Spartina alterniflora invasion affects soil carbon in a C3 plant-dominated tidal marsh

The carbon cycle is significantly affected by Spartina alterniflora invasion through its impact on blue carbon in many salt marshes. To determine the impacts on soil organic carbon (SOC), we studied the vertical and horizontal distribution of SOC. And stable carbon isotopes were used to explore the impact of the age of S. alterniflora invasion on SOC in Chongming Dongtan wetland located in the Yangtze River estuary, China. The results showed that the SOC concentration was higher in the S. alterniflora community than that in the native Phragmites australis community. The age of invasion and the SOC concentration increased with increasing elevation, while the SOC concentration decreased with increasing soil depth. The δ13C value became less negative at greater depth, which was related to the contribution from 13C- enriched carbon sources after 3 years of invasion. After 7 and 10 years, the δ13C value became more negative at greater depth in both communities. S. alterniflora had a positive effect on the soil carbon pool, and its contribution was related to soil depth. In the low tidal marshes, the contribution of S. alterniflora was negatively correlated with soil depth, while it was positively correlated with soil depth in the high tidal marshes. The results from this study will contribute to improved understanding of future ecological consequences.

Stable carbon isotopes have been employed as tracers of the sources of accumulated organic carbon in salt marsh sediments [16][17][18][19][20][21][22][23][24] . This technique is based on differences in stable carbon isotope ratios, such as those between the soil carbon inputs of S. alterniflora (C 4 source) and Phragmites australis (C 3 source). C 4 and C 3 sources are readily distinguished by their distinctive δ 13 C values. C 3 plants have been reported to exhibit δ 13 C values between −35 and −20‰, while those of C 4 plants range between −19 and −9‰ [25][26][27] . These values are widely used to study the source of soil organic matter. If the climate and environmental conditions are the same for all sampled plots in a given study, the variation in the soil isotope composition can be employed to measure the newly generated SOC after S. alterniflora invasion.
In this study, we hypothesized that time since S. alterniflora invasion would significantly affect the soil carbon pool in different soil fractions in the Chongming Dongtan tidal marsh. Analyzing the stable carbon isotopic composition and carbon concentration in the soil to a depth of 50 cm in areas where the invasive plant S. alterniflora and the native plant P. australis occurred allowed us to better understand the influence of invasive C 4 plants on soil carbon. The major purposes of this study were to 1) understand the different vertical and horizontal distributions of the soil carbon pool in this Yangtze River estuary salt marsh; 2) quantify the contribution of S. alterniflora invasion to SOC; and 3) determine the influence of different S. alterniflora invasion times (3, 7 and 10 years) on the soil carbon pool.

Results
Plant biological traits. Along the southern transect, the aboveground biomass of the S. alterniflora community was 3500, 3889 and 3412 g m −2 for the low, middle and high tidal marshes, respectively, but these values did not differ significantly. These biomass values were approximately 1.5 times higher than those recorded for the P. australis community. Along the northern transect, the corresponding aboveground biomass of the S. alterniflora community was 3525, 2800 and 3104 g m −2 for the low, middle, and high tidal marshes, respectively; and these values were more than three times higher than those recorded for the P. australis community (Table 1).

Distribution patterns of SOC in Dongtan wetland. Horizontal distribution of SOC.
Along the southern transect, the average SOC concentrations in all soil layers rapidly increased from approximately 2% to 8% across an area about 200 m from the mudflat to the vegetated tidal marsh which the distance reached 1000 m. Then, the SOC values remained at a high level, starting at approximately 8%, increased slightly when the distance reached approximately 1700 m. Along the entire northern transect, the soil total carbon content quickly increased rapidly from the mudflat to the vegetated area (Fig. 1).
Along the southern transect, the SOC of the S. alterniflora community increased during the period of plant invasion, ranging from 3.00% to 7.79%. Compared with the SOC of the mudflat, the SOC of the S. alterniflora-invaded plots (S-L-S) was increased by 128%. Additionally, the SOC of the S. alterniflora-invaded plots increased by 43.25% and 24.00% in the middle and high marshes, respectively, compared with that associated with P. australis at the same elevation. The SOC along the northern transect ranged from 3.87% to 11.16%. Compared with the SOC of the mudflat, the SOC of the S. alterniflora-invaded sample plots (N-L-S) was increased by 55.64%. The SOC measured in the S. alterniflora plots was generally higher than that in the P. australis plots, except in the N-M-S sample plot along the northern transect (Fig. 1b).
