1 Introduction

Coastal wetlands, also known as blue carbon (C) ecosystems, consist of a wide variety of saltmarshes, mangrove forests, and seagrass meadows. Although these ecosystems occupy a relatively small fraction of the earth’s surface, they play a disproportionately large role in global C cycling (Milton et al. 2018; Najjar et al. 2018). In the coastal ecosystem of China, saltmarsh is the most prominent wetland type, widely distributed across the entire coastline of China (Cao et al. 2013; Wang et al. 2023). Previous studies on saltmarshes in China suggested that the total soil C stock amounted to 7.5 ± 0.6 Tg, and the average carbon accumulation rate (CAR) reached 201 g C m−2 yr−1 – much higher than the global average (168 g C m−2 yr−1) (Fu et al. 2021; Macreadie et al. 2019; Wang et al. 2019), highlighting coastal wetlands of China as an important C pool in global C budget (Wang et al. 2021a, b).

Unfortunately, over recent decades, the landscape of saltmarshes in China has been undergoing rapid changes, mainly due to species invasion and land use changes (Osland et al. 2018; Wang et al. 2022a, b, 2021a). All these processes strongly modified the distribution and composition of saltmarshes, with potentially far-reaching impact on coastal C stock due to the variable C sequestration capacity of different marsh types. For instance, Spartina alterniflora (S. alterniflora), an exotic marsh type, is aggressively invading the coastal mudflats of China (Mao et al. 2019). This marsh type is characterized by high biomass and large dense roots that may contribute to coastal C sink (Dan et al. 2015; Huang et al. 2018). For example, previous studies conducted in the Yellow River Delta, Yancheng, and Jiuduansha saltmarshes have indicated that the expansion of S. alterniflora led to an increase in soil organic carbon (SOC) storage (Huang et al. 2015; Liao et al. 2007; Zhang et al. 2018). In contrast, Suaeda salsa (S. salsa), a native marsh type, is steadily replaced by Phragmites australis (P. australis) in some areas due to soil desalination, a process that may undermine regional C sink (Wang et al. 2022a, b). The vegetation transition was in strong agreement with the desalination trend, which was attributed to the colonization of S. alterniflora at the low-elevation mudflats increased sediment deposition and decreased the salt inundation from seawater (Hughes et al. 2019; Liu et al. 2007; Redfield 1972; Zhou et al. 2015). As a result, the decreasing salinity of the habitat made S. salsa less competitive, and its living area was primarily occupied by P. australis (Wang et al. 2022a, b). This effect caused the expansion of P. australis seaward, continuously outcompeting and replacing S. salsa (Li et al. 2021; Wang et al. 2022a, b). This long-distance interaction may undermine regional C sink potential and require careful management of saltmarshes to sustain their ecological and C sequestration functions.

Coastal land reclamation has become a common global practice, with China leading the way in massive land reclamation initiatives. The demand for coastal reclamation in China has risen significantly over the years, driven mainly by economic growth (Ma et al. 2014). China has reclaimed around 11,162 km2 of land between 1979 to 2014 (Meng et al. 2017), while losing approximately 22,000 km2 of coastal wetlands since 1949 (Sengupta et al. 2018). Land reclamation poses a risk of greatly decreasing the previously sequestered and stored C pools of saltmarshes (Bu et al. 2015). Coastal reclamation is often implemented to satisfy the increasing demand for agricultural land. However, studies have shown that converting salt marshes into croplands can lead to a significant decline in SOC pools over a period of decades (Fernandez et al. 2010; Iost et al. 2007). Such land use conversion causes roughly 0.15–1.02 Pg yr−1 of CO2 released from coastal ecosystems (Pendleton et al. 2012).

These emerging phenomena observed in coastal wetlands call for a re-assessment of the blue C sink in China. However, current knowledge on the spatiotemporal distribution of major marsh types in China remains fragmented (Meng et al. 2019), posing obstacles for accurate assessment of coastal C pool, especially on a regional-scale. Here we systematically evaluated the spatiotemporal dynamics of marsh types and the associated C stock change in Yancheng, the largest saltmarsh ecosystem in China that has been heavily impacted due to the invasion of S. alterniflora and land reclamation, and have experienced significant uncertainties in their blue C sink capacity (Li et al. 2023). Specifically, we (1) mapped changes in marsh types between 1982 and 2021 using Remote Sensing; (2) examined the dynamics of C stock and CAR associated with different marsh types by meta-analysis; and (3) evaluated the blue C sink capacity following the long-term trajectory of marsh cover change across the study region. We anticipate that the insights generated from this study will update the current understanding of coastal C sink in China, and the study may also serve as a template for exploring C stock change in other coastal regions of the world, where vegetation shift is underway.

