil solution at the root surface. This is a seeminglysimple matter. But there are well-known difficulties in obtaining unequivocalinformation (Marschner, 1995; Tinker and Nye, 2000). The main problem is thatthe influx of the nutrient is closely regulated by the plant and depends sensitivelyon the current nutrient content of the plant as well as the external concentrationthe root is exposed to. Over time the plant will adjust its intake to the new exter-nal concentration, so the measured influx will be a function of how long the planthas been exposed to the new concentration. Measurements should therefore bemade as rapidly as possible following exposure to the new concentration.

Currently the best available technique for this for N absorption by roots usesthe short-lived tracer 13N. This is a strong γ -emitter and so can be assayedvery accurately and rapidly in fresh root tissue and thereby N fluxes across rootmembranes measured rapidly and non-destructively. Wang et al. (1993a, b) andKronzucker et al. (1998a, b, 1999, 2000) have used this technique to study NH4

+and NO3

− absorption by rice roots. In the following sections I discuss theseresults at some length. In brief the procedure is as follows. Plants were grownfor 3 to 4 weeks in hydroponic cultures with different concentrations of N, thenexposed briefly (<10 min) to solutions containing 2 to 1000 µM of N labelledwith 13N, and the kinetics of influx deduced from the accumulation of 13N in the

Page 193: The Biogeochemistry of Submerged Soils

Nutrient Absorption Properties of Wetland Plant Roots 185

plants. In other experiments the kinetics of 13N efflux out of previously labelledroots were followed, and from the results the partitioning of the 13N betweendifferent subcellular compartments was inferred.

Effects of External and Internal Nutrient Concentrations

Wang et al. (1993a, b) studied the kinetics and regulation of NH4+ absorption by

rice roots using 13N. In common with other plants and ions, this revealed at leasttwo transport systems for NH4

+ influx: one active and operating at low externalconcentrations of NH4

+ (<1000 µM); the other passive and operating at higherconcentrations and associated with a significant efflux of NH4

+ into the externalsolution. Figure 6.11 shows results for the high affinity concentration range withplants grown for 3 weeks at three different NH4

+ concentrations. The data forthe different concentrations fitted Michaelis–Menten-type equations:

V = VmaxCLa

Km + CLa(6.11)

where Vmax is the maximum influx in moles per unit root fresh weight perunit time, Km a constant and CLa the concentration in solution at the root sur-face. Table 6.1 gives the values of Vmax and Km; Vmax was four-fold larger andKm six-fold smaller for plants grown in 2 µM NH+

4 solutions than for those in1000 µM solutions.

It is apparent that the roots have considerable flexibility in their response tothe external N concentration. Influx of NH4

+ is ‘up-regulated’ as the plant’sinternal N status decreases, but suppressed as the N status increases. Hence there

[NH4+] in external solution (µM)

NH

4+ in

flux

(µm

ol g

−1 F

W h

−1)

0 200 400 600 800 10000

4

8

12

16

2 µM

100 µM

1000 µM

Figure 6.11 Concentration dependence of steady-state NH4+ influx into rice roots grown

at a range of external NH4+ concentrations. Prior to the influx measurements, the plants

were grown in solutions at the concentrations indicated on the curves (Wang et al., 1993b).Reproduced by permission of the American Society of Plant Biologists

Page 194: The Biogeochemistry of Submerged Soils

186 Processes in Roots and the Rhizosphere

Table 6.1 Parameters for Michaelis–Menten equationsfitted to NH4

+ absorption data in Figure 6.11

Initial [NH4+]

(µM)Vmax

(µmol g−1fresh wt h−1)Km

(µM)

2 12.8 32.2100 8.2 90.21000 3.4 122.1

Source: Wang et al. (1993b). Reproduced by permission of the AmericanSociety of Plant Biologists.

is unused capacity in the root transporters in plants that are not very low inN. Furthermore there is considerable efflux of absorbed NH4

+ back out of theroots, implying a futile cycling of N across the root membrane. From studiesof the kinetics of efflux of 13NH4

+ out of the roots, during a 30 min exposureto 13NH4

+ of plants grown in 100 µM NH4+, 20 % of the 13N was assimilated,

20 % sequestered in the vacuole, 40 % retained in the cytoplasm and 20 % lostthrough efflux (Wang, 1993b). Concentrations of NH4

+ in the cytoplasm werelarge—15 to 20 mM—with no sign of toxicity.

These results indicate it may be possible to improve the efficiency of absorptionand assimilation by altering the process of regulation. However the mechanismsgoverning regulation are poorly understood. It is not known whether the regula-tion is linked to the concentration of NH4

+ or NO3− itself or to the concentrations

of products of N assimilation ‘downstream’ from NH4+ or NO3

−, such as par-ticular amino acids. Nor is it known what the targets of the resulting feedbackmechanisms are.

Effects of Anoxia

The above studies were made in aerated growth media. To simulate anoxic con-ditions in submerged soil, Kronzucker et al. (1998a) grew plants for 3 weeks inaerated nutrient solutions and then transferred the plants to solutions bubbled withN2–O2 mixtures to give O2 concentrations from 100 to 15 % of air-saturation.They found that the capacity for NH4

+ absorption remained large, even at verysmall external O2 concentrations. In the early stages of exposure to hypoxia,NH4

+ absorption actually increased, but subsequently it declined reaching asteady state after a few days (Figure 6.12). Thus as the plants adapted to hypoxiainflux of NH4

+ into the roots was both up-regulated and down-regulated. Atsteady state, the maximum influx (Vmax) varied with the degree of hypoxia butthe affinity for NH4

+(Km) was constant (Table 6.2).The rate and extent of these changes are consistent with metabolic adaptations

to hypoxia rather than impairment of uptake due the changes in root morphol-ogy. Thus Kronzucker et al. (1998a) argue that the initial up-regulation of NH4

+influx was a response to cytoplasmic acidosis involving decarboxylation of N

Page 195: The Biogeochemistry of Submerged Soils

Nutrient Absorption Properties of Wetland Plant Roots 187

Duration of hypoxia (days)

Influ

x of

NH

4+

(µm

ol g

−1 F

W h

−1)

0 1 2 3 4 5 6 70

1

2

3

4

5

6

7

Figure 6.12 Effect of hypoxia on influx of NH4+ into rice roots. Seedlings were culti-

vated in nutrient solutions containing 100 µM NH+4 , aerated for 3 weeks then at 15 % O2

for indicated times (Kronzucker et al., 1998a). Reproduced by permission of the AmericanSociety of Plant Biologists

Table 6.2 Effect of hypoxia on Michaelis–Menten parame-ters for NH4

+ absorption by rice. Plants were grown in nutrientsolutions containing 100 µM NH+

4 , aerated for 21 days andthen exposed to hypoxia for 7 days

O2 pressure(% of air-saturated)

Vmax(µmol g−1fresh wt h−1)

Km(µM)

100 5.22 31.850 8.21 38.935 5.41 32.615 4.69 46.9

Source: Kronzucker et al. (1998a).

compounds to neutralize acidity, and the subsequent down-regulation of influxis as a result of restrictions in ATP supply. Though they did not measure thechanges in root morphology in response to hypoxia, other work shows thatwithin 7 days aerenchyma formation and other changes would have occurred(Kronzucker, unpublished). But since influx was at steady state after 4 days atabout 50 % of the pre-hypoxia level this had no very dramatic effect on thecapacity for NH4

+ absorption.

Ammonium versus Nitrate Absorption

Kronzucker et al. (1999, 2000) have found that lowland rice (cv IR72) grownhydroponically is exceptionally efficient in absorbing NO3

−, raising the possi-bility that rice growing in flooded soil may absorb significant amounts of NO3

−formed by nitrification of NH4

+ in the rhizosphere. This is important because(a) this NO3

− is otherwise lost through denitrification in the soil bulk (Reddy

Page 196: The Biogeochemistry of Submerged Soils

188 Processes in Roots and the Rhizosphere

et al., 1989), and (b) plant growth and yield are generally improved when plantsabsorb their nitrogen as a mixture of NO3

− and NH4+ compared with either on

its own (Layzell and Turpin, 1990; Taiz and Zeiger, 2002).Previous field research has shown the large potential for losses of NO3

−but not the potential advantages of NO3

− nutrition. Because NO3− is rapidly

reduced in the plant, and there are no simple methods for measuring NO3−

fluxes into the plant, it is difficult to quantify the extent of NO3− absorption

under field conditions.Three lines of evidence from Kronzucker et al., suggest unusually efficient

NO3− absorption. First, steady-state influx of NO3

− and NH4+ followed

Michaelis–Menten kinetics over the relevant concentration range (Figure 6.13a),and Vmax for NO3

− was some 40 % larger than that for NH4+ and Km50 %

smaller. Second, induction of the root NO3− transporters following its re-supply

to plants deprived of NO3− for 24 h was exceptionally rapid, peaking within

2 h (Figure 6.13a). For comparison, in white spruce, which is not well adaptedto using NO3

−, full induction takes several days, and in barley, which isconsidered one of the most efficient NO3

− users, full induction takes up to24 h (references in Kronzucker et al., 2000). Third, subcellular pool sizes andfluxes, estimated from the kinetics of 13N efflux out of labelled roots indicatedhighly efficient NO3

− use: while similar proportions of incoming NH4+ and

NO3− were channelled into assimilation and to the vacuole, the proportion of

NO3− translocated to the shoot was larger and that lost through efflux out of the

roots smaller (Figure 6.14).

[N] in external solution (µM)

0 100 200 300 400 500

N in

flux

(µm

ol g

−1 F

W h

−1)

0

1

2

3

4

5

6

7

8

9

NO3−

NH4+

Time after exposure to NO3− (h)

0 5 10 15 20 25

4 wk

3 wk

(a) (b)

Figure 6.13 Influx of N into roots of intact rice plants grown on 100 µM N as eitherNO3

− or NH4+: (a) concentration dependence of NO3

− and NH4+ influx in 4-week-old

plants; (b) induction of NO3− uptake in 3- or 4-week-old plants deprived of N for 24 h

before re-supply at 100 µM for the indicated periods (Kronzucker et al., 2000). Repro-duced by permission of Blackwell Publishing

Page 197: The Biogeochemistry of Submerged Soils

Nutrient Absorption Properties of Wetland Plant Roots 189

0

1

2

3

4

5

6

7

8

9

assimilation/vacuolexylemefflux

NO3−

NH4+

NO3− + NH4

+

3.18

N fl

ux (

µmol

g−1

h−1

)

2.28

0.52

0.99

1.05

2.04

2.86NO3

− = 1.12

4.19NO3

− = 1.37

0.83NO3

− = 0.32

NH4+ = 0.51

NH4+ = 2.82

NH4+ = 1.74

0.09

0.07

0.19

0.17

0.08

0.11

Figure 6.14 Fluxes of NO3− and NH4

+ within rice roots measured by analysing 13Nefflux kinetics. Plants were grown on 100 µM NO−

3 , NH4+ or NH4NO3 (i.e. [NO3

−]= [NH4

+] = 50 µM) for 3 weeks. Efflux kinetics were measured following 60 min expo-sure of roots to 13N-labelled solutions of the same composition (data from Kronzuckeret al., 1999)

When NO3− and NH4

+ were provided together in the nutrient solution at thesame total N concentration (100 µM, i.e. [NO3

−] = [NH4+] = 50 µM), NO3

−influx, accumulation and metabolism were repressed (Figure 6.14). However,plasma membrane fluxes of NH4

+, NH4+ accumulation in the cytosol and NH4

+assimilation were larger than with solely NH4

+ at 100 µM, and NH4+ efflux

was smaller. Because very little free NH4+ is translocated to the shoot, enhanced

translocation of 13N derived from 13NH4+ in the presence of NO3

− indicatesthat NH4

+ assimilation was stimulated by NO3−. As a result, net N acquisition

and translocation to the shoot were much larger than when NO3− or NH4

+ wasprovided alone.

The extent of NO3− absorption by soil-grown plants will depend on its rate of

formation and loss in the rhizosphere (this is considered in Section 6.5). Trans-porters for amino acids have also been found in plant roots, and concentrationsof amino acids in the soil solution in flooded soils can be appreciable. Thereforeit seems likely that some N is also absorbed as amino acids, but as yet we donot have the necessary data to quantify this.

Prospects for Improving the Efficiency of Absorption

These results suggest various possibilities for increasing the efficiency of Nabsorption and assimilation. The fact that the NH4

+ and NO3− transport systems

Page 198: The Biogeochemistry of Submerged Soils

190 Processes in Roots and the Rhizosphere

are down-regulated for the most part indicates that there is no need to incor-porate additional transporters. Rather, efforts should focus on manipulating theregulation of influx or decreasing efflux or both. Depending on the specific Ncompounds that trigger down-regulation, and on their subcellular location, itmight be possible to dampen down-regulation by directing N to different bio-chemical pathways or subcellular compartments or both, and, as necessary, tostore the additional N taken up in cell vacuoles. The rate of assimilation ofNH4

+, which must to a large extent occur in the root, may be limited by thesupply of carbon skeletons. The extent to which carbon skeletons are limitingwill depend on the C:N ratio of the amino acid or amide used to transport Nto the shoot, and this may be manipulable. In addition any process that led to alowering of cytoplasmic NH4

+ concentrations would presumably lower efflux.

6.4 ROOT-INDUCED CHANGES IN THE SOIL

The following root-induced changes in the soil occur as an inevitable consequenceof the nature of submerged soils and plant adaptations to them.

(1) Organic compounds are released from the root into the soil. As for drylandplants, this may account for up to 10–15 % of the photosynthate, dependingon the plant’s nutritional status and other factors. It includes microbial sub-strate, as exudate or sloughed-off material; extracellular enzymes, actively orpassively released; and complexing agents, also actively or passively released.

(2) Oxygen diffusing down through the root’s internal aerenchyma leaks out intothe soil, which has a lower O2 concentration.

(3) Mobile inorganic reductants in the soil are oxidized, particularly Fe2+ whichis precipitated as Fe(OH)3 on or near the root. As a result the concentrationof Fe2+ near the root falls and more Fe2+ diffuses in from the bulk soil. Thisis then oxidized resulting in a zone of Fe(OH)3 accumulation near the root.

(4) The oxidation of inorganic reductants generates H+:

4Fe2+ + O2 + 10H2O = 4Fe(OH)3 + 8H+

so the pH in the oxidation zone tends to fall.(5) Because the main form of plant-available N in anaerobic soil is NH4

+, theroot absorbs an excess of cations (NH4

+, K+, Na+, Ca2+, Mg2+) over anions(H2PO4

−, Cl−, SO42−). Consequently H+ is released by the root to maintain

electrical neutrality, tending to further decrease the soil pH. Note that if anyN is taken up as NO3

− as a result of nitrification of NH4+ in the rhizosphere,

the net acid–base change is the same because, although the root exports 2 molless H+ for each mol of NO3

− replacing a mol of NH4+, 2 mol of H+ are

formed in the nitrification of each mol of NH4+. Note also that Si, which is

taken up in large quantities by rice plants, crosses the root as the unchargedH4SiO4 molecule (pK1 = 9.46 at 25 ◦C).

Page 199: The Biogeochemistry of Submerged Soils

Root-Induced Changes in the Soil 191

(6) Because very large concentrations of dissolved CO2 develop in submergedsoil, in spite of root respiration the CO2 pressure outside the root may begreater than that inside it, resulting in a flow of CO2 from the soil to the atmo-sphere through the aerenchyma. Net removal of CO2 by the root decreasesthe concentration of the acid H2CO3 near the root, and this may offset theacidity produced in oxidation and excess cation uptake.

The net effects of these processes will depend on their rates versus the rates atwhich the resulting changes are buffered by processes in the soil. In the followingsections I give available information for different soil conditions.

6.4.1 OXYGENATION OF THE RHIZOSPHERE

The extent to which wetland roots oxygenate their rhizospheres is a matter ofcontention. There is little doubt that some O2 is released: reddish-brown ferricoxide deposits are frequently observed on the surfaces of wetland roots. But themagnitude of release is debated and measured rates of release vary by more thantwo orders of magnitude (Bedford et al., 1991; Sorrell and Armstrong, 1991;Kirk and Le van Du, 1997). The flux of O2 across a particular portion of theroot depends not only on the rate of O2 transport through the root—which isitself complicated by the effects of root type, age and condition—but also on thestrength of the sink presented by the external medium. In soil, the strength ofthe sink depends on the rate of O2 diffusion into the soil, its rate of consumptionby microbes and reaction with mobile reductants such as Fe2+, and the rate ofFe2+ diffusion towards the oxidation zone. There are also differences along theroot length. As a root grows through a portion of soil, a zone of Fe2+ depletionarises where oxidation is intense in the region of the root tip, but is rapidlyfilled in when the O2 supply decreases as the root grows passed. Re-reductionof Fe(III) is slow compared with oxidation of Fe(II). Hence the root tips aregenerally white and free of ferric oxide deposits, whereas the older parts arecoloured orange-brown.

The calculations in Section 6.2 indicate that the root system as a whole can sus-tain considerable rates of O2 loss to the rhizosphere without compromising theirinternal O2 requirements. The standard O2 flux in the calculations in Section 6.2was 0.5 nmol dm−2 (root) s−1 for the parts releasing O2. For rice roots grown insoil, Begg et al. (1994) obtained values of 0.1–1.2 nmol dm−2 (root) s−1 fromrates of Fe2+ oxidation and Fe(III) accumulation near planar layers of rice rootsin anaerobic soil, and Kirk and Bajita (1995) obtained 0.1–0.2 nmol dm−2 (root)s−1 with the same experimental system but a soil with a smaller ferrous ironcontent. These values probably underestimate the total O2 release because theydid not allow for O2 consumed by soil microbes. Revsbech et al. (1999) obtainedvalues of 1–3 nmol dm−2 (root) s−1 from measurements of O2 gradients madewith a microelectrode near rice roots in the soil used by Kirk and Bajita (1995).These values are in the middle of the range described above.

Page 200: The Biogeochemistry of Submerged Soils

192 Processes in Roots and the Rhizosphere

Figures 6.15 and 6.16 give the profiles of Fe(II) and Fe(III) concentrationand pH measured by Begg et al. (1994) and Kirk and Bajita (1995). Blocks ofreduced soils were placed in contact with planar layers of rice roots, with theroots separated from the soil by fine nylon mesh which they could not penetrate.

0.00

0.05

0.10

0.15

0.20

0.25

4.0

4.5

5.0

5.5

6.0

6.5

7.0

3 days

4.0

4.5

5.0

5.5

6.0

6.5

7.0

0.00

0.05

0.10

0.15

0.20

0.25

5 days

Distance from root plane (mm)

pH

[Fe(

II)],

[Fe(

III)]

(m

ol k

g−1)

0 2 4 6 8 10 120.00

0.05

0.10

0.15

0.20

0.25

4.0

4.5

5.0

5.5

6.0

6.5

7.0

10 days

pH

Fe(II)

Fe(III)

Figure 6.15 Profiles of ferrous and ferric iron and pH in blocks of two reduced soils incontact with planar layers of rice roots for indicated times in Iloilo soil. Iloilo soil is ahighly weathered sandy loam, org C = 1.2 %, aerobic pH = 3.4, reducible Fe = 80 mmolkg−1 (Begg et al., 1994). Reproduced by permission of Blackwell Publishing

Page 201: The Biogeochemistry of Submerged Soils

Root-Induced Changes in the Soil 193

0.00

0.03

0.06

0.09

0.12

0.15

0.18

7.0

7.1

7.2

7.3

7.4

pH

Fe(II)

Fe(III)

0.00

0.03

0.06

0.09

0.12

0.15

0.18

7.0

7.1

7.2

7.3

7.4

Distance from root plane (mm)

0 2 4 6 8 10 120.00

0.03

0.06

0.09

0.12

0.15

0.18

7.0

7.1

7.2

7.3

7.4

12 days

6 days

3 days

pH

[Fe(

II)],

[Fe(

III)]

(m

ol k

g−1)

Figure 6.16 As Figure 6.15 but with Maahas soil. Maahas soil is a dark humic clay,org C = 18 %, aerobic pH = 5.9, reducible Fe = 30 mmol kg−1 (Kirk and Bajita, 1995).Reproduced by permission of Blackwell Publishing

Over 2 weeks of root–soil contact, Fe2+ close to the roots was oxidized by O2

from the roots, and substantial quantities of iron were transferred towards theroot plane producing a well-defined zone of Fe(OH)3 accumulation. The pH inthe oxidation zone fell. In both the soils studied, the amount of H+ formed in

Page 202: The Biogeochemistry of Submerged Soils

194 Processes in Roots and the Rhizosphere

Fe(II) oxidation was comparable to that released from the roots to balance excesscation uptake. But in both soils the H+ generated in these two processes exceededthe acidification calculated from the pH profile and the soil pH buffer powers.This was possibly because of CO2 uptake by the roots and, in the Maahas soil,where acidity diffusion was fast because of the high pH and high CO2 pressure,because the acidification spread beyond the zone of soil analysed.

6.4.2 THE pH PROFILE ACROSS THE RHIZOSPHERE

The pH profile across the rhizosphere resulting from the above processes dependson the rate at which acidity is generated versus the rate at which the resulting pHchange is propagated away through the soil. In the Iloilo soil (Figure 6.15), whichis a highly weathered sandy loam, the pH at the root surface fell by more than2 units, whereas in the Maahas soil (Figure 6.16), a less-weathered humic clay,it fell by only 0.2 units. Apart from differences in the rate of acidity generation,the Maahas soil had a greater initial pH and a greater organic C content resultingin a greater CO2 pressure when flooded. These together result in faster aciditydiffusion through the soil.

The continuity equation for diffusion of acidity in the soil surrounding a rootis (Equation 2.32 expressed in radial coordinates)

∂pH

∂t= 1

r

∂r

(rDHS

d pH

dr

)(6.12)

where DHS is the soil acidity diffusion coefficient, given by:

DHS = 2.303θLfL

bHS(DLH[H3O+] + DLC[HCO3

−]) (6.13)

The explanation of Equation (6.13) is that a small portion of soil near the rootmay gain acidity either by access of H3O+(S— + H3O+ = S—H+ + H2O) orby dissociation of CO2 and removal of HCO3

− though the soil solution (S— +H2CO3 = S—H+ + HCO3

−) (Section 2.2). Since [H3O+] and [HCO3−] are both

sensitive to pH, the rate at which pH changes are propagated is also sensitiveto pH and passes through a minimum in the pH range 4.5–6.0 where [H3O+]and [HCO3

−] are both small. In this pH range, therefore, a flux of acid or basethrough the soil causes steep pH gradients.

Figure 6.17 shows pH profiles calculated with Equations (6.12) and (6.13) withparameters appropriate for a rice root in anaerobic soil. The pH change at theroot surface is small at pHs above neutral because DHS is large, and it increasestowards the pH at which DHS is minimal. Because of this, the combined effect ofthe two sources of acidity in the rice rhizosphere is greater than the sum of theireffects in isolation. So in spite of θLfL and the CO2 pressure both being largein submerged soils, if the rate of generation of acidity is sufficiently large theremay be substantial pH changes at the root surface. These effects are summarized

Page 203: The Biogeochemistry of Submerged Soils

Root-Induced Changes in the Soil 195

Distance from root surface (mm)

pH

0.0 0.5 1.0 1.5 2.0 2.5 3.04.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

Figure 6.17 Calculated pH profiles around a rice root exporting acid into anaerobic soilat different initial pHs. Other parameter values: H+ flux across root = 2.5 nmol dm−2 s−1,root radius = 0.2 mm, pH buffer power = 0.05 mol dm−3pH−1, CO2 pressure = 1 kPa,θL = 0.7, fL = 0.4, time = 12 days (after Nye, 1981)

initial pH

6.0 6.5 7.0 7.5−1.50

−1.25

−1.00

−0.75

−0.50

−0.25

0.00

pH buffer power (mol kg−1 pH−1)

0.00 0.02 0.04 0.06 0.08 0.10

CO2 pressure (kPa)0.0 0.5 1.0 1.5 2.0

flux of H+ across root (nmol dm−2 s−1)

0 1 2 3 4 5

fluxof H+

initial pH

CO2pressure

pH bufferpower

Cha

nge

in p

H a

t ro

ot s

urfa

ce

Figure 6.18 Sensitivity of pH change at root surface to important variables, varied indi-vidually. Standard values as in Figure 6.17 with initial pH = 6.75

Page 204: The Biogeochemistry of Submerged Soils

196 Processes in Roots and the Rhizosphere

in Figure 6.18, which shows the sensitivity of the pH change at the root surfaceto the flux of H+ across the root and the important soil variables.

6.5 CONSEQUENCES OF ROOT-INDUCED CHANGES

Effects of the root-induced changes on the general microbiology of submergedsoils are discussed in Chapter 5 and effects on methane production and consump-tion are discussed in Chapter 8. I here discuss specific effects on plant nutrients.

6.5.1 NITRIFICATION–DENITRIFICATION IN THE RHIZOSPHERE

As we have seen in Section 6.4, wetland rice is particularly efficient at absorbingNO3

−. Kirk and Kronzucker (2000) developed a model to calculate the extentto which rice growing in submerged soil can capture NO3

− formed in the rhi-zosphere before it diffuses away and is denitrified in the soil bulk. The modelallows for the following processes.

(1) Transport of O2 away from a root and its consumption in microbial pro-cesses—in addition to nitrification—and oxidation of mobile reductants suchas Fe2+. Microbial O2 consumption is described with Michaelis–Mentenkinetics and Fe2+ oxidation with first-order kinetics with respect to both[O2] and [Fe2+].

(2) Transport of NH4+ towards the root and its consumption in nitrification and

uptake at the root surface. Nitrification is described with dual Michaelis–Menten kinetics allowing for [O2] and [NH4

+].(3) Transport of NO3

− formed from NH4+ towards the root and its consumption

in denitrification and uptake by the root. Denitrification is described withMichaelis–Menten kinetics with an inhibition function related to [O2].

Uptake of NH4+ and NO3

− into the root are described by Michaelis–Mentenrelations with the parameter values discussed in Section 6.3.

Figure 6.19 shows the concentration profiles of O2, Fe2+, NH4+ and NO3

−near a root calculated with this model for realistic flooded soil conditions andrealistic rates of O2 release (last section); Figure 6.20 shows the correspondingfluxes of O2 out of the root and NH4

+ and NO3− in. The amount of N denitrified

in 10 days in the calculations corresponds to about 10 % of the NH4+ initially in

the volume of soil influenced by the root. This is of the order of maximum ratesof denitrification reported in the literature for rice in flooded soil, indicating thatthe model parameter values are indeed realistic.

The calculations indicate that quite large amounts of NO3− may be absorbed

by rice in flooded soils, perhaps as much as a third of the total N absorbedif soil conditions and water management prevent very thorough soil reduction.This may explain, for example, the benefit of the elaborate water management

Page 205: The Biogeochemistry of Submerged Soils

Consequences of Root-Induced Changes 197

Distance from root surface (mm)

Sol

utio

n O

2, N

H4+ ,

NO

3− co

nc. (

µM)

Sol

utio

n F

e2+ c

onc.

(µM

)

0 1 2 30

20

40

60

80

950

960

970

980

990

1000

Fe2+

NH4+

NO3−

O2

Figure 6.19 Calculated concentration profiles of O2, NO3−, NH4

+ and Fe2+ in a floodedsoil near a rice root after 10 days of root–soil contact. The parameter values used in thecalculations are realistic for a healthy root growing in an unexceptional lowland ricesoil (Kirk and Kronzucker, 2000). Reproduced by permission of IRRI

Time (days)

Flu

x of

O2

out o

r N

H4+ ,

NO

3− or

N in

(nm

ol d

m−2

s−1

)

0 2 4 6 8 100.0

0.2

0.4

0.6

0.8

1.0

1.2O2

N

NH4+

NO3−

Figure 6.20 Calculated fluxes of O2, NO3− and NH4

+ across the root over time. Param-eter values as in Figure 6.19

schemes practiced for rice in parts of China and Japan, involving intermittentdrainage of water from the fields during the season (Section 7.2 and Figure 7.4).But further research is needed to quantify how far mixed NH4

+ –NO3− nutrition

operates under field conditions and its benefits to rice growth.

6.5.2 SOLUBILIZATION OF PHOSPHATE

Deficiency of P is often the main nutrient limitation in natural wetlands, thoughit is rarely important in wetland rice soils that have at least some history of P

Page 206: The Biogeochemistry of Submerged Soils

198 Processes in Roots and the Rhizosphere

fertilization. Phosphate tends to be solubilized by the electrochemical changesfollowing soil submergence, but with prolonged submergence it may becomere-immobilized as reduced phases are precipitated (Section 4.3). Some of thisP might be re-solubilized by root-induced oxidation of the rhizosphere soil. Onthe other hand, precipitation of amorphous Fe(OH)3 close to and on rice rootsmight be expected to immobilize P from the soil solution, impeding its access tothe roots.

Saleque and Kirk (1995) measured concentration profiles of P and other root-induced changes near planar layers of rice roots growing in a highly weatheredP-deficient soil to which different amounts of P had been added (Figure 6.21). Inboth P-fertilized and -unfertilized soil, the quantity of readily plant-availableP was negligible, so it was necessary for the plants to solubilize P. Some90 % of the P taken up was drawn from acid-soluble pools, probably associ-ated with Fe(II) carbonates and hydroxides. There were also narrow zones of Paccumulation in an alkali-soluble pool which coincided with zones of Fe(OH)3accumulation near the roots. The zone of P depletion coincided with a zone ofacidification, caused by the processes discussed in Section 6.4. Kirk and Saleque(1995) showed with a model of this system that the acidification and the P-solubilizing effect of acidity in the soil were sufficient to account for the Pmobilized and absorbed by the roots. Solubilization accounted for at least 80 %of the P taken up in both the P-unfertilized and -fertilized soil, though onlyabout half the P solubilized was absorbed because the rest diffused away fromthe roots. The amount of P solubilized greatly exceeded the amount immobilizedon Fe(OH)3 precipitated near the roots. This is an extreme example, involvingparticularly large pH changes, but it indicates the magnitude of the effects thatare possible.

By contrast, when a submerged soil is dried and oxidized, immobilization of Pon the ferric oxide formed may be the dominant process and plants may becomeseverely P deficient (Section 4.3). Huguenin-Elie et al. (2003) investigated themechanisms by which rice growing in alternately submerged and drained soilsextract P by measuring uptake from moist, flooded or flooded then moist soilsand comparing the results with model calculations allowing for solubilizationby various means. In all three water regimes the plants relied on solubilizationfor most of their P. The roots were not mycorrhizal, as they will often not bein intermittently flooded soils. In the moist soil, the uptake was only a thirdof that in the flooded soils and was consistent with solubilization by organicanion excretion from the roots, which appears to be the mechanism by whichupland rice in aerobic soil extracts P (Kirk et al., 1999; Trolove, 2000). In thesubmerged then moist soil, uptake declined sharply as the soil dried becauseP became immobilized. The final uptake was similar to that in the continuouslymoist soil, indicating that some of the immobilized P was re-solubilized by roots,possibly by excretion of organic anions.

Page 207: The Biogeochemistry of Submerged Soils

Consequences of Root-Induced Changes 199

P e

xtra

cted

(µm

ol g

−1)

0.0

0.4

0.8

1.2

1.6

P e

xtra

cted

(µm

ol g

−1)

0.8

1.2

1.6

2.0

2.4P

ext

ract

ed (

µmol

g−1

)

1.5

1.6

1.7

1.8

1.9

2.0

P e

xtra

cted

(µm

ol g

−1)

0.3

0.4

0.5

0.6

0.7

P50

P15

P0

P50

P15

P0

P50

P15

P0

mm from root plane

0 2 4 6 8 10 0 2 4 6 8 10

Fe(

II) (

µmol

g−1

)

5

15

25

35

45

Fe(

III)

(µm

ol g

−1)

45

55

65

75

85

pH

3.5

4.0

4.5

5.0

5.5

6.0

P0

P50

P15

Acid-1-P

Alkali-Po

Fe(III)Fe(II)

Acid-2-P

Alkali-Pi

Figure 6.21 Profiles of P, Fe and pH near a planar layer of rice roots in contact withflooded soil fertilized containing 0, 15 or 50 mg P kg−1 for 6 weeks. The P pools mea-sured sequentially were: readily available P extracted by anion-exchange resin (negligibleat all P levels and therefore not shown); readily acid-soluble P (Acid-1-P), extractedby anion-exchange resin + H+-form cation-exchange resin; alkali-soluble inorganic P(Alkali-Pi), by 0.1 M NaOH; alkali-soluble organic P (Alkali-Po), by digesting the previousextract and subtracting the alkali-soluble inorganic P; the more recalcitrant acid-solubleP (Acid-2-P), by 1 M HCl + 1 M H2SO4. (Differences between P levels not significantfor Fe and Alkali-Po) (Saleque and Kirk, 1995). Reproduced by permission of BlackwellPublishing

Page 208: The Biogeochemistry of Submerged Soils

200 Processes in Roots and the Rhizosphere

6.5.3 SOLUBILIZATION OF ZINC

Zinc is often highly insoluble in submerged soils and Zn deficiency is an impor-tant constraint to rice production throughout Asia (Chapter 7). In similar experi-ments to those in the last section, Kirk and Bajita (1995) measured changes in Znpools near rice roots in anaerobic soil (Figure 6.22) and simultaneous changes inFe(II), Fe(III) and pH (Figure 6.16). As for P in the soil in Figure 6.21, the con-centration of easily extractable Zn in the soil was negligible following floodingand it was necessary for the plants to solubilize Zn to meet their requirements.Zinc was mobilized from highly insoluble forms in the soil and re-precipitatedwith Fe(OH)3 and organic matter within 4–5 mm of the roots. The accumulationcontinued over time but simultaneously there was a substantial depletion of theaccumulated fractions within 2 mm of the roots. The zones of accumulation anddepletion coincided with zones of Fe(III) accumulation and soil acidification. Theauthors concluded that Fe oxidation released Zn from highly insoluble forms andthat this Zn was re-adsorbed on Fe(OH)3 and on organic matter in forms thatwere acid-soluble and therefore accessible to the plants.

