Effects of environmental variation on stable isotope abundances during typical seasonal floodplain dry season litter decomposition

Science of the Total Environment 630 (2018) 1205–1215

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Effects of environmental variation on stable isotope abundances during typical seasonal floodplain dry season litter decomposition Guangshuai Zhang a,b, Xiubo Yu a,b,⁎, Jun Xu c, Houlang Duan a,b, Loretta Rafay d, Quanjun Zhang a,b, Ya Li a,b, Yu Liu a, Shaoxia Xia a a

Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China University of Chinese Academy of Sciences, Beijing 100049, China Donghu Experimental Station of Lake Ecosystem, State Key Lab of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, Hubei 430072, China d School of Environmental and Forest Sciences, University of Washington, Seattle, Washington 98195, USA b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Groundwater levels controlled changes in soil variables and lignin decay rate. • Changes in stable isotope abundances were significantly correlated with lignin decay rates. • Soil variables affected changes in stable isotope abundances during decomposition. • Isotopes abundances related to an interaction between soil and lignin degradation.

a r t i c l e

i n f o

Article history: Received 11 December 2017 Received in revised form 23 February 2018 Accepted 25 February 2018 Available online xxxx Editor: Jay Gan Keywords: Isotopic variation Litter decomposition Lignin degradation Variation partitioning analysis Poyang Lake Wetland

a b s t r a c t Unique hydrological characteristics and complex topography can create wide-ranging dry season environmental heterogeneity in response to groundwater level across China's Jiangxi Province Poyang Lake wetland. Soil traits are one of several fluctuating environmental variables. To determine the effects of soil variables on stable isotope (δ13C and δ15N) abundances during decomposition, we performed a field experiment using Carex cinerascens along a groundwater level gradient (GT-L: −25 to −50 cm, GT-LM: −15 to −25 cm, GT-MH: −5 to −15 cm, GT-H: 5 to −5 cm) in a shallow lake. Twelve soil properties—including total organic carbon (TOC), nitrogen (N), pH, moisture, bulk density, clay, silt, sand, peroxidase, cellulase, microbial biomass carbon (MBC), and microbial biomass nitrogen—were measured in surface soil samples to assess soil environmental conditions. Analyses were performed to determine the effects of soil traits and lignin degradation on changes in stable isotope abundances. This study revealed that stable isotope abundances were significantly lower at high groundwater levels than at low groundwater levels. Lignin degradation was associated with a decrease in both δ13C and δ15N abundances. These two stable isotopes were positively related with soil N and bulk density, but negatively with pH and microbial quotient (MBC/TOC). Variation partitioning analysis (VPA) showed that soil variables and lignin decay rates explained 80.1% of the δ13C variation and 42.8% of the δ15N variation. Soil chemical and biological variables exhibited significant interactions with lignin decay rates, indicating they may affect stable isotope abundances via complex mechanisms. Our results indicate that the change in stable isotope abundances during decomposition may be affected directly by soil variables or indirectly through lignin degradation. Our results

⁎ Corresponding author at: Key Laboratory of Ecosystem Network Observation and Modeling, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China. E-mail address: [email protected] (X. Yu).

https://doi.org/10.1016/j.scitotenv.2018.02.298 0048-9697/© 2018 Elsevier B.V. All rights reserved.

1206

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

provide useful insight for understanding the roles of litter decomposition and soil traits in changing environmental conditions of seasonal floodplain wetlands. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Litter decomposition, a critical process for ecosystem function, is essential to nutrient cycling (Mitsch and Gosselink, 2000; Xiao et al., 2017). This process is regulated by the relationship between leaf litter biochemical quality, site environment, and decomposer community composition. Environments with drying and rewetting cycles considerably accelerate leaf litter degradation (Abbott et al., 2013; Shi and Marschner, 2014). To explore the general mechanisms of the decomposition process, especially carbon (C) fluxes, it is essential to identify environmental effects and determine the common drivers of decomposition across experimental gradients (Brovkin et al., 2011). At the regional and global scales, most studies have shown that environmental conditions, especially temperature and precipitation, can explain N51% of variation in decomposition (Xie et al., 2017). However, at a local scale, no consistent correlation has been recorded. Trinder et al. (2008) suggest that a high groundwater level significantly reduces decay rate. In contrast, Wiedermann et al. (2017) demonstrate that the decomposition process is restricted by low groundwater levels. There is increasing recognition that soil properties associated with specialized decomposer communities may be important for characterizing local scale decomposition processes (Ågren et al., 2013; Waddington et al., 2015; Charman et al., 2015). Poyang Lake is the largest freshwater lake in China, and its exposed dry season floodplain is one of the most important wetland ecosystems in the world (Finlayson et al., 2010). The hydrology, soil, and vegetation of the Poyang Lake wetlands have formed a unique ecosystem with complex environmental mechanisms and processes. In the dry season when water levels fall to their lowest, large quantities of isolated subterranean lakes and mudflats emerge, providing diverse habitat and abundant food resources crucial to biodiversity. Because most plants wither up and begin to decompose just as subterranean lakes become isolated from the main lake (Nie et al., 2016), substantial nutrients (i.e. N, P, and organics) in plant litter are released into the water, resulting in a relatively high nutrient concentration. These changes affect resource utilization in the subterranean lakes (Zhang et al., 2017a, 2017b). Due to these unusual conditions, it is important to study leaf litter decomposition and its relationship to the surrounding environment during the dry season. Isotopic approaches are increasingly used to study matter and energy fluxes in ecosystems. Decomposer discrimination between δ13C and δ15N during decomposition, as well as the relationship of δ13C and δ15N abundances with rates of litter mass loss, provide an excellent intrinsic tracer by which to characterize the mechanism of litter decomposition (Xu et al., 2011; Ngao and Cotrufo, 2014). Fractionation of stable isotopes is affected not only by the chemical composition of plant litter, but also by variability in environmental factors (Schmidt et al., 2011). Ember et al. (1987) found decreasing δ13C values during 18 months of leaf litter decomposition in anoxic conditions, but no changes under oxic conditions. In another decomposition experiment, Lehmann et al. (2002) found 15N enrichment under oxic conditions, with depletion under anoxic conditions. Several studies linked this variation of isotopic abundances to either the microbial transformation of lignin and non-lignin C, or the incorporation of C originating in microbial biomass into the transformed litter (Osono et al., 2008; Piao et al., 2006; Hobbie and Högberg, 2012; Gautam et al., 2016). Fernandez et al. (2003) attributed the changes in δ13C abundance to the ability of microorganisms to selectively discriminate against 13C. Ngao and Cotrufo (2014) detected that lignin controlled changes in C isotope abundance during decomposition. Bacterial dominance in soil