2 Methods and materials

2.1 Study area

The Yancheng Natural Reserve (32°34′-34°28′N, 119°27′-121°16′E), located in central to northern Jiangsu Province, is the largest continuous coastal wetland in China. Its core area is situated between Xinyanggang and Doulonggang estuaries. This region is dominated by a subtropical monsoon climate with a mean annual air temperature of 14.9 °C and mean annual precipitation of 1,010 mm. The study area represents an intertidal silt-muddy coastal saltmarsh landscape subject to irregular semidiurnal tides (Wang et al. 2022a, b). With little anthropogenic disturbance in this region, the vegetation retains the original ecological succession sequence of the coastal zone. From sea to land, the sequence was: mudflat, S. alterniflora, S. salsa, and P. australis (Gao et al. 2012, 2016; Liu et al. 2007; Xu 2015; Zang 2019). The obtained sampling sites are shown in Fig. 1. The sampling sites were primarily in Sheyang county and Dafeng city, with the rest in the core areas of the Yancheng Natural Reserve.

Fig. 1
figure 1

The specific position of data collection points and the habitat status in Yancheng saltmarshes, where (a) the location of the sampling sites with different marsh types in Yancheng city, Jiangsu Province, China. MF was mudflat, PA was P. australis, SA was S. alterniflora, and SS was S. salsa, (b) the natural saltmarsh habitat with S. alterniflora expanding into mudflats, and (c) the saltmarsh habitat with dike construction that inhibited the S. alterniflora expansion

Full size image

2.2 Saltmarsh classification

We generated 8 time-series coastal vegetation maps between 1982 and 2021 using cloud-free Landsat images (30 by 30 m, Landsat 5–8) downloaded from the National Aeronautics and Space Administration (https://www.nasa.gov/) and GS Cloud website (http://www.gscloud.cn/). All maps were created using supervised classification with human–computer interaction, and each map includes six coastal landcover types, referring to mudflat, S. alterniflora, S. salsa, P. australis, water, and human land use. Specifically, we acquired all clear-sky Landsat images during low tide between July and November – the low-growth season when inter-vegetation contrast is maximized for optimal spectral separation between classes (Wang et al. 2022a, b). We validated the resulting maps using visual interpretation and high-resolution aerial images based on published in situ studies of saltmarshes, and then computed the overall classification accuracy and Kappa coefficient associated with each map for accuracy assessment. During the processes, 70% of samples were used for model training and the remaining 30% of samples for assessing accuracy.

2.3 Data acquisition and processing

We carried out a literature search in 2021 for articles concerning the C cycling of Yancheng saltmarshes (mainly concentrated on articles from 2008 to 2020) in Web of Science (https://www.webofscience.com/wos) and China National Knowledge Infrastructure (CNKI, https://www.cnki.net) employing the following string: (coastal wetland OR tidal marsh OR saltmarsh OR tidal flat) AND Yancheng AND carbon. Ultimately, we completed our search on January 1, 2022, with 106 papers. The original data was obtained directly from the table, and the graphs were digitally extracted using GetData Graph Digitizer 2.26.

We analyzed soil bulk density (BD) data from all studies selected, and for those without BD measurements, we computed the BD data from the SOC measurements using the empirical equation below:

$$\mathrm{OC}=0.04\times\mathrm{LOI}+0.0025\times\mathrm{LOI}^2$$
(1)
$$\mathrm{BD}=\;1/\left[\mathrm{LOI}/{\mathrm k}_{{}^1}\:+\:(1\:-\:\mathrm{LOI})/{\mathrm k}_2\right]$$
(2)

where OC was the content of organic carbon (%), LOI was the loss of ignition (%), BD was the bulk density (g cm−3), and k1 and k2 were the self-packing densities of the pure organic and inorganic components (Craft et al. 1991; Morris et al. 2016).

Most of our data analysis was performed in R 4.1.2 (R Core Team 2021). The permutation test was used to verify the differences among the C stock, CAR, total C storage, and C burial amount of Yancheng saltmarshes, mostly using the R package ‘RVAideMemoire’ (Hervé, 2021). The general linear and generalized additive models were constructed to fit the relationship of the variables by using the R package ‘mgcv’ (Wood 2011). The regulatory factors of C pools in Yancheng saltmarshes were discussed through the structural equation model (SEM), using the R package ‘piecewiseSEM’ (Lefcheck 2016).

