National Environmental Research InstituteMinistry of the Environment . Denmark

Hydrology, nutrientprocesses and vegetationin floodplain wetlandsPhD thesisHans Estrup Andersen

[Blank page]

National Environmental Research InstituteMinistry of the Environment . Denmark

Hydrology, nutrientprocesses and vegetationin floodplain wetlandsPhD thesis2003

Hans Estrup Andersen

Data sheet

Title: Hydrology, nutrient processes and vegetation in floodplain wetlandsSubtitle: PhD thesis

Author: Hans Estrup AndersenDepartment: Department of Freshwater Ecology, National Environmental Research InstituteUniversity: The Royal Veterinary and Agricultural University, Department of Agricultural Sci-

ence, Laboratory for Agrohydrology and Bioclimatology

Publisher: National Environmental Research Institute Ministry of the Environment

URL: http://www.dmu.dk

Date of publication: January 2003Editing complete: December 2002

Referees: Henry E. Jensen and Søren Hansen, The Royal Veterinary and Agricultural Univer-sity, Department of Agricultural Sciences, Laboratory for Agrohydrology and Bio-climatology

Financial support: National Environmental Research Institute; The Royal Veterinary and AgriculturalUniversity, Department of Agricultural Sciences

Please cite: Andersen, H.E. 2003: Hydrology, nutrient processes and vegetation in floodplainwetlands. PhD thesis. National Environmental Research Institute, Denmark. 36 pp.http://afhandlinger.dmu.dk

Reproduction is permitted, provided the source is explicitly acknowledged.

ISBN: 89-7772-796-7

Number of pages: 36

Internet-version: The report is available only in electronic form from NERI’s homepagehttp://www.dmu.dk/1_viden/2_Publikationer/ovrige/rapporter/phd_hea.pdf

Contents

Danish summary 5

English Summary 6

Review of hydrology, nutrient processes and vegetation infloodplain wetlands 71. Introduction 7

1.1 Interactions between a river and its surroundings 71.2 The floodplain wetland 7

2. Hydrology of wetlands 82.1 Hydraulic properties of peat 9

Structure of peat 9Water retention in peat 10Compressibility 10Storage 12Saturated hydraulic conductivity 13Measurement of saturated hydraulic conductivity 14Non-Darcian flow in peat? 15Unsaturated hydraulic conductivity 15

2.2 Evapotranspiration from wetlands 16Canopy radiation balance 16Energy balance of an evaporating surface 17Wetland evapotranspiration rates 18

2.3 Wetland water balances 203. Nutrient processes in wetlands 21

3.1 Chemical conditions 21Redox conditions and pH 21Nitrogen 22Phosphorus 23

4. Riparian wetland vegetation 234.1 Factors controlling plant species distribution in riparian wet-

lands 23References 25

Paper 1Andersen, H. E.Hydrology and nitrogen balance of a seasonally inundated Danishfloodplain wetland 32Hydrological Processes, accepted.

Paper 2Andersen, H. E., Hansen, S. and Jensen, H. E.Evapotranspiration from a riparian fen wetland 33Nordic Hydrology, submitted.

Paper 3Schröder, H. K., Andersen, H. E. and Kiehl, K.Rejecting the mean – estimating the response of wetland plant speciesto environmental factors by non-linear quantile regression 34Journal of Vegetation Science, submitted

Paper 4Andersen, H. E. and Svendsen, L. M. (1997)Suspended sediment and total phosphorus transport in a major Dan-ish river: methods and estimation of the effects of a coming majorrestoration 35Aquatic Conservation, Vol. 7, 265-276.

National Environmental Research Institute

5

Dansk Sammenfatning

Denne afhandling afslutter mit PhD-studium udført ved dels Den Kongelige Veterinær- og Landbohøjskole(KVL), Afdeling for Jordbrugsvidenskab, Laboratoriet for Agrohydrologi og Klimatologi, dels ved Dan-marks Miljøundersøgelser, Afdeling for Ferskvandsøkologi. Mine vejledere var professor Henry E. Jensen oglektor Søren Hansen, begge fra KVL. Sektionsleder Lars M. Svendsen var min interne DMU-vejleder.

Formålet med studiet var at beskrive og kvantificere hydrologiske processer og disses betydning for næ-ringsstoffer og vegetation i hyppigt oversvømmede ådale.

Afhandlingen omfatter fire videnskabelige artikler og en sammenfatning. En artikel er publiceret i AquaticConservation, en artikel er accepteret for publicering i Hydrological Processes og en artikel er indsendt tilNordic Hydrology. Den fjerde artikel foreligger som et manuskript, der er forberedt til indsendelse til Jour-nal of Vegetation Science.

Tre af artiklerne omhandler arbejde udført i et vandløbsnært, hyppigt oversvømmet vådområde på den ned-re del af Gjern Å. Den første artikel karakteriserer vådområdet og indeholder en analyse af de styrende for-hold for vand- og kvælstofbalancen. Der er lagt vægt på umættede og mættede hydrauliske karakteristika afvådområdesedimenterne. Der er en ringe tilstrømning af grundvand til vådområdet og vandudvekslingmed atmosfæren er dominerende. Store vandmængder oversvømmer vådområdet, men pga. sedimenterneskarakteristika er infiltrationen ringe. Denitrifikationen udgør 71 kg kvælstof pr. år og er begrænset af til-førslen af nitrat. 75% af denitrifikationen udgøres af reduktion af nitrat, der diffunderer ned i sedimentetunder oversvømmelse af vådområdet med åvand.

Den anden artikel fokuserer på evapotranspirationen fra vådområdet. Evapotranspirationen er estimeretudfra kontinuerte målinger med en Bowen ratio-opsætning gennem hele vækstsæsonen i 1999. Evapotran-spirationen er højere end de fleste publicerede værdier for vådområder med en gennemsnitlig rate forvækstsæsonen på 3.6 mm dag-1. Gennemsnitligt over vækstsæsonen udgør evapotranspirationen 128% afreference-evapotranspirationen beregnet med Penman-Monteith-formlen som anbefalet af FAO. De høje ra-ter forklares dels ved at de kapillære egenskaber af sedimenterne opretholder nær-mættede forhold i rodzo-nen gennem hele vækstsæsonen, dels ved lokal advektion.

Den tredje artikel indeholder en analyse af plantefordelende faktorer i vådområdet. Quantile regression, enny metode til analyse af økologiske data, testes og viser sig velegnet til at reducere effekten af umålte fakto-rer, hvorved effekten af de målte faktorer tydeliggøres. Artiklen indeholder ligninger, der kvantificerer re-sponsen af 18 plantearter til 6 miljøfaktorer. Det vises, at basemætningsgrad, fosfatindhold, oversvømmel-seshyppighed, samt grundvandsspejls-amplituden er dominerende faktorer i den rumlige fordeling ogdækningsgrad af plantearter i vådområdet.

Den fjerde artikel omhandler arbejde udført på en større skala, nemlig et moniteringsstudie udført på dennedre del af Skjern Å forud for restaureringen af de nederste 18 km. Der udvikles empiriske modeller foråens transport af suspenderet sediment og total-fosfor. Det vurderes, at effekten af restaureringen vil væreen reduktion i tilledningen til Ringkøbing Fjord af suspenderet stof og total-fosfor på hhv. 37% og 20% pga.sedimentation i søer og på oversvømmede enge.

I sammenfatningen beskrives faktorer af betydning for hydrologien, næringsstofprocesserne og vegetationeni hyppigt oversvømmede ådale, som ikke er blevet tilstrækkeligt belyst i artiklerne. De særlige forhold, deradskiller tørv fra mineraljorde omtales grundigt. Resultaterne præsenteret i artiklerne sættes i forhold til deninternationale litteratur. Samtidig perspektiveres resultaterne ved sammenligning med resultater fra et stu-die udført i et vådområde, som også ligger i Gjern Å-systemet, men som adskiller sig væsentligt fra det i ar-tikel 1 - 3 undersøgte.

6

English Summary

This thesis represents the conclusion on my PhD study carried out jointly at The Royal Veterinary and Agri-cultural University (KVL), Department of Agricultural Sciences, Laboratory for Agrohydrology and Biocli-matology, Copenhagen, and at The National Environmental Research Institute (DMU), Department ofFreshwater Ecology, Silkeborg. My supervisors were Professor Henry E. Jensen and associate Professor Sø-ren Hansen, both from the KVL. Senior scientist Lars Moeslund Svendsen was my internal DMU-supervisor.

The purpose of the study was to describe and quantify hydrological processes, and the implication of thesefor nutrient processes and vegetation in floodplain wetlands.

The thesis comprises four scientific papers and a review. One of the papers is published in Aquatic Conser-vation, one paper is accepted for publication in Hydrological Processes, and one paper is submitted to Nor-dic Hydrology. The fourth paper is a manuscript prepared for Journal of Vegetation Science.

Three of the papers are on work done in a floodplain wetland in the lower part of the river Gjern. The firstpaper characterises the wetland and comprises an analysis of the controls on the water and nitrogen balan-ces. Emphasis is put on unsaturated and saturated hydraulic characteristics of the wetland sediments. Thereis a minimal inflow of groundwater to the wetland and water exchange with the atmosphere is dominant.Large amounts of water flood the wetland, however, due to the characteristics of the sediments, infiltrationis low. Denitrification amounts to 71kg nitrate per year and is limited by the supply of nitrate. Reduction ofnitrate diffusing into the sediments during flooding of the wetland constitutes 75% of total denitrification.

The second paper focuses on evapotranspiration from the wetland. Evapotranspiration was estimated fromcontinuos measurements with a Bowen ratio-set up throughout the growing season of 1999. With an averagerate for the growing season of 3.6 mm day-1 evapotranspiration was higher than most published values forwetlands. The wetland evapotranspiration comprised 128% of reference-evapotranspiration calculated bythe Penman-Monteith formula as prescribed by FAO. The high rates are explained partly by the capillarycharacteristics of the wetland sediments, which sustain near-saturated conditions in the rootzone throughoutthe growing season, and partly by local advection.

The third paper contains an analysis of plant species distributing factors in the wetland. Quantile regression,a new method for analysing ecological data, was tested. The method was evaluated as adequate for reducingthe influence of multiple combined factors and thus to clarify the relation to single factors. A set of equationsquantifying the response of 18 floodplain wetland species to six environmental factors is given in the paper.It is shown that that degree of base-saturation, exchangeable phosphate, groundwater amplitude and floo-ding duration are major factors in determining plant species distribution and cover in the wetland.

In the fourth paper the scale is enlarged to the subcatchment-level. The results of a monitoring study on thelower 18 km of the river Skjern, prior to the river restoration project, is described. Empirical models for rive-rine transport of suspended sediment and total phosphorus is developed. Assessment of the effects of therestoration, based on measured transport and estimated retention rates for suspended sediment and totalphosphorus for different area types of the river system, revealed that suspended sediment and total phos-phorus will be reduced by 37 and 20%, respectively.

Factors of importance for the hydrology, nutrient processes and vegetation in floodplain wetlands, andwhich have not been considered sufficiently in the papers, are described in the review. Emphasis is put onthe differences between peat and mineral soils regarding hydraulic properties. The results presented in thepapers are discussed relative to the international literature and compared to results from a wetland studyalso in the river Gjern system, but with a hydrology and nutrient turnover deviating from that of the flood-plain wetland analysed in papers 1 – 3.

7

Review of hydrology, nutrient processes and vegetation in floodplain wet-lands

1. Introduction

1.1 Interactions between a river and its surround-ings

Riparian ecosystems encompass the stream chan-nel between the low- and high-water marks. Theyalso encompass the terrestrial landscape above thehigh-water mark where vegetation may be influ-enced by elevated water tables or flooding and bythe ability of the soils to hold water (Naiman andDécamps, 1997). River systems and their riparianzones can be viewed as open ecosystems dynami-cally linked longitudinally, laterally and vertically(Ward, 1989). From a hydrological point of viewstreams and rivers as surface-water bodies are in-tegral parts of groundwater flow systems. It isgenerally assumed that topographically high areasare groundwater recharge areas and topographi-cally low areas are groundwater discharge areas.However, this is primarily true for regional flowsystems (Winter, 1999). Local flow systems associ-ated with surface-water bodies are superposi-tioned on the regional framework resulting incomplex interactions between groundwater andsurface-water regardless of regional topographicposition. Despite of this an overview of ground-water surface-water interactions within a catch-ment is provided by the following conceptualmodel, Fig. 1 (Nilsson et al., 2002). The model as-sumes a homogenous geology and equal precipi-tation within the entire catchment and divides awatercourse into sub-reaches of characteristic hy-drogeology and geomorphology. Fig. 1 shows alongitudinal transect of a river together with anassumed linear course of the regional water divide.

The difference in elevation between the water di-vide and the river along a groundwater streamlineillustrates the hydraulic potential, which drives theregional groundwater flow towards the surfacewater. It is apparent from the Fig. that the largesthydraulic potential is found along the middlereach of the river. Thus it is expected that the larg-est and most stable regional groundwater dis-charge will take place on this section. Upstreamand downstream the hydraulic potential dimin-ishes and the discharge of deep groundwater willbe less. In the headwater area local groundwaterflow systems will dominate and contribute younggroundwater. The flow system will depend onprecipitation and be unstable. The upper reach willbe dominated by local and intermediary flow sys-tems. Along the lower reach the deep groundwaterdischarge has diminished, local and intermediarysystems can contribute, and flooding occur due tothe lower slope of the river.

There has been considerable focus on riparianwetlands in Denmark during the past decade dueto their ability to modify diffuse pollution of sur-face waters. In fact, the Danish Parliament in 1998decided on the restoration of 16,000 hectares ofwetlands with the chief purpose of reducing agri-cultural nitrate loading to the aquatic environment.A number of Danish studies on the hydrology andnutrient processes of riparian wetlands located onthe upper and middle reaches of water courseshave been carried out (Brüsch and Nilsson, 1990;Hoffmann et al., 1993; Dahl, 1995; Paludan, 1995;Hoffmann, 1998a; Blicher-Mathiesen, 1998, Hoff-mann et al., 2000). Less work has been done on thelower reaches subject to frequent flooding – thefloodplains (Hoffmann et al., 1998); Hoffmann,1998b; Kronvang et al., 2001, Andersen, 2002; An-dersen et al., 2002). However, knowledge about thecharacteristics and functioning is important in or-der to restore and maintain floodplain wetlands.

1.2 The floodplain wetland

Internationally as well has there been considerableinterest in the function and value of wetlands (e.g.Carter, 1986). Floodplain wetlands, in their naturalstate, have been cited to be of particular valuesince they have a high biodiversity, provide criticalhabitats for many plants and animals, and are animportant, natural element in the maintenance ofwater quality (Mitsch and Gosselink, 1986; Whit-ing and Pomeranets, 1997; Takatert et al., 1999;Hupp, 2000).

vstreg-sort-0,5/4

Head-water Upper

reach

Middlereach

Lowerreach

Sea

Water divide

Mineral soilOrganic sedimentsGyttja

Fig. 1. Division of a water course into sub-reaches withcharacteristic hydrogeology and geomorphology (afterNilsson et al., 2002).

8

The characteristics and function of a floodplain is in-timately linked to the river that flows upon it. Whenunregulated, floodplains are highly dynamic like mostfluvial landforms, and frequently inundated by over-bank floods. Compared to the upper parts of thecatchments groundwater level is shallow – near orabove soil surface – facilitating the build up of peat(Grootjans, 1985). Typically, the river is bordered bylevees consisting of relatively coarse material. The lev-ees are often the highest points on the floodplain,which may otherwise be extremely flat. This has theeffect that elevation differences of just few centimeterscreate differences in hydroperiod (length of inunda-tion), and hence a profound zonation in vegetationpattern (Mitsch and Gosselink, 1986).