microorganism communities can reduce 15N signatures in residual plant litter because bacteria have a greater potential for immobilizing nitrate depleted of 15N (Bragazza and Iacumin, 2010). Groundwater level gradients are a comprehensive expression of fluctuations in water availability, thereby representing environmental changes driven by the hydrological regime (Kamiri et al., 2013; Miao et al., 2014). Groundwater lever significantly affects soil pH, nutrients, and texture, which in turn regulate composition of the microbial community (Battle and Golladay, 2001). Thus, microbial communities associated with different environmental conditions could use different metabolic pathways to decompose the same organic compound or to assimilate inorganic matter from their environment. This could result in different isotope fractionations. For example, soil with higher water content is correlated with greater abundances of 13C (Wang et al., 2015). The influences of groundwater level and its concomitant environmental factors on decomposition have been thoroughly assessed in other research. However, few studies have focused on whether these factors influence isotope variation in a direct or indirect way during the decomposition of leaf litter. The current study was undertaken to evaluate the effects of groundwater level and its related environmental factors on variations in carbon and nitrogen isotope abundances during dry season Poyang Lake wetland leaf litter decomposition. The objectives of this investigation were: (1) to reveal the relationships between isotopic compositions and decay rates during early decomposition stages, and (2) to determine the extent to which soil environmental factors influence changes in δ13C and δ15N abundances. We hypothesized that the variation of δ13C and δ15N abundances in early periods of litter decomposition can be explained by soil environmental variables depending on groundwater level. This study sheds light on the response of litter decomposition to changing environmental conditions along groundwater level gradients and provides helpful information for management of wetland meadows during the dry season. 2. Materials and methods 2.1. Study area The Poyang Lake Wetland (28°22′–29°45′N, 115°47′–116°45′E) is located in the middle of the Yangtze River basin in northern Jiangxi Province, China. The lake catchment is in a subtropical wet climate zone with an annual mean precipitation of 1680 mm that mainly falls between April and June, and an annual mean temperature of 17.5 °C. The bottomlands are almost submerged during wet season hydrology. However, after October, these submerged areas gradually become exposed with the onset of the dry season. Because of the fluctuating water levels, the plant communities typically form a ringed pattern along elevation gradients. The dominant plant species in the Poyang Lake Wetland are Carex cinerascens (C. cinerascens), Phragmites australis and Triarrhena sacchariflora (Wang et al., 2011; Wang et al., 2014). For this study, Baisha Lake, one of the shallow lakes in the Poyang Lake Wetland, was selected as the site for the decomposition simulation field experiment. Our research focused on leaf litter of the dominant graminoid C. cinerascens in four northwest lakeshore zones spanning the complete groundwater level gradient across a total area of 200 m × 300 m (Fig. 1). We divided the study area into four zones according to the following groundwater levels: −25 to −50 cm (GT-L), −15 to −25 cm (GT-LM), −5 to −15 cm (GT-MH) and 5 to −5 cm (GT-H). These groundwater level zones exhibited significant differences in soil physical, chemical, and biological properties (Zhang et al., 2018). The groundwater level in all four zones was relatively stable during the

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

October through March dry season. During this experimental period, soil surfaces in the GT-L, GT-LM, and GT-MH zones were higher than water level and remained in a dry state. However, soil surfaces in the GT-H zone remained submerged until mid-November, at which point it experienced cyclical drying and wetting in response to precipitation. Because soil environmental variability in the GT-L zone is much smaller than that in the other zones, only three subplots were established in this zone. Five subplots were established in the GT-LM zone, and six subplots were established in both the GT-MH and GT-H zones. 2.2. Decomposition assay C. cinerascens leaf litter was collected from a high elevation meadow in the Poyang Lake wetland. To avoid confounding influences on leaf litter decomposition processes, leaf litter was cut into 1-cm-long sections and pretreated by washing with deionized water to remove surface impurities. The litter was then oven dried at 85 °C for 72 h. The dried plant litter was stored in a dryer until used. We measured total C, total N, and isotope abundance in ten samples from the mixed leaf litter for initial component values. The components of the C. cinerascens litter (C, N, δ13C, and δ15N content) are presented in Appendix A. The decomposition potential in plots along the groundwater level gradients was determined using a litterbag experiment, as described by Rejmánková and Houdková (2006). 5 g of air-dried C. cinerascens leaves were placed in each litterbag. The litterbags were made of Nytex 80 μm mesh, which minimized root in-growth and the influence of invertebrates. The bags were placed on the lakeshore on October 15, 2016 and were loosely attached to stakes with nylon cord to maintain their position near the sediment surface. We used a PVC tube with an area of 0.15 m2 and a length of 1 m to house the leaf litter bags for a relatively stable environment. The PVC tubes were inserted into 0.5 m underground. Three replicate bags were collected from each plot on days 15, 30, 60, 90, 120, and 150. The experiment was terminated on April 15, 2017 when the increased water level began to submerge the bags. A PVC tube (diameter, 2.5 cm) was inserted beside the experimental plots to a depth of about 0.7 m to measure the height of the groundwater level. Groundwater measurement followed the methods of Li et al. (2011) and Zhang et al. (2012), using a steel rule and plumb line. At the time of sample collection, all litter was removed from the litterbags and a mixed subsample from all three replicates was dried, weighed, and ground prior to further analysis. Litter C and N were determined using an element analyzer (Elementar, Vario Max CN; Hanau,

1207

Germany). Lignin content was determined by sequential analysis of lignin fiber. About 1 g of sample biomass was extracted with a neutral detergent, followed by application of an acid detergent and 72% sulfuric acid, using the methods of Loranger et al. (2002). 2.3. Soil samples and measurement of physical, chemical, and microbiological properties Soil samples were collected on October 10, 2016; December 15, 2016; and February 15, 2017 using a 20 mm diameter stainless steel soil augur. Samples were collected to a depth of 20 cm from three subplots at each experimental site to analyze site micro-ecological environment. For each sublot, soil samples were mixed into one composite sample and stored at 4 °C in a polyethylene sealed bag to await analysis. After carefully removing fine roots and other organic detritus, each soil sample was divided into two parts. One part was stored at 4 °C for microbial assay, and the other part was air-dried for analysis of soil chemical properties. Soil total organic C (TOC) and total N (TN) were determined using an element analyzer (Elementar, Vario Max CN) (Wang et al., 2014). Soil moisture content was determined by weighing soil before and after oven drying at 105 °C for 72 h, following the methods of Wang et al. (2014). Soil particle fractions were analyzed with a LongbenchMastersizer 2000 instrument (Malvern Instruments, Malvern, England) using deionized water as the solvent and sodium metaphosphate (SMP) as a dispersing agent. Soil pH was determined using a pH meter (Sartorius, Germany). Soil cellulase activity was measured using nitrosalicylic acid colorimetry (Guan, 1986). The amount of glucose released over 72 h was assayed colorimetrically at 540 nm and expressed as mg glucose g−1 dry soil. Soil peroxidase was measured following the methods of Tang et al. (2005), based on the oxidation of veratryl alcohol to veratraldehyde monitored at 310 nm. Microbial biomass C (MBC) was measured using a TOC analyzer (TOC-VCPH/CPNTNM-1, Shimadzu) and microbial biomass N (MBN) was measured using the salicylic method, after a pretreatment with the fumigation extraction method (Joergensen and Mueller, 1996; Baaru et al., 2007; Spohn et al., 2016). 2.4. Isotopic analysis The δ13C and δ15N abundances of the plant litter residue were determined using a mass spectrometer (Finnegan MAT 253, Thermo

Fig. 1. Location of sampling sites in Poyang Lake Wetland, China. The dotted line boxes indicate the approximate areas used for the decomposition experiment in this study. Photographs present the landscape of the experimental area (Photo 1 shows a low groundwater level site, Photo 2 shows a high groundwater level site) for decomposition.