2.4 Calculation of C stock and C burial

The plant C stock per unit area was calculated by multiplying the biomass per unit area by the plant C content as shown in the equation below:

$$\mathrm{Plant}\;\mathrm C\;\mathrm{stock}\;\mathrm{per}\;\mathrm{unit}\;\mathrm{area}\;(\mathrm g\;\mathrm C\;\mathrm m^{-2}\;)\;=\;\mathrm{Plant}\;\mathrm{biomass}\;(\mathrm g\;\mathrm m^{-2})\;\times\;\mathrm C\;\mathrm{contents}\;\mathrm{in}\;\mathrm{plant}\;(\%)$$
(3)
$$\text{Plant}\;\text C\;\text{storage}\;(\text{Mg}\;\mathrm C)\;=\;\text{Plant}\;\text C\;\text{stock}\;\text{per}\;\text{unit}\;\text{area}\;(\text{Mg}\;\text C\;\text{ha}^{-1})\;\times\;\text{area}\;(\text{ha})$$
(4)

SOC stock was determined by SOC content and soil depth in a particular region. Therefore, it is necessary to calculate SOC content within 0–1 m soil depth after sampling at intervals (n is the total number of soil intervals), and then obtain the C stock (Mg C ha−1) through area multiplication and unit transformation.

$$\mathrm C\;\mathrm{density}\;(\mathrm g\;\mathrm{cm}^{-3})\;=\;\mathrm{Bulk}\;\mathrm{density}\;(\mathrm g\;\mathrm{cm}^{-3})\;\times\;\mathrm C\;\mathrm{content}\;(\%)/100\;$$
(5)
$$\mathrm C\;\mathrm{stock}\;\mathrm{in}\;\mathrm{single}\;\mathrm{soil}\;\mathrm{depth}\;(\mathrm g\;\mathrm{cm}^{-2})\;=\;\mathrm{Soil}\;\mathrm C\;\mathrm{density}\;(\mathrm g\;\mathrm{cm}^{-3})\;\times\;\mathrm{Soil}\;\mathrm{depth}\;(\mathrm{cm})\;$$
(6)
$$\mathrm C\;\mathrm{stock}\;\mathrm{per}\;\mathrm{unit}\;\mathrm{area}\;(\mathrm g\;\mathrm{cm}^{-2})\;={\textstyle\sum_{\mathrm i=1}^{\mathrm n}}\;\mathrm C\;\mathrm{stock}\;\mathrm{in}\;{\mathrm{depth}}_{\mathrm i}\;\left(\mathrm g\;\mathrm{cm}^{-2}\right)$$
(7)
$$\text{SOC}\;\text{storage}\;(\text{Mg}\;\text C)\;=\;\text{SOC}\;\text{stock}\;\text{per}\;\text{unit}\;\text{area}\;(\text{Mg}\;\text C\;\text{ha}^{-1})\;\times\;\text{area}\;(\text{ha})$$
(8)

The content of radionuclides in sediments has been widely used to determine the age of sediments in estuaries, coastal wetlands, and continental shelf regions, reveal depositional processes and trace biogeochemical sources (Bellucci et al. 2007; Xia et al. 2016). From the measurement of radionuclides, soil accretion rate (SAR) is obtained, and the CAR in coastal wetlands is usually calculated by SAR and C density. The annual C burial is equal to CAR multiplied by area.

$$\mathrm{CAR}\;(\mathrm g\;\mathrm C\;\mathrm m^{-2}\;\mathrm{yr}^{-1})\;=\;\mathrm C\;\mathrm{density}\;(\mathrm g\;\mathrm{cm}^{-3})\;\times\;\mathrm{SAR}\;(\mathrm{mm}\;\mathrm{yr}^{-1})/10000$$
(9)
$$\mathrm C\;\mathrm{burial}\;(\mathrm g\;\mathrm C\;\mathrm{yr}^{-1})\;=\;\mathrm{CAR}\;(\mathrm g\;\mathrm C\;\mathrm m^{-2}\;\mathrm{yr}^{-1})\;\times\;\mathrm{area}\;(\mathrm m^2)\;$$
(10)

We combined remote sensing with data synthesis to determine the C sequestration of different vegetations and the distribution dynamics in Yancheng saltmarshes from 1982 to 2021. This way, we were able to clarify the C storage and C burial in the Yancheng area over a broad spatiotemporal scale, and predict the direction and magnitude of future C stock changes across the study region. The complete procedure is illustrated in Fig. S1.

3 Results

3.1 Changes of Yancheng saltmarsh area

The distribution of different marsh types in Yancheng between 1982 and 2021 was mapped in time-series (Fig. 2a), with an overall classification accuracy between 87 and 98%, and Kappa Coefficient between 0.82 and 0.90 (Table S1). From 1982 to 2021, the total area of Yancheng saltmarsh decreased by 43.09 km2 (Table 1, Fig. 2c). The net loss was mainly induced by the reduction of S. salsa, which reduced from 411.8 km2 to 43.11 km2 (net loss of 90%). While the area of S. salsa underwent a dramatic decrease, that of S. alterniflora significantly increased by 287.5 km2 over an expansive period of 40 years. The area occupied by P. australis also increased from 164.9 km2 to 203.1 km2.

