Effect of Wetland Reclamation on Soil Organic Carbon Stability in Peat Mire Soil Around Xingkai Lake in Northeast China

Content and density of soil organic carbon (SOC) and labile and stable SOC fractions in peat mire soil in wetland, soybean field and rice paddy field reclaimed from the wetland around Xingkai Lake in Northeast China were studied. Studies were designed to investigate the impact of reclamation of wetland for soybean and rice farming on stability of SOC. After reclamation, SOC content and density in the top 0–30 cm soil layer decreased, and SOC content and density in soybean field were higher than that in paddy field. Content and density of labile SOC fractions also decreased, and density of labile SOC fractions and their ratios with SOC in soybean field were lower than that observed in paddy field. In the 0–30 cm soil layer, densities of labile SOC fractions, namely, dissolved organic carbon (DOC), microbial biomass carbon (MBC), readily oxidized carbon (ROC) and readily mineralized carbon (RMC), in both soybean field and paddy field were all found to be lower than those in wetland by 34.00% and 13.83%, 51.74% and 35.13%, 62.24% and 59.00%, and 64.24% and 17.86%, respectively. After reclamation, SOC density of micro-aggregates (< 0.25 mm) as a stable SOC fraction and its ratio with SOC in 0–5, 5–10, 10–20 and 20–30 cm soil layers increased. SOC density of micro-aggregates in the 0–30 cm soil layer in soybean field was 50.83% higher than that in paddy field. Due to reclamation, SOC density and labile SOC fraction density decreased, but after reclamation, most SOC was stored in a more complex and stable form. Soybean farming is more friendly for sustainable SOC residence in the soils than rice farming.


Introduction
Soil organic carbon (SOC) is an important source or sink of atmospheric CO 2 .A slight change in the SOC pool can induce a significant change in the CO 2 concentration in the atmosphere (Lal, 2004;Eglin et al., 2010).The SOC pool is heterogeneous, consisting of labile carbon with a short turnover time and stable carbon with a long turnover time (Haynes, 2005;Davidson and Janssens, 2006).The labile pool is sensitive to climate change and human activity, and determines the SOC circulation.The stable pool can control and maintain SOC as a long-term sink of atmospheric CO 2 .This contributes to maintenance of the SOC for long periods of time, thus mitigating climate change (Trumbore et al., 1990;Von Lützow et al., 2007;Budge et al., 2011).Cur-rently, quantitative determination of labile and stable SOC fractions involves physical, chemical and mineralization incubation methods (von Lützow et al., 2007;Puissant et al., 2017).Dissolved organic carbon (DOC), microbial biomass carbon (MBC), readily oxidized carbon (ROC) and readily mineralized carbon (RMC) as labile fractions can be used to indicate activity of SOC (Plante et al., 2011;Huo et al., 2013).More labile SOC is newly formed in macro-aggregates (> 0.25 mm) than in micro-aggregates (< 0.25 mm) (Six et al., 2004;Nie et al., 2014).SOC in macro-aggregates decomposes more quickly, that in micro-aggregates can remain for longer time (Six et al., 2002;Ewing et al., 2006).Micro-aggregates provide better protection for SOC than macro-aggregates (Six et al., 2002;Nie et al., 2014).The presence of SOC found in micro-aggregates indicates the stability of SOC.
SOC is susceptible to organic matter input and soil environmental conditions, such as vegetation types, soil temperature and moisture.Land use change is an important factor that influences the SOC quantity and quality (Poeplau and Don, 2013;Zhao et al., 2016;Francaviglia et al., 2017).SOC stocks in topsoil are known to decrease as a consequence of the conversion of natural ecosystem to plantation or cropland (Benbi et al., 2015a;Sheng et al., 2015;Francaviglia et al., 2017).Compared with stable SOC fractions, labile SOC fractions can respond more rapidly to land use change than the total SOC (Guidi et al., 2014;Gabarrón-Galeote et al., 2015).Labile and stable SOC fraction content and the ratio with total SOC vary with different land use changes.Francaviglia et al. (2017) suggest that SOC and its fraction content are higher in natural and undisturbed ecosystems.Labile (non-humified carbon and MBC) and stable (humic and fulvic acid carbon) fraction contents were ordered from high to low as follows: hay pasture > hay crop > former vineyard.However, in Malaysia, topsoil labile SOC content (0-15 cm) increased by 18% and 6% after forest were converted to oil palm plantation and pineapple orchard, respectively (Nahrawi et al., 2012).In Sergipe, Brazil, SOC content and active humic acid concentration in surface soil did not differ between a 12-year-old integrated coconut plantation and an adjacent remnant native Atlantic Forest (Guimarães et al., 2013).Due to human activities, Zoigê alpine wetland degraded from swamp, to swampy meadow, to meadow with observable degradation.However, residue of imported animals and plants decomposed more thoroughly in the 0-50 cm soil layers.After degradation SOC, DOC and LFOC (light fraction organic carbon) contents decreased, and most SOC was stored in more complex and stable forms (Huo et al., 2013).After cropland afforestation across Europe, the SOC shifted from stable to labile pools (Poeplau and Don, 2013).However, conversion of corn cropland to forage and grass land increased the total SOC and decreased the percentages of labile fractions (DOC, ROC and MBC) to total SOC in the Songnen Plain in China (Yu et al., 2017).In semiarid subtropical regions, labile and stable SOC content increased in undisturbed soil (undisturbed for more than 40 years) compared with cultivated soil.The ratio of stable fractions to total SOC was higher in undisturbed soil than in agroforestry, maize-wheat and sugarcane agro-ecosystem (Benbi et al., 2015a).Direction and magnitude of the SOC and SOC fractions in topsoil following land use change can significantly differ among biomes, geographical regions and with human activities.The response of SOC fractions to land conversion also depends on the type of land use change (Sheng et al., 2015).
The Sanjiang Plain is the largest distribution area of freshwater wetland in China, and an important SOC pool in Northeast China (Bao et al., 2011).SOC density was very high in peat mire soil, and bottomland around Xingkai Lake was the main distribution area of peat mire soil ① .Xingkai Lake Farm was constructed in 1955 and, since that time, wetland has been changed to farmland.This alteration in land use changed the carbon cycle model and influenced the wetland SOC pool.After reclamation, in the first 5 to 7 years, the SOC decreased quickly; then in the subsequent 15 to 20 years, the SOC tended to be stable (Song et al., 2004).After reclamation, soybean and paddy field were the typical land use patterns.In this study, wetland, and soybean and paddy field reclaimed from the wetland around Xingkai Lake with a long reclamation period were chosen as research plots, in which labile and stable SOC fractions were determined.Stability characteristics of the SOC in wet-land, soybean and paddy field were explored.The results presented here may enrich the carbon cycle research in the Sanjiang Plain, and provide a support to accurately estimate SOC pools after dramatic land use change.Such information is helpful in optimizing regional land use structures while considering both carbon emission and food production.