An additional benefit of acidification of the soil close to the root may be tolower the concentration of HCO3

− in solution. High HCO3− impairs Zn absorp-

tion (Dogar and Hai, 1980), its translocation to the shoot (Forno et al., 1975),root growth (Yang et al., 1994), or all three. The modest decrease in pH near theroots in Figure 6.16 (from pH 7.35 to 7.1) is equivalent to a two-fold increasein H+ concentration and if the CO2 pressure is constant a two-fold decrease inHCO3

− concentration.There may also be effects via the concentrations of competing cations at the

root surface. In studies of short-term uptake of 65Zn by rice from nutrient solutionscontaining realistic Zn2+ concentrations, Giordano and Mortvedt (1974) founduptake was inhibited by various metabolic inhibitors and by Fe2+, Mn2+, Ca2+and Mg2+ as Cl− salts at typical concentrations in flooded soil solutions. Translo-cation of absorbed Zn was also inhibited by Fe2+ and Mn2+ but not by Ca2+ orMg2+. Cayton et al. (1985) also found antagonistic effects of competing cationson Zn uptake. Absorption involves preferential binding of Zn2+ at cation exchangesites in the root cell walls prior to active uptake across the plasma membrane.Preferential binding concentrates Zn2+ at the sites of active uptake, and would besensitive to the concentrations of competing cations. But it is not clear whethercompetition for exchange sites or for transporters in the plasma membrane is themore important (Reid et al., 1996).

6.5.4 IMMOBILIZATION OF CATIONS

Metal cations in the soil solution may be immobilized by sorption onto iron‘plaque’ on root surfaces in submerged soils, in the same way that solubilizedZn2+ was re-adsorbed on ferric oxide in the experiments in Figure 6.22. Seques-tering of metals on the external surfaces of wetland roots in this way limits uptake

Page 209: The Biogeochemistry of Submerged Soils

201

mm

fro

m r

oot

plan

e6 da

ys12

day

s

02

46

810

120

24

68

1012

02

46

810

12

[Zn] in indicated pool (mmol kg−1)

0.0

0.1

0.2

0.3

0.4

(a)

(b)

(c)

0 da

ysC

uAc

Aci

dN

H4O

x

CB

DH

ypot

hetic

al p

rofil

eif

no u

ptak

e

KC

l

Fig

ure

6.22

Profi

les

ofZ

nfr

actio

nsin

anae

robi

cso

ilne

ara

plan

arla

yer

ofri

cero

ots

afte

rin

dica

ted

times

ofro

ot–

soil

cont

act.

The

dash

edlin

esar

eth

ehy

poth

etic

alpr

ofile

sif

noZ

nha

dbe

enre

mov

edby

the

plan

ts,

estim

ated

byin

terp

olat

ion

ofth

elin

esbe

yond

the

depl

etio

nzo

nes.

Cor

resp

ondi

ngpr

ofile

sof

Fe(I

I),

Fe(I

II)

and

pHar

ein

Figu

re6.

16(K

irk

and

Baj

ita,

1995

).R

epro

duce

dby

perm

issi

onof

Bla

ckw

ell

Publ

ishi

ng

Page 210: The Biogeochemistry of Submerged Soils

202 Processes in Roots and the Rhizosphere

of metal pollutants into the vegetation (Otte et al., 1989; St-Cyr and Crowder,1990; Ye et al., 1997a, b; Hansel et al., 2001).

In soils with small pH buffer powers, root-induced acidification may alsoimpede access of cations to roots, though it may diminish sorption onto oxidesas the surface negative charge decreases (Razafinjara, 1999). This is because theoverall concentration of the soil solution in a submerged soil depends largelyon the concentration of HCO3

−, buffered by dissolved CO2. Therefore, if thepH close to the root decreases below about 6.0, the concentration of anions insolution also decreases and so the concentration of cations in solution and hencetheir rate of diffusion to roots must decrease [Bouldin (1989) gives calculationsof this effect]. This may happen, for example, in ‘iron toxic’ soils which developlarge concentrations of Fe2+ in solution. High rates of Fe2+ oxidation and asso-ciated H+ generation result in a low pH in the rhizosphere, especially if the soilis already acid or has a small pH buffer power. Hence the need to exclude toxicFe2+ from the root by oxidizing it in the rhizosphere may impair the absorptionof nutrient cations by the root. Consistent with this the symptoms of iron toxicityare often alleviated by applications of K salts.

A further complication is that the lowering of the rhizosphere pH and con-sequent depression of HCO3

− means that any Fe2+ entering the root will beaccompanied by a proportion of Cl− or SO4

2− rather than HCO3−. When Fe2+

enters with HCO3−, the acidity generated in Fe2+ oxidation in the plant is neutral-

ized by conversion of HCO3− to CO2, which is assimilated or lost. However when

Fe2+ enters with a non-volatile anion, Fe2+ oxidation will produce the equiva-lent amount of free H+ in the plant, with damaging effects on plant tissues (vanMensvoort et al., 1985).

6.6 CONCLUSIONS

This chapter has shown the complexity of the chemical and biological processesaround wetland plant roots and the effects of the extreme electrochemical gradientbetween the root surface and surrounding soil. Models of nutrient uptake byplants in aerobic soil, which treat the root as a simple sink to which nutrientsare delivered by mass flow and diffusion but the root not otherwise influencingthe surrounding soil, work reasonably well for the more soluble nutrient ions.However, the complexity of the wetland root environment is such that suchmodels are inadequate and more elaborate treatments are necessary. Many of themechanisms involved are still poorly defined and speculative, but their potentialsignificance is clear.

Page 211: The Biogeochemistry of Submerged Soils

7 Nutrients, Toxins and Pollutants

This chapter is concerned with the different types of wetland soil as sources,sinks and transformers of nutrients, particular nutrient deficiencies and mineraltoxicities that commonly arise following submergence, and the fate of pollutantsthat are commonly added to submerged soils, both accidentally and intentionally.

7.1 NUTRIENT AND ACIDITY BALANCES

As discussed in Chapter 1, the nutrient supplies of submerged wetland soils havea number of special features compared with upland soils. Nutrient removal inleaching and erosion tend to dominate the nutrient balance of upland soils. Inwetlands, inputs from inflowing water tend to exceed leaching losses, and erosiondoes not occur. Rates of fixation of nitrogen from the atmosphere under wetconditions with a good supply of other nutrients are also often greater, though sotoo are losses through denitrification. Wetlands are therefore often net sinks fornutrients, and this is reflected in their generally high productivity. However thereare large differences between the different wetland types and between locations.

7.1.1 NUTRIENT BALANCES IN RICEFIELDS

The fact that rice production has sustained huge human populations on the riverdeltas of Asia for millennia is an indication of the favourable nutrient balance inwetland ricefields. Prior to 1960 and the green revolution, yields were sustainedwithout artificial inputs of nutrients, other than by recycling through manures andnight soil. The requirements of the crop were met by the inflow of nutrients andfertile sediments with floodwaters and by nitrogen fixation. The greatly increasedyields since 1960 are sustained by inputs of mineral fertilizers. Figure 7.1 showshow increases in the yields of rice (and wheat) have been paralleled by increasesin the use of nitrogen fertilizers since 1960 to keep pace with world population.

Greenland (1997) has compiled realistic average annual nutrient balances forwetland ricefields pre- and post-1960 from probable inputs and outputs. Inputscome from rainfall, R, irrigation and floodwater, F , sediments, S, nitrogen fix-ation, N , and manures and fertilizers, M . Outputs are due to crop removals in

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 212: The Biogeochemistry of Submerged Soils

204 Nutrients, Toxins and Pollutants

Year AD1860 1880 1900 1920 1940 1960 1980 2000

0

1

2

3

Pop

ulat

ion

(bill

ions

)

Ara

ble

or ir

rigat

ed a

rea

(Gha

)

Fer

tiliz

er N

(M

t)

Wor

ld a

vera

ge r

ice

and

whe

at y

ield

s (t

ha−1

)

4

5

6

0.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

3.6

4.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0

20

40

60

80

100

120

140

160

180

arable area

rice

wheat

population

fertilizerN

irrigatedarea

Figure 7.1 Increases in world population, arable area, average yields of rice and wheat,amount of fertilizer N used, and the irrigated area of the world (Evans, 1997). Reproducedby permission of the Royal Society

rice grain and straw and legumes, C, seepage and percolation, S, and gaseousemission, G. Therefore the balance, B, is

B = (R + F + S + N + M) − (C + P + G) (7.1)

Macronutrients

The macronutrient balances for irrigated rice are shown in Figure 7.2. Pre-1960,there were no inputs from mineral fertilizers though manures were applied, butthere were substantial inputs from sediments borne by the irrigation water, whichwas obtained by stream diversion. The sediment supplied the bulk of the annualP and K inputs, and the bulk of the N was derived from biological fixation inthe soil and floodwater during the rice crop and in legume crops following therice. The calculation is for a single rice crop of 2 t ha−1 of grain and 3 t ha−1

of straw and 0.5 t ha−1 of legume, all removed from the field. The figure showsthat the balances of N and K are positive but that of P substantially negative. Apositive balance for P is attained with a crop of 1 t ha−1 of rice grain if the strawis retained.

Page 213: The Biogeochemistry of Submerged Soils

Nutrient and Acidity Balances 205

0

100

200

300

400

M

N

P

Cl

Cs

Cg

G

0

10

20

30

40

50

0

50

100

150

200

250

300

Cl

Cs

Cg

P

M

FR

Cl

Cs

Cg

P

R

Irrigated rice after 1960

Irrigated rice before 1960

0

20

40

60

80

100

N (

kg h

a−1)

N (

kg h

a−1)

P (

kg h

a−1)

P (

kg h

a−1)

K (

kg h

a−1)

K (

kg h

a−1)

120

140

M

N

S

FR

P

Cl

Cs

Cg

G 0

2

4

6

8

10

12

14

0

20

40

60

80

100

120

140

160

180

Cl

Cs

Cg

PM

S

R

M

S

F

R

Cl

Cs

Cg

P

F

In Out In Out In Out

In Out In Out In Out

F

RF

Figure 7.2 Typical annual nutrient balances for irrigated rice soils pre- and post-1960calculated from probable inputs (left side of each graph) and outputs (right side) (datafrom Greenland, 1997). Inputs come from R = rainfall, F = floodwater and irrigation,S = sediments, N = nitrogen fixation and M = manures and fertilizers. Outputs are dueto removals in Cg = rice grain, Cs = rice straw, Cl = legume crop, S = seepage and per-colation and G = gaseous emission

Post-1960 and the green revolution, much larger quantities of nutrients areremoved in intensive double and sometimes triple rice cropping. The irrigationwater is now often obtained by storage in reservoirs behind high dams and deliv-ered through artificial channels so the sediment settles out of the water and nolonger reaches the ricefields with its load of nutrients. However there are nowmuch larger additions of nutrients in mineral fertilizers. A greater proportion ofthe mineral fertilizer N is lost through gaseous emissions, and, because undermultiple cropping the soil is flooded for more of the year, percolation losses aregreater, particularly of K.

Page 214: The Biogeochemistry of Submerged Soils

206 Nutrients, Toxins and Pollutants

The minimum quantity of nitrogen that the crop must accumulate to produce1 t of rough rice is of the order of 16–18 kg, though efficiencies are lower atlow yield levels (Cassman et al., 1998; Dobermann et al., 2002). Thus 8 t ofrice per ha requires 128–144 kg N ha−1. Experiments without additions of Nin a wide range of rice environments over many years have shown that from50–80 kg N ha−1 can be accumulated from soil reserves that are replenished bybiological N fixation and crop residues. For higher yields, additional N must beprovided from sources outside the ricefield, either as organic manures or mineralfertilizers.

The calculations in Figure 7.2 are for a rice–rice–legume cropping systemremoving 8 t ha−1 of rice grain (3 in the wet season, 5 in the dry), 8 t ha−1 ofstraw (i.e. a harvest index of 0.5 compared with 0.4 in the pre-1960 varieties) and1 t ha−1 of legume. A positive balance is maintained for N and P, but becausethere are no additions of K fertilizer and no additions with sediment, there is asteady depletion of soil K reserves: 195 kg ha−1 year−1 if straw is removed or35 kg ha−1 year−1 if retained. Such a picture is not unrealistic, and yield limi-tations due to K deficiency in irrigated rice systems are becoming increasinglycommon across Asia (Dobermann et al., 1996). The importance of mineral fer-tilizers in sustaining intensive rice production is evident.

Figure 7.3 shows corresponding balances for rainfed rice systems. Here theyields are more modest: 2, 3 and 0.5 t ha−1 of rice grain, straw and grain legume,respectively, for rainfed lowland rice; 2 and 8 t ha−1 of grain and straw for flood-prone rice; and 1, 2 and 0.5 t ha−1 of grain, straw and grain legume for uplandrice. The nutrient budgets are correspondingly smaller. In rainfed lowland rice,as for the pre-1960 irrigated rice, positive balances are maintained for N and Kbut a negative balance for P; a positive P balance is maintained if only 1 t ha−1

of rice is harvested and the straw retained in the fields. The supplies of P andK in sediment are critical. In flood-prone rice the balances of all three nutrientsare negative if straw is removed, but roughly balanced if it is retained. Nitrogenfixation in the deep, low P water is diminished, and the supplies of N, P andK in sediment are crucial. In upland rice, the N, P and K balances are againall negative, even for yields of only 1 t ha−1. Few soils are sufficiently fertile towithstand such withdrawals for very long, hence the importance of the restorativefallow for several years in shifting cultivation systems based on upland rice.

Secondary and Micronutrients

In the colluvial and alluvial soils of the main rice producing areas, the amounts ofcalcium and magnesium deposited in irrigation and floodwaters, especially if theycarry sediment, usually far exceed crop removals (Greenland, 1997; Dobermannand Fairhurst, 2000). Deficiencies do however occur in rainfed lowland rice,especially on highly weathered soils where they are compounded by deficienciesof other nutrients. The only secondary nutrient that is commonly removed ingreater amounts than it is supplied in inflowing water is sulfur. Deficiencies of

Page 215: The Biogeochemistry of Submerged Soils

Nutrient and Acidity Balances 207

0

20

40

60

80

N

Cs

Cg

0

2

4

6

8

10

12

0

50

100

150

200

250

Cs

Cg

F

R

Cs

Cg

F

Flood-prone rice

Rainfed lowland rice

0

20

40

60

80

100

120

M

N

SFR

Cl

Cs

Cg

V 0

2

4

6

8

10

12

0

20

40

60

80

100

120

140

160

Cl

Cs

Cg

M

S

R

M

S

FR

Cl

Cs

Cg

0

20

40

60

80

100

N

R

P

Cl

Cs

Cg

G0

2

4

6

8

10

0

20

40

60

80

Cl

Cs

Cg

PR R

Cl

Cs

Cg

P

Upland rice

S

R

S

FR

S

In OutIn OutIn Out

In OutIn OutIn Out

In OutIn OutIn Out

N (

kg h

a−1)

N (

kg h

a−1)

N (

kg h

a−1)

P (

kg h

a−1)

P (

kg h

a−1)

P (

kg h

a−1)

K (

kg h

a−1)

K (

kg h

a−1)

K (

kg h

a−1)

Figure 7.3 Typical annual nutrient balances for rainfed rice soils from probable inputs(left side of each graph) and outputs (right side) (data from Greenland, 1997). Inputscome from R = rainfall, F = floodwater and irrigation, S = sediments, N = nitrogenfixation and M = manures and fertilizers. Outputs are due to removals inCg = rice grain, Cs = rice straw, Cl = legume crop, S = seepage and percolation andG = gaseous emission

Page 216: The Biogeochemistry of Submerged Soils

208 Nutrients, Toxins and Pollutants

sulfur have become more widespread as rice yields and the intensity of croppinghave increased (Dobermann et al., 1998). However, while deficiencies of sulfurin crops in Europe have increased in recent decades as deposition from industrialpollution has decreased, deficiencies are being offset in rice in large parts of Asiaby increasing emissions from industry and the large-scale burning of forest.

Though it is not strictly an essential nutrient, rice plants accumulate very largeamounts of silicon–shoot contents typically exceeding 5 %–and this has variousbeneficial effects on the plant (Savant et al., 1997). Silicon adds to the mechan-ical strength of cell walls, confers resistance to certain pests and diseases, andis thought to offset abnormalities in the supply of certain other nutrients. Defi-ciencies occur in highly weathered soils from which the soluble silicon has beenleached, and in organic soils with low mineral reserves. More widespread inci-dences are expected as rice cropping continues to intensify (Dobermann andFairhurst, 2000).

Of the essential micronutrients (Fe, Mn, Zn, Cu, B and Cl), deficiency ofzinc is the most commonly reported (Quijano-Guerta et al., 2002). Generally theamounts of zinc and other micronutrients brought in with irrigation water, rainfalland sediments are more than sufficient to offset crop removals. Deficienciesarise as a result of particular changes in the soil following submergence causingimmobilization (Section 7.2). Such problems become more acute the greater therate of removal in cropping, and increases in incidences of Zn deficiency areexpected with the advent of Zn-dense rice for improved human nutrition (Welchand Graham, 1999).

7.1.2 ACIDITY BALANCES IN RICEFIELDS

A further reason for the long-term sustainability of wetland rice farming is thatthe soils tend not to become acid after continuous, intensive cultivation. In culti-vated upland soils, acidification occurs because of the leaching of the NO3

− ion.Nitrification of NH4

+ added in mineral fertilizers produces 2 mol of H+ per molof NH4

+ nitrified (Table 7.1, Process 3). If some of the NO3− is subsequently

leached from the soil accompanied by an exchangeable cation, 2H+ are left behindper mol of NO3

− leached, acidifying the soil. Because in general little or no NO3−

is leached through submerged rice soils, any NO3− entering the soil or formed

in it being either absorbed by the crop or denitrified, this process does not occur.The removal of fertilizer N in the crop as NH4

+ does not lead to acidification.Hydrolysis of urea fertilizer—by far the main form of N fertilizer used in wetlandrice, together with ammonium bicarbonate in some countries—consumes 1 molof H+ per mol of NH4

+ formed (Table 7.1, Process 1). So although absorptionof N as NH4

+ leads to a net export of H+ from the roots to balance the resultingexcess intake of cations over anions (Table 7.1, Process 5), this acidity is matchedby the H+ consumed in urea hydrolysis. Likewise there is no net generation ofacidity as a result of NH3 volatilization, although 1 mol of H+ is left behind permol of NH4

+ converted to NH3 (Table 7.1, Process 3).

Page 217: The Biogeochemistry of Submerged Soils

Nutrient and Acidity Balances 209

Table 7.1 Acid–base changes in nitrogen transformations in ricefields

Process�[H+]

�[N]

1. Hydrolysis of urea fertilizerCO(NH2)2 + 2H+ + H2O → 2 NH4

+ + CO2 −1

2. NH3 volatilizationNH4

+ → NH3 + H+ +1

3. NitrificationNH4

+ + 2CO2 + H2O → NO3− + 2CH2O + 2H+ +2

4. Denitrification4NO3

− + 5CH2O + 4H+ → 2N2 + 5CO2 + 7H2O −1

5. Removal of NH4+ in cropa

426CO2 + 12 NH4+ + H2PO4

− + 408H2O →C426H855O426N12P + 414O2 + 11H+ +0.92

a Based on data of Dobermann and Fairhurst (2000) for mean mineral content of grain (forclarity, K+, Ca2+, Mg2+, SO4

2−, Cl−, etc. omitted).

Nitrification of the NH4+ does not cause net acidification, whether the NO3

−is absorbed by the crop or denitrified. Two mol of H+ are formed per mol ofNH4

+ nitrified, resulting in a net addition of one H+ per urea-N hydrolysed. Ifthe NO3

− is absorbed by the crop, 1 mol of OH− is exported from the roots permol of NO3

− absorbed, or if it is denitrified, 1 mol H+ is formed per mol of Ndenitrified (Table 7.1, Process 4).

Note that the anaerobic processes causing pH changes following submer-gence—reduction of ferric iron tending to increase the pH of acid soils andaccumulation of CO2 tending to decrease the pH of alkaline soils—are reversedupon drainage and reoxidation of the soil. Thus, unless there has been substantialmovement of acid or base out of the soil during submergence, as generally therewill not have been, the pH changes are reversed. Permanent changes of pH onlyoccur if the concentration of acid or base in the water entering the ricefield differsfrom that leaving it.

Ferrolysis

An exception to these general rules is the formation of so-called ferrolysedsoils (Brinkman, 1970; van Breemen and Buurman, 1998). These occur underparticular hydrological and geological conditions in which there is prolongedseasonal waterlogging of the soil and subsequent drainage as the regional watertable falls. Dissolved Fe2+ formed in soil reduction displaces cations from thesoil exchange complex, and these are subsequently removed in leaching or runoff.When the soil dries, adsorbed Fe2+ is oxidized producing acidity which reactswith the soil. The resulting H+-saturated clay is unstable and partly decomposesto give Al-saturated clay and silica. Repeated cycles of this process can lead to

Page 218: The Biogeochemistry of Submerged Soils

210 Nutrients, Toxins and Pollutants

near complete removal of weatherable minerals. The effects are most marked insoils derived from acidic rocks or strongly weathered sediments. Large wetlandareas in south and south-east Asia have soils of this sort (Brammer and Brinkman,1977; Brinkman, 1977a, b), but doubt has recently been cast on the importanceof ferrolysis in European soils previously thought to have been formed by it (vanRanst and De Coninck, 2002).

7.1.3 PEAT BOGS

At the opposite end of the fertility scale from ricefields are peat bogs in pluviallandscapes. Nutrient inputs come almost entirely from rainfall, and the nutrientreserves in the organic matter buffering the soil solution are small (Moore andBellamy, 1974). The chemistry of peat bogs is therefore precarious and changesin the composition of the rainfall can have a large effect on the composition ofthe soil solution.

Bogs are naturally acid. This is an inevitable consequence of their development.The principal source of acidity is the intake of mineral nutrients by the vegetation.Because the main forms of N absorbed under the anoxic conditions are NH4

+ orN2 fixed from the air, the plants absorb more cations than anions and consequentlyexport H+ from their roots to maintain electrical neutrality. At steady state theinput of organic matter from primary production at the bog surface is balanced byloss of organic matter by decomposition throughout the profile (Clymo, 1984).But because by definition bogs largely comprise undecomposed plant material,at steady state there is a substantial and effectively permanent accumulationof alkalinity in the organic matter and of acidity in the soil. The process is selfreinforcing: the greater the acidity that develops, the more the CEC of the organicand mineral matter in the peat is dominated by H+ and the weaker its abilityto retain nutrient cations. Deposition of nitric or sulfuric acids in rainfall willadd to the acidity. Subsequent denitrification and sulfate reduction generate anequivalent amount of base, so this acidity is neutralized. The acidity of wetlandpeats is discussed by Ross (1995).

An example of the fragile nutrient balance in peat bogs is given in Table 7.2 fora blanket bog in northern England. The table shows the losses of N through ero-sion of the peat, and the losses of nutrient cations through leaching and removalin stream water.

7.1.4 RIPARIAN WETLANDS

Riparian wetlands are those lands that are periodically inundated with waterfrom adjacent rivers, streams, lakes or other freshwater bodies, and by runofffrom upland areas. Large fluxes of energy and nutrients pass through riparianwetlands and they are important sinks and transformers of nutrients. In water-sheds with extensive riparian wetlands, the composition of the river water may

Page 219: The Biogeochemistry of Submerged Soils

Nutrient and Acidity Balances 211

Table 7.2 Nutrient balance for blanket peat bog in the Pennines, north-ern England

Amount (kg ha−1 year−1)

Na K Ca P N

InputPrecipitation 25.5 3.1 9.0 0.69 8.2OutputSale of sheep 0.0 0.0 0.0 0.01 0.1Dissolved in stream 45.3 9.0 53.8 0.39 2.9Peat erosion in stream 0.3 2.1 4.8 0.45 14.6Net loss 20.0 8.0 49.7 0.15 9.5

Source: data from Crisp (1966).

be dominated by transformations in the wetlands (Olson, 1992; Mitsch and Gos-selink, 2000). There are large seasonal variations depending on the water flow andthe state of the riparian vegetation. But in general inorganic forms of nutrientsare transformed to organic forms and nitrate is denitrified. These transforma-tions and losses have important consequences for the productivity of aquaticsystems downstream.

The terms in the nutrient balance of a riparian wetland are essentially the sameas those in a traditional wetland ricefield in a river floodplain or delta, thoughof course the magnitudes differ. Inputs are delivered along the stream course asdissolved material and sediment, and in lateral runoff from neighbouring upland;dissolved and particulate material is filtered, absorbed, adsorbed and variouslytransformed in the wetland, and flows out in runoff and percolation.

As for ricefields, additions from sediment vary very widely. The amount carriedand its composition will depend on the landscape through which the river hasflown and its soils and geology, and rainfall characteristics. The amount depositedwill in turn depend on the local landscape and conditions, including the nature ofthe vegetation filtering and trapping the sediment. Table 7.3 compares additionsof P in sediment in riparian wetlands in North America with those estimated forricefields in Asia. The generally greater additions in North American sedimentsand the very large variations are apparent.

Riparian wetlands are effective though not infinite sinks for nitrate and phos-phate from agricultural runoff. Strips of wetland a few tens of metres wide havebeen shown to remove the bulk of nitrate and phosphate entering in runoff andgroundwater, though the limits to this under different circumstances are not wellquantified (Baker and Maltby, 1995; Mitsch and Gosselink, 2000).

7.1.5 TIDAL WETLANDS

The balance in tidal wetlands is complicated by the tidal inflow and outflow ofwater across the submerged sediments and the greater influence of subsurfaceleaching under the large tidal head of water.

Page 220: The Biogeochemistry of Submerged Soils

212 Nutrients, Toxins and Pollutants

Table 7.3 Additions of P in sediments in ricefields in Asia and riparian wetlandsin North America

Rate of deposition(g m−2 year−1)

Ricefields in AsiaCalculated from contents of tropical soils and assumed

high sedimentation rate (1 kg m−2 year−1) 0.4–0.7Calculated from measured sediment contents and

assumed high sedimentation rate 0.4–1.1Measured additions

Guangdong, China 0.06Bangladesh deepwater sites 0.2Mekong delta 0.1

Riparian wetlands in North AmericaSouthern Illinois 3.6Central Florida 3.25North Carolina 0.17North Carolina from input–output balance 0.32–0.73Northwestern Illinois 1.36

Sources : ricefields, Greenland (1997); others, Mitsch and Gosselink (2000).

Nutrient balances in tidal wetland systems have been studied at length andthe picture is variable. In general the net exchanges are small in relation to theoverall nutrient budgets, though nutrients may be transformed between dissolvedand particulate inorganic and organic forms and oxidation states (Nixon, 1980;Childers et al., 2000; Mitsch and Gosselink, 2000). Nitrogen generally flows intothe marsh largely as nitrate but is exported in dissolved and particulate reducedforms, and is denitrified. Table 7.4 shows a representative N budget, obtainedin a large tidal salt marsh in Massachusetts. Though surface water inflows werenot measured, there is a rough balance between inflows and outflows. The tidalexchange is far greater than any of the other components. Tidal marshes tend tobe net sinks for total phosphorus, entering in estuarine water in dissolved organicand inorganic forms but there may be a remobilization of inorganic phosphateleached out of sediments by saline water.

7.2 TOXINS

7.2.1 ACIDITY

In general high acidity and resulting high concentrations of toxic aluminium insolution do not occur in submerged soils because the electrochemical changesaccompanying submergence tend to neutralize acidity present in the unfloodedsoil. Acid sulfate soils are a notable exception.

Page 221: The Biogeochemistry of Submerged Soils

Toxins 213

Table 7.4 Nitrogen budget of a tidal salt marshin Massachusetts

Flow(g N m−2 year−1)

InputsSurface water ?Rainfall 0.8Biological N fixation 6.5Groundwater 27.0Tidal inflow 116.0

150.3OutputsDenitrification 13.3Tidal outflow 140.0

153.3

Source: data from Mitsch and Gosselink (2000).

Acid Sulfate Soils

Acid sulfate soils are an especially difficult class of acid soil formed in formermarine sediments that have been drained. The acidity is generated from the oxi-dation of pyrite in the soil resulting in acute aluminium toxicity, iron toxicity,and deficiencies of most nutrients, especially phosphate which becomes immobi-lized in ferric oxide. The development and management of acid sulfate soils arediscussed in detail in Dost and van Breemen (1983) and Dent (1986).

In brief, the steps in the formation of pyrite in marine sediments are:

(1) reduction of Fe(III) in the sediment to soluble Fe2+ and reduction of SO42−

from seawater to S2−;(2) partial oxidation of S2− to elemental S or polysulfide, S2

2−;(3) formation of pyrite, FeS2, either directly from Fe2+ and S2

2− or via FeSformed from Fe2+ and S2− and subsequent reaction with S.

The overall reaction is

2Fe2O3 + 8SO42− + 16CH2O + O2 → 4FeS2 + 16HCO3

− + 8H2O (7.2)

The necessary conditions are sources of iron oxide, dissolved SO42− and organic

matter, and sufficiently reducing conditions for reduction of SO42− coupled to

intermittent or localized oxidizing conditions to produce elemental S or poly-sulfide. Potential acidity develops by the removal of alkalinity (represented byHCO3

− in Equation 7.2) from the sediment by diffusion and tidal action. What

Page 222: The Biogeochemistry of Submerged Soils

214 Nutrients, Toxins and Pollutants

little is known about rates of pyrite formation under natural conditions indicatesrates of a few kg S m−3 of sediment in 100 years under mangrove vegetation instationary or slowly aggrading coastal plains.

Potential acid sulfate soils ripen into actual acid sulfate soils as a result ofdrainage and oxidation of the pyrite, forming sulfuric acid. The reaction betweenpyrite and oxygen is slow, but oxidation of FeS2 by Fe(III) in solution is fastproducing Fe(II). This process is catalysed at low pH by the bacterium Thiobacil-lus ferrooxidans which mediates the oxidation of sulfur species and Fe(II), soregenerating Fe(III) and facilitating further FeS2 oxidation. The process requiresacid conditions because Fe(III) is insufficiently soluble at pH greater than about 4and because the growth of T. ferrooxidans is inhibited at higher pH. The overallreaction is

4FeS2 + 15O2 + 14H2O → 4Fe(OH)3 + 8SO42− + 16H+ (7.3)

Most of the Fe(III) eventually crystallizes as reddish-brown ferric oxide in mot-tles, coatings and nodules. Under strongly oxidizing severely acid conditions,pale yellow coatings of the mineral jarosite, KFe3(SO4)2(OH)6, may form onped faces. At higher pH, jarosite is hydrolysed to goethite. Hence ripe acidsulfate soils often have a layer of yellow jaorosite mottling adjacent overlyinga still-reduced pyrite layer but overlain by layers from which acidity has beenleached, and hence dominated by reddish-brown goethite. These features are usedto assess the ripening of the soil.

7.2.2 IRON TOXICITY

Iron toxicity is a syndrome of disorders associated with large concentrations ofFe2+ in the soil solution. It is only found in flooded soils. A wide range of con-centrations produce the symptoms, from 1000 to only 10 mg L−1 in soils withpoor nutrient status—especially of P or K—or with respiration inhibitors such asH2S. There are large differences in tolerance between rice varieties. The effectsinclude internal damage of tissues due to excessive uptake of Fe2+; impairednutrient uptake, especially of P, K, Ca and Mg; and increased diseases associatedwith imbalanced nutrition, such as brown leaf spot (caused by Helminthospo-rium oryzae), sheath blight (caused by Rhizoctonia solani ) and blast (caused byPyricularia oryzae).

The circumstances of the toxicity are quite well established, though some ofthe details of the mechanisms involved are uncertain. Three main groups of Fetoxic soils are distinguished:

• acid sulfate soils, in which extremely large concentrations of Fe2+ in the soilsolution arise as a result of the soils’ peculiar mineralogy;

• poorly drained sandy soils in valleys receiving interflow water from adjacenthigher land with highly weathered sediments; and

Page 223: The Biogeochemistry of Submerged Soils

Toxins 215

• more clayey, acid, iron-rich soils in sediments derived from highly weatheredsoils and which give iron toxicity without interflow.

Where Fe toxicity is associated with interflow the concentrations of dissolvedFe in the upwelling water have been found to be too small to account for thelarge concentrations of Fe2+ in the root zone, and most of the Fe2+ is apparentlyformed in situ. Therefore the interflow aggravates toxicity by some mechanismother than bringing in Fe2+, possibly involving depletion of other nutrients andupsetting the plant’s ability to exclude Fe (Section 6.5).

It is often a symptom of imbalanced nutrition rather than high Fe2+ in the soilsolution per se. Thus the soil solution Fe2+ concentrations at which it is reportedvary from 10 to 1000 mg L−1, and it is more often associated with low levels ofP, K, Ca and Mg and impaired ability of roots to exclude Fe2+.