Sediment trapping on a floodplain is strongly in-fluenced by the length of the hydroperiod, andhence by the height of the levees. The floodplainmay act as a very important sink for nutrients car-ried by the river. Permanent deposition of Phos-phorus (P) during flooding of a Danish lowlandriver floodplain was found to be 100 kg P ha-1year-1,equalling the loss of P from 200 ha agriculturalland (Kronvang et al., 2001). In a study prior to therestoration of the lower 18 km of the river Skjern,Denmark, mean annual riverine transport of sus-pended sediment (SS) and P was determined to12,200 t SS and 100 t P, respectively (Andersen andSvendsen, 1997). The effect of the restoration,mainly by allowing frequent flooding of riparianareas, was estimated to be a reduction in the trans-port of SS and P of 37% and 20%, respectively.

Analysis of time series for all major Danishstreams has shown an increasing trend over thepast 80 years in annual mean runoff (statisticallysignificant in 6 out of 32 stations). This increase ismost likely caused by the increase in precipitationof 0.76 mm yr-1 observed over the period 1874-1998(p=0.02%) (Ovesen et al., 2000). If these are con-tinuing trends the importance of floodplains asbuffers for water and sediment will increase.

2. Hydrology of wetlands

Hydrologically, a wetland is distinguished fromadjacent upland areas by the presence of water,which creates alternatingly or permanently satu-rated conditions. The consequential effect is sub-stantial water storage within wetlands, and thedevelopment of a readily identifiable wetlandvegetation, which is adapted to periodic anoxicconditions (Bradley and Gilvear, 2000). Thus, thewater-budget provides the framework from whichto investigate environmental conditions in a wet-land (Lent et al., 1997) and linkages between up-land, wetland, and river (Drexler et al., 1999), Fig.2. In spite of this only relatively few comprehen-

sive water balance studies of wetlands exist (e.g.Gilvear et al., 1993; Hyashi et al., 1998; Raisin et al.,1999). Since the surface and subsurface hydrologicprocesses within a wetland are inseparable (Rou-let, 1990) a water balance needs to quantify bothsurface and subsurface water fluxes. The main in-puts of water include precipitation (P), influentriver seepage (qi), overbank floods (qov), andgroundwater inflow (qgr_i). Outflows of water areevapotranspiration (ET), effluent river seepage (qo),surface runoff (qsu), and groundwater outflow(qgr_o). Problems in quantifying the water fluxes bydirect measurements have often been articulated inthe literature, especially concerning the ground-water flux (e.g. Gilvear et al., 1993; McKillop et al.,1999). The physical determination of this waterflux is commonly based on observations of hy-draulic head in piezometers and point-measurements of hydraulic conductivity, neitherof which necessarily can be extrapolated. A directmeans of determining the groundwater flux is bytracer study (e.g. Blicher-Mathiesen, 1998). In othercases groundwater flow models have been appliedinstead of measurements (e.g. Bradley, 1996; Zeeband Hemond, 1998; Stewart et al.,1998; Restrepo etal., 1998). Direct measurement of evapotranspira-tion is also often omitted and substituted by em-pirical formulas or determined as the residual inthe water-budget equation (Carter, 1986). Somewetlands are permanently water saturated, how-ever many wetlands exhibit partly unsaturatedconditions, typically during the growing season.The unsaturated zone is potentially very importantfor wetland ecology for a number of reasons: (i)water content in the root zone is a strong determi-nant of species composition of vegetation commu-nities due to different adaptions (e.g. Bridghamand Richardson, 1993; Grevilliot et al.1998; Silver-town et al., 1999, Schröder et al., 2002); (ii) the de-gree of saturation and thus aeration substantiallyinfluences decomposition of organic matter andmineralisation of nutrients (Ponnamperuma, 1984);(iii) the water-budget is affected since evapotran-spiration is dependent on soil water content(Brandyk et al., 1995), and with the existence of asubstantial unsaturated zone some or all precipi-tation infiltrating the surface will be retained hererather than percolate (Winter and Rosenberry,1995). However, in water balance studies of wet-lands the unsaturated zone is generally neglected.Exceptions are e.g. Bradley and Gilvear (2000),who described conditions in the unsaturated zonevia modelling, and Andersen (2002) who under-took measurements of saturated as well as unsatu-rated water balance components. The literature,however, contains examples on studies directedmore specifically towards the unsaturated charac-teristics of wetland soils; e.g. Bloemen (1982);Brandyk et al.(1985); da Silva et al.(1993); Weiss etal.(1998).

9

2.1 Hydraulic properties of peat

The hydraulic characteristics of wetland sedimentscan vary significantly over short horizontal andvertical distances (e.g. Zeeb and Hemond, 1998).This reflects the coincidence, in floodplain wet-lands, of alluvial coarse sediments and fine silt andclay from overbank deposition with heterogenousorganic deposits of peat (Bradley, 1996). The pres-ence of peat is a characteristic feature of manyfloodplains (Mitsch and Gosselink, 1986). Peat ac-cumulates where the decomposition of plant re-mains is retarded due to prevalent anaerobic con-ditions as in a permanently or intermittently watersaturated soil (Ingram, 1983). The hydraulic prop-erties of peat vary in relation to species composi-tion, humification and inorganic content. Further,the extreme compressibility of peat relative tomineral sediments have significant hydrologicalconsequences which so far are scarcely recognized(Price and Schlotzhauer, 1999). Peat is thus of spe-cial interest and will be treated here in some detail.

Peat is by definition (Vedby, 1984) a biogenic ma-terial containing at least 12 - 18% organic carbon inthe form of completely or partly decomposedplants deposited in a more or less anaerobic envi-ronment. Soil Taxonomy (Creutzberg, 1975)groups organic soils into three categories based ontheir fiber content (a fiber is a fragment of planttissue larger than 0.15 mm): fibric soils have a fibercontent of at least 3/4 of the soil volume, hemicsoils are intermediate between fibric and sapricsoils, where the latter has a fiber content less than

1/6 of the soil volume. A fibric peat is an unde-composed peat whereas a sapric peat is stronglyhumified. Since bulk density increases with humi-fication this parameter has also been used to char-acterize peat.

Structure of peat

Studies of the structure of peat (e.g. Loxham andBurghardt, 1986) reveal that peat is a highlystructured material with both connected and deadend pores. Very large pores can be present. Twoextreme types of peat can be described betweenwhich all other peat gradations exist. These twotypes are the highly humified ‘amorphous granu-lar’ peat, in which the soil particles are mainly ofcolloidal size and most of the pore water is ab-sorbed around the grain structure, and the unde-composed ‘fibrous peat’, which has essentially anopen structure with interstices filled with a secon-dary structural arrangement of non-woody finefibrous material. Porosity is thus at least to someextent a function of degree of humification. Boelter(1965) measured the variation in porosity of a mosspeat profile. Values ranged from 0.966 of the unde-composed peat to 0.833 of the decomposed peat.Vedby (1984) measured porosity of different peats.Values ranged from 0.757 to 0.972, highest for veryfibrous peat and lowest for less fibrous peat andpeat with a high amount of minerals.

The decrease in porosity with increasing humifica-tion is reflected in increasing bulk density (Boelter,1969; Paivanen, 1973). In the study of Vedby (1984)dry bulk density ranged from extremely low val-

Riv

er le

vel

fluct

uatio

ns

max.

min.

River flow reflects regional evapotranspiration and precipitation

The height of levees controls flooding

Evapotranspiration varies with vegetation type

Fall in water table reduces evapotranspiration – dependent on capillary and storage characteristics of floodplain sediments

Water table regime controlled by precipitation, evapotranspiration, groundwater input and river water level

Water table

Water ponding controlled by flooding, precipitation and infiltration capacity of top-layer sediments

Fig. 2. Key controls upon the wetland water balance along a floodplain transect.

10

ues in undecomposed Sphagnum peat; 0.034 - 0.041g cm-3, to 0.2 - 0.3 g cm-3 for very decomposed peat.However, given the higher specific density of min-erals (approx. 2.6 g cm-3) compared to peat (1.31 -1.38 g cm-3 (Galvin, 1976); 1.5 g cm-3 (Vedby, 1984) )the bulk density of peat is highly dependent on thesize of a possible inorganic fraction. Presence ofminerals can thus obscure the correlation betweenbulk density and degree of humification.

Water retention in peat

It is not only the total pore volume that changesupon humification, but also the pore size distribu-tion and hence the water retaining capabilities. Inan undecomposed peat the large pores will domi-nate whereas with increasing degree of humifica-tion the amount of fine pores will increase alongwith a decrease in total pore volume (Eggelsmann,1971). This means that total water content at satu-ration will be the highest for low humified peats.For high tension values (pF 3 - 4) on the other handthe water content is higher in the more humifiedpeat (Fig. 3). This implies that decomposition ofpeat apart from an increase in fine pores also re-sults in the formation of smaller particle sizeshaving larger specific surfaces facilitating wateradsorption since at high tension water moleculesare adsorped rather than being held by capillarity(e.g. Jury et al., 1991). For comparison with mineralsoils Fig. 3 also contains retention data from asandy and a loamy soil (Jacobsen, 1989).

For a non-swelling soil the relationship betweenpore water pressure and water content is repre-sented by the water retention curve (Fig. 3). Porewater pressure can be measured in the field by atensiometer. Ignoring hysteresis there thus exists aunique relation between a tensiometer reading and

actual water content. For a swelling soil such aspeat (see below) the relationship between porewater pressure and the water content is morecomplicated because the water content also de-pends on the load applied to the soil matrix(Towner, 1981). For a given water content the porewater pressure increases (becomes less negative)the higher the applied load. Thus there is a seriesof curves, one for each specified load applied tothe point of observation in the soil matrix, as sug-gested in Fig. 4. In order that the appropriate curvein Fig. 4 can be selected the applied load must alsobe known.

For the determination of total soil water potential,ψT, in a swelling soil, e.g. for flow calculations,however, one need not to know the applied load.This is due to the fact that the pressure potential,ψp, measured with the tensiometer encompasses alleffects on soil water other than gravity (specifiedin the gravitational potential, ψg) including the ap-plied load at the point of observation (e.g. Jury etal., 1991).

Compressibility

The analysis of transient flow in a saturated po-rous medium requires introduction of the conceptof compressibility. Equally important is the con-cept of compressibility in the understanding of theconsolidation (natural or man induced) of peatwhich is accompanied by drastic changes in per-meability. Following Hillel (1980) the term com-pression comprises two different processes orphases, viz. compaction - the compression of anunsaturated soil body resulting in reduction of thefractional air volume - and consolidation - thecompression of a saturated soil by squeezing outwater.

P0 = 0

P1 P2

0

θ

ψP

Fig. 4 Relationship for a swelling soil between the watercontent, θ, and pore water pressure, ψp, under differentload pressures, P0 < P1 <P2 on the soil matrix. AfterTowner (1981).

0

1.0

2.0

3.0

4.0

5.0

0 20 40 60 80 100

Vol. water (%)

pF

Humified peatUn-decomposed peatLoamSand

Fig. 3 Retention curves drawn from data published inLoxham and Burghardt (1986) and Jacobsen (1989). Theun-decomposed peat has a bulk density of 0.09 g cm-3,while the humified peat has a bulk density of 0.36 g cm-3.The sandy soil is “Jyndevad, 90 cm” having 2.5% clayand 0.5% silt. The loam is “Askov, 50 cm” having 24.4%clay and 11.5% silt.

11

Compression is defined (e.g. Freeze and Cherry,1979) as strain/stress - dε/dσ - the change in vol-ume, or strain, induced in a material under an ap-plied stress. The term is utilized for both elasticand nonelastic materials. When a stress (N m-2) isapplied to a saturated porous material there arethree mechanisms by which a reduction in volumecan be achieved: (i) by compression of water in thepores, (ii) by compression of the individual build-ing blocks of the soil skeleton (sand grains in thecase of a sandy deposit, organic molecules in thecase of a peat), and (iii) by a rearrangement of thesoil skeleton into a more closely packed configura-tion. The first of the mechanisms is controlled bythe fluid compressibility β, which can be consid-ered as a constant having a value for water of 4.4 x10-10 m2 N-1. The second mechanism is normally as-sumed negligible for mineral constituents; we willapply the same assumption regarding the com-pressibility of the constituents of a peat soil. In or-der to treat the third mechanism of volume reduc-tion we will have to introduce the Terzaghi-principle of effective stress (e.g. Skempton, 1961):Considering the stress on a normal plane in a satu-rated porous medium (Fig. 5) σT is the total stressacting downward on the plane. It is due to theweight of overlying solids and water. This stress isborne in part by the solid skeleton of the porousmedium and in part by the fluid pressure ψp of thewater in the pores. This is in contrast to shearstress which when applied to a porous mediummust be carried exclusively by interparticle forcesin the solid skeleton, since as first proposed byTerzaghi (1936) the water (and air) phase cannotcarry shear stress but only normal stress. The por-tion of the total stress that is not borne by the fluidis called the effective stress σe. This is the stress,that is actually applied to the constituents of theporous medium. Rearrangement of the solids andthe resulting compression of the solid skeleton iscaused by changes in the effective stress, not bychanges in the total stress. The two are related bythe equation

(1) σT = σe + ψp

As an example we will calculate the effective stressin depth z in a saturated soil without externalloading: The normal stress is the weight of theoverlying solids and water divided by the surfacearea of a normal plane

(2) σT = ρb g z

whereρb = (wet) bulk densityg = acceleration due to gravity

If there is no flow the water pressure is hydrostatic

(3) ψp = ρ g z

whereρ = density of water.

Thus effective stress is growing downwards in thesoil proportionally to the density difference

(4) σe = σT - ψp = (ρb - ρ) g z

For cases where the total stress does not changewith time - which holds true for many transientsubsurface flow problems according to Freeze andCherry (1979) - the effective stress at any point inthe system, and the resulting volumetric deforma-tions there, are controlled by the fluid pressures atthat point, as can be seen from (1).

Equation (1) has been shown (Skempton, 1961) tobe an excellent approximation to the truth for satu-rated soils. Experimental support for the validityof (1) also for negative pore water pressures hasbeen given (Bishop and Eldin, 1950), provided thesystem remains apparently saturated (that is, satu-rated by all practical measurable means). Thus, anegative pore water pressure contributes posi-tively to the effective stress in the solid frameworkof the soil system. For significantly unsaturatedsoils anomalies arise, because both air and waterare present in the voids, and in general these areunder different pressures owing to the curvatureof air-water interfaces. This means that a hydro-static pressure distribution does not exist in an un-saturated soil. However, an expression similar to(1) has been proposed for unsaturated soils:

(5) σT = σe + χψp

ψPFluid

pressure

σeEffective

stress

Total stressσT

Fig. 5 Total stress, effective stress, and fluid pressure on anormal plane through a saturated porous medium. After Freezeand Cherry (1979).

12

where χ denotes the degree of saturation. The im-plication of (5) is that also for negative pore waterpressures will a decrease in pore water pressure(more negative) increase the effective stress actingon the solid skeleton given a constant externalload. Experimental evidence for (5) has been given(Bishop, 1961; Jennings, 1961; Aitchison, 1961),however its validity has also been questioned (e.g.Bradford and Gupta, 1986). Van der Molen (1975)reports, based on experience with shrinkage due todessication of peats in the Netherlands, thatequating mechanical stresses to soil moisturestresses gives good results. The connection be-tween shrinkage and negative pore water pressurehas also been observed by Vedby (1984) while de-termining water retention characteristics on peatsamples.