1208

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

Scientific, USA) coupled with an Elemental Analyzer (Flash EA 1112, Thermo Scientific, USA). The C and N isotope ratios are expressed as δ13C and δ15N (units: per thousand ‰), respectively, according to the following equation:  δð‰Þ ¼

 Rsample −1  1000 Rstandard

where Rsample and Rstandard refer to either 13C:12C or 15N:14N abundance ratios of sample to standard material, respectively. The reference standards were Vienna Pee Dee Belemnite (VPDB) for C isotopes and atmospheric N2 for N isotopes. The standard deviation of replicates was ± 0.1‰ for δ13C and ±0.4‰ for δ15N. 2.5. Data analysis The differences in soil physical and chemical properties, microbial activities, MBC, and MBN — as well as proportions of initial dry leaf biomass, leaf C, and leaf N— across the groundwater level gradients were analyzed using one-way analysis of variance combined with Dunnett's T3 post-hoc pairwise comparisons (α = 0.05). Changes in decomposing litter C and N isotope abundances in relation to C concentration and decay rates were estimated using a linear regression. The effects of groundwater level gradients and decomposition period on the variation of δ13C and δ15N abundances were evaluated using a general linear model across the 150-day experiment. The decay rate of dry litter mass and lignin (k) was calculated using the following simple exponential model (Olson, 1963; Zhang et al., 2008): Mt ¼ α e−kt ; where Mt (%) is the amount of dry litter mass remaining at time t, t is the decomposition time (days), α is a modified coefficient, and k is the decay constant. The length of time needed to decompose 50% (t0.5) and 95% (t0.95) of the initial dry litter mass and lignin were calculated by the following equations, respectively (Tripathi et al., 2013): t 0:5 ¼ − ln ð0:5=0:9362Þ=k t 0:95 ¼ − ln ð0:05=0:9362Þ=k Principal component analysis (PCA) was conducted to identity the main components by describing their variability in the experimental plots and objectively separating the environmental gradients based on soil physical, chemical, and microbiological variables. Data of eight physical and chemical variables, as well as six microbiological variables, were collected and the correlation matrix was generated to acquire the loading values for the first two components. The correlations between each soil variable and the principle component were indicated by Pearson's r and P-value. One previous study (Rejmánková and Houdková, 2006) has reported that the effects of microbes on leaf litter decomposition in wetland conditions increased rapidly in the first 30–90 days, then remained stable for the next 200 days. On the basis of those results, we decided to use the decay rates and isotopic values recorded during the first 60 days to characterize the decomposition patterns along the groundwater level gradients. Correlation coefficients between isotopic signatures with decay rates of lignin and select environmental parameters were calculated using a linear regression equation. Variation partitioning analysis (VPA) was used to test significance of the effects of soil variables and lignin degradation on isotopic signature (Burgess et al., 2001) To obtain the most parsimonious model and minimize redundancy of explanatory variables, we used the first principle component to represent soil chemical and microbial variables, then used the second principle component to represent soil physical variables,

according to the loading values. For PCA and VPA analyses, we mainly concentrated on the period with the most significant differences (P b 0.01) in decay rates under different soil environmental conditions. The mean value of soil variables and isotopic signatures during this period (0–90 days) were used to illustrate the variation of soil environmental conditions and stable isotopes along the groundwater level gradients. All statistical analyses were performed with R v.3.4.0 (R Development Core Team, 2013). Visualizations were performed using Origin Pro 8 (Origin Lab, US). Maps of the spatial distribution of δ13C and δ15N abundance values were drawn using ArcMap 10.1 based on the kriging interpolation method.

3. Results 3.1. Litter decomposition along groundwater level gradients Litter examined in this study did not show any significant differences in dry mass and total remnant C (Table 1, P N 0.05) among groundwater level gradients. However, C:N remnants in leaf litter of the GT-H zone was significantly lower than that of other groundwater levels after 60 days. Lignin remnants in the GT-H zone litter were the lowest of all zones after 15 days (Table 1, P b 0.05). After 30 and 60 days of decomposition, the proportion of lignin remaining in the GT-H zone was 27.21% and 33.36% lower, respectively, than in the GTL zone (Table 1, P b 0.05). Dry litter mass, total C, lignin, and C: N values declined exponentially for all groundwater level gradients during litter decomposition (Fig. 2, R2 N 0.2, P b 0.01). Decay rates of lignin and C: N in the GT-H zone showed the highest values. According to the exponential models (Fig. 2), the time to decompose 50% (t0.5) and 95% (t0.05) of the initial litter mass was estimated to be about 184 and 856 days, respectively, at all four groundwater level gradients. The time needed to decompose half (t0.5) of the initial lignin was estimated to be about 628, 314, 314, and 79 days for the GT-L, GT-LM, GT-MH, and GT-H zones, respectively. The time taken to decompose 95% (t0.05) of the initial lignin was 2930, 1465, 1465 and 367 days for the GT-L, GT-LM, GT-MH, and GT-H zones, respectively. Spatial variation of lignin decay rates along groundwater level gradients is shown in Appendix B-a.

3.2. δ13C and δ15N dynamics during litter decomposition Groundwater level and decomposition period significantly affected δ13C abundances in residual leaf litter (P b 0.001). However, δ15N abundances were only significantly influenced by groundwater level (P b 0.001). The decline of groundwater level was associated with a decrease in both δ13C (Fig. 3a, Appendix B-c) and δ15N (Fig. 3b, Appendix B-d) abundances. δ13C abundances revealed a general trend of fluctuations as leaf litter decomposed. (Fig. 3a; P b 0.001). The variation coefficients of δ13C abundances with decomposition periods decreased in the following order: GT-H (Cv = 0.91 ± 0.29‰) N GT-MH (Cv = 0.57 ± 0.09‰) N GT-LM (Cv = 0.33 ± 0.07‰) N GT-L (Cv = 0.18 ± 0.04‰). δ13C abundances were significantly and positively related with leaf litter C concentration during decomposition (Fig. 3c; R2 = 0.53, P b 0.001). The δ15N abundances behaved differently in two phases, showing depletion (0–60 days) followed by substantial enrichment (60–150 days) (Fig. 3b). However, the relationship between δ15N and N concentration was not significant (Fig. 3d; P N 0.05). We investigated whether the isotopic abundances were related to lignin decay rates. Significance of the relationships between lignin decay rates and δ13C or δ15N abundances increased in the first 90 days and then began to decline (Fig. 4). Both δ13C and δ15N abundances were significantly and negatively correlated with lignin decay rates from 60 to 90 days (R2 = 0.60–0.70, P b 0.001 for δ13C and R2 = 0.22–0.43 P b 0.05 for δ15N).

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

1209

Table 1 Statistical values for decomposition dynamics of dry mass, total C, C: N and lignin among different groundwater level gradients. Values shown are mean ± standard deviation. Different lowercase letters indicate significant differences (Dunnett's T3 post-hoc pairwise comparisons) among the four water table levels from GT-L to GT-H. Different capital letters indicate significant differences among the six decomposition periods (**P b 0.01; *0.01 b P b 0.05; ns = not significant at P ≥ 0.05). Variables

Gradients

15 days

30 days

60 days

90 days

120 days

150 days

F value

Dry mass (%)

GT-L GT-LM GT-MH GT-H F value GT-L GT-LM GT-MH GT-H F value GT-L GT-LM GT-MH GT-H F value GT-L GT-LM GT-MH GT-H F value

81.54 ± 5.87Aa 77.33 ± 7.22Aa 75.40 ± 5.49Aa 82.85 ± 4.27Aa 2.01 ns 81.55 ± 5.00Aa 77.19 ± 8.71Aa 75.62 ± 5.45Aa 76.42 ± 9.73Aa 0.41 ns 19.78 ± 1.50Aa 20.02 ± 2.89Aa 19.12 ± 1.41Aa 18.88 ± 1.37Aa 0.43 ns 82.67 ± 8.05Ac 85.48 ± 0.96Abc 91.28 ± 3.02Ab 94.18 ± 8.11Aa 12.15***