7.2.3 ORGANIC ACIDS

Decomposition of organic matter in submerged soils produces phytotoxic com-pounds such as aliphatic and phenolic acids (Takajima, 1964; Tsutsuki and Pon-namperuma, 1987). Many of the potentially phytotoxic compounds are producedonly transiently, so it is difficult to assess the damage they cause. The most cer-tain effects are seen in rice soils to which large quantities of organic manuresare added, especially under temperate conditions. As a result, in ricefields inJapan and China, where organic manures are widely used, to avoid accumula-tion of phytotoxins in the root zone, greater percolation rates are maintained—5to 10 mm day−1 compared with 1 to 5 mm day−1 elsewhere—and water is oftencompletely drained from the fields midseason (Figure 7.4). This is less critical intropical countries because higher temperatures allow the toxins to be decomposedmore rapidly, and the use of organic manures is less.

90%8070

12345 cm (Submergence) (Intermittent

drainage)

(Midsummer drainage)

(Intermittent irrigation)

Soilsurface

Sept.AugustJulyJuneMay

60

Soi

lm

oist

ure

Wat

erde

pth

50

Figure 7.4 Water management in ricefields in Japan and parts of China (modifiedfrom Yukawa, 1989). Reproduced by permission of the Japanese Society of Irrigation,Drainage and Reclamation Engineering

Page 224: The Biogeochemistry of Submerged Soils

216 Nutrients, Toxins and Pollutants

Research on the effects of phenolic acids on rice has measured effects onseed germination and rates of root elongation and seedling growth. This hasshown that concentrations greater than a few mM of the more noxious acids arerequired to substantially impair root elongation (Olofsdotter et al., 2002). Undertropical conditions, without large additions of organic matter, total concentrationsof alkali-soluble phenolic acids in the soil during rice cropping are generallyless than this (Tsutsuki and Ponnamperuma, 1987), though concentrations maybe greater locally, such as in the rhizosphere. Factors that exacerbate the toxiceffects of phenolic acids include low plant N status (Vaughan and Ord, 1990); lowsoil pH, the undissociated acid having greater membrane permeability (Tanakaand Navaesero, 1967); and the nature of the acid, cinnamic acid derivatives beingmore inhibitory than benzoic acid derivatives, and lipid-soluble acids being moreinhibitory than lipid-insoluble (Glass, 1973). Little research has been done on theeffects of phenolic acids on ion uptake by rice roots, but ion uptake is presumablyimpaired at much smaller concentrations than root elongation.

Recent experiments at the International Rice Research Institute (IRRI) on themaximum yields of high-yielding rice cultivars developed over the last 40 yearshave shown that the older cultivars no longer perform as well as they usedto (Peng et al., 2000). The first modern high-yielding cultivar released by IRRI,IR8, which often produced 9–10 Mg of grain ha−1 at the IRRI farm in the1960s under optimal management, now yields only 7–8 Mg ha−1 under sim-ilar conditions, whereas its most recent successors yield 9–10 Mg ha−1. Thisdoes not appear to be due to deterioration in seed stocks over the years throughrepeated multiplication cycles (Peng, unpublished), or to new or increased pestsor diseases. It appears that new abiotic stresses have arisen over the 40 yearsof continuous rice cropping on the IRRI farm, with two or three crops per yearin flooded soil, and the more recent cultivars may be better adapted to thesestresses. Changes in soil conditions have been observed, most notably changesin the nature of the soil organic matter associated with prolonged soil flood-ing and more strongly reducing conditions in the soil, particularly increasedconcentrations of phenolic compounds (Olk and Senesi, 2000). The decrease ingrain yield of the older cultivars is associated with poor grain filling and harvestindex, and impaired acquisition of soil nitrogen. A possible explanation is thatthe changes in soil conditions have led to impaired root function in the older cul-tivars through toxins associated with prolonged soil flooding, such as phenolicor other organic acids.

7.2.4 SALINITY

Salinity occurs in coastal areas affected by seawater, in areas receiving saltywater by lateral flow from salt-bearing rocks upstream, and also in otherwisenon-saline environments as a result of soil waterlogging, mainly due to highground water. In waterlogged land, whatever dissolved salt is brought in with the

Page 225: The Biogeochemistry of Submerged Soils

Toxins 217

water necessarily accumulates. Even good quality water contains 200 mg L−1 ofsoluble salts, so, for example, a ricefield receiving 1000 mm of irrigation waterwill accumulate 2 t ha−1 of salt per year (Greenland, 1997). Also, salts in subsoilsand groundwater in waterlogged land may be brought into the surface by massflow and diffusion.

In land that is flooded for part of the year but drains naturally after the flood-water recedes, accumulated salt is removed with the draining water and there isa natural renewal of the land. Percolation and lateral drainage at the start of thefollowing rainy season but before the land is re-flooded also wash out accumu-lated salt. However this natural recovery is prevented if the water table remainsabove or close to the surface. This may happen in depressions, but also wherelarge reservoirs have been established at a higher elevation than the flood-proneland, or where an unlined canal that carries a large volume of water has beenbuilt on permeable soil (Greenland, 1997). Such problems are more common inarid or semi-arid areas where there is both less leaching of the soil and ground-and irrigation-waters are more likely to be saline.

A further problem associated with waterlogging and salinity is sodicity. Whenthe quantity of Na+ in the soil exceeds about 15 % of the CEC, the soil maybecome dispersed, i.e. aggregation is lost, and it will then dry to large tough clods.Salts in groundwater are often high in Na+, so sodicity may be a problem eventhough salinity is not. Sodic soils are not necessarily problematic for wetland ricecultivation, though rates of percolation may be sub-optimal but they are difficultto cultivate for following dryland crops.

About 5 Mha of tidal wetlands are cultivated with rice (IRRI, 2002). Althoughthis is only a small part of the 147 Mha of land currently cropped with riceglobally, there is some scope for expanding the area in response to increasingdemand for rice and loss of more favourable land to non-rice uses (Greenland,1997). There is also a need to improve the productivity of existing tidal riceareas to relieve pressure to clear marginal lands. The principal soil chemicalstress in tidal wetlands is high salinity. Rice is only moderately tolerant of saltand is more sensitive to it than some cereals (Flowers and Yeo, 1981; Yoshidaet al., 1983). The sensitivity varies over the growth cycle being most acute in theseedling stage and again during flowering. Although there are large differencesin tolerance between cultivars, none are tolerant of salt throughout the growingseason (Flowers et al., 2000; Gregorio et al., 2002). The suitability of rice forcoastal saline areas therefore arises from its tolerance of soil submergence ratherthan exceptional salt tolerance: because it can tolerate soil submergence, andbecause flooding with less saline water dilutes the salt in the soil, rice will growon land not able to support dryland crops. In addition, various other soil chemicalstresses are prevalent. These include alkalinity, acidity, Fe toxicity, and deficien-cies of P, Zn and other nutrients, typically in combination. Quijano-Guerta andKirk (2002) discuss the tolerance of rice germplasm to the multiple soil chemicalstresses in tidal wetlands.

Page 226: The Biogeochemistry of Submerged Soils

218 Nutrients, Toxins and Pollutants

7.3 TRACE ELEMENTS

7.3.1 GLOBAL CYCLING OF TRACE ELEMENTS

Wetlands are important in the cycling of trace metals and metalloids, firstlybecause they are often in high rainfall areas and receive correspondingly highadditions by wet deposition; secondly because fluxial and phreatic wetlandsreceive large volumes of surface and ground water bearing dissolved and par-ticulate trace elements; and thirdly because of their particular biogeochemistrywhich results in transformations and accumulation of trace elements.

Concentrations of trace elements in surface and ground waters are controlledby deposition from the atmosphere and dissolution from soils and bedrock.Concentrations in the atmosphere arise from anthropogenic sources—fossil fuelcombustion, cement production, extractive metallurgy—as well as through nat-ural processes—windborne soil, volcanic ejecta, forest fires, biogenic processes.Depending on the metal, transport through the atmosphere and subsequent wetor dry deposition may exceed transport through surface and ground water. Ele-ments for which atmospheric transport predominates are termed atmophile andthose for which transport by water predominates are termed lithophile (Stummand Morgan, 1996). Many atmophile metals are volatile or can become volatilethrough methylation, especially the B-type metals (Section 3.1) Hg, As and Pb.By contrast the A-type metals—Mn, Co, Cr, V and Ni—are lithophile. Hence B-type metals tend to be enriched in the environment through diffuse atmosphericpollution (Table 7.5). They also tend to be pernicious toxins because of theirtendency to react with soft bases, such as –SH and –NH groups in enzymes.

7.3.2 TRANSPORT OF TRACE ELEMENTS THROUGH SOIL AND INTOPLANT ROOTS

The factors controlling the transport through soil include:

• the concentration of the free ion in solution;• complexation with organic and inorganic ligands in solution;• redox reactions;• sorption on organic matter, clay minerals and oxides in the soil solid in outer-

sphere and highly insoluble inner-sphere complexes;• precipitation and co-precipitation in insoluble compounds, particularly hydrox-

ides, carbonates, sulfides or phosphates;• sorption, precipitation and co-precipitation in suspended colloids;• conversion to volatile forms.

Because the importance of these factors differs between the different traceelements, predicting mobilities is complicated. The tendency to form organic

Page 227: The Biogeochemistry of Submerged Soils

Trace Elements 219

Table 7.5 Global emissions of trace metals to the atmosphere and concentrationsin freshwater

Emissions (kt year−1)a,b Concentration infreshwater (µ g L−1)c

Natural Anthropogenic Total Mean (range)

Antimony 2.6 3.5 6.1 0.2 (0.01–5)Arsenic 12 19 31 0.5 (0.2–230)Cadmium 1.4 7.6 9.0 0.1 (0.01–3)Chromium 43 31 74 1 (0.1–6)Cobalt 6.1 35 41 0.2 (0.04–8)Copper 28 332 360 3 (0.2–30)Lead 12 38 50 3 (0.06–120)Manganese 317 3.6 320 8 (0.02–130)Mercury 2.5 52 54 0.1 (0.0001–2.8)Molybdenum 3.0 6.3 9.3 0.5 (0.03–10)Nickel 29 5.1 34 0.5 (0.02–27)Selenium 10 5.1 15 0.2 (0.02–1)Vanadium 28 86 114 0.5 (0.01–20)Zinc 45 132 177 15 (0.2–100)

Sources :a Nriagu (1989).b Nriagu and Pacyna (1988).c Bowen (1979).

complexes increases in the order (Chapter 3):

Cd < Zn < Co � Ni = Cu < Hg

and the tendency to form strongly sorbed, inner-sphere complexes with oxidesand clays increases in the order:

Cd < Ni < Co < Zn � Cu < Hg

The tendency to co-precipitate in secondary minerals also differs. Typical co-precipitates are (Sposito, 1983):

Fe oxides V, Mn, Ni, Cu, Zn, MoMn oxides Fe, Co, Ni, Zn, PbCa carbonates V, Mn, Fe, Co, CdClay minerals Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Pb

These differences are exemplified in Figure 7.5 which shows the results of anexperiment in which Cd2+, Zn2+, Ni2+ and Cu2+ salts were applied on the surfaceof an acid soil, with and without lime, and the soil leached with 0.01 M CaCl2 forseveral hours (McBride, 1994). In the unlimed soil, Cd2+, Zn2+ and Ni2+ movedreadily to depth, but Cu2+ remained near the surface because it was stronglysorbed on soil solids. In the limed soil, with pH 6.5, increased sorption and

Page 228: The Biogeochemistry of Submerged Soils

220 Nutrients, Toxins and Pollutants

Relative concentration of adsorbed metal

0

10

20

Dep

th (

cm)

30

40

unlimed limed

Cu

NiZn

Cd

Cu

Cd

Zn/Ni

Figure 7.5 Profiles of surface-applied metals in acid soil, with and without lime andleached with 0.01 M CaCl2 (McBride, 1994). Reproduced by permission of Oxford Uni-versity Press

precipitation of all four metals resulted in retarded leaching. The attenuation ofCd2+, Zn2+ and Ni2+ leaching could be accounted for with a model allowing forsimple, pH-dependent cation exchange. The results for Cu2+ required allowancefor more-selective chemisorption and chelation reactions with highly nonlinearconcentration dependence.

The bioavailability of trace elements is further complicated by differences inthe factors controlling transport to plant roots. These are:

• desorption or dissolution from the soil solid, which may be slow compared withtransport to roots, and complexation in solution; all of these may be affectedby root-induced changes in the soil, which may both increase and decrease thesolubilities of trace elements;

• diffusion through the soil solution, especially as complexes with carrier ligands;• absorption across the root surface by passive and active transporters, including

for ions complexed with carrier ligands;• translocation from root to shoot: cationic trace elements especially may accu-

mulate on and in roots.

7.3.1 MOBILITIES OF INDIVIDUAL TRACE ELEMENTS

In the following sections the biogeochemistries of important trace metals andmetalloids in submerged soils are discussed. They are important either becauseof their redox chemistries or because they are particularly affected by soil redox

Page 229: The Biogeochemistry of Submerged Soils

Trace Elements 221

conditions. The list is not exhaustive but it serves to illustrate the importantprocesses. The properties of the elements are summarized in Table 7.6 andTable 7.7 gives the important redox equilibria.

Zinc

Zinc occurs in soils exclusively in the +2 oxidation state and in solution asthe B-type cation Zn2+. It is weakly complexed by the main functional groupsin organic matter, but under sulfate-reducing conditions forms insoluble sulfides(ZnS, pK = 24.7). In intermittently submerged soils, such as wetland rice soils,ZnS probably generally does not form because FeS and FeS2 are precipitatedat a higher pe + pH and hence will form preferentially if the redox is poisedby Fe(II) (Sajwan and Lindsay, 1986). Under such conditions Zn2+ forms solidsolutions in oxides and clay minerals (see below). Hence it tends to be highlyimmobile under anaerobic conditions, but under acid oxidizing conditions it isreleased in soluble and mobile forms.

Zinc solubility in soils tends to show a minimum at near neutral pH(Figure 3.14). At low pH the free ions are only weakly sorbed on charged soilsurfaces, but at pH > 7, as concentrations of dissolved organic ligands increase,soluble Zn—organic complexes may form, raising the total concentrations ofZn in solution even though the activity of the free ion may be extremely smallthrough sorption reactions. At high pH Zn2+ forms solid solutions in Ca and Mgcarbonates, and mixed hydroxy-carbonates, so it is immobile and unavailable toplants in alkaline or calcareous soils.

Zinc deficiency is widespread in wetland rice affecting up to 50 % of thearea (Katyal and Vlek, 1985; Welch et al., 1991; Batten et al., 1992; Neue andLantin, 1994). Zinc relations in rice have therefore been studied extensively. Thedeficiency is most often associated with poor drainage and perennial soil wetness.The soils typically have weak profile development, reflecting the poor drainage,and much of the Zn is in primary minerals or in other highly insoluble forms. Itis also often associated with high soil organic matter content, high pH and highMg:Ca ratios in the soil.

All of these factors are present in the toposequence at Tiaong, Quezon Province,Philippines shown in Figure 7.6, which has been used for many years by the IRRIto screen rice for Zn deficiency tolerance (Quijano-Guerta et al., 2002).

The toposequence is on the gently sloping foot-slope of a young inactive vol-cano, Mt Banahaw. The drainage is poor across the toposequence due to perennialupwelling of artesian water. The extent of Zn deficiency increases down the slope,as do soil wetness, organic C content, CaCO3 content and CEC. The pH, claycontent and extractable Zn are uniform. The upwelling water contains high con-centrations of Ca2+, Mg2+, HCO3

− and H4SiO4 of volcanic origin. As it reachesthe surface, CO2 degasses causing the pH to rise. This results in precipitationof CaCO3 and possibly de novo synthesis of Mg smectities, on and in which

Page 230: The Biogeochemistry of Submerged Soils

222

Tabl

e7.

6Pr

oper

ties

oftr

ace

met

als

and

met

allo

ids

impo

rtan

tin

subm

erge

dso

ils

23V

24C

r25

Mn

26F

e27

Co

28N

i29

Cu

30Z

n

VA

VIA

VIIA

VIII

AV

IIIA

VIII

AIB

IIB

IIIB

IVB

VB

VIB

[Ar]

3d3 4s

2

1.63

4

,5[A

r]3d

5 4s1

1.66

3

,6[A

r]3d

5 4s2

1.55

2

,3,4

[Ar]

3d6 4s

2

1.83

2

,3[A

r]3d

7 4s2

1.88

2

(,3)

[Ar]

3d8 4s

2

1.91

2

(,3)

[Ar]

3d10

4s1

1.90

(

1,)2

[Ar]

3d10

4s2

1.65

2

33A

s[A

r]3d

104s

2 p3

2.18

3,5

34S

e[A

r]3d

104s

2 p4

2.55

−2,4

,6

48C

d[K

r]4d

105s

2

1.69

2

51S

b[K

r]4d

105s

2 p3

2.05

3,5

80H

g[X

e]4f

145d

106s

2

2.00

(1,

)2

81T

I[X

e]4f

145d

106s

2 p1

2.04

1,

3

5 B[H

e]2s

2 p1

2.04

3

82P

b[X

e]4f

145d

106s

2 p2

2.33

2(

,4)

Gro

ups

clas

sifie

dby

orig

inal

IUPA

Csy

stem

.Fo

rea

chel

emen

t:se

cond

row

give

sel

ectr

onco

nfigu

ratio

n,th

ird

give

sel

ectr

oneg

ativ

ityan

dim

port

ant

oxid

atio

nst

ates

.

Page 231: The Biogeochemistry of Submerged Soils

Trace Elements 223

Table 7.7 Equilibrium constants of reduction half-reactions of trace elements in sub-merged soils compared with those for Fe and Mn

pe0 pe0∗

pH 5 pH 7

V

VO2+ + 2H+ + e− = VO2+ + H2O 16.9 6.9 2.9

VO2+ + 2H+ + e− = V3+ + H2O 5.7 −4.3 −8.3

(note also VO2+ + 2H2O = VO(OH)3 + H+, pK = 3.3; VO(OH)3 = VO2(OH)2

−+H+, pK = 4.0; VO2(OH)2

− = VO3(OH)2− + H+, pK = 8.55)

Cr13 HCrO4

− + 4H+ + e− = 13 Cr(OH)3 + 1

3 H2O 18.9 10.6 7.9

(note also HCrO4− = H+ + CrO4

2−, pK = 6.5)

Mn12 Mn3O4(s) + 4H+ + e− = 3

2 Mn2+ + 2H2O 30.8 16.3 8.3

MnOOH(s) + 3H+ + e− = Mn2+ + 2H2O 25.3 14.0 8.012 MnO2(s) + 2H+ + e− = 1

2 Mn2+ + 2H2O 21.8 13.7 9.7

Fe

Fe(OH)3(s) + 3H+ + e− = Fe2+ + 3H2O 16.5 4.5 −1.5

α-FeOOH(s) + 3H+ + e− = Fe2+ + 2H2O 11.3 −0.7 −6.7

Co

Co3+ + e− = Co2+ 30.6 30.6 30.6

Ni12 NiO2(s) + 2H+ + e− = 1

2 Ni2+ + H2O 29.8 22.3 18.3

Cu

Cu2+ + e− = Cu+ 2.6 2.6 2.6

Hg

Hg2+ + e− = 12 Hg2

2+ 15.4 15.4 15.412 Hg2+ + e− = 1

2 Hg 14.4 14.4 14.4

Tl

Tl3+ + e− = Tl2+ 21.3 21.3 21.3

Pb12 PbO2(s) + 2H+ + e− = 1

2 Pb2+ + H2O 24.8 17.3 13.3

As

(continued overleaf )

Page 232: The Biogeochemistry of Submerged Soils

224 Nutrients, Toxins and Pollutants

Table 7.7 (continued )

pe0 pe0∗

pH 5 pH 7

12 HAsO4

2− + 2H+ + e− = 12 H3AsO3 + 1

2 H2O 14.9 4.9 0.9

(note also H3AsO4 = H2AsO4− + H+, pK = 2.24; H2AsO4

− = HAsO42− + H+,

pK = 6.94; HAsO42− = AsO4

3− + H+, pK = 11.5; H3AsO3 = H2AsO3− + H+,

pK = 9.29)

Sb12 SbO3

− + 32 H+ + e− = 1

2 Sb(OH)3 11.3 3.8 0.8

(note also Sb(OH)3 + H+ = Sb(OH)2+ + H2O, pK = −1.42; Sb(OH)3 + H2O =

Sb(OH)4− + H+, pK = 11.82)

Se12 SeO4

2− + H+ + e− = 12 SeO3

2− + 12 H2O 14.9 9.9 7.9

14 SeO3

2− + 32 H+ + e− = 1

4 Se(s) + 34 H2O 14.8 6.0 3.0

16 SeO3

2− + H+ + e− = 16 Se2− + 1

2 H2O 5.3 0.3 -1.7

(note also H2SeO3 = HSeO3− + H+, pK = 2.4; HSeO3

− = SeO32− + H+, pK = 7.9;

SeO42− + H+ = HSeO4

−, pK = −1.7)

I

IO3− + 6H+ + 6e− = I− + 3H2O 18.3 13.3 11.3

12 I2(aq) + e− = I− 10.5 10.5 10.5

Sources : pe0 values calculated with Equation (4.8) using �G0f values from Garrels and Christ (1965). pe0∗

values calculated with Equation (4.22) for conditions in submerged soil solutions: for trace element ions, (ion) =10 µM, (Mn2+) = 0.2 mM, (Fe2+) = 1 mM. Constants for hydrolysis equilibria from Baes and Maesmer (1976).

Zn2+ becomes strongly immobilized as solid solutions (van Breemen et al., 1980;Scharpenseel et al., 1983). The formation of [Ca,Mg,Zn]CO3 solid solutions insubmerged rice soils would explain the association between Zn deficiency andsoils with high Mg:Ca ratios.

Cadmium

Like Zn, Cd is a Group IIB element and occurs in soils exclusively in the +2oxidation state as the Cd2+ cation. Cadmium and zinc are often co-precipitatedwith each other in sulfide minerals in rocks (pKCdS = 27.0). Hence Cd tendsto be highly immobile under anaerobic sulfate-reducing conditions, but underacid, oxidizing conditions it is released in soluble and mobile forms. Hence soils

Page 233: The Biogeochemistry of Submerged Soils

225

100

200

300

400

m

Zn

defic

ienc

y(s

tron

g, w

eak,

non

e)N

o ric

e gr

own

(shr

ubs

and

gras

ses)

Ext

ent o

f Zn

defic

ienc

y

CaC

O3

cont

ent (

%)

Org

anic

C c

onte

nt (

%)

SW

NEm above sea level

Zn

defic

ienc

y(w

eak,

mod

erat

e, s

tron

g, v

ery

stro

ng)

Exp

erim

enta

l fie

lds

Art

esia

n w

ell

0

<0.5

<0.5

<1

<1

3.5

−4

3−4

4−5

3−4

2−3

1−2

Dep

th o

f obs

erva

tion

2.5

−3.5

1.5

−2.5

0.5

−1.5

45 44 43 42 41 45 44 43 42 41 45 44 43 42 41

Fig

ure

7.6

Topo

sequ

ence

atT

iaon

g,Q

uezo

nPr

ovin

ce,

Phili

ppin

es,

show

ing

extr

eme

Zn

defic

ienc

yin

rice

(van

Bre

emen

etal

.,19

80).

Rep

rodu

ced

bype

rmis

sion

ofK

luw

erA

cade

mic

Publ

ishe

rs

Page 234: The Biogeochemistry of Submerged Soils

226 Nutrients, Toxins and Pollutants

that contain toxic concentrations of Cd when aerobic may be entirely suitable forwetland rice cultivation (Takijima and Katsumi, 1973; Bingham et al., 1976). TheCd content of rice grain in Cd-contaminated soil has been found to be correlatedwith the number of days the soil is drained prior to harvest (Page et al., 1981).

Cobalt, Nickel and Copper

Cobalt and nickel are Group VIIIA and copper Group IB elements. They occurpredominantly in the +2 oxidation state in soils as divalent cations, though Co2+may be oxidized to Co3+ forming very insoluble compounds with Mn oxides, andCu2+ may be reduced to Cu+, especially if soft bases such as halides and S2− arepresent to stabilize the Cu+ ion. All are chalcophiles and tend to form insolublesulfides in anaerobic conditions (pKs = 21.3–25.6, 19.4–26.6 and 36.1, respec-tively). They therefore tend to have low mobilities in submerged soils, especiallyCu2+, and accumulate.

All are strongly sorbed on soil surfaces, increasingly as the pH increases, andare more strongly bound to functional groups in organic matter than Cd2+ orZn2+. They therefore tend to show the minimum in solubility at near neutralpH discussed for Zn. Ni2+ and Cu2+ form highly stable complexes with organicmatter, especially with ligands containing N and S, and they therefore tend toaccumulate in organic soils.

Mercury

Mercury occurs in soils predominantly in the +2 oxidation state. Elemental Hgin the atmosphere is oxidized to Hg2+ and deposited in rainfall. It is a strongchalcophile and under anaerobic conditions forms the extremely insoluble sul-fide cinnabar (HgS, pK = 52.7). Nonetheless it is not entirely immobilized underanaerobic conditions because it is reduced to volatile Hg0 or methylated to volatilemethyl mercury compounds by microbial action, and so returned to the atmo-sphere. The methylation is mediated by various bacteria, especially methanogens,through the reactions:

Hg2+ methylation−−−−−→ CH3Hg+

CH3Hg+ methylation−−−−−→ (CH3)2Hg ↑

Other volatile methyl mercury compounds, such as (C6H5)2Hg, are also formed.The CH3Hg+ unit is very inert with respect to decomposition. Therefore, onceformed, methyl mercury compounds are not readily demethylated. The bio-geochemistry of Hg in the environment is reviewed by Ridley et al. (1977)and Mason et al. (1993).

Page 235: The Biogeochemistry of Submerged Soils

Trace Elements 227

The chemistry of Hg in aerobic soils is also complicated and so it is difficult tomake general predictions about its mobility. The Hg2+ cation is B-type and formsstrong bonds with soft ligands such as the sulfhydryl group (–SH) and S2− anion,but not with the main functional groups in organic matter (Section 3.1). In aerobicsoil it is therefore likely to be immobile at trace concentrations but moderatelymobile at greater concentrations (McBride, 1994). However the mobility alsodepends on pH. At pH > 4, the predominant form in solution is Hg(OH)2

0,which is little sorbed on soil surfaces, but at pH > 7 Hg2+ is precipitated asHg(OH)2 and HgCO3.

Concentrations of Hg in the global atmosphere (Slemr et al., 1985; Slemr andLanger, 1992; Mason et al., 1994) and deposited in ice and lake sediments (Weisset al., 1971; Swain et al., 1992) are increasing, probably due to industrial activity.Accumulations of Hg in soils and sediments tend to correlate with soil organicmatter content, and the greatest natural accumulations are in peaty and submergedsoils. Though submerged soils are sinks for Hg as HgS, they are also the mainsource of methyl mercury in the environment (St Louis et al., 1994).

Vanadium

Vanadium occurs in soils predominantly as the +5 vanadate species (VO(OH)30,

VO2(OH)2− and VO3(OH)2−) and under reducing conditions as the +4 vanadyl

cation (VO2+). Less commonly V3+ may also form and substitute for Fe3+in minerals. Interchange between these oxidation states with redox conditionsgreatly alters the solubility of V in soils.

The vanadate equilibria are given in Table 7.7. The VO2(OH)2− and

VO3(OH)2− anions are sorbed on positively charged sites on oxides and sili-cates at low pH, but sorption decreases with pH as the surface positive chargedecreases. Consequently V is quite soluble at high pH and less soluble at low pH.

Reduced V(IV) is much less soluble. The VO2+ cation behaves like Cu2+and forms strong complexes with organic ligands and is chemisorbed on oxidesand silicate clays. The V4+ ion is isomorphously substituted for Si4+ and Al3+in kaolinite (Gehring et al., 1993). Hence the mobility of V under reducing oracid conditions is expected to be low. Reduction of VO2

+ to VO2+ occurs atpe0∗ = 6.9 pH 5 and 2.9 at pH 7 (Table 7.7) and so requires only weakly reduc-ing conditions. A wide range of heterotrophic bacteria and fungi is capable ofreducing V(V) (Bautista and Alexander, 1972).

Chromium

Chromium occurs in soils predominantly as the immobile +3 chromiccation (Cr3+), but may be oxidized to or added as +6 chromate species(CrO4

2−, HCrO4−). Chromate is weakly sorbed on soils and is highly toxic to

Page 236: The Biogeochemistry of Submerged Soils

228 Nutrients, Toxins and Pollutants

plants and animals, whereas Cr(III) is more strongly sorbed, forms complexeswith organic matter and precipitates as insoluble Cr(OH)3 at high pH, and is farless toxic. Oxidation of Cr(III) to Cr(VI) by O2 is slow, but oxidation by Mnoxides is thermodynamically favourable under acid conditions: in Table 7.8, pe0∗values at pH 5 for reduction of Mn(III,IV) oxides are greater than for reductionof Cr(VI). This process is catalysed by sorption of Cr(III) onto Mn oxidesurfaces (Eary and Rai, 1987). However it is slower than analogous reactionswith other species that have been studied (As(III) → As(V) and Se(IV) →Se (VI)) because oxidation of Cr(III) is less thermodynamically favourable andbecause Cr3+ is less strongly sorbed on positively-charged Mn oxides at lowpH (Scott and Morgan, 1996).

Chromate is reduced to Cr(III) in dissimilatory microbial reactions, but this pro-cess is inhibited at moderate concentrations of C(VI) and so is probably of limitedvalue in detoxifying soils contaminated with Cr(VI) (Lovley, 1993). However,Cr(VI) can also be reduced to Cr(III) abiotically by oxidation of Fe(II): Fe(III)in ferric oxide is reduced to Fe(II) biotically:

4Fe(OH)3 + CH2O + 8H+ −−−→ 4Fe2+ + CO2 + 11H2O

subsequently Cr(VI) is reduced abiotically to Cr(III) as Fe(II) is re-oxidizedto Fe(III):

3Fe2+ + HCrO4− + 8H2O −−−→ 3Fe(OH)3 + Cr(OH)3 + 5H+

Wielinga et al. (2001) demonstrated this process by incubating goethite anaer-obically at pH 7 with lactate and an iron-reducing bacterium, and introducingCr(VI) after commencement of Fe(III) reduction (Figure 7.7). In treatments with-out Cr(VI), accumulation of Fe(II) in solution continued, but in the treatmentswith Cr(VI) it was reversed; in abiotic controls there was no accumulation ofFe(II). Chromate can also be reduced abiotically by sulfide.

Boron

Boron occurs in soil predominantly as the uncharged B(OH)3, though at high pHB(OH)3 is converted to B(OH)4

−(pK = 9.0), which forms insoluble Ca salts. Itmay therefore be deficient to plants in acid soils in humid regions, as a resultof B(OH)3 leaching, or in calcareous soils as a result of precipitation in Casalts. By contrast, in alkaline soils in arid regions, soluble Na borate salts mayaccumulate. Boron toxicity in rice is quite commonly reported where irrigationwater is obtained from deep groundwater in dry seasons (Ponnamperuma andYuan, 1966; Cayton, 1985; Ayers and Westcot, 1989).

Thallium

Thallium occurs in soils in both +3 and +1 oxidation states. Tl3+ behaves muchlike Al3+, but hydrolyses even more readily and insoluble Tl(OH)3 is formed

Page 237: The Biogeochemistry of Submerged Soils

Trace Elements 229

Time (h)

0 20 40 60 800

10

20

30

[Fe

2+ ]

in s

olut

ion

(mg

L−1)

40

50

60

70

Abiotic control

Cr(VI) added

Cr(VI) not added

Start of Cr(VI)addition

Figure 7.7 Abiotic reduction of toxic Cr(VI) to Cr(III) by Fe(III) (Wielinga et al., 2001).Reproduced by permission of the American Chemical Society

at pH < 2 and remains stable to pH > 10. The mobility of Tl3+ in aerobic soilis therefore expected to be low. Under moderate reducing conditions Tl3+ isreduced to Tl+ (pe0 = 21.3, independent of pH). The reduced Tl+ behaves verydifferently, acting more like an exchangeable alkali metal cation. However incor-poration into sulfide minerals may limit its solubility and mobility.

Lead

Lead occurs mainly in the +2 oxidation state in soils, but it may be oxidized toPb4+. It is the least mobile heavy metal in soils. In aerobic soils it is chemisorbedon clays and oxides; forms complexes with organic matter, especially with S-containing functional groups; and forms insoluble hydroxides, carbonates andphosphates. All of these increase with pH, so solubility is greatest under acidconditions. In anaerobic soils it is precipitated as the highly insoluble sulfidegalena (PbS, pK = 27.5). It may also be methylated into volatile forms.

Arsenic and Antimony

Arsenic and antimony are Group VB elements and both occur in soils predomi-nantly in +3 and +5 oxidation states and they have similar redox and sorptionbehaviour. The oxidized forms are rather insoluble in soils and the reduced formsmuch more soluble.