Fig. 2
figure 2

The dynamics of saltmarshes in Yancheng, in which (a) the map of Yancheng saltmarsh with different types from 1982 to 2021, (b) the variations in the area of Yancheng saltmarsh over recent four decades, and (c) the dynamics of total C storage in Yancheng saltmarsh covering with different vegetation types

Full size image
Table 1 The area of Yancheng saltmarsh covered by different vegetation types, and the net change of carbon storage over the past four decades. PA was P. australis, SA was S. alterniflora, and SS was S. salsa
Full size table

Further analysis found that the massive loss of the S. salsa marsh area was transformed to P. australis marsh, land reclamation, and non-vegetated mudflats, which accounted for 27%, 25%, and 39%, respectively (Table S2). The vastly increased S. alterniflora marsh was mainly attributed to its colonization in mudflats (about 85%), with a small proportion invading the native saltmarshes. In terms of P. australis marsh, land reclamation occupied its area of 49.88 km2, accounting for 80% of the net reduction. However, the rapid expansion of P. australis in S. salsa marsh made up for its loss by land reclamation, and achieved a net increase.

3.2 The C stock and CAR

We analyzed marsh-specific plant biomass per unit area, and identified significant differences among marsh types (P < 0.05). Specifically, S. alterniflora marsh has the highest plant biomass per unit area, followed by P. australis, and S. salsa (Fig. 3a). However, the C content in plant tissues showed different patterns. The aboveground C content of S. salsa was significantly lower than that of the other two marshes, while the belowground C content of S. alterniflora was significantly lower than the other two types (Fig. 3a; P < 0.05). The plant C stock per unit area measured for different marsh types followed the order of S. alterniflora (1270 ± 122.0 g C m−2) > P. australis (789.4 ± 74.55 g C m−2) > S. salsa (98.29 ± 6.79 g C m−2) (Fig. 3b; P < 0.05).

Fig. 3
figure 3

The plant C of Yancheng saltmarsh, in which (a) the plant biomass and its C contents of vegetated saltmarshes in Yancheng, and (b) the plant C storage per unit area in Yancheng saltmarshes with different vegetations. PA was P. australis, SA was S. alterniflora, and SS was S. salsa. AGB was above-ground biomass, BGB was below-ground biomass, AGC stock was above-ground C stock, and BGC stock was below-ground C stock. The asterisks indicated significantly lower C contents than the other two vegetations, and the different lowercase letters indicated significant differences among saltmarshes types (permutation test; P < 0.05)

Full size image

At the 0–100 cm soil depth, no significant difference was found in SOC stock among marsh types (Fig. 4a; P > 0.05). However, the SOC stock per unit area at different soil depths showed distinct patterns. Particularly, at the 0–20 cm soil depth, the SOC stock in S. alterniflora was significantly higher than that in P. australis and S. salsa, while at the 50–100 cm soil depth, the SOC stock in P. australis was significantly higher than that in the other two marsh types (Fig. 4b, d; P < 0.05). However, at 20–50 cm soil depth, no significant difference was detected among marsh types (Fig. 4c; P > 0.05).

Fig. 4
figure 4

The SOC storage covered different vegetations in Yancheng saltmarsh, in which (a) the C storage at the 0–100 cm soil depth, (b) the C storage at the 0–20 cm soil depth, (c) the C storage at the 20–50 cm soil depth, and (d) the C storage at the 50–100 cm soil depth. PA was P. australis, SA was S. alterniflora, and SS was S. salsa. The different lowercase letters indicated significant differences among saltmarshes types (permutation test; P < 0.05)

Full size image

The regulators of plant biomass and SOC stock at the 0–20 cm and 20–50 cm soil depths were explored by SEM. The results showed that aboveground biomass (AGB) was mainly affected by pH, total potassium (TK), plant height, and coverage, while belowground biomass (BGB) was regulated by NO3−, NH4+, and plant coverage (Fig. S2a; P > 0.05). The main regulatory factors of SOC stock were soil water content (SWC), dissolved organic carbon (DOC), the ratio of carbon to nitrogen (C/N), available phosphorous (AP), AGB, and carbonate (CO32−) at the 0–20 cm soil depth (Fig. S2b; P > 0.05), but BGB, DOC, and pH influenced the SOC stock at the 20–50 soil depth (Fig. S2c; P > 0.05).

Both C density and CAR of S. alterniflora were significantly higher than that of P. australis and S. salsa (Table 2; P < 0.05). Specifically, the CAR was 242.2 ± 11.26 g C m−2 yr−1 in S. alterniflora, followed by 108.5 ± 5.00 g C m−2 yr−1 in S. salsa, and 94.77 ± 7.80 g C m−2 yr−1 in P. australis (Table 2).

