organic carbon accumulation on soil mineral surfaces in paddy soils derived from tidal wetlands

Geoderma 228–229 (2014) 90–103

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma

Organic carbon accumulation on soil mineral surfaces in paddy soilsderived from tidal wetlands

Livia Wissing a,⁎, Angelika Kölbl a, Peter Schad a, Tino Bräuer b, Zhi-Hong Cao c, Ingrid Kögel-Knabner a,d

a Lehrstuhl für Bodenkunde, Department Ecology and Ecosystem Management, Center of Life and Food Sciences Weihenstephan, Technische Universität München,D-85350 Freising-Weihenstephan, Germanyb Leibniz-Labor für Altersbestimmung und Isotopenforschung, Christian-Albrechts Universität zu Kiel, Max-Eyth Str. 11, D-24118 Kiel, Germanyc Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, PR Chinad Institute for Advanced Study, Technische Universität München, Lichtenbergstrasse 2a, D-85748 Garching, Germany

⁎ Corresponding author. Tel.: +49 8161 71 3734; fax:E-mail addresses: [email protected] (L. Wissing),

[email protected] (P. Schad), [email protected]@126.com (Z.-H. Cao), [email protected]

0016-7061/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.geoderma.2013.12.012

a b s t r a c t
a r t i c l e i n f o
Article history:Received 6 December 2012Received in revised form 12 December 2013Accepted 17 December 2013Available online 24 January 2014

Keywords:PedogenesisChronosequencePaddy rice cultivationIron oxidesSpecific surface areaDecalcification

Westudied organic carbon (OC) accumulation in organo-mineral associations during soil development on calcar-eous parent material. Two chronosequences in the Zhejiang Province, PR China, were investigated; one underpaddy cultivation with a maximum soil age of 2000 years, and the other under upland crops where the oldestsoil was 700 years old. Bulk soils and soil fractions of the uppermost A horizonswere analyzed for OC concentra-tions, radiocarbon (14C) contents, total pedogenic ironoxide concentration and oxalate extractable proportions ofiron (FeOX) oxides. The specific surface area of soil minerals was measured with the Brunauer–Emmett–Teller(BET-N2) method on four conditions: untreated, after organic matter removal, after iron oxide removal andafter removal of both. Initial soil formation on calcareous marine sediments includes soil decalcification and OCaccumulation. Paddy soils are characterized by an accelerated decalcification, higher contents of OC and FeOXoxides, and a pronounced accumulation of modern OC. The mineral constitution of the soil material indicatedalready a certain degree of weathering since the earliest stages of pedogenesis and remained unchanged inpaddy and non-paddy soils. The study provides no evidence of formation of new clay-sized minerals duringsoil development, which could supply new surfaces for OC accumulation. However, the study revealed higherOC coverage on mineral surfaces in decalcified paddy soils. Therefore, we assume the specific surface areaand the specific affinity of FeOX oxides for OC storage to play an important role for OC accumulation in organo-mineral associations. In contrast, the surface area of minerals in non-paddy soils, in which decalcification andthe proportion of FeOX oxides were much lower, showed significantly lower OC coverage. Selective removal ofSOM or iron oxides clearly showed that iron oxides and SOM protect each other in organo-mineral associationsprimarily in paddy fine clay-sized fraction. Thus, we explained the higher OC coverage on mineral surfaces bycomplex association between clay minerals, iron oxides and SOM in paddy soils.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The accumulation of organic matter is an important soil-formingprocess. It is considered that the structural organization of organicmatter and mineral constitutions in organo-mineral associationsare the major mechanism of soil organic matter (SOM) stabilization(Balabane and Plante, 2004; Eusterhues et al., 2005). A positive correla-tion of the SOM content and mineral constitutions is known from theliterature (Bosatta and Agren, 1997; Dümig et al., 2012; Jenkinson andRayner, 1977; Jenkinson et al., 1987; Kaiser and Guggenberger, 2000;Kiem and Kögel-Knabner, 2002). This is explained by the large mineralsurface area of soil minerals and the possibility of SOM adsorption on

+49 8161 71 [email protected] (A. Kölbl),-kiel.de (T. Bräuer),.de (I. Kögel-Knabner).

ghts reserved.

their surface (Balabane and Plante, 2004; Tipping, 1981; Wagai andMayer, 2007) or by their ability to interconnect particles to aggregates(Eusterhues et al., 2005). Organo-mineral associations are moreresistant to biodegradation (Chenu and Plante, 2006), resulting in aslow turnover time of the clay-bound organic matter (Balesdent et al.,1987). However, the understanding of the degree and mechanisms oforganic coverage on soil mineral surfaces is still incomplete (vonLützow et al., 2006; Wagai et al., 2009). The abundance of the mineralsurface areas and the adsorption of organicmatter on thesemineral sur-faces seem to be an important process in SOM stabilization (Kaiser andGuggenberger, 2003). In this context, clay-sizedminerals, in particularlyexpandable phyllosilicates and sesquioxides are known to be providersof highest mineral surface areas for adsorption of organic matter(Bartoli et al., 2007; Cornell and Schwertmann, 2003; Ransom et al.,1998; Theng, 1979). Clay-sized minerals, such as smectites and vermic-ulites, provide a high surface area up to 800 m2 g−1 (Carter et al., 1986;Robert and Chenu, 1992) and in general stabilize more SOM than sand-

http://crossmark.crossref.org/dialog/?doi=10.1016/j.geoderma.2013.12.012&domain=pdf

http://dx.doi.org/10.1016/j.geoderma.2013.12.012

mailto:[email protected]






http://dx.doi.org/10.1016/j.geoderma.2013.12.012

http://www.sciencedirect.com/science/journal/00167061

91L. Wissing et al. / Geoderma 228–229 (2014) 90–103

sized minerals, which can be attributed to higher adsorption to mineralsurfaces (Balabane and Plante, 2004).Most of the previouslymentionedstudies were done on relatively acid forest and agricultural soils orsandy subsoils. Investigations on the importance of clay-sized mineralsand iron oxides for organic carbon (OC) accumulation are not availablefor paddy soil development starting from calcareous parent materialwith successive decalcification. At Hangzhou Bay, PR China, newagricultural land has been created by consecutive land reclamationwith protective dikes over the past 2000 years, and it has been usedfor cultivation of flooded rice or non-inundated upland crops. Theparent material is a calcareous marine sediment, which is influencedby suspended Yangtze River load. The management of the paddy soils(e.g., flooding and drainage) produces special soil properties. Floodingand discharge lead to a gradual carbonate loss and subsequent pHdecrease during rice cultivation (Wissing et al., 2011). Thus, the pedo-genesis of paddy soils differs remarkably from that of correspondingnon-inundated croplands (Li et al., 2005) where decalcification ismuch slower. Faster decalcification caused by the paddy managementaccelerates other pedogenic processes. Paddy soil developmentfavors OC accumulation due to high inputs of OC (Gong and Xu, 1990;Tanji et al., 2003) and/or retards the OC decomposition because of theperiodic anoxic conditions that lead to enhanced SOM accumulation(Lal, 2002; Neue et al., 1997; Sahrawat, 2004; Wissing et al., 2011;Wu, 2011; Zhang and He, 2004). This is also confirmed by a higherradiocarbon (14C)-documented replacement of “old” carbon by “modern”carbon over time than in non-inundated non-paddy soils due to plowpandevelopment (Bräuer et al., 2012). The accumulation of SOM by associa-tion with oxalate-extractable iron (FeOX) has already been identified asa relevant feature in paddy soils (Pan et al., 2003a, 2003b). Wissinget al. (2013) found higher FeOX contents in paddy soils than in non-inundated upland soils. The authors pointed out that the higher propor-tion of FeOX seems to be responsible for a large proportion of mineral-associated SOM in paddy soils. Thus, the ability to stabilize OC is morepronounced in paddy soils compared to corresponding upland soils(Wissing et al., 2013).

In the present study, a chronosequence approach was applied toinvestigate soil development from calcareous parent material in orderto assess the organic coverage on mineral surfaces. Our major research

Table 1Basic soil parameters of mudflat (0 years), 30-year-old marshland, and paddy (P) and non-paconcentrations of inorganic carbon (IC) and organic carbon (OC), dithionite- and oxalate-extrgive the arithmetic mean (n = 3) with standard errors. The effect of the soil age was testedare indicated by italic letters. The clay content from soil texture analysiswas determined in a sinof all age classes of paddy and non-paddy soils over 700 and 2000 years pedogenesis. * repres

Site Depth(cm)

Horizona

(FAO)pHb

(KCl)ICc

(mg g1)OCc

(mg

Mudflat (0 years)f 0–30 – 8.2 ± – 5.0 ± – 5.1 ±Marsh (30 years)f 0–13 – 7.8 ± – 4.0 ± – 10.9P 50 0–7 Alp 7.4 ± 0.0 a 1.4 ± 0.3 a 17.8P 100 0–9 Alp1 5.0 ± 0.2 bc b.d.l ± – b 17.6P 300 0–18 Alp 5.8 ± 0.3 b b.d.l ± – b 22.6P 700 0–10 Alp1 6.7 ± 0.1 a b.d.l ± – b 22.3P 1000 0–10 Alp 5.2 ± 0.3 bc b.d.l ± – b 14.0P 2000 0–15 Alp 5.1 ± 0.1 c b.d.l ± – b 30.0Mean (50–700 years) 6.2 20.1*Mean (50–2000 years) 5.9 20.7*NP 50 0–9 Ap 7.3 ± 0.1 a 1.6 ± 0.1 a 10.6NP 100 0–14 Ap1 7.3 ± 0.1 a 0.7 ± 0.1 ab 10.8NP 300 0–11 Ap 7.0 ± 0.2 a 0.1 ± 0.2 b 10.5NP 700 0–11 Ap1 5.9 ± 0.3 b b.d.l ± – 11.0Mean (50–700 yrs) 6.9 10.7

a Guidelines for soil profile description (FAO, 2006).b pH values, determined by the laboratory of the Institute for Agricultural and Nutritional

Wissing et al. (2013).c Carbon data, FeDCB and FeOX were published in Wissing et al. (2011, 2013).d BS of the paddy and non-paddy soils was already published in Kölbl et al. (2014–in this issue)e Clay content from soil texture analysis was published inWissing et al. (2013). Statistics of pH

detection limit. The effect of the soil management on the bulk soil parameter IC was not assignabf pH values, carbon data and base saturation (BS) of mudflat and marshland from Kalbitz et al.

interest was to elucidate the evolution of organo-mineral associationsduring initial soil formation and if the OC coverage on mineral surfacearea is different in paddy agricultural systems with periodical sub-mergence and drainage compared to soils under dryland cropping.A paddy and a non-paddy soil chronosequence in Zhejiang Province,PR Chinawere used, to compare soil development andOCaccumulation,starting from calcareous parentmaterialwith successive decalcification.Therefore, we addressed the following three research questions:

(i) Does decalcification control organic matter accumulation andaccelerate the formation of clay minerals and iron oxides inpaddy soils?

(ii) Is the OC coverage on soil minerals (claymineral and iron oxide)higher in paddy soils, and does theOC coverage increasewith soildevelopment?

(iii) Which soil component accumulates more OC per gram fraction-clay minerals or iron (hydr)oxides?

For this approach, samples were taken from differently aged paddy(50–2000 years) and non-paddy soils (50–700 years). A particle sizefractionation of the Ap horizon was applied to the soils in order toisolate the clay-sized fractions, and their iron oxides were extracted byusing the dithionite–citrate–bicarbonate (DCB) method. The specificsurface area (SSA) of the b20 μm fraction was measured by theBrunauer–Emmett–Teller (BET-N2) method (Brunauer et al., 1938). Toinvestigate the accessibility of those mineral surfaces for OC coverageduring pedogenesis, we used selective removal of organic matterand iron oxides by combining hydrogen peroxide (H2O2) and DCBtreatments.