Page 238: The Biogeochemistry of Submerged Soils

230 Nutrients, Toxins and Pollutants

The oxidized form of As, arsenate, As(V), which is present as HAsO42− at

neutral pH (pK values in Table 7.8), is sorbed on soil surfaces in a similar way toorthophosphate. The reduced form arsenite, As(III), which is present in solutionlargely as H3AsO3(pK1 = 9.29), is only weakly sorbed, hence mobility tends toincrease under reducing conditions. Mobility will also increase without reductionof As(V) because, as for phosphate, reductive dissolution of iron oxides resultsin desorption of HAsO4

2− into the soil solution. Under prolonged submergenceAs(III) may be co-precipitated with sulfides.

Additionally, like Hg, As is converted into volatile compounds in microbiallycatalysed reactions that are sensitive to pH and redox conditions (Fowler, 1983).Under strongly reducing conditions methyl- and alkyl-arsine compounds mayform, resulting in loss of As to the atmosphere. The resulting cycle of reactionsis complex but involves in anaerobic soil:

H3AsO3reduction−−−−→ AsH3 ↑

H3AsO3

methylation−−−−−→ (CH3)3As ↑

and on diffusion of (CH3)3As into oxic zones,

(CH3)3Asmethylation−−−−−→ (CH3)3HAsO ↑

Also in oxic zones,

AsO43− methylation−−−−−→ CH3HAsO3

methylation−−−−−→ (CH3)3HAsO ↑

Re-oxidation of As(III) to As(V) under oxidizing conditions is fast and is catal-ysed by sorption onto oxides with the oxide metal acting as oxidant (Scott andMorgan, 1995).

Reductive dissolution of Fe oxyhydroxides holding sorbed As appears toexplain the very large concentrations of As in water from wells drilled intoalluvial sediments of the Brahmaputra and Ganges Rivers in Bangladesh andWest Begal (Nickson et al., 1998, 2000). Dissolved As has accumulated fromthe reduction of As-rich Fe oxyhydroxides formed upstream of the contaminatedareas by weathering of As-rich base metal sulfides. The reduction is driven bysedimentary organic matter in the deposits. Release of As from oxidation of pyritein shallow wells contributes little to the water contamination because any As(IV)released would be re-sorbed on Fe oxides formed in pyrite oxidation.

The HAsO42− ion is taken up by plant roots by the same transport systems

as H2PO4−, leading to excessive uptake and toxicity in plants growing on soils

with high arsenate levels (Meharg and Hartley-Whitaker, 2002). However theextent of uptake varies greatly between soil conditions and plant species (Marinet al., 1993; Onken and Hossner, 1995; Schmoger et al., 2000). Presumably under

Page 239: The Biogeochemistry of Submerged Soils

Trace Elements 231

anaerobic conditions As(OH)3 is absorbed passively in the transpiration stream,and so because of the far greater solubility in soil of As(OH)3 than HAsO4

2−, Asshould accumulate in plants more rapidly under flooded conditions than underdrained aerobic conditions. In rice, the disease ‘straighthead’ is repeatedly linkedto moderate As concentrations in the plant (Horton et al., 1983) and this maylimit the accumulation of As in rice grain.

The redox and sorption behaviour of Sb is similar, but no volatile forms areproduced. The oxidized form of Sb, antimonite, Sb(V), has the anionic formSb(OH)6

− at pH > 4, and is sorbed on oxides and silicate clays. The reducedform, antimonite, Sb(III), is present as the uncharged Sb(OH)3 molecule exceptat very low or very high pH where the Sb(OH)2

+ cation and Sb(OH)4− anion

form, respectively. The uncharged Sb(OH)3 is little sorbed on soil surfaces.

Selenium

Selenium has a complex chemistry in the environment because of its multipleoxidation states and variable surface adsorption properties. Qualitatively it isanalogous to sulfur occurring in the oxidation states +6 (selenate, SeO4

2−), +4(selenite, SeO3

2−), 0 (elemental selenium) and −2 (Se2−, selenide) The Se2−anion closely resembles S2− (radii 0.20 and 0.185 nm, respectively) and is oftenassociated with sulfide minerals. Also, like S, Se is subject to volatilizationthrough biological methylation.

Under oxidizing conditions SeO42− is the thermodynamically favoured form

(for the reduction SeO42− → SeO3

2−, pe0∗ = 7.9 at pH 7—Table 7.8) and undermildly reducing conditions SeO3

2− is the favoured form (for SeO32− → Se,

pe0∗ = 3.0 at pH 7). Since SeO32− is strongly sorbed on oxides and precipi-

tates as Fe2(SeO3)3, whereas SeO42− is only weakly sorbed, especially at high

pH, this leads to large changes in solubility. Hence toxic concentrations of Setend to occur in alkaline soils in arid and semi-arid regions, and irrigation ofsuch soils may move SeO4

2− into groundwater. Microbially mediated reductionsof SeO4

2− to SeO32− and of SeO3

2− to elemental Se have been documentedin a wide variety of soils and aquatic sediments (Lovley, 1993). Reduction isinhibited by NO3

− and Mn(IV), which are preferred electron acceptors. In puresystems rates of transformation of SeO3

2− to SeO42− and vice versa are slow

but, as for oxidation of Cr(III) and As(III), rapid oxidation of SeO32− sorbed on

Mn oxides can occur, with Mn(IV) acting as oxidant (Scott and Morgan, 1996).In reducing environments Se is present as Se2− which forms insoluble com-

pounds with metals, especially Fe(II), or, if metal concentrations are insufficient,the foul-smelling poisonous gas H2Se may be formed. Reduction of SeO3

2− toSe2− (pe0∗ = 0.3 at pH 5 and −1.7 at pH 7) is microbially mediated at low pH(Lovley, 1993).

Though large concentrations of Se can develop in poorly drained soils asa result of accumulation of insoluble Se2− compounds, Se is also lost under

Page 240: The Biogeochemistry of Submerged Soils

232 Nutrients, Toxins and Pollutants

anaerobic conditions by methylation into volatile compounds. The reactionsinclude (Reamer and Zoller, 1980):

HSeO3− methylation−−−−−→ CH3SeO3H

reduction−−−−→ CH2SeO2−

CH3SeO2− methylation−−−−−→ (CH3)2SeO2 ↑ or CH3SeOOCH3 ↑ reduction−−−−→ (CH3)2Se ↑

CH3SeO2− reduction−−−−→ CH3SeH or CH3SeOH −−−→ CH3SeSeCH3 ↑

The different steps are mediated by a consortium of microbes with tolerances tothe various form of Se, resulting in removal of toxic Se from the soil thoughenhancing atmospheric Se transport.

Iodine

Iodine is essential in the mammalian diet to produce the thyroid hormone thy-roxine; deficiency in humans causes goitre. Collectively, deficiencies of iodine,iron, zinc and vitamin A in humans are thought to be at least as widespread anddebilitating as calorie deficiencies (Welch and Graham, 1999). The main sourceof iodine in soils is oceanic salts rather than parent rock, and so deficiency ismost widespread in areas remote from the sea (Fuge, 1996). In principle defi-ciency is easily corrected with dairy supplements. However in practice this is notalways feasible. Addition of iodate to irrigation water has successfully correctedwidespread iodine deficiency in parts of China where the usual methods of sup-plementation had failed (Cao et al., 1994; Jiang et al., 1997). However there isnot much information on the behaviour of iodine in soil and water systems.

Iodine is present in the environment predominantly in the oxidation states −1(I−, iodide) and +5 (IO3

−, iodate). Reduction of IO3− to I− occurs at pe0 =

13.3 at pH 5 and pe0 = 11.3 at pH 7. Hence I− is expected to predominatein the soil solution except in oxic alkaline soils (Whitehead, 1984). HoweverYuita (1992) found predominantly IO3

− in acid Japanese soils contaminated withiodine: the concentrations in solution were some 20 times those of I− and I2.On flooding the soils, the total concentration of I in solution increased 10- to50-fold, predominantly as I−. The concentrations of sorbed I were not measured,but both IO3

− and I− are expected to be bound to organic matter and oxides andhence their concentrations in solution are expected to increase with reductivedissolution reactions. Further, for a given concentration in solution, I− is morerapidly absorbed by plants than IO3

− (Mackowiak and Grossl, 1999). Henceflooding is expected to increase accumulation in plants both through increasedsolubility and increased absorption.

Page 241: The Biogeochemistry of Submerged Soils

8 Trace Gases

This chapter considers the extent, mechanisms and possibilities for control ofemissions of trace gases from submerged soils. The focus is on ricefields becausethis is where research has been most intense and because ricefields are the focusof the greatest scrutiny for possibilities to reduce emissions.

8.1 METHANE

8.1.1 GLOBAL BUDGET

Table 8.1 shows recent estimates of the global methane budget made by theIntergovernmental Panel on Climate Change (IPCC) (Prather et al., 2001). Thereare substantial emissions from natural sources, particularly wetlands. But anthro-pogenic sources account for 60 % of the total emission and the abundance of CH4

in the atmosphere is now more than double its pre-industrial value–1745 ppb(molar mixing ratio in the troposphere) compared with 700 ppb in 1750. The rateof increase has been near exponential over the last 300 years. However the annualrate of increase has been highly variable and in the last 20 years it has declinedfor reasons that are not fully understood. The current percentage rate of increaseis comparable to that of CO2 –about 0.4 % year−1. The radiative forcing effect ofCH4 is about 21 times that of CO2 per mole of gas. Currently, increases in CO2

account for about 50 % of global warming and increases in CH4 about 20 %.There is therefore political pressure to decrease man-made emissions of CH4

by whatever means possible. Since CH4 has a short lifetime in the atmosphere(8 years compared with 50–200 years for CO2), modest decreases in emissionscan quickly have a large effect on atmospheric abundance. Adjusting the globalCO2 balance requires larger percentage as well as absolute changes. Further,it is argued that the fossil fuels responsible for most of the atmospheric CO2

increase also produce aerosols that have negative radiative forcing effects inthe troposphere (Hansen et al., 2000). These include aerosols of non-absorbingsulfates, which both directly reflect radiation and increase reflection by clouds.They therefore to some extent mitigate the effects of increased CO2, and thiscomplicates calculations of the relative importance of CO2 emissions versus othergreenhouse gases.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 242: The Biogeochemistry of Submerged Soils

234 Trace Gases

Table 8.1 Estimates of the global methane budget (Tg CH4 year−1) from differentsources and sinks

Reference Funget al.

(1991)

Heinet al.

(1997)

Olivieret al.

(1999)

Lelieveldet al.

(1998)

Mosieret al.

(1998b)

Caoet al.

(1998)

Houwelinget al.

(1999)

Pratheret al.

(2001)

Base year 1980s — 1990 1992 1994 — — 1998

Natural sourcesWetlands 115 237 225b 92 145Termites 20 — 20 20Oceans 10 — 15 15Hydrates 5 — 10 —AnthropogenicEnergy 75 97 109 110 89Landfills 40 35 36 40 73Ruminants 80 90a 93a 115 89 93Waste treatment — a a 25 14 —Rice agriculture 100 88 60 b 25–54 53 —Biomass burning 55 40 23 40 34 40Other — — — 15 20

Total source 500 587 600 598SinksTropospheric OH 10 — 30 44 30 30Stratosphere 450 489 510 506Soils — 46 40 40

Total sink 460 535 580 576

a Waste treatment included under ruminants.b Rice included under wetlands.Source: adapted from Prather et al. (2001).

Estimates of CH4 emissions from ricefields have improved greatly in thepast decade and the contribution of ricefields to the global CH4 budgetis far smaller than originally thought (Table 8.2). However, there is stillconsiderable uncertainty. Recent estimates compiled by the IPCC range from 25to 60 Tg CH4 year−1 out of a total global emission of about 600 Tg CH4 year−1

(Table 8.1), but credible estimates of less than 10 Tg CH4 year−1 are also made(Table 8.2). These compare with 100–111 Tg CH4 year−1 from fossil fuels,80–115 Tg CH4 year−1 from ruminants, and 35–75 Tg CH4 year−1 from landfills.Rice therefore ranks about fourth among anthropogenic sources of methane.

8.1.2 PROCESSES GOVERNING METHANE EMISSIONS FROM RICE

Reviewers of this topic include Schutz et al. (1989a,b), Conrad (1993), Neueand Roger (1994) and Segers (1998). The rate of emission depends on the linkedrates of CH4 production, transport and oxidation, which are sensitive to a hostof soil, plant, climate and management variables. Production occurs through

Page 243: The Biogeochemistry of Submerged Soils

Methane 235

Table 8.2 Estimates of global CH4 emissions from ricefields in chronological order

Method Source strength(Tg year−1)

1 Methane production in incubated samples of rice soilsmultiplied by estimated amounts of soil (Koyama,1963)

190

2 Uniform emission factor based on flux measurementsmultiplied by harvest area of rice (Cicerone andShetter, 1981; Holzapfel-Pschorn and Seiler, 1986;Schutz et al., 1989a, respectively)

59, 70–170, 50–150

• Excluding upland rice area −12 %• Allowing for average temperature during growing

season (IPCC, 1997)60–105

3a Methane emission proportional to net primaryproduction, e.g. 3–7 % (Aselman and Crutzen, 1989),5 % (Taylor et al., 1991)

60–140

• Allowing for soil CH4 emission potential (Bacheletand Neue, 1993)

47

3b Methane emission proportional to carbon returned to thesoil: 30 % of the carbon retuned emitted as CH4(Neue et al., 1990)

63

• Allowing for soil CH4 emission potential (Bacheletand Neue, 1993)

52

4 Specific emission factors for specific ecosystems,regions or management, or all (IPCC, 1997)

• Rice ecosystem-specific emission factors (Neue andSass, 1998)

30–50

• Country-specific emission factors (Neue and Sass,1998)

32

• Regional rice statistics (Yao et al., 1996) 15 (China only)5 Empirical models using data from national statistics

linked to GIS (Kern et al., 1997)10 ± 3 (China only)

6 Mechanistic models using weather, soil, agronomic andother data linked to GIS

Cao et al. (1996) 53Huang et al. (1998) 7.2–13.6 (China only)Matthews et al. (2000a) 6.5–17.4 (70 % of area)

Source: adapted from van der Gon et al. (2000). Reproduced with kind permission of Kluwer Academic Publishers.

the anaerobic decomposition of organic matter, mostly after inorganic terminalelectron acceptors have been exhausted (Section 5.3). Transport occurs by ebul-lition (Section 2.3), by diffusion through the soil to root surfaces and then viathe plant aerenchyma to the atmosphere (Section 6.2), and to a lesser extent bydiffusion directly to the floodwater and atmosphere. Oxidation is also microbiallymediated, mainly by obligate aerobes in the oxic floodwater–soil and root–soilinterfaces. The sensitivity of these processes to many variables suggests the pos-sibility of interventions to decrease emissions. It also complicates the predictionof emissions from readily measurable parameters.

Page 244: The Biogeochemistry of Submerged Soils

236 Trace Gases

Differences between Rice Production Systems

Early measurements were made mostly in temperate countries (Ciceron et al.,1983; Seiler etal., 1984; Holzapfel-Pschorn and Seiler, 1986; Schutz et al.,1989a,b), but a large programme of measurements was conducted in Asia inthe 1990s by the IRRI and partners using a common measurement system(summarized in the book by Wassmann et al., 2000b). This revealed largedifferences in emissions per season–more than an order of magnitude–betweendifferent climatic zones across Asia, between types of rice culture, and betweenmanagement practices, particularly management of crop residues and use oforganic manures. For example, mean emissions in China are large because of thewidespread use of organic manures, in spite of the moderating effect of lowertemperature compared with tropical Asia; whereas in India, crop residues areoften largely removed from the fields and emissions are correspondingly smaller.

Seasonal emissions from irrigated rice are generally two- to four-fold greater thanfrom rainfed lowland rice under similar climates. Mean fluxes from deepwater riceare smaller than from irrigated rice, but because of the far longer growing season,the total seasonal emission may be similar. Extrapolating from measured seasonalemissions for the different rice ecosystems and the area of each planted annually,irrigated rice accounts for 70–80 % of global CH4 emissions from rice, rainfedlowland rice for 15 % and deepwater rice for 10 % (Wassmann et al., 2000d).

Differences within Seasons

Emissions from irrigated ricefields show distinct diurnal and seasonal variationswhich illustrate the interactions between the governing processes. The diurnalvariation includes a maximum during the day and a minimum at night, and ismainly linked to changes in the temperature of the soil solution which drivechanges in rates of CH4 production and solubility and therefore changes in emis-sion, whether through the plant or by ebullition (Schutz et al., 1990; Yagi et al.,1994; Wang et al., 1999). The seasonal variation often has two peak periods:one early in the season corresponding primarily to the decomposition of addedorganic matter and ending abruptly when the organic matter has been used up;and a second later in the season corresponding to emissions via the plant fuelledby root exudation and turnover (Holzapfel-Pschorn et al., 1986; Schutz et al.,1989a,b; Yagi and Minami, 1990; Neue, 1997). Figure 8.1 shows the typicalpattern for a field to which organic matter has been added.

Various factors alter this basic pattern:

(1) The amount of organic matter added. This is highly variable across riceproduction systems, depending on such factors as the time available betweencrops, mechanization allowing residue incorporation, alternative requirementsfor organic matter, and so forth. Often straw is entirely removed from the

Page 245: The Biogeochemistry of Submerged Soils

Methane 237

Time (days after planting)

−20 0 20 40 60 80 100 1200

5

10

15

20R

ate

of e

mis

sion

(kg

C h

a−1 d

ay−1

)

25

Figure 8.1 Seasonal variation in CH4 emission. Rice straw (t ha−1) was incorporated inthe soil 14 days before planting the crop (data from Wassmann et al., 2000a). Reproducedby permission of Kluwer Academic Publishers

field following a crop and there are no additions to the succeeding crop. Theearly-season peak may then be absent.

(2) The availability of inorganic electron acceptors, both in the soil constituentsand added in mineral fertilizers. This affects the time course of soil reductionand hence the rate at which sufficiently reducing conditions for methanogen-esis develop.

(3) Temperature and radiation. The temperature regime affects rates of CH4 pro-duction, transport and oxidation, and generally high temperature favours highrates of emission. The main effect of radiation is through its influence oncrop growth. Hence dry season emissions at a particular site are often muchsmaller than wet season emissions, and a well-managed crop under optimalconditions of temperature and radiation emits less CH4 (see Section 8.1.5).

(4) Water regime, particularly where there is mid-season drainage allowingescape of entrapped CH4 but also oxidation of the soil. The water regime isalso affected by soil texture as this affects percolation rates; high percolationrates tend to decrease emissions because less reducing conditions aremaintained in the soil. Texture may also affect gas entrapment and ebullition.

So the picture is complicated.

8.1.3 MODELLING METHANE EMISSION

Given the complexity, some form of mechanism-based modelling is required tounderstand or predict emissions in given circumstances. Approaches to modellingemissions from rice are reviewed by van der Gon et al. (2000). A complete

Page 246: The Biogeochemistry of Submerged Soils

238 Trace Gases

treatment is given by Arah and Kirk (2000) in a general transport-reaction model,which they simplify to focus on emissions fuelled by root exudation and deathand transmission through the plant. This model is now outlined.

Following Equation (2.6), the concentration profile with depth z of any non-adsorbed substrate in an areally homogeneous system is given by:

∂C

∂t= ∂

∂z

(D

∂C

∂z− vCL

)+ O + P − Q − R − S (8.1)

where O is the root-mediated influx, P is production, Q is consumption, R isthe root-mediated efflux and S is ebullition. The terms D, v, O, P , Q, R, S

and C are effective areal averages at depth z and time t : they subsume withinthemselves any areal heterogeneity in the real system. Temperature is an implicitvariable in Equation (8.1), influencing the instantaneous rates of all transport andreaction processes.

Diffusion depends on the bulk concentration C; leaching and consumptionon the solution-phase concentration CL; and root-mediated efflux and ebullitionon the gas-phase concentration CG. Root-mediated influx and production areindependent of C, CL and CG, though they may depend on other properties ofthe system. The concentrations C, CL and CG are easily inter-converted assumingequilibrium between solution and gas phases:

CL = HCG (8.2)

where H is a dimensionless Henry’s law constant. The bulk concentration isgiven by

C = θGCG + θLCL (8.3)

where θG is the air-filled porosity and θL is the volumetric water constant. Hence

CG = C

θG + HθL, i.e.

∂θCG

∂C= 1

θG + HθL(8.4)

and

CL = HC

θG + HθL, i.e.

∂θCL

∂C= H

θG + HθL(8.5)

The diffusion coefficient allows for both gas and liquid phase diffusion. It isgiven by (Stephen et al., 1998a,b):

D = DGθG + DLHθL

(θG + HθL)f(8.6)

where f is a tortuosity factor, approximately equal to unity in a well puddled soil.Root-mediated influx may be represented as an exchange process in which

only the gas phase moves. Most simply this can be expressed (Stephen et al.,1998a,b):

O = κDGCG0 (8.7)

Page 247: The Biogeochemistry of Submerged Soils

Methane 239

where CG0 is the concentration at z = 0 and κ is a transmission constant whichdepends on such variables as the root length density, root porosity, the perme-ability of root tips and laterals, and root architecture. Similarly for root-mediatedefflux at a particular depth:

R = κDGCG (8.8)

The rate of ebullition, S, of a particular substance depends on its gas-phaseconcentration. Most simply this is expressed:

S = σCG (8.9)

where σ is a rate constant.The boundary conditions for solving Equation (8.1) are (a) for volatile solutes

that the concentration at the surface is known and (b) for non-volatile solutesthat the flux is zero.

Parameter Values

With appropriate values for H, DG, DL, v, C0 and the depth-profiles of θG, θL, κ

and σ , Equations (8.1) to (8.9) apply to any non-adsorbed substance. To simulateCH4 production, transport, oxidation and emission, we need to consider at leasttwo mobile substances–O2 and CH4 –and at least three reactions:

Oxic respiration CH2O + O2 −−−→ CO2 + H2O

Methanogenesis CH2O + CH2O −−−→ CO2 + CH4

CH4 oxidation CH4 + 2O2 −−−→ CO2 + 2H2O

Here CH2O represents oxidizable organic matter. In reality the reactions withinorganic terminal electron acceptors, particularly Fe(III) and SO4

2−, should alsobe considered. But in the absence of a complete understanding of these processes(see Chapter 5), and for the sake of simplicity, we exclude them.

Production, P . Methanogenesis is inhibited by solution-phase O2:

PCH4 = IVM (8.10)

where VM (z, t) is the CH4 production potential and I (z, t) is an inhibitionfunction which we take to be:

I = 1

1 + ηCLO2

0 ≤ I ≤ 1 (8.11)

where η is an inhibition efficiency constant. No reaction produces O2, i.e. PO2 = 0.

Page 248: The Biogeochemistry of Submerged Soils

240 Trace Gases

Consumption, Q . Methane oxidation is described with dual-substrate Michae-lis–Menten kinetics:

QCHh = VO

(CLO2

KO1 + CLO2

)(CLCH4

KO2 + CLCH4

)(8.12)

where VO (z, t) is the oxidation potential and KO1 and KO2 are Michaelis con-stants. Oxygen is consumed in respiration and CH4 oxidation, the latter requiringtwo molecules of O2 per molecule of CH4. Hence, assuming Michaelis–Mentenkinetics and no carbon limitation:

QO2 = VR

(CLO2

KR + CLO2

)+ 2QCH4 (8.13)

where VR (z, t) is the respiration potential and KR a Michaelis constant.

Reaction Potentials. The reaction potentials VM, VO and VR are the rates atwhich methanogenesis, CH4 oxidation and oxic respiration would proceed in situwere all enzymes saturated with the necessary substrates. They depend on in situenzyme concentrations and hence on in situ microbial populations. They changeover time.

Equations (8.1)–(8.13) can be solved to provide transient- or steady-state pro-files of O2 and CH4 concentration, reaction rates and surface fluxes for anycombination of the controlling variables θG, θL, v, κ, σ, VM, VO and VR. Where,as is usual, one or more of the controlling variables may be further simplified,approximated or neglected, process-based simulation of CH4 emission becomespossible using a relatively limited set of input data.

Simplified Model Focusing on Effects of the Plant

Arah and Kirk (2000) make the following assumptions to develop a simpli-fied model:

(1) the soil is saturated and air-filled porosity external to roots is negligible, i.eθG = 0;

(2) water content, θL, is uniform with depth;(3) leaching is negligible, i.e. v = 0;(4) root transmissivity is proportional to root-length density, i.e. κ = kTLV;(5) ebullition is negligible, i.e. σ = 0;(6) oxidation potential, VO, is constant;(7) CH4 production potential is proportional to respiration potential, VM = VR/ 50;(8) respiration potential is proportional to root-length density, i.e. VR = kVLV;(9) root-length density is normally distributed with depth, with maximum value

LV max at depth zmax, and standard deviation equal to zmax/ 2.

Page 249: The Biogeochemistry of Submerged Soils

Methane 241

Assumptions 1–9 are ad hoc simplifications introduced in order to define astandard system with characteristics that can be explored. Some of the assump-tions (1–3, 5) are relatively uncontroversial; others (7–9) depend on an under-lying supposition that root-mediated processes dominate.

Results. Figure 8.2 gives steady-state profiles of O2 and CH4 and the corre-sponding reaction rates calculated with the model for the fixed root system definedin Assumption 9. Net O2 consumption is 460 µmol m−2h−1, net CH4 emission is480 µmol m−2 h−1, the fractions of the O2 and CH4 fluxes through the plant are0.84 and 0.97, respectively, and the fraction of CH4 oxidized prior to emissionis 0.13. These are all credible numbers.

Figure 8.3 shows the consequences of varying the root transmissivity factorkT and the substrate supply factor kV. It shows that, other things being equal, theCH4 flux increases with kV but decreases with kT. The latter, perhaps counter-intuitive, result reflects the fact that transport through roots allows O2 into thesystem as well as CH4 out. Enhanced O2 concentrations in the rhizosphere inhibitmethanogenesis and promote oxidation, and the combined effect of these twoprocesses more than compensates for the greater ease with which CH4 can escape.The model also shows, unsurprisingly, that the fraction of the CH4 transmittedthrough the plants increases as root transmissivity increases and decreases assubstrate supply increases.

Figure 8.3(b) shows that the fraction of CH4 that is oxidized before reach-ing the atmosphere is a sensitive function of kV and kT. Increasing kV reducesthe fraction oxidized, presumably because the oxidation potential VO is heldconstant in these simulations and increased production simply overwhelms theoxidation capacity; increasing kT increases the fraction oxidized where transmis-sivity is low and decreases it where transmissivity is high, presumably reflecting

Concentration (mM)

0.0 0.1 0.2 0.3

Dep

th (

cm)

0

10

20

30

40

50

oxygenmethane

Reaction rate (mmol m−3 h−1)

0 5 10 15 20 25 30

(a) (b)

Figure 8.2 Calculated profiles of O2 and CH4 concentrations (a) and reaction rates(b) (Arah and Kirk, 2000). Reproduced by permission of Kluwer Academic Publishers

Page 250: The Biogeochemistry of Submerged Soils

242 Trace Gases

1000

10001000

1000

100

10

10

101

1

Multiple of standard transport factor kT

0.1 1 10

Mul

tiple

of

stan

dard

sup

ply

fact

or k

V

0.1

1

10

0.1

0.1

0.1

0.1

0.20.2

0.2

0.2

0.30.3

0.3

0.3

0.40.4

0.4

0.4

0.5

0.60.6 0.6

0.7

0.1 1 10

0.1

1

10(a) CH4 flux (µmol m−2 h−1) (b) Fraction of CH4 oxidized

0.50.5 0.5

0.7

100

100

100

Figure 8.3 Calculated effects of the CH4 supply factor kV (Assumptions 7 and 8 ofthe simplified model) and the root transmissivity factor kT (Assumption 4) on (a) theCH4 flux and (b) the fraction of CH4 oxidized. Lines are contours of CH4 flux or frac-tion oxidized (Arah and Kirk, 2000). Reproduced by permission of Kluwer AcademicPublishers

the intricate balance between the twin effects of O2, namely inhibiting CH4

production and promoting CH4 oxidation (and thereby anaerobiosis, and therebyCH4 production).

Conclusions. The model shows that the bulk of mid to late season emissionoccurs via the plant; where organic matter has been added, large emissions earlyin the season must occur by ebullition. For any given root-length density profile:

(1) rice cultivars with high specific substrate supply rates will lead to increasedCH4 emissions;

(2) cultivars with high specific transmissivities will decrease CH4 emissions;(3) drainage leading to an air-filled porosity of just 0.01 decreases CH4 emissions

practically to zero.

These findings broadly agree with experimental observations. Measured ratesof CH4 oxidation in the rice rhizosphere range widely from 5 to 90 % of theCH4 transported (Holzapfel-Pschorn et al., 1985; Epp and Chanton, 1993; vander Gon and Neue, 1996). This agrees with the model. Rates of O2 flow throughrice roots to the rhizosphere are of the order of a few mmol O2m−2 (soil surface)h−1 (Section 6.4), which is sufficient to account for the rates of oxidation calcu-lated with the model. Measured differences in emissions between rice cultivarsare largely due to differences in root biomass (Lu et al., 1999); the effects ofdifferences in root porosity are smaller (Aulakh et al., 2001a,b).

What little is known about the microbiology of CH4 oxidation inthe rice rhizosphere indicates complicated kinetics and competition effects.

Page 251: The Biogeochemistry of Submerged Soils

Methane 243

Studies of microbial growth rates indicate that VR >> VO but KO1 < KR inEquations (8.12) and (8.13) (van Bodegom et al., 2001). Hence heterotrophs willout compete methanotrophs for O2 except at very small O2 concentrations.This may constrain the spatial distribution of CH4 oxidation in the rhizosphere.Nutrient concentrations may also be limiting, in which case fertilizer may enhanceoxidation and lessen emissions, depending on interactions with plant growthand other variables. Bodelier et al. (2000) found enhanced CH4 oxidation inthe rice rhizosphere on adding N, contrary to expectations for aerobic soilswhere it is found that methanotrophs preferentially oxidize NH4

+ over CH4.In the rice rhizosphere in flooded soil, CH4 concentrations are much higher and,close to roots, NH4

+ concentrations smaller. Likewise Lu et al. (1999) founddiminished CH4 emissions from rice when P was added to a P-deficient floodedsoil. Salinity also impairs CH4 oxidation and to a greater extent than it impairsmethanogenesis (van der Gon and Neue, 1995).

Field-scale Model of Emissions

Matthews et al. (2000a) have developed a field-scale model of emissions basedon the above approach. In addition to the processes discussed above, the field-scale model allows for added organic matter and soil organic matter separately,and for the effects of inorganic terminal electron acceptors. Figure 8.4 shows thatthe model was capable of predicting seasonal emissions at a particular site frommodel parameter values measured independent of the emission data.

Time (days after planting)

−20 0 20 40 60 80 100 1200

5

10

15

Rat

e of

em

issi

on (

kg C

ha−1

day

−1)

20

25

Figure 8.4 Measured (points) and calculated (line) seasonal variation in CH4 emission.Measured data as in Figure 8.1; calculation described in text (Matthews et al., 2000a).Reproduced by permission of Kluwer Academic Publishers

Page 252: The Biogeochemistry of Submerged Soils

244 Trace Gases

8.1.4 ESTIMATING EMISSIONS AT THE REGIONAL SCALE

The uncertainty in estimates of the global CH4 emission from rice remains large:the mean estimate for the 1990s in Table 8.1 is 60 ± 30 Tg year−1, equivalent to arange of 5 to 15 % of the total emission. This reflects the diversity of conditions inwhich rice is grown and the large effects of management. To date, most estimateshave involved a down-scaling approach in which knowledge and understandingat the field or local scale are used to extrapolate to the regional scale and above;this approach is limited by the availability of reliable data at the required scale.An alternative approach is to work in the opposite direction, down-scaling byinterpolation from measurements of overall terrestrial emissions made in globalair sampling networks (Heimann and Kaminski, 1999). Further improvements inestimates are likely to come from a meeting in the middle of these approaches.I here outline their pros and cons.

Up-scaling using Mechanistic Models and GIS

Knox et al. (2000) and Matthews et al. (2000b) have coupled a field-scale modelof CH4 emissions from rice to GIS systems, and used available regional data onweather, soil, agronomic management and other variables to make regional-scaleestimates of emissions. The model is based on the approach described earlier.The extrapolation is based on the following framework.

Basic polygons for attributing vector datasets were derived from a digitizedmap of national and provincial or state boundaries for China (31 polygons), India(31), Indonesia (26), Philippines (80) and Thailand (73). At least one polygonwas defined for each province or state. Data on crop production and croppedarea under the four rice ecosystems were obtained for each polygon using themap of Huke and Huke (1997), mainly from 1990. Weather data in individualpolygons were obtained from the nearest of 46 weather stations within the appro-priate agroecological zone. Soil data were obtained from FAO-DSMW soil units(1:5 000 000) with supplementary data for individual soil units in top and sub-soils: pH, organic C, Fe content, texture and available water capacity. Each soilproperty was given a weighted mean value for the polygon based on the distri-bution of FAO soil units. Locally recommended crop management was assumed.Due to lack of data, no allowance was made for differences in applications oforganic manures, and this will probably have caused underestimates in emissions.The resulting global estimates are in the range 10–25 Tg CH4 year−1.

This approach is of course limited by the availability of reliable data and theresolution of the data. An inherent problem in the ‘up-scaling’ process is theinteraction between variance in input parameters and non-linearity in models.This may produce chaotic behaviour. van Bodegom et al. (2002) discuss thisin relation to CH4 emission from rice. The point at which input data are aver-aged before making model runs may also be limited by the available computing

Page 253: The Biogeochemistry of Submerged Soils

Methane 245

power. Nonetheless the great strength of a mechanism-based modelling approachis that the sensitivity to different variables can be tested and the effects of time-dependent variables, feedback mechanisms and the consequences of changes inthe system can be explored.