The compressibility of a porous medium is de-fined as

(6)e

TT

dσ

/V-dV = α

whereVT = total volume = Vs + Vv

Vs = volume of solidsVv = volume of water-saturated voidsσe = effective stress

Typical values of compressibility (m2 N-1) are forclay 10-6 - 10-8 and for sand 10-7 - 10-9 (Freeze andCherry, 1979). Galvin (1976) made measurementson a number of different irish peat deposits. Hereported values of compressibility ranging from 8x 10-6 to 2 x 10-5; the younger the peat the larger thevalue of compressibility.The high compressibility of peat and the increasein effective stress downwards in a peat profile (4)combined with the continuous increase in totalstress due to the increase in load in the form of ac-cumulating plant residues on the peat surface, re-sult – together with decomposition of the peat - inthe often observed increase in bulk density andcorresponding decrease in porosity and perme-ability with depth (e.g. Boelter, 1965; Vedby, 1984,Gafni and Brooks, 1990). The application of a loadincreases on the first hand the pore water pressure.During the process of consolidation the pore waterpressure in excess of hydrostatic pressure slowlyequalizes by a release of water possibly in the formof upward flow as observed by Dasberg andNeuman (1977).

Partly de-watering of a peat deposit consolidatesthe remaining saturated part by two mechanisms:(i) according to equation (1) a decrease in porewater pressure increase the effective stress actingon the deeper layers. A lowering of the water tableof 1 m decreases the hydrostatic pressure by 9807N m-2 (1 m H2O) and hence increases effectivestress with the same amount. (ii) De-watering theupper layer of the peat increases effective stressacting on the deeper layers by increasing the effec-tive weight of the unsaturated peat. Since there isno continuous hydrostatic contact between thewater in the unsaturated zone and the water in thesaturated layer, the weight of the water in the un-saturated zone cannot be transferred to the waterin the saturated layer. Instead the water stored inthe unsaturated zone adds to the weight of thepeat here. The water content at field capacity isconsiderable (Fig. 3) meaning that de-wateringincreases the effective weight significantly. Theeffective weight of the unsaturated peat is calcu-lated as the sum of the dry bulk density and thewater content at the actual pore water pressure.This effective weight is unreduced transferred tothe solid skeleton of the saturated peat as effectivestress.

Storage

Another property of a porous medium to be con-sidered when dealing with transient flow proc-esses is the storage capacity of the medium. For asaturated aquifer the specific storage Ss (m

-1) is de-fined as the volume of water that a unit volume ofaquifer releases from storage under a unit declinein hydraulic head, hT. Since hT = p + z (p = ψp/ρg),and z being constant at the point in question, a de-crease in hydraulic head infers a decrease in fluidpressure and an increase in effective stress σe. Thewater that is released from storage under condi-tions of decreasing hT is produced by two mecha-nisms: (i) the compaction of the aquifer caused byincreasing σe, and (ii) the expansion of the watercaused by decreasing ψp. The first of these mecha-nisms is controlled by the aquifer compressibility αand the second by the fluid compressibility β. Thespecific storage Ss is given by (7)

(7) Ss = ρg(α + nβ)

wheren = porosity.

Since water is not significantly compressible underthe range of pressures encountered in shallowsystems (7) reduces to Ss = ρgα.

13

For a confined aquifer (i.e. the pore water pressureat the top of the aquifer is higher than atmosphericpressure) the storativity (or storage coefficient) S isdefined as Ssb, where b is the thickness of the aqui-fer. The storativity is normally determined by indi-rect methods such as pumping test analysis. Ac-cording to Neuman and Witherspoon (1972) thespecific storage Ss can also be inferred from con-solidation curves. Specific storages range from 10-4

m-1 (clay) to 10-6 m-1 (sand) (Rosbjerg, 1987). Das-berg and Neuman (1977) report specific storage forpeat in the Hula Basin in Israel ranging from 0.086m-1 to 0.118 m-1, with the lowest values for deepermore compacted peat. The implication is that aunit volume of saturated peat can release about 105

times more water due to compression than a unitvolume of sand, when the hydraulic head drops byone unit.

For an unconfined aquifer (i.e. under water tableconditions where the pore water pressure in thetop of the aquifer is at atmospheric pressure) thestorage term is traditionally known as the specificyield, Sy. Sy is defined as the volume of water thatan unconfined aquifer releases from storage perunit surface area of aquifer per unit decline in thewater table. The usual range of Sy for mineral soilsis 0.01 - 0.3 (Rosbjerg, 1987) and can be inferredfrom the water retention curve as the difference inwater content between saturation and field capac-ity. Accepting field capacity as represented by anegative pore water pressure of 1 m H2O Sy valuesfor peat ranging from 0.1 - 0.3 can be read out ofpublished water retention data by Dasberg andNeuman (1977) and Loxham and Burghardt (1986).Price (1996), from a Canadian cutover bog, re-ported a range from 0.048 to 0.55. Lower values arefor decomposed peat, and higher values are forundecomposed or living mosses. Andersen (un-published results) found for floodplain peat de-posits of sedges and herbaceous plants and with aconsiderable fraction of inorganic material a morenarrow range from 0.09 to 0.26 for hemic to fibricpeat. The relatively high values for unconfinedaquifers reflect the fact that releases from storagein unconfined aquifers represent an actual de-watering of the soil pores, whereas releases fromstorage in confined aquifers represent only secon-dary effects of water expansion and aquifer com-paction. Since for mineral soils Sy is normally somuch higher than Ss in unconfined units, Ss is oftenignored in the overall aquifer storativity, where Stot

= Sy + bSs. However, also in studies involving peathydrology the significance of storage changescaused by peat compaction has typically been ig-nored (Price and Schlotzhauer, 1999). Price andSchlotzhauer (1999) measured Sy and Ss in an un-

confined cutover peat deposit with a fluctuatingwater table and found Ss was in fact 67 to 170%larger than Sy. Considerable hysteresis was ob-served with Ss averaging 0.094 m-1 during dryingperiods, but only 0.026 m-1 on rewetting. Thus, asalso recommended by Dasberg and Neuman(1977), compressibility must not be neglected indealing with fluid flow through peat layers, irre-spective of whether these layers are confined orunconfined.

In the unsaturated zone storage is given by themoisture content θ. Changes in storage are accom-panied by changes in the pressure head p, throughthe θ (p) relationship displayed on the water re-tention curve (Fig. 3). The slope of this curve rep-resents the unsaturated storage property of a soil,where the specific moisture capacity C is definedas

(8) C = dθ / dp

Saturated hydraulic conductivity

Permeability or hydraulic conductivity, Ks, wasearly recognized as a crucial hydraulic variable inthe study of the hydrology of peatlands. As a re-sult, Ks became the most studied parameter of or-ganic soils according to Gafni and Brooks (1990).However, different methods applied coupled withthe variation among organic soils mentioned abovehave yielded a wide range of values. In a reviewby Rycroft et al. (1975a) some seven orders ofmagnitude of Ks- values are reported for peat. Sev-eral researchers who have studied the variation inhydraulic conductivity have noted a negative ef-fect of depth and degree of decomposition on Ks

(Boelter, 1965; Paivanen, 1973; Gafni and Brooks,1990). Boelter (1969) found on basis of 119 samplesfrom 12 Minnesota bogs a linear relationship be-tween Ks and the degree of decomposition ex-pressed as fibric, hemic and sapric peat, respec-tively. Table 1 lists the range in Ks-values togetherwith bulk density and porosity. Table 2 lists simi-lar characteristics, but for floodplain depositswhere peat is mixed with inorganic sediments de-rived from floodings with surface water.

Table 1. Range in physical characteristics of fibric, hemicand sapric peat materials from northern Minnesota bogs.After Boelter (1969).Organicmaterial

Bulk density Totalporosity

Hydraulicconductivity

g cm-3 % 10-7 m s-1

Fibric < 0.075 > 90 > 180Hemic 0.075 – 0.195 85 - 90 2.1 - 180Sapric > 0.195 < 85 < 2.1

14

Table 2. Physical characteristics of fibric and hemic peatfrom a Danish floodplain wetland. Organic content was67% and 40% for fibric and hemic peat, respectively (af-ter Andersen, 2002).Organicmaterial

Bulk density Totalporosity

Hydraulicconductivity

g cm-3 % 10-7 m s-1

Fibric 0.16 90.8 1100Hemic 0.36 82.5 0.6

Table 1 and Table 2 demonstrate the range inphysical characteristic encountered for peat mate-rials in various degrees of decomposition. Alsoillustrated is the effect on bulk density of inclusionof inorganic material originating from flooding ofthe floodplain wetland. From the tables it is clearthat the hydraulic conductivity can be indeed verylow for well decomposed peat even though totalporosity remains high. Various explanations havebeen offered for this apparently anomalous be-haviour: clogging of the pores by fine peat parti-cles and colloidal humus (Rycroft et al., 1975a),compression of the peat matrix due to overburdenpressure stress (Hemond and Goldman, 1985), oc-clusion of the pores by bubbles of methane gasgenerated by in situ anaerobic decomposition ofplant biomass (Mathur and Levesque, 1985). Allexplanations appear plausible. Chow et al., (1992)measured Ks in a compaction experiment on amoderately decomposed fibrous Sphagnum peat. Ks

decreased 3 orders of magnitude when dry bulkdensity was increased from 0.1 g cm-3 to 0.2 g cm-3.This was primarily due to a reduction of theamount of macropores: the proportion of pores 148µm – 594 µm decreased from 17.5% to less than 2%,while the proportion of pores smaller than 59 µmremained relatively constant. It has been shown(Loxham and Burghardt, 1986) using chemicaltracer tests on undisturbed peat samples, that atsaturation water flows through a few intercon-nected channels bypassing the rest of the pores.Also, experimental evidence for the probability ofthe theory of occlusion by methane gas bubbleshas recently been obtained in laboratory experi-ments by Reynolds et al. (1992). They comparedthe course of water content, θ, and Ks over timebetween sterilized and non-sterilized peat columnsflowed continuously under positive pressure. Overtime periods ranging from 44 to 78 days, Ks de-creased on average by 1.6 orders of magnitude, θ(measured in situ with TDR) decreased by 20 per-centage points, and gaseous methane increased 12-fold in the non-sterillized columns. In the sterilizedcolumn on the other hand, hydraulic conductivityand water content increased to stable, relativelyhigh values (explained by the gradual removal ofartifact carbon dioxide gas produced by the sterili-zation process), while only trace levels of methane

was detected. By flushing the non-sterilized col-umns with low pressure helium gas and resatu-rating them it was possible to regenerate the initialhigh values of water content and hydraulic con-ductivity, indicating that the observed decline wasnot due to changes in the physical peat matrix.

Measurement of saturated hydraulic conductivityEven though the literature displays examples onlaboratory measurements of hydraulic conductiv-ity in peat, the bulk of the published data are de-rived from field measurements. When field andlaboratory measurements have been comparedthose obtained in the laboratory have usually beenhigher (Boelter, 1965; Paivanen, 1973; Dasberg andNeuman, 1977; Schlotzhauer and Price, 1999). Thisis explained by possible damage to the peat uponsampling, increased decomposition in the labora-tory, leakage along the interface between the in-side wall of the sampling cylinder, and a possibleeffect of removing vertical stress.

In the field hydraulic conductivity has tradition-ally been measured conducting a head recoverytest where water is added to (slug test) or removedfrom (slug withdrawal or bail test) a standpipepiezometer and the recovery to the original waterlevel is recorded. The theory describing the re-sponse time of open hydraulic piezometers in rigidsoil was developed independently by Kirkham(1945) and Hvorslev (1951). Hvorslev (1951) devel-oped a solution for equalisation during a head re-covery test for any shaped piezometer and gaveshape factors describing the geometry of the flowfield around the piezometer. His theory was de-veloped assuming an incompressible and isotropicsoil or an anisotropic soil with the principal direc-tions of anisotropy coinciding with the horizontaland vertical plane. In the case of anisotropy theratio of horizontal to vertical hydraulic conductiv-ity need to be known. The method of Hvorslev(1951) has predominantly been used in aquiferstudies. The piezometer method of Kirkham (1945)has mainly been employed in agronomic researchand peatland studies. In this method (as describedin Luthin and Kirkham, 1949) a standpipe pie-zometer with a fairly small diameter is installed inthe soil and a cavity is augered below the pipe(Fig. 6). The value of Ks is calculated by the equa-tion (Kirkham,1945) :

(9) t

o2

yyln

trA

K π=

wherey0 = initial change in water levely = water level at time t

15

r = radius of cavityA = shape factor depending on the ge-

ometry of the system.

For the Kirkham piezometer method shape factorshave been developed using electrical analogues(Frevert and Kirkham, 1948; Luthin and Kirkham,1949; Youngs, 1968) and more recently by finiteelement modelling (Brown and Hodgson, 1988).

Anisotropy of the soil, usually considered in verti-cal and horizontal direction only, can be estimatedby comparing Ks obtained with methods that differin the direction in which Ks is measured - prefera-bly on the same soil region to minimise the effectof soil heterogeneity (Bouwer and Jackson, 1974).Topp and Sattlecker (1983) claimed to measurevertical and horizontal conductivity directly byusing pipes either open only at their basal end oropen only at a perforated section, but closed at thebottom. According to Boelter (1965) certain peatmaterials have a very pronounced horizontallaminar structure, particularly the herbaceouspeats formed from remains of reed- and sedge-type plants. Thus anisotropy might be expected.Boelter (1965), however, in his work on peat invarious degrees of decomposition found no sig-nificant anisotropy. Schlotzhauer and Price (1999)measured Ks in the laboratory on core samples of aSphagnum peat and found that vertical conductiv-ity was 4 times less than horisontal conductivity.

Non-Darcian flow in peat?

A number of workers have reported apparentlyanomalous head recovery test results which seemthat hydraulic conductivity is time dependent andincreases with the applied head. This was inter-preted as non-Darcian flow behaviour in peat (Daiand Sparling, 1973; Ingram et al., 1974; Rycroft etal., 1975b; Waine et al., 1985). Hemond and Gold-man (1985) proposed that much of the apparentnon-Darcian behaviour in peat was an artifact ofthe inappropriate application of steady-state pie-zometer formulae to transient experiments in amaterial having significant non-zero elastic stora-tivity, Ss (see above): during a piezometer test withfalling or rising head pore water pressures near theporous/slotted tip are altered. One result is achange in effective stress, σe, since total verticalstress, σT, is constant (equation 1) leading to com-pression or expansion of the peat matrix and tran-sient changes in water content of the peat. Hence afraction of the water measured initially in the pie-zometer test represents pore water coming into orout of storage near the well - not a steady-stateflow field. This leads to higher rate of water levelchange in the early stages of a rising or fallinghead test and to higher calculated values of Ks overshort periods. The observed dependency in Ks onapplied head was explained by expansion of hori-zontal flow paths in the peat under decreased ef-fective stress when pore water pressure is in-creased, i.e. as long as the elevated head is main-tained in the piezometer

Unsaturated hydraulic conductivity

The literature exhibits only few examples on thesubject of unsaturated hydraulic conductivity ofpeat. The interest in the subject has primarily beenin investigating the ability of a drained unsatu-rated peat layer to supply plants with water froman underlying water table by capillarity.