Study area
Study sites are located at the Xingkai Lake Wetland Experimental Station, Chinese Academy of Sciences (45°21′55″N, 132°20′00″E) (Fig. 1).This station belongs to the temperate continental monsoon climate zone and the humid and sub-humid region.This region experiences a mean annual temperature of 3.1 and a ℃ mean annual precipitation of 750 mm, of which about 70% occurs in summer.The coldest month is January with a mean monthly temperature of -19.2 , and the ℃ warmest month is July with a mean monthly temperature of 21.2 .There are a lot of ℃ snowstorms in winter and freeze-up period is from November to the following March (Huo et al., 2015).There are natural wetland with typical plant mannagrass (Glyceria spiculosa), and soybean (Giycine max) and paddy (Oryza sativa) fields reclaimed from wetland.G. max and O. sativa are both continuously cropped once a year and just litter and stubble are remained.The soybean and paddy fields are ploughed with a depth of 20 cm at the end of autumn.Inorganic nitrogen (N) fertilizer, phosphate (P) fertilizer and potash (K) fertilizer are used, with no organic fertilizer.

Experimental details
Soil samples were collected from wetland, 50 year-old soybean field and 15-year-old paddy field.It is important to note that the rice paddy field existed previously as soybean field for 35 years and then served as paddy field for 15 years.Both soybean and paddy field were reclaimed from wetland.The soil type of all the three adjacent plots is peat mire soil.After harvesting, for each sample plot, 5 sampling sites were arranged in an 'S' shape.In each sampling site, soil samples at 0-5, At each soil layer in one sampling site, five samples were mixed as one soil sample and then sealed in a polyethylene bag.Fresh soil samples were sieved through a 2 mm mesh to determine DOC, MBC and RMC; dry soil samples were sieved through a 0.150 mm mesh to determine SOC and ROC.Undisturbed dry soil was used to determine SOC in aggregates.The soil bulk density was also determined.