2. Materials and methods

2.1. Study area and soil description

The study area was located in the eastern part of the PR China, nearthe city of Cixi (30°10′N, 121°14′E), Zhejiang Province. The investigationregion is affected by river runoff and tide and the parentmaterial consistsof estuarine sediment, which originated from the Yangtze (Changjiang)River. With a sediment load of ca. 480 million tons per year (Milliman

ddy (NP) bulk soils (uppermost A horizons): Depths, horizon denominations, pH values,actable iron oxides (FeDCB and FeOX, respectively) and base saturation (BS). All numbersand significant differences (p ≤ 0.05) within the paddy and non-paddy chronosequenceglemeasurement. The effect of the soilmanagementwas tested by comparingmean valuesents significant differences (p ≤ 0.05) between paddy and non-paddy soils.

g−1)FeDCBc

(mg g−1)FeOXc

(mg g−1)BSd

(% of CECpot)Clay content b2.0 µme

(g kg−1)

– 5.3 ± – 4.1 ± – N100 –

± – 6.2 ± – 2.4 ± – N100 –

± 0.5 b 7.0 ± 0.1 a 2.6 ± 0.6 a N100 290± 1.0 bc 6.8 ± 0.4 a 3.0 ± 0.3 a 72 279± 2.0 bcb 6.6 ± 0.1 a 2.2 ± 0.3 a 71 245± 2.2 b 6.8 ± 0.2 a 3.3 ± 0.7 a 82 286± 0.8 c 7.3 ± 0.2 a 3.0 ± 0.4 a 69 277± 0.9 ab 4.4 ± 0.1 b 2.8 ± 0.2 a 58 242

6.8 2.8* 2756.5 2.8* 270*

± 0.0 a 6.3 ± 0.2 a 1.5 ± 0.3 a N100 208± 0.0 a 7.0 ± 0.3 a 1.3 ± 0.2 a N100 272± 0.0 a 6.3 ± 0.1 a 0.7 ± 0.2 a 95 237± 0.3 a 5.0 ± 0.9 a 1.5 ± 0.6 a 74 190

6.2 1.3 227

Sciences at the Martin-Luther University in Halle-Wittenberg, were already published in

. BS only meaningful for carbonate free horizons and BS saturation N100% due to carbonates.values, OC, FeDCB and FeOX were published inWissing et al. (2011, 2013). b.d.l. = below thele due to less than two IC values of paddy soils.(2013).

Table 2Properties of the uppermost A horizons of the paddy (P) and non-paddy (NP) soil fractions (b20 μm). Data of mass proportion and mass recovery (paddy n = 3; non-paddy n = 2) aregiven as the arithmetic mean with standard deviations. Data of H2O2-resistant OC were determined in a single measurement. The contributions of FeDCB and FeOX to the bulk soil weredetermined as a single measurement with standard error over time. The effect of the soil management was tested by comparing mean values of all age classes of paddy and non-paddy soils over 700 and 2000 years pedogenesis. * represents significant differences (p ≤ 0.05) between paddy and non-paddy soils. The effect of the soil agewas tested and • representsstatistically significant differences (p ≤ 0.01/0.05) among the different soil ages.

Soil fractiona

20–6.3 μm 6.3–2 μm 2–0.2 μm b0.2 μm

Site Depth (cm) Mass recovery (%) Mass (g kg−1)

P 50b 0–7 97 ± 0.6 422 ± 3.5 113 ± 2.1 185 ± 5.8 55 ± 5.5P 100b 0–9 96 ± 0.0 479 ± 5.7 112 ± 1.4 138 ± 3.5 51 ± 2.1P 300b 0–18 94 ± 1.2 402 ± 22.7 113 ± 13.6 134 ± 7.6 54 ± 0.6P 700b 0–10 97 ± 0.0 474 ± 5.9 114 ± 4.6 170 ± 2.6 57 ± 4.5P 1000b 0–10 99 ± 2.1 521 ± 31.8 116 ± 26.9 148 ± 2.1 62 ± 2.1P 2000b 0–15 98 ± 0.0 459 ± 8.7 128 ± 5.1 148 ± 6.0 46 ± 1.5Mean (50–700 years) 444 113* 157 54Mean (50–2000 years) 459 116* 154 54NP 50c 0–9 94 ± 0.8 430 ± 15.8 74 ± 3.8 114 ± 8.5 36 ± 6.7NP 100c 0–14 95 ± 0.1 446 ± 48.8 94 ± 4.2 153 ± 9.9 59 ± 3.5NP 300c 0–11 98 ± 0.1 530 ± 9.2 97 ± 0.7 129 ± 1.4 65 ± 2.1NP 700c 0–11 99 ± 0.1 536 ± 0.7 63 ± 1.4 86 ± 6.4 42 ± 0.7Mean (50–700 years) 485 82 120 50

Soil fractiona

20–6.3 μm 6.3–2 μm 2–0.2 μm b0.2 μm

Site Depth (cm) OC contribution to the bulk soil (%)

P 50b 0–7 0.8• 2.5 4.7 2.1P 100b 0–9 1.7• 2.8 4.9 2.0P 300b 0–18 2.0• 4.2 6.4 2.5P 700b 0–10 2.2• 4.6 8.2 2.5P 1000b 0–10 1.1• 1.9 4.1 1.8P 2000b 0–15 4.5• 4.7 8.2 2.5Mean (50–700 years) 1.7 ± 0.6 3.6 ± 1.0* 6.0 ± 1.6* 2.3 ± 0.2*Mean (50–2000 years) 2.1 ± 1.3 3.5 ± 1.2* 6.1 ± 1.8* 2.2 ± 0.3*NP 50 0–9 1.6 1.3 2.5 0.8NP 100 0–14 1.3 0.9 2.0 0.9NP 300 0–11 1.1 1.8 3.2 1.5NP 700 0–11 0.9 1.8 2.4 1.2Mean (50–700 years) 1.2 ± 0.3 1.4 ± 0.4 2.5 ± 0.5 1.1 ± 0.3

Site Depth (cm) FeDCB contribution to the bulk soil (mg g−1)

P 50 0–7 1.0 1.0 3.8 2.1P 100 0–9 1.4 1.3 2.6 2.0P 300 0–18 1.1 1.4 2.4 2.5P 700 0–10 1.2 1.5 3.5 2.5P 1000 0–10 1.1 1.1 3.3 1.8P 2000 0–15 0.6 0.8 1.8 2.5Mean (50–700 years) 1.2 ± 0.2* 1.3 ± 0.2 3.1 ± 0.7 2.3 ± 0.2*Mean (50–2000 years) 1.1 ± 0.3* 1.2 ± 0.3 2.9 ± 0.8 2.2 ± 0.3*NP 50 0–9 0.1 0.9 2.9 0.8NP 100 0–14 0.1 1.0 3.3 0.9NP 300 0–11 0.1 1.0 2.8 1.5NP 700 0–11 0.1 0.6 1.8 1.2Mean (50–700 years) 1.0 ± 0.0 0.8 ± 0.2 2.7 ± 0.7 1.1 ± 0.3

Site Depth (cm) FeOX contribution to the bulk soil (mg g−1)

P 50 0–7 0.3 0.5 2.0 1.3P 100 0–9 0.4 0.6 1.5 0.7P 300 0–18 0.2 0.5 1.2 0.6P 700 0–10 0.7 0.8 2.2 0.9P 1000 0–10 0.4 0.5 1.7 0.9P 2000 0–15 0.4 0.6 1.4 0.5Mean (50–700 years) 0.4 ± 0.2* 0.6 ± 0.1* 1.7 ± 0.5* 0.9 ± 0.3*Mean (50–2000 years) 0.4 ± 0.2* 0.6 ± 0.1* 1.7 ± 0.4* 0.8 ± 0.3*NP 50 0–9 0.0 0.2 0.9 0.4NP 100 0–14 0.0 0.2 0.8 0.5NP 300 0–11 0.0 0.1 0.3 0.4NP 700 0–11 0.0 0.2 0.5 0.4Mean (50–700 years) 0.0 ± 0.0 0.2 ± 0.0 0.6 ± 0.3 0.4 ± 0.1

Site Depth (cm) H2O2-resistant OC (%)

P 50 0–7 n.d. 7 7 4P 100 0–9 17 6 8 2P 300 0–18 12 6 5 3P 700 0–10 17 6 4 3

92 L. Wissing et al. / Geoderma 228–229 (2014) 90–103

Table 2 (continued)

Site Depth (cm) H2O2-resistant OC (%)

P 1000 0–10 36 9 7 4P 2000 0–15 6 3 4 2NP 50 0–9 54 11 9 4NP 100 0–14 31 11 15 5NP 300 0–11 25 8 9 3NP 700 0–11 31 8 8 4

a Medium silt (20–6.3 pm), fine silt (6.3–2 pm), coarse clay (2–0.2 pm), fine clay (b0.2 pm).b Mass distribution and OC contribution of paddy soils was already published byWissing et al. (2011).c Mass distribution of non-paddy soils was published by Wissing et al. (2013); n.d. = not detected.


and Meade, 1983; Milliman and Syvitski, 1992), the Yangtze is thedominant source of sediment delivered to the East China Sea (Wanget al., 2008). After passing the Yangtze delta, the sediments are re-deposited into the Hangzhou Bay under the influence of the TaiwanWarm Current and the Zhejiang–Fujian Coastal Current (Guo et al.,2000; Jilan and Kangshan, 1989; Xie et al., 2009). During the past2000 years several dikes had been built for land reclamation, whichresulted in a chronosequence of soils under agricultural use. Parts ofthe land were used for paddy rice cultivation under flooded conditions,followed by a winter crop (paddy soils). Other parts were used fora variety of non-inundated upland crops (non-paddy soils). Thepaddy (50–1000 years) and non-paddy (50–700 years) soil sequenceswere determined according to the records in the county annals of theZhejiang Province. The description of the 2000-year-old paddy site canbe found in Zou et al. (2011). Information in Chinese is obtainable athttp://www.cixi.gov.cn/ (Cheng et al., 2009). The paddy chronosequencecontained a succession of 50, 100, 300, 700, 1000 and 2000 years of soildevelopment. The non-paddy chronosequence consisted of four differentage stages: 50, 100, 300 and 700 years. Soils are abbreviated as P 50,P 100, etc. and NP 50, NP 100, etc., respectively. Additionally, one profilewas situated at the mudflat (estuarine sediment; tidal wetland),representing the parent material (day 0 of terrestrial soil development);another was in a nearby 30-year-old marsh land, which had not beenunder agricultural use (Kölbl et al., 2014–in this issue). It wasmentionedbyKölbl et al. (2014–in this issue) that the similarity in soil texture acrossa whole chronosequence is a strong indicator that all soils developedfrom similar parent materials. Furthermore, the authors used lipid bio-markers to show the homogeneity of the original coastal sedimentsand to enable the reconstruction of a consistent land use history forboth chronosequences.

Paddy and non-paddy soils were sampled in triplicate from adjacentindependent fields, described by the FAO Guidelines for Soil Description

Table 3The proportion of clay minerals in the b2.0 μm soil fraction of paddy (P) and non-paddy (NP)

14 Å

Soil fractiona Site Depth (cm) Chlorite (%)

2.0–0.2 μm P 50 0–7 19P 300 0–18 15P 700 0–10 19P 2000 0–15 10

2.0–0.2 μm NP 50 0–9 17NP 300 0–11 20NP 700 0–11 15

b0.2 μm P 50 0–7 13P 300 0–18 22P 700 0–10 14P 2000 0–15 12

b0.2 μm NP 50 0–9 19NP 300 0–11 21NP 700 0–11 22

a Coarse clay (2.0–0.2 μm); fine clay (b0.2 μm).