68.39 ± 0.90Ba 69.90 ± 6.28ABa 75.03 ± 2.25Aa 76.33 ± 3.92Aa 1.03 ns 72.39 ± 1.84Ba 72.41 ± 8.30ABa 77.58 ± 2.54Aa 72.13 ± 6.83Aa 1.15 ns 16.19 ± 0.68Ba 16.76 ± 1.88Ba 17.19 ± 0.82Ba 17.07 ± 0.61Ba 0.60 ns 80.38 ± 12.89Aa 79.00 ± 16.33Aa 64.34 ± 11.09Bb 58.51 ± 7.00Bc 3.96*

61.02 ± 4.82BCa 67.38 ± 5.89ABa 70.55 ± 8.58Aa 69.59 ± 8.21Ba 1.21 ns 61.91 ± 7.21Ca 66.18 ± 2.83ABa 66.02 ± 4.94Ba 63.66 ± 9.90Ba 1.27 ns 15.95 ± 2.21Ba 15.00 ± 1.20Ba 15.36 ± 1.34Ca 15.37 ± 1.11Ca 0.30 ns 65.04 ± 10.65Ba 69.31 ± 7.49Ba 62.46 ± 17.38Bb 43.35 ± 2.97Cc 35.40***

65.80 ± 2.99BCa 64.12 ± 5.62ABa 69.26 ± 5.51Aa 68.83 ± 6.69Ba 0.97 ns 68.64 ± 3.31BCa 65.03 ± 6.04ABa 66.79 ± 6.76Ba 56.59 ± 9.92Ba 0.45 ns 15.96 ± 0.99Ba 15.67 ± 1.06Ba 15.24 ± 0.94Cb 14.14 ± 1.16Cc 4.21* 66.99 ± 3.76Ba 71.51 ± 2.22Ba 62.37 ± 11.86Bb 43.62 ± 3.51Cc 37.96***

57.97 ± 3.79Ca 62.90 ± 6.47Ba 62.28 ± 5.70BCa 62.25 ± 5.19BCa 0.57 ns 60.19 ± 5.40Ca 64.94 ± 6.35ABa 61.98 ± 5.84BCa 63.11 ± 5.88ABa 2.61 ns 16.85 ± 1.35Ba 16.41 ± 1.86Bab 15.87 ± 1.08Cb 13.81 ± 1.05Cc 4.65* 66.23 ± 6.17Ba 63.44 ± 6.93Bab 62.98 ± 5.92Bb 43.73 ± 2.35Cc 38.20***

57.30 ± 0.61Ca 58.67 ± 7.60Ba 56.18 ± 4.56Ba 57.97 ± 2.79Ca 0.27 ns 60.01 ± 0.77Ca 60.00 ± 7.42Ba 57.64 ± 6.41Ca 59.95 ± 3.65Ba 0.25 ns 16.67 ± 1.14Ba 16.18 ± 3.46Bab 15.57 ± 1.31Cb 14.31 ± 0.77Cc 4.56* 66.45 ± 2.97Ba 66.67 ± 5.14Ba 62.46 ± 6.01Ba 42.86 ± 2.35Cb 65.24***

17.94*** 4.88** 10.65*** 17.57***

Total C (%)

C: N

Lignin (%)

10.81*** 3.97** 11.83*** 6.76*** 4.69* 3.87* 15.00*** 12.74*** 4.79* 3.46* 14.14*** 167.70***

Fig. 2. Dynamics of initial dry mass (a), total C (b), C:N (c), and lignin (d) remaining in decomposing litter. Solid lines are trend lines that best describe the relationships between the measured decomposition parameters and time with exponential equations of y = 0.876e−0.004t (R2 = 0.681, P b 0.001, F = 746.884), y = 0.866e−0.003t (R2 = 0.588, P b 0.001, F = 1105.323), y = 0.883e−0.003t (R2 = 0.672, P b 0.001, F = 1795.286), y = 0.915e−0.003t (R2 = 0.777, P b 0.001, F = 2676.760) for the relationship between remaining dry mass and decomposition time for GT-L, GT-LM, GT-MH, and GT-H respectively; and y = 0.881e-0.003t (R2 = 0.658, P b 0.001, F = 805.515), y = 0.867e−0.003t (R2 = 0.557, P b 0.001, F = 1089.076), y = 0.882e−0.003t (R2 = 0.675, P b 0.001, F = 1752.019), y = 0.848e−0.003t (R2 = 0.452, P b 0.001, F = 688.140) for the relationship between total C remaining and decomposition time; and y = 18.435e−0.001t (R2 = 0.195, P b 0.001, F = 881.288), y = 18.748e−0.002t (R2 = 0.268, P b 0.001, F = 1020.525), y = 18.770e−0.002t (R2 = 0.618, P b 0.001, F = 3926.585), y = 18.727e−0.002t (R2 = 0.638, P b 0.001, F = 3572.684) for the relationship between C:N ratio and decomposition time; and y = 0.885e−0.001t (R2 = 0.261, P b 0.001, F = 1006.092), y = 0.887e−0.002t (R2 = 0.273, P b 0.001, F = 1115.806), y = 0.847e−0.002t (R2 = 0.202, P b 0.001, F = 601.388), y = 0.932e−0.008t (R2 = 0.711, P b 0.001, F = 507.537) for the relationship between lignin remaining and decomposition time.

1210

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

Fig. 3. Temporal dynamics of leaf litter (a) δ13C and (b) δ15N abundances and their relationships with (c) carbon and (d) nitrogen concentrations under different groundwater levels.

3.3. Effects of soil environmental factors on δ13C and δ15N abundances All soil physicochemical properties showed significant differences across groundwater level gradients (Appendix C). Compared to sites with lower groundwater levels, the highest groundwater level (GT-H) zone was characterized by significantly lower TOC, TN, silt, and clay contents, as well as significantly higher pH, moisture, and sand content (P b 0.05). In terms of soil microbial variables, the GT-H zone was characterized by significantly higher activity of cellulase, higher values of microbial biomass (MBC and MBN) (P b 0.01), and higher turnover rates for both organic C (MBC/TOC) and N (MBN/TN) (P b 0.01). The PCA indicated that eight soil physical and chemical variables, as well as six microbiological variables, could be synthesized into two PCs, which together accounted for 97.70% of the variation. The first PC accounted for 78.56% of the variation, with roughly equal loadings for soil TN, pH for soil physicochemical indicators, and MBC/TOC for soil microbial indicators (r N 0.90; Table 2). The spatial variation of PC1 along groundwater levels is shown in Appendix B-b. The second PC accounted for 19.14% of the variation, with soil bulk density as the most highly weighted variable (r = 0.57). Indicators with high loadings of PC1 and PC2 were selected to indicate overall soil properties. The δ13C indicator was significantly and positively correlated with soil TN (Fig. 5; R2 = 0.62, P b 0.001, F = 31.96) and bulk density (R2 = 0.62, P b 0.001, F = 31.36), but was negatively correlated with soil pH (R2 = 0.68, P b 0.00, F = 41.12) and MBC/TOC (R2 = 0.60, P b 0.00, F = 29.50). The δ15N indicator was significantly and positively correlated with soil TN (R2 = 0.33, P = 0.01, F = 10.29) and bulk density (R2 = 0.25, P = 0.02, F = 7.20), but was