Table 2 The soil C density (extracted from reported articles, which were mostly focused on surface soil) and CAR (carbon accumulation rate) of Yancheng saltmarsh with different vegetation coverings, where PA was P. australis, SA was S. alterniflora, and SS was S. salsa. Different superscript lowercases within a row indicated significant differences among vegetation types at P < 0.05. The values were means ± SE (The original data of soil C density were sourced from Bao et al. 2016; Fu et al. 2021; Liu et al. 2010, 2013; Qi et al. 2019; Zang 2019)
Full size table

3.3 Dynamics of C storage and C burial

Over the 40 years succession, the total area of saltmarsh in Yancheng decreased by 43.09 km2, while the total C storage and C burial increased from 3152 to 3529 Gg, and from 60.35 Gg yr−1 to 93.57 Gg yr−1, respectively (Fig. 5a). Among them, the total soil C storage, plant C storage, and the C burial increased by 17.98 Gg, 359.0 Gg, and 33.22 Gg yr−1, respectively (Fig. 5b, c, d). The proportion of the plant C pool in the total C pool also increased from 5 to 15%, owing to drastic rise in plant C storage induced largely by S. alterniflora expansion (increased by 210.2%).

Fig. 5
figure 5

The net changes of area and C sequestration in Yancheng saltmarsh, in which (a) the net exchanges of the total area, total C storage, and total C burial over the last 40 years, (b) the dynamics of total soil C storage from 1982 to 2021, (c) the dynamics of total plant C storage from 1982 to 2021, and (d) the dynamics of C burial from 1982 to 2021

Full size image

3.4 Effects of S. alterniflora expansion

The area of S. alterniflora tremendously expanded from 1982 to 2021 (Fig. 6a), and the total plant C storage and C burial also greatly increased (Fig. 5c, d). The linear regression results proved that the elevated plant C storage and C burial were positively correlated with the expansion of S. alterniflora (Fig. 6c, d; P < 0.01). S. alterniflora expansion also promoted surface SOC stock, but the relationship was not linear. The generalized additive model suggested that the SOC storage in areas colonized by S. alterniflora increased rapidly in the first few years, and peaked at the 10th year of S. alterniflora establishment (Fig. 6b; P < 0.01), after which the SOC storage remained relatively stable over time. In addition, the SOC at peak storage was 4.86 times higher than that at the beginning of propagation, showing the contribution of S. alterniflora to C sequestration.

Fig. 6
figure 6

The map of S. alterniflora expansion and its influence on C storage and C burial, in which (a) the expansive trend of S. alterniflora revised from the remote-sensing image in 1982, 1992 and 2012, (b) the relationship between the soil C storage at the 0–20 cm soil depth and the expansion year of S. alterniflora, (c) the relationship between the total plant C storage in Yancheng saltmarsh and the area of S. alterniflora, and (d) the relationship between the total C burial rate in Yancheng saltmarsh and the area of S. alterniflora

Full size image

4 Discussion

4.1 Changes of Yancheng saltmarsh area

Saltmarshes are sensitive to anthropogenic activities, especially coastal reclamation, affecting the distribution of saltmarshes directly or indirectly (Chen et al. 2022). From 1982 to 2021, changes in the area of Yancheng saltmarshes showed a “U-shape” tendency that first fell and then rose (Fig. S3). The total area of Yancheng saltmarshes reached 533.7 km2 in 2021, basically returning to the level of the 1980s. The total area did not change much, but the area of different marsh types altered considerably, mainly due to the effects of the anthropogenic reclamation and the conversion of mudflat to S. alterniflora saltmarsh.

From 1985 to 2010, about 7552 km2 of coastal wetlands were lost in mainland China caused by reclamation, and the shoreline was therefore vastly altered (Liu et al. 2020). Similarly, the total area of Yancheng saltmarsh experienced a sharp decrease from 1982 to 2007, dropping by 28.14%, from 576.8 km2 to 414.5 km2. In the 1980s and early 1990s, conversion to aquaculture ponds became the dominant driver of reclamation, especially in northern China (Wang et al. 2020). During late 1990s to 2010s, urban, industrial, and port expansions were the main driving forces of reclamation, and the reclamation projects reached a maximum in 2009 (Gu et al. 2018). Anthropogenic reclamation activities pushed the coastline regression (Fig. S4), which directly decreased the area of P. australis by 38% till the 2010s. The decline continued until the government implemented protection policies to conserve and restore coastal wetlands in 2009 (Tang et al. 2009).