Detection methods
DOC determination: Fresh soil equivalents of 5 g of dry soil were shaken with 50 mL deionized water for 30 min in 100 mL polypropylene bottles on a reciprocating shaker at a speed of 200 rpm at room temperature.Then, the soil extracts were centrifuged at 8000 rpm for 10 min, and the supernatant was filtered through a 0.45 µm filter membrane.The organic C content of leach liquor was determined by a Shimadzu TOC-VCPH made in Japan to determine the DOC concentration (Jones and Willett, 2006).MBC determination: Fresh soil equivalents of 50 g of dry soil were fumigated with CHCl 3 vapor for 24 h in a desiccator.After residual CHCl 3 was removed by evacuation, fumigated portions were extracted with 0.5 mol/L K 2 SO 4 .Similar non-fumigated soil portions were extracted at the time fumigation commenced.The soil extracts were filtered through a 0.45 µm filter membrane, and total organic C in the extracts was determined by the Shimadzu TOC-VCPH.D-value (carbon) between total organic carbon concentrations of fumigation and non-fumigation extracts was determined.Then, the MBC was calculated by the formula MBC = D-value (carbon) / 0.45 (Powlson et al., 1987;Joergensen, 1996).
ROC determination: Dry soil containing 15-30 mg SOC was shaken with 25 mL of KMnO 4 (333 mol/L) for 1 h in 50 mL covered plastic centrifuge tubes on a reciprocating shaker at a speed of 200 rpm at room temperature.Soil extracts were centrifuged at 2000 rpm for 5 min.Then 1 mL of the supernatant was diluted 250 times by deionized water, and analyzed in a Shimadzu ultraviolet spectrophotometer at 565 nm.KMnO 4 concentration with sample and blank control were both calculated and D-value (carbon) was determined.KMnO 4 concentration decreased by 1 mmol/L with C decreasing by 9 mg.ROC (mg/kg) = (D-value (carbon) × 25 × 250 × 9) / (soil sample weight × 1000) (Loginow et al., 1987;Blair et al., 1995).
RMC determination: RMC was measured by incubating a fresh soil (60% field capacity) equivalent of 20 g dry soil at 25 o C in 500 mL jar without any addition of substrate.The RMC was measured as CO 2 production determined by the Los Gatos Research DLT-100 Green House Analyzer (GGA) with Syringe Injection Model Number 908-0011(w/ Syringe Inject Option) made in America.RMC was described by CO 2 -C produced over 10 days (Feller et al., 2001;Mclauchlan and Hobbie, 2004).
Determination of SOC in aggregates: 25 g of dry soil was sieved by a homemade Soil Aggregates Analyzer composed of 3 sieves (1, 0.25, 0.053 mm) in distilled water.Aggregates of > 1.000, 0.250-1.000,0.053-0.250and < 0.053 mm were gained.Aggregates were ground and sieved through a 0.150 mm mesh to determine the SOC (Huo and Lu, 2011).
The SOC content in total soil and aggregates were determined using the potassium dichromate-external heating method (Walkley and Black, 1934).Bulk density was determined by the cutting-ring method.
The soil horizon C density was calculated using the following equation (Bockheim, 2003): where HCD is horizon SOC density (g C/m 2 ); D is the bulk density (g/cm 3 ); C is the soil organic carbon concentration (g/kg); and T is the thickness of the horizon (cm).

Statistical analysis
Software SPSS19.0 developed by SPSS Company in America was used to analyze data, and OriginPro 8.6 developed by OriginLab Company in America was used to fit the plot.

SOC content in wetland, soybean and paddy field
After reclamation, SOC content decreased, and that in soybean field was higher than that in paddy field.The SOC contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy field were 74.29% and 75.66%, 68.15% and 75.25%, 73.09% and 77.28%, and 74.17% and 83.19% lower than those in wetland, respectively (Fig. 2).