Down-scaling using Inverse Modelling

The spatial and temporal variation in sources and sinks of a trace gas are reflectedin the spatial and temporal variation of its mixing ratio in the atmosphere. Ininverse modelling, observed net emissions over a region are apportioned to knownsources and sinks according to a priori assumptions about their relative impor-tance. The resulting magnitudes of the sources and sinks and their distributionsare then used to calculate the net flux a posteriori using models of atmospherictransport and chemistry. The agreement between the a priori and a posteriorivalues indicates the accuracy of the a priori assumptions.

This approach has been applied to global emissions of CO2, CH4, N2O, halo-carbons and CO, which have sufficiently long lifetimes and well understood atmo-spheric chemistries (Heimann and Kaminski, 1999). Methane emissions havebeen studied by Hein et al. (1977) and Houweling et al. (1999). van der Gonet al. (2000) used inverse modelling to test the effects of different a prioriassumptions about the magnitude of CH4 emissions from rice. In Scenario A awidely accepted standard range of 50–80 Tg CH4 year−1 (Lelieveld et al., 1998)is used, and in Scenario B their own best estimate of 15–30 Tg CH4 year−1. Thesame total emission is assumed for the two scenarios, and the same combinedflux from wetlands and ricelands. The global emission is apportioned to rice andother sources according to these assumptions, and calculations made for the wholeglobe, the northern and southern hemispheres, and an area roughly correspond-ing to the part of Asia where rice is most important. The results, summarizedin Table 8.3, show that the assumed and calculated results for the rice area aremuch closer for the lower emission scenario, indicating that it is more realistic.

Table 8.3 Global distributions of CH4 emissions (Tg CH4 year−1) calculated usinginverse modelling. In Scenario A rice contributes 50–80 Tg year−1 and in B15–30 Tg year−1; the net contribution of natural wetlands and ricelands is constant

Globe Northernhemisphere

Southernhemisphere

Asiaa

Scenario Aassumed 528 ± 90 405 ± 81 123 ± 40 111 ± 56calculated 505 ± 24 340 ± 19 165 ± 18 77 ± 23

Scenario Aassumed 528 ± 24 384 ± 66 143 ± 38 74 ± 31calculated 508 ± 24 342 ± 16 166 ± 17 66 ± 18

a 10◦ N, 75◦ W to 40◦ N, 135◦ W.Source: van der Gon et al. (2000). Reproduced with kind permission of Kluwer Academic Publishers.

Page 254: The Biogeochemistry of Submerged Soils

246 Trace Gases

However, as yet monitoring networks on atmospheric mixing ratios are notsufficiently extensive or close to continental CH4 sources to make this approachentirely reliable for rice.

8.1.5 POSSIBILITIES FOR DECREASING EMISSIONS

Methane emissions from rice are expected to increase over the next few decadesas rice production increases to meet projected increases in population and demand.Intensified production to produce more rice on a smaller land area, with morecrops per year, greater use of fertilizers, greater quantities of crop residues tobe disposed of, legislation against burning residues, and mechanization allowingincorporation of residues are all likely to exacerbate emissions.

However the better crop management necessary to increase yields will of itselftend to lessen emissions. Hence van der Gon et al. (2002) found that emissionsover five years at an irrigated site in the Philippines were inversely related tograin yield (Figure 8.5). They found that large emissions were associated withsmall ratios of grain to biomass, particularly in the wet season, and hypothesizedthat this caused greater CH4 production from root carbon released into the soil.In a greenhouse experiment, removing spikelets to reduce the plants’ capacity tostore photosynthate in grains increased CH4 emissions, possibly via more carbonentering the soil. Unfavourable conditions for spikelet formation in the wet sea-son may similarly explain high CH4 emissions. For similar reasons, modern highyielding rice varieties generally emit less CH4 than traditional varieties (Neueet al., 1997; Aulakh et al., 2001a,b), and there are possibilities for breeding vari-eties with low emission potentials, exploiting differences in biomass partitioningand gas transport.

Grain yield (t ha−1)

3 4 5 6 7 8 9

Sea

sona

l CH

4 em

issi

on(k

g C

H4-

C h

a−1)

0

100

200

300

400

500

dry seasonwet season

y = −61x + 558 r 2 = 0.94

Figure 8.5 Seasonal CH4 emission as a function of grain yield in wet and dry seasonsat Maligaya, Luzon, Philippines (van der Gon et al., 2002). Reproduced by permission ofNational Academy of Sciences, USA

Page 255: The Biogeochemistry of Submerged Soils

Nitrogen Oxides 247

Water management might also be manipulated to lessen emissions. Currentlydirect seeding and other water-conserving practices are being adopted in manyparts of Asia in response to shortages of water and labour (Guerra et al., 1998). Asingle, well-timed period of drainage in the early season can decrease emissionsby 50 % without compromising yield (Sass et al., 1992; Neue, 1997; Wassmannet al., 2000c). However the timing is critical so that entrapped CH4 that wouldotherwise be oxidized is not released, and N is not lost from the soil throughnitrification–denitrification, especially if conditions are such that nitrous oxideforms.

8.2 NITROGEN OXIDES

8.2.1 GLOBAL BUDGET

Nitrous oxide (N2O) is an important greenhouse gas with a radiative forcingeffect 310 times that of CO2 and a lifetime in the troposphere of approxi-mately 120 years. Part of the N2O is converted to NO in the stratosphere, andso contributes to depletion of ozone. Nitric oxide (NO) is very reactive in theatmosphere and has a lifetime of only 1–10 days. It contributes to acidificationand to reactions leading to the formation of ozone in the troposphere, and so alsoto global warming.

Table 8.4a shows estimates of the global nitrous oxide budget from differentsources and sinks (Prather et al., 2001) and Table 8.4b estimates for NOx(NO +NO2). As for CH4, there are substantial emissions from natural sources, par-ticularly the ocean and humid tropical forests for N2O and lightning and soilprocesses for NOx . However anthropogenic sources account for 40 % of thetotal emission in both cases. For NOx , this is largely from road transport andpower plants, but for N2O 60 % is from agricultural soils. The post-industrialincrease in N2O abundance in the atmosphere is smaller than that of CH4 –314 ppb(molar mixing ratio in the troposphere) compared with 270 ppb in 1750–andthe current percentage increase (0.25 % year−1) is smaller than that of CH4.However because of its large radiative forcing effect, the increase is highly sig-nificant.

Irrigated ricefields are not expected to be major sources of N2O if the fieldsare kept continuously submerged during the growing season (Buresh and Austin,1988; De Datta, 1995; Hou et al., 2000). Rates of nitrification of NH4

+ in rice-fields and subsequent denitrification can be substantial (Chapter 5). In general,conditions are sufficiently reducing and the availability of organic substrate suf-ficiently large that denitrification proceeds as far as N2 with little intermediaryN2O produced en route. However, under fluctuating water regimes, as in ricesystems in which the soil is deliberately drained in the middle of the season,conditions may be ideal for N2O emission. This is often done to remove toxicproducts of anaerobic metabolism or simply to save water. A common practice

Page 256: The Biogeochemistry of Submerged Soils

248 Trace Gases

Table 8.4a Estimates of the global nitrous oxide budget (Tg N year−1; mean and range)from different sources and sinks

Reference Olivier et al.(1998)

Mosier et al., (1998a);Kroeze et al., (1999)

Prather et al.(2001)

Base year 1990 1994 1990s

Natural sourcesOcean 3.6 (2.8–5.7) 3.0 (1–5)Atmosphere (NH3

oxidation)0.6 (0.3–1.2) 0.6 (0.3–1.2)

Tropical soilsWet forest 3.0 (2.2–3.7)Dry savanna 1.0 (0.5–2.0)

Temperate soilsForests 1.0 (0.1–2.0)Grasslands 1.0 (0.5–2.0)

All soils 6.6 (3.3–9.9)AnthropogenicAgricultural soils 1.9 (0.7–4.3) 4.2 (0.6–14.8)Biomass burning 0.5 (0.2–0.8) 0.5 (0.2–1.0)Industrial sources 0.7 (0.2–1.1) 1.3 (0.7–1.0)Cattle and feedlots 1.0 (0.2–2.0) 2.1 (0.6–3.1) 6.9Total source 14.9 (7.7–24.5) 17.7 (6.7–36.6)SinksTotal sink (stratosphere) 12.3 (9–16) 12.6Implied total sourcea 16.2 16.4

a Total sink + atmospheric increase.Source: adapted from Prather et al. (2001).

Table 8.4b Estimates of the global NOx(NO + NO2)budget (Tg N year−1; mean and range)

Base year 1990

Natural sourcesAll soils 5.5 (4–12)Lightning 12.2 (2–20)NH3 oxidation in atmosphere 0.9 (0–1.6)N2O destruction in stratosphere 0.7 (0.4–1)AnthropogenicFossil fuel combustion 23.4 (13–31)Biomass burning 7.7 (3–15)Total source 50.4 (22–81)

Source: Olivier et al. (1998).

in southern China is to apply large quantities of nitrogen fertilizer early in thegrowing season to stimulate crop growth and production of tillers, and to thenabruptly drain the soil to arrest tillering by driving off nitrogen through nitrifi-cation–denitrification. Here and in other systems, large emissions of N2O are to

Page 257: The Biogeochemistry of Submerged Soils

Nitrogen Oxides 249

be expected. Current efforts to make water use in rice production more efficientwill undoubtedly increase N2O emissions unless steps are taken to avoid this.

Rice can also be an indirect source of N2O (and NO) emissions via depositionof volatilized ammonia on natural ecosystems, particularly wet tropical forests,which are one of the main ‘natural’ sources of N2O (Table 8.4a).

8.2.2 PROCESSES GOVERNING NITROUS AND NITRIC OXIDEEMISSIONS FROM RICE

Emissions of nitric and nitrous oxides are the result of microbial nitrification anddenitrification in soils, controlled principally by soil water and mineral N contents,labile organic carbon, and temperature. Nitric oxide is a direct intermediate ofboth nitrification

NH4+ −−−→ NH2OH −−−→ NO −−−→ NO2

− −−−→ NO3−

and denitrification,

NO3− −−−→ NO2

− −−−→ NO −−−→ N2O −−−→ N2

It is thought that little net NO is produced in denitrification, it being readilyreduced to N2O, and nitrification is therefore the main source of NO (Andersonand Levine, 1986; Skiba et al., 1993). Nitrous oxide is also produced in bothnitrification and denitrification. At low O2 concentrations in otherwise aerobicsoil, small amounts of N2O are formed as a by-product of nitrification, N2O notitself being reduced to NO2

−. In denitrification, the proportion of N2O producedrelative to N2 increases as the availability of O2 increases and that of carbondecreases (Tiedje, 1988). In general only a small fraction of the N nitrified ordenitrified in these pathways is released as NO or N2O. The emission is thereforesensitive to the amount of mineral N in the system, which is driven principally byadditions of nitrogen fertilizers and deposition of nitrogen from the atmosphere.

As discussed in Chapter 5, in submerged soils nitrification occurs in aerobicsites at the floodwater–soil and root–soil interfaces. Denitrification occurs upondiffusion of the NO3

− to the anaerobic bulk soil. Denitrification is favoured overdissimilatory reduction to NH4

+(NO3− → NO2

− → NH4+) because of the large

ratio of available carbon to electron acceptors in submerged soils. Denitrificationis likely to proceed completely to N2 with little accumulation of N2O because ofthe very large sink and therefore steep concentration gradient of O2, and becausecarbon is less likely to be limiting (Section 5.1).

However this will not be the case when a submerged soil is drained and airenters, leading to gradients of oxidation from the surfaces of soil cracks towardsthe anaerobic interiors of soil clods. Now conditions may be ripe for productionof nitrous and nitric oxides.

Hence there is a fine balance between conditions favouring methanogenesis andthose favouring nitrous oxide production. This is nicely illustrated in Figure 8.6,

Page 258: The Biogeochemistry of Submerged Soils

250 Trace Gases

CH

4 flu

x (m

g m

−2 h

−1)

−40

−30

−20

−10

0

10

20

30

40

N2O

flu

x (µ

g m

−2 h

−1)

−40

−30

−20

−10

0

10

20

30

40CH4

N2O

−56 −189 −266 −279452 387 333 289 280 169 145

EH (mV)

Figure 8.6 Fluxes of CH4 and N2O from ricefields during cropping and fallows plottedagainst the corresponding soil EH (Hou et al., 2000). Reproduced by permission of SoilSci. Soc. Am.

which shows emissions of CH4 and N2O from a ricefield in northern Chinameasured from March to December for 2 years, plotted against the correspondingsoil redox potentials, EH (Hou et al., 2000). It shows that emissions of CH4 andN2O were strongly correlated with EH. Significant CH4 emission only occurred atEH < −100 mV, whereas N2O emission was only significant at EH > +200 mV.The results suggest the possibility of using management practices to maintain theredox potential in a range where both N2O and CH4 emissions are low.

8.2.3 DIFFERENCES BETWEEN RICE PRODUCTION SYSTEMS

Bronson et al. (1997a,b) made continuous measurements of CH4 and N2O emis-sions from ricefields over a period that included two dry season and one wetseason irrigated rice crops and the two intervening fallow periods. The soilwas clayey and poorly drained. Figure 8.7 shows that during the growing sea-sons, N2O fluxes were generally barely detectable although small emissions(≤ 3.5 mg N m2 day−1) occurred after N fertilizer applications. Methane fluxes,on the other hand, were substantial throughout the rice-growing seasons. Thetotal emission of CH4 over the season decreased three- to four-fold when Nwas supplied as (NH4)2SO4 rather than urea at 200 kg N ha−1, but emission ofN2O was 2.5-fold greater with (NH4)2SO4. Mid-season drainage suppressed CH4

emission by ≤ 60 %, but markedly increased N2O emissions.Figure 8.8 shows the results for the fallow periods. These lasted 5 to 11

weeks and were weedy. The soil was generally aerobic, and moderate amounts ofNO3

− accumulated (7–20 kg N ha−1). Moderately high, continuous N2O emis-sions occurred, apparently during nitrification of mineralized organic N in thetopsoil and possibly also during denitrification in the wet subsoil. The flux ofN2O was greatest immediately after rainfall and after the field was flooded forrice at the end of the fallow, as a result of denitrification of accumulated NO3

−.

Page 259: The Biogeochemistry of Submerged Soils

Nitrogen Oxides 251

CH

4 flu

x (m

g cm

−2 d

ay−1

)

0

200

400

600

800

Days after transplanting

0 20 40 60 80 100 0 20 40 60 80 100

N2O

flu

x (m

g nm

−2 d

ay−1

)

0.0

0.5

1.0

1.5

2.0

0

7

14

21

28

35

0

2

4

6

(d) urea(c) straw

(b) urea(a) straw

continuousflooding drainage

(arrows)

Figure 8.7 Emissions of CH4 and N2O during a rice crop with different water, straw andfertilizer managements. Single upward arrow = drainage; double downward arrow = floodirrigation (Bronson et al., 1997a). Reproduced by permission of Soil Sci. Soc. Am.

CH

4 flu

x (m

g cm

−2 d

ay−1

)

0

1

2

3

4

5

Geen manureStraw

Days after harvest

0 6 12 18 24 30 36 0 6 12 18 24 30 36

N2O

flu

x (m

g nm

−2 d

ay−1

)

0

20

40

60

80

0

1

2

3

4

5

Urea

0

20

40

60

80

Ammonium sulfate

Figure 8.8 Emissions of CH4 and N2O during a fallow between rice crops. Singlearrow = rainfall; double arrow = flood irrigation (Bronson et al., 1997b). Reproduced bypermission of Soil Sci. Soc. Am.

Page 260: The Biogeochemistry of Submerged Soils

252 Trace Gases

Little CH4 was emitted during the fallows. This study demonstrates that rice soilsin the fallow periods can be significant sources of N2O.

A common cropping sequence in the rainfed lowlands is wet season rice fol-lowed by a dry season upland crop on residual soil moisture or supplementalirrigation, followed by a 60- to 70-day fallow during the dry-to-wet transition.Alternate soil wetting and drying in this system create particular difficulties forthe conservation of nitrogen in the soil (Buresh et al., 1993a; George et al., 1993,1994, 1995). Soil N mineralized and nitrified at the onset of rains in the fallowmay be lost by leaching and by denitrification when the soil becomes submerged.Commonly high-value vegetable crops are grown in the dry season with heavyapplications of fertilizers, leaving substantial amounts of residual nitrate in thesoil. This situation leads to large losses of N before the wet season rice is estab-lished. Studies of N balances in an intensified rainfed lowland system of thissort in the Philippines have shown N losses of up to 550 kg ha−1 year−1 throughnitrate leaching and denitrification (Tripathi et al., 1997).

8.3 AMMONIA

8.3.1 GLOBAL BUDGET

Ammonia has a lifetime of only a few hours to a few days in the atmosphere. Itand its reaction products are transported through the atmosphere and depositedon terrestrial surfaces elsewhere. It is the main gaseous alkaline species in theatmosphere and neutralizes a large part of the acid produced in oxidation of sulfurand nitrogen oxides, probably up to a half though its dry-deposition is much fasterthan that of NOx and SO2 (Dentener and Crutzen, 1994). Dry- and wet-depositionof ammonia contribute to soil acidification because 2 mol of H+ are produced inthe nitrification of 1 mol of NH4

+. Also a large part of the ammonia depositedon moist forest soils may be re-emitted as N2O (Section 8.2).

Table 8.5 shows a global inventory of ammonia emissions compiled by Bouw-man et al. (1997). The main sources are the excreta of domestic animals (40 %),use of nitrogen fertilizers (17 %), the oceans (15 %) and biomass burning (11 %).About half of the global emission comes from Asia, and 70 % is from foodproduction. Europe, the Indian subcontinent and eastern China have the largestemission rates, reflecting the densities of domestic animals and the types andintensities of fertilizer use. Anthropogenic emissions have probably increasedthree-fold since 1950 in line with the increase in global population and foodproduction.

Ammonia volatilization from fertilizers is a function of the type of fertilizer,soil conditions, meteorological conditions–temperature, wind speed, precipita-tion–and fertilizer management. Table 8.6 shows the global use of nitrogenousfertilizers and the corresponding NH3 emissions based on empirical emission fac-tors for different fertilizer types in temperate and tropical conditions (Bouwman

Page 261: The Biogeochemistry of Submerged Soils

Ammonia 253

Table 8.5 Estimates of global ammonia emissions (Tg N year−1) from different sources

Reference Schlesinger andHartley (1992)

Dentener andCrutzen (1994)

Bouwman et al.(1997)

AnimalsCattle including buffaloes 19.9 14.2 14.0Pigs 2.0 2.8 3.4Horses, mules, asses 1.8 1.2 0.5Sheep, goats 4.1 2.5 1.5Poultry 2.4 1.3 1.9Wild animals a 2.5 0.1Total animals 32.3 24.5 21.7OthersSynthetic fertilizers 8.5 6.4 9.0Undisturbed ecosystems 10 5.1 2.4Crop plants — — 3.6Biomass burning 5 2.0 5.7Human excrement 4 — 2.6Sea surface 13 7.0 8.2Fossil fuel combustion 2.2 — 0.1Industry — — 0.2Total emission 75 45.0 53.6

a Included in undisturbed ecosystems.Source: Bouwman et al. (1997). Reproduced by permission of American Geophysical Union.

Table 8.6 Global use of nitrogenous fertilizers and corresponding NH3 emissions basedon empirical emission factors for different fertilizer types and uses

Type of nitrogenous Global use Emission (Gg N year−1)

fertilizer (Tg N year−1)

Temperate Tropical Total

Urea 29.2 1632 4137 5769Ammonium bicarbonate 9.5 802 1189 1991Ammonium nitrate 8.2 25 141 166NPK 6.6 40 219 259Anhydrous ammonia 5.2 18 190 209Nitrogen solutions 4.2 11 93 104Calcium ammonium nitrate 4.1 9 72 82Ammonium phosphates 3.7 35 113 147Ammonium sulfate 2.6 34 169 203Others 3.7 20 85 105Total 77.0 2626 6409 9035

Source: Bouwman et al. (1997). Reproduced by permission of American Geophysical Union.

et al., 1997). Seventy per cent of the emission is from developing countries inthe tropics; of this 65 % is from urea and 19 % from the volatile hydrolysis prod-uct of urea, ammonium bicarbonate, which is widely used in China. Based ondata for rice area and yield by country (IRRI, 2002), the approximate relation

Page 262: The Biogeochemistry of Submerged Soils

254 Trace Gases

between rice yield and N fertilizer use (Figure 7.1), and the emission factorsused by Bouwman et al., I estimate a total emission of NH3 from wetland riceof very roughly 3.6 Tg N year−1, of which 1.2 Tg N year−1 is from China and0.6 from India. This compares with a total global emission from N fertilizer of9.0 Tg N year−1. Clearly wetland rice is an important source of NH3.

8.3.2 PROCESSES GOVERNING AMMONIA EMISSIONS FROM RICE

Urea is the main form of N fertilizer used in rice, together with, in China, ammo-nium bicarbonate. At least two applications are generally made by broadcastingthe fertilizer onto the floodwater: the first 14–21 days after planting the cropand a second at the maximum tillering stage 45–55 days after planting. The firstis subject to high rates of loss by volatilization (De Datta and Patrick, 1986;De Datta, 1995). Losses are smaller once the crop canopy and root system areestablished, because turbulence and hence gas exchange at the water surfaceare less and absorption by the crop is greater. Rates of volatilization during theearly period measured by bulk aerodynamic and micrometeorological methodsoften account for 30–40 % and sometimes as much as 60 % of the fertilizerapplied (Simpson et al., 1984; Cai et al., 1986; Fillery et al., 1986 Freney et al.,1990; De Datta et al., 1989). Losses during the later stages are typically less thanhalf this, depending on how well matched the application is with crop demand.

Urea broadcast into the ricefield floodwater is hydrolysed to ammonium, bicar-bonate and hydroxyl ions; the reaction is catalysed by the enzyme urease:

CO(NH2)2 + 2H2O + H+ −−−→ 2 NH4+ + HCO3

One mol of H+ is consumed in this reaction for every 2 mol of NH4+ formed.

In subsequent volatilization of NH3, 1 mol of H+ is produced for every mol ofNH4

+ converted to NH3:

NH4+ −−−→ NH3 + H+

Because urease activities are much greater in the soil than in the floodwater, theNH4

+ is largely formed in the soil as the urea moves downward by mass flowand diffusion. The NH4

+, H+ and other reactants will also move between thefloodwater and soil–both upward and downward–with NH3 being lost from thefloodwater by volatilization. The recovery of N in the crop therefore depends onthe rate of movement of urea and its reaction products through the soil and onthe rate at which the roots remove N from the downward moving pool.

Rachhpal-Singh and Kirk (1993a,b) developed a model of these processesbased on equations for the transport and reaction of urea, ammoniacalspecies (NH4

+, NH3, NH4OH), carbonate species (H2CO3, HCO3−, CO3

2−)

and mobile acid–base pairs (H2CO3 –HCO3−, HCO3

− –CO32−, NH4

+ –NH3,NH4

+ –NH4OH, H2O–OH−). The equations are of the form of Equation (2.6)

Page 263: The Biogeochemistry of Submerged Soils

255

[NH

4-N

] (m

mol

dm

−3 s

oil)

Soi

l pH

6.8

7.0

7.2

7.4

7.6

7.8

[Ure

a-N

] (m

mol

dm

−3 s

oil)

01

23

45

01

23

45

67

Depth in soil (cm)

0 1 2 3 4 5

2 da

ys5

days

10 d

ays

Tim

e (d

ays)

02

46

810

%N volatilized

0102030405060

Fig

ure

8.9

Profi

les

ofur

ea-N

,am

mon

iaca

l-N

and

pHw

ithde

pth

follo

win

gbr

oadc

ast

appl

icat

ion

ofur

eaon

rice

field

flood

wat

er,

and

the

corr

espo

ndin

gra

tes

ofN

H3

vola

tiliz

atio

n(c

alcu

late

dw

ithth

em

odel

ofR

achh

pal-

Sing

han

dK

irk,

1993

a,b)

Page 264: The Biogeochemistry of Submerged Soils

256 Trace Gases

with terms for N, C and base added or removed in urea hydrolysis, organic C andN mineralization, and root uptake. The dynamics of CO2 in the floodwater andthe coupled transfer of CO2 and NH3 across the air–water interface (Section 3.5)are allowed for.

The model shows that cumulative volatilization of NH3 is sensitive to theinitial distribution of urea in the soil, its rate of hydrolysis, and the rate ofabsorption of N by rice roots. It is largely insensitive to other parameters. Forexample, it might be thought that addition of organic matter to the soil to acidifythe floodwater should lessen NH3 volatilization. However, the model shows thatalthough increased CO2 production affects the diurnal change in floodwater pH, itlittle affects the daily average pH and hence NH3 volatilization. This is becausethe relative rates of movement of carbonate species and acidity between thesoil and floodwater are such that the increased alkalinization of the floodwaterresulting from increased CO2 loss is not matched by an equal inflow of acidityfrom the soil.

The model shows that the spread of urea and NH4+ into the soil is typi-

cally only a centimetre or two in a week (Figure 8.9). The recovery of broad-cast fertilizer N in the crop must therefore depend entirely on the superficialroot system in the soil–floodwater interface. The good recovery of broadcastfertilizer N obtained if the fertilizer is added when the crop demand is maxi-mal (Peng and Cassman, 1998) therefore indicate rapid uptake by roots in thesoil–floodwater interface.

8.4 SULFUR COMPOUNDS

8.4.1 GLOBAL BUDGET

Submerged soils are important sinks for atmospheric sulfur (Howarth et al.,1992). Sulfate washed into wetlands or deposited from the atmosphere is largelyreduced to sulfide by sulfate-reducing bacteria. Subsequent precipitation withmetals, especially as FeS, results in more or less permanent removal of the Sfrom the global S cycle.

Little sulfur is re-emitted from wetlands into the atmosphere. Table 8.7 givesestimates of global emissions of volatile sulfur compounds from different sources.Total emissions are in the range 98 to 120 Tg (S) year−1; 75 % is anthropogenic,mainly from fossil fuel combustion in the northern hemisphere. The main naturalsources are the oceans and volcanoes. Wetlands and soils contribute less than3 % of the total emission.

8.4.2 EMISSIONS FROM RICEFIELDS

The main source of S emissions from ricefields is the burning of crop residues,during which most of the sulfur in the residues is converted to volatile oxides (Fox

Page 265: The Biogeochemistry of Submerged Soils

257

Tabl

e8.

7E

stim

ates

ofgl

obal

sulf

urem

issi

ons

(Tg

Sye

ar−1

)

Sour

ces

H2S

CH

3SC

H3

CS

2O

CS

SO2

SO4

Tota

la

Foss

il-fu

elco

mbu

stio

n+

indu

stry

Tota

lre

duce

dS

=2.

270

2.2

71–

77(6

8/6)

(mid

1980

s)B

iom

ass

burn

ing

<0.

01?

—<

0.01

?0.

075

2.8

0.1

2.2

–3.

0(1

.4/1

.1)

Oce

ans

<0.

315

–25

0.08

0.08

—40

–32

015

–25

(8.4

/11.

6)b

Wet

land

s0.

006

–1.

10.

003

–0.

680.

0003

–0.

06—

——

0.01

–2

(0.8

/0.2

)Pl

ants

+so

ils

0.17

–0.

530.

05–

0.16

0.02

–0.

05—

—2

–4

0.25

–0.

78(0

.3/0

.2)c

Vol

cano

es0.

5–

1.5

——

0.01

7–

82

–4

9.3

–11

.8(7

.6/3

.0)

Tota

lan

thro

poge

nic

73–

80To

tal

natu

ral

(exc

ludi

ngse

asa

ltan

dso

ildu

st)

25–

40

Tota

l98

–12

0

aN

umbe

rsin

pare

nthe

ses

are

fluxe

sfr

omno

rthe

rn/s

outh

ern

hem

isph

eres

.bE

xclu

ding

cont

ribu

tions

from

sea

salt.

cE

xclu

ding

cont

ribu

tions

from

soil

dust

.So

urce

:Se

infe

ldan

dPa

ndis

(199

8).

Rep

rodu

ced

bype

rmis

sion

ofW

iley,

New

Yor

k.

Page 266: The Biogeochemistry of Submerged Soils

258 Trace Gases

and Hue, 1986; Lefroy et al., 1992). However these may be rapidly returned tothe soil in rainfall, particularly during tropical wet seasons, though not necessarilyat the same site. Sulfur may also be lost by volatilization as H2S, but in mostsoils any S2− formed in reduction is promptly precipitated as FeS or other sulfides(Chapter 4), and so net losses are small.

There are also modest emissions of organic-S compounds. Minami et al. (1993)measured S emissions from field lysimeters treated with S-containing compoundsin amounts found in crop residues and organic manures. Emission of H2S,OCS (carbonyl sulfide), CH3SH (methyl mercaptan), CH3SCH3 (dimethylsul-fide, DMS), CS2 (carbon disulfide) and CH3SSCH3 (dimethyl disulfide) weredetected, with DMS by far the largest. Emissions of DMS ranged from 4.1 to7.3 mg (S) m−2 year−1, and varied diurnally and seasonally in ways indicatingmediation by the rice plants. The type of soil had little effect. The measured emis-sions multiplied by the total global rice area indicate a potential global emissionfrom ricefields of 0.004 to 0.01 Tg (S) year−1. This compares with emissionsfrom all wetlands of 0.003–0.68 Tg (S) year−1, from other plants and soils of0.05–0.16 Tg year−1, and from oceans of 15–25 Tg year−1. Likewise emissionsof CS2, OCS and other organic-S compounds from ricefields are small in com-parison with other known sources.

8.5 CARBON SEQUESTRATION

Table 1.3 gives estimates of the global distribution of carbon in soils. Wetlandsare the single largest organic C pool, account for roughly 40 % of the total soilcarbon. Destruction of wetlands therefore results in significant loss of the globalterrestrial carbon store. There have been large losses of wetlands in developedcountries in the past–almost half the original wetland area in the US in the last200 years for example–but this is now largely under control. However preser-vation of wetlands is less of a priority in developing countries. About half theglobal area of natural wetlands is in the tropics.

Despite the burning of crop residues in the productive, irrigated rice areas oftropical and subtropical Asia, and their removal for other purposes in the low-producing rainfed rice areas, soil carbon levels are largely constant (Bronsonet al., 1998). In any case, the amount of carbon in the shallow puddled layer ofricefields amounts to only a few per cent of the amount in natural wetlands.

Page 267: The Biogeochemistry of Submerged Soils

References

Achtnich C, Schuhmann A, Wind T, Conrad R. 1995. Role of interspecies H2 transfer tosulfate and ferric iron-reducing bacteria in acetate consumption in anoxic paddy soil.FEMS Microbiology Ecology 16: 61–69.

Ahmad AR, Nye PH. 1990. Coupled diffusion and oxidation of ferrous iron in soils. I.Kinetics of oxygenation of ferrous iron in soil suspension. Journal of Soil Science 41:395–409.

Aller RC. 1980a. Diagenesis near the sediment water interface of Long Island Sound. I.Decomposition and nutrient element geochemistry (S,N,P). Advances in Geophysics 22:235–348.

Aller RC. 1980b. Quantifying solute distributions in the bioturbated zone of marine sed-iments by defining an average microenvironment. Geochimica et Cosmochimica Acta44: 1955–1965.

Aller RC. 1994. Bioturbation and remineralization of sedimentary organic-matter–effectsof redox oscillation. Chemical Geology 114: 331–345.

Allison LE. 1947. Effects of microorganisms on the permeability of soils under prolongedsubmergence. Soil Science 63: 439–450.

Anderson IC, Levine JS. 1986. Relative rates of nitric oxide and nitrous oxide productionby nitrifiers, denitrifiers and nitrate respirers. Applied and Environmental Microbiology51: 938–945.

Arah JRM, Kirk GJD. 2000. Modeling rice plant-mediated methane emission. NutrientCycling in Agroecosystems 58: 221–230.

Arden TV. 1950. The solubility products of ferrous and ferrosic hydroxides. Journal ofthe Chemical Society 882–885.

Armentano TV, Verhoeven JTA. 1990. Biogeochemical cycles: global. In: Patten BC,et al., eds. Wetlands and Shallow Continental Water Bodies . Vol. 1. Human and NaturalRelationships. The Hague: SPB Academic Publishers, 281–311.

Armstrong J, Armstrong W, Beckett PM, Halder JE, Lythe S, Holt S, Sinclair A. 1996.Pathways of aeration and the mechanisms and beneficial effects of humidity- andVenturi-induced convections in Phragmites australis (Cav.) Trin. ex Steud. AquaticBotany 54: 177–198.

Armstrong W, Armstrong J, Beckett PM. 1990. Measurement and modelling of oxygenrelease from roots of Phragmites australis . In: Cooper SC, Findlater BC, eds. The Useof Constructed Wetlands in Water Pollution Control. Oxford: Pergamon, 41–52.

Armstrong W, Beckett PM. 1987. Internal aeration and the development of stelar anoxiain submerged roots. A multishelled mathematical model combining axial diffusion ofoxygen in the cortex with radial losses to the stele, the wall layers and the rhizosphere.New Phytologist 105: 221–245.