S. alterniflora invasion is an important issue in the coastal regions of China, and previous studies usually indicated that S. alterniflora outcompetes native marshes, and then spreads to areas unsuitable for most native plants (Nie et al. 2022). However, our time-series remote sensing analyses suggested that mudflat was the predominant habitat colonized by S. alterniflora in the region. The establishment and expansion of S. alterniflora at the low-elevation mudflats can modify water flow, increase sediment deposition, and promote the desalinization of high-elevation saltmarshes, thus influencing the interspecific competition between P. australis and S. salsa (Wang et al. 2022a, b; Zhou et al. 2015). As a result, decreased soil salinity enables the seaward movement of P. australis, leading to the displacement of S. salsa marshes over time, while the S. alterniflora remains stable and expands into mudflats (Tang et al. 2022; Wang et al. 2022a, b; Yang et al. 2016). This encroachment of land reclamation on the landward and S. alterniflora expansion on the seaward have kept the total saltmarsh area stable over the past several decades. However, there has been a significant shift in the distribution of plant communities in saltmarshes, with a considerable reduction in the area occupied by S. salsa due to the transition to P. australis through interspecific competition (Li et al. 2021). Therefore, our study further proved that S. alterniflora did not directly invade the area where native plants grew from the landscape scale, but indirectly exacerbated interspecific competition among native plants by transforming the mudflats into S. alterniflora marsh.

Besides the competition between P. australis and S. salsa that expedited the decline of S. salsa marsh area, climate change and anthropogenic reclamation were vital to S. salsa marsh shrinking (Kirwan and Mudd 2012). When sea-level rise exceeds the C deposition in sediments, sub-tidal saltmarshes will gradually degrade to unvegetated mudflats (Saintilan 2021; Törnqvist et al. 2019). Meanwhile, dams built with cement along the coastline inhibit the landward migration of saltmarshes, further leading to the loss of saltmarsh area. Anthropogenic reclamation, such as shipping, fishing, and tourism, greatly influences on the distribution of saltmarshes (Poppe and Rybczyk 2021). The temporal reduction of saltmarshes showed a trend similar to that in reclamation areas in southern Hangzhou Bay, indicating that saltmarshes loss was closely related to reclamation activities (Chen et al. 2022). Our results also demonstrated that more than half of the S. salsa marsh area has degraded to non-vegetated mudflats and encroached by anthropogenic reclamation (Table S2).

4.2 The C stocks and CAR

Halophytes directly contribute to the C cycle in the saltmarsh ecosystem because they are major OC producers (Wang et al. 2022a, b). In this study, we found that AGB and BGB in S. alterniflora were far higher than those in P. australis and S. salsa, consistent with previous observations that S. alterniflora had higher plant biomass than other types of saltmarshes (Zhou et al. 2015). The SEM results showed that the biomass was mainly regulated by vegetation characteristics (plant height and coverage) and soil physicochemical properties (pH, TK, NH4+, etc.). On the one hand, S. alterniflora was a C4 plant, while P. australis and S. salsa were C3 plants. Generally, C4 plants had higher photosynthetic efficiency than C3 plants (Shang et al. 2014; Zhou et al. 2007). On the other hand, the higher plant height and coverage (Liao et al. 2007; Zuo et al. 2012), the more vital propagation ability (Li et al. 2009; Zhao et al. 2020), and the longer growing season of S. alterniflora (Tang et al. 2022) contributed to its accumulation of plant C and expansion of living space. S. alterniflora was competitively superior to the native species in terms of nutrient uptake and utilization efficiency, and the competitive advantages enabled its C acquisition and storage (Li et al. 2021; Xie et al. 2019). From the proportion of plant C pool to the total C pool (Table S3), the vegetation types were differentiated in the plant C sequestration capacity. S. alterniflora owned the most extraordinary capacity in plant C sequestration, and it was 1.52- and 9.42-fold of P. australis and S. salsa, which supported the potential of S. alterniflora in plant C stabilization.

As the largest OC pool in saltmarshes, soil C cycling is complex and variable (Diao et al. 2020; Spencer et al. 2016). Elsey-Quirk et al. (2011) found that SOC stock was the central part of total C stock in saltmarsh ecosystem, which was also consistent with our findings. In our findings, SOC stock accounted for more than 80% of the total C stock in the vegetated saltmarshes, and it was up to 98% in S. salsa saltmarshes. There was no significant difference in SOC stock among different vegetations. However, for different soil depths, we found that the SOC stock of S. alterniflora at the 0–20 cm soil depth was higher than that of other vegetated saltmarshes, while it was significantly lower than that of P. australis at the 50–100 cm soil depth. We speculated that such a result was related to the distribution of plant roots in different vegetations. S. alterniflora has sufficiently proliferated roots distributed mainly in the surface soil layer, especially in the 0–20 cm depth, occasionally extending to 50–100 cm belowground (Cheng et al. 2006; Liu et al. 2017). In addition, the SOC stock in the newly developed S. alterniflora saltmarshes was similar to the mudflat (Liu et al. 2017; Wang and Wall 2010), indicating that the SOC increment in S. alterniflora saltmarshes took time to accumulate. Thus, the significantly lower SOC stock in S. alterniflora saltmarshes than that of P. australis saltmarshes in the deeper soil depth was partly owing to the short duration of S. alterniflora colonization.