(FAO, 2006) and classified according to International Union ofSoil Sciences Working Group (2007). The following soil types wereidentified (revised from Wissing et al., 2011): Endogleyic AnthraquicCambisols (P 50, 100, 300) and Endogleyic Hydragric Anthrosols(P 700, 1000, 2000). Non-paddy soils were classified as EndogleyicHyposalic Endofluvic Cambisol (NP 50), Endogleyic Cambisol (NP 100)andHaplic Cambisols (NP 300, 700). This studydiscusses the uppermostA horizon: in non-paddy soils, the plowed horizon and in paddysoils, the puddled horizon (Alp). The meaning of the lowercase lettersis l = mottling as in capillary fringes and p = altered by plowing(FAO, 2006). If the upper part of the Alp horizon had more pronouncedredox features, the horizonwas subdivided into Alp1 and Alp2, and onlyAlp1 was analyzed in the present study. Similarly, thick Ap horizonsin non-paddy soils were subdivided into Ap1 and Ap2, with only Ap1analyzed here. Paddy soils were sampled under similar soil moistureconditions. OC input data for the chronosequences with up to 700 or2000 years of agricultural use are not available. However, samplingsites with similar cropping history during the last decades were chosenfor the respective paddy and non-paddy chronosequences. The OCcontent of the A(l)p1 and A(l)p2 horizons was similar because of theperiodical puddling and plowing. Bulk soil samples were air-dried andsieved to a size of b2 mm for further analyses.

2.2. Soil properties

The cation exchange capacity (CECpot) remained constant over timein the uppermost A horizon (see Kölbl et al., 2014–in this issue). CECpotin the paddy soils ranged between 171 and 215 mmolc kg−1 soil and innon-paddy soils between 159 and 185 mmolc kg−1 soil (data fromKölbl et al., 2014–in this issue). The AlDCB concentrations throughout allsamples range between 0.9 and 1.1 mg g−1 without any chronologicaltrend. The 14C concentration was determined in single measurement at

soils (uppermost A horizons). Data were determined in a single measurement.

10–14 Å

Mixed layer minerals (%) Illite (%) Kaolinite (%)

12 63 612 67 611 64 646 40 420 58 521 55 420 63 334 48 538 32 845 37 437 32 1923 54 329 46 424 52 2

http://www.cixi.gov.cn/)

Table 4BET-N2 specific surface area (SSA) of paddy (P) and non-paddy (NP) b20 μm soil fractions (uppermost A horizons): SSA on untreated samples, SSA after hydrogen peroxide (H2O2)treatment, SSA after dithionite–citrate–bicarbonate (DCB) extraction, H2O2 treatment and DCB extraction and the SSA covered by organic carbon (OC). SSA data were determined in asingle measurement and some samples were measured multiple with standard error. The effect of the soil management was tested by comparing mean values of all age classes ofpaddy and non-paddy soils over 700 and 2000 years pedogenesis. * represents significant differences (p ≤ 0.05) between paddy and non-paddy soils. The effect of the soil age was testedand • represents statistically significant differences (p ≤ 0.01/0.05) among the different soil ages.

SSA SSA covered by OC

Soil fractiona Site Depth untreated H2O2 DCB H2O2 + DCB

(cm) (m2 g−1) (m2 g−1) (m2 g−1) (m2 g−1) (%)

20–6.3 μm P 50 0–7 3 ± 0.2 2 ± 0.1 2• 5 ± 0.1 –

P 100 0–9 3 ± 0.0 4 2• 12 ± 1.2 21P 300 0–18 3 ± 0.1 3 2• 7 ± 0.5 11P 700 0–10 4 ± 0.0 3 2• b38 ± 0.6 –

P 1000 0–10 3 ± 0.3 3 3• 10 ± 0.0 –

P 2000 0–15 3 ± 0.4 4 3• b21 ± 2.0 33Mean (50–700 years) 3 3 2 16Mean (50–2000 years) 3 3 2 16NP 50 0–9 3 ± 0.1 5 2 5 ± 0.8• 34NP 100 0–14 4 ± 0.2 3 2 ± 0.3 5 ± 0.9• –

NP 300 0–11 4 ± 0.2 4 ± 0.3 2 7 ± 0.6• –

NP 700 0–11 4 ± 0.1 3 2 ± 0.4 23 ± 2.7•b –

Mean (50–700 years) 4 4 2 10P 50 0–7 13 ± 0.5• 15 8 ± 0.3 10 ± 0.2 13P 100 0–9 12 ± 0.3• 16 10 ± 0.0 24 ± 3.5 27P 300 0–18 12 ± 0.5• 15 14 ± 2.7 20 ± 2.2 18P 700 0–10 10 ± 0.5• 17 9 9 ± 0.5 38P 1000 0–10 10 ± 0.6• 10 9 ± 0.4 15 ± 1.3 –

P 2000 0–15 7 ± 0.2• 12 7 ± 0.2 12 ± 1.8 376.3–2 μm Mean (50–700 years) 8 16 10 16

Mean (50–2000 years) 11 14 9 15NP 50 0–9 16 ± 1.4 11 10 20 ± 3.2 –

NP 100 0–14 14 ± 1.5 12 7 15 ± 1.9 –

NP 300 0–11 13 ± 1.3 12 7 30 ± 4.2 –

NP 700 0–11 11 ± 0.8 9 10 13 ± 1.1 –

Mean (50–700 years) 13 11 8 20P 50 0–7 44 ± 1.7b 61 28 ± 0.9 40 ± 3.0 28P 100 0–9 27 ± 0.3b 55 ± 2.8 27 ± 0.9 27 ± 2.0 50P 300 0–18 34 ± 0.8 52 ± 0.5 26 48 ± 3.4 35P 700 0–10 29 ± 0.2 41 23 62 ± 1.3 28P 1000 0–10 39 ± 0.5 45 22 34 ± 0.3 13P 2000 0–15 25 ± 0.3 49 23 35 ± 4.4 50

2–0.2 μm Mean (50–700 years) 34 52 26 44Mean (50–2000 years) 33 50* 25 41NP 50 0–9 45 ± 3.3 39 23 47 ± 0.9 –

NP 100 0–14 47 ± 0.7 41 23 ± 0.5 52 ± 1.9 –

NP 300 0–11 42 ± 0.9 37 26 36 ± 0.6 –

NP 700 0–11 39 ± 0.2 48 26 53 ± 2.3 18Mean (50–700 years) 43* 41 24 47P 50 0–7 80 ± 2.9 117b 69• 123 ± 1.3 31P 100 0–9 61 ± 1.7 122b 65• 146 ± 3.0 49P 300 0–18 46 ± 2.4 105b 56• 77 ± 23.2 56P 700 0–10 37 ± 0.7 118b 60 ± 2.3• 84 ± 3.6 68P 1000 0–10 74 ± 3.5b 125b 63 ± 3.0• 319 ± 23.0 40P 2000 0–15 25 ± 0.1 100 39• 308 ± 29.2 75

b0.2 μm Mean (50–700 years) 56 116 63 108Mean (50–2000 years) 54 115 59 176NP 50 0–9 79 ± 2.3 112b 85 105 ± 1.8 30NP 100 0–14 99 ± 0.6b 121b 79 68 ± 3.7 18NP 300 0–11 88 ± 0.5 116b 81 ± 2.2 118 ± 3.6 24NP 700 0–11 72 ± 0.8 123b 69 104 41Mean (50–700 years) 85 118 78* 99

a Medium silt (20–6.3 μm); fine silt (6.3–2 μm); coarse clay (2–0.2 μm); fine clay (b2 μm).b Micropore (b2 nm) surface area was measured in this samples.


the Leibniz-Laboratory at the Christian-Albrechts University in Kiel with a3 × 106 V HVE Tandetron AMS (accelerator mass spectrometry) system(Bräuer et al., 2012), with a 1 − σ precision of about 0.25% moderncarbon (pMC) (Nadeau et al., 1997). Paddy bulk soil samples are charac-terized by a 14C content (Appendix) ranging from 99 pMC (P 700) to112 pMC (P 2000) and in non-paddy bulk soils from 98 pMC (NP 100)to 103 pMC (NP 700) (see Appendix). The mudflat and the marshlandhad the lowest 14C content and revealed therefore the highest proportionof “old” inherited C (Appendix). Total carbon (Ctot) concentrations of bulksoils were determined in duplicate by dry combustion at 950 °C on a

Vario EL elemental analyzer (Elementar Analysensysteme, Hanau,Germany). The OC concentration increased during paddy soil devel-opment from 17.8 mg g−1 (paddy 50 years) to 30 mg g−1 (paddy2000 years), whereas the OC concentration remained constant(11 mg g−1) during non-paddy soil development (Table 1). The Ctot

concentrations of particle size fractions were measured in duplicateby dry combustion (EuroEA Elemental Analyzer 3000, HEKAtech,Wegberg, Germany). Both clay-sized fractions are characterized by thehighest OC concentration out of all soil fractions b20 μm, wherein theOC concentration in paddy clay fractions (mean: 4.0%) was twice as

Table 5BET-N2 specific surface area (SSA) of paddy (P) and non-paddy (NP) across the soil fractions b20 μm (uppermost A horizons): SSA on untreated samples, SSA after hydrogen peroxide(H2O2) and their difference to the bulk soil SSA (untreated and H2O2-treated).

SSAuntreated (m2 g−1) SSAuntreated (m2 g−1)

Soil fractionsa Bulk soil

Site 20–6.3 μm 6.3–2 μm 2–0.2 μm b0.2 μm Sum (m2 g−1) Difference (m2 g−1)

P 50 1.4 1.5 8.1 6.6 18 14 −3P 100 1.4 1.3 3.8 3.1 10 12 2P 300 1.1 1.4 4.5 2.5 10 11 1P 700 1.8 1.2 5.0 2.1 10 8 −2P 1000 1.8 1.2 5.8 4.6 13 13 −1P 2000 1.3 0.9 3.6 1.2 7 6 −1NP 50 1.3 1.1 5.1 2.6 10 16 6NP 100 2.0 1.3 7.2 5.8 16 18 1NP 300 2.1 1.2 5.4 5.7 14 11 −3NP 700 1.9 0.7 3.4 3.0 9 11 2

SSAH2O2 (m2 g−1) SSAH2O2 (m

2 g−1)

Soil fractionsa Bulk soil

Site 20–6.3 μm 6.3–2 μm 2–0.2 μm b0.2 μm Sum (m2 g−1) Difference (m2 g−1)

P 50 1.0 1.7 11.2 5.0 19 23 4P 100 1.8 1.8 7.5 2.8 14 15 1P 300 1.3 1.7 7.0 2.8 13 11 −2P 700 1.4 1.9 6.9 2.3 12 11 −2P 1000 1.6 1.1 6.6 2.8 12 11 −1P 2000 1.9 1.5 7.3 2.3 13 10 −3NP 50 2.0 0.8 4.3 3.7 11 9 −2NP 100 1.4 1.1 6.2 7.1 16 15 −1NP 300 2.0 1.2 4.7 7.5 15 8 −8NP 700 1.8 0.5 4.1 5.1 12 5 −7

a Medium silt (20–6.3 μm), fine silt (6.3–2 μm), coarse clay (2–0.2 μm), fine clay (b0.2 μm).


high as those in non-paddy clay fractions (mean 2.2% and 2.3%). The OCconcentrations of bulk soil and the soil fractions were published inWissing et al. (2011, 2013). The OC contribution of the soil fractions to

P 50: 20-6.3 µmP100: 20-6.3 µmP 300: 20-6.3 µmP 700: 20-6.3 µmP 1000: 20-6.3 µmP 2000: 20-6.3 µm

NP 50: 20-6.3 µmNP 100: 20-6.3 µmNP 300: 20-6.3 µmNP 700: 20-6.3 µm

P 50: 6.3-2 µmP 100: 6.3-2 µmP 300: 6.3-2 µmP 700: 6.3-2 µmP 1000: 6.3-2 µmP 2000:6.3-2 µm

NP 50:6.3-2 µmNP 100: 6.3-2 µmNP 300: 6.3-2 µmNP 700: 6.3-2 µm

0

20

40

60

80

100

120

140

SS

A (

m2

g-1

)

FeDCB (mg g-1)0 5 10 15 20 25 30 35

a

Fig. 1.Relation betweendithionite-extractable iron oxides (FeDCB) and BET-N2 specific surface arthe uppermost A horizons (20–6.3 μm = medium silt; 6.3–2.0 μm = fine silt; 2.0–0.2 μm =

the bulk soil was determined for paddy soils (Wissing et al., 2011)and non-paddy soils by multiplying the mass proportion with theOC concentration of the individual fraction (Table 2). Other basic

P 50: 2-0.2 µmP 100: 2-0.2 µmP 300: 2-0.2 µmP 700: 2-0.2 µmP 1000: 2-0.2 µmP 2000: 2-0.2 µm


P 50: <0.2 µmP 100: <0.2 µmP 300: <0.2 µmP 700: <0.2 µmP 1000: <0.2 µmP 2000: <0.2 µm

NP 50: <0.2 µmNP 100: <0.2 µmNP 300: <0.2 µmNP 700: <0.2 µm

0

20

40

60

80

100

120

140

0 5 10 15 20 25 30 35

FeDCB (mg g-1)

SS

AH

2O2

(m2

g-1

)

b

ea (SSA) (a) untreated and b)H2O2-treated paddy (P) andnon-paddy (NP) soil fractions ofcoarse clay; b0.2 μm = fine clay).