negatively correlated with soil pH (R2 = 0.41, P b 0.01, F = 14.27) and MBC/TOC (R2 = 0.37, P b 0.01, F = 12.07). We used Variation partitioning analysis (VPA) to test effects of soil variables and lignin decay rates on variation in decomposing leaf litter δ13C and δ15 abundances (Fig. 6). All selected variables combined explained 80.10% of the variance in δ13C abundances and 42.8% of the variance in δ15N abundances. PC1, which was used to indicate soil chemical and microbial properties, played a dominant role in δ13C changes (with a total contribution of 77.2%, P b 0.05). The interaction soil chemical and microbial properties with lignin degradation significantly affected both δ13C and δ15N abundances with contributions of 72.60% and 46.90%, respectively (P b 0.01). 4. Discussion 4.1. Decomposition dynamics The decomposition process in wetland ecosystems is remarkably different from that observed in terrestrial ecosystems (Li et al., 2008). Decomposition of bulk litter in a subtropical plantation located in the same region as the current study reported that t0.5 for dry mass ranged from 11.4 to 26.3 months while t0.95 ranged from 53.9 to 114.9 months (Li et al., 2008; Xu et al., 2011). However, our estimates (t0.5 = about 7 months and t0.95 = about 29 months) showed a faster decay rate in the wetland ecosystem than decomposition in a subtropical forest (Li et al., 2008; Xu et al., 2011). During the first 60 days, 31.96% of total residual leaf litter dry mass was degraded (Table 1). However, Xu et al. (2011) found that leaf litter dry mass in a subtropical plantation only decreased by 12.97%. This suggests that degradation occurs faster, and

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

1211

Fig. 4. Relationships between isotopic signatures and lignin decay rate during different decomposition periods (n = 20). a: relationships between δ13C with lignin decay rates in the first 30 and 60 days; b: relationships between δ13C with lignin decay rates in the 90 and 120 days; c: relationships between δ15N with lignin decay rates in the first 30 and 60 days; d: relationships between δ15N with lignin decay rates in the 90 and 120 days. Circles and triangles represent decomposition in the first 30 and 60 days, respectively, and diamonds and squares represent decomposition in the first 90 and 120 days, respectively. Solid lines represent the linear fit between isotopic signatures and lignin decay rates with equations of y = 0.009x − 29.81 (R2 = 0.14, P = 0.05, F = 4.09), y = 0.046x − 29.68 (R2 = 0.60, P b 0.001, F = 28.86), y = 0.038x − 29.60 (R2 = 0.70, P b 0.001, F = 45.77), y = 0.030x − 30.02 (R2 = 0.02, P = 0.42, F = 0.68) for the relationship between δ13C and lignin decay rates during the first 30, 60, 90 and 120 days respectively; and y = 0.002x + 5.34 (R2 = 0.05, P = 0.95, F = 0.004), y = 0.098x + 6.03 (R2 = 0.22, P = 0.02, F = 6.40), y = 0.18x + 6.52 (R2 = 0.43, P b 0.01, F = 15.36), y = 0.15x + 6.09 (R2 = 0.10, P = 0.10, F = 3.00) for the relationship between δ15N and lignin decay rates during the first 30, 60, 90 and 120 days respectively.

to a greater extent, in this wetland ecosystem than in a terrestrial ecosystem of the same climatic zone (Dignac et al., 2010; Berg and McClaugherty, 2014). However, the mechanism underlying the more

Table 2 Correlation coefficients between soil indicators and principal component scores. Factor loadings in bold are highly weighted. Variables

Soil physicochemical indicators

Soil microbial indicators

PC1 (78.56%)

TOC TN pH Moisture Bulk density Clay Silt Sand Peroxidase Cellulase MBC MBN MBC/TOC MBN/TN

PC2 (19.14%)

Pearson's r

P

Pearson's r

P

−0.86 −0.90 0.93 0.83 −0.73

b0.001 b0.001 b0.001 b0.001 b0.001

−0.33 −0.19 −0.14 −0.18 0.57

0.15 0.42 0.57 0.46 0.01

−0.80 −0.78 0.77 0.52 0.82 0.84 0.84 0.93 0.91

b0.001 b0.001 b0.001 0.03 b0.001 b0.001 b0.001 b0.001 b0.001

0.05 0.46 −0.25 −0.33 −0.40 0.25 0.40 0.31 0.33

0.83 0.04 0.28 0.15 0.08 0.30 0.08 0.18 0.16

rapid litter decomposition process of more hydrologic conditions is not clear. It has been speculated that the physical environment and litter chemical properties may favor litter decomposition under wetland ecosystems. Both oxygen and moisture availabilities for litter decomposition might be optimized in wetlands under periodic saturation (Bedford, 2005). Lower soil water content in upland sites might limit litter moisture availability during decomposition (Yoon et al., 2014). In addition, the Ncontent and C/N ratio of initial leaf litter may in part determine early-stage litter decomposition rates (Zhang et al., 2008). Lignin has traditionally been considered a recalcitrant compound that prevents biotic breakdown of organic matter (Williams and Yavitt, 2003). However, recent studies of soil systems indicate that lignin is considerably more bioavailable and decomposable than was previously thought. For example, Klotzbücher et al. (2011) reported that strong lignin degradation occurred in the first 41 days of litter decomposition in a forest river. In this study, lignin easily decomposed during the first 60 days, but then its decomposition rate decreased significantly, with an average overall decomposition rate of 39.96%. Compared to the lowest groundwater level zone, a high groundwater level in an alternating dry–wet environment promoted lignin degradation, indicating that higher groundwater level and intermittent submergence accelerate litter decomposition through stimulated microbial metabolism and greater leaching losses (Battle and Golladay, 2001). The high

1212

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

Fig. 5. Relationships between isotopic signatures (δ13C, δ15N) and soil environmental factors (TN, pH, bulk density, and MBC/TOC) (n = 20). Solid lines are trend lines that best describe the relationships between δ13C and soil indicators, and dashed lines are trend lines that best describe the relationships between δ15N and soil indicators.

groundwater level area with cyclic drying and wetting patterns provided an ideal environment for maximum degradation; that is, suitable microbial community, soil pH, and soil texture for promoting bioavailability and turnover of organic C (Foulquier et al., 2013). In addition, our results also show that C and N turnover rates indicated

by MBC/TOC and MBN/TN ratios were enhanced at a relatively high groundwater level (Appendix C). This is because sites with higher groundwater level have adequate soil water and a moderately acid environment, which may be more suitable for microbial growth and metabolism (Mentzer et al., 2006).

Fig. 6. Variation partitioning analysis of isotopic discrimination explained by soil environmental variables, lignin decay rates, and their interactions. PC1: The first principal component of soil variables with high loadings of TN, pH, and MBC/TOC; PC2: The second principal component of soil variables with high loadings of bulk density and silt content. Values presented are percentages explained by the selected factors. (**: P b 0.01; *: P b 0.05).