SOC density in wetland, soybean and paddy field
SOC densities in the 0-30 cm soil layer in wetland, soybean and paddy field were 15 979.07 g/m 2 , 12 738.51g/m 2 and 11 335.53 g/m 2 , respectively.The data showed that SOC densities in soybean and paddy fields were 20.28% and 29.06% lower than that in wetland.SOC density in the 0-30 cm soil layer in soybean field was higher than that in paddy field.After reclaiming to soybean field, the SOC densities in 0-5, 5-10 and 20-30 cm soil layers decreased by 16.26%, 16.43% and 34.88%, respectively; in paddy field, those in 0-5, 5-10, 10-20 and 20-30 cm soil layers decreased by 19.35%, 17.98%, 21.07% and 44.44%, respectively.The SOC density in 5-10 cm soil layer was ordered from high to low as follows: soybean field > wetland > paddy field (Table 1).

DOC in wetland, soybean and paddy field
After reclamation, the DOC content decreased, and the DOC content in soybean field was lower than that in paddy field.DOC contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy field were lower than those in wetland by 75.10% and 74.50%, 75.43% and 72.67%, 75.72% and 71.50%, and 81.36% and 77.13%, respectively (Fig. 3).The DOC densities in the 0-30 cm soil layer in wetland, soybean and paddy fields were 24.59 g/m 2 , 16.23 g/m 2 and 21.19 g/m 2 , respectively.The DOC densities in the 0-30 cm soil layer in soybean field and paddy field were lower than that in wetland by 34.00% and 13.83%, respectively.In 0-5, 5-10, 10-20 and 20-30 cm soil layers, the ratio of DOC to SOC was ordered from high to low as follows: paddy field > wetland > soybean field (Fig. 3).

MBC in wetland, soybean and paddy field
Due to reclamation, the MBC content decreased.The MBC contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy fields were lower than those in wetland by 75.72% and 83.92%, 87.04% and 89.13%, 82.57% and 83.26%, and 86.37% and 72.56%, respectively.In 0-5, 5-10 and 10-20 cm soil layers, the MBC contents in soybean field were higher than those in paddy field; however, in the 20-30 cm soil layer, the case was the opposite (Fig. 4).MBC densities in the 0-30 cm soil layer in wetland, and soybean and paddy fields were g/m 2 , 65.71 g/m 2 and 88.33 g/m 2 , respectively.The MBC densities in the 0-30 cm soil layer in soybean field and paddy field were lower than that in wetland by 51.74% and 35.13%, respectively.In the 0-5 cm soil layer, the ratio of MBC to SOC was or-dered from high to low as follows: wetland > soybean field > paddy field; in the 5-10 and 10-20 cm soil layers, wetland > paddy field > soybean field; in the 20-30 cm soil layer, paddy field > wetland > soybean field (Fig. 4).

ROC in wetland, soybean and paddy field
After reclamation, the ROC content decreased.The ROC contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy fields were lower than those in wetland by 84.44% and 85.77%, 85.49% and 88.10%, 88.43% and 86.47%, and 87.89% and 90.20%, respectively.In 0-5, 5-10 and 20-30 cm soil layers, the ROC contents in soybean field was higher than those in paddy field; however, in the 10-20 cm soil layer, the case was the opposite (Fig. 5).The ROC densities in 0-30 cm soil layer in wetland, and soybean and paddy fields were 46.35 g/m 2 , 17.50 g/m 2 and 19.00 g/m 2 , respectively.The ROC densities in the 0-30 cm soil layer in soybean field and paddy field were lower than that in wetland by 62.24% and 59.00%, respectively.Overall, after reclamation the ratio of ROC to SOC decreased.In 0-5 cm soil layer the ratio was higher in soybean field than in paddy field, in other three soil layers those in soybean field were lower (Fig. 5).

RMC in wetland, soybean and paddy field
After reclamation, RMC content also decreased.It was lower in soybean field than in paddy field.RMC contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy fields were lower than those in wetland by 88.87% and 79.34%, 82.46% and 76.15%, 89.65% and 74.79%, and 88.00% and 74.00%, respectively (Fig. 6).RMC density in the 0-30 cm soil layer in wetland, and soybean and paddy fields were 43.96 g/m 2 , 15.72 g/m 2 and 36.11g/m 2 , respectively.The RMC densities in the 0-30 cm soil layer in soybean field and paddy field were lower than that in wetland by 64.24% and 17.86%, respectively.Overall, in the 0-5 and 5-10 cm soil layers, the ratio of RMC to SOC was ordered from high to low as follows: wetland > paddy field > soybean field; in the 10-20 and 20-30 cm soil layers the ratio was ordered as: paddy field > wetland > soybean field (Fig. 6).