It is experimentally demonstrable (Loxham andBurghardt, 1986) that shrinkage, which accompa-nies the early stages in drainage, is attributable tocollapse of the large channels without significantchange in the fine (micro-pore) capillaries. Hence,at least in less decomposed peat with a relativelylarge proportion of macro-pores an initial sharpdecrease in hydraulic conductivity must be antici-pated upon de-saturation.

Only laboratory measurements have been en-countered in the literature collected for this study:Bartels and Kunze (1973) using a double-membrane apparatus; da Silva et al. (1993) and

H

hc

S

yPipe

Cavity

Impermeable layer

2 r

Fig. 6 Geometry and symbols for the piezometer method.Typical values for r and hc are 0.75 - 1.5 cm and 5 - 10 cm,respectively. After Bouwer and Jackson (1974).

16

Olszta (1977) using steady state flux control meth-ods. Bartels and Kunze (1973) examined the effect fdecomposition, flux direction, mineral content andbulk density. The correlation to unsaturated con-ductivity decreased in the sequence: decomposi-tion, flux direction, mineral content, bulk density.The authors found conductivity to be dependenton flux direction, i.e. anisotropy, with horizontalunsaturated conductivity being 3 - 4 times largerthan vertical, the ratio increasing up to pF = 2.5.However for pF = 3 they found no difference be-tween horizontal and vertical conductivity. As ex-pected, it was shown that unsaturated hydraulicconductivity generally is larger in the less humi-fied peat. Even at high tension there are consider-able amounts of water in the peat (e.g. Fig. 3: 40 %at pF = 3) meaning that there are many water filledpores. From the retention curve it is seen, that wa-ter in the less humified peat is stored in relativelylarger pores (i.e. lower tension at the same volu-metric water content) than in the more humifiedpeat. According to Poiseuilles Law the fluxthrough a capillary tube is proportional to thesquared radius indicating a higher unsaturatedconductivity in the less humified peat.

Saturated hydraulic conductivity of mineral soils isgenerally larger than of peat – except for the un-decomposed peat. However, the picture is re-versed for negative pore water pressures. The rea-son for the higher unsaturated conductivities inpeat is the larger amount of water filled poresrelative to mineral soils at the same tension. Inwater filled pores water will flow due to a givengradient at a higher velocity than water creepingalong the hydration films over the particle surfacesin desaturated pores (Hillel, 1980).

Measurement of unsaturated hydraulic conductiv-ity is time consuming , involving many technicaldifficulties and is subject to measurement errorsdue to the swelling/shrinking properties of peat.In recent years efforts have been made to applyvarious mathematical functions to describe andpredict unsaturated hydraulic conductivity. Thesefunctions have in general been developed for min-eral soils but also by some researchers soughttransferred to organic soils: Bloemen (1983) - basedon 227 measurements of air entry value, saturatedhydraulic conductivity, Ks, and bulk density - de-veloped a modified Brooks and Corey-expressionrequiring only bulk density to be known. Brandyk(1985), da Silva et al. (1992), Brandyk (1995) usedthe van Genuchten formula for unsaturated hy-draulic conductivity which requires two parame-ters that can be derived from the retention curve,besides Ks as a matching factor. da Silva et al.

(1993) compared estimated unsaturated hydraulicconductivities to measured values and found verygood agreement.

2.2 Evapotranspiration from wetlands

Evapotranspiration (the combined processes ofevaporation and transpiration) is a fundamentaland often a major component of the hydrologicalcycle of wetlands (e.g. Lafleur, 1990; Doss, 1993;Campbell, 1997). Furthermore, evapotranspirationinfluences on nutrient cycling by affecting subsur-face water flow-pattern. Water lost by evapotran-spiration may be replenished by capillary rise re-sulting in a more upward flow, whereby nutrientsare transported into soil layers with e.g. a highernutrient turnover or binding capacity (Hoffmannet al., 1993; Andersen, 2002).

The most important factors influencing evapotran-spiration are: (i) input of energy as global radiationand advection; (ii) vapour pressure deficit betweensurface and air; (iii) air movement; (iv) vegetationspecific factors: stomates, root depth, crop speciesand crop structure; (v) soil specific factors: poros-ity, pore size distribution, permeability (deter-mining the movement of water in the soil).

Canopy radiation balance

The radiant energy available at the surface of theearth to drive physical and biological processesoriginates from the sun. At the canopy level, theproperties of the canopy influence the energy ex-change and thereby determine the energy availablefor canopy processes.

At the top of the atmosphere the incoming solarradiation (wave lengths 0.3 - 3 µm) is approxi-mately constant 1373 W m-2 (the solar constant). Onaverage 25% of this radiation is attenuated due tomolecular and particle scattering and absorptionparticularly by water vapour on its way throughthe atmosphere. Part of the remaining radiationreaches the ground in a direct solar beam (Sb) andpart of it (typically 15 - 25% in clear-sky conditionsand up to 100% in overcast conditions) in a diffuseform after being scattered (Sd).

Part of the short-wave radiation is reflected. Thereflection coefficient, ρ, of a particular surface isstrongly dependent on the solar elevation. The re-flection coefficient is the average reflectivity overall wave lengths in the solar spectrum (0.3 - 3 µm).For natural surfaces it is often called the albedo. Ingeneral, ρ for forests is 0.11 - 0.16, and for agricul-tural crops the values are higher, ρ = 0.15 - 0.26. In

17

a review Burba et al. (1999) found the followingalbedo values for wetlands: 0.12 – 0.16 for Sphag-num-sedge bogs; 0.12 for swamp forest; and 0.11 –0.17 for open Sphagnum fen.

The net short-wave radiation (Sn) is the part of thetotal incident short-wave radiation (termed globalradiation, St) that is captured at the ground sur-face. Sn is given by (10)

(10) Sn = (1 - ρ) St = (1 -ρ) (Sb + Sd)

where Sb and Sd are direct and diffuse radiation,respectively.

At longer wave lengths (3 - 100 µm) radiation isexchanged between the ground and the atmos-phere by blackbody (or full radiator) emission. Asthe ground, on average, is warmer than the atmos-phere, and as the atmosphere is not a full radiator(only clouds are full radiators) the net result is aloss of energy as thermal radiation from theground.

The net long-wave radiation (Ln) is given by (11)

(11) Ln = Li - Lo = εaσTa

4 - εsσTs

4

Li is the incoming long-wave radiation to the sur-face with temperature Ts, and Lo is the outgoinglong-wave radiation to the atmosphere with thetemperature Ta. σ is the Stefan-Boltzman constant,and εa (< 1) and εs (≈ 1) are the apparent emissivi-ties of the atmosphere and the surface of the earth,respectively. For most surfaces 100 W m-2 may be agood average Fig. for the net loss to a clear sky,whereas in overcast weather the net loss ap-proaches 0 W m-2.

The net flux of all radiation across unit area of aplane is called the net radiation, Rn. Rn is given by(12)

(12) Rn = (1 -ρ) St + Li - Lo

Due to the throughflow of cold groundwater andthe higher water content and the consequentlyhigher soil heat capacity, the soil temperature ofwetlands will be lower than that of mineral uplandsoils. This leads to relatively lower outgoing long-wave radiation resulting in higher net radiationover the wetland. Andersen et al., (2002) foundthat the lower soil temperature and lower albedoof a fen wetland increased net radiation by 20%relative to a cultivated grass field.

The radiant energy intercepted by a plant is partlyabsorbed, partly reflected, and partly transmittedas seen from the identity (13)

(13) α + ρ + τ = 1

where α is the absorption coefficient, and τ is thetransmission coefficient. The radiative transferwithin canopies depends on the architecture of theplants and is very complex to describe. A simpli-fied way of describing the distribution of radiantenergy within a canopy is to use Beer’s law and anempirical extinction coefficient, K. The averageirradiance at any level z in the canopy is related tothe irradiance above the canopy, S(0), and to theaccumulated leaf area L(z) from the top of the can-opy down to level z:

(14) S(z) = S(0) . exp(-K . L(z))

Energy balance of an evaporating surface

The energy budget of a vegetated surface can beexpressed in terms of available energy being parti-tioned into latent and sensible heat. Considering avolume of a plant stand of unit cross section, ex-tending from a plane below the soil surface, wheretemperature changes are negligible, to a plane atreference level above the canopy, where the netradiation balance, Rn, is determined, the energybalance is given by (15)

(15) Rn - G - J - M - Ad = λE + H

G is the heat conducted into the soil (soil heat flux),J is the sensible and latent heat temporarily storedwithin the volume, M is the net energy absorbedby metabolism (photosynthesis minus respiration),and Ad is the net loss of energy due to horizontaladvection by air movement. E is the rate of evapo-ration from the soil surface and the vegetation, λ isthe latent heat of vaporisation of water, and H isthe upward flux of sensible heat by thermal con-vection.

M and J are often ignored since they only consti-tute a small fraction of the overall energy balance.In ‘oasis’ conditions the horizontal transport (ad-vection) may be large, but generally Ad is neglectedin order to obtain a simplified one-dimensionalenergy balance:

(16) Rn - G = λE + H

However, Devitt et al.(1998) found the energy bal-ance for a riparian corridor in Southern Nevada tobe dominated by advection in one out of two years

18

of study. Andersen et al. (2002) found evapotran-spiration of a floodplain fen wetland to be en-hanced by local advection in periods with highevaporative demand and low precipitation. Duringthese periods warm, dry air was formed over drierupland areas. This is a condition opposite to thatreported in a number of other wetland studies (e.g.Lafleur, 1990; Souch et al., 1998) in which proxim-ity to large water bodies result in humid sea/lakebreezes flowing across the wetland during thedaytime thus preventing advective enhancementof evapotranspiration rates.

The soil heat flux, G, is an energy storage term,being the change in sensible heat in the soil de-pending on the soil temperature gradient and onthe thermal properties (specific heat, cs, and ther-mal conductivity, k) of the soil. These propertiesdepend on soil type and soil water content. Duringthe day, G is largest around midday when the airtemperature and net radiation to the ground are atmaximum. In wetlands with open water G can be aconsiderable sink of energy during daytime (G/Rn

= 0.2 – 0.4) (Souch et al., 1998; Burba et al., 1999).On a daily (24 hr) basis Burba et al., (1999) found Gto be small, whereas Souch et al. (1998) even on adaily basis found G to consume 30% of Rn. Ander-sen et al.(2002) found G of a floodplain wetland toconstitute only 5% and 1.5% of Rn, respectivelyduring daytime and on a daily (24 hr) basis. Theyattributed the low values to the presence a mat ofdead vegetation and a dense canopy, which effec-tively isolated the soil surface from incident radia-tion.

The rate of transfer of sensible heat (H) from a sur-face into the atmosphere is proportional to thetemperature gradient between the surface and theenvironment and is governed by the magnitude ofan exchange coefficient that depends on the tur-bulent conditions above the surface.

Similarly, the rate of evaporation (λE) is propor-tional to the vapour pressure gradient between theevaporating surface and the environment, and isgoverned by the turbulent exchange conditions forlatent heat.

Wetland evapotranspiration rates

Two reviews of wetland hydrology (Linacre, 1976,and Ingram, 1983) conclude that studies of evapo-transpiration over a wide range of wetland typeshave produced a number of conflicting and incon-clusive findings. Of particular controversy hasbeen the influence of vascular vegetation cover onwetland evapotranspiration. Some authors suggest

that the presence of vegetation increases evapo-transpiration above lake (i.e. open water) evapora-tion, E0, while others maintain that evapotranspi-ration from vegetated wetlands is always less thanlake-evaporation. Ingram (1983) who extensivelyreviewed the literature on evapotranspiration fromfens and bogs, concluded that vegetation coverswith differing characteristics (e.g. species compo-sition, stand density and height) have varying de-grees of influence on evapotranspiration. On theother hand, Linacre (1976) stated in his review ofevapotranspiration from swamps, that the pres-ence and nature of vegetation have relatively mi-nor influence on evapotranspiration rates, com-pared with regional climate and local advection ofsensible heat. One important difference betweenSphagnum-dominated wetlands and wetlands withvascular vegetation is, that the roots of the vascu-lar vegetation are capable of extracting soil waterso as to maintain a high transpiration rate eventhough the water table drops. This is in contrast tothe situation in Sphagnum-dominated wetlands.Several studies report that the level of the watertable affects the evaporative capacity of the Sphag-num surface (Ingram, 1983, Phersson and Petter-son, 1997). This is due to the lack of vascular tissueand low matric potentials, which means that muchlarger surface resistances are imposed as watertable declines below the ground surface.

There are significant correlations between stomatalconductance and habitat (Roberts, 2000), withwetland plants having a relatively high stomatalconductance. However, very little is known aboutthe water requirements and consumption of wet-land species (Tabacchi et al., 2001), except for man-influenced, homogeneous woody communities(poplar plantations, coppiced willows). Table 3 is areview of evapotranspiration rates measured inwetland studies. In order to compare studiesacross wetland types and different climates Pen-man’s potential open water evaporation, E0, hastraditionally been used as a reference. E0 , however,cannot be considered a true reference since net ra-diation has been actually measured at the experi-mental sites over vegetated surfaces with albedosdiffering from a free water surface. Table 4 listsresults from three Danish wetland studies in whichmeasured evapotranspiration rates have beencompared to an independent reference. In thestudies of Vedby (1984) and Hoffmann et al. (1993)evapotranspiration was measured using vegetatedlysimeters, which could be weighed, 1 m deep andwith a surface area of 1000 cm2.

19

Table 3 Mean daily evapotranspiration rates including range, E/E0 and λE/Rn reported in wetland studies.

Wetland type E (mm day-1) E/E0 λE/RnAuthor

Floodplain wetland, Den-mark

3.6a (0.8 – 5.6) 1.10 0.82 Andersen et al. (2002)

Sub-Arctic coastal wetland Dry site

Wet site

2.6b (1.0 – 4.5)

3.1b (1.4 – 6.0)

0.74

0.90

0.50

0.57

Lafleur (1990)

Quaking fens, Netherlands 2.5c (1.0 – 4.1) 0.77 Koerselman and Beltman (1988)Transition peatland, Japan 2.5d ( - 4.6) 0.37 – 0.68e Tagaki et al.(1999)Open water marsh/sedgemeadow, Indiana DunesNational Lakeshore

3.58f 1.03 0.44 Souch et al. (1998)

Prairie wetland, Nebraska 3.75g (0.5 – 6.5) 0.8 – 1.0h

0.6 – 0.8i

0.3 j

Burba et al., (1999)

Lakeshore Typha marsh,Ontario

4.9k (3.5 – 6.3) 1.0 0.78 Price (1994)

Czechoslovakia

Willow carr

sedge-grass marsh

3.5l (2.4 – 4.9)

3.0l (2.3 – 3.7)

0.86

0.72

Priban and Ondok (1986)

Raised peat bog, New Zea-land, dry canopy

1.54m ( - 2.13) 0.34 0.23 Campbell and Williamson (1997)

aDaylight values, April – SeptemberbDaylight values, May – AugustcDaily values, April – October. Average for three vegetation typesdDaily values, June – OctobereRead from Fig. 6 in Tagaki et al.(1999)fDaylight values, JunegDaily values, June – OctoberhEarly and peak growthiDuring senescencejAfter senescencekHours 600 - 1800, June – AugustlDaily values, July – September. Range of monthly meansmDaylight values, November – March.

Table 4. Ratio of actual to reference evapotranspiration reported in Danish wetland studies.

Wetland type E/Eref AuthorFloodplain wetland, frequently flooded. Sedge-herbs-grass community. 1.28a Andersen et al. (2002)Undrained riparian meadow vegetated with herbs and grasses. 1.04b Hoffmann et al. (1993)

Sphagnum mire hummock hollow

0.93c

1.21 c

Vedby (1984)

Riparian fen wetland, frequently flooded. Vegetated with grasses and wil-low shrubs.