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

4.2. Isotopic dynamics during decomposition As decomposition proceeded, δ13C abundances fluctuated significantly, indicating alternations of enrichment and depletion (Fig. 3a). Compared with initial abundances, δ13C abundances declined rapidly in the first 15 days. Residual litter maintained a lower δ13C abundance as groundwater level increased (Fig. 3a) These isotopic shifts may be related to the loss of soluble C fractions depleted in 13C, or to microbial C mixing (in labile and recalcitrant carbohydrates) with residual litter during transformation (Connin et al., 2001). Several studies have reported δ13C enrichment (Gautam et al., 2016), depletion (Osono et al., 2008), or even little difference at all (Ngao and Cotrufo, 2014) in residual litter by tracking changes in δ13C signatures arising from C isotopic fractionation during decomposition. Connin et al. (2001) suggested that isotopic discrimination is characteristic of early litter decomposition stages. We also found a significant positive relationship between δ13C signatures of decomposing litter and the remaining C concentration, a result that may be related to different immobilization– mineralization phases (Fig. 3c). This supported previous findings that increasing δ13C values during decomposition are based on microbial transformation as well as mixing of external C (Gautam et al., 2016). The lignin fraction was suggested to be a controlling factor of C isotopic dynamics in both anaerobic environments (Osono et al., 2008; Ngao and Cotrufo, 2014) and in aerobic environments (Benner et al., 1987). In the present study, the correlation between δ13C abundance and lignin decay rates provides further evidence for lignin control on C isotope dynamics in wetland plants (Fig. 4). Dynamics of δ15N patterns were inconsistent with δ13C patterns in our study (Fig. 3b). We found a pattern of 15N depletion during the initial 2 months of litter decomposition, and 15N enrichment from the third to fifth months. Xu et al. (2011) found a similar non-linear 15N pattern in a terrestrial ecosystem, with 15N depletion during the initial 7 months, and 15N enrichment from the eighth to twentieth months. One possible explanation for 15N depletion in the early decomposition period is that decomposers need additional N to attack litter with high C: N ratios (Hodge et al., 2000). Incorporation of exogenous N via soil microbial metabolism maybe leads to 15N depletion in remaining litter (Nadelhoffer and Fry, 1988). In the subsequent stage, microbial discrimination could perform a dominant role and thus lead to 15N enrichment. This indicates that microbial metabolic rates in wetland ecosystems are more rapid than those in terrestrial ecosystems. Our linear regression analyses demonstrate no significant correlations between δ15N and N concentrations remaining in decomposing litter (Fig. 3d). This suggests that factors other than N content were also related to δ15N discrimination (Bragazza and Iacumin, 2010). An indirect effect of N concentration through the processes of nitrification, immobilization, and mineralization appears more plausible (Connin et al., 2001). Several studies reported that 15N changes during decomposition could be attributed to variation in the dominant microbial community (Hobbie and Hobbie, 2008; Bragazza and Iacumin, 2010). Fellerhoff et al. (2003) showed that the preferential uptake of 14NH4 causes a decrease in δ15N values, while the microbial assimilation of dissolved inorganic N enriched with 15N would cause an increase. Macko and Estep (1984) reported that some bacteria were depleted in 15N relative to the substrate they were growing on; however, some were enriched, and others showed no significant change. 4.3. Relevance of isotopic variation to soil environmental variables In our previous study, decomposer bacteria, fungi, and other microorganisms were generally found to be sensitive to soil physical and chemical properties, especially soil pH, particle distribution, bulk density, and soil C and N (Zhang et al., 2018). In the current study, the sand content increased while silt and clay content decreased along the groundwater level gradients, resulting in the decline of soil bulk density. Because the rise in groundwater levels can lead to a series of

1213

environmental changes, a decrease in soil acidity was found to increase enzymatic activities and turnover rates of organic matter. These results were consistent with earlier observations of changes in soil properties along a waterlogging gradient (Yu, 2010; Callaway et al., 2012; Lyu et al., 2015). The magnitude of isotopic variation during litter decomposition is related to biotic and environmental factors (Wang et al., 2015). In the present study, groundwater level regulated changes in stable isotopes directly through soil environmental conditions or indirectly through lignin degradation. On the one hand, in the study area, sites with a moderate pH were reported to promote turnover rates of organic C (MBC/TOC) and N (MBN/TN) (Zhang et al., 2018). Thus, a higher groundwater level provided an appropriate environment for microbial metabolism and the depletion of 13C and 15N. In addition, microbial abundance increased significantly when soil pH was near neutral, which could promote mineralization of 13C/15N-enriched carbohydrates. Therefore, both δ13C and δ15N abundances were negatively related with soil pH and MBC/TOC (Fig. 5). Wedin et al. (1995) hypothesized that variations in isotopic signatures during litter decomposition were due to mixing external C from soil organic matter with original C via microbial populations. We detected a positive relationship between both δ13C and δ15N with soil N over the first 90 days of the decomposition period (Fig. 5). Preston et al. (2009) argue that microbial biomass constitutes only a small fraction of decomposing litter, and enrichment results from site effects such as C sorption with diverse isotopic signatures. For the current study, it is difficult to distinguish whether the mineralization or the assimilation of soil microbes contributes more to changes in stable isotope abundances. On the other hand, another reason for isotopic variation is related to lignin decay rates. We measured proportions of lignin remaining over time across groundwater level gradients. The results indicate that edaphic factors changed along the groundwater level gradients, especially under alternating wet–dry conditions, in which lignin decomposed more rapidly (Scott et al., 1996; Klotzbücher et al., 2011). Generally, soil chemical and microbial variables, together with the selective degradation of internal carbohydrates, interactively affected the change of δ13C and δ15N abundances (Fig. 6). Moreover, variation in δ13C abundances along different groundwater levels can be explained by soil environmental factors significantly more than was seen with δ15N variation. Whether soil factors affected the decomposition processes directly through the incoming flow of external C, or indirectly through the outgoing flow from internal carbohydrates, the result can be indicated by the change in δ13C abundance. In our study, we found that δ13C, rather than δ15N, could better reflect changes in soil environmental factors and decomposition processes. Considering the fast and marked response of stable isotopes to the environmental factors under different groundwater levels, we suggest that δ13C abundances in residual leaf litter can be used as an indicator to reflect microbial activity and carbon turnover rates in lieu of performing long-term decomposition experiments. There exists a critical and complicated link between environmental conditions and plant litter decomposition in wetland meadows, which includes physical, chemical, and biological processes (Foulquier et al., 2013; Shi and Marschner, 2014; Xiao et al., 2017). However, these ecological processes strongly controlled the change of stable isotope abundances, especially that of δ13C. Our study reveals that changes in stable isotopes and leaf litter decay rates were driven by a cyclic drying and wetting pattern. This implies that there exists an active zone with high microbial activity and strong biochemical reactions in the wetland meadow during dry seasons which may affect the carbon cycle in seasonal flood plain wetland ecosystems. 5. Conclusion We conducted a 150d litterbag experiment in four groundwater level zones to understand the relationships between isotopic abundances in leaf litter residue and lignin decay rates, as well as soil