SOC in aggregates in wetland, soybean and paddy field
The SOC density in aggregates was calculated according to the ratio of aggregate weight to the total soil, SOC content in aggregates, soil bulk density and soil layer thickness (Table 2).After reclamation, the SOC densities  in micro-aggregates (< 0.25 mm) in the 0-5, 5-10, 10-20 and 20-30 cm soil layers increased, and those in soybean field were higher than in paddy field.The SOC densities in micro-aggregates (< 0.25 mm) in the four soil layers in soybean field were higher than those in the wetland by 349.89%, 156.46%, 74.51% and 86.40%, respectively, those in paddy field were 132.37%, 3.83%, 74.05% and 7.67% higher, respectively.SOC density of micro-aggregates (< 0.25 mm) in the 0-30 cm soil layer in soybean field was 50.83% higher than that in paddy field.The ratios of the SOC densities in micro-aggregates (< 0.25 mm) to the SOC densities in the four soil layers increased.Likewise, the ratio of SOC densities in micro-aggregates (< 0.25 mm) to the SOC densities in soybean and paddy fields were higher than those in wetland by 362.75% and 156.86%, 137.50% and 16.25%, 95.24% and 134.29%, and 177.50% and 102.50% in four soil layers, respectively.Except for the 10-20 cm soil layer, the ratios of the SOC density in micro-aggregates to that in soil in soybean field was higher than that in paddy field.

Effect of reclamation on SOC
The SOC accumulation process included organic carbon inputs and outputs.The balance between the inputs of primary production and carbon outputs (decomposition, leaching, erosion, etc.) from the soil determined the size of SOC reserves (Amundson, 2001;Schlesinger and Bernhardt, 2013).After reclamation, litter and stubble input into the soil decreased dramatically.Litter and roots in wetland were (1230.42± 125.67) g/m 2 and (3072.00± 572.23) g/m 2 ; litter and stubble in soybean field were (624.17 ± 107.97) g/m 2 and (214.61 ± 21.09) g/m 2 ; and stubble in paddy field were (771.56 ± 64.32) g/m 2 (Huo et al., 2015).The wetland soil was flooded and under anoxic conditions.In an anaerobic environment, animal and plant residues is difficult to decompose, more soil organic matter tends to accumulate in wetland than in soybean or paddy fields (Huo et al., 2013, Gabriel andKellman, 2014).Litter in soybean field was greater than in paddy field by 624.17 g/m 2 .Stubble in paddy field was greater than in soybean field by 556.95 g/m 2 , but stubble in paddy field was difficult to decompose and was cleaned away before ploughing otherwise that will influence cultivation.Therefore, the SOC content and density in soybean field was higher than in paddy field.