P 50: 20-6.3 µmP100: 20-6.3 µmP 300: 20-6.3 µmP 700: 20-6.3 µmP 1000: 20-6.3 µmP 2000: 20-6.3 µm


P 50: 6.3-2 µmP 100: 6.3-2 µmP 300: 6.3-2 µmP 700: 6.3-2 µmP 1000: 6.3-2 µmP 2000:6.3-2 µm

NP 50:6.3-2 µmNP 100: 6.3-2 µmNP 300: 6.3-2 µmNP 700: 6.3-2 µm

P 50: 2-0.2 µmP 100: 2-0.2 µmP 300: 2-0.2 µmP 700: 2-0.2 µmP 1000: 2-0.2 µmP 2000: 2-0.2 µm


P 50: <0.2 µmP 100: <0.2 µmP 300: <0.2 µmP 700: <0.2 µmP 1000: <0.2 µmP 2000: <0.2 µm

NP 50: <0.2 µmNP 100: <0.2 µmNP 300: <0.2 µmNP 700: <0.2 µm

R² = 0.28

R² = 0.31

R² = 0.83

R² = 0.96

0

20

40

60

80

100

120

140

OC (mg g-1)

SS

A (

m2

g-1

)

R² = 0.14

R² = 0.23

R² = 0.07

R² = 0.39

0

20

40

60

80

100

120

140

0 1 2 3 4 5 6 0 1 2 3 4 5 6

OC (mg g-1)

SS

AH

2O2

(m2

g-1

)

a b

Fig. 2. Relation between organic carbon (OC) content and BET-N2 specific surface area (SSA) (a) untreated and b) H2O2-treated paddy (P) and non-paddy (NP) soil fractions of theuppermost A horizons (20–6.3 μm = medium silt; 6.3–2.0 μm = fine silt; 2.0–0.2 μm = coarse clay; b0.2 μm = fine clay).


information of the bulk soil samples (depths, horizon denominations,pH values, inorganic carbon (IC) concentration, base saturation andclay content) are summarized in Table 1.

2.3. Particle size fractionation

Before fractionation, all soils were pre-treated with 0.1 M HCl (finalpH value = 4.8) to remove carbonate. HCl and the dissolved Ca2+ ionswere removed by centrifugation (20 min, at 3423 g). The particle sizefractionation was done according to Wissing et al. (2011). Briefly, 15 gof air-dried bulk soil (b2 mm) was suspended with demineralizedwater (1:5 ratio soil (g) to water (ml)). To disrupt sand-sized macro-aggregates, ultrasonic treatment with an energy input of 60 J ml−1

(56.03 W × 80.03 s) was used, followed by wet sieving to separatesand and particulate organic matter (2000–63 μm) from the resi-due. The complete dispersion of microaggregates (b200 μm) wasachieved by an additional ultrasonic treatment with an energyinput of 440 J ml−1 (56.03 W × 2356 s). The b20 μm fraction wasfurther separated by sedimentation in Atterberg cylinders (silt-sizedfractions) and centrifugation (clay-sized fractions). The followingsix particle size fractions were isolated: sand and particulate organicmatter (2000–63 μm), coarse silt (63–20 μm), medium silt (20–6.3 μm), fine silt (6.3–2.0 μm), coarse clay (2.0–0.2 μm) and fine clay(b0.2 μm). Compared to physical fractionation, soil texture analysisled on average in both soils to slightly higher amounts of clay and finesilt (paddy: 7% (b2 μm); 4% (6.3–2.0 μm) and non-paddy 6% (b2 μm);3% (6.3–2.0 μm)) but to less proportions of medium silt. All fractionswere freeze-dried and weighed to obtain the mass proportion of eachfraction to the bulk soil.

A high mass recovery was obtained from physical fractionation,varying from 94% to 99% for both paddy and non-paddy soils(Table 2). The mass distribution of the soil fractions was already pub-lished in Wissing et al. (2011, 2013) and was summarized in Table 2.

We tested the effect of the soil management (see Section 2.7) on theparameter mass distribution and obtained significant differences forthe 6.3–2 μm fraction between paddy and non-paddy soils. No signifi-cant differences were found by time series statistics (see Section 2.7)for the mass distribution among the different soil ages of the b0.2 mfraction.

2.4. Soil mineralogy

Total pedogenic iron oxides (dithionite-extractable iron [FeDCB]) ofthe bulk soil and the soil fractions b20 μm were determined accordingto Mehra and Jackson (1960) by using the DCB method. The proportionof oxalate soluble iron oxides (FeOX) was analyzed using the oxalatemethod of Schwertmann (1964). The iron concentration wasmeasuredby inductively coupled plasma optical emission spectroscopy (ICP-OES)(Vista-Pro, CCD simultaneous, Varian, Darmstadt, Germany). FeDCB andFeOX oxide concentrations of the bulk soil samples and the soil fractionswere published byWissing et al. (2013) and this datawere summarizedin Tables 1 and 2. The contribution of the FeDCB and FeOX oxides wascalculated by multiplying the mass proportion with the FeDCB and FeOXoxide concentration of the individual fraction (Table 2). We testedthe effect of the soil management (see Section 2.7) on the parameterFeDCB- and FeOX-contribution to the bulk soil and obtained significantdifferences between paddy and non-paddy soils (Table 2). No significantdifferences were identified by time series statistics (see Section 2.7) forthe FeDCB- and FeOX-contribution to the bulk soil among the differentsoil ages.

Aliquots of both clay fractions (2.0–0.2 μm and b0.2 μm) were pre-treated with H2O2 (30%) to remove organic matter (see Section 2.5).The X-ray diffraction (XRD) patterns were recorded with a Co radiationsource using a Philips PW 1070 diffractometer. Random powdersamples were measured from 5 °C to 60 °C in steps of 0.02 °C with acounting time of 5 s for each step. To evaluate expandable clayminerals,

a

b

Specific surface area change …

External specific surface areawhich is measurable by BET-N2Clay mineral Organic matterFeDCB oxides

Ca2+ Ca2+

Ca2+

bridges

c

d

Ca2+

Ca2+

Ca2+

Ca2+

non-paddy soilspaddy soils

… after H2O2 - and DCB-treated

untreated

… after H2O2-treated

…after DCB-treated

Fig. 3. Schematic overview of the specific surface area change affected by the chemical treatments in comparison to untreated paddy and non-paddy soils. The different letters indicatesa) the untreated b2 μm fraction, b) b2 μm fraction after organic matter oxidation by hydrogen peroxide (H2O2), c) b2 μm fraction after removal of FeDCB oxides by dithionite–citrate–bicarbonate (DCB) treatment and d) b2 μm fraction treated by H2O2 and the subsequent removal of FeDCB oxides. The figure comparing the mean values of all age classes of paddy andnon-paddy soils over 700 years pedogenesis.


the samples were treated with glycol as well as with KCl and stepwiseheated to 100 °C, 200 °C and 550 °C. Clay minerals were identifiedaccording toMoore andReynolds (1989) and thedifferentmineral phaseswere determined by semi-evaluation. Due to the low FeDCB concentra-tions of the paddy and non-paddy soil fractions (0.1% and 3.3% (Wissinget al., 2013)), iron oxides could not be detected by XRD analyses.

2.5. Organic matter oxidation

A H2O2 treatment was applied to the b20 μm soil fractions in orderto oxidize organic matter. Aliquots of the 20–6.3 μm (2 g), 6.3–2.0 μm(500 mg), 2.0–0.2 μm (300 mg) and b0.2 μm (300 mg) soil fractionswere weighed into beakers to which 20 ml (soil fractions) of 30%H2O2 was added. To accelerate the oxidation reaction, soil sampleswere heated to 60 °C on a sand bath. Samples were placed in an ovenat 60 °C for 24 h to remove the remaining H2O2. Silt-sized fractionswere centrifuged for 20 min at 4470 g, and the clay fraction for 1 h at

4470 g. Afterwards, samples were freeze-dried and retained for XRDand BET-N2 analyses and OC measurement.

2.6. Determination of the SSA

The SSA of b20 μm soil fractions was determined by the BET-N2

method (Brunauer et al., 1938). Gas adsorption of 11 points was mea-sured in the relative pressure range of 0.05 to 0.3 with an Autosorb-1analyzer (Quantachrome, Syosset, NY, USA). To avoid adsorbed waterand volatile substances, samples were outgassed for at least 24 h invacuum under helium flow at 40 °C before measurement. Also beforemeasurement, all soil samples were weighed to calculate the massproportion of each sample. Analyses of SSA were carried out on fourdifferent conditions:

(i) on untreated soil samples (SSAuntreated),(ii) after oxidation of organic matter by H2O2 treatment (SSAH2O2)

(see Section 2.5),


(iii) after removal of total iron oxides by DCB extraction (SSADCB)(see Section 2.4), and

(iv) after H2O2 treatment followed by DCB extraction (SSAH2O2 + DCB).

The mass loss after H2O2 treatment due to organic matter extractionwas 0%–3% in the 20–6.3 μm fraction; 5%–11% in the 2.0–0.2 μmfraction; 2%–22% in the 2.0–0.2 μm fraction and 0%–9% in the b0.2 μmfraction.

The percentage proportion of the SSA (m2 g−1) covered by OC wascalculated as:

% SSA covered by OC ¼ SSAH2O2−SSAuntreated

� �=SSAH2O2

� 100: ð1Þ

The SSA proportion across the soil fractions was calculated byadding up the SSA of the individual untreated and H2O2-treated frac-tions b20 μm. Afterwards, the sum of the fractional SSA was subtractedfrom the untreated bulk soil SSA and the H2O2-treated bulk soil SSA.

2.7. Statistical analyses

Statistical analyses were done by SPSS Statistics 19 (IBM SPSSCompany). We conducted different statistical tests in order to evaluateon the one hand the effect of the soil management (mean values ofpaddy versus mean value of non-paddy of all age classes) and on theother hand the effect of the soil age (different age classes of bothchronosequences) on diverse bulk soil and soil fraction parameters.The effect of the soil age on pH values, IC, OC, FeDCB and FeOX concentra-tions were tested by analysis of variance (ANOVA) with the post hoctests Tukey-B and Bonferroni. All data were analyzed for homogeneityof variances by Levene test, and the analysis of normalitywasperformedwith the Shapiro–Wilk test. Significant differences (p ≤ 0.05) withinthe paddy and non-paddy chronosequence are indicated by italicletters.