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 268: The Biogeochemistry of Submerged Soils

260 References

Armstrong W, Beckett PM, Justin SHFW, Lythe S. 1991a. Modelling, and other aspectsof root aeration by diffusion. In: Jackson MB, Davies DD, Lambers H, eds. Plant LifeUnder Oxygen Deprivation. Ecology, Physiology and Biochemistry. The Hague: SPBAcademic Publishers, 267–282.

Armstrong W, Cousins D, Armstrong J, Turner DW, Beckett PM. 2000. Oxygen dis-tribution in wetland plant roots and permeability barriers to gas-exchange with therhizosphere: a microelectrode and modelling study with Phragmites australis . Annalsof Botany 86: 687–703.

Armstrong W, Justin SHFW, Beckett PM, Lythe S. 1991b. Root adaptation to soil water-logging. Aquatic Botany 39: 57–73.

Aselmann I, Crutzen PJ. 1989. Global distribution of natural fresh-water wetlands and ricepaddies: their net primary productivity, seasonality and possible methane emissions.Journal of Atmospheric Chemistry 8: 307–358.

Aulakh MS, Wassmann R, Bueno C, Kreuzwieser J, Rennenberg H. 2001a. Characteriza-tion of root exudates at different growth stages of ten rice (Oryza sativa L.) cultivars.Plant Biology 3: 139–148.

Aulakh MS, Wassmann R, Rennenberg H. 2001b. Methane emissions from ricefields–Quantification, mechanisms, role of management, and mitigation options.Advances in Agronomy 70: 193–260.

Ayers RS, Westcot DW. 1989. Water Quality for Agriculture. Rome: FAO.Baas-Becking LGM, Kapala IR, Moore D. 1960. Limits of the natural environment in

terms of pH and oxidation-reduction potentials. Journal of Geology 68: 243–284.Bachelet D, Neue HU. 1993. Methane emissions from wetland rice areas of Asia.

Chemosphere 26: 219–237.Baes CF, Mesmer RE. 1976. The Hydrolysis of Cations. A Critical Review of Hydrolytic

Species and Their Stability Constants in Aqueous Solution. New York: Wiley-Interscience.

Baker CJ, Maltby E. 1995. Nitrate removal by river marginal wetlands: factors affect-ing the provision of a suitable denitrification environment. In: Hughes JMR, Heath-waite AL, eds. Hydrology and Hydrochemistry of British Wetlands. Chichester: Wiley,291–313.

Batten GD, Dowling V, Short C, Blakeney AB. 1992. Mineral content of shoots of Aus-tralian rice crops. Communications in Soil Science and Plant Analysis 23: 1195–1208.

Bautista EM, Alexander M. 1972. Reduction of inorganic compounds by microorganisms.Soil Science Society of America Proceedings 36: 918–920.

Beckett PM, Armstrong W, Justin SHFW, Armstrong J. 1988. On the relative importanceof convective and diffusive gas flows in plant aeration. New Phytologist 110: 463–468.

Bedford BL, Bouldin DR, Beliveau BD. 1991. Net oxygen and carbon dioxide balancesin solutions bathing roots of wetland plants. Journal of Ecology 79: 943–959.

Bedford BL, Walbridge MR, Aldous A. 1999. Patterns in nutrient availability and plantdiversity of temperate North American wetlands. Ecology 80: 2151–2169.

Begg CBM, Kirk GJD, Mackenzie AF, Neue HU. 1994. Root-induced iron oxidation andpH changes in the lowland rice rhizosphere. New Phytologist 128: 469–477.

Berger R, Libby WF. 1969. Equilibration of atmospheric carbon dioxide with sea water:possible enzymatic control of the arte. Science 164: 1395–1398.

Berner RA. 1980. Early Diagenesis: A Theoretical Approach. Princeton: Princeton Uni-versity Press.

Page 269: The Biogeochemistry of Submerged Soils

References 261

Bingham FT, Page AL, Mahler RL, Gange TJ. 1976. Cadmium availability to rice insludge-amended soil under ‘flood’ and ‘nonflood’ culture. Soil Science Society of Amer-ica Journal 40: 715–719.

Blesa MA, Marinovich HA, Baumgartner EC, Maroto AJG. 1987. Mechanism of dis-solution of magnetite by oxalic acid ferrous iron solutions. Inorganic Chemistry 26:3713–3717.

Bloomfield C. 1951. Experiments on the mechanism of gley formation. Journal of SoilScience 2: 196–211.

Bodelier PLE, Roslev P, Henckel T, Frenzel P. 2000. Stimulation by ammonium-basedfertilizers of methane oxidation in soil around rice roots. Nature 403: 421–424.

Bolin B, Cook RB. 1983. The Major Biogeochemical Cycles and Their Interactions. NewYork: Wiley-Interscience.

Bond WJ, Verburg K. 1997. Comparison of methods for predicting ternary exchange frombinary isotherms. Soil Science Society of America Journal 61: 444–454.

Bouldin DR. 1966. Speculations on the oxidation–reduction status of the rhizosphere ofrice roots in submerged soils. In: IAEA Technical Report 655. Vienna: IAEA, 128–139.

Bouldin DR. 1968. Models for describing the diffusion of oxygen and other mobile con-stituents across the mud–water interface. Journal of Ecology 56: 77–87.

Bouldin DR. 1989. A multiple ion uptake model. Journal of Soil Science 40: 309–319.Bouldin DR, Alimago BV. 1976. NH3 volatilization from IRRI paddies following broad-

cast applications of fertilizer nitrogen. IRRI Miscellaneous Report. Manila: InternationalRice Research Institute.

Bouwman AF, Lee DS, Asman WAH, Dentener FJ, Van der Hoek KW, Olivier JGJ. 1997.A global high-resolution emission inventory for ammonia. Global BiogeochemicalCycles 11: 561–587.

Bowden WB. 1987. The biogeochemistry of nitrogen in fresh-water wetlands. Biogeo-chemistry 4: 313–348.

Bowen HJM. 1979. Environmental Chemistry of the Elements. London: Academic Press.Brammer H. 1971. Soil Survey Project, Bangladesh: agricultural development possibili-

ties. Technical Report No. 2. Rome: FAO.Brammer H, Brinkman R. 1977. Surface-water gley soils in Bangladesh: environment,

landforms and soil morphology. Geoderma 17: 19–109.Brandon AM, Mikkelsen DS. 1979. Phosphorus transformations in alternately flooded

California soils. I. Cause of plant phosphorus deficiency in rice rotation crops andcorrection methods. Soil Science Society of America Journal 43: 989–994.

Brinkman R. 1970. Ferrolysis, a hydromorphic soil forming process. Geoderma 3:199–206.

Brinkman R. 1977a. Problem hydromorphic soils in north-east Thailand. 2. Physical andchemical aspects, mineralogy and genesis. Netherlands Journal of Agricultural Science25: 170–180.

Brinkman R. 1977b. Surface-water Gley soils in Bangladesh. Genesis. Geoderma 17:111–144.

Brinkman R. 1985. Mineralogy and surface properties of the clay fraction affecting soilbehaviour and management. In: Soil Physics and Rice. Manila: International RiceResearch Institute, 161–182.

Bronson KF, Cassman KG, Wassmann R, Olk DC, van Noordwijk M, Garrity DP. 1998.Soil carbon dynamics in different cropping systems in principal ecoregions of Asia. In:

Page 270: The Biogeochemistry of Submerged Soils

262 References

Lal R, Kimble J, Follett RF, Stewart BA, eds. Management of Carbon Sequestration inSoil. Boca Raton, FL: CRC Press, 35–57.

Bronson KF, Neue HU, Singh U, Abao EBJ. 1997a. Automated chamber measurementof methane and nitrous oxide flux in a flooded rice soil: I. Residue, nitrogen, and watermanagement. Soil Science Society of America Journal 61: 981–987.

Bronson KF, Singh U, Neue HU, Abao EBJ. 1997b. Automated chamber measurementof methane and nitrous oxide flux in a flooded rice soil: II. Fallow period emissions.Soil Science Society of America Journal 61: 988–993.

Bucher S. 2001. Nitrogen and phosphorus availability in irrigated rice as influenced bysoil drying during the fallow period, straw incorporation, and tillage. PhD thesis, ETH-Zurich.

Buresh RJ, Austin ER. 1988. Direct measurement of dinitrogen and nitrous oxide flux inflooded rice fields. Soil Science Society of America Journal 52: 681–688.

Buresh RJ, Casselman ME, Patrick WH, Jr. 1980. Nitrogen fixation in flooded soil sys-tems. Advances in Agronomy 33: 149–192.

Buresh RJ, Chua TT, Castillo EG, Liboon SP, Garrity DP. 1993a. Fallow and Sesbaniaeffects on soil nitrogen dynamics in lowland rice-based cropping systems. AgronomyJournal 85: 316–321.

Buresh RJ, Patrick WH, Jr. 1981. Nitrate reduction to ammonium and organic nitrogenin an estuarine sediment. Soil Biology and Biochemistry 13: 279–283.

Buresh RJ, Samson MI, De Datta SK. 1993b. Quantification of denitrification in floodedsoils as affected by rice establishment method. Soil Biology and Biochemistry 25:843–848.

Buresh RJ, Woodhead T, Shepherd KD, Flordelis E, Cabangon RC. 1989. Nitrate accu-mulation and loss in a mungbean/lowland rice cropping system. Soil Science Society ofAmerica Journal 53: 477–482.

Butterbach-Bahl K, Papen H, Rennenberg H. 2000. Scanning electron microscopy analy-sis of the aerenchyma in two rice cultivars. Phyton Annales rei Botanicae 40: 43–55.

Cai GX, Zhu ZL, Trevitt ACF, Freney JR, Simpson JR. 1986. Nitrogen loss from ammo-nium bicarbonate and urea fertilizers applied to flooded rice. Fertiliser Research 10:203–215.

Cao MK, Gregson K, Marshall S. 1998. Global methane emission from wetlands and itssensitivity to climate change. Atmospheric Environment 32: 3293–3299.

Cao MK, Gregson K, Marshall S, Dent JB, Heal OW. 1996. Global methane emissionsfrom rice paddies. Chemosphere 33: 879–897.

Cao XY, Jiang XM, Kareem A, Dou ZH, Rakeman MA, Zhang ML, Ma T, O’Donnell K,DeLong N, DeLong GR. 1994. Iodination of irrigation water as a method of supply-ing iodine to a severely iodine-deficient population in Xinjiang, China. Lancet 344:107–110.

Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann A, Singh U. 1998.Opportunities for increased nitrogen use efficiency from improved resource managementin irrigated rice systems. Field Crops Research 56: 7–38.

Catling D. 1992. Rice in Deep Water. London: MacMillan.Cayton MT, Reyes ED, Neue HU. 1985. Effect of zinc fertilization on the mineral nutri-

tion of rices differing in tolerance to zinc deficiency. Plant and Soil 87: 319–327.Cayton MTC. 1985. Boron toxicity in rice. IRRI Research Paper Series No. 113 . Manila:

International Rice Research Institute.

Page 271: The Biogeochemistry of Submerged Soils

References 263

Childers DL, Day JW, McKellar HN. 2000. Twenty more years of marsh and estuarineflux studies: revisiting Nixon (1980). In: Weinstein MP, Kreeger DA, eds. Conceptsand Controversies in Tidal Marsh Ecology. Dordrecht: Kluwer, 389–421.

Cicerone RJ, Shetter JD. 1981. Sources of atmospheric methane: measurements in ricepaddies and a discussion. Journal of Geophysical Research 86: 7203–7209.

Cicerone RJ, Shetter JD, Delwiche CC. 1983. Seasonal variation of methane flux from aCalifornia rice paddy. Journal of Geophysical Research 88: 11022–11024.

Clymo RS. 1984. The limits to peat bog growth. Philosophical Transactions of the RoyalSociety of London 303B: 605–654.

Conrad R. 1989. Control of methane production in terrestrial ecosytems. In: Andreae MO,Schimel DS, eds. Exchange of Trace Gases Between Terrestrial Ecosystems and theAtmosphere. Dahlem Conference. Chichester: Wiley, 39–58.

Conrad R. 1993. Mechanisms controlling methane emission from wetland rice fields. In:Oremland R, ed. Biogeochemistry of Global Change. Radiatively Active Gases. NewYork: Chapman & Hall, 317–355.

Conrad R, Frenzel P. 2002. Flooded soils. In: Bitton G, ed. Encyclopedia of Environmen-tal Microbiology. New York: John Wiley & Sons, 1316–1333.

Conrad R, Schink B, Phelps TJ. 1986. Thermodynamics of H2-producing and H2-con-suming metabolic reactions in diverse methanogenic environments under in-situ con-ditions. FEMS Microbiology Ecology 38: 353–360.

Craft CB. 2001. Biology of wetland soils. In: Richardson JL, Vepraskas MJ, eds. Wet-land soils: Genesis, Hydrology, Landscapes, and Classification. Boca Raton, FL: LewisPublishers, 107–135.

Crank J, McFarlane NR, Newby JC, Paterson GD, Pedley JB. 1981. Diffusion Processesin Environmental Systems. London: MacMillan.

Crisp DT. 1966. Input and output of minerals for an area of Pennine moorland: theimportance of precipitation, drainage, peat erosion and animals. Journal of AppliedEcology 3: 327–348.

Dacey JWH. 1980. Internal winds in water lilies: an adaptation for life in anaerobicsediments. Science 210: 1017–1019.

Dacey JWH. 1981. Pressurized ventilation in the yellow water lily. Ecology 62:1137–1147.

Damgaard LR, Revsbech NP, Reichardt W. 1998. Use of an oxygen-insensitivemicroscale biosensor for methane to measure methane concentration profiles in a ricepaddy. Applied and Environmental Microbiology 64: 864–870.

Davis RB. 1974. Tubificids alter profiles of redox and pH in profundal lake sediment.Limnology and Oceanography 19: 342–346.

De Datta SK. 1995. Nitrogen transformations in wetland rice ecosystems. FertiliserResearch 42: 193–203.

De Datta SK, Buresh RJ. 1989. Integrated nitrogen management in irrigated rice. Advancesin Soil Science 10: 143–169.

De Datta SK, Patrick WH, Jr, eds. 1986. Nitrogen Economy of Flooded Rice Soils. Dor-drecht: Martinus Nijhoff.

De Datta SK, Trevitt ACF, Freney JR, Obcemea WN, Real JG, Simpson JR. 1989. Mea-suring nitrogen losses from lowland rice using bulk aerodynamic and nitrogen-15balance methods. Soil Science Society of America Journal 53: 1275–1281.

Page 272: The Biogeochemistry of Submerged Soils

264 References

De Gee JC. 1950. Preliminary oxidation-reduction potential determination in a ‘Sawah’profile near Bogor (Java). In: Transactions of the 4th International Congress of SoilScience. Amsterdam: International Society of Soil Science, 300–303.

Deer WA, Howie RA, Zussman J. 1992. An Introduction to the Rock-Forming Minerals.Harlow: Longman Scientific and Technical.

Degens BP, Harris JA. 1997. Development of a physiological approach to measuring thecatabolic diversity of soil microbial communities. Soil Biology and Biochemistry 29:1309–1320.

Dent D. 1986. Acid Sulphate Soils: A Baseline for Research and Development. Wagenin-gen: International Institute for Land Reclamation and Improvement.

Dentener FJ, Crutzen PJ. 1994. A three-dimensional model of the global ammonia cycle.Journal of Atmospheric Chemistry 19: 331–369.

Devai I, Delaune RD. 1995. Evidence for phosphine production and emission fromLouisiana and Florida marsh soils. Organic Geochemistry 23: 277–279.

Dixon RK, Brown S, Houghton RA, Solomon AM, Trexler MC, Wisniewski J. 1994.Carbon pools and flux of global forest ecosystems. Science 263: 185–190.

Dobermann A, Cassman KG, Mamaril CP, Sheehy JE. 1998. Management of phosphorus,potassium and sulfur in intensive, irrigated lowland rice. Field Crops Research 56:113–138.

Dobermann A, Dawe D, Roetter R, Cassman KG. 2000. Reversal of yield decline in along-term continuous cropping experiment. Agronomy Journal 92: 633–643.

Dobermann A, Fairhurst T. 2000. Rice Nutrient Disorders and Nutrient Management.Manila: Institute and International Rice Research Institute.

Dobermann A, Santa Cruz PC, Cassman KG. 1996. Fertilizer inputs, nutrient balance, andsoil nutrient-supplying power in intensive, irrigated rice systems. I. Potassium uptakeand K balance. Nutrient Cycling in Agroecosystems 46: 1–10.

Dobermann A, Witt C, Dawe D, Abdulrachman S, Gines HC, Nagarajan R, Satawa-thananont S, Son TT, Tan PS, Wang GH, Thoa VTK, Phung CV, Stalin P, Muthukr-ishnan P, Ravi V, Babu M, Chatuporn S, Sookthongsa J, Sun Q, Fu R, Simbahan G,Adviento MA. 2002. Site-specific nutrient management for intensive rice cropping sys-tems in Asia. Field Crops Research 74: 37–66.

Dogar MA, Hai TV. 1980. Effect of P, N and HCO−3 levels in the nutrient solution on

rate of Zn absorption by rice roots and Zn content in plants. Zeitschrift fur Pflanzen-physiologie 98: 203–212.

Dost H, van Breemen N, eds. 1983. Proceedings of the Bangkok Symposium on AcidSulphate Soils. Wageningen: International Institute for Land Reclamation and Improve-ment.

Drew MC, Saker LR. 1986. Ion transport to the xylem in aerenchymatous roots of Zeamays (L.). Journal of Experimental Botany 37: 22–33.

Driessen PM, Moormann FR. 1985. Soils on which rice-based cropping systems are prac-ticed. In: Soil Physics and Rice. Manila: International Rice Research Institute, 45–56.

Drits VA, Manceau A. 2000. A model for the mechanism of Fe3+ to Fe2+ reduction indioctahedral smectites. Clays and Clay Minerals 48: 185–195.

Eary LE, Rai. 1987. Kinetics of chromium (III) oxidation to chromium (VI) by reactionwith manganese oxide. Environmental Science and Technology 21: 1187–1193.

Elgawhary SM, Lindsay WL, Kemper WD. 1970. Effect of complexing agent and acidson the diffusion of Zn to a simulated root. Soil Science Society of America Proceedings34: 211–214.

Page 273: The Biogeochemistry of Submerged Soils

References 265

Ellis JR. 1998. Post flood syndrome and vesicular-arbuscular mycorrhizal fungi. Journalof Production Agriculture 11: 200–204.

Emerson S. 1975. Chemically enhanced CO2 gas exchange in a eutrophic lake: a generalmodel. Limnology and Oceanography 20: 743–753.

Epp MA, Chanton JP. 1993. Rhizospheric methane oxidation determined via the methyl-fluoride inhibition technique. Journal of Geophysical Research-Atmospheres 98:18413–18422.

Evans LT. 1993. Crop Evolution, Adaptation and Yield. Cambridge: Cambridge UniversityPress.

Evans LT. 1997. Adapting and improving crops: the endless task. Philosophical Transac-tions of the Royal Society of London Series B 352: 901–906.

FAO 1990. FAO-UNESCO Soil Map of the World. Revised Legend. Soils Bulletin 60. Rome:FAO.

Favre F, Tessier D, Abdelmoula M, Genin JM, Gates WP, Boivin P. 2002. Iron reductionand changes in cation exchange capacity in intermittently waterlogged soil. EuropeanJournal of Soil Science 53: 175–183.

Fillery IRP, Simpson JR, De Datta SK. 1986. Contribution of ammonia volatilization tototal nitrogen loss after applications of urea to wetland rice fields. Fertiliser Research8: 193–202.

Fitter AH, Hay RKM. 2002. Environmental Physiology of Plants. San Diego: AcademicPress.

Flessa H, Fischer WR. 1993. Plant-induced changes in the redox potential of rice rhizo-spheres. Plant and Soil 143: 55–60.

Flowers TJ, Koyama ML, Flowers SA, Sudhakar C, Singh KP, Yeo AR. 2000. QTL: theirplace in engineering tolerance of rice to salinity. Journal of Experimental Botany 51:99–106.

Flowers TJ, Yeo AR. 1981. Variability in the resistance to sodium chloride salinity withinrice (Oryza sativa L.) varieties. New Phytologist 88: 363–373.

Forno DA, Asher CJ, Yoshida S. 1975. Zinc deficiency in rice. II. Studies on two varietiesdiffering in susceptibility to zinc deficiency. Plant and Soil 42: 551–563.

Fowler BA. 1983. Biological and Environmental Aspects of Arsenic. Amsterdam: Elsevier.Fox RL, Hue NV. 1986. Sulfur cycling in the tropics and sulfur requirements for agricul-

ture. In: Portch S, Hussain SG, eds. Sulfur in Agricultural Soils. Washington, DC: TheSulfur Institute, 139–162.

Freney JR, Trevitt ACF, De Datta SK, Obcemea WN, Real JG. 1990. The relative impor-tance of denitrification and ammonia volatilization as loss processes in flooded rice inthe Philippines. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft 60, 211–216.

Frost T, Upstill-Goddard RC. 1999. Air–sea gas exchange into the millennium: Progressand uncertainties. Oceanography and Marine Biology 37: 1–45.

Fry JC. 1982. Interactions between bacteria and benthic invertebrates. In: Nedwell DB,Brown CM, eds. Sediment Microbiology. London: Academic Press, 171–202.

Fuge R. 1996. Geochemistry of iodine in relation to deficiency diseases. In Appleton JD,Fuge R, McCall GJH, eds. Environmental Geochemistry and Health. London: Geolog-ical Society, 201–212.

Fung I, John J, Lerner J, Matthews E, Prather M, Steele LP, Fraser PJ. 1991. 3-Dimen-sional model synthesis of the global methane cycle. Journal of Geophysical Research-Atmospheres 96: 13033–13065.

Garrels RM, Christ CLA. 1965. Solution. New York: Harper and Row.

Page 274: The Biogeochemistry of Submerged Soils

266 References

Gates WP, Stucki JW, Kirkpatrick RJ. 1996. Structural properties of reduced Upton mont-morillonite. Physics and Chemistry of Minerals 23: 535–541.

Gehring AU, Fry IV, Luster J, Sposito G. 1993. The chemical form of vanadium(IV) inkaolinite. Clays and Clay Minerals 41: 662–667.

George T, Ladha JK, Buresh RJ, Garrity DP. 1993. Nitrate dynamics during the aerobicsoil phase in lowland rice- based cropping systems. Soil Science Society of AmericaJournal 57: 1526–1532.

George T, Ladha JK, Garrity DP, Buresh RJ. 1994. Legumes as nitrate catch crops dur-ing the dry-to-wet transition in lowland rice cropping systems. Agronomy Journal 86:267–273.

George T, Ladha JK, Garrity DP, Torres R. 1995. Nitrogen dynamics of grain legume–weedy fallow–flooded rice sequences in the tropics. Agronomy Journal 87: 1–6.

Gilbert B, Assmus B, Hartmann A, Frenzel P. 1998. In situ localization of two methan-otrophic strains in the rhizosphere of rice plants. FEMS Microbiology Ecology 25:117–128.

Giordano PM, Mortvedt JJ. 1974. Response of several rice cultivars to zinc. AgronomyJournal 66: 220–223.

Glass ADM. 1973. Influence of phenolic acids on ion uptake. I. Inhibition of phosphateuptake. Plant Physiology 51: 1037–1041.

Glasstone S, Laidler KJ, Eyring H. 1941. The Theory of Rate Processes. New York:McGraw-Hill.

Graham D, Reed ML, Patterson BD, Hockley DG. 1984. Chemical properties, distribu-tion, and physiology of plant and algal carbonic anhydrase. Annals of the New YorkAcademy of Science 429: 222–237.

Grant IF, Roger PA, Watanabe I. 1986. Ecosystem manipulation for increasing biologi-cal N2 fixation by blue-green algae (Cyanobacteria) in lowland rice fields. BiologicalAgriculture & Horticulture 3: 299–315.

Greenland DJ. 1981. Recent progress in studies of soil structure, and its relation to prop-erties and management of paddy soils. In: Proceedings of Symposium on Paddy Soil.Berlin: Springer-Verlag, 42–58.

Greenland DJ. 1997. The Sustainability of Rice Farming. Wallingford: CAB International.Gregorio GB, Senadhira D, Mendoza RD, Manigbas NL, Roxas JP, Guerta CQ. 2002.

Progress in breeding for salinity tolerance and associated abiotic stresses in rice. FieldCrops Research 76: 91–101.

Guerra LC, Bhuiyan SI, Tuong TP, Barker R. 1998. Producing more rice with less waterfrom irrigated systems. IRRI Discussion Paper Series No 29. Manila: International RiceResearch Institute.

Hansel CM, Fendorf S, Sutton S, Newville M. 2001. Characterization of Fe plaque andassociated metals on the roots of mine-waste impacted aquatic plants. EnvironmentalScience & Technology 35: 3863–3868.

Hansen J, Sato M, Ruedy R, Lacis A, Oinas V. 2000. Global warming in the twenty-firstcentury: An alternative scenario. Proceedings of the National Academy 97: 9875–9880.

Harrison WH, Aiyer PAS. 1920. The gases of swamp rice soils. I. Their composition andrelationship to the crop. Memoirs of Department of Agriculture of India 3: 65–104.

Heimann M, Kaminski T. 1999. Inverse modelling approaches to infer surface trace gasfluxes from observed atmospheric mixing ratios. In: Bouwman AF, ed. Approaches toScaling of Trace Gas Fluxes in Ecosystems. Amsterdam: Elsevier.

Page 275: The Biogeochemistry of Submerged Soils

References 267

Hein R, Crutzen PJ, Heimann M. 1997. An inverse modeling approach to investigate theglobal atmospheric methane cycle. Global Biogeochemical Cycles 11: 43–76.

Holzapfel-Pschorn A, Conrad R, Seiler W. 1985. Production, oxidation and emission ofmethane in rice paddies. FEMS Microbiology Ecology 31: 343–351.

Holzapfel-Pschorn A, Conrad R, Seiler W. 1986. Effects of vegetation on the emissionof methane from submerged paddy. Plant and Soil 92: 223–233.

Holzapfel-Pschorn A, Seiler W. 1986. Methane emission during a cultivation period froman Italian rice paddy. Journal of Geophysical Research-Atmospheres 91: 1803–1814.

Horton DK, Frans RE, Cothren RT. 1983. MSMA-induced straighthead in rice (Oryza-sativa) and effect upon metabolism and yield. Weed Science 31: 648–651.

Hou AX, Chen GX, Wang ZP, van Cleemput O, Patrick WH. 2000. Methane and nitrousoxide emissions from a rice field in relation to soil redox and microbiological processes.Soil Science Society of America Journal 64: 2180–2186.

Houghton RA, Skole DL. 1993. Carbon. In: Turner BL, Clark WC, Kates RW,Richards JF, Mathews JT, Meyer WB, eds. The Earth as Transformed by HumanAction: Global and Regional Changes in the Biosphere Over the Past 300 Years.Cambridge: Cambridge University Press.

Houweling S, Kaminski T, Dentener F, Lelieveld J, Heimann M. 1999. Inverse modelingof methane sources and sinks using the adjoint of a global transport model. Journal ofGeophysical Research-Atmospheres 104: 26137–26160.

Howarth RW, Stewart JWB, Ivanov MV. 1992. Sulfur Cycling on the Continents: Wet-lands, Terrestrial Ecosystems and Associated Water Bodies. Chichester: Wiley.

Howeler RH, Bouldin DR. 1971. The diffusion and consumption of oxygen in submergedsoils. Soil Science Society of America Proceedings 36: 202–208.

Huang Y, Sass RL, Fisher FM. 1998. Model estimates of methane emission from irrigatedrice cultivation of China. Global Change Biology 4: 809–821.

Huang Y-M. 1990. The effects of precipitation of calcium carbonate on soil pH followingurea application. DPhil thesis, University of Oxford.

Huguenin-Elie O, Kirk GJD, Frossard E. 2003. Phosphorus uptake by rice from soil thatis flooded, drained or flooded then drained. European Journal of Soil Science 54: 77–90.

Huke RE, Huke EH. 1997. Rice Area by Type of Culture: South, Southeast, and East Asia.A Revised and Updated Database. Manila: International Rice Research Institute.

Ilag LL, Rosales AM, Elazegui FA, Mew TW. 1987. Changes in the population of infec-tive endomycorrhizal fungi in a rice-based cropping system. Plant and Soil 103: 67–73.

Inskeep WP, Bloom PR. 1986. Kinetics of calcite precipitation in the presence of water-soluble organic ligands. Soil Science Society of America Journal 50: 1167–1172.

IPCC. 1997. Revised IPCC Guidelines for National Inventories of Sources and Sinks ofGreenhouse Gases. Workbook. Paris: OECD.

IRRI. 1964. 1964 Annual Report. Manila: International Rice Research Institute.IRRI. 1965. 1965 Annual Report. Manila: International Rice Research Institute.IRRI. 2002. Rice Almanac, 3rd edn. Manila: International Rice Research Institute.Jackson MB, Armstrong W. 1999. Formation of aerenchyma and the processes of plant

ventilation in relation to soil flooding and submergence. Plant Biology 1: 274–287.Jeffrey JWO. 1961. Defining the state of reduction of a paddy soil. Journal of Soil Science

12: 172–179.Jiang XM, Cao XY, Jiang JY, Ma T, James DW, Rakeman MA, Dou ZH, Mamette M,

Amette K, Zhang ML, DeLong GR. 1997. Dynamics of environmental supplementation

Page 276: The Biogeochemistry of Submerged Soils

268 References

of iodine: Four years’ experience of iodination of irrigation water in Hotien, Xinjiang,China. Archives of Environmental Health 52: 399–408.

Katyal JC, Vlek PLG. 1985. Micronutrient problems in tropical Asia. Fertilizer Research7: 69–94.

Keller JB. 1980. Liesegang rings: theory of fast reaction and slow diffusion. In: StewartWE, Ray WH, Conley CC, eds. Dynamics and Modelling of Reactive Systems. NewYork: Academic Press, 211–224.

Kern JS, Gong ZT, Zhang GL, Zhuo HZ, Luo GB. 1997. Spatial analysis of methaneemissions from paddy soils in China and the potential for emission reduction. NutrientCycling in Agroecosystems 49: 181–195.

Kiene RP. 1991. Production and consumption of methane in aquatic systems. In:Rogers JE, Whitman WB, eds. Microbial Production and Consumption of GreenhouseGases: Methane, Nitrogen Oxides, and Halomethanes. Washington, DC: AmericanSociety for Microbiology, 111–146.

Kikuchi E, Furusaka C, Kurihara Y. 1975. Surveys of the fauna and flora in the water andsoil of paddy fields. Report of the Institute of Agricultural Research. Tohuku: TohukuUniversity.

Kikuchi E, Kurihara Y. 1977. In vitro studies on the effects of tubificids on the biolog-ical, chemical, and physical characteristics of submerged ricefield soil and overlyingfloodwater. Oikos 29: 348–356.

Kikuchi E, Kurihara Y. 1982. The effects of the oligochaete Branchiura sowerbi Beddard(Tubificdae) on the biological and chemical characteristics of overlying floodwater andsoil in a submerged ricefield soil system. Hydrobiologia 97: 203–208.

Kirk GJD. 2003. Rice root properties for internal aeration and efficient nutrient acquisitionin submerged soil. New Phytologist 159: 185–194.

Kirk GJD, Ahmad AR, Nye PH. 1990a. Coupled diffusion and oxidation of ferrous ironin soils. II. A model of the diffusion and reaction of O2, Fe2+, H+ and HCO3

− in soilsand a sensitivity analysis of the model. Journal of Soil Science 41: 411–431.

Kirk GJD, Bajita JB. 1995. Root-induced iron oxidation, pH changes and zinc solubiliza-tion in the rhizosphere of lowland rice. New Phytologist 131: 129–137.

Kirk GJD, Bouldin DR. 1991. Speculations on the operation of the rice root systemin relation to nutrient uptake. In: Penning de Vries FWT, van Laar HH, Kropff MJ,eds. Simulation and Systems Analysis for Rice Production (SARP). Wageningen: Pudoc,195–203.

Kirk GJD, Kronzucker HJ. 2000. Nitrogen uptake by rice roots. In: Kirk GJD, Olk DC,eds. Carbon and Nitrogen Dynamics in Flooded Soils. Manila: International RiceResearch Institute, 147–162.

Kirk GJD, Le van Du. 1997. Changes in rice root architecture, porosity, and oxygen andproton release under phosphorus deficiency. New Phytologist 135: 191–200.

Kirk GJD, Rachhpal-Singh. 1992. Simultaneous transfer across an air-water interfaceof gases that interact through acid–base reactions. Atmospheric Environment 26A:1651–1660.

Kirk GJD, Saleque MA. 1995. Solubilization of phosphate by rice plants growing inreduced soil: prediction of the amount solubilized and the resultant increase in uptake.European Journal of Soil Science 46: 247–255.

Kirk GJD, Santos EE, Santos MB. 1999. Phosphate solubilization by organic anion excre-tion from rice growing in aerobic soil: rates of excretion and decomposition, effects on

Page 277: The Biogeochemistry of Submerged Soils

References 269

rhizosphere pH and effects on phosphate solubility and uptake. New Phytologist 142:185–200.

Kirk GJD, Solivas JL. 1994. Coupled diffusion and oxidation of ferrous iron in soils: III.Further development of the model and experimental testing. European Journal of SoilScience 45: 369–378.