Effect of reclamation on labile SOC fractions
DOC is a labile SOC fraction derived from plant residue leaching, SOC decomposing, root exudation and microbial metabolism (Muller et al., 2009;Asensio et al., 2014).Residue soaked in water is an important DOC source, followed by DOC production and migration into the deep soil layer (Wickland et al., 2007).After reclamation, the residue decreased influencing the microbial energy source, thus the DOC content and density shows a decrease (Qualls and Richardson, 2003;Gerke et al., 2016).Similarly, DOC stock in the 0-20 cm topsoil decreased by 29% and 78% following the conversion of natural forest to plantation and orchard, respectively (Sheng et al., 2015).Compared with soybean field, paddy field was waterlogged before harvest, which was good for DOC leaching.Large amounts of DOC were produced and migrated downward in paddy field (Kögel-Knabner et al., 2010).Additionally, fibrous root systems of rice produced more root exudates as DOC resources (Wickland et al., 2007;Asensio et al., 2014).Therefore, the DOC content and density in paddy field were higher than those in soybean field.Wetland and paddy field were waterlogged, causing the ratio of DOC to SOC to be higher than in soybean field.This is consistent with data of Huo et al. (2013), who suggested that as soil moisture content decreased, the DOC content and the percentage compared with total SOC decreased.
MBC is an important labile SOC fraction, which can be influenced by microbial activities.Microbial quantity, activity and community structure are sensitive to environmental change, residue quantity and quality (Zhang et al., 2010;Xu et al., 2013).The quantity of MBC in the 0-20 cm soil layer deceased by 44% due to the conversion of natural forestland to lowland rice field in the Paresar area in Northwest Iran (Raiesi and Beheshti, 2014).Litter and stubble in soybean and paddy fields were lower than litter in wetland by 391.64 g/m 2 and 458.86 g/m 2 , respectively.Energy and nutriment for microbes decreased and the MBC content and density decreased (Huo et al., 2015).Yu et al. (2017) found that the higher MBC content observed in alfalfa grassland compared with corn cropland could be due to the continuous input of leaf litter and root material, which provide sufficient energy and substrates for soil microor-ganisms, thus stimulating growth and activity of microbial population.In addition, changes of soil porosity, pH and temperature due to human activities can make the habitat more suitable or unsuitable for microorganisms.Soil MBC content was higher in ecosystems that were natural or less disturbed by human activities (Yu et al., 2014;Francaviglia et al., 2017).After reclamation, because of ploughing and single chemical fertilizer application, macro-aggregates were broken, causing an increase in soil bulk density.In addition, large quantities of pesticide were used (Ladd et al., 1983;Nyamadzawo et al., 2009;Pabst et al., 2013;Benbi et al., 2015b).Thus, the environmental habitat for microbes deteriorated and MBC decreased.Perishable soybean fallen leaves and loose soil provided increased substrate and better environmental conditions for microbes to live.As such, the MBC contents in the 0-5, 5-10 and 10-20 cm soil layers in soybean field were higher than those in paddy field.In the 20-30 cm soil layer, woody soybean stubble is difficult for microbes to utilize, but fibrils of rice provided more exudates and root shedding to microbes as substrate (Liao and Boutton, 2008;Xu et al., 2013;Wu et al., 2014).Thus, the MBC content was higher in paddy field.MBC density in the 0-30 cm soil layer was higher than that in soybean field by 30.69%.The ratio of MBC to SOC was higher in paddy field than in soybean field.
ROC was also the labile SOC fraction influenced by quantity, quality and decomposition degree of soil organic matter (Blair et al., 1995;Benbi et al., 2015b).Litter in wetland was greater and wetland plant roots were more developed; therefore, more labile primary soil organic carbon was produced (Benbi et al., 2015b;Chen et al., 2017).After reclamation, the ROC content and density decreased.This result is generally consistent with the results of Benbi et al. (2015a) in semiarid subtropical regions where they found that ROC content in undisturbed soil (undisturbed for more than 40 years) was higher compared with cultivated soil due to the higher inputs of above-and below-ground biomass in undisturbed soil.Sheng et al. (2015) also found that the ROC density decreased due to the conversion of natural forest to plantation and orchard.The ROC density in the 0-30 cm soil layer in the paddy field was higher than that in the soybean field by 24.11%.The ratio of ROC to SOC in paddy field was also higher than that in soybean field.This could be attributed to a combination of crop types and tillage management patterns.
Under the same moisture and temperature condition, microbial quantity and activity, SOC quality (e.g.labile and stable fractions) and residue quality (e.g.C : N, both the easy and difficult decomposing components) would affect SOC mineralization (Poeplau and Don, 2013;Wickings et al., 2012).In our study, SOC mineralization experiments at the same moisture and temperature indicated that after reclamation, the RMC content and density decreased and those in soybean field were lower than those in paddy field.The ratio of RMC to SOC in soybean field was also lower.Some researchers have suggested that MBC/MBN reflects the ratio of the amount of fungus to that of bacteria, such that higher MBC/MBN reflects much more fungus (Fauci and Dick, 1994;Huang et al., 2013).According to our investigation in 2015, MBC/MBN decreased due to reclamation, and MBC/MBN was lower in soybean field than in paddy field.That indicated that reclamation changed microbial structure and microbial quantity, and the quality also decreased (Huo et al., 2015).In addition, C : N of the underground part of the residue was ordered from high to low as follows: paddy field > wetland > soybean field; C : N of the aerial part was: soybean field > paddy field > wetland.These were all important indicators for RMC change.