Vertical distribution of SOC. The average SOC content in Dongtan wetland over a depth of 0-50 cm increased with increasing distance from the mudflat, but the rate of change varied in each soil layer. At the edge of the coastal marsh along the southern transect, the SOC content remained almost unchanged at a depths of 0-20 cm (P > 0.05) but decreased from 20 to 50 cm (P < 0.05) in the mudflat and the low and middle marshes (Fig. 2a). In the high marsh and near the dike, the SOC content decreased with increasing soil depth. The northern transect   showed the same pattern (Fig. 2b). In addition, there was a greater decrease in SOC with depth in the high marsh than that in the low and middle marshes.
Isotope composition and soil carbon stabilization in Dongtan wetland. In this study, the δ 13 C value of S. alterniflora ranged from −13.68 to −14.5‰ ( Table 2). The δ 13 C value of S. alterniflora leaves was much lower than those of the roots and litter (P < 0.05), while the δ 13 C values of the roots and litter were similar. The average δ 13 C value for the leaves, roots, and litter of −13.47‰, is regarded as representative for S. alterniflora. Along the southern transect, the δ 13 C value of the S. alterniflora community was less at greater soil depth in the low tidal marsh. The rate of increase of the δ 13 C value was 0.24 (R 2 = 0.91, P < 0.05). In the middle and high tidal marshes, the δ 13 C value of the S. alterniflora community was more negative at greater soil depth. The rates of increase in the δ 13 C value were −0.59 (R 2 = 0.88, P < 0.05) and −0.56 (R 2 = 0.96, P < 0.05) for the middle and high tidal marshes, respectively. There were significant differences in the δ 13 C values between the S. alterniflora and P. australis communities (P < 0.05). Along the northern transect, the δ 13 C value of the S. alterniflora community of the mudflat and high tidal marsh exhibited the same pattern as for the southern transect, while the δ 13 C value of the S. alterniflora community did not change in the middle tidal marsh. In the high tidal marsh, the δ 13 C value of SOC in the S. alterniflora community was more negative as the stem density of S. alterniflora increased. The rate of decrease in the δ 13 C value was −1.16 (P < 0.05, R2 = 0.97). However, there were no significant differences in the δ 13 C values between the P. australis and S. alterniflora communities (Fig. 3).

Discussion
Plant invasion has been found to alter various components of nutrient cycles in wetland ecosystems 28 . Our results showed that the seaward invasion of S. alterniflora could play a major role in increasing carbon fixation in Chongming Dongtan wetland (Fig. 1). SOC was generally higher in soil dominated by S. alterniflora than that in soil dominated by native P. australis (Fig. 1). Higher SOC concentrations have been found in S. alterniflora communities due to high levels of rhizome and root biomass and elevated plant residue decomposition rates 15 . Our results showed that the aboveground biomass of S. alterniflora was significantly higher than that of P. australis (P < 0.05) ( Table 1), indicating that S. alterniflora fixed more carbon and potentially provided more residues to the wetland soil 6,29,30 . In terms of the soil SOC distribution in the S. alterniflora community along the southern transect, the SOC content in the S-M-S plot was the highest but was not significantly different from that in the S-H-S plot (Fig. 2a). This showed that SOC stabilized after 7-10 years of S. alterniflora invasion. Wang et al. 31 reported that SOC was saturated after 5-12 years of S. alterniflora invasion in marshes along the coast of Xinyanggang in North Jiangsu Province, China. However, SOC gradually increased with increasing time since invasion along the northern transect (Fig. 2b), similar to the results of previous research. Zhang et al. 11 revealed that in Wanggang, Jiangsu Province, the carbon sequestration capacity of the topsoil remained unsaturated after 14 years of S. alterniflora invasion. The differences between the northern and southern transects could be related to the sedimentation conditions and organic carbon input from decomposition (e.g., the tidal influx rate and organic matter content in tidewater), environmental factors (e.g., soil moisture, redox potential, heat, bioturbation, and many associated factors) and the characteristics of species development in different research areas. Long-term monitoring of changes in SOC and controlled experiments are needed in future studies.