The effect of the different management techniques (paddy versusnon-paddy) on bulk soil and soil fraction parameters (e.g., mass distri-bution, OC contributions) was tested by comparing the mean valuesof the chronosequences. In this case, homogeneity of variances andnormality of data were not given. Therefore we used a non-parametricMann–Whitney U-test instead of ANOVA. This applies also for thePearson correlation which was used to compare SSA data along thechronosequence. The mean values of paddy and non-paddy soils(mean of 50–700 years and mean of 50–2000 years of pedogenesis)were compared and “*” represents statistically significant differences(p ≤ 0.05) between both management systems. Time series statisticswere accomplished by Pearson correlation coefficients to evaluate theeffect of the soils age on the soil fraction parameters SSA (untreated,after H2O2- and DCB-treatment and H2O2 + DCB), mass distribution, con-tribution of OC, FeDCB and FeOX to the bulk soil. “*” represents statisticallysignificant differences (p ≤ 0.01/0.05) between the different soil ages.

3. Results

3.1. Soil mineralogy of the soil fractions

Results from XRD analyses (Table 3) showed that kaolinite, illite,chlorite and secondary chlorite are clay minerals in both soil sequences.Illite was the major clay mineral and varied between 40%–67% (P) and55%–63% (NP) in the 2.0–0.2 μm fraction. The b0.2 μm fraction wascharacterized by less illite, 32%–48% (P) and 46%–54% (NP) (Table 3).Thosemineral constituents of the soil material indicate already a certaindegree of weathering before they were re-deposited at the HangzhouBay. Secondary clay minerals were components of the soil materialsince the earliest stages of paddy and non-paddy soil development.

3.2. Properties of the SSA of different b20 μm soil fractions

The SSAuntreated of the soil fractions (Table 4) increasedwith decreas-ing particle size in both chronosequences. Time series statistics by Pear-son correlation revealed significant (p ≤ 0.01) decreasing untreatedSSA values in the 6.3–2.0 μm fraction during paddy soil development,ranging from 13 m2 g−1 (P 50) to 7 m2 g−1 (P 2000) in paddy soilsand from 16 m2 g−1 (NP 50) to 11 m2 g−1 (NP 700) in non-paddysoils. All other soil fractions showed varying SSAuntreated values, reachingin almost all cases the lowest values in the oldest soil (P 2000 andNP 700). For example, the SSAuntreated values ranged from 44 m2 g−1

(P 50) to 25 m2 g−1 (P 2000) in the 2.0–0.2 μm fraction and from 45m2 g−1 (NP 50) to 39 m2 g−1 (NP 700) in non-paddy soils (Table 4).The values decreased also in the b0.2 μm fraction (P: 80–25 m2 g−1

and NP: 79–72 m2 g−1) (Table 4). Compared to SSAuntreated, the organicmatter removal led to larger SSAs in paddy (11%–75%) and non-paddysoils (18%–41%) (SSA covered by OC; Table 4). In particular, the paddysoil clay fractions showed markedly larger SSAH2O2 values (Table 4),which increased during soil development in both fractions (2.0–0.2 μm: from 28% to 50% and from 31% to 75% in the N0.2 μm fraction).When compared to the SSAuntreated values, the fractions 20–6.3 μm(P + NP), 6.3–2.0 μm and 2.0–0.2 μm from non-paddy soils werecharacterized by almost no increase of themineral surface after organicmatter removal (Table 4). The removal of pedogenic iron oxides by DCBtreatment (SSADCB) resulted in the lowest SSA values in paddy and non-paddy soils (Table 4). Clay-sized fractions of paddy soils showeddecreasing SSADCB values with soil age. The SSADCB of the 2.0–0.2 μmfraction declined with soil age from 28 m2 g−1 (P 50) to 23 m2 g−1

(P 2000) and that of the b0.2 μm fraction declined significantly(p ≤ 0.05, Pearson correlation) with paddy soil age from 69 m2 g−1

(P 50) to 39 m2 g−1 (P 2000). With increasing non-paddy soil age, theb0.2 μm fraction also showed decreasing SSADCB values with soil agefrom 85 m2 g−1 (NP 50) to 69 m2 g−1 (NP 700), whereas no signifi-cance was calculated by Pearson correlation. No consistent trend wasdetected for the silt-sized and 2.0–0.2 μm fractions (Table 4). Time se-ries statistics by Pearson correlation showed significantly (p ≤ 0.05) de-creasing SSADCB values in the 20–6.3 μm fraction during paddy soildevelopment. The combined treatment of organic matter oxidationand removal of FeDCB led to the largest SSA values overall (Table 4),and the values were the highest in the b0.2 μm fractions of both soils,ranging from 77 m2 g−1 (P 300) up to 319 m2 g−1 (P 1000) in paddysoils and from 68 m2 g−1 (NP 100) to 118 m2 g−1 (NP 300) in non-paddy soils without a chronological trend. Pearson correlationshowed significant (p ≤ 0.05) increasing SSAH2O2- and SSADCB-valuesin the 20–6.3 μm fraction during non-paddy soil development. Theuntreated and H2O2-treated SSA across the fractions and their dif-ference to the bulk soil SSA (untreated and H2O2-treated) is summa-rized in Table 5. We found for the paddy soil chronosequence that theSSAH2O2 samples (bulk soil and soil fractions) were only slightlydifferent compared to the untreated bulk soil and soil fractions. TheSSAuntreated sum of the soil fractions varied between −21% (P 100)to 27% (P 2000) and −36% (NP 50) to 28% (NP 300) compared to theSSAuntreated of the bulk soils (Table 5). We calculated a difference ofthe SSAH2O2 sum of the soil fractions compared to the SSAH2O2 of thebulk soils, ranging in paddy from −19% in P 50 years to 30% in P2000 years and in non-paddy soils from 9% in NP 100 years to 148% inNP 700 years (Table 5). The H2O2 treatment released up to 98% of SOCin the N2.0 μm fraction (Table 2). The amount of H2O2-resistant OCranged from 8% (P 100) to 2% (P 100, P 2000) in paddy clay-sized frac-tion and from 15% (NP 100) to 3% (NP 300) in non-paddy soils. Theamount of H2O2-resistant OC was highest in the 20–6.3 μm fraction,ranging from 6% up to 36% in paddy soils and from 31% to 54% in non-paddy soils (Table 2).

The relation between the FeDCB content and the SSA before and afterorganicmatter removal is shown in Fig. 1a and b. The FeDCB contributionwas highest in both clay-sized fractions (Table 2). Therefore, increasing

30 3000 100 2000100070050years

30 3000 100 70050years

proportion of inheritedcarbon

decalcification

content of Feox

proportion of inheritedcarbon

decalcification

content of Feox

Paddy soil develoment

without redox cycle

with redox cycle

Non-paddy soil development

OC accummulation

OC accummulation

Fig. 4. Chronological changes of dominant soil forming processes in paddy versus non-paddy soils.


SSA values were related to the decrease of the particle size fraction(Fig. 1a), but we could not confirm after organicmatter removal the for-mation of additional mineral surfaces from/due to iron oxides duringsoil development (Fig. 1b). The relation between the OC content andthe SSA before and after organic matter removal is summarized inFig. 2a and b. The OC content of 6.3–2.0 μm fraction and both clay-sized fractions strongly increased with decreasing particle size as wellaswith increasing soil age (except for P 1000). This was strongly relatedto thedecline of the SSA (R2 increasedwith decreasing particle size from0.28 to 0.96; see Fig. 2a). After removal of OCwith H2O2, the correlationbetween SSA and the original OC content of the respective soil fractionwas much lower (Fig. 2b) compared to the untreated SSA (Fig. 2a).However, both paddy and non-paddy soil fractions can be fittedby the same regression line because the relation between the SSA andthe OC content is irrespective of the soil management (Fig. 2a, b).

4. Discussion

4.1. Paddy management accelerates soil decalcification

In the Cixi coastal region of Zhejiang Province, where land reclama-tion has been in effect for at least 2000 years, the substrate for all paddyand non-inundated non-paddy soils is predominantly marine tidal flatsediment (Kölbl et al., 2014–in this issue). The coastal sediments werediked and drained and gradually becameoxic. Portnoy (1999) describedan acceleration of the initial soil forming processes due to soil diking anddrainage, especially sulfide oxidation that causes changes in redoxcondition, lowers soil pH and alters mobility of iron (Lord, 1980;Zottoli, 1973), and it may also accelerate initial soil-forming processes(Portnoy, 1999). In addition, management of those marshlandsoils may further influence their soil properties. In this study, paddy

soils were characterized by an accelerated decalcification (Table 1)induced by specific soil management. Carbonatewas leached by period-ical flooding and drainage of the paddy soils during rice cultivation(Wissing et al., 2011). Kalbitz et al. (2013) mentioned thatthe carbon losses due to carbonate weathering were faster inpaddy soils (155–220 kg C ha−1 a−1) than in non-paddy soils(80–116 kg C ha−1 a−1) in the first 300 years. The loss rate of carbonwas even 395–545 kg C ha−1 a−1 during the first 50 years of paddysoil development (Kalbitz et al., 2013). However, howmuch the intenseand accelerated decalcification of paddy soils affects pedogenesis,including OC accumulation and mineral formation, is still unclear. Inparticular, the formation of mineral surface areas and their accessibilityfor OC coverage may provide additional information on organic matteraccumulation processes in paddy soils.

4.2. Providers of mineral surfaces

4.2.1. Organic matter associations with different mineral surfacesTo illustrate our understanding of soil mineral surface areas, their

coverage with organic matter and the effect of different chemical treat-ments, we designed a schematic overview, showing the effect of organicmatter oxidation by H2O2 (Fig. 3b), removal of iron by DCB treatment(Fig. 3c) and H2O2 treatment combined with the subsequent removalof iron (Fig. 3d) in paddy andnon-paddy clay-sized fractions. The higherorganic matter content in paddy soils results in smaller untreated SSAsin clay-sized fraction (Table 4; Fig. 2a) than in correspondingnon-paddysoils. The removal of organic matter with H2O2 leads to large OC losses(P: 92%–98% and NP: 85%–97%) for the clay-sized fractions (H2O2-resistant OC; Table 2) and to a distinct increase of the BET-N2 specificsurface area (Table 4). Such results have been reported by numerousauthors (Burford et al., 1964; Theng et al., 1999; Mayer and Xing,


2001; Kahle et al., 2002; Kiem and Kögel-Knabner, 2002; Wagai et al.,2009; Dümig et al., 2012). The higher SSA values due to the H2O2 treat-ment were particularly large for the fine mineral fractions (see alsoBurford et al., 1964; Mayer and Xing, 2001), especially for clay-sizedfractions of paddy and non-paddy soils (Table 4). Mineral surfacesthat were covered by organic matter (Fig. 3a), did not contribute signif-icantly to theN2 area (Fig. 3; compare external SSAwhich ismeasurableby BET-N2), but the surface areas were accessible to N2 after organicmatter removal (Chiou, 1990; Theng et al., 1999) (Fig. 3b). This observa-tion can be explained by the blocking of pores (Weiler and Mills, 1965;Williams et al., 1967) or the coating of clay aggregates (Burford et al.,1964) by organic matter, which restricted N2 molecules from enteringthe micropores (b2 nm) of clay domains (Theng et al., 1999). Some ni-trogenmay have been adsorbed by coarse organicmatter, but the loss ofsurface due to organic matter removal was in all instances less than theincrease due to areas rendered accessible (Burford et al., 1964). So theremoval of organic matter enhances the accessibility of microporesurfaces to nitrogen (Theng et al., 1999). The schematic overviewseems to give the impression that H2O2 treatmentmost likely had no ef-fect on iron oxides and clay minerals without organic matter (Fig. 3b)and also no effect on organic matter, which was protected by soil min-erals. But it has to bementioned that e.g., the abundance of smaller par-ticles and microporosity present in soils strongly controls mineralsurfaces and the H2O2 treatment may alter the mineralogy properties.This issue becomes more serve if the H2O2 treatment was done at hightemperature. The literature gives not much consistency in H2O2 usebefore SSA determination of the mineral phase (see Mikutta et al.,2005). It is suggested that organic matter sorption to mineral surfacesreduces effectively the SSA due to micropores, which suggests thatthese are the preferred sorption sites (Kaiser and Guggenberger,2003). Thus, SOM is most strongly bound to micropores in soil min-eral matrix and such organic matter is more resistant to chemicaloxidation. Kaiser and Guggenberger (2003) mentioned that theloss of the SSA from pores of different size on sorption of organicmatteris highly variable and the organic matter sorption on themineral surfacedoes not necessarily render all exposed surfaces. In that case, the SSAafter H2O2 treatment may likely underestimate the true mineral surfacearea especially for soils which contain considerable amounts of H2O2-resistant organicmatter. However, possible alternations of SSA and espe-cially a loss of microporosity due to heating during H2O2 treatment(Kaiser and Guggenberger, 2003; Mikutta et al., 2005) could not be con-firmed by our study. In contrast, (i) the H2O2-treated soils revealedhigher mineral surface areas compared to the untreated samples and(ii) micropores were measured almost exclusively after H2O2 treatment(Table 4). This may indicate a negligible influence of the H2O2 treatmenton remainingmineral surface area. This could be attributed to the gener-ally low amounts of short-range order minerals and almost neutral pHvalues (Mikutta et al., 2005), making the investigated soils of bothchronosequences less prone to SSA alterations. Further, micropore sur-face area was detected (i) in few clay-sized fractions in untreatedpaddy andnon-paddy soils, (ii) after H2O2 treatment in the b0.2 μmfrac-tion of paddy and non-paddy soils, and (iii) after H2O2 and the subse-quent removal of iron oxides by DCB treatment in the 20–6.3 μmfraction of paddy soils (Table 4). No samples contain micropores afterextraction of iron oxides. This was already found by other authors andEusterhues et al. (2005) came to the conclusion that micropores origi-natemainly fromDCB-solubleminerals. Further, organicmatter removalby H2O2 treatment with moderate heating can alter oxide surface prop-erties and thus gives rise to artifacts, at least in soils rich in poorly crystal-line components (Kaiser and Guggenberger, 2003; Mikutta et al., 2005).