1214

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215

environmental factors, during Carex cinerascens dry season litter decomposition. Our results suggest that a relatively high groundwater level promotes lignin degradation and strongly affects soil physical, chemical, and biological factors (pH, TN, bulk density, and MBC/TOC). We found that isotopic abundance variation is related to either lignin degradation or soil environmental factors. δ13C abundances are more sensitive to the interaction of lignin degradation and soil factors (TN, pH, and MBC/TOC) than are δ15N abundances. However, the causal mechanism of these relationships remains speculative and requires further investigation. Both field studies and experiments under controlled conditions are needed to disentangle the effects of environmental variables and the internal degradation process on isotopic abundances. Additionally, isotopic signatures in surface soil should be observed to reveal the isotope transformation between leaf litter and soil during the decomposition. Overall, the current study lends support for optimizing groundwater level management strategies to minimize litter decomposition rates and the mineralization of carbon to prevent carbon release from the system. δ13C abundances in decomposing litter can be used to detect microenvironments which are sensitive to groundwater levels. Acknowledgements The authors would like to express their sincere thanks to the Nanji Wetland National Nature Reserve Agency for permitting us to collect samples. This work was supported by the National Natural Science Foundation of China (Grant No. 41471088). We also thank Alex Boon, PhD, for editing the English text of an early draft of this manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2018.02.298. References Abbott, G.D., Swain, E.Y., Muhammad, A.B., Allton, K., Belyea, L.R., Laing, C.G., Cowie, G.L., 2013. Effect of water-table fluctuations on the degradation of sphagnum, phenols in surficial peats. Geochim. Cosmochim. Acta 106, 177–191. Ågren, G.I., Hyvönen, R., Berglund, S.L., Hobbie, S.E., 2013. Estimating the critical n:c from litter decomposition data and its relation to soil organic matter stoichiometry. Soil Biol. Biochem. 67, 312–318. Baaru, M., Mungendi, D., Bationo, A., Verchot, L., Waceke, W., 2007. Soil microbial biomass carbon and nitrogen as influenced by organic and inorganic inputs at Kabete, Kenya. In: Bationo, A., Waswa, B., Kihara, J., Kimetu, J. (Eds.), Advances in Integrated Soil Fertility Management in Sub-Saharan Africa: Challenges and Opportunities. Springer, Dordrecht, pp. 827–832. Battle, J.M., Golladay, S.W., 2001. Hydroperiod influence on breakdown of leaf litter in cypress-gum wetlands. Am. Midl. Nat. 146, 128–145. Bedford, A.P., 2005. Decomposition of phragmites australis litter in seasonally flooded and exposed areas of a managed reedbed. Wetlands 25, 713–720. Benner, R., Fogel, M.L., Sprague, E.K., Hodson, R.E., 1987. Depletion of 13C in lignin and its implications for stable carbon isotope studies. Nature 329, 708–710. Berg, B., McClaugherty, C., 2014. Plant Liter: Decomposition, Humus Formation, Carbon Sequestration. Third edition. Springer Berlin Heidelberg, New York, USA. Bragazza, L., Iacumin, P., 2010. Seasonal variation in carbon isotopic composition of bog plant litter during 3 years of field decomposition. Biol. Fertil. Soils 46, 877–881. Brovkin, V., Bodegom, P.M.V., Kleinen, T., Wirth, C., 2011. Plant-driven variation in decomposition rates improves projections of global litter stock distribution. Biogeosci. Discuss. 8, 565–576. Burgess, R.M., Ryba, S.A., Cantwell, M.G., Gundersen, J.L., 2001. Exploratory analysis of the effects of particulate characteristics on the variation in partitioning of nonpolar organic contaminants to marine sediments. Water Res. 35, 4390–4404. Callaway, J.C., Borgnis, E.L., Turner, R.E., Milan, C.S., 2012. Carbon sequestration and sediment accretion in San Francisco bay tidal wetlands. Estuar. Coasts 35, 1163–1181. Charman, D.J., Amesbury, M.J., Hinchliffe, W., Hughes, P.D.M., Mallon, G., Blake, W.H., Daley, T.J., Gallego-Sala, A.V., Mauquoy, D., 2015. Drivers of Holocene peatland carbon accumulation across a climate gradient in northeastern North America. Quat. Sci. Rev. 121, 110–119. Connin, S.L., Feng, X., Virginia, R.A., 2001. Isotopic discrimination during long-term decomposition in an arid land ecosystem. Soil Biol. Biochem. 33, 41–51. Dignac, M.F., Rumpel, C., Thevenot, M., 2010. Fate of lignins in soils: a review. Soil Biol. Biochem. 42 (1200–121). Ember, L.M., Williams, D.F., Morris, J.T., 1987. Processes that influence carbon isotope variations in salt marsh sediments. Mar. Ecol. Prog. Ser. 36, 33–42.

Fellerhoff, C., Voss, M., Wantzen, K.M., 2003. Stable carbon and nitrogen isotope signatures of decomposing tropical macrophytes. Aquat. Ecol. 37, 361–375. Fernandez, I., Mahieu, N., Cadisch, G., 2003. Carbon isotopic fractionation during decomposition of plant materials of different quality. Glob. Biogeochem. Cycles 17, 1075. Finlayson, M., Harris, J., Mccartney, M., Young, L., Chen, Z., 2010. Report on Ramsar Visit to Poyang Lake Ramsar Site, Ramsar Site. P.R. China. pp. 1–34. Foulquier, A., Volat, B., Neyra, M., Bornette, G., Montuelle, B., 2013. Long-term impact of hydrological regime on structure and functions of microbial communities in riverine wetland sediments. FEMS Microbiol. Ecol. 85, 211–226. Gautam, M.K., Lee, K.S., Song, B.Y., Lee, D., Bong, Y.S., 2016. Early-stage changes in natural (13)c and (15)n abundance and nutrient dynamics during different litter decomposition. J. Plant Res. 129, 463–476. Guan, S., 1986. Soil Enzymes and Their Research Methods. Agriculture Press, Beijing. Hobbie, E.A., Hobbie, J.E., 2008. Natural abundance of 15 n in nitrogen-limited forests and tundra can estimate nitrogen cycling through mycorrhizal fungi: a review. Ecosystems 11, 815. Hobbie, E.A., Högberg, P., 2012. Nitrogen isotopes link mycorrhizal fungi and plants to nitrogen dynamics. New Phytol. 196, 367–382. Hodge, A., Robinson, D., Fitter, A., 2000. Are microorganisms more effective than plants at competing for nitrogen? Trends Plant Sci. 5, 304–308. Joergensen, R.,.G., Mueller, T., 1996. The fumigation-extraction method to estimate soil microbial biomass: calibration of the k EN value. Soil Biol. Biochem. 28, 33–37. Kamiri, H., Kreye, C., Becker, M., 2013. Effect of management and soil moisture regimes on wetland soils total carbon and nitrogen in Tanzania. Egu General Assembly. 15. Klotzbücher, T., Kaiser, K., Guggenberger, G., Gatzek, C., Kalbitz, K., 2011. A new conceptual model for the fate of lignin in decomposing plant liter. Ecology 92, 1052–1062. Lehmann, M.F., Bernasconi, S.M., Barbieri, A., Mckenzie, J.A., 2002. Preservation of organic matter and alteration of its carbon and nitrogen isotope composition during simulated and in situ early sedimentary diagenesis. Geochim. Cosmochim. Acta 66, 3573–3584. Li, Z.C., Xu, D.Y., Yang, X.S., Fu, M.Y., Sun, X.Z., Xi, J.R., 2008. Carbon dynamics of litter decomposition in six forest stands of subtropical China. For. Res. 21, 675–680 (Print in Chinese with English abstracts). Li, L.I., Lei, G.C., Gao, J.Q., Cai, L.U., Yan, Z., Jia, Y.F., Yang, M., Suolangduoerji, 2011. Effect of water table and soil water content on methane emission flux at carexmuliensismarshes in Zoige plateau. Wetland Sci. 9, 173–178 (in Chinese). Loranger, G., Ponge, J.F., Imbert, D., Lavelle, P., 2002. Leaf decomposition in two semievergreen tropical forests: influence of litter quality. Biol. Fertil. Soils 35, 247–252. Lyu, X., Yu, J., Zhou, M., Ma, B., Wang, G., Zhan, C., Han, G., Guan, B., Wu, H., Li, Y., Wang, D., 2015. Changes of soil particle size distribution in tidal flats in the yellow river delta. PLoS One 10, e0121368. Macko, S.A., Estep, M.L.F., 1984. Microbial alteration of stable nitrogen and carbon isotopic compositions of organic matter. Org. Geochem. 6, 787–790. Mentzer, J.L., Goodman, R.M., Balser, T.C., 2006. Microbial response over time to hydrologic and fertilization treatments in a simulated wet prairie. Plant Soil 284, 85–100. Miao, G., Noormets, A., Domec, J., Trettin, C.C., Mcnulty, S.G., Sun, G., King, J.S., 2014. The effect of water table fluctuation on soil respiration in a lower coastal plain forested wetland in the southeastern U.S. J. Geophys. Res. Biogeosci. 118, 1748–1762. Mitsch, W.J., Gosselink, J.E., 2000. Wetlands. 3rd ed. John Wiley & Sons, New York, p. 920. Nadelhoffer, K.J., Fry, B., 1988. Controls on natural 15N and 13C abundances in forest soil organic matter. Soil Sci. Soc. Am. J. 52, 1633–1640. Ngao, J., Cotrufo, M.F., 2014. Carbon isotope discrimination during litter decomposition can be explained by selective use of substrate with differing δ13C. Biogeosci. Discuss. 8, 51–82. Nie, L., Qin, W.U., Yao, B., Shan, F.U., Qiwu, H.U., 2016. Leaf litter and soil carbon, nitrogen, and phosphorus stoichiometry of dominant plant species in the Poyang Lake wetland. Acta Ecol. Sin. 36, 1898–1906 (in Chinese). Olson, J.S., 1963. Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44, 322–331. Osono, T., Takeda, H., Azuma, J.I., 2008. Carbon isotope dynamics during leaf litter decomposition with reference to lignin fractions. Ecol. Res. 23, 51–55. Piao, H.C., Zhu, J.M., Liu, G.S., Liu, C.Q., Tao, F.X., 2006. Changes of natural 13C abundance in microbial biomass during litter decomposition. Appl. Soil Ecol. 33, 3–9. Preston, C.M., Nault, J.R., Trofymow, J.A., 2009. Chemical changes during 6 years of decomposition of 11 litters in some Canadian forest sites. part 2. 13C abundance, solid-state 13 C NMR spectroscopy and the meaning of “lignin”. Ecosystems 12, 1078–1102. R Development Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Austria, Vienna http://www.r-project.org (accessed 20.05.13). Rejmánková, E., Houdková, K., 2006. Wetland macrophyte decomposition under different nutrient conditions: what is more important, litter quality or site quality? Biogeochemistry 80, 245–262. Schmidt, M.W., Torn, M.S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I.A., Kleber, M., Kogel-Knabner, I., Lehmann, J., Manning, D.A.C., Nannipieri, P., Rasse, D.P., Weiner, S., Trumbore, S.E., 2011. Persistence of soil organic matter as an ecosystem property. Nature 478, 49. Scott, N.A., Cole, C.V., Elliott, E.T., Huffman, S.A., 1996. Soil textural control on decomposition and soil organic matter dynamics. Soil Sci. Soc. Am. J. 60, 1102–1109. Shi, A., Marschner, P., 2014. Drying and rewetting frequency influences cumulative respiration and its distribution over time in two soils with contrasting management. Soil Biol. Biochem. 72, 172–179. Spohn, M., Klaus, K., Wanek, W., Richter, A., 2016. Microbial carbon use efficiency and biomass turnover times depending on soil depth – implications for carbon cycling. Soil Biol. Biochem. 96, 74–81. Tang, L., Zeng, G.N., Wang, h., Shen, G.L., Huang, D.L., 2005. Enzym. Microb. Technol. 36, 960.