In the middle and high marshes, the SOC content was higher in the surface (0~20 cm) than that in the deeper soil layers. However, the highest SOC in the vegetation edge zone, the transition from the mudflat to the tidal marsh, occurred in the subsurface soil layer (5-20 cm) (Fig. 2). These results indicated that a considerable amount of partly decomposed S. alterniflora residue remained within soil depths of 5-20 cm. At the vegetation edge zone, the soil formation process began on the dead S. alterniflora trapped by the tidally transported sediment. Farther from the mudflat, tidal forcing is weak, resulting in less sediment deposited on the surface and a higher SOC content in the near-surface layer. Aboveground parts (i.e., standing desiccated parts) decompose in the air or fall to the ground. The tide carries away decomposed soluble carbon and tiny carbon granules, transporting them inshore, where they are deposited on the soil surface or leach into the soil 6,32,33 .
The photosynthetic pathways of different plants make different relative contributions to the net primary productivity of surface soil, and the δ 13 C values of SOC vary 8 . Therefore, under conditions of long-term stability of plant species, the δ 13 C value of SOC is strongly influenced by plant δ 13 C 26,33-35 . Our stable isotopic analysis suggested that the δ 13 C value of SOC along the soil profile differed between the S. alterniflora and P. australis communities. The δ 13 C value was less negative at greater depth, which was related to the contribution from 13 C-enriched carbon sources, after 3 years of invasion by S. alterniflora (Fig. 3). This finding is consistent with previously reported δ 13 C distributions in soil profiles 36,37 . However, after 7 and 10 years, the δ 13 C value was typically more negative at greater depth in both S. alterniflora and P. australis communities. This result can be explained by the presence of SOC derived from 13 C-depleted C 3 vegetation 32 . Cheng et al. 15 considered that 13 C-depleted C 4 S. alterniflora contributed less to SOC than native species in Jiuduansha tidal marsh because of a short invasion (i.e.,<7 years) before which Scirpus maiqueter monoculture was dominant on the island for a 30-year period.
Along the southern transect, the rates of the decrease in δ 13 C values along the 0-20 cm soil profile in the S. alterniflora community were 0.88 (P < 0.01, R 2 = 0.99) and 0.78 (P < 0.05, R 2 = 0.80) for the middle and high marshes, respectively. In addition, the rates of the decrease along the 20-50 cm soil profile in the S. alterniflora community were 0.18 (P < 0.05, R 2 = 0.80) and 0.33 (P < 0.01, R 2 = 0.99), respectively. These results likely occurred because SOC decomposes rapidly in surface soil. The SOC was almost completely decomposed when the δ 13 C value reached a plateau. Subsequently, the decomposition rate of SOC slowed, as did the change in the δ 13 C value. In this study, the S. alterniflora residue contributed predominantly to SOC at a depth of 20 cm compared with a lower contribution to surface SOC due to increased oxygenation that resulted in greater decomposition of organic carbon 38 . S. alterniflora contributed less to deep SOC because of less root exudation and weaker osmosis of soluble organic carbon 34 .
In the present study, the sediment of the mudflat originated mainly from tidally transported sediment, which is likely why the δ 13 C value of the mudflat sediments was approximately −24.33‰, which is similar to that of terrestrial C 3 plants, such as P. australis. A double-variable hybrid model was used to calculate the contribution of S. alterniflora to surface SOC; with increasing time since S. alterniflora invasion, the content and proportion of the contributed SOC increased (Fig. 5). During the seaward invasion of S. alterniflora, the SOC content and proportion attributed to this species showed a clear positive correlation with time since invasion (P < 0.01), revealing that the SOC contributed by S. alterniflora increased continuously. The proportion of SOC contributed by S. alterniflora over 10 years of invasion reached 27.3 and 33.7% along the southern and northern transects, respectively.