1.29 c Vedby (1984)

Cutover Sphagnum mire vegetated with shrubs and trees. 0.86 c Vedby (1984)aDaylight values, April – SeptemberbDaily values, full yearcDaily values, April – October.

20

Reference evapotranspiration, Eref , was calculatedusing the formula of Penman (1956) with parame-ters measured at a nearby climate station over avegetation cover of short, well-watered grass. An-dersen et al. (2002) estimated evapotranspirationusing the Bowen ratio energy balance approach.Following the recommendation by FAO (Allen etal., 1998) they calculated a reference evapotranspi-ration using the Penman-Monteith method for ahypotethical crop with an assumed height of 0.12m, with surface resistance of 70 s m-1 and an albedoof 0.23. Net radiation was calculated using meas-ured values of incident shortwave radiation, airtemperature and relative humidity as prescribedby Allen et al. (1998). It should be noted, however,that the method of Allen et al., (1998) has beenshown (Detlefsen and Plauborg, 2001) to yield re-sults which in Denmark are 7 – 13% below evapo-transpiration calculated by the Penman formula(Penman, 1956), when the Penman formula wasapplied as in Vedby (1984) and Hoffmann et al.(1993).

2.3 Wetland waterbalances

In order to illustrate the range in hydrologic con-trols on riparian wetlands follows a presentation oftwo water balance studies. The studies were car-ried out in the same river system. One study wascarried out on a tributary stream to the river Gjern,Denmark (Dahl, 1995) (Fig. 7), while the other wasdone further downstream on the floodplain of thelower Gjern (Andersen, 2002) (Fig. 8).

The stratigraphy of the wetland on the tributarystream is 50 cm of fibric, highly conductive peat ontop of 2 m mixed layers of peaty sand underlain bygyttja. The wetland is situated below a steep slope.Due to the topography and sedimentology in-flowing groundwater is forced upwards and formsoverland flow at the foot of the hillslope. Becausethe wetland is fed by a constant and largegroundwater influx of both local and regional ori-gin the surface water level is very stable (10 cmabove the ground surface).

Groundwaterrecharge

8377

Groundwaterdischarge

833

Precipitation810

Actual eva-potranspiration

624

Overlanddischarge

7731

Water level all year

Ground surface

Fig. 7. Water balance of a riparian wetland located on atributary stream to river Gjern, Denmark. Numbers aremean values for a dry year, 1992, and a wet year, 1993(mm yr-1). (After Dahl, 1995).

The floodplain wetland is also underlain by gyttja.On top of the gyttja is 4 –5 m peaty-sandy deposits.The peaty sand is overlain by 1.5 m mainly hemicpeat, and a 40 cm silty-clayey layer of sedimentsdeposited by flooding river water constitutes thesurface layer. The gyttja prevents hydraulic contactbetween the floodplain and deeper groundwateraquifers.

Flooding returning to riverca. 10000

Precipitation1004

Actual eva-potranspiration

666

Floodingca. 10000

Winter

Summer

Storage change +422

Storage change -10Groundwater

recharge46

Riverseepage

54Water level

Fig. 8. Water balance for a floodplain wetland on thelower river Gjern, Denmark for the calendar year 1999(mm yr-1). (After Andersen, 2002).

Hydraulic conductivity of the sediments below theriver bottom is also low meaning that exchange ofwater between the river and the floodplain takesplace only through and above the riverbank. Thereis a small influx of groundwater of local originfrom the hillslope to the floodplain. During thegrowing season an unsaturated zone developswith a maximum depth of 60 cm. Evapotranspira-tion is maintained at a high rate because precipita-tion is supplemented by capillary rise of waterfrom storage within the underlying peat layer.During autumn, winter and spring the floodplainis frequently flooded by the river. Large amountsof water flow across the riverbank during flooding

21

events, and even though the vast majority returnsto the river the levees, which rise 40 cm above thefloodplain, trap some of the water. Due to the lowconductivity of the silty-clayey layer infiltration islow and water ponds on the floodplain most of thetime outside the growing season.The wetland on the tributary stream had a waterturnover rate of 8 times per year corresponding toa hydraulic residence time of 44 days. This wet-land is characterised by having a stable flow pat-tern controlled by groundwater recharge and be-ing hydrologic open. The floodplain wetland hasthree storages, namely the surface water storage,the soil water storage and the groundwater stor-age. During a flooding event the hydraulic resi-dence time of the surface water storage is in theorder of hours. Water ponds on the surface in be-tween flooding with an average depth of 17 cm.Since the average number of floods is 4 per yeardistributed over 7 months hydraulic residencetime is ca. 50 days. The soil water storage has aresidence time in the same order of magnitude, 46days, whereas flow through the groundwater stor-age is very slow with a residence time of 11.7years. Thus the floodplain wetland is hydrologi-cally closed concerning groundwater flow andgoverned by vertical water exchange with the at-mosphere and with ponding water. Large amountsof surface water flow across this wetland duringflooding situations and concerning surface waterexchange this floodplain wetland is a very opensystem.

3. Nutrient processes in wetlands

3.1 Chemical conditions of a waterlogged soil

Redox conditions and pH

The determining characteristic of a waterloggedsoil is the absence of oxygen (Reddy et al., 1980).Since oxygen has a low solubility in water (11.28

mg O2 l-1) and the diffusivity of oxygen in water is

10,000 times lower than in air (Armstrong, 1978)the input of oxygen via groundwater, flooding sur-face water and diffusion from the atmosphere can-not fulfil the respiratoritive demands from plantroots and microbial metabolism. Within a fewhours of flooding an air-dry soil, the bulk of thesoil is rendered practically devoid of molecularoxygen (Ponnamperuma, 1984). Consequently,aerobes are replaced by facultative anaerobes,which in turn are superseded by strict anaerobes.Soil microorganisms oxidize organic matter in or-der to get energy and building materials formaintaining cell metabolism and building up newbiomass. Under anaerobiosis microbial metabolismuses other substances and compounds as terminalelectron acceptors as alternatives to oxygen inthermodynamic sequence (Table 5): NO3

-, man-ganic compounds, ferric hydroxides, SO4

2-, andCO2. Compared to aerobic respiration the energyoutput from anaerobic respiration is lower. De-composition of organic matter in an intermittentlysaturated soil is thus slow and organic matter ac-cumulates. The redox potential of a soil is a quali-tative measure for whether compounds will bereduced or oxidized. Depending on temperatureand pH reduction and oxidation of compoundswill take place in a specific redox interval. Ele-ments which are not reduced or oxidized them-selves can be influenced by changed redox condi-tions if they are constituents of a compound con-taining elements which are affected. An example isthe release of phosphate from sparsely solubleFe(III)-compounds when Fe(III) is reduced to themore soluble Fe(II). Some of the reductants result-ing from the anaerobic respiration have a provenphytotoxic effect: the reduced form of manganese,ferrous iron, sulphides and organic products ofvarious types (Armstrong, 1978).

pH in anaerobic soils generally lies between 6 and7. However, humic acids produced by microbialmetabolism may lower pH (Ponnamperuma, 1972).

Table 5. Summary of different types of respiration and corresponding redox potential (after Hoffmann, 1998): 1. Oxygenrespiration, 2. Nitrate respiration (heterotrophic denitrification), 3. Manganese respiration, 4. Iron respiration, 5. Sulphaterespiration, 6. Nitrate respiration with pyrite (autotrophic denitrification), 7. Methanogenesis.

Redox potential(mV)

OH6 + S3 + CO6 SO3 + OHC 5

OH66 + 24Fe + CO6 H48 + )24Fe(OH + OHC 4

OH18 + 12Mn + CO6 H24 + MnO12 + OHC 3

OH42 + N12 + CO30 H24 + NO24 + OHC5 2

OH6 + CO6 O6 + OHC 1

2-2

2-2

46126

22

2+

36126

22

2+

26126

222+-

36126

2226126

→

→

→

→

→

+

+

700 – 300

300 – 100

200 – 100

100 - -100

-100 - -200

H2 + SO4 + 2FeOOH + N2 OH2 + NO6 + FeS2 6 +-2422

-32 →

OH2 + CH CO + H4 7 2422 → -200 - -300

22

Likewise will acid soils low in organic matter or inactive iron attain lower pH-values (less than 6.5).pH is a complex parameter since it interacts withchemical processes due to the involvment of hy-drogen ions in chemical equilibria (e.g. Table 5)and determines the solubility of many substances.The effect of pH on hydroxide equilibria can bedescribed by (Ponnamperuma, 1984)

(17) pH + ½ log M2+ = K

where M is Fe, AlOH, Zn, or Cu; and K = 5.4 forFe, 2.2 for AlOH, 3.0 for Zn, and 1.6 for Cu. From(17) it follows that a decrease in pH from 7.0 to 6.5in ferruginous soil may increase the concentrationof water-soluble iron from 35 to 350 mg l-1 andcause iron toxicity.

Nitrogen

Nitrogen occurs in soils and sediments chiefly ascomplex organic substances, ammmonia, molecu-lar nitrogen, nitrite, and nitrate. The transforma-tions that they undergo are largely microbiologicalinterconversions regulated by the physical andchemical environment.

The mineralization of organic nitrogen in sub-merged soils stops at the ammonia stage becauseof the lack of oxygen to carry the process via nitriteto nitrate leading to an accumulation of ammoniain anaerobic soils (Ponnamperuma, 1972). Al-though aerobic decomposition of organic mattermay be more rapid than the anaerobic process,inorganic nitrogen is released in larger quantitiesand faster in anaerobic soils due to less immobili-zation in anaerobic soils (Broadbent and Reyes,1971). Nitrogen immobilization in anaerobic soilsis less because the microbiological activity is lessintense and thus the requirement for nitrogenlower than in aerobic soils. Soils vary widely intheir capacity to produce ammonium dependingon the content of organic matter.

Nitrate can be formed by nitrification of ammo-nium in the oxidized layer or imported convec-tively by ground water or by diffusion fromflooding surface water. Nitrate moves by convec-tion or diffusion to the reduced zone, where it isdenitrified (converted to nitrous oxide, N2O, orgaseous nitrogen, N2 (Knowles, 1982)). The processis microbial respiration and takes only place underanoxic conditions (Table 5). Denitrification is nor-mally a fast process; within a few days the bulk ofnative or added nitrate disappears from a sub-merged soil (Ponnamperuma, 1972). Denitrifyingorganisms need a source of H+ ions and electrons

to reduce nitrate and a carbon source and ammo-nia for cell synthesis. In soils, organic matter is thesource of all these ingredients. Thus denitrificationis absent or slow in soils low in organic matter(Ponnamperuma, 1972). Other substances can alsobe used, however, including pyrite (Table 5) (e.g.Hoffmann, 1998). The process is termed hetero-trophic denitrification when a carbon source deliv-ers energy to the process and autotrophic denitrifi-cation when the energy is derived from e.g. pyrite.For a Danish freshwater fen wetland Paludan(1995) attribute 65% of nitrate reduction in thewetland to be autotrophic using pyrite, while het-erotrophic denitrifiers are responsible for the re-maining reduction. Apart from the abscence ofoxygen and the presence of organic matter or e.g.pyrite, the process of denitrification depends onthe presence of nitrate. Denitrification rate is oftenlimited by the nitrate concentration (Reddy andPatrick, 1984). Studies of denitrification in wetmeadows and fens in Danish river valleys showrates ranging from 57 to 2179 kg NO3

- ha-1 yr-1 cor-responding to 56 – 99% of NO3

- input to the area(review by Hoffmann (1998)).

Uptake by plants can also be an important sink fornitrogen, and subsequent grazing or hay harvestcan lead to a net export of nitrogen from the areain question. Andersen (unpublished results) meas-ured 89 and 156 kg N ha-1 in above ground biomassin two vegetation zones on a non-fertilised flood-plain wetland dominated by Deschampsia caespitosaand Glyceria maxima, respectively. Hoffmann et al.(1993) measured 70 – 100 kg N ha-1 in aboveground biomass harvested on a wet meadowdominated by species characteristic of a nutrientrich soil.

The N cycle is driven by processes that occur on orat the interface of particulate material (e.g. Piany etal., 2002). Critical for the ability of a particularwetland to retain nitrogen is thus the degree ofcontact between nitrate-containing groundwater orsurface water and the wetland sediments. Conse-quently, understanding the hydrologic controls ona wetland is crucial in understanding its behaviourregarding nitrogen retention. The importance ofhydrology can be illustrated by two Danish wet-land studies. The wetland in the study of Dahl(1995) is illustrated in Fig. 7. The wetland is a dis-charge area for groundwater of both local and re-gional origin (see also Fig. 7). Due to a largethrough-flow of nitrate-rich groundwater the de-nitrification rate is very high, 2179 kg NO3-N ha-1

yr-1. Andersen (2002) studied a floodplain wetlandon the lower reach of the river Gjern (Fig. 8). Thiswetland received only a small input of local

23

groundwater, whereas the main water exchangeswas through precipitation, evapotranspiration andflooding by the river. The floodplain sedimentshave a very high potential for nitrate reduction,however denitrification was limited by supply ofnitrate. Andersen (2002) reported a rate of 71kgNO3-N ha-1 yr-1 of which reduction of nitrate dif-fusing into the sediments during ponding ac-counted for 75%.

PhosphorusPhosphorus exists in soils as (i) dissolved in pore-water, (ii) precipitated with Fe, Al or Ca to poorlysoluble compounds, (iii) adsorped to clay, CaCO3,and Al- and Fe-oxides, and (iv) incorporated inplant and microbial tissue and in peat. Only dis-solved phosphate is bio-available.

Richardson (1985) showed that the concentrationof Al-oxides was a more important factor regard-ing P-sorption than content of organic matter, pHand concentration of Fe-oxides. Sediments with alarge mineral content thus have a higher P-adsorption capacity than organogenic sediments.Since the bulk density of organic matter-rich soilsis much smaller than of soils with a high mineralcontent, the per hectare P-sorption capacity ismuch lower in the former soils. The presence ofreducible Fe(III)-compounds is considered to beimportant in controlling the concentration ofporewater dissolved phosphate (Paludan, 1995).When Fe(III) is reduced to Fe(II) in anoxic sedi-ments, phosphate bound to reducible Fe(III)-compounds is mobilised and enters the porewater.

Due to the low content of mineral material in anorganogenic soil, storage of P in plants, microbialtissue and peat is relatively more important than ina mineral soil. Thus Richardson and Marshall(1986) found vegetation and microorganisms in anorganic fen to efficiently retain P from porewaterand store a substantial portion of the available Ppool throughout the growing season. Immobilisa-tion in plants and microorganisms is temporarywhile immobilisation in peat is the only permanentbiotic storage. Richardson and Marshall (1986) re-ported that 35% of P stored in plants in a fen peat-land was returned to the surface water the firstyear and almost all of P stored in microbial tissue.Accumulation of P via peat formation amounts to0.05 – 2.4 kg P ha-1 yr-1 (Richardson, 1985), whileplant uptake is reported to 10 – 25 kg P ha-1 yr-1 inabove ground biomass (Richardson and Marshall(1986), Hoffmann et al. (1993)). Some wetlands aredepending on P release and recirculation fromdead organic matter since the external loading issmall relative to plant requirement (e.g. Paludan,

1995). However, external supply of nitrate can en-hance mineralisation of organic matter by hetero-trophic denitrification and thus mobilise P. Palu-dan and Bilcher-Mathiesen (1996) found that theorganic matter decomposition rate more than dou-bled in an organic sediment when the availabilityof nitrate was high. Autotrophic denitrification, onthe other hand, may improve the phosphate ad-sorption capacity since autotrophic denitrifiers canutilize reduced Fe and S compounds as energysources, which eventually results in the formationof Fe(III) (Table 5, reaction 6) (Paludan, 1995).