The removal of FeDCB oxides (SSADCB) by DCB treatment led in somecases to similar SSA values as in the untreated clay fractions (Table 4).Losses of surface area may be induced if iron oxides occurred withoutbeing protected by organic matter (Fig. 3a versus c) or if the surfacearea of clay minerals, covered by iron oxides, was smaller than theentire outer surface area of the iron oxides. The decrease in SSA after

DCB treatment was also reported by Kahle et al. (2002). Kaiser andGuggenberger (2000) attributed this to the removal of iron oxides,which are known to have a high large external SSA. It should be notedthat changes in SSA by DCB treatment cannot be attributed to the lossof iron oxides alone. Dithionite is considered to have an impact onother soil mineral phases (Borggaard, 1982) and dissolution of OCfrom organo-Fe(III) complexes by dithionite extraction is possible(Wagai and Mayer, 2007). A single dithionite extraction is sufficient toremove all reducible iron and dissolves some of the soil organic matterdue to its alkalinity and the OM-iron associations present in soil (Wagaiand Mayer, 2007). The interpretation of the SSA data after DCBtreatment is even more complex because the citrate contained in DCBsolution, could sorb ontomineral surfacewhile dithionite dissolves ped-ogenic iron oxides. Therefore, residual organic acid makes quantitativeinterpretation problematic (Filimonova et al., 2006).

The subsequent removal of FeDCB oxides after H2O2 treatment(SSAH2O2 + DCB: silicate surface area without organic matter and ironoxide minerals) caused an increase in SSA and presented the largestmineral surface for almost all silt-sized fractions of paddy and non-paddy soils (Table 4), whereas no consistent trend was observed forthe paddy and non-paddy in the coarse clay fraction (Table 4). However,most fine clay-sized fractions of the paddy chronosequence, and espe-cially the b0.2 μm fraction of P 1000 and P 2000 are characterized bythe largest mineral surface area after the subsequent removal of FeDCBoxides after H2O2 treatment (Table 4). Kahle et al. (2002) concludedthat H2O2 and subsequent treatment with DCB removed coatings fromthe clay minerals and left them with clean surfaces. Hodson et al.(1998) explained that the removal of oxyhydroxides may break upthe aggregation with the clay-sized particles and cause a marked in-crease in SSA. It is known that short-range ordered forms of iron oxides(ferrihydrite) and also nanocrystalline goethite are present in soils withfrequent changes in redox conditions (Mansfeldt et al., 2011; van derZee et al., 2003), which are present in the b0.2 μm fraction and maycontribute to a strong aggregation of fine clay particles, but may haveonly negligible effects on the aggregation of larger particles. This be-comes especially obvious in the b0.2 μm fraction of the two oldestpaddy soils (P 1000 and P 2000) which have participated the mostredox cycles. In the SOM-rich paddy topsoils of the two oldest soils (P1000 and P 2000), the described breakup of soil aggregates was onlyachieved after the removal of protective SOM (Table 4). Therefore, thecombined treatment (H2O2 + DCB) showed a higher effect on SSAthan each treatment alone. That confirms that iron oxides and SOMpro-tect each other in organo-mineral associations, leading to an incompleteremoval of iron–SOM-compounds after single treatments, and thus toan underestimation of SSA in b0.2 μm fractions of paddy soils. Thehigh SSA increase after the combined treatment underlines the largeprotective effect of SOM bound to iron oxides. If this was not the case,the H2O2 treatment alone would cause the highest SSA due to removalof organic matter, leading to uncovered surfaces of iron oxides andclay minerals. Further, it should be mentioned that the concentrationsof OC and iron oxides alone provide no indication about the SSA andtheir coverage by iron oxides and organic matter. The configurationand how iron oxides are connected with silicate mineral surfaces andorganic matter (distinct concretions or oxide layers covering silicateparticles) cannot be answered alone by the concentration.

4.2.2. Formation of mineral surface area during soil developmentSorption of organic molecules to the surface area of the mineral

phase is mainly due to expandable phyllosilicates (Kennedy et al.,2002; Theng et al., 1999) and poorly crystalline iron oxides (Borggaard,1982; Kennedy et al., 2002). In general, increasing soil evolution is char-acterized by advanced mineral weathering. Silicate weathering reducessizes of primary particles combined with the formation of clay-sizedsecondary minerals (Hodson et al., 1998). Such trends are commonlyobserved in chronosequences (Merritts et al., 1991; Muhs, 1982)and cause increases of the mineral surface area with time. In both


investigated chronosequences, although the proportion of SSAwhich is covered by OC increased, the SSA after organic matter re-moval did not increase with soil age (Table 4). Therefore we assumelow weathering of minerals in paddy and non-paddy soils with time.Smaller particle sizes correlated with larger FeDCB contents (Fig. 1a),but we observed no increase of the FeDCB oxide proportion with de-velopment of both soil groups. Fig. 1b supports this statement, be-cause there is no chronological arrangement of the symbols in thefigure. The relation between the FeDCB oxide concentration and theSSA without organic matter (Fig. 1b) separates the individual frac-tions but there was no evidence for the formation of additional min-eral surfaces. The mineral surface areas after H2O2 and DCBtreatment (Table 4) likewise showed no consistent trend with timeand may give no indication for further formation of clay-sized min-erals. Therefore, we could not confirm our (i) research question ifthe formation of soil silicate minerals and oxides is morepronounced in decalcified paddy soils. Our results provide no evi-dence that the decalcification of (paddy) soils resulted in the forma-tion of additional mineral surfaces during the time span of theinvestigated chronosequences. Both chronosequences do not indicatea further formation of clay-sizedminerals after deposition of the estuarinesediment from theYantze River. This is supported by unchanged soilmin-eralogy (Wissing et al., 2013), and the slightly acidic pH values may nothave been sufficient for extensive mineral weathering. The soil materialof the chronosequences was already characterized by a certain degree ofweathering before it was re-deposited at the Hangzhou Bay. Therefore,secondary clay minerals such as mixed layer minerals and kaolinitewere already components of the soil material in the tidal flat(0 years), marshland (30 years) (Kölbl et al., 2014–in this issue) andsince the earliest stages of paddy and non-paddy soil development.Dümig et al. (2012) investigated a relatively young chronosequence(15–120 years) that developed after glacier retreat, and they couldlikewise not confirm a chronological trend of SSA values over time.The authors explained that the OC accumulation may be faster thanthe supply ofmineral surfaces byweathering during initial soil develop-ment, which appears at later stages of pedogenesis. White et al. (1996)and Hodson et al. (1998) found increasing SSA values with ongoing soildevelopment in much older soil chronosequences (0.2–3000 ky;80–13,000 years), indicating decreasing particle sizes due to progres-sive weathering (Hodson et al., 1998).

4.2.3. Contribution of silicate (clay) and iron oxide surfacesfor OC accumulation

The OC concentrations in the b2.0 μm fraction increased withsoil age, resulting in decreasing surface areas of the mineral phase(Fig. 2a). The OC accumulation was strongest in the b0.2 μm frac-tion (having the highest mineral surface area). In the b0.2 μm frac-tion of paddy soils, the proportion of SSA that was covered by OCrose from 31% in P 50 to 75% in P 2000 (Table 4). In contrast, theOC increase in non-paddy soils was much less, from about 30% (NP50) to 41% in NP 700 (Table 4). This was reflected by a strong linearrelationship of the SSAuntreated values and the OC contents of paddy(and to a lesser extent: non-paddy clay-sized fractions; Fig. 2a). Thecorrelation increased with decreasing particle size and was highestin the b0.2 μm fraction. In summary, the results supported the (ii)research question and we can confirm higher organic matter cover-age onmineral surfaces in decalcified paddy soils. However, the sec-ond part of the (ii) research question can primarily only beconfirmed for paddy soils. An increase of OC coverage on mineralsurfaces with soil development was not observed for non-paddysoils.

Iron oxides determined by oxalate extraction, e.g., ferrihydrite, areknown to stabilize OCbecause of their large and highly reactive externalsurface area, with values between 200 and 1200 m2 g−1 (Borggaard,1982; Bracewell et al., 1970; Parfitt, 1989). Recent studies withMössbauer spectroscopy revealed that chemical extraction methods

are not mineral selective. For example, acid-ammonium oxalate dis-solves not only ferrihydrite but also for the most part, if notcompletely, goethite with very small particle size, i.e., nano-goethite (see Cornell and Schwertmann, 2003; Thompson et al.,2006, 2011). Van der Zee et al. (2003) found a nano-goethite witha particle size of only 5 nm and mentioned that these sizes are com-parable to the values 1–3 nm for natural and synthetic two lineferrihydrites. Mansfeldt et al. (2011) had repeatedly pointed outthat the extraction with oxalate is not specifically selective for ferri-hydrite. Iron associated with organic matter will also be extractedby oxalate. This might be of major importance in the uppermostSOM-rich soil horizons in paddy soils of the present study.Wissing et al. (2013) mentioned that higher contents of FeOX oxidesin paddy soils are associated with higher proportions of mineral-associated SOM than in non-paddy soils, underlining the high ability ofFeOX oxides to stabilize organic matter. This is further supportedby the fact that the treatment with H2O2 and the subsequentremoval of FeDCB oxides revealed the largest surface area in themost fine clay fractions, indicating a large interactive protectionbetween organic matter and iron oxides. Therefore, we assumecomplex associations of clay minerals, iron oxides and SOM toplay a decisive role in OC accumulation (see research question(iii)) during paddy soil formation.