The soil C stock in saltmarshes is mainly due to the rapid accumulation of OC-rich sediments and conditions that promote the preservation of OC (Miller et al. 2022; Xia et al. 2021a, b). It was estimated that the introduction of S. alterniflora increased the CAR seven-fold more than adjacent mudflats (Huang et al. 2018). In this study, the CAR of S. alterniflora was more than twice as much as the P. australis and S. salsa saltmarshes, further corroborating its strong C accumulation capability. First, S. alterniflora has higher plant biomass and litter yield than native plants, boosting plant C input to soils (Gao et al. 2016; Yang et al. 2016). Second, the several advantageous traits of S. alterniflora, such as densely growing canopies and roots, are beneficial for attenuating waves and maintaining sedimentation, thus sequestrating more C (Koch et al. 2009; Yang 2019). Third, the litter decomposition rate of S. alterniflora, particularly the root residues, was much slower than that of P. australis litter due to its lower litter quality (i.e., higher C: N ratio) (Duan et al. 2018). The increased C input, C sequestration, and slower litter decomposition derived from S. alterniflora reduced the C turnover in saltmarshes and promoted C accumulation (Xia et al. 2021a, b).

4.3 Expansion of S. alterniflora

S. alterniflora originated from the east coast of North America and the Gulf of Mexico (Cheng et al. 2006). The strong adaptability and tolerance of S. alterniflora, together with intentional introduction and plantation, enable it to be widely distributed in saltmarshes of different geographical and climatic regions (Zuo et al. 2012). As a typical halophyte, it is strongly tolerant to salinity and flooding (Daehler and Strong 1996). Since its introduction in 1979, S. alterniflora has been continuously expanding in the coastal areas of China (Mao et al. 2019; Zhang et al. 2017).

The expansion of S. alterniflora significantly altered C cycling in saltmarshes. The unique traits of S. alterniflora, such as high primary productivity, enriched nutrient utilization efficiency, and strong C captivate capacity, make it efficient in fixing C in both biomass and soils to strengthen coastal C sink (Choi and Wang 2004; Kaštovská and Šantrůčková, 2007; Kirwan and Mudd 2012; Tang et al. 2022). We found that, compared to native marsh types (P. australis and S. salsa), S. alterniflora possessed greater surface soil C stock, plant C stock, and CAR. In addition, the promotion on surface SOC stock showed a cumulative effect along with the S. alterniflora colonization duration and peaked in the 10th year (the peak increased the SOC stock by almost five times compared with the early stage). To summarize, the blue C sequestration capacity in Yancheng saltmarshes was enhanced along with the expansion of S. alterniflora.

The area of S. alterniflora had expanded by 287.5 km2 in Yancheng from 1982 to 2021. Along with the expansion of S. alterniflora, the proportion of S. alterniflora in the total area of Yancheng saltmarshes continuously increased (Fig. S5), resulting in strengthened C storage and C burial. The strongly positive relationships of the total plant C storage and C burial with the area of S. alterniflora confirmed that S. alterniflora drove the C sequestration in Yancheng saltmarshes over temporal scales, especially in the aspect of plant C storage and C burial (Fig. 6c, d). Greater CAR is usually associated with greater SOC stock, but the relationship is not always true in reality. The total SOC storage was correlated with the total area of Yancheng saltmarshes rather than the area of S. alterniflora (Fig. S6a, b) because no significant differences existed in the SOC stock (0–100 cm) among different vegetations. The increment of SOC stock requires a continuous high CAR for some time (Ouyang and Lee 2020). Thus, the future SOC stock of S. alterniflora would be more extensive than that of the other two native plants under such a great CAR. S. alterniflora is prospected to play an increasingly essential role as a C sink for its continuous expansion and promoting SOC storage.