In contrast to the N-cycle gas-flux is insignificantin the P-cycle. Consequently, influx and efflux of Pmainly take place in the liquid phase (Paludan,1995). Mass balance studies of P transport in Dan-ish freshwater wetlands (mainly irrigated withdrainage or stream water) generally show that P isretained, albeit with a variation ranging from 0%to 100% of the P loading (review by Hoffmann(1998)). One study of a natural fen wetland, how-ever, showed a large P export of 16.55 kg P ha-1 yr-1.Denitrification rate in this wetland was very high.Mobilisation of P as a result of enhanced decom-position of organic matter driven by heterotrophicdenitrification was thought to be the explanationfor the high P loss rate (Paludan and Hoffmann,1996).

Eutrophication of many shallow surface waters iscontrolled by diffuse loss of P from arable land(e.g. Grant et al.,1996). Deposition of P-rich sedi-ments on riparian areas by flooding stream watercan be a significant mechanism of P retention. An-dersen and Svendsen (1997) estimated the P reten-tion rate resulting from a planned major restora-tion of the lower river Skjern, where the riverwould be allowed to frequently flood riparian ar-eas, to be 2 – 20 kg P ha-1 yr-1 corresponding to a20% reduction in riverine P transport. Kronvang etal., (2001) measured a permanent deposition of Pduring flooding of a Danish lowland river flood-plain to be 100 kg P ha-1year-1.

4. Riparian wetland vegetation

4.1 Factors controlling plant species distribution inriparian wetlands

Plants growing in riparian wetlands are subjectedto the influence of both floods and the water tablelevel. These two hydric factors strongly influencethe floristic richness and diversity (e.g. Bridghamand Richardson, 1993, van der Valk et al., 1994,David, 1996, Oomes et al., 1996, Grevilliot et al.,1998, Silvertown et al., 1999), and often a distinct

24

zonation across very short (metres) lateral dis-tances in species distribution following the hydricgradient is seen (Mitsch and Gosselink, 1986,Ward, 1998, Hupp, 2000).

The strong control of soil water content owes tothe often total lack of molecular oxygen in the soilsof these areas and the production of phytotoxinesassociated with anaerobic respiration. This meansthat plants have to be able to tolerate phytotoxinesand at the same time fulfil their own energy re-quirement in a soil devoid of free oxygen. Despitethese potentially harmful properties of the water-logged soil, many plants are endemic to wetlandsites. Many others can withstand some degree ofsoil anoxia yet are unable to compete successfullywith wetland species. Presumably the degree towhich individual species have adapted to anoxia-related stresses controls the distinct changes invegetation composition.

As anoxic conditions are absolutely detrimental toplant roots even for a short time, internal ventila-tion has emerged as a vital property of the wetlandroot system, as has the phenomenon of rhizos-phere amelioration by radial oxygen loss throughthe root wall. Indeed, tolerance to flooding is de-termined primarily by the capacities of plants toaerate their root systems (Armstrong, 1978). Phy-totoxines are immobilised by direct oxidation bythe molecular oxygen within, or leaking from, theroot system. Anoxia tolerance is not a constantcharacteristic of a plant species but can vary dur-ing the life span of the plant and over the year(Crawford and Braendle, 1996) which means thatthe response of plants to water level changes isdependent on the conditions under which thesechanges occur. Of determining influence is thelength of time the rooting zone is waterlogged(Wierda et al., 1997), the time when the rootingzone is drained in spring (Crawford and Braendle,1996) and maximum depth to the water table in thevegetative period (Wierda et al., 1997, Kotowski etal., 1998).

Grevilliot et al. (1998) found that vegetation in fre-quently flooded and thus often waterlogged areascould be characterised by high aerial biomass andlow species density. Dominating species (e.g. Glyc-eria maxima and Phalaris arundinacea) showed char-acteristics of a competitive strategy: erect and tallstature, large leaf areas and a high growth rate.Density and height of the canopy is responsible forthe disappearance of regeneration niches whichare necessary for the maintenance of species rich-ness (Grubb, 1977). This illustrates the concept ofrealised niche vs. fundamental niche (Kotowski et

al., 1998, Blanch et al., 1999): the presence of a plantspecies will depend on its tolerance to prevailingecological factors, but also on the presence ofother, competing species. However, even thoughspecies diversity locally may be low due to stressand competition, on the scale of the floodplainspecies diversity is greater than in upslope habitats(Gregory et al., 1991). This is caused by the highdiversity of microsites and complex, high-frequency disturbance regimes found along thefloodplain.

The need for nutrients of plants is basic, howeverthe craving for various compounds varies amongspecies depending on among other things growthrate and plant size. Availability and concentrationof nutrients therefore influences species distribu-tion. In wetlands focus is often on nitrogen andphosphorus both because these as macro nutrientsare necessary in large quanta and because theymight be supplied from bordering agricultural ar-eas or by flooding stream water. Changes in theavailability of N have been found to alter speciesdistribution (Grootjans et al., 1985). Similar resultshave been found regarding P availability (Boyerand Wheeler, 1989, Walbridge, 1991). Bridghamand Richardson (1993) found that a gradient innutrient concentration effectively differentiated theplant community composition of two wetlands.Fertilisation is known to increase biomass produc-tion and decrease species diversity (e.g. Vermeerand Berendse, 1983). The disappearance or regres-sion of species following fertilisation is the resultof two linked phenomena (Grevilliot et al., 1998):(i) the increase in soil fertility which encouragescompetitive species with high growth rates andwhich are able to respond quickly to an increase innutrient supplies like Alopecurus pratensis and Poapratensis, and (ii) the earlier and more frequentmowing enabled by nitrogen application. This dis-favors tall forbs and sedge species like Filipendulaulmaria, Phalaris arundinacea and Carex acuta, whichcannot survive in heavily-cut meadows.

Grime (1979) devised a model for the relationshipbetween species richness and biomass production.According to this theory an optimum curve existswith maximum species number occuring at inter-mediate biomass production. In vegetation of lowbiomass there will be severe stresses or distur-bances operating leading to conditions in whichonly a few species can survive. Up to a certainlimit there will be a positive correlation betweenthe availability of nutrients and species number.When the availability of nutrients exceed this limit,dominance of only a few species will occur as aresult of competition for light and space. Vermeer

25

and Berendse (1983) studied nine different plantcommunities ranging from dry grassland to wet-lands and were able to demonstrate an optimumcurve. They found that species richness increasedwith increasing biomass at low production levels(< 400-500 g m-2) and decreased with increasingproductivity at higher production levels (>400-500g –2). Similar results were found by Berendse et al.(1992) who analysed results of long term meadowfertilisation experiments. Grevilliot et al., (1998)found the principle to be valid also regarding theeffect of soil moisture. Thus they observed themaximum species number at intermediate soilmoisture levels, i.e. soils regularly flooded butwhere water does not stand for a long time.

One aim of vegetation science is to develop modelsfor predicting vegetation changes due to a chang-ing environment. For computer aided ecologicalmodelling, quantitative descriptions of the plantspecies response to environmental factors areneeded (Schröder et al., 2002). Consider a fieldstudy on the effect of one or more environmentalfactors on plant species distribution. Vegetationcould be recorded as e.g. the percentage cover ofthe present plant species in a number of samples.At the sample sites would also be measured envi-ronmental factors, e.g. exchangeable phosphate.Plotting the percentage cover for individual spe-cies vs. the environmental factor would often forma diffuse cloud with data points widely scatteredbeneath an upper limit. In the classical approach aregression function would be modelled throughthe centre of the distribution to describe speciesresponse to the environmental factor (Huisman etal. 1993). However, as argued by Thomson et al.(1996), most ecological information in such a graphresides in the upper limit, because the most ex-treme response of the species to the measuredfactor could be considered least affected by un-measured or unknown factors. In ecological situa-tions, numerous factors can intervene, and it willoften be impossible to account for them all. Thusthe upper limit describes the action of the meas-ured variable as a limiting factor, and the interiorof the distribution is where other factors intervene.Cade et al. (1999) introduced the quantile regres-sion approach to ecological science. This approach,which to some extent allows unknown factors in-fluencing the response variable without hamper-ing the result of the regression analysis, have beenused in econometrics for more than two decades(Koenker, 1978). Schröder et al. (2002), in an at-tempt to quantitatively describe the response offloodplain wetland plant species to environmentalfactors, tested the quantile regression approach.They used the 95 % regression quantile and found

the approach adequate to reduce the influence ofmultiple combined factors and thus to clarify therelation to single factors. They gave a set of equa-tations quantifying the response of 18 floodplainwetland species to six environmental factors.

During the last century floodplains have oftenbeen subjected to intensive agricultural use in-cluding drainage, fertilisation and frequent mow-ing. The resulting impoverishment and changes infloristic composition are well documented, e.g.Berendse et al., (1992). However, experiments haveshown that species replacement after cessation offertilisation and re-wetting is a very slow processand pointed out the difficulty to restore speciesdiversity (review by Grevilliot et al., 1998). Ber-endse et al., (1992) concluded that low productivitylevels are essential but cannot guarantee succesfulrestoration of species-rich meadows. Absence of aseed bank is a possible explanation for the slowrecovery (Hald and Vinther, 2000). Likewise is acertain level of disturbance to create sites suitablefor germination and seedling establishments nec-essary (Berendse et al., 1992).

References

Aitchison, G. D. (1961): Relationships of moisturestress and effective stress functions in unsaturatedsoils. In: Pore pressure and suction in soils. But-terworth, London, 47-52.

Allen, R. G., Pereira, L. S., Raes, D. and Smith, M.(1998): Crop evapotranspiration. Guidelines forcomputing crop water requirements. FAO irriga-tion and drainage paper no. 56. Food and Agri-culture Organisation of United Nations, Rome.

Andersen, H. E. (2002): Hydrology and nitrogenbalance of a seasonally inundated Danish flood-plain wetland. Submitted to Hydrological Proc-esses.

Andersen, H. E., Hansen, S. and Jensen, H. E.(2002): Evapotranspiration from a riparian fenwetland. Submitted to Nordic Hydrology.

Andersen, H.E. and Svendsen, L.M. (1997): Sus-pended sediments and total phosphorus transportin a major Danish river: methods and estimation ofthe effects of a coming major restoration. AquaticConservation, vol. 7, 265-276.

Armstrong, W. (1978): Root aeration in the wetlandcondition. In: D. D. Hook and R. M. M. Crawford(eds.): Plant life in anaerobic environments, 269-296. Ann Arbor Science.

26

Bartels, R. & Kuntze, H. (1973): Torfeigenschaftenund ungesättigte hydraulische Leitfähigkeit vonMoorböden. Z. Pflanzenernaehr. Duengung Bo-denkd. 134 (2): 125-135.

Berendse, F., Oomes, M. J. M., Altena, H. J. andElberse, W. T. (1992): Experiments on the restora-tion of species-rich meadows in The Netherlands.Biological Conservation, 62, 59-65.

Bishop, A. W. (1961): The measurements of porepressure in the triaxial test. In: Pore pressure andsuction in soils. Butterworth, London, 38-46.

Bishop, A. W. and Eldin, G. (1950): Undrained tri-axial tests on saturated sands and their significancein the general theory of shear strength. Geotech-nique, 2, 13-33.

Blanch, S. j., Ganf, G. G. and Walker, K. F. (1999):Tolerance of riverine plants to flooding and expo-sure indicated by water regime. Regulated Rivers:Research and Management, 15, 43-62.

Blicher-Mathiesen, G. (1998): Nitrogen removal inriparian areas. PhD-thesis. Nat. Env. ResearchInst., Denmark.

Bloemen, G.W. (1983): Calculation of hydraulicconductivities and steady state capillary rise inpeat soils from bulk density and solid matter vol-ume. Z. Planzenernaehr. Bodenk., 146, 460-473.

Boelter, D.H. (1965): Hydraulic Conductivity ofPeats. Soil Science, vol. 100, No. 4, pp 227-231.

Boelter, D.H. (1969): Division S-6 — Soil and WaterManagement and Conservation. Physical Proper-ties of Peats as Related to Degree of Decomposi-tion. Soil Sci. Amer. Proc., vol. 33, pp.606-609.

Bouwer, H. & Jackson, R.D. (1974): Determing SoilProperties. In Schilfgaarde, J.V. (ed). Drainage forAgriculture (1974). Am. Soc. of Agronomy, pp.611-672.

Boyer, M. L. H. and Wheeler, B. D. (1989): Vegeta-tion patterns in spring-fed calcareous fens: calciteprecipitation and constraints on fertility. J. of Ecol-ogy, 77, 597-609.

Bradley, C. (1996): Transient modelling of water-table variation in a floodplain wetland, Narbor-ough Bog, Leicestershire. J. of Hydrology, 185, 87-114.

Bradley, C. and Gilvear, D. J. (2000): Saturated andunsaturated flow dynamics in a floodplain wet-land. Hydrological Processes, 14, 2945-2968.

Bradford, J. M. and Gupta, S. C. (1986): Com-pressibility. In: Klute, A. (ed.): Methods of soilanalysis, part 1. Soil Sci. Soc. Am., Inc. Publisher.

Brandyk, T.J., Dodd, V.A. & Grace, P.M. (1985): AnAssessment of Some Unsaturated Hydraulic Con-ductivity Characteristics for a Range of Irish PeatSoils. Ir. J. agric. Res. 24: 237-251.

Brandyk, T., Szunirwicz, J., Skapsi, K. and Szaty-lowicz, J., 1995. The soil moisture regime study offen peat soils in the Middle Biebrza Basin as a basisfor soil protection. Z. f. Kulturtechnik und Lan-dentwicklung, 36, 78-83.

Bridgham, S. D. and Richardson, C. J. (1993): Hy-drology and nutrient gradients in North Carolinapeatlands. Wetlands, 13, 207-218.

Broadbent, F. E. and Reyes, O. C. (1971). Uptake ofsoil and fertilizer nitrogen by rice in some Philip-pine soils. Soil Science, 112 (3), 200-205

Brown, J. M. B. & Hodgson, M. P. (1988): FiniteElement Computation of Shape Factors for Kirk-ham’s Piezometer Tube Method for HydraulicConductivity of Saturated Soils. Hydrological Pro-cesses, 2: 219-225.

Brüsch, W. and Nilsson, B. (1990): Nitratomsæt-ning og vandbevægelse i et vådområde. NPo-forskning fra Miljøstyrelsen, Nr. C15 (in Danish).

Burba, G. G., Verma, S. B. and Kim, J. (1999): Sur-face energy fluxes of Phragmites australis in a prai-rie wetland. Agricultural and Forest Meteorology,94, 31-51.

Cade, B. S., Terrell, J. W. and Schroeder, R. L.(1999): Estimating effects of limiting factors wihtregression quantiles. Ecology, 80 (1), 311-323.

Campbell, D. I. And Williamson, J. L. (1997):Evaporation from a raised peat bog. Journal ofHydrology, 193, 142-160.

Carter, V. (1986): An overview of the hydrologicconcerns related to wetlands in the United States.Can. J. Bot., vol. 64, 364-374.

Chow, T. L., Rees, H. W., Ghanem, I. and Cormier,R. (1992): Compactibility of cultivated Sphagnum

27

peat material and its influence on hydrologicalcharacteristics. Soil Science, 153 (4), 300-306.