4.3. Does decalcification increase the accessibility of soil mineral surfacesfor OC coverage in paddy soils?

One of the first soil-forming processes in calcareous parent materialis decalcification, which has to be completed before secondary mineralsstart to form (Talibudeen and Arambarri, 1964). The carbonate loss inmany soils which derived from marine sediments is accelerated bysulfideoxidation bywhich sulfuric acid is formed and calciumcarbonateis dissolved (Brümmer et al., 1971). The loss of carbonate leads to adecrease in pH inmany soils (van den Berg and Loch, 2000). In this con-text, our ideawas that the faster dissolution of carbonate and leaching ofCa2+ ions in paddy soils may promote dispersion of soil aggregates,making additional clay mineral surfaces available for OC coverage. Thisidea was supported by the lower OC coverage on mineral surfaces innon-paddy soils, in which decalcification was much lower. We assumethat potentially strong aggregation by carbonate Ca2+ ion bridgesmost likely obstructs the organic matter coverage on mineral surfacesand with this the organic matter accumulation in non-paddy soils.In contrast, the higher OC coverage in the investigated paddy soilsmay be promoted by enhanced accessibility of mineral surface areadue to disaggregation during decalcification. However, our studyrevealed generally higher OC contents in paddy soils. Paddymanagement with rice cropping under irrigated conditions isconsidered to favor OC accumulation (Sahrawat, 2004; Zhangand He, 2004) due to retarded OC decomposition (Sahrawat,2004) and/or the fact that paddy soils receive large carbon inputvia organic fertilizers and plant residues (Gong and Xu, 1990;Tanji et al., 2003).

4.4. Paddy soils — soils with special soil-forming processes

Due to our results, we designed paddy-specific soil-formingprocesses (Fig. 4) which were markedly different from those ofnon-paddy soils:

(1) The decalcification of the topsoil horizons was already finishedwithin 100 years of paddy soil development, whereas thedecalcification of non-paddy soils required less than 700 years(Fig. 4).

(2) A marked proportion of modern carbon (confirmed by higherOC accumulation) was added to the inherited carbon during100 years of paddy soil development (Fig. 4), whereas non-

Site Depth (cm) 14C (corrected pMC) Conventional age

Bulk soil b0.2 μm Bulk soil b0.2 μm

Mudflata 0–2 62 ± 0.2 – 3900 ± 30 BP –

18–20 61 ± 0.2 – 4000 ± 30 BP –

Marshb 2–3 90 ± 0.3 – 840 ± 30 BP –

P 50 0–7 102 ± 0.4 92 ± 0.3 Modern 640 ± 20BPP 100 0–9 106 ± 0.3 104 ± 0.3 Modern ModernP 300 0–18 101 ± 0.3 106 ± 0.4 Modern ModernP 700 0–10 99 ± 0.3 97 ± 0.3 110 ± 20 BP 270 ± 20 BPP 1000 0–10 103 ± 0.3 102 ± 0.3 Modern ModernP 2000 0–15 112 ± 0.3 108 ± 0.3 Modern ModernNP 50 0–9 101 ± 0.4 98 ± 0.3 Modern 170 ± 20 BPNP 100 0–14 98 ± 0.3 85 ± 0.3 150 ± 30 BP 1300 ± 30 BPNP 300 0–11 101 ± 0.3 96 ± 0.3 Modern 300 ± 20 BPNP 700 0–11 103 ± 0.3 97 ± 0.3 Modern 280 ± 20 BP

14C data of paddy and non-paddy soils were determined by the Leibnitz-Laboratory at


paddy soils generally preserved their carbon signature(Appendix A), indicating a slower accumulation of modernOC during pedogenesis (Fig. 4).

(3) The periodical changes in redox conditions and the higheramounts of organic matter in paddy soils (Fig. 4) allowedthe persistence of FeOX oxides (Wissing et al., 2013), whichin turn act as sorbent for organic matter.

Crystallization of iron oxides may be retarded and inhibited in soilswith water-soluble constituents such as organic compounds, whichhave a high affinity towards iron oxide surfaces (Eusterhues et al.,2008; Schwertmann, 1966; Schwertmann et al., 1982). In contrast,non-paddy soils may provide less mineral surface area in the form ofFeOX oxides due to missing redox cycles and as a consequence oflow OC concentrations. Soils under periodically anoxic conditionsare generally characterized by higher OC accumulation due tothe reduced decomposition (Sahrawat, 2004). The OC accumula-tion during paddy soil development on marshlands may promotehigher proportions of SOM–FeOX oxide association since initialstages of pedogenesis.

the Christian-Albrechts-University in Kiel. aData from Bräuer et al. (2012). bPersonalcommunication with Tino Bräuer (unpublished data). The P 700 site was characterizedby a low 14C content, which is attributed to a markedly high proportion of n-alkenesdue to contamination with fossil fuels (Müller-Niggemann et al., 2012).

5. Conclusions

We investigated the soil development on calcareous marinesediments using two chronosequences with similar parent mate-rial; one used for paddy rice cultivation, the other for uplandcrops. Within 2000 years of pedogenesis, no change in clay min-eral composition and mineral surface area was observed. Butthe soils differed in the degree of decalcification and OC accu-mulation and in the formation of iron oxides. Paddy soils arecharacterized by specific soil-forming processes which weremarkedly different from those of non-paddy soils. Paddy soil man-agement led to an enhanced decalcification and larger OC accumu-lation. Management-induced redox cycles led to a higherproportion of FeOX oxides. Their large SSA, added to the surfacearea of clay-sized minerals, provided additional options for OCcoverage.

Selective removal of organicmatter byH2O2 treatment or iron oxidesby DCB showed that iron oxides and SOM protect each other in organo-mineral associations. Only the combined treatment of H2O2 and DCBleads to completely uncovered mineral surface areas and revealed thecomplex association between clay minerals, iron oxides and SOM.These associations play a decisive role in OC accumulation andpromote the higher organic matter coverage on mineral surfacesin paddy soils.

Acknowledgments

The authors thank Rui Yin and the Institute of Soil Science, Chi-nese Academy of Sciences, in Nanjing for support during samplingat the chronosequence site around Cixi and for logistic handling.We thank especially Reinhold Jahn and Vanessa Vogelsang for pro-viding us with inorganic carbon data and Pieter Grootes for determi-nation of the radiocarbon content. Markus Steffens is acknowledgedfor assisting with statistical analysis and Werner Häusler for thesupport concerning the soil mineralogy measurements. We thankMonika Heilmeier for technical assistance and Martina Bauer,Carolin Botond, Robert U. Hagemann, Tahereh Javaheri, JulianeTeichmann and Maria Vonach for student research assistance. Weare grateful to the Deutsche Forschungsgemeinschaft (DFG) fortheir generous funding of Research Unit FOR 995 “Biogeochemistryof paddy soil evolution.” The authors also thank the two anony-mous reviewers for their helpful comments which greatly improved themanuscript.

Appendix A

Radiocarbon (14C) content in percentmodern carbon (pMC) and theconventional carbon age of mudflat, marshland, paddy (P) and non-paddy (NP) bulk soil and the b0.2 μm fraction (uppermost A horizons).

References

Balabane, M., Plante, A.F., 2004. Aggregation and carbon storage in silty soil using physicalfractionation techniques. Eur. J. Soil Sci. 55, 415–427.

Balesdent, J., Mariotti, A., Guillet, B., 1987. Natural 13C abundance as a tracer for studiesof soil organic matter dynamics. Soil Biol. Biochem. 19, 25–30.

Bartoli, F., Poulenard, A.J., Schouller, B.E., 2007. Influence of allophane and organic mattercontents on surface properties of Andosols. Eur. J. Soil Sci. 58, 450–464.

Borggaard, O.K., 1982. The influence of iron oxides on the surface area of soil. J. Soil Sci. 33,443–449.

Bosatta, E., Agren, G.I., 1997. Theoretical analyses of soil texture effects on organic matterdynamics. Soil Biol. Biochem. 29, 1633–1638.

Bracewell, J.M., Campbell, A.S., Mitchell, B.D., 1970. An assessment of some thermaland chemical techniques used in the study of the poorly-ordered aluminosilicatesin soil clays. Clay Miner. 8, 325–335.

Bräuer, T., Grootes, P.M., Nadeau, M.-J., Andersen, N., 2012. Downward carbon transport ina 2000-year rice paddy soil chronosequence traced by radiocarbon measurements.Nucl. Instr. Meth. B. http://dx.doi.org/10.1016/j.nimb.2012.07.012.

Brümmer, G., Grünwaldt, H.S., Schroeder, D., 1971. Beitrage zur Genese undKlassifizierung derMarschen II. Zur Schwefelmetabolik in Schlicken und Salzmarschen.Z. Pflanzenernähr. Bodenkd. 128, 208–220.

Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multimolecular layers.J. Am. Chem. Soc. 60 (2), 309–319.

Burford, J.R., Deshpande, T.L., Greenland, D.J., Quirk, J.P., 1964. Influence of organicmaterials on the determination of the specific surface area of soils. J. Soil Sci. 15 (2),192–201.

Carter, D.L., Mortland, M.M., Kemper,W.D., 1986. Methods of soil analysis, Part 1. Physicaland Mineralogical Methods—Agronomy Monograph2nd ed. 9.

Cheng, Y., Yang, L.-Z., Cao, Z.-H., Ci, E., Yin, S., 2009. Chronosequential changes of selectedpedogenic properties in paddy soils as compared with non-paddy soils. Geoderma151, 31–41.

Chenu, C., Plante, A.F., 2006. Clay-sized organo-mineral complexes in a cultivationchronosequence: revisiting the concept of the ‘primary organo-mineral complex’.Eur. J. Soil Sci. 57 (4), 596–607.

Chiou, C.T., 1990. The surface area of soil organic matter. Environ. Sci. Technol. 24,1164–1166.

Cornell, R.M., Schwertmann, U., 2003. The Iron Oxides. Structure, Properties, Reactions,Occurrences and Uses, 2nd edn. Wiley-VHC, Weinheim.

Dümig, A., Häusler, W., Steffens, M., Kögel-Knabner, I., 2012. Clay fractions from a soilchronosequence after glacier retreat reveal the initial evolution of organo-mineralassociations. Geochim. Cosmochim. Acta 85, 1–18.

Eusterhues, K., Rumpel, C., Kögel-Knabner, I., 2005. Organo-mineral associations insandy acid forest soils: importance of specific surface area, iron oxides and micro-pores. Eur. J. Soil Sci. 56, 753–763.

Eusterhues, K., Wagner, F.E., Häusler, W., Hanzlik, M., Knicker, H., Totsche, K.U., Kögel-Knabner, I., Schwertmann, U., 2008. Characterization of ferrihydrite-soils organicmatter coprecipitates by X-ray diffraction and Mössbauer spectroscopy. Environ.Sci. Technol. 42, 7891–7897.

FAO, 2006. Guidelines for Soil Description. FAO, Rome (97 pp.).Filimonova, S.V., Knicker, H., Kögel-Knabner, I., 2006. Soil micro- and mesopores studied

by N2 adsorption and 129Xe NMR of adsorbed xenon. Geoderma 130, 218–228.

http://refhub.elsevier.com/S0016-7061(13)00440-0/rf0005














http://dx.doi.org/10.1016/j.nimb.2012.07.012




































Gong, Z.T., Xu, Q., 1990. Paddy soils. Soils of China.Science Press, Beijing 233–260.Guo, Z.G., Yang, Z.S., Qu, Y.H., Fan, D.J., 2000. Study on comparison sedimentary

geochemistry of mud area on East China Sea continental shelf. Acta Sedimentol. Sin.18, 284–289.

Hodson, M.E., Langan, S.J., Kennedy, F.M., Bain, D.C., 1998. Variation in soil surface area ina chronosequence of soils from Glen Feshie, Scotland and its implications for mineralweathering rate calculations. Geoderma 85, 1–18.

International Union of Soil Sciences Working Group, 2007. World reference base for soilresources 2007. World Soil Resources Reports, vol. 103. FAO, Rome.

Jenkinson, D.S., Rayner, J.H., 1977. The turnover of soil organic matter in some of theRothamsted classical experiments. Soil Sci. 123, 298–305 (cited from Balabane andPlante, 2004).

Jenkinson, D.S., Hart, P.B.S., Rayner, J.H., Parry, L.C., 1987. Modelling the turnover oforganic matter in long-term experiments at Rothamsted. INTECOL Bull. 15, 1–8(cited from Eusterhues et al., 2005).

Jilan, D., Kangshan, W., 1989. Changjiang river plume and suspended sediment transportin Hangzhou Bay. Cont. Shelf Res. 9, 93–111.