A study conducted in the Jiuduansha tidal marsh in Shanghai showed that after 8 years of seaward invasion by S. alterniflora in a S. mariqueter community, the contribution to surface SOC ranged from 5 to 10% 15 . In addition, after 14 years of landward invasion by S. alterniflora in a Suaeda salsa community in the Wanggang tidal marsh in Jiangsu Province, the contribution of S. alterniflora to surface SOC was 18.7% 31 . First of all, soil organic carbon input was affected by the high primary productivity and extensive root system of S. alterniflora. In addition, S. alterniflora had higher primary productivity in Chongming Dongtan, ranging from 2800 to 3889 g·m −2 , compared with that in Wanggang and Jiuduansha tidal marshes, which ranged from 1442 to 2596 g·m −2 39,40 . It can immobilize carbon from the atmosphere and surrounding areas during growth; this carbon then enters the soil via litter and root secretions, which can directly influence the SOC content. And its well-developed roots, S. alterniflora can weaken the power of the tide, solidifying sediments in the flat and sedimenting tiny granules, which greatly improves adsorption due to the high specific surface area of the tiny granules, facilitating an increase in SOC. Second, our research site is located in the Yangtze River estuary, which carried a large number of organic matters through the tidal action, and deposited in Chongming Dongtan tidal marsh. Wang et al. 31 considered the biomass of S. alterniflora and the sediment particle size to be the dominant factors determining the content of SOC. Additionally, the SOC content derived from S. alterniflora showed little correlation with S. alterniflora biomass but presented a clear positive correlation with particle size, which revealed a complicated process between the plant and soil that was greatly affected by soil characteristics.
The organic carbon input of salt marsh ecosystem mainly has two kinds of ways, including endogenous and exogenous input. Endogenous input contains soil organic matter and litter decomposition 41 . The exogenous inputs include suspended particles and dissolved organic carbon brought by tidal waters 42 . In Chongming Dongtan, suspended organic matter and sediment entered the soil carbon pool through horizontal flux. S. alterniflora has dense stalks and developed underground roots. Its ability to attenuate wave energy and reduce flow contributes to the sedimentation and deposition of silt 43 . Therefore, in this study of the effects of invasion on the carbon balance of ecosystem, the lateral flux caused by tidal activity is a component as well.
In this study, the tide had more influence on the contribution on soil carbon pool during the early period of invasion (after 3 years of invasion). With invasion time extension (after 7 years of invasion), the increase of soil carbon pool mainly depends on the plant (Fig. 4). The tidal flow is affected by surface roughness, elevation, and friction from plant stems and leaves 44 . After 3 years of invasion, the S. alterniflora patch adjacent to the bare flat had less tidal energy consumption and greater biomass loss (such as litter and lodged plants) compared with observations recorded after 7 and 10 years of invasion. In coastal marsh, the frequent cyclical effects of tide brought a non-negligible sea source input to soil carbon pool and showed an influence on the early stage of S. alterniflora invasion 45 .

Conclusions
In summary, we found that S. alterniflora invasion in Chongming Dongtan wetland over the past 10 years was associated with increased SOC concentrations. SOC was generally higher in the S. alterniflora than in the native P. australis community. The contribution of S. alterniflora to the soil carbon pool exhibited an obvious gradient in terms of its horizontal distribution. In the high marsh, which was invaded 10 years ago, the contribution was positively correlated with soil depth, while in the low marsh, which was invaded by S. alterniflora within the past 3 years, the contribution of S. alterniflora to the soil carbon pool was negatively correlated with depth. This result might have occurred because litter is frequently washed away by the tide in the low marsh. Thus, the main source of SOC is likely the root exudates of S. alterniflora.

Materials and Methods
Study area. This study was conducted in the Shanghai Chongming Dongtan Wetland National Nature Reserve (31°25′-31°38′N, 121°50′-122°05′E) located at the eastern end of Chongming Island. Chongming Dongtan is formed by alluvial silt and soil from the upper reaches of the Yangtze River and is the largest and youngest estuarine tidal marsh in the Yangtze River estuary. Chongming Dongtan has a northern subtropical ocean climate, with average annual precipitation of 1117.1 mm and a mean annual temperature of 15.3 °C (Fig. 6a). The native species Scirpus mariqueter and Phragmites australis have dominanted for over 20 years, while invasive S. alterniflora has rapidly spread in this tidal marsh and is replacing native species due to its considerable seed production and high germination rate 46 . Currently, P. australis and S. alterniflora now show zonation in their distribution in Chongming Dongtan 46 .