Crawford, R. M. M. and Braendle, R. (1996): Oxy-gen deprivation stress in a changing environment.J. of Experimental Biology, 47, 145-159.

Creutzberg, D: (ed.) (1975): Field extract of SoilTaxonomy. International Soil Museum, Wagenin-gen, The Netherlands.

Dahl, M. (1995): Flow dynamics and water balancein two freshwater wetlands. PhD-thesis, Nat. Env.Research Inst., Denmark (partly in Danish).

Dai, T. S. & Sparling, J. H. (1973): Measurement ofHydraulic Conductivity of Peats. Can. J. Soil Sci.53: 21-26

Dasberg, S. & Neuman, S.P. (1977): Peat Hydrol-ogy in the Hula Bain, Israel: I. Properties of Peat.Journal of Hydrology, 32: 219-239.

David, P. G. (1996): Changes in plant communitiesrelative to hydrologic conditions in the Florida Ev-erglades. Wetlands, 16 (1), 15-23.

Detlefsen, N. K. and Plauborg, F. (2001): Reconsid-eration of methods for calculation of referenceevapotranspiration (potential evapotranspiration).Danish Institute of Agricultural Sciencses. DJFRapport Markbrug no. 61 (in Danish).

Devitt, D. A., Sala, A., Smith, S. D., Cleverly, J.,Shaulis, L. K. and Hammett, R. (1998): Bowen ratioestimates of evapotranspiration for Tamarix ramo-sissima stands on the Virgin River in southern Ne-vada. Water Res. Research, 34 (9), 2407-2414.

Doss, P. K. (1993): The nature of a dynamic wa-tertable in a system of non-tidal, freshwater coastalwetlands. J. of Hydrology, 141, 107-126.

Drexler, J. Z., Bedford, B. L. DeGaetano, A. T. andSiegel, D. I. (1999): Quantification of the waterbudget and nutrient loading in a small peatland. J.of the American Water Ressources Association, 35(4), 753-769.

Eggelsmann, R. (1971): On the hydrological influ-ence of a bog. Telma 1, 37-38. Hanover.

Freeze, R.A. & Cherry, J.A. (1979): Groundwater.Prentice-Hall, Inc.

Frevert, R.K. & Kirkham, D.: A Field Method forMeasuring the Permeability of Soil Below a WaterTable. Highway Res. Board. Proc. 28, pp 433-442.

Gafni, A. & Brooks, K.N. (1990): Hydraulic Char-acteristics of Four Peatlands in Minnesota. Can. J.Soil Sci. 70, pp 239-253.

Galvin, L.F. (1976): Physical Properties of IrishPeats. Ir. J. Agric. Res. 15, pp. 207-221.

Gilvear, D. J., Andrews, R., Tellam, J. H., Lloyd, J.W. and Lerner, D. N. (1993): Quantification of thewater balance and hydrogeological processes inthe vicinity of a small groundwater-fed wetland,East Anglia, UK. J. of Hydrology, 144, 311-334.

Grant, R., Laubel, A., Kronvang, B., Andersen, H.E., Svendsen, L. M. and Fuglsang, A. (1996): Lossof dissolved and particulate phosphorus from ar-able catchments by subsurface drainage. Wat. Res.,Vol. 30 (11), 2633-2642.

Gregory, S. V., Swanson, F. J., McKee, W. A. andCummins, K. (1991): An ecosystem perspective ofriparian zones. BioScience, 41, 540-551.

Grevilliot, F., Krebs, L. and Muller, S. (1998): Com-parative importance and interference of hydrologi-cal conditions and soil nutrient gradients in floris-tic biodiversity in flood meadows. Biodiversityand Conservation, 7, 1495-1520.

Grime, J. P. (1979). Plant strategies and vegetationprocesses. Chiseler: John Wiley and Sons.

Grootjans, A. P. (1985): Changes of groundwaterregime in wet meadows. Doctoral thesis. Rijksuni-versiteit te Groningen.

Grootjans, A. P., Schipper, P. C. and van derWindt, H. J. (1985): Influence of drainage on N-mineralisation and vegetation response in wetmeadows. I. Calthion palustris stands. Acta Eocol.Eocologia Plant, 6, 403-417.

Grubb, P. J. (1977): The maintenance of species-richness in plant communities: the importance ofthe regeneration niche. Biol. Rev., 52, 107-145.

Hald, A. B. and Vinther, E. (2000): Restoration of aspecies-rich fen-meadow after abandonment: re-sponse of 64 plant species to management. Ap-plied Vegetation Science, 3, 15-24.

28

Hemond, H. F. & Goldman, J. C. (1985): On Non-Darcian Water Flow in Peat. J. of Ecology, 73: 579-584.

Hillel, D. (1980): Fundamentals of Soil Physiscs.Academic Press, Inc. pp 413.

Hoffmann, C. C. (1998a): Nutrient retention in wetmeadows and fens. PhD-thesis. Nat. Env. ResearchInst. Denmark.

Hoffmann, C. C. (1998b): Nitrate removal in aregularly flooded riparian meadow. Verh. Internat.Verein. Limnol., 26, 1352-1358.

Hoffmann, C. C., Dahl, M., Kamp-Nielsen, L. andStryhn, H. (1993): Vand- og stofbalance i ennatureng. Miljøprojekt nr. 231. Miljøstyrelsen (InDanish).

Hoffmann, C. C., Pedersen, M. L. and Laubel, A.(2000): Headwater restoration of the river Gudenå– 2. Implications for nutrients in riparian areas.Verh. Internat. Verein. Limnol.,27, 602-609.

Hoffmann, C. C., Pedersen, M. L., Kronvang, B.and Øvig, L. (1998): Restorations of rivers Brede,Cole and Skerne: A joint Danish and British EU-LIFE Demonstration Project, IV – Implications fornitrate and iron transformation. Aquatic Conser-vation, 8 (1), 223-241.

Hvorslev, M. J. (1951): Time Lag and Soil Perme-ability in Ground-water Obsservations. Bulletinno. 36. Waterways Experiment Station. Corps ofEngineers, U.S. Army.

Huismann, J., Olff, H. and Fresco, L. F. M. (1993):A hierarchical set of models for species responseanalysis. Journal of Vegetation Science, 4, 37-46.

Hupp, C. R. (2000): Hydrology, geomorphologyand vegetation of Coastal Plain rivers in the south-eastern USA. Hydrological Processes, 14, 2991-3010.

Hyashi M., van der Kamp, G. and Rudolph, D. L.(1998): Water and solute transfer between a prairiewetland and adjacent uplands, 1. Water balance. J.of Hydrology, 207, 42-55.

Ingram, H. A. P. (1983): Hydrology. In: Ecosystemsof the world, vol. 4.

Ingram, H. A. P., Rycroft, D. W. & Williams, D. J.A. (1974): Anomalous Transmission of Waterthrough certain Peats. J. of Hydrology, 22: 213-218.

Jacobsen, O. H. (1989): Unsaturated hydraulic con-ductivities for some Danish soils (in Danish). Be-retning nr. S 2030, Statens Planteavlsforsøg.pp. 60.

Jennings, J. E. B. (1961): A revised effective stresslaw for use in the prediction of the behaviour ofunsaturated soils. In: Pore pressure and suction insoils. Butterworth, London, 26-30.

Jury, W.A., Gardner, W.R. & Gardner, W.H. (1991):Soil Physics. Fifth Edition. John Wiley & Sons, Inc.pp 328.

Kirkham, D. (1945): Proposed Method for FieldMeasurement of Permeability of Soil Below theWater Table. Soil Sci., Amer. Proc., Vol. 10, pp. 58-68.

Knowles, R. (1982): Denitrification. MicrobiologicalReviews, 46 (1), 43-70.

Koenker, R. (1978): Regression quantiles.Econometrica 46, 33-50.

Koerselman, W. and Beltman, B. (1988): Evapo-transpiration from fens in relation to Penman’spotential free water evaporation (Eo) and panevaporation. Aquatic Botany, 31, 307-320.

Kotowski, W., van Digglen, R. and Kleinke, J.(1998): Behaviour of wetland plant species along amoisture gradient in two geographically distantareas. Acta Bot. Neerl., 47 (3), 337-349.

Kronvang, B., Falkum, Ø., Svendsen, L. M. andLaubel, A. (2001): Deposition of sediment andphosphorus during overbank flooding. Verh. In-ternat. Verein. Limnol. In press.

Lafleur, P. M. (1990): Evapotranspiration fromsedge-dominated wetland surfaces. Aquatic Bot-any, 37, 341-353

Lent, R. M., Weiskel, P. K., Lyford, F. P. and Arm-strong, D. S. (1997): Hydrological indices for non-tidal wetlands. Wetlands, 17 (1), 19-30.

Linacre, E. T. (1976): Swamps. In: J.L.Monteith(editor), Vegetation and the Atmosphere, vol. 2,Case Studies. Academic Press, New York, NY, 329-347

Loxham, M. and Burghardt, W. (1986): Saturatedand unsaturated permeabilities of North Germanpeats. Peat Water 37-59.

29

Luthin, J.N. & Kirkham, D. (1949): A PiezometerMethod for Measuring Permeability of Soil in SituBelow a Water Table. Soil Science, Vol. 68, pp. 349-358.

Mathur, S.P. & Lévesque, M. (1985): Negative Ef-fect of Depth on Saturated Hydraulic Conductivityof Histosols. Soil Science, vol. 140, No. 6, pp. 462-466.

McKillop, R., Kouwen, N. and Soulis, E. D. (1999):Modelling the rainfall-runoff response of a head-water wetland. Water Resources Research, 35 (4),1165-1177.

Mitsch, W. J. and Gosselink, J. G. (1986): Wetlands.Van Nostrand Reinhold Company Inc.

Naiman, R. J. and Décamps, H. (1997): The ecologyof interfaces: riparian zones. Annual Review ofEcology and Systematics, 28, 621–658.

Neuman, S. P. and Witherspoon, P. A. (1972): Fielddetermination of the hydraulic properties of leakymultiple aquifer systems. Water Resour. Res., 8,1284-1298

Nilsson, B., Refsgård, J. C., Kronvang, B., Christen-sen, S., Langhoff, J. H., Dahl, M., Andersen, H. E.,Hoffmann, C. C., Rasmussen, K. R. and Møller, I.(2002): Hydrokemisk interaktion mellem grund-vand og overfladevand (HYGRO). En metode tilklassificering af ådale i typeområder. Arbejdsrap-port fra Miljøstyrelsen (in Danish). In press.

Olszta W. (1977): The Influence of Drainage Depthon Hydraulic Conductivity of Peat Soils. ZeszytyProblemowe Postepow Nauk Rolniczych.

Oomes, M. J. M., Olff, H. and Altena, H. J. (1996):Effects of vegetation management and raising thewater table on nutrient dynamics and vegetationchange in a wet grassland. J. of Applied Ecology,33, 576-588.

Ovesen, N. B., Iversen, H. L., Larsen, S. E., Muller-Wohlfeil, D.-I., Svendsen, L. M., Blicher, A. S. andJensen, P. M. (2000): Discharge in Danish streams.Technical Report no. 340. Nat. Env. Research Inst.,Denmark (in Danish).

Päivänen, J. (1973): Hydraulic Condutivity andWater Retention in Peat Soils. Acta ForrestaliaFennica, Vol. 129, pp. 9-28.

Paludan, C. (1995): Phosphorus dynamics in wet-land sediments. PhD-thesis. Nat. Env. ResearchInst., Denmark.

Paludan, C. and Blicher-Mathiesen, G. (1996):Losses of inorganic carbon and nitrous oxide froma temperate freshwater wetland in relation to ni-trate loading. Biogeochemistry, 35, 305-326.

Paludan, C. and Hoffmann, C. C. (1995): Fate ofphosphorus in a Danish minerotrophic wetland.In: Sediment and phosphorus, erosion and deliv-ery, transport and fate of sediments and sediment-associated nutrients in watersheds (Kronvang, B.,Svendsen, L. M. and Sibbesen, E., eds.). Proceed-ings from an international workshop held in Silke-borg, Denmark, October 9-12, 1995. Nat. Env. Re-search Inst., Denmark, Technical report no. 178.

Penman, H. L. (1956): Evaporation: an introductorysurvey. Neth. J. Agric. Sci. 4, 8-29.

Phersson, M. and Pettersson, O., 1997. Energy andwater balances of a bog in central Sweden. NordicHydrology, 28 (4/5), 263-272.

Pinay, G., Clément, J. C. and Naiman, R. J. (2002):Basic principles and ecological consequences ofchanging water regimes on nitrogen cycling in flu-vial systems. Environmental Management, 30 (4),481-491.

Ponnamperuma, F. N. (1972): The chemistry ofsubmerged soils. Advances in Agronomy, 24, 29-96.

Ponnamperuma, F. N. (1984): Effects of floodingon soils. In: T. T. Kozlowski (ed.): Flooding andplant growth. 10-45. Academic Press Inc.

Priban, K. and Ondok, J. P. (1986): Evapotranspi-ration of a willow carr in summer. Aquatic Botany,25, 203-216.

Price, J. S. (1994): Evapotranspiration from a lake-shore Typha marsh on Lake Ontario. Aquatic Bot-any, 48, 261-272.

Price, J. S. and Schlotzhauer, S. M. (1999): Impor-tance of shrinkage and compression in determin-ing water storage changes in peat: the case of amined peatland. Hydrological Processes, 13, 2591-2601.

Raisin, G. Bartley, J. and Croome, R. (1999):Groundwater influence on the water balance andnutrient budget of a small natural wetland in

30

Northeastern Victoria, Australia. Ecological Engi-neering, 12, 133-147.

Reddy, K. R. and Patrick, W. H. (1984): Nitrogentransformations and loss in flooded soils andsediments. CRC Crit. Rev. Envir. Control, 13 (4),273-309.

Reddy, K. R., Rao, P. S. and Patrick, W. H. (1980):Factors influencing oxygen consumption rates inflooded soils. Soil Sci. Soc. Am. J., 44, 741-744.

Restrepo, J. I., Montoya, A. M. and Obeysekera J.(1998): A wetland simulation module for theMODFLOW ground water model. Ground Water,36 (5), 764-770.

Reynolds, W.D., Brown. D.A., Mathur. S.P. &Overend, R.P. (1992): Effect of in-situ Gas Accu-mulation on the Hydralic Conductivity of Peat.Soil Science, Vol. 153, No. 5, pp. 397-408.

Richardson, C. J. (1985): Mechanisms controllingphosphorus retention capacity in freshwater wet-lands. Science, 228, 1424-1427.

Richardson, C. J. and Marshall, P. E. (1986): Proc-esses controlling movement, storage and export ofphosphorus in a fen peatland. Ecological Mono-graphs, 56 (4), 279-302.

Roberts, J. (2000): The influence of physical andphysiological characteristics of vegetation on theirhydrological response. Hydrological Processes, 14(16-17), 2885-2903.

Rosbjerg, D. (1987): Strømning i mættet og umæt-tet jord. Noter til Hydrologi, Danmarks TekniskeHøjskole (in Danish).

Roulet, N. (1990): Hydrology of a headwater basinwetland: groundwater discharge and wetlandmaintenance. Hydrological Processes, 4, 387-400.

Rycroft, D.W., Williams, D.J.A. & Ingram, H.A.P.(1975b): The Transmission of Water through Peat.II. Field Experiments. Importance of the HydraulicConductivity in Studies of Peat Lands. J. Ecol. 63a:535-556.