Effect of reclamation on SOC in aggregates
After reclamation, > 1 mm macro-aggregates content (aggregates content = weight of aggregates / weight of total soil × 100%) decreased.However, the 0.250-1.000mm macro-aggregates, and the 0.053-0.250mm and < 0.053 mm micro-aggregates increased.The same result was found in the case of conversion from grassland or forest to arable land (Xu et al., 2017).That study showed that the < 0.25 mm micro-aggregates contents in 0-5, 5-10, 10-20 and 20-30 cm soil layers in soybean and paddy field were higher than those in the wetland by 32.70% and 22.52%, 20.12% and 8.92%, 17.63% and 14.28%, and 25.08% and 11.95%, respectively.And that in soybean field was higher than that in paddy field.The SOC contents in macro-aggregates and micro-aggregates decreased, and those in soybean field were higher than those in paddy field except for < 0.053 mm microaggregates in the 0-5 and 10-20 cm soil layers.The SOC density in aggregates was determined by the content of the aggregates, SOC content in aggregates and soil bulk density.Types of human activities and the relative extent of human use in different land use types might lead to different results in SOC density in aggregates.The SOC density in < 0.25 mm micro-aggregates in the 0-15 cm soil layer was 2.1 times higher in the plant cover olive oil orchard with higher litter input than in bare soil olive oil orchard with low litter input (Vicente-Vicente et al., 2017).Sheng et al. (2015) found the SOC density in 0.053-0.250mm micro-aggregates in 0-20 cm topsoil decreased by 21%, 53% and 51% after natural forest conversion to plantation, orchard and sloping tillage, respectively.However, in this study, after wetland were reclaimed to soybean and paddy fields, litter input decreased, SOC content in micro-aggregates decreased, but the SOC density in micro-aggregates increased.This was caused by an increase in the quantity of micro-aggregates (Table 2).The SOC density in micro-aggregates in soybean field was higher than that in paddy field.This was due to micro-aggregates content and the SOC content in micro-aggregates, which were both higher in the soybean field.More SOC distributed in < 0.25 mm micro-aggregates in the soybean field.In paddy field, SOC density in micro-aggregates and the ratio of SOC density in micro-aggregates to SOC density were both lower than those in soybean field.Micro-aggregates provided better protection to SOC than macro-aggregates (Six et al., 2002;Ewing et al., 2006;Nie et al., 2014).The SOC in micro-aggregates indicated stability of the SOC.Thus, soybean farming is more friendly for sustainable SOC residence in the soil than rice farming.

Conclusions
After reclamation, the SOC content decreased.The SOC content in soybean field was higher than that in paddy field.The SOC densities in the 0-30 cm soil layer in wetland, soybean and paddy field were 15 979.07 g/m 2 , 12 738.51g/m 2 and 11 335.53 g/m 2 , respectively, such that the SOC densities in soybean and paddy fields were lower than that in wetland by 20.28% and 29.06%, and higher in soybean field than in paddy field.DOC, MBC, ROC and RMC contents and densities decreased, and those in soybean field were lower than those in paddy field.The ratios of the labile SOC fractions to SOC in soybean field were also lower than those in paddy field.Compared with wetland, the SOC density in < 0.25 mm micro-aggregates and the ratio of SOC density in micro-aggregates (< 0.25 mm) to SOC density were higher in farmland, and that in soybean field was higher than that in paddy field.Due to reclamation, the total SOC density and labile fractions density decreased.After reclamation, most of the SOC was stored in a more complex and stable form.Soybean farming is thus more friendly for sustainable SOC residence in the soil than rice farming.

Fig. 1
Fig. 1 Location of study area in Sanjiang Plain, Northeast China

Fig. 2
Fig. 2 SOC contents in different soil layers in wetland, soybean and paddy fields.Bar: standard error, n = 5 Fig. 3 DOC content (a) and ratio of DOC to SOC (b) in different soil layers in wetland, soybean and paddy field.Bar: standard error, n = 3

Fig. 4 3 Fig. 5
Fig. 4 MBC content (a) and ratio of MBC to SOC (b) in different soil layers in wetland, soybean and paddy fields.Bar: standard error, n = 3

Fig. 6
Fig. 6 RMC content (a) and ratio of RMC to SOC (b) in different soil layers in wetland, soybean and paddy fields.Bar: standard error, n = 3

Table 1
Density of SOC in wetland, soybean and paddy fields

Table 2
SOC density in micro-aggregates and ratio to SOC density in soils of wetland, and soybean and paddy fields (n = 3) Soil layer (cm) Sample plot SOC density in micro-aggregates (< 0.25 mm) (g/m 2 )Ratio of SOC density in micro-aggregates (< 0.25 mm) to that in soil (%)