Using ZY-3 satellite high-resolution remote sensing images and GPS coordinates, two transects were established using a combination of factors specific to each site, such as vegetation type, invasion time, water content, soil physical factors (pH, ORP, Salinity) ( Table 3). The southern and northern transects were located in southeastern and northeastern Dongtan, respectively. The salt marsh directly landward of the study area was reclaimed in 1998. Both transects were established so that they extended through a variety of plant communities containing S. alterniflora and P. australis. The southern transect-with westernmost geographic coordinates of 31°50′N and 121°13′E and easternmost coordinates of 31°32′N and 121°58′E -was approximately 1800 m long. The northern transect-with westernmost coordinates of 31°91.67ʹN and 121°08.85′E and easternmost coordinates of 31°90.77′N and 121°95.52′E-was approximately 700 m long. Both transects were divided into three parts according to the tidal characteristics, i.e., low, middle and high tidal marshes. Low marsh refers to areas of the saltmarsh below mean high water level of neap tide. Middle marsh refers to areas of the saltmarsh between mean water level and mean high water level. High marsh refers to areas of the saltmarsh above mean high water. According to the elevations of two tidal gauges near Chongmin Dongtan (Jigujiao Gauge and Wusong Gauge), the mean high water level of Chongming Dongtan is 1.46 ± 0.01 m, the mean high water level during neap tide is 0.95 ± 0.05 m, and the mean water level is approximately 0.2 ± 0.03 m. These two chronosequenced transects extended through remnant P. australis and the S. alterniflora communities that replaced P. australis in 2001   (10 years previously), 2004 (7 years) and 2008 (3 years). Plots were established every 200-300 m along the transects. Each plot consisted of two 20 m × 20 m quadrats, and samples were collected from five replicate 0.5 m × 0.5 m subquadrats randomly established within each quadrat (Fig. 6b).
Field sample collection. Soil, leaf, litter, and root samples were collected from the S. alterniflora and P.
australis communities. Cores were extracted using a stainless-steel peat core sampler and were divided into 5 cm sections in the field. Each section was sealed in multiple plastic bags with as much air squeezed out as possible and stored on ice or at 4 °C to avoid volatilization losses and minimize microbial activity until analysis.
Laboratory analyses. Plant biomass was sampled in all plots in September 2011, when the aboveground plant biomass was the greatest according to a previous study in a nearby wetland 47 . All aboveground plant biomass was harvested and then oven-dried at 50 °C to a constant weight. Plant material and organic debris were first gently removed from the soil samples using forceps, and the leaf, root, litter, and soil samples were then dried at 60 °C to a constant weight and ground to pass through 100-mesh sieves 34,48 . After thorough mixing, approximately 10 g of dried soil was treated with 0.5 mol L −1 hydrochloric acid (HCl) for 24 h at room temperature to remove any inorganic carbon, and the unhydrolyzed residue was categorized as SOC 48 . Approximately 3 mg of the leaf, root, and litter subsamples was weighed and analyzed to determine the total carbon contents of the plant materials. Non-HCl-treated soil subsamples (approximately 50 mg) were weighed and analyzed to determine soil total carbon. These measurements were performed using a soil element analyzer (Flash, EA, 1112 Series, USA).
The carbon isotope ratio of organic materials (leaves, roots, litter and HCl-treated soil) was measured using an isotope ratio mass spectrometer (Thermo Fisher, Delta V Advantage, Flash, EA, 1112 Series, USA). The precision of δ 13 C was ±0.1‰, which was based on repeated measurements of a laboratory working standard. The δ 13 C value of the organic matter was calculated using the following equation: Calculation of carbon inputs. The proportion f is the proportion of S. alterniflora-derived carbon in the soil. This calculation was based on the δ 13 C values of the soil before and after S. alterniflora invasion (δ 13 C old and δ 13 C new , respectively) and the δ 13 C value of S. alterniflora (δ 13 C mix ), which was the average of the values for the leaves, roots, and litter ( Statistical analyses. Statistical analyses were performed using the statistical software package SPSS version 18 for Windows 7. Data were tested for normality prior to statistical analysis. One-way ANOVA at a 5% significance level was employed to determine the differences in SOC and the biomass, height and density of the plants. Significant differences in plant δ 13 C values between the two species were analyzed using the t-test and the least significant difference at P < 0.05. δ 13 C data were arcsine transformed to ensure homoscedasticity before the ANOVAs were conducted. Regression analyses were performed to test the relationships between the accumulative rates of SOC or S. alterniflora-derived carbon and the ages of the plant invasions.
Data availability statement. The datasets generated during the current study are available from the corresponding author upon reasonable request.