Kahle, M., Kleber, M., Jahn, R., 2002. Predicting carbon content in illitic clay fractions fromsurface area, cation exchange capacity and dithionite-extractable iron. Eur. J. Soil Sci.53, 639–644.

Kaiser, K., Guggenberger, G., 2000. The role of DOM sorption to mineral surfaces in thepreservation of organic matter in soils. 31, 711–725.

Kaiser, K., Guggenberger, G., 2003. Mineral surfaces and soil organic matter. Eur. J. Soil Sci.54 (2), 219–236.

Kalbitz, K., Kaiser, K., Fiedler, S., Kölbl, A., Amelung, W., Bräuer, T., Cao, Z.-H., Don, A.,Grootes, P., Jahn, R., Schwark, L., Vogelsang, V., Wissing, L., Kögel-Knabner, I.,2013. The carbon count of 2000 years of rice cultivation. Glob. Chang. Biol. 19,1107–1113.

Kennedy, M.J., Pevear, D.R., Hill, R.J., 2002. Mineral surface control of organic carbon inblack shale. Science 295, 657–660.

Kiem, R., Kögel-Knabner, I., 2002. Refractory organic carbon in particle-size fractions ofarable soils II: organic carbon in relation to mineral surface area and iron oxides infractions b6 μm. Org. Geochem. 33, 1699–1713.

Kölbl, A., Schad, P., Jahn, R., Amelung, W., Bannert, A., Cao, Z.H., Fiedler, S., Kalbitz, K.,Lehndorff, E., Müller-Niggemann, C., Schloter, M., Schwark, L., Vogelsang, V.,Wissing, L., Kögel-Knabner, I., 2014. Accelerated soil formation due to paddy manage-ment on marshlands (Zhejiang Province, China). Geoderma 228, 67–89 (in this issue).

Lal, R., 2002. Soil carbon sequestration in China through agricultural intensification, andrestoration of degraded and desertified ecosystems. Land Degrad. Dev. 13, 469–478.

Li, Z.-P., Zhang, T.-L., Li, D.-C., Velde, B., Han, F.-X., 2005. Changes in soil properties ofpaddy fields across a cultivation chronosequence in subtropical China. Pedosphere15 (1), 110–119.

Lord III, C.J., 1980. The Chemistry and Cycling of Iron, Manganese, and Sulfur in Salt MarshSediments. PhD dissertation Univ. Delaware, Newark, USA (cited from Portnoy andGiblin, 1997).

Mansfeldt, T., Schuth, S., Häusler, W., Wagner, F.E., Kaufhold, S., Overesch, M., 2011. Ironoxide mineralogy and stable iron isotope composition in a Gleysol with petrogleyicproperties. J. Soils Sediments 12 (1), 97–114.

Mayer, L.M., Xing, B., 2001. Organic matter-surface area relationship in acid soils. Soil Sci.Soc. Am. J. 65, 250–258.

Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by a dithionite–citrate system buffered with sodium bicarbonate. Clays and Clay Minerals, Proceed-ings of the 7th National Conference, pp. 317–327.

Merritts, D.J., Chadwick, O.A., Hendricks, D.M., 1991. Rates and processes of soil evolutionon uplifted marine terraces, northern California. Geoderma 51, 241–275.

Mikutta, R., Kleber, M., Kaiser, K., Jahn, R., 2005. Review: organic matter removal fromsoils using hydrogen peroxide, sodium hypochlorite, and disodium peroxodisulfate.Soil Sci. Soc. Am. J. 69, 120–135.

Milliman, J.D., Meade, R.H., 1983. World-wide delivery of river sediment to the oceans.J. Geol. 91, 1–21.

Milliman, J.D., Syvitski, J.P.M., 1992. Geomorphic/tectonic control of sediment discharge tothe ocean: the importance of small mountainous rivers. J. Geol. 100, 525–544.

Moore, D.M., Reynolds, R.C., 1989. X-ray Diffraction and the Identification and Analysis ofClay Minerals. Oxford University Press, London.

Muhs, D.R., 1982. A soil chronosequence on Quaternary marine terraces San ClementeIsland, California. Geoderma 28, 257–283.

Müller-Niggemann, C., Bannert, A., Schloter, M., Lehndorff, E., Schwark, L., 2012. Intra-versus inter-site macroscale variation in biogeochemical properties along a paddysoil chronosequence. Biogeosciences 9 (3), 1237–1251.

Nadeau, M.-J., Schleicher, M., Grootes, P.M., Erlenkeuser, H., Gottdang, A., Mom, D.J.W.,Samthein, J.M., Willkomm, H., 1997. Nucl. Instrum.Methods Phys. Res. Sect. B 123, 22.

Neue, H.U., Gaunt, J.L., Wang, Z.P., Becker-Heidmann, P., Quijano, C., 1997. Carbon intropical wetlands. Geoderma 79, 163–185.

Pan, G., Li, L., Wu, L., Zhang, X., 2003a. Storage and sequestration potential of topsoilorganic carbon in China's paddy soils. Glob. Chang. Biol. 10, 79–92.

Pan, G.X., Li, L.Q., Zhang, X.H., Dai, J.Y., Zhou, Y.C., Zhang, P.J., 2003b. Soil organic carbonstorage of China and the sequestration dynamics in agricultural lands. Adv. EarthSci. 18, 609–618 (in Chinese).

Parfitt, R.L., 1989. Phosphate reactions with natural allophane, ferrihydrite and goethite.J. Soil Sci. 40, 359–369.

Portnoy, J.W., 1999. Saltmarsh diking and restoration: biochemical implications of alteredwetland hydrology. Environ. Manag. 24, 111–120.

Portnoy, J.W., Giblin, A.E., 1997. Effects of historic tidal restrictions on salt marsh sedimentchemistry. Biogeochemistry 36, 275–303.

Ransom, B., Kim, D., Kastner, M., Wainwright, S., 1998. Organic matter preservationon continental slopes: importance of mineralogy and surface area. Geochim.Cosmochim. Acta 62, 1329–1345.

Robert, M., Chenu, C., 1992. Interactions between soil minerals and microorganisms. In:Stotzky, G., Bollag, J.M. (Eds.), Soil Biochemistry, vol. 7. Marcel Dekker, New York,pp. 307–404.

Sahrawat, K.L., 2004. Organic matter accumulation in submerged soils. Adv. Agron. 81,169–201.

Schwertmann, U., 1964. Differenzierung der Eisenoxide des Bodens durch Extraktion mitAmmoniumoxalat-Lösung. Z. Pflanzenernaehr. Düngung Bodenkd. 105 (3), 194–202.

Schwertmann, U., 1966. Inhibitory effect of soil organic matter on crystallization ofamorphous ferric hydroxide. Nature 212, 645–646.

Schwertmann, U., Schulze, D.G., Murad, E., 1982. Identification of ferrihydrite in soilsby dissolution kinetics, differential X-ray diffraction, and Mössbauer spectroscopy.Soil Sci. Soc. Am. J. 46, 869–875.

Talibudeen, O., Arambarri, P., 1964. The influence of the amount and the origin of calciumcarbonates on the isotopically exchangeable phosphate in calcareous soils. J. Agric.Sci. (Camb.) 62, 93–97.

Tanji, K.K., Gao, S., Scardaci, S.C., Chow, A.T., 2003. Characterization redox status of paddysoils with incorporated rice straw. Geoderma 114, 333–353.

Theng, B.K.G., 1979. Formation and Properties of Clay–Polymer Complexes. ElsevierScience Publishing Co., New York.

Theng, B.K.G., Ristori, G.G., Santi, C.A., Percival, H.J., 1999. An improved method fordetermining the specific surface areas of topsoils with varied organic matter content,texture and clay mineral composition. Eur. J. Soil Sci. 50, 309–316.

Thompson, A., Chadwick, O.A., Rancourt, D.G., Chorover, J., 2006. Iron-oxide crystallinityincreases during redox oscillations. Geochim. Cosmochim. Acta 70, 1710–1727.

Thompson, A., Rancourt, D.G., Chadwick, O.A., Chorover, J., 2011. Iron solid-phasedifferentiation along a redox gradient in basaltic soils. Geochim. Cosmochim. Acta75, 119–133.

Tipping, E., 1981. The adsorption of aquatic humic substances by iron oxides. Geochim.Cosmochim. Acta 45 (2), 191–199.

van den Berg, G.A., Loch, J.P.G., 2000. Decalcification of soils subject to periodicwaterlogging. Eur. J. Soil Sci. 51, 27–33.

van der Zee, C., Roberts, D.R., Rancourt, D.G., Slomp, C.P., 2003. Nanogoethite is thedominant reactive oxyhydroxide phase in lake and marine sediments. Geology 31,993–996.

von Lützow, M., Kögel-Knabner, I., Ekschmitt, K., Matzner, E., Guggenberger, G.,Marschner, B., Flessa, H., 2006. Stabilization of organic matter in temperate soils:mechanisms and their relevance under different soil condition — a review. Eur.J. Soil Sci. 57, 426–445.

Wagai, R., Mayer, L.M., 2007. Sorptive stabilization of organic matter in soils by hydrousiron oxides. Geochim. Cosmochim. Acta 71, 25–35.

Wagai, R., Mayer, L.M., Kitayama, K., 2009. Extent and nature of organic coverage of soilmineral surfaces assessed by a gas sorption approach. Geoderma 149 (1–2), 152–160.

Wang, H., Zuosheng, Y., Saito, Y., Liu, J.P., 2008. Reconstruction of sediment flux from theChangjiang (Yangtze River) to the sea since the 1860s. J. Hydrol. 349, 318–332.

Weiler, R.R., Mills, A.A., 1965. Surface properties and pore structure of marine sediments.Deep-Sea Res. 12, 511–529.

White, A.F., Blum, A.E., Schulz, M.S., Bullen, T.D., Harden, J.W., Peterson, M.L., 1996. Chem-ical weathering rates of a soil chronosequence on granitic alluvium: I. Quantificationof mineralogical and surface area changes and calculation of primary silicate reactionrates. Geochim. Cosmochim. Acta 60 (14), 2533–2550.

Williams, B.G., Greenland, D.J., Quirk, J.P., 1967. The effect of polyvinyl alcohol on thenitrogen surface area and pore structure of soils. Aust. J. Soil Res. 5, 77–83.

Wissing, L., Kölbl, A., Vogelsang, V., Fu, J., Cao, Z., Kögel-Knabner, I., 2011. Organic carbonaccumulation in a 2000-year chronosequence of paddy soil evolution. Catena 87,376–385.

Wissing, L., Kölbl, A., Häusler, W., Schad, P., Cao, Z., Kögel-Knabner, I., 2013. Management-induced organic carbon accumulation in paddy soils: the role of organo-mineralassociations. Soil Tillage Res. http://dx.doi.org/10.1016/j.still.2012.08.004.

Wu, J., 2011. Carbon accumulation in paddy ecosystems in subtropical China: evidencefrom landscape studies. Eur. J. Soil Sci. 62, 29–34.

Xie, D., Wang, Z., Gao, S., De Vriend, H.J., 2009. Modeling the tidal channelmorphodynamics in a macro-tidal embayment, Hangzhou Bay, China. Cont. ShelfRes. 29, 1757–1767.

Zhang, M., He, Z., 2004. Long-term changes in organic carbon and nutrients of an Ultisolunder rice cropping in southeast China. Geoderma 118, 167–179.

Zottoli, R., 1973. Introduction to Marine Environments. CV Mosby Company, Saint Louis,Mosby, USA (cited from Portnoy and Giblin, 1997).

Zou, P., Fu, J., Cao, Z., 2011. Chronosequence of paddy soils and phosphorus sorption–desorption properties. J. Soil Sediment. 11, 249–259.











































































































































http://dx.doi.org/10.1016/j.still.2012.08.004












organic carbon accumulation on soil mineral surfaces in paddy soils derived from tidal wetlands

Documents