Kirk GJD, Solivas JL. 1997. On the extent to which root properties and transport throughthe soil limit nitrogen uptake by lowland rice. European Journal of Soil Science 48:613–621.

Kirk GJD, Solivas JL, Alberto MC. 2003. The effects of redox conditions on solute dif-fusion in soil. European Journal of Soil Science 54: 617–624.

Kirk GJD, Yu T-R, Choudhury FA. 1990b. Phosphorus chemistry in relation to waterregime. In: Phosphorus Requirements for Sustainable Agriculture in Asia and Oceania.Manila: International Rice Research Institute, 211–223.

Knox JW, Matthews RB, Wassmann R. 2000. Using a crop/soil simulation model andGIS techniques to assess methane emissions from rice fields in Asia. III. Databases.Nutrient Cycling in Agroecosystems 58: 179–199.

Komadel P, Madejova J, Stucki JW. 1995. Reduction and reoxidation of nontronite: ques-tions of reversibility. Clays and Clay Minerals 43: 105–110.

Kostka JE, Luther GWI. 1994. Partitioning and speciation of solid phase iron in saltmarshsediments. Geochimica et Cosmochimica Acta 58: 1701–1710.

Kostka JE, Wu J, Nealson KH, Stucki JW. 1999. The impact of structural Fe(III) reduc-tion by bacteria on the surface chemistry of smectite clay minerals. Geochimica etCosmochimica Acta 63: 3705–3713.

Koyama T. 1963. Gaseous metabolism in lake sediments and paddy soils and the pro-duction of atmospheric methane and hydrogen. Journal of Geophysical Research 68:3971–3973.

Kroeze C, Mosier A, Bouwman L. 1999. Closing the global N2O budget: A retrospectiveanalysis 1500–1994. Global Biogeochemical Cycles 13: 1–8.

Kronzucker HJ, Glass ADM, Siddiqi MY, Kirk GJD. 2000. Comparative kinetic analysisof ammonium and nitrate acquisition by tropical lowland rice: implications for ricecultivation and yield potential. New Phytologist 145: 471–476.

Kronzucker HJ, Kirk GJD, Siddiqi MY, Glass ADM. 1998a. Effects of hypoxia on13NH4

+ fluxes in rice roots. Plant Physiology 116: 581–587.Kronzucker HJ, Schjoerring JK, Erner Y, Kirk GJD, Siddiqi MY, Glass ADM. 1998b.

Dynamic interactions between root NH4+ influx and long-distance N translocation in

rice: insights into feedback processes. Plant and Cell Physiology 39: 1287–1293.Kronzucker HJ, Siddiqi MY, Glass ADM, Kirk GJD. 1999. Nitrate-ammonium synergism

in rice. A subcellular flux analysis. Plant Physiology 119: 1041–1045.Krylova NI, Conrad R. 1998. Thermodynamics of propionate degradation in methanogenic

paddy soil. FEMS Microbiology Ecology 26: 281–288.Kyuma K. 1978. Mineral composition of rice soils. In: Soils and Rice. Manila: Interna-

tional Rice Research Institute, 219–235.Kyuma K. 2003. Paddy Soil Science. Kyoto: Kyoto University Press.Ladha JK, Dawe D, Ventura TS, Singh U, Ventura W, Watanabe I. 2000. Long-term

effects of urea and green manure on rice yields and nitrogen balance. Soil ScienceSociety of America Journal 64: 1993–2001.

Page 278: The Biogeochemistry of Submerged Soils

270 References

Landolt H, Bornstein R, Borschers H, Bartels J, Eucken A. 1960. Landolt-BornsteinZahlenwerte und Funktionen aus Physik, Chemie, Astronomie, Geophysik, Technik.Berlin: Springer.

Layzell DB, Turpin DH. 1990. Plant Physiology, Biochemistry and Molecular Biology.Harlow: Longman Scientific & Technical.

Lefroy RDB, Mamaril CP, Blair GJ, Gonzales PB. 1992. Sulfur cycling in rice wetlands.In: Howarth RW, Stewart JWB, Ivanov MV, eds. Sulfur Cycling on the Continents.New York: Wiley, 279–299.

Lelieveld J, Crutzen PJ, Dentener FJ. 1998. Changing concentration, lifetime and climateforcing of atmospheric methane. Tellus Series B-Chemical and Physical Meteorology50: 128–150.

Lerman A. 1979. Geochemical Processes. Water and Sediment Environments. New York:Wiley-Interscience.

Li Y-H, Gregory S. 1974. Diffusion of ions in sea water and in deep-sea sediments.Geochimica et Cosmochimica Acta 38: 703–714.

Liesack W, Schnell S, Revsbech NP. 2000. Microbiology of flooded rice paddies. FEMSMicrobiology Reviews 24: 625–645.

Lindsay WL. 1979. Chemical Equilibria in Soils. New York: Wiley.Liss PS, Slater PG. 1974. Flux of gases across the air-sea interface. Nature 247: 181–184.Lovley DR. 1991. Dissimilatory reduction of Fe(III) and Mn(IV). Microbiological Reviews

55: 259–287.Lovley DR. 1993. Dissimilatory metal reduction. Annual Review of Microbiology 47:

263–290.Lovley DR. 1997. Microbial Fe(III) reduction in subsurface environments. FEMS Micro-

biology Reviews 20: 305–313.Lovley DR, Goodwin S. 1987. Hydrogen concentration as an indicator of the predominant

terminal electron-accepting reactions in aquatic sediments. Geochimica et Cosmochim-ica Acta 52: 2953–3003.

Lu Y, Wassmann R, Neue HU, Huang C. 1999. Impact of phosphorus supply on rootexudation, aerenchyma formation and methane emission of rice plants. Biogeochemistry47: 203–218.

Mackowiak CL, Grossl PR. 1999. Iodate and iodide effects on iodine uptake and partition-ing in rice (Oryza sativa L.) grown in solution culture. Plant and Soil 212: 135–143.

Mahieu N, Olk DC, Randall EW. 2000a. Accumulation of heterocyclic nitrogen in humi-fied organic matter: a N-15-NMR study of lowland rice soils. European Journal of SoilScience 51: 379–389.

Mahieu N, Olk DC, Randall EW. 2000b. Analysis of phosphorus in two humic acid frac-tions of intensively cropped lowland rice soils by P-31-NMR. European Journal of SoilScience 51: 391–402.

Mahieu N, Olk DC, Randall EW. 2002. Multinuclear magnetic resonance analysis of twohumic acid fractions from lowland rice soils. Journal of Environmental Quality 31:421–430.

Mangold DC, Tsang C-F. 1991. A summary of subsurface hydrological and hydrochem-ical models. Reviews of Geophysics 29: 51–79.

Marin AR, Masscheleyn PH, Patrick WH, Jr. 1993. Soil redox-pH stability of arsenicspecies and its influence on arsenic uptake by rice. Plant and Soil 152: 245–253.

Marschner H. 1995. Mineral Nutrition of Higher Plants. Academic Press, London.

Page 279: The Biogeochemistry of Submerged Soils

References 271

Mason RP, Fitzgerald WF, Hurley J, Hanson AKJr, Donaghay PL, Sieburth JM. 1993.Mercury biogeochemical cycling in a stratified estuary. Limnology and Oceanography38: 1227–1241.

Mason RP, Fitzgerald WF, Morel FMM. 1994. The biogeochemical cycling of elementalmercury: anthropogenic influences. Geochimica et Cosmochimica Acta 58: 3193–3198.

Matsuo T, Hoshikawa K. 1993. Science of the Rice Plant. I. Morphology. Tokyo: Foodand Agriculture Policy Research Center.

Matthews RB, Wassmann R, Arah JRM. 2000a. Using a crop/soil simulation model andGIS techniques to assess methane emissions from rice fields in Asia. I. Model devel-opment. Nutrient Cycling in Agroecosystems 58: 141–159.

Matthews RB, Wassmann R, Knox JW, Buendia LV. 2000b. Using a crop/soil simulationmodel and GIS techniques to assess methane emissions from rice fields in Asia. IV.Upscaling to national levels. Nutrient Cycling in Agroecosystems 58: 201–217.

McBride MB. 1994. Environmental Chemistry of Soils. New York: Oxford UniversityPress.

McGillis WR, Edson JB, Ware JD, Dacey JWH, Hare JE, Fairall CW, Wanninkhof R.2001. Carbon dioxide flux techniques performed during GasEx-98. Marine Chemistry75: 267–280.

Meharg AA, Hartley-Whitaker J. 2002. Arsenic uptake and metabolism in arsenic resistantand nonresistant plant species. New Phytologist 154: 29–43.

Melchior DC, Bassett RL. 1990. Chemical Modeling of Aqueous Systems II. Washington,DC: American Chemical Society.

Mikkelsen DS, De Datta SK, Obcemea WN. 1978. Ammonia volatilization losses fromflooded rice soils. Soil Science Society of America Journal 42: 725–730.

Miller SP. 2000. Arbuscular mycorrhizal colonization of semi-aquatic grasses along awide hydrologic gradient. New Phytologist 145: 145–155.

Minami K, Kanda K-I, Tsuruta H. 1993. Emission of biogenic sulfur gases from ricepaddies in Japan. In: Oremland R, ed. Biogeochemistry and Global Change. New York:Chapman & Hall, 403–418.

Mitsch WJ, Gosselink JG. 2000. Wetlands. New York: Wiley.Moore PD, Bellamy DJ. 1974. Peatlands. London: Elek Science.Moormann FR, van Breemen N. 1978. Rice: Soil, Water, Land. Manila: International Rice

Research Institute.Morel FMM, Herring JG. 1993. Principles and Applications of Aquatic Chemistry. New

York: Wiley-Interscience.Mosier AR, Duxbury JM, Freney JR, Heinemeyer O, Minami K. 1998a. Assessing and

mitigating N2O emissions from agricultural soils. Climatic Change 40: 7–38.Mosier AR, Duxbury JM, Freney JR, Heinemeyer O, Minami K, Johnson DE. 1998b.

Mitigating agricultural emissions of methane. Climatic Change 40: 39–80.Neue HU. 1997. Fluxes of methane from rice fields and potential for mitigation. Soil Use

and Management 13: 258–267.Neue HU, Becker-Heidmann P, Scharpenseel HW. 1990. Organic matter dynamics, soil

properties, and cultural practices in rice lands and their relationship to methane pro-duction. In: Bouwman AF, ed. Soils and the Greenhouse Effect. New York: Wiley,457–466.

Neue HU, Bloom PR. 1989. Nutrient kinetics and availability in flooded rice soils. In:Progress in Irrigated Rice Research. Manila: International Rice Research Institute,173–190.

Page 280: The Biogeochemistry of Submerged Soils

272 References

Neue HU, Lantin RS. 1994. Micronutrient toxicities and deficiencies in rice. In: Yeo AR,Flowers TJ, eds. Soil Mineral Stresses: Approaches to Crop Improvement. Berlin:Springer-Verlag, 175–200.

Neue HU, Roger PA. 1994. Potential of methane emission in major rice ecologies. In:Zepp RG, ed. Climate Biosphere Interaction. New York: Wiley, 65–93.

Neue HU, Sass RL. 1998. The budget of methane from ricefields. IGACtivities Newsletter12: 3–11.

Neue HU, Scharpenseel HW. 1987. Decomposition pattern of 14C-labeled rice straw inaerobic and submerged soils of the Philippines. Science of the Total Environment 62:431–434.

Neue HU, Wassmann R, Kludze HK, Wang B, Lantin RS, Wang BJ. 1997. Factors andprocesses controlling methane emissions from rice fields. Nutrient Cycling in Agroe-cosystems 49: 111–117.

Nevin KP, Lovley DR. 2002. Mechanisms for accessing insoluble Fe(III) oxide duringdissimilatory Fe(III) reduction by Geothrix fermentans . Applied and EnvironmentalMicrobiology 68: 2294–2299.

Nickson R, McArthur J, Burgess W, Ahmed KM, Ravenscroft P, Rahman M. 1998.Arsenic poisoning of Bangladesh groundwater. Nature 395: 338.

Nickson RT, McArthur JM, Ravenscroft P, Burgess WG, Ahmed KM. 2000. Mechanismof arsenic release to groundwater, Bangladesh and West Bengal. Applied Geochemistry15: 403–413.

Nixon SW. 1980. Between coastal marshes and coastal waters-A review of twenty years ofspeculation and research on the role of salt marshes in estuarine productivity and waterchemistry. In: Hamilton P, MacDonald KB, eds. Estuarine and Wetland Processes. NewYork: Plenum, 437–525.

Nriagu JO. 1989. A global assessment of natural sources of atmospheric trace metals.Nature 338: 47–49.

Nriagu JO, Pacyna JM. 1988. Quantitative assessment of worldwide contamination of air,water and soils with trace metals. Nature 333: 134–139.

Nye PH. 1972. The measurement and mechanism of ion diffusion in soils. VIII. A theoryfor the propagation of changes of pH in soils. Journal of Soil Science 23: 82–92.

Nye PH. 1979. Diffusion of ions and uncharged solutes in soils and soil clays. Advancesin Agronomy 31: 225–272.

Nye PH. 1981. Changes in the pH across the rhizosphere induced by roots. Plant and Soil61: 7–26.

Nye PH, Ameloko AY. 1986. A comparison of measured and theoretical soil aciditydiffusion coefficients over a wide range of pH. Journal of Soil Science 37: 191–196.

Nye PH, Staunton S. 1994. The self-diffusion of strongly adsorbed anions in soil: a two-path model to simulate restricted access to exchange sites. European Journal of SoilScience 45: 145–152.

O’Connor GA, Lindsay WL, Olsen SR. 1971. Diffusion of iron and iron chelates in soil.Soil Science Society of America Proceedings 35: 407–410.

Olesen T, Moldrup P, Yamaguchi T, Rolston DE. 2001. Constant slope impedance factormodel for predicting the solute diffusion coefficient in unsaturated soil. Soil Science166: 89–96.

Olivier JGJ, Bouwman AF, Berdowski JJM, Veldt C, Bloos JPJ, Visschedijk AJH, vander Maas CWM, Zasndveldt PYJ. 1999. Sectoral emission inventories of greenhouse

Page 281: The Biogeochemistry of Submerged Soils

References 273

gases for 1990 on a per country basis as well as on 1 × 1. Environmental Science andPolicy 2: 241–263.

Olivier JGJ, Bouwman AF, Van der Hoek KW, Berdowski JJM. 1998. Global air emissioninventories for anthropogenic sources of NOx, NH3 and N2O in 1990. EnvironmentalPollution 102: 135–148.

Olk DC, Brunetti G, Senesi N. 1999. Organic matter in double-cropped lowland rice soils:Chemical and spectroscopic properties. Soil Science 164: 633–649.

Olk DC, Cassman KG, Mahieu N, Randall EW. 1998. Conserved chemical propertiesof young humic acid fractions in tropical lowland soil under intensive irrigated ricecropping. European Journal of Soil Science 49: 337–349.

Olk DC, Cassman KG, Randall EW, Kinchesh P, Sanger LJ, Anderson JM. 1996.Changes in chemical properties of organic matter with intensified rice cropping intropical lowland soil. European Journal of Soil Science 47: 293–303.

Olk DC, Senesi N. 2000. Properties of chemically extracted organic matter in intensivelycropped lowland rice soils. In: Kirk GJD, Olk DC, eds. Carbon and Nitrogen Dynamicsin Flooded Soils. Manila: International Rice Research Institute, 65–87.

Olofsdotter M, Rebulanan M, Madrid A, Wang DL, Navarez D, Olk DC. 2002. Why phe-nolic acids are unlikely primary allelochemicals in rice. Journal of Chemical Ecology28: 229–242.

Olson RK, ed. 1992. The role of created and natural wetlands in controlling nonpointsource pollution. Ecological Engineering 1: 1–170.

Onken BM, Hossner LR. 1995. Plant uptake and determination of arsenic species in soilsolution under flooded conditions. Journal of Environmental Quality 24: 373–381.

Otte ML, Rozema J, Koster L, Haarsma MS, Broekman RA. 1989. Iron plaque in rootsof Aster tripolium L.: interaction with zinc uptake. New Phytologist 111: 309–317.

Ottow JCG. 1973. Bacterial mechanism of iron reduction and gley formation. In: Schlicht-ing E, Schwertmann U, eds. Pseudogley and Gley. Weinheim: International Soil ScienceSociety, 29–35.

Page AL, Bingham FT, Chang AC. 1981. Cadmium. In: Lepp NW, ed. Effect of HeavyMetal Pollution on Plants . Vol. 1. London: Applied Science, 77–109.

Painuli DK, Woodhead T, Pagliai M. 1988. Effective use of energy and water in rice-soilpuddling. Soil and Tillage Research 12: 149–161.

Pasricha NS, Ponnamperuma FN. 1976. Ionic equilibria in flooded saline, alkali soils: theK+ − (Ca2+ + Mg2+) exchange equilibria. Soil Science 122: 315–320.

Pasricha NS, Ponnamperuma FN. 1977. Na+ − (Ca2+ + Mg2+) exchange equilibria undersubmerged soil conditions. Soil Science 123: 220–223.

Patrick WH, Jr. 1966. Apparatus for controlling the oxidation–reduction potential ofwaterlogged soil. Nature 212: 1278–1279.

Patrick WH, Jr, Delaune RD. 1972. Characterization of the oxidized and reduced zonesin flooded soils. Soil Science Society of America Proceedings 36: 573–576.

Patrick WH, Jr, Gambrell RP, Faulkner SP. 1996. Redox measurements of soils. In:Sparks DL, Page AL, Helmke PA, Loeppert RH, Soltanpour PN, Tabatabai MA, John-ston CT, Sumner ME, eds. Methods of Soil Analysis. Madison, WI: Soil Science Societyof America Inc., 1255–1273.

Patrick WH, Jr, Mahapatra JC. 1968. Transformation and availability to rice of nitrogenand phosphorus in waterlogged soils. Advances in Agronomy 20: 323–359.

Page 282: The Biogeochemistry of Submerged Soils

274 References

Patrick WH, Jr, Williams BG, Moraghan JJ. 1973. A simple system for controlling redoxpotential and pH in soil suspensions. Soil Science Society of America Proceedings 37:331–332.

Pearsall WH, Mortimer CH. 1939. Oxidation-reduction potentials in waterlogged soils,natural waters and muds. Journal of Ecology 27: 483–501.

Peng S, Cassman KG. 1998. Upper thresholds of nitrogen uptake rates and associatednitrogen fertilizer efficiencies in irrigated rice. Agronomy Journal 90: 178–185.

Peng S, Laza RC, Visperas RM, Sanico AL, Cassman KG, Khush GS. 2000. Grain yieldof rice cultivars and lines developed in the Philippines since 1966. Crop Science 40:307–314.

Phillips DL, Lovley DR, Roden EE. 1993. Composition of non-microbially reducibleFe(III) in aquatic sediments. Applied and Environmental Microbiology 59: 2727–2729.

Pinner A, Nye PH. 1982. A pulse method for studying effects of dead-end pores, slowequilibration and soil structure on diffusion of solutes in soil. Journal of Soil Science33: 25–35.

Ponnamperuma FN. 1955. The chemistry of submerged soils in relation to the growth andyield of rice. PhD thesis, Cornell University.

Ponnamperuma FN. 1972. The chemistry of submerged soils. Advances in Agronomy 24:29–96.

Ponnamperuma FN. 1985. Chemical kinetics of wetland rice soils relative to soil fertility.In: Wetland Soils: Characterization, Classification and Utilization. Manila: InternationalRice Research Institute, 71–89.

Ponnamperuma FN, Yuan WL. 1966. Toxicity of boron to rice. Nature 211: 780–781.Prather M, Ehhalt D, Dentener F, Derwent R, Dlugokencky E, Holland E, Isaksen I,

Katima J, Kirchhoff V, Matson P, Midgley P, Wang M. 2001. Atmospheric chemistryand greenhouse gases. In: Houghton JT, Ding Y, Griggs DJ, Noguer M, van derLinden PJ, Dai X, Maskell K, Johnson CA, eds. Climate Change 2001: The ScientificBasis. Contribution of Working Group I to the Third Assessment Report of theIntergovernmental Panel on Climate Change. Cambridge: Cambridge University Press,239–287.

Preston-Mafham J, Boddy L, Randerson PF. 2002. Analysis of microbial community func-tional diversity using sole-carbon-source utilisation profiles—a critique. FEMS Micro-biology Ecology 42: 1–14.

Quijano-Guerta C, Kirk GJD. 2002. Tolerance of rice germplasm to salinity and other soilchemical stresses in tidal wetlands. Field Crops Research 76: 111–121.

Quijano-Guerta C, Kirk GJD, Portugal AM, Bartolome VI, McLaren GC. 2002. Toler-ance of rice germplasm to zinc deficiency. Field Crops Research 76: 123–130.

Rachhpal-Singh, Kirk GJD. 1993a. A model for predicting the fate of nitrogen fertilizerin lowland ricefields. I. Theory. Journal of Soil Science 44: 271–283.

Rachhpal-Singh, Kirk GJD. 1993b. A model for predicting the fate of nitrogen fertilizerin lowland ricefields. II. Predicted dynamics of inorganic carbon, nitrogen and acidityin the soil and floodwater. Journal of Soil Science 44: 285–297.

Raven JA. 1996. Into the voids: the distribution, functions, development and maintenanceof gas spaces in plants. Annals of Botany 78: 137–142.

Razafinjara AL. 1999. Cation–anion balances and chemical changes in the rhizosphereof rice in an iron toxic soil. PhD thesis, University of the Philippines at Los Banos.

Reamer DC, Zoller WH. 1980. Selenium biomethylation products from soil and sewagesludge. Science 208: 500–502.

Page 283: The Biogeochemistry of Submerged Soils

References 275

Reddy KR, Patrick WH, Jr., Lindau CW. 1989. Nitrification–denitrification at the plantroot–sediment interface in wetlands. Limnology and Oceanography 34: 1004–1013.

Reddy KR, Rao PSC, Patrick WH, Jr. 1980. Factors influencing oxygen consumptionrates in flooded soils. Soil Science Society of America Journal 44: 741–744.

Redfield AC. 1934. On the proportions of organic derivatives in seawater and their relationto the composition of plankton. In: Daniel RJ, ed. James Johnson Memorial Volume.Liverpool: Liverpool University Press, 176–192.

Reichardt W, Inubushi K, Tiedje JM. 2000. Microbial processes in C and N dynamics.In: Kirk GJD, Olk DC, eds. Carbon and Nitrogen Dynamics in Flooded Soils. Manila:International Rice Research Institute, 101–145.

Reichardt W, Mascarina GB, Padre B, Doll J. 1996. Microbial communities of contin-uously cropped, irrigated rice fields. Applied and Environmental Microbiology 63:233–238.

Reid RJ, Brookes JD, Tester MA, Smith FA. 1996. The mechanism of zinc uptake inplants–Characterisation of the low-affinity system. Planta 198: 39–45.

Revsbech NP, Pedersen O, Reichardt W, Briones A. 1999. Microsensor analysis of oxy-gen and pH in the rice rhizosphere under field and laboratory conditions. Biology andFertility of Soils 29: 379–385.

Richardson JL, Vepraskas MJ, eds. 2001. Wetland Soils: Genesis, Hydrology, Landscapes,and Classification. Boca Raton, FL: Lewis Publishers.

Ridley WP, Dizikes LJ, Woord JM. 1977. Biomethylation of toxic elements in the envi-ronment. Science 197: 329–332.

Ritz K. 2004. Underview: origins and consequences of belowground biodiversity. In:Usher MB, Hopkins DW, Bardgett RG, eds. Soil Biodiversity and Ecosystem Function.Oxford: Blackwell Scientific, in press.

Robinson RA, Stokes RH. 1959. Electrolyte Solutions. London: Butterworth.Roger PA. 1996. Biology and Management of the Floodwater Ecosystem in Ricefields.

Manila: International Rice Research Institute.Roger PA, Heong KL, Teng PS. 1991. Biodiversity and sustainability of wetland rice pro-

duction: role and potential of microorganisms and invertebrates. In: Hawksworth DL,ed. The Biodiversity of Microorganisms and Invertebrates: Its Role in Sustainable Agri-culture. Wallingford: CAB International, 117–136.

Roger PA, Ladha JK. 1992. Biological N2 fixation in wetland rice fields–estimation andcontribution to nitrogen balance. Plant and Soil 141: 41–55.

Ross SM. 1995. Overview of hydrochemistry and solute processes in British wetlands. In:Hughes JMR, Heathwaite AL, eds. Hydrology and Hydrochemistry of British Wetlands.Chichester: Wiley, 133–181.

Roth CB, Jackson ML, Syers JK. 1969. Deferration effect on structural ferrous-ferric ironration and CEC of vermiculites and soils. Clays and Clay Minerals 17: 253–264.

Roy R, Conrad R. 1999. Effect of methanogenic precursors (acetate, hydrogen, propi-onate) on the suppression of methane production by nitrate in anoxic rice field soil.FEMS Microbiology Ecology 28: 49–61.

Sah RN, Mikkelsen DS, Hafez AA. 1989. Phosphorus behavior in flooded-drained soils.III. Phosphorus desorption and availability. Soil Science Society of America Journal 53:1729–1732.

Sajwan KS, Lindsay WL. 1986. Effects of redox on zinc deficiency in paddy rice. SoilScience Society of America Journal 50: 1264–1269.

Page 284: The Biogeochemistry of Submerged Soils

276 References

Saleque MA, Kirk GJD. 1995. Root-induced solubilization of phosphate in the rhizosphereof lowland rice. New Phytologist 129: 325–336.

Sanchez PA. 1976. Properties and Management of Soils in the Tropics. New York: Wiley.Sanchez PA. 1981. Soil management in the Oxisol savannahs and Ultisol jungles of trop-

ical South America. In: Greenland DJ, ed. Characterization of Soils. Oxford: OxfordUniversity Press, 214–252.

Sass RL, Fisher FM, Wang YB, Turner FT, Jund FT. 1992. Methane emission from ricefields: the effect of flood water management. Global Biogeochemical Cycles 6: 249–262.

Savant NK, Snyder GH, Datnoff LE. 1997. Silicon management and sustainable rice pro-duction. Advances in Agronomy 58: 151–199.

Scharpenseel HW. 1997. Preface to workshop ‘Management of carbon in tropical soilsunder global change: science, practice and policy’. Geoderma 79: 1–8.

Scharpenseel HW, Eichwald E, Haupenthal Ch, Neue HU. 1983. Zinc deficiency in asoil toposequence, grown to rice, at Tiaong, Quezon Province, Philippines. Catena 10:115–132.

Schlesinger WH, Hartley AE. 1992. A global budget for atmospheric NH3. Biogeochem-istry 15: 191–211.

Schmoger MEV, Oven M, Grill E. 2000. Detoxification of arsenic by phytochelatins inplants. Plant Physiology 122: 793–801.

Schutz H, Holzapfel-Pschorn A, Conrad R, Rennenberg H, Seiler W. 1989a. A 3-yearcontinuous record on the influence of daytime, season and fertilizer treatment onmethane emission rate from an Italian rice paddy. Journal of Geophysical Research94: 16405–16416.

Schutz H, Seiler W, Conrad R. 1989b. Processes involved in formation and emission ofmethane in rice paddies. Biogeochemistry 7: 33–53.

Schutz H, Seiler W, Conrad R. 1990. Influence of soil temperature on methane emissionfrom rice paddy fields. Biochemistry 11: 77–95.

Scott MJ, Morgan JJ. 1995. Reactions on oxide surfaces. 1. Oxidation of As(III) by syn-thetic birnessite. Environmental Science and Technology 29: 1898–1905.

Scott MJ, Morgan JJ. 1996. Reactions at oxide surface. 2. Oxidation of Se(IV) by syntheticbirnessite. Environmental Science and Technology 30: 1990–1996.

Segers R. 1998. Methane production and methane consumption: a review of processesunderlying wetland methane fluxes. Biogeochemistry 41: 23–51.

Seiler W, Holzapfel-Pschorn A, Conrad R, Scharffe D. 1984. Methane emission from ricepaddies. Journal of Atmospheric Chemistry 1: 241–268.

Seinfeld JH, Pandis SN. 1998. Atmospheric Chemistry and Physics. From Air Pollutionto Climate Change. New York: Wiley.

Setter TL, Ellis M, Laureles EV, Ella ES, Senadhira D, Mishra SB, Sarkarung S, Datta S.1997. Physiology and genetics of submergence tolerance in rice. Annals of Botany 79:67–77.

Shioiri M. 1943. The Chemistry of Paddy Soils. Tokyo: Japanese Agricultural Society(Dainihan-Nokai).

Simpson IC, Roger PA, Oficial R, Grant IF. 1993a. Density and composition of aquaticoligochaete populations in different farmers’ ricefields. Biology and Fertility of Soils16: 34–40.

Simpson IC, Roger PA, Oficial R, Grant IF. 1993b. Impact of agricultural practices onaquatic oligochaete populations in ricefields. Biology and Fertility of Soils 16: 27–33.

Page 285: The Biogeochemistry of Submerged Soils

References 277

Simpson IC, Roger PA, Oficial R, Grant IF. 1994a. Effects of nitrogen fertilizer andpesticide management on floodwater ecology in a wetland ricefield. 2. Dynamics ofmicrocrustaceans and dipteran larvae. Biology and Fertility of Soils 17: 138–146.

Simpson IC, Roger PA, Oficial R, Grant IF. 1994b. Effects of nitrogen fertilizer and pes-ticide management on floodwater ecology in a wetland ricefield. 3. Effects on benthicmolluscs. Biology and Fertility of Soils 18: 219–227.

Simpson JR, Freney JR, Wetselaar R, Muirhead WA, Leuning R, Denmead OT. 1984.Transformations and losses of urea nitrogen after application to flooded rice. AustralianJournal of Agricultural Research 35: 189–200.

Skelton NJ, Alloway WG. 1996. Oxygen and pressure changes in situ during floodingin roots of the grey mangrove Avicennia marina (Forskk) Vier. Aquatic Botany 54:165–175.

Skiba U, Smith KA, Fowler D. 1993. Nitrification and denitrification as sources of nitricoxide and nitrous oxide in a sandy loam soil. Soil Biology and Biochemistry 25:1527–1536.

Slemr F, Langer E. 1992. Increase in global atmospheric concentration of mercury inferredfrom measurements over the Atlantic Ocean. Nature 355: 434–437.

Slemr F, Schuster G, Seiler W. 1985. Distribution, speciation, and budget of atmosphericmercury. Journal of Atmospheric Chemistry 3: 407–434.

Snoeyenbos-West OL, Nevin KP, Anderson RT, Lovley DR. 2000. Enrichment ofGeobacter species in response to stimulation of Fe(III) reduction in sandy aquifersediments. Microbial Ecology 39: 153–167.

So HB, Nye PH. 1989. The effect of bulk density, water content and soil type on thediffusion of chloride in soil. Journal of Soil Science 40: 743–749.

Sorrell BK, Armstrong W. 1991. On the difficulties of measuring oxygen release by rootsystems of wetland plants. Journal of Ecology 82: 177–183.

Sparks DL. 2003. Environmental Soil Chemistry. San Diego: Academic Press.Sposito G. 1983. The chemical forms of trace metals in soils. In: Thornton I, ed. Applied

Environmental Geochemistry. London: Academic Press, 123–170.Sposito G. 1984a. The future of an illusion: Ion activities in soil solutions. Soil Science

Society of America Journal 48: 531–536.Sposito G. 1984b. The Surface Chemistry of Soils. New York: Oxford University Press.Sposito G. 1994. Chemical Equilibria and Kinetics in Soils. New York: Oxford University

Press.St-Cyr L, Crowder AA. 1990. Manganese and copper in the root plaque of Phragmites

australis (Cav.) Trin. ex Steudel. Soil Science 149: 191–198.St Louis VL, Rudd JWM, Kelly CA, Beaty KG, Bloom NS, Flett RJ. 1994. Importance of

wetlands as sources of methyl mercury to boreal forest ecosystems. Canadian Journalof Fisheries and Aquatic Sciences 43: 2472–2503.

Staunton S. 1990. A comparison of the surface impedance factors of Ca, Na, Rb and Csderived from their self-diffusion coefficients in various soils. Journal of Soil Science41: 643–653.

Staunton S, Nye PH. 1983. The self-diffusion of sodium in a naturally aggregated soil.Journal of Soil Science 34: 263–269.

Staunton S, Nye PH. 1987. The contribution of sorbed sodium and calcium to the self-diffusion of these cations in soils with different clay mineralogies. Journal of SoilScience 38: 651–658.

Page 286: The Biogeochemistry of Submerged Soils

278 References

Stein LY, La Duc MT, Grundl TJ, Nealson KH. 2001. Bacterial and archaeal populationsassociated with freshwater ferromanganous micronodules and sediments. EnvironmentalMicrobiology 3: 10–18.

Stern KH. 1954. The Liesegang phenomenon. Chemical Reviews 54: 79–99.Stephen KD, Arah JRM, Daulat W, Clymo RS. 1998a. Root-mediated gas transport in

peat determined by argon diffusion. Soil Biology and Biochemistry 30: 501–508.Stephen KD, Arah JRM, Thomas KL, Benstead J, Lloyd D. 1998b. Gas diffusion coeffi-

cient profile in peat determined by modelling mass spectrometric data: implications forgas phase distribution. Soil Biology and Biochemistry 30: 429–431.