The expansion of S. alterniflora in mudflats not only resulted in a substantial enhancement of the blue C sinks in saltmarshes but also initiated a succession of plant communities that altered the original saltmarsh ecosystems' blue C sinks. (Liu et al. 2017; Nie et al. 2022). The establishment of S. alterniflora at the low-elevation mudflats enhanced the accumulation of organic matter and increased the sediment elevation due to the strong ability to capture fine-grained sediments. This led to progressively shorter inundation and lower seawater input to the landward P. australis and S. salsa marshes, altering their plant communities (Redfield 1972; Wang et al. 2022a, b). The halophyte S. salsa was strongly dependent on the tidally driven episodic accumulation of salt, which killed its competitors (Cheng et al. 2006; Tang et al. 2022). The adaptability and preference of different halophytes to salt altered the competition between P. australis and S. salsa (Cao et al. 2022; Yan et al. 2022; Zhao et al. 1999), resulting in the modification of plant structure in Yancheng saltmarshes after S. alterniflora invasion (Fig. S7). Consequently, the desalination trend made it more competitive, causing it to migrate seaward and compress the S. salsa zone. The substitution of S. salsa by P. australis was also helpful in strengthening the blue C sinks in saltmarshes because P. australis had a significantly higher capacity in plant C stock than S. salsa (Wang et al. 2022a, b). This vegetation transition has critical implications for the long-term C sequestration potential of saltmarshes and their ability to act as natural C sinks. Further studies are necessary to understand the underlying mechanisms driving these plant community successions and their implications for coastal ecosystem management.

4.4 Implication

Our study provided novel insights into the succession of plant landscape patterns and the saltmarsh C sequestration in coastal wetlands of China. The colonization of S. alterniflora mainly occupied the non-vegetated mudflats (Mao et al. 2019). The shrinking zone of native S. salsa was not caused by direct competition of exotic S. alterniflora, but by the occupation of native P. australis, which was indirectly driven by the promoted soil sedimentation of S. alterniflora expansion to mudflats (Wang et al. 2022a, b).

The widespread and rapid expansion of S. alterniflora to mudflats along the coast of China promoted primary productivity and plant C inputs (Li et al. 2009; Liao et al. 2007), and slowed water flow that accelerated sediment accumulation (Liu et al. 2017), favoring C storage and C accumulation. S. alterniflora saltmarshes had 2.8- and 6.6-fold more total C stock and CAR than mudflats, respectively (Table S4). China has 5,300 km2 of mudflats (Wang et al. 2021a, b). There is a tremendous potential to increase the blue C sink in these mudflats. S. alterniflora has been regarded as an invasive species in China and human intervention has been implemented to avoid its broader expansion. However, S. alterniflora colonization also strengthened C sequestration, prevented the beaches from erosion, and did not compete with native species directly. Follow-up studies are needed to explore whether the carbon sequestration potential of S. alterniflora should be exploited to reach China’s carbon–neutral goal.

Developing environmental management strategies to balance the negative impacts of S. alterniflora and its strong capacity for C sequestration to attain carbon neutrality is a complex challenge. Firstly, it is needed to comprehensively reassess the invasive behavior of S. alterniflora by considering its capacity for blue C sinks and its adverse environmental impacts. Additionally, it may be beneficial to explore methods for utilizing the carbon captured by S. alterniflora in a sustainable way, such as through bioenergy production or carbon storage. Collaborative efforts, involving scientists, policymakers, and local communities, may be needed to develop efficient and sustainable management strategies that balance the benefits of C sequestration with the protection of native ecosystems. The adoption of such strategies will require significant investment in research, monitoring, and management efforts, but will ultimately be essential for the attainment of carbon neutrality and the protection of our environment.

5 Conclusion

This study systematically analyzed the SOC stock, the plant C stock, and the soil C accumulation rate associated with decades of marsh cover change in Yancheng, China. We found that S. alterniflora had an exceptionally high C sink capacity in its soil surface C storage, plant C storage, and C burial. In addition, the expansion of S. alterniflora contributed to an increase in the surface (0–20 cm) SOC stock, which had a cumulative effect with incrementing expansion years, and tended to be stable after peaking at the 10th year. Although Yancheng saltmarsh’s total area reduced by 43.09 km2 over the past four decades, the total C storage and C burial were strengthened to different degrees due to the high C sequestration of S. alterniflora. The colonization of S. alterniflora in mudflats increased the saltmarsh area and promoted carbon accumulation, therefore facilitating the expansion of the native P. australis. Largely expanded P. australis led to massive shrinkage of the S. salsa marsh area, which also offset the compressed living zone by land reclamation. The direct and indirect carbon gain by S. alterniflora invasion overcompensated carbon stock loss due to land reclamation and the areal loss of carbon-dense marsh of S. salsa to the marsh of P. australis, a species with a lower carbon accumulation rate (from 108.5 ± 5.00 g C m−2 yr−1 to 94.77 ± 7.80 g C m−2 yr−1). Our findings provide a theoretical basis for the C neutrality of Yancheng saltmarshes. The spatiotemporal assessment on the C sequestration of Yancheng saltmarshes is of great significance for better understanding of the blue C ecosystems of China.