G. Zhang et al. / Science of the Total Environment 630 (2018) 1205–1215 Trinder, C.J., Johnson, D., Artz, R.R.E., 2008. Interactions among fungal community structure, leaf litter decomposition and depth of water table in a cutover peatland. FEMS Microbiol. Ecol. 64 (3), 433–448. Tripathi, G., Deora, R., Singh, G., 2013. The influence of litter quality and micro-habitat on litter decomposition and soil properties in a silvopasture system. Acta Oecol. 50, 40–50. Waddington, J.M., Morris, P.J., Kettridge, N., Granath, G., Thompson, D.K., Moore, P.A., 2015. Hydrological feedbacks in northern peatlands. Ecohydrology 8, 113–127. Wang, Y., Yu, X., Xu, J., 2011. Erratum to: decreased trophic position as a function of increasing body size of a benthic omnivorous fish from the largest freshwater lake in China. Environ. Biol. Fish 91, 513. Wang, X., Han, J., Xu, L., Wan, R., Chen, Y., 2014. Soil characteristics in relation to vegetation communities in the wetlands of Poyang Lake, China. Wetlands 34, 829–839. Wang, G., Jia, Y., Li, W., 2015. Effects of environmental and biotic factors on carbon isotopic fractionation during decomposition of soil organic matter. Sci. Rep. 5, 11043. Wedin, D.A., Tieszen, L.L., Dewey, B., Pastor, J., 1995. Carbon isotope dynamics during grass decomposition and soil organic matter formation. Ecology 76, 1383–1392. Wiedermann, M.M., Kane, E.S., Potvin, L.R., Lilleskov, E.A., 2017. Interactive plant functional group and water table effects on decomposition and extracellular enzyme activity in sphagnum, peatlands. Soil Biol. Biochem. 108, 1–8. Williams, C.J., Yavitt, J.B., 2003. Botanical composition of peat and degree of peat decomposition in three temperate peatlands. Ecoscience 10, 85–95. Xiao, L., Zhu, B., Nsenga, K.M., Jiang, S., 2017. Plant soaking decomposition as well as nitrogen and phosphorous release in the water-level fluctuation zone of the three gorges reservoir. Sci. Total Environ. 592.

1215

Xie, Y., Xie, Y., Xiao, H., Chen, X., Li, F., 2017. Controls on litter decomposition of emergent macrophyte in Dongting lake wetlands. Ecosystems 20, 1383–1389. Xu, S., Liu, Y., Cui, Y., Pei, Z., 2011. Litter decomposition in a subtropical plantation in Qianyanzhou, China. J. For. Res. 16, 8–15. Yoon, T.K., Noh, N.J., Han, S., Lee, J., Son, Y., 2014. Soil moisture effects on leaf litter decomposition and soil carbon dioxide efflux in wetland and upland forests. Soil Sci. Soc. Am. J. 78, 1804–1816. Yu, J.G.E.S., 2010. Relationships among plants, soils and microbial communities along a hydrological gradient in the New Jersey pinelands, USA. Ann. Bot. 105, 185–196. Zhang, D., Hui, D., Luo, Y., Zhou, G., 2008. Rates of litter decomposition in terrestrial ecosystems: global patterns and controlling factors. J. Plant Ecol. 1, 85–93. Zhang, Q.J., Yu, X.B., Qian, J.X., 2012. Distribution characteristics of plant communities and soil organic matter and main nutrients in the Poyang Lake Nanji Wetland. Acta Ecol. Sin. 32, 3656–3669 (in Chinese). Zhang, G.S., Yu, X.B., Zhang, Q.J., Li, Y., Liu, Y., Duan, H.L., 2017a. Variation in the distribution of soil microbial community structure along ground water level gradients in the Poyang Lake Wetland. Acta Ecol. Sin. https://doi.org/10.5846/stxb201705220946 (Print in Chinese with English abstracts). Zhang, H., Yu, X., Wang, Y., Xu, J., 2017b. The importance of terrestrial carbon in supporting molluscs in the wetlands of Poyang Lake. Chin. J. Oceanol. Limnol. 35, 1–8. Zhang, G.S., Yu, X.B., Zhang, Q.J., Li, Y., Liu, Y., Duan, H.L., 2018. Variation in the distribution of soil microbial community structure along ground water level gradients in the Poyang Lake Wetland. Acta Ecol. Sin. https://doi.org/10.5846/stxb201705220946 (Print in Chinese with Englishabstracts).