Straub KL, Benz M, Schink B. 2001. Iron metabolism in anoxic environments at nearneutral pH. FEMS Microbiology Ecology 34: 181–186.

Stucki JW. 1988. Structural iron in smectites. In: Stucki JW, Goodman BA, Schwert-mann U, eds. Iron in Soils and Clay Minerals. Dordrecht: D. Reidel, 625–675.

Stucki JW, Auerswald K, Stanjek H, Bigham J. 1997. Redox processes in smectites: soilenvironmental significance. Advances in Geoecology 30: 395–406.

Stucki JW, Golden DC, Roth CB. 1984. Effect of reduction and reoxidation of structuraliron on the surface charge and dissolution of dioctahedral smectites. Clays and ClayMinerals 32: 350–356.

Stumm W, Lee GF. 1961. Oxygenation of ferrous iron. Industrial and Engineering Chem-istry 53: 143–146.

Stumm W, Morgan JJ. 1996. Aquatic Chemistry: Chemical Equilibria and Rates in Natu-ral Waters. New York: Wiley-Interscience.

Sturgis MB. 1936. Changes in the oxidation–reduction equilibrium in soils as related tothe physical properties of the soil and the growth of rice. Louisiana Bulletin No 271.Baton Rouge, LA: Louisiana State University and A & M College.

Sumner ME, ed. 2000. Handbook of Soil Science. Boca Raton: CRC Press.Swain EB, Engstrom DR, Brigham MF, Henning TA, Brezonik PL. 1992. Increasing rates

of atmospheric Hg deposition in midcontinental north America. Science 257: 784–787.Taiz L, Zeiger E. 2002. Plant Physiology. Sunderland, MA: Sinauer Associates.Takai Y. 1952. Microbiological studies on paddy soils. Journal of the Science of Soil and

Manure Japan 23: 37–41.Takajima Y. 1964. Studies on organic acids in paddy field soils with reference to their

inhibitory effects on rice plants. Soil Science and Plant Nutrition 10: 212–219.Takijima Y, Katsumi F. 1973. Cadmium contamination of soils and rice plants caused by

zinc mining. I. Production of high cadmium rice on the paddy fields in lower reachesof the mine station. Soil Science and Plant Nutrition 19: 29–38.

Tamura H, Goto K, Nagayama Y. 1976. The effect of ferric hydroxide on the oxygenationof ferrous iron in neutral solution. Corrosion Science 16: 197–207.

Tanaka A, Navasero SA. 1967. Carbon dioxide and organic acids in relation to the growthof rice. Soil Science and Plant Nutrition 12: 25–30.

Taylor JA, Brasseur GP, Zimmermann PR, Cicerone RJ. 1991. A study of the sources andsinks of methane and methyl chloroform using a global three-dimensional Lagrangiantropospheric tracer transport model. Journal of Geophysical Research 96D: 3013–3044.

Tiedje JM. 1988. Ecology of denitrification and dissimilatory nitrate reduction to ammo-nium. In: Zehnder AJB, ed. Biology of Anaerobic Microorganisms. New York: Wiley-Interscience, 179–244.

Tinker PB, Nye PH. 2000. Solute Movement in the Rhizosphere. New York: Oxford Uni-versity Press.

Page 287: The Biogeochemistry of Submerged Soils

References 279

Tipping E. 1994. WHAM–A chemical equilibrium model and computer code for waters,sediments and soils incorporating a discrete site/electrostatic model of ion-binding byhumic substances. Computing in Geosciences 20: 973–1023.

Tipping E. 2002. Cation Binding by Humic Substances. Cambridge: Cambridge UniversityPress.

Tripathi BP, Ladha JK, Timsina J, Pascua SR. 1997. Nitrogen dynamics and balance inintensified rainfed lowland rice-based cropping systems. Soil Science Society of AmericaJournal 61: 812–821.

Trolldenier G. 1988. Visualization of oxidizing power of rice roots and of possible partici-pation of bacteria in iron deposition. Zeitschrift fur Pflanzenernahrung und Bodenkunde151: 117–121.

Trolove SN. 2000. Root–soil–phosphate interactions in rice growing in aerobic soil. PhDthesis, Massey University.

Tsutsuki K, Ponnamperuma FN. 1987. Behavior of anaerobic decomposition products insubmerged soils. Effect of organic material amendment, soil properties, and temperature.Soil Science and Plant Nutrition 33: 13–33.

Tsuzuki M, Miyachi S. 1989. The functions of carbonic anhydrase in aquatic photosyn-thesis. Aquatic Botany 34: 85–104.

USDA. 1999. Soil Taxonomy. A Basic System of Soil Classification for Making and Inter-preting Soil Surveys. Washington, DC: US Department of Agriculture, Natural ResourcesConservation Service.

Usher MB, Hopkins DW, Bardgett RG, eds. 2004. Soil Biodiversity and Ecosystem Func-tion. Oxford: Blackwell Scientific, in press.

van Bodegom PM, Goudriaan J, Leffelaar P. 2002. A mechanistic model on methaneoxidation in a rice rhizosphere. Biogeochemistry 55: 145–177.

van Bodegom PM, Stams F, Mollema L, Boeke S, Leffelaar P. 2001. Methane oxidationand the competition for oxygen in the rice rhizosphere. Applied and EnvironmentalMicrobiology 67: 3586–3597.

van Breemen N. 1969. The effect of ill-defined ferric oxides on redox characteristics offlooded soils. Netherlands Journal of Agricultural Science 17: 256–260.

van Breemen N. 1988. Redox processes of iron and sulfur involved in the formation ofacid sulfate soils. In: Stucki JW, Goodman BA, Schwertmann U, eds. Iron in Soils andClay Minerals. Dordrecht: D. Reidel, 825–841.

van Breemen N, Buurman P. 1998. Soil Formation. Dordrecht: Kluwer.van Breemen N, Quijano CC, Sen LN. 1980. Zinc deficiency in wetland rice along a

toposequence of hydromorphic soils in the Philippines. I. Soil conditions and hydrology.Plant and Soil 57: 203–214.

van der Gon HACD, Kropff MJ, van Breemen N, Wassmann R, Lantin RS, Aduna E,Corton TM, van Laar HH. 2002. Optimizing grain yields reduces CH4 emissions fromrice paddy fields. Proceedings of the National Academy of Sciences 99: 12021–12024.

van der Gon HACD, Neue HU. 1995. Methane emission from a wetland rice field asaffected by salinity. Plant and Soil 170: 307–313.

van der Gon HACD, Neue HU. 1996. Oxidation of methane in the rhizosphere of riceplants. Biology and Fertility of Soils 22: 359–366.

van der Gon HACD, van Bodegom PM, Houweling S, Verburg PH, van Breemen N.2000. Combining upscaling and downscaling of methane emissions from rice fields:Methodologies and preliminary results. Nutrient Cycling in Agroecosystems 58:285–301.

Page 288: The Biogeochemistry of Submerged Soils

280 References

van Mensvoort MEF, Lantin RS, Brinkman R, van Breemen N. 1985. Toxicities of wet-land soils. In: Wetland Soils: Characterization, Classification, and Utilization. Manila:International Rice Research Institute, 123–138.

van Noordwijk M, Brouwer G. 1993. Gas-filled porosity in response to temporary lowoxygen supply in different growth stages. Plant and Soil 152: 187–199.

van Noordwijk M, Cerri C, Woomer PL, Nugroho K, Bernoux M. 1997. Soil carbondynamics in the humid tropical forest zone. Geoderma 79: 187–225.

van Raalte MH. 1944. On the oxidation of the environment by the roots of rice (Oryzasativa, L.). Annales du Jardin Botanque de Buitenzorg, Hors Serie 1: 15–34.

van Ranst E, De Coninck F. 2002. Evaluation of ferrolysis in soil formation. EuropeanJournal of Soil Science 53: 513–519.

van Rees KCJ, Reddy KR, Rao PSC. 1996. Influence of benthic organisms on solutetransport in lake sediments. Hydrobiologia 317: 31–40.

Vartapetian BB, Jackson MB. 1997. Plant adaptations to anaerobic stress. Annals of Botany79: 3–20.

Vaughan D, Ord B. 1990. Influence of phenolic acids on morphological changes in rootsof Pisum sativum . Journal of the Science of Food and Agriculture 52: 289–299.

Vinograd JR, McBain JW. 1941. Diffusion of electrolytes and of ions in their mixtures.Journal of the American Chemical Society 63: 2008–2015.

Wang B, Neue HU, Samonte HP. 1999. Factors controlling diel patterns of methane emis-sion via rice. Nutrient Cycling in Agroecosystems 53: 229–235.

Wang YM, Siddiqi MY, Ruth TJ, Glass ADM. 1993a. Ammonium uptake by rice roots.I. Fluxes and subcellular distribution of 13 NH4

+. Plant Physiology 103: 1249–1258.Wang YM, Siddiqi MY, Ruth TJ, Glass ADM. 1993b. Ammonium uptake by rice

roots. II. Kinetics of 13 NH4+ influx across the plasmalemma. Plant Physiology 103:

1259–1267.Wassmann R, Buendia LV, Lantin RS, Bueno CS, Lubigan LA, Umali A, Nocon NN,

Javellana AM, Neue HU. 2000a. Mechanisms of crop management impact on methaneemissions from rice fields in Los Banos, Philippines. Nutrient Cycling in Agroecosys-tems 58: 107–119.

Wassmann R, Lantin RS, Neue HU, eds. 2000b. Methane Emissions from Major RiceEcosystems in Asia. Dordrecht: Kluwer.

Wassmann R, Lantin RS, Neue HU, Buendia LV, Corton TM, Lu Y. 2000c. Characteri-zation of methane emissions from rice fields in Asia. III. Mitigation options and futureresearch needs. Nutrient Cycling in Agroecosystems 58: 23–36.

Wassmann R, Neue HU, Lantin RS, Makarim K, Chareonsilp N, Buendia LV, Rennen-berg H. 2000d. Characterization of methane emissions from rice fields in Asia. II.Differences among irrigated, rainfed, and deepwater rice. Nutrient Cycling in Agroe-cosystems 58: 13–22.

Watanabe I, Furusaka C. 1980. Microbial ecology of flooded rice soils. In: Alexander M,ed. Advances in Microbial Ecology , Vol. 4. New York: Plenum, 125–168.

Wehrli B. 1990. Redox reactions of metal ions at mineral surfaces. In: Stumm W, ed.Aquatic Chemical Kinetics. New York: Wiley-Interscience, 311–336.

Weiss HV, Koide M, Goldberg ED. 1971. Mercury in a Greenland icesheet: Evidence ofrecent input by man. Science 174: 694.

Welch RM, Allaway WH, House WA, Kubota J. 1991. Geographic distribution of traceelement problems. In: Mortvedt JJ, Cox FR, Shuman LM, Welch RM, eds. Micronu-trients in Agriculture. Madison, WI: Soil Science Society of America, 31–57.

Page 289: The Biogeochemistry of Submerged Soils

References 281

Welch RM, Graham RD. 1999. A new paradigm for world agriculture: meeting humanneeds. Productive, sustainable and nutritious. Field Crops Research 60: 1–10.

Whitehead DC. 1984. The distribution and transformation of iodide in the environment.Environment International 10: 321–339.

Widdel F. 1988. Microbiology and ecology of sulfate- and sulfur-reducing bacteria. In:Zehnder AJB, ed. Biology of Anaerobic Microorganisms. New York: Wiley-Interscience,469–585.

Wielinga B, Mizuba MM, Fendorf S. 2001. Iron promoted reduction of chromate by dis-similatory iron-reducing bacteria. Environmental Science and Technology 35: 522–527.

Willett IR. 1979. The effects of flooding for rice culture on soil chemical properties andsubsequent maize growth. Plant and Soil 52: 373–383.

Willett IR. 1991. Phosphorus dynamics in acidic soils that undergo alternate flooding anddrying. In: Deturck P, Ponnamperuma FN, eds. Rice Production on Acid Soils of theTropics. Kandy, Sri Lanka: Institute of Fundamental Studies, 43–49.

Witt C, Cassman KG, Olk DC, Biker U, Liboon SP, Samson MI, Ottow JCG. 2000. Croprotation and residue management effects on carbon sequestration, nitrogen cycling andproductivity of irrigated rice systems. Plant and Soil 225: 263–278.

Yagi K, Minami K. 1990. Effect of organic matter application on methane emission fromsome Japanese paddy fields. Soil Science and Plant Nutrition 36: 599–610.

Yagi K, Tsuruta H, Minami K, Chairog P, Cholitkul W. 1994. Methane emission fromJapanese and Thai paddy soils. In: Minami K, Mosier AR, Sass RL, eds. CH4 and N2 O:Global Emissions and Controls from Rice Fields and Other Agricultural and IndustrialSources. Tsukuba: National Institute for Agroenvironmental Sciences, 41–53.

Yang X, Romheld V, Marschner H. 1994. Effect of bicarbonate and root growth andaccumulation of organic acids in Zn-inefficient and Zn-efficient rice cultivars (Oryzasativa L.). Plant and Soil 164: 1–7.

Yao H, Conrad R. 1999. Thermodynamics of methane production in different rice paddysoils from China, the Philippines and Italy. Soil Biology and Biochemistry 31: 463–473.

Yao H, Conrad R. 2000. Electron balance during steady-state production of CH4 and CO2

in anoxic rice soil. European Journal of Soil Science 51: 369–378.Yao H, Conrad R, Wassmann R, Neue HU. 1999. Effect of soil characteristics on sequen-

tial reduction and methane production in sixteen rice paddy soils from China, thePhilippines, and Italy. Biogeochemistry 47: 269–295.

Yao H, Zhuang YB, Chen ZL. 1996. Estimation of methane emission from rice paddiesin mainland China. Global Biogeochemical Cycles 10: 641–649.

Ye ZH, Baker AJM, Wong MH, Willis AJ. 1997a. Copper and nickel uptake, accumula-tion and tolerance in Typha latifolia with and without iron plaque on the root surface.New Phytologist 136: 481–488.

Ye ZH, Baker AJM, Wong MH, Willis AJ. 1997b. Zinc, lead and cadmium tolerance,uptake and accumulation in Typha latifolia . New Phytologist 136: 469–480.

Yoshida S, Ogawa M, Suenaga K, He C. 1983. Induction and selection of salt-tolerantmutant rices by tissue culture–recent progress at IRRI. In: Cell and Tissue CultureTechniques for Cereal Crop Improvement. Beijing: Science Press, 237–254.

Yu T-R. 1985. Physical Chemistry of Paddy Soils. Berlin: Springer-Verlag.Yuita K. 1992. Dynamics of iodine, bromine, and chlorine in soil. II. Chemical forms of

iodine in soil solutions. Soil Science and Plant Nutrition 38: 281–287.Yukawa K. 1989. Historical development of rice cultivation and its water management.

Irrigation Engineering and Rural Planning 16: 60–70.

Page 290: The Biogeochemistry of Submerged Soils

282 References

Zehnder AJB, Stumm W. 1988. Geochemistry and biogeochemistry of anaerobic habi-tats. In: Zehnder AJB, ed. Biology of Anaerobic Microorganisms. New York: Wiley-Interscience, 1–38.

Zinder B, Furrer G, Stumm W. 1986. The coordination chemistry of weathering. II. Dis-solution of Fe(III) oxides. Geochimica et Cosmochimica Acta 50: 1861–1869.

Zinder SH. 1993. Physiological ecology of methanogens. In: Ferry JG, ed. Methanogen-esis: Ecology, Physiology, Biochemistry and Genetics. New York: Chapman & Hall,128–206.

Page 291: The Biogeochemistry of Submerged Soils

Index

absorption see nutrient absorptionacetaldehyde 166acetate 144, 146acidification 200, 208acidity

see also pHpeat bogs 210ricefields 208–10sulfate soils 213–14toxins 212–14trace metals 78

acids 46–7, 215–16acid–base reactions 35–7, 46–7, 48,

64acid sulfate soils 213–14adsorption 76–9aeration

nutrient absorption model 172–7roots 170–1

aerenchyma 167–9, 171aerobic processes 147–50air–water interface

see also waterCO2 transfer 61–4gas transport 58–64

Alfisols 12–15algae 154–5, 156alkalinity

changes after flooding 109–12water 57

alluvial soils 13, 14, 15, 69aluminium toxicity 213amino acids 46, 180, 190aluminosilicates 68ammonia 7–8

emissions from rice 254–6global budget 252–4volatilization 64, 148, 252, 254–6

ammonium 121–2absorption 178–80, 185–6, 187–9anoxia effects 186–7formation 41–2nitrates 187–9reduction 141

anaerobic conditions 139, 140, 151,165–7

anoxiadecomposition 75–6, 144–7ion transport 186–7root processes 167–70

antimony 231aqueous solutions

carbonates 49ferrous iron oxidation 128–9

Armstrong and Beckett’s model 170–1arsenic 231atmospheric methane 6autotrophs 104–6

bacteriasee also microbiological processesautotrophs 104biological nitrogen fixation 157–8cyanobacteria 154–5, 156, 157heterotrophic 106, 143, 157methane oxidation 149nitrification 148phototrophic 157reduction 137, 142–3

bases 46–7see also acids; acid–base reactionsbiological fixation 70depletion 70

bicarbonate 28, 30, 111air–water interface 61–3rhizosphere 200, 202

The B iogeochemistry of Submerged Soils Guy Kirk 2004 John Wiley & Sons, Ltd ISBN: 0-470-86301-3

Page 292: The Biogeochemistry of Submerged Soils

284 Index

bioavailability of trace elements 220biodiversity 163–4biological nitrogen fixation (BNF)

156–9biological processes

macrobiological 150–62microbiological 135–50

biomass 154, 155biotic catalysis 137biotic factors 156bioturbation 40–4, 161–2bluegreen algae see cyanobacteriaBNF see biological nitrogen fixationbogs 1, 2, 6, 8boron 228buffer power 36

see also derivativesbulk density

depth gradient 25–6impedance factor 27–8

burrowing invertebrates 40–4,161–2

cadmium 224–6calcification 70calcite precipitation 85, 86calculated changes in exchangeable

cations 89–91carbon

balances 4–6, 7, 75organic 4, 7, 9oxidation state 146sequestration 258transformations 120

carbon dioxideaccumulation in rice soils 137–40air–water transfer 61–4anaerobic incubation 139–40diurnal changes 58ebullition 38–9floodwater dynamics 56–8hydration 55–6pressure increase 111

carriers, membranes 182–4catalysis of redox reactions 102–6,

136–7cation exchange capacity (CEC) 28,

30, 32, 116

cationscalculated changes 89–91concentration 66immobilization 200–2radii 84

CEC see cation exchange capacitychannels, membranes 182, 183charges, pH dependent 73–4chemoautotrophs 104–6chromium 227–8clays 65–8, 69

structural charge 67surface property changes 71–4

climate 155co-precipitation, solid solutions 79,

82–4coastal plains 14, 15cobalt 226complexes 48–50, 76–9

adsorption 78inner- and outer-sphere 77

continuity equations 17–18copper 226cyanobacteria 154–5, 156, 157

Darcy’s law 20dCL/dC derivative 33–5decomposition 150–1

anoxic 75–6methanogenesis 144–7organic matter 120

deficienciessulfur 206–8zinc 208, 221, 225

denitrificationanaerobic conditions 141–2, 148nitrogen oxides 249rhizosphere 196–7

depth changes 106–9diffusion 22–38, 59, 81–2

pH changes 35–8simultaneous oxidation 131–4

diffusion coefficients 22–35bulk density 25–6derivative dCL/dC 33–5free solution 23–5impedance factor 26–33solutes 22–3

Page 293: The Biogeochemistry of Submerged Soils

Index 285

disaggregation 19dissimilatory reduction 141, 143dissolution of coatings 71–2dissolved oxygen 153–4diurnal variations in ricefields 58, 236divalent cation radii 84diversity, biological 163–4

ebullition 38–9ecosystems 3

see also wetlandsEH see redox potentialelectrode curves 118, 119electron acceptors 135, 137–8, 141electron activities 94–9electron balances 146, 147electron donors 141emissions

see also methane emissionsammonia 252–6nitrogen oxides 247–52sulfur compounds 256–8trace elements 219

energetics, microbe-mediated reactions102–6

Entisols 12–15equilibrium

calculations 50–2gas–water 54–8

equilibrium constantsacids–bases 47carbonates 49mineral phases in reduced soils 112reduction half-reactions 95–6, 103trace elements 223–4

exchangeable cations 89–91extracellular transport 180

fauna 40–4, 159–62fens 1, 2, 6, 8fermentation 94, 165–7ferric iron 28, 32

hydroxides 70–1, 113oxides 70–1reduced soil 132reduction 70–1, 72–3rhizosphere 192–4

structural Fe(III) reduction 70–1,72–3

ferrihydrite 101ferrolysis 209–10ferrous iron 31, 32

calculated changes 113–16changes after flooding 110, 112–13depth distribution 107–9hydroxides 113oxidation kinetics 128–31reduced soil 132rhizosphere 192–4toxicity 214–15

fertilizers 253–6Fick’s laws 18, 23flooding

see also ricefieldscation concentration 66changes in pe, pH and Fe2+ 109–13derivative dCL/dC 34–5EH distribution 107–9impedance factor 27–32, 33soil mineralogy 69–71surface properties 71–4

floodplains 2, 6, 8, 13–15floodwater

see also riparian wetlands; waterCO2 dynamics 56–8flora 154–9macrobiological processes 150–62microbiological processes 135–50properties 152–4salinity 216–17soil zones 151–2

flora 154–9fluxes, CO2 61–4fluxial wetlands 4, 5free energy changes 94–7, 105, 136,

139functional analysis 163–4

Gaines and Thomas equation 88Gapon equation 88gas transfer

see also aeration; diffusion; emissionsair–water transport 58–64ebullition 38–9roots 167–70, 171, 174

Page 294: The Biogeochemistry of Submerged Soils

286 Index

gasessee also individual gasesequilibrium with water 54–8global budgets 233–4, 247–9,

252–4greenhouse gases 233trace gases 233–58

genotypic analysis 163GIS systems 244global budgets

ammonia 252–4methane 233–4nitrogen oxides 247–9sulfur compounds 256

glucose oxidation 165–7

Henry’s law 60heterotrophs 106, 143, 157, 243Histosols 14hydration of CO2 55–6hydraulic conductivity 20hydrogen 137, 144hydroxides of iron 70–1, 113hypoxia 186–7

immobilization 198, 200–2impedance factor

liquid phase 26–32solid phase 32–3

Inceptisols 13–15inland valleys 12–13, 14inorganic nutrients 154inputs to ricefields 203–8

see also fertilizers; nutrientsinterchanges, water 45–64intracellular transport 180inverse modelling 245–6invertebrates 40–4, 159–60iodine 232ion transport (roots) 180–90

absorption 184–5, 187, 189–90ammonium 187anoxia effects 186–7membrane transport 181–4nitrate 1287nutrient concentrations 185–6

ions 51see also cations

ironsee also ferric iron; ferrous ironpyrite oxidation 213–14reduction 142–3rice soils content 70–1structural 70–1, 72–3toxicity 214–15

irrigated rice soils 204, 205

kineticsabsorption 184–5ferrous iron oxidation 128–31redox reactions 94–106

lakes 2, 6landforms–soils relationship 12–15lateral roots 173–4, 181layer silicates 65–8lead 229ligands 47–50lignin 76lowland ricefields 1–3

macrobiological processes 150–62fauna 159–62floodwater flora 154–9floodwater properties 152–4floodwater–soil system 151–2net primary production and

decomposition 150–1macrofauna 159–62macronutrients 204–6macrophytes 154–5, 156manganese 107–9, 142–3marshes 2, 6mass flow 19–22mechanistic models 244–5meiofauna 159membranes

root transport processes 181–4transporters 182–4

mercury 226–7metal ions 51metals 9methane 233–47

see also methanogenesisatmospheric 6

Page 295: The Biogeochemistry of Submerged Soils

Index 287

ebullition 38–9global budget 233–4, 245oxidation 149–50, 240

methane emissionsfield-scale model 243modelling 237–43reduction 246–7regional-scale estimation 244–6from ricefields 234–7, 246–7, 251simplified model 240–3

methanogenesisprocesses 144–7, 239–40rice soils 137–9ricefields 249–50

microbiological processes 135–50see also bacteriaaerobic processes 147–50arsenic reduction 230iron and manganese reduction

142–3mercury reduction 226nitrate reduction 141–2redox reactions catalysis 102–6,

136–7respiration 28sequential reduction 136–40sulfate reduction 143–4

micronutrients 206–8mineral soils 10–12mineralogy of rice soils 69–70mixing by soil animals 39–44modelling

down-scaling 245–6methane emissions 237–43nutrient absorption 172–7root aeration 170–1, 172–7transport 17up-scaling 244–5

net flux 21net primary production (NPP) 4, 6,

150–1nickel 226nitrates

absorption v. ammonium 187–9concentration 148reduction 141–2

nitrification 148, 196–7, 249

nitrogenacid–base changes 209biological fixation 6–8, 156–9ebullition 38–9ricefields 206root uptake 178–80tidal soil marsh 213transformations 120–2

nitrogen oxidesemissions 249–52global budget 247–9

NPP see net primary productionnutrient absorption (roots) 180–90

ammonium 187–9anoxia effects 186–7efficiency improvement 189–90external and internal concentrations

185–6kinetics measurement 184–5membrane transport processes

181–4nitrates 187–9root surfaces requirement 177–80

nutrient balances 203–12peat bogs 210, 211ricefields 203–8riparian wetlands 210–11tidal wetlands 211–12, 213

nutrientsinorganic 154transformations 119–26

Nye’s theory 35–8

oligochaetes 40, 161organic acids 45, 111, 137–41, 215–16organic carbon 4, 7, 9, 74–5organic matter see soil organic matterorganic soils 9–10outputs from ricefields 203–8oxidation

see also redox reactionscarbon 146Fe2+ kinetics 128–31glucose 165–7methane 149–50, 240organic matter 136–40pyrite 213–14reduced soil 127–34

Page 296: The Biogeochemistry of Submerged Soils

288 Index

oxidation (continued )selenium 231simultaneous diffusion 131–4

oxides 68–9, 71–2, 81oxygen

see also aeration; aerobic processescalculated changes 115–16consumption 127–8, 169–70, 175dissolved 153–4root transport 169–70, 175

oxygenation of rhizosphere 191–4

pecalculated changes 113–16pe–pH diagrams 99–102redox couples concentration 97–9redox reactions 94–7time changes 109–11

peat bogs 210, 211percolation rates 19–22permeability 19pH

buffer capacity 53–4calculated changes 113–16charge changes 73–4floodwater CO2 dynamics 56–8pe–pH diagrams 99–102propagation of changes 35–8redox conditions 28, 30reduced soil 132rhizosphere 192–6, 202time changes 109–12

phenolic acids 216phenotypic analysis 163Philippines, rice Zn deficiency tolerance

221–4, 225phosphorus 9

diffusion 34immobilization 125, 126, 198solubilization 197–9transformations 124–6transport 42–4wetlands 212

photoautotrophs 104phototrophic bacteria 157phreatic wetlands 4, 5phytotoxic compounds 215plasma membranes 181–4

pluvial wetlands 4, 5pollutants 9

see also toxinsprecipitation 79–82

calcite 86co-precipitation 79, 82–4inhibition 85–7rates 80–1

primary production 155–6productivity of wetland rice systems

135, 154–5puddling 12, 19–20pumps, membranes 183, 184pyrite oxidation 213–14

radiation 152–3rainfed rice soils 206, 207redox conditions

derivative dCL/dC 34–5impedance factor 27–32, 33rice soils 70–1

redox potential (EH)

conditions 28, 30distribution after flooding 107–9electrodes 97measurement 116–19

redox reactions 94–134arsenic 230catalysis 102–6, 136–7free energy changes 105microbe mediation 102–6nutrient element transformations

119–26pe and redox couples 97–9pe–pH diagrams 99–102reduced soil oxidation 127–34soil conditions 106–19

redox species 136redoxymorphic features 10–11reduced soil, oxidation 127–34reduction

see also methanogenesis; redoxreactions

equations 114equilibrium constants 96half-reactions 95–6, 103, 223–4iron and manganese 142–3nitrates 141–2

Page 297: The Biogeochemistry of Submerged Soils

Index 289

sequential 136–40structural Fe(III) 70–1sulfates 143–4

rhizosphere 147–50, 152, 165–202see also root architecture; root

processesmethane emissions 241, 242methane oxidation 149–50nitrification–denitrification 148,

196–7oxygenation 191–4pH 192–6, 202

rice plantsroot system 171, 173yields 216zinc deficiency tolerance 221–4, 225

rice soils see wetland rice soilsricefields 1–3, 6, 8

acidity balances 208–10ammonia emissions 254–6biological nitrogen fixation 158carbon dioxide pressure 63compacted soil pan 11invertebrates 160macronutrients 204–6mass flow 21–2methane emissions 234–7, 246–7,

251micronutrients 206–8nitrogen oxides emissions 249–52nutrient balances 203–8production systems 236, 250–2productivity 135radiation 152–3salinity 217seasonal differences 236–7sulfur emissions 256–8temperatures 152–3water management 215, 247zinc deficiency 208, 221, 225

riparian wetlands 210–11river basins 13–15root architecture 171–80

aeration/nutrient absorption model172–7

aerenchyma 167–9, 171individual roots and laterals 173–4,

181

length densities 171, 241, 242oxygen budget 171–2, 175structure of root system 172–3surface required for nutrient

absorption 177–80root processes 165–202

see also wetland rootsaeration 170–1anoxia adaptations 167–70fermentation 165–7gas transfer 167–70, 171, 174methane emissions 240–3nutrient absorption properties

180–90root-induced changes 190–202transport 177–9, 180–4

root-induced changes 190–202cation immobilization 200–2nitrification–denitrification 196–7phosphate solubilization 197–9rhizosphere oxygenation 191–4zinc solubilization 200

rootssee also rhizosphereaerenchyma 167cortical porosity 176solute pathways 180–1tissue morphology 168

salinity 216–17saturation 10, 11

see also floodingseasonal variations in ricefields 236–7secondary nutrients 206–8selenium 231self-diffusion coefficients 24sequential reduction 136–40shallow lakes 2, 6silicates 65–8silicon 208sodicity 217soil animals 39–44, 159–62soil classification 12–5soil organic matter (SOM) 11, 69,

74–6decomposition 75–6, 120, 144–7oxidation 136–40

Page 298: The Biogeochemistry of Submerged Soils

290 Index

soil profiles 10–11see also depth changes

soil redox conditions 106–19depth changes 107–9Fe2+ changes 112–13nutrient elements transformations

119–26pe and pH changes 109–11redox potential 116–19reduction, calculated changes

113–16soils 65–91

acidity 155–6carbon 4, 7floodwater 151–2landforms relationship 12–15mixing by fauna 39–44moisture content 25–6solid surfaces 65–76surface oxidation 129–31surface property changes 71–4types 9–12

solid phases 74solid surfaces 65–76solid–solution interactions 76–91

adsorption 76–9co-precipitation 79, 82–4equations 87–91precipitation 79–82, 85–7rates 81

solubilizationphosphate 197–9zinc 200, 201

solutesdiffusion coefficients 22–3soils 65–91water 45–64

SOM see soil organic mattersorption equations 35speciation 47–50still water see ricefieldsstructural iron reduction 70–1, 72–3submergence see floodingsulfates

acid soils 213–14reduction 143–4

sulfurbiogeochemical characteristics 8

distribution after flooding 107–9emissions from ricefields 256global budget 256transformations 122–4

surface complexes 76–8see also complexes

surfaces 65–76ferrous iron oxidation 129–31property changes 71–4wetland rice soils 69–71

swamps 1, 2, 6, 8

temperatures 152–3thallium 228–9thermodynamics, redox reactions

94–106tidal wetlands 211–12, 213, 217toxins 212–17

acidity 212–14iron 214–15organic acids 215–16pollutants 9salinity 216–17

trace elements 218–32see also individual elementsequilibrium constants 223–4global cycling 218, 219important properties 222mobilities of elements 220–32soil and plant transport 218–20

trace gases 233–58see also individual gases

trace metals 78transformations of nutrient elements

119–26transport processes 17–44

diffusion 22–38ebullition 38–9mass flow 19–22mixing by soil animals 39–44modelling 17roots 177–9, 180–4trace elements 218–20

tropical wetlands 1tubificids 161, 162

Ultisols 12–15uptake see nutrient absorption

Page 299: The Biogeochemistry of Submerged Soils

Index 291

urea 253hydrolysis 86, 208, 254

vanadium 227

water 45–64see also floodwateracids and bases 46–7air interface, gas transport 58–64composition 45–52equilibrium calculations 50–2equilibrium with gas phase 54–8major inorganic species 52pH buffer capacity 53–4ricefields management 215, 247saturation 10, 11speciation 47–50

waterlogging see flooding; saturationwet tillage 19wetland rice soils 11–12

see also ricefieldsiron contents 70–1organic carbon 74–5

organic matter 76phosphorus 126solid surfaces 69–71

wetland rootsarchitecture 171–80ion transport 184–90nutrient absorption 180–90

wetlandssee also ricefieldsbiogeochemical characteristics 3–9definitions 2global extent 1–3hydrology 5invertebrates 159soils and landforms 9–15

worm burrows 40–4, 160–2

yields of rice 203–4, 216

zincdeficiency 208, 221–4, 225solubility 80solubilization 200, 201

With kind thanks to Indexing Specialists, Hove, East Sussex, UK for compilationof this Index.