Rycroft, D.W., Williams, D.J.A. & Ingram, H.A.P.(1975a): The Transmission of Water through Peat.I. Review. Importance of the Hydraulic Conduc-tivity in Studies of Peat Lands. J. Ecol. 63a: 535-556.

Schlotzhauer, S. M. and Price, J. S. (1999): Soil wa-ter flow dynamics in a managed cutover peat field,

Quebec: Field and laboratory investigations. WaterResources Research, vol. 35 (12), 3675-3683.

Schröder, H., Andersen, H. E. and Kiehl, K. (2002):Rejecting the mean – estimating the response ofwetland plant species to environmental factors bynon-linear quantile regression. Manuscript pre-pared for Journal of Vegetation Science.

da Silva, F. F., Wallach, R. and Chen, Y. (1993):Hydraulic properties of Sphagnum peat moss andtuff (scoria) and their potential effects on wateravailability. Plant and Soil, 154, 119-126.

Silvertown, J., Dodd, M. E., Gowing, D. J. G. andMountford, J. O. (1999): Hydrologically definedniches reveal a basis for species richness in plantcommunities. Nature, 400, 61-63.

Skempton, A. W. (1961): Effective stress in soils,concrete and rocks. In: Pore pressure and suctionin soils. Butterworth, London, 4-16.

Souch, C., Grimmond, C. S. B. and Wolfe, C.(1998): Evapotranspiration rates from wetlandswith different disturbance histories: Indiana DunesNational Lakeshore. Wetlands, 18 (2), 216-229.

Stewart, L. K., Hudak, P. F. and Doyle, R. D.(1998): Modelling hydrologic alterations to a de-veloping wetland in an abandoned borrow pit. J. ofEnvironmental Management, 53, 231-239.

Tabacchi, E., Lambs, L., Guilloy, H., Planty-Tabacchi, A.-M., Muller, E. and Decamps, H.(2000): Impacts of riparian vegetation on hydro-logical processes. Hydrological Processes, 14, 2959-2976.

Takagi, K., Tsuboya, T., Takahashi, H. and Inoue,T. (1999): Effect of the invasion of vascular plantson heat and water balance in the Sarobetsu Mire,northern Japan. Wetlands, 19 (1), 246-254.

Takatert, N., Sanchez-Perez, J. M. and Tremo-lieres, M. (1999): Spatial and temporal variations ofnutrient concentration in the groundwater of afloodplain: effect of hydrology, vegetation andsubstrate. Hydrological Processes, 13, 1511-1526.

Terzaghi, K. (1936): The shearing resistancse orsaturated soils and the angle between planes ofshear. Proceedings at the first international confer-ence at the Soil Mechanics and Foundations Engi-neering. Cambridge, Mass., 1936, vol.1, pp. 54-56.

31

Thomson, J. D., Weiblen, G., Thomson, B. A. Al-faro, S. and Legendre, P. (1996): Untangling multi-ple factors in spatial distributions: lilies, gophers,and rocks. Ecology, 77 (6), 1698-1715.

Topp, G.C. & Sattlecker, S. (1982): A Rapid Meas-urement of Horizontal and Vertical Components ofSaturated Hydraulic Conductivity. Can. Agric.Eng. 25, pp 193-197.

Towner, G.D. (1981): The Correction of in situ Ten-siometer Readings for Overburden Pressures inSwelling Soils. Journal of Soil Science. 32, pp. 499-504.

van der Molen, W.H. (1972): Subsidence of PeatSoils after Drainage. Hydrology of Marsh-RiddenAreas. Proceedings of the Minsk Symposium, June1972. pp 183-186

van der Valk, A. G., Squires, L. and Welling, C. H.(1994): Assessing the impacts of an increase inwater level on wetland vegetation. Ecological Ap-plications, 4 (3), 525-534.

Vedby, S. (1984): Vådbundsarealer i Susåområdet.Hydrologi og jordbund. Rapport Suså-T1, Susåun-dersøgelsen. pp 332 (in Danish).

Vermeer, J. G. and Berendse, F. (1983): The rela-tionship between nutrient availability, shoot bio-mass and species richness in grasslands and wet-land communities. Vegetatio, 53, 121-126.

Waine, J., Brown, J. M. B. & Ingram H. A. P. (1985):Non-Darcian transmission of Water in certainHumified Peats. J. of Hydrology, 82: 327-339.

Walbridge, M. R. (1991): Phosphorus availability inacid organic soils of the lower North Carolinacoastal plain. Ecology, 72 (6), 2083-2100.

Ward, J. V. (1989): The four-dimensional nature oflotic ecosystems. Journal of the North AmericanBenthological Society, 8, 2-8.

Ward, J. V. (1998): Riverine landscapes: biodiver-sity patterns, disturbance regimes, and aquaticconservation. Biological Conservation, 83 (3), 269-278.

Weiss, R., Alm, J., Laiho, R. and Laine, J. (1998):Modelling moisture retention in peat soils. Soil Sci.Soc. Am. J., 62, 305-313.

Whiting, P. J. and Pomeranets, M. (1997): A nu-merical study of bank storage and its contributionto streamflow. J. of Hydrology, 202, 121-136.

Wierda, A., Fresco, L. F. M., Grootjans, A. P. andvan Diggelen, R. (1997): Numerical asessment ofplant species as indicators of groundwater regime.J. of Vegetation Science, 8, 707-716.

Winter, T. C. (1999): Relation of streams, lakes, andwetlands to groundwater flow systems. Hydro-geology Journal, 7, 28-45.

Winter, T. C. and Rosenberry, D. O. (1995): Theinteraction of ground water with prairie potholewetlands in the Cottonwood Lake area, East-central North Dakota, 1979-1990. Wetlands, 15 (3),193-211.

Youngs, E.G. (1968): Shape Factors for Kirkham’sPiezometer Method for Determining the HydraulicConductivity of Soil in situ for Soils Overlying anImpermeable Floor or Infinitely Permeable Stra-tum. Soil Scince, Vol. 106, No. 3, pp. 235-237.

Zeeb, P. J. and Hemond H. F. (1998). ClimaticChange, 40, 211-227.

32

Paper 1, accepted by Hydrological Processes

Hydrology and nitrogen balance of a seasonally inundated Danish flood-

plain wetland

Hans Estrup Andersen1*

1National Environmental Research Institute, Vejlsøvej 25, P.O. Box 314, DK-8600 Silkeborg, Denmark

*Corresponding author; E-mail: hea@dmu.dk

Abstract

The present paper characterises a seasonally inundated Danish floodplain wetland in a state close to natu-ralness and includes an analysis of the major controls on the wetland water and nitrogen balances. Emphasisis put on unsaturated and saturated hydraulic characteristics of floodplain sediments. The main inputs of water are precipitation and percolation during ponding and unsaturated conditions.Lateral saturated subsurface flow is low. The studied floodplain owes its wetland status to the hydraulicproperties of its sediments: the low hydraulic conductivity of a silt-clay deposition layer on top of the flood-plain maintains ponding water during winter, and parts of autumn and spring. A capillary fringe extends tothe soil surface, and capillary rise from groundwater during summer maintains near-saturated conditions inthe root zone, and allows a permanently very high evapotranspiration rate; average for the growing seasonof 3.6 mm day-1 and a peak rate of 5.6 mm day-1. In summer, the evapotranspiration is to a large degree sup-plied by subsurface storage in a confined peat layer underlying the silt-clay. The floodplain sediments are in a very reduced state indicated by low sulphate concentrations. All nitratetransported into the wetland is thus denitrified. However, due to modest water exchange with surroundinggroundwater and surface water, denitrification is low; 71 kg NO3-N ha-1 during the study period of 1999. Re-duction of nitrate diffusing into the sediments during water ponding accounts for 75% of nitrate removal.Biomass production and nitrogen uptake in above-ground vegetation is high – 8.56 t dry matter ha-1 yr-1 and103 kg N ha-1 yr-1. Subsurface ammonium concentrations are high, and convective upward transport into theroot zone driven by evapotranspiration amounted to 12.8 kg N ha-1yr-1. The flodplain wetland sediments ha-ve a high nitrogen content, and conditions are very favourable for mineralisation. Mineralisation thus con-stitutes 72% of above ground plant uptake.

33

Paper 2, submitted to Nordic Hydrology

Evapotranspiration from a riparian fen wetland

Hans Estrup Andersen1*, Søren Hansen2 and Henry E. Jensen2

1National Environmental Research Institute, Vejlsøvej 25, P.O. Box 314, DK-8600 Silkeborg, Denmark2Royal Veterinary and Agricultural University, Bülowsvej 13, 1870 Frederiksberg C, Denmark

*Corresponding author; E-mail: hea@dmu.dk

Abstract

Evapotranspiration rates were measured in a riparian fen wetland dominated by vascular vegetation andsurrounded by open agricultural areas and forests. The wetland is situated on a floodplain in central Den-mark. Measurements were taken throughout the growing season (April – September) of 1999. Evapotranspiration rates were higher than those published for most other wetlands, with an average of 3.6mm day-1 during the growing season and a peak rate of 5.6 mm day-1. Daily average evapotranspiration was110% of Penman’s potential open water evaporation, and considerably higher than results published forother wetland types. Evapotranspiration was the dominant sink in the energy balance of the wetland studied. During the day,evapotranspiration accounted for 82% of the available radiant energy, Rn. Due to the presence of depositedfine-grained sediments, soil-water availability was kept high at all times which resulted in moderate canopyresistances, rc (overall mean = 32 s m-1). Evapotranspiration was controlled by a combination of driving for-ces: Rn , saturation vapour pressure deficit, D, and rc . The results presented in this study are conditioned by the proximity of the wetland to drier upland areas.During periods with high evaporative demand and low precipitation, warm, dry air is formed over theseareas, and wetland evapotranspiration rates are enhanced by local advection. Although the absolute mag-nitude of the results reported is only directly relevant to similar sites in Denmark, the processes and controlsdescribed are believed as being representative of riparian wetlands subjected to frequent flooding and/orwith a high groundwater table, with vascular vegetation, and being narrow corridors in open agriculturallandscapes.

34

Paper 3, submitted to Journal of Vegetation Science

Rejecting the mean - estimating the response of wetland plant species toenvironmental factors by non-linear quantile regression

Henning Schröder1*, Hans Estrup Andersen2 and Kathrin Kiehl3

1Institute of Geography, Friedrich-Alexander-University Erlangen-Nürnberg, Kochstr. 4/4, D-91054 Erlangen, Germany;2National Environmental Research Institute, Vejlsøvej 25, P.O. Box 314, DK-8600 Silkeborg, Denmark;3Vegetation Ecology, Technische Universität München, Am Hochanger 6, D-85350 Freising-Weihenstephan, Germany;

*Corresponding author; Fax +49 9131 8522013, E-mail: henning.schroeder@geographie.uni-erlangen.de

Abstract

Data-sets in vegetation field studies are characterised by a large number of zeros and they are incomplete inrespect to the factors which possibly influence plant species distribution. Thus it is problematic to relateplant species abundance to single environmental factors by the ordinary least square regression technique ofthe conditional mean. The non-parametric quantile regression is a promising approach for this kind of re-gression problem. In this article we employ non-linear regression quantiles in the analysis of plant speciescover in relation to environmental factors. 18 wetland species and six factors were selected (flooding durati-on, groundwater amplitude, soil organic matter, S-value, soil content of exchangeable phosphate and potas-sium). 95 % quantiles were used in order to reduce the impact of multiple unmeasured factors in the regres-sion analysis. Our results show that the standard regression of the conditional mean underestimates the ra-tes of change of species abundance due to the factor in focus in comparison to upper regression quantiles.The parameters of the response functions are given for each species and factor. The fitted response curvesindicated a general broad tolerance of the studied species to different flooding durations but a narrowerrange concerning the groundwater amplitude. Soil exchangeable potassium had only a minor influence onplant species cover whereas there was a distinct relation between species cover and the soil content ofexchangeable phosphate and the base-richness (S-value).

35

Paper 4Aquatic Conservation: Marine and Freshwater Ecosystems, Vol. 7, 265-276 (1997)

Suspended sediment and total phosphorus transport in a major Danish ri-ver: methods and estimation of the effects of a coming major restoration

H. E. ANDERSEN and L. M. SVENDSEN

Department of Streams and Riparian Areas, National Environmental Research Institute, P.O. Box 314, Vejlsøvej 25,DK-8600 Silkeborg, Denmark

ABSTRACT1. Planned restoration of the lowermost 18 kilometres of the Skjern river system (catchment area 2,490

km2) through remeandering the river to its former course with consequent frequent flooding of riparian are-as and the creation of a shallow lake and ponds is the hitherto largest river restoration project in Europe. Animportant aspect of the project planning and design has been to measure suspended sediment (SS) and totalphosphorus (TP) transport in the project area, and to assess the inter-annual variation.

2. SS and TP concentrations were measured continuously (every fourth hour) from 1993 to 1995 in the ri-ver Skjern and its main tributary, the river Omme, using automatic sampling equipment (ISCO). In addition,discrete samples were collected monthly in the remaining five smaller tributaries. Estimated SS transport inthe Skjern river system in 1994 and 1995 determined on the basis of continuous sampling was approx. 60%greater than that determined on the basis of discrete sampling. Empirical models for SS and TP transportwere developed based on the data collected in this study and applied to a 31-year time series of daily dis-charge values. Mean annual transport amounted to 12,220 tonnes SS and 100 tonnes TP corresponding to 5tonnes SS km-2 yr-1 and 41 kg TP km-2 yr-1, respectively.

3. Assessment of the effects of the coming restoration project based on measured transport and estimatedSS and TP retention rates for different areas of the lower river system revealed that SS and TP transport inthe river will be reduced by 37% and 20%, respectively. Restoration will therefore considerably enhance thenatural self-purification capacity of the river system. In addition, restoration will reduce nitrogen and ochreloading of Ringkjøbing Fjord, thereby improving environmental conditions, and remeandering will improvehabitat quality and diversity in the river system. The study stresses the importance of considering streamsand riparian areas as an entity when evaluating the effects of restoration activities.

National Environmental Research Institute

The National Environmental Research Institute, NERI, is a research institute of the Ministry of the Environment. In Danish, NERI is called Danmarks Miljøundersøgelser (DMU). NERI’s tasks are primarily to conduct research, collect data, and give advice on problems related to the environment and nature.

Addresses: URL: http://www.dmu.dk

National Environmental Research Institute ManagementFrederiksborgvej 399 Personnel and Economy SecretariatPO Box 358 Research and Development SectionDK-4000 Roskilde Department of Policy AnalysisDenmark Department of Atmospheric Environ mentTel: +45 46 30 12 00 Department of Marine EcologyFax: +45 46 30 11 14 Department of Environmental Chemistry and Microbiology Department of Arctic Environment

National Environmental Research Institute Environmental Monitoring Co-ordination SectionVejlsøvej 25 Department of Terrestrial EcologyPO Box 314 Department of Freshwater EcologyDK-8600 Silkeborg Department of Marine EcologyDenmark Project Manager for Surface WatersTel: +45 89 20 14 00Fax: +45 89 20 14 14

National Environmental Research Institute Department of Wildlife Biology and BiodiversityGrenåvej 12-14, KaløDK-8410 RøndeDenmarkTel: +45 89 20 17 00Fax: +45 89 20 15 15

Publications:NERI publishes professional reports, technical instructions, and the annual report. A R&D projects’ catalogue is available in an electronic version on the World Wide Web.Included in the annual report is a list of the publications from the current year.

[Blank page]

National Environmental Research Institute ISBN 87-7772-796-7Ministry of the Environment

Hyd

rology, nutrient processes and vegetation in flood

plain wetland

s