Human activity has increasingly affected recent carbon accumulation in Zhanjiang mangrove wetland, South China

Summary Mangrove wetlands are an important component of blue carbon (C) ecosystems, although the anthropogenic impact on organic C accumulation rate (OCAR) in mangrove wetlands is not yet clear. Three sediment cores were collected from Zhanjiang Gaoqiao Mangrove Reserve in Southern China, dated by 210Pb and 137Cs, and physico-chemical parameters measured. Results show that the OCARs in mangroves and grasslands have significantly increased by 4.4 and 1.3 times, respectively, since 1950, which is consistent with the transformation of organic C sources and the increase of sedimentation rate. This increment is due to increased soil erosion and nutrient enrichment caused by land use change and the discharge of fertilizer runoff and aquaculture wastewater. This study provides clear evidence for understanding the changes in organic C accumulation processes during the Anthropocene and is conducive to promoting the realization of C peak and neutrality targets.


INTRODUCTION
Mangroves have high primary productivity and well-developed roots, which can promote the sedimentation of suspended solids.Oxygen deficiency and sulfate-rich conditions slow down carbon (C) mineralization, which can be preserved in mangroves for thousands of years. 1,2This makes mangroves the most C-rich forests in the tropics, 3 reducing greenhouse gas emissions and effectively mitigating the impacts of climate change. 4he C pool of mangroves consists of sediment C, living and dead biomass including leaves, stems, branches, litter, and roots over the short term. 5Sediment C is much larger than biomass C, accounting for about 49-98% of the total C stock of mangrove ecosystems. 3Globally, sediment organic C stock and accumulation rates in mangroves are spatially highly heterogeneous and are related to biological factors, geomorphology settings, hydrology, and physical and chemical parameters. 6,7ecent studies have found that human management has altered the physical and chemical properties and organic matter (OM) sources of mangrove sediments, gradually becoming a driving factor for the accumulation of organic C in mangroves. 8,9Over the past 50 years, mangroves have been large-scale reclaimed for agriculture and aquaculture, resulting in a loss of mangrove area. 10,11Moreover, when mangroves are converted to other land use types, they not only lose the ability to store C from the atmosphere but also lead to the removal of C from the sediment in the form of carbon dioxide (CO 2 ). 12 At the same time, along with the discharge of fertilizer runoff and aquaculture wastewater, nutrient overloads occurred. 13,14However, the impact of eutrophication on organic C accumulation is still unclear.Nutrient enrichment has been suggested in some regions to increase the deposition of phytoplankton and benthic microalgae, 15 as well as stimulate microbial respiration and mineralization of OM, 16 which reduced the ability of local mangrove communities to retain organic C. 17 But for nutrient-limited mangrove ecosystems, nitrogen (N) and phosphorus (P) inputs are essential to promote sediment C inputs by stimulating plant growth. 18Eutrophication not only changes the dynamics of organic C by affecting plant biomass and microbial processes, but also drives the marsh loss due to the reduction of geomorphic stability. 19Other human disturbances, such as river damming, reservoir construction, and deforestation, can also alter the salinity and supply of sediment. 20,21In summary, the interactions between multiple anthropogenic activities are very complex; thus, understanding the changes in organic C accumulation rate (OCAR) and sources is fundamental to predicting the C sequestration potential in mangrove wetlands.
A large area of mangrove wetlands is distributed along the coast of Southern China.Zhanjiang City in Guangdong Province is home to China's largest mangrove nature reserve.Over the past few decades, 5,587 ha of aquaculture ponds have been spread around its edges. 22t the end of each farming cycle, nutrient-rich pond sediments are discharged through the gates into the adjacent mangroves, making them a sink for nutrients and pollutants and causing serious environmental risks. 22,23e hypothesize that human activity, including regional aquaculture and land use change, has affected the C accumulation rate of Zhanjiang mangrove wetlands.This impact has increased with urbanization and aquaculture development over the past decades.To test such a

Chronology and sedimentation rates
The excess 210 Pb ( 210 Pb ex ) values in GQ20-Man1 and GQ20-GL1 varied exponentially with depth (R 2 = 0.72 and 0.95, respectively) (Figure 1).The maximum 210 Pb ex values of GQ20-Man1 and GQ20-GL1 both occurred at 0.5 cm, on the surface layer, while the maximum value at GQ20-Man2 appeared at 18.5 cm.Vertical mixing occurred in the GQ20-Man2 with an irregular jagged change.The history was traced back to 1874 in GQ20-Man1, 1905 in GQ20-Man2, and 1887 in GQ20-GL1.In general, the first peak of 137 Cs occurred in 1963, and the differences between the ages calculated in this study using 210 Pb ex were within the error range of G10 years (Figure 1).
The sedimentation rates of GQ20-Man1 ranged from 0.04 to 1.81 cm yr À1 , with an average of 0.57 G 0.39 cm yr À1 .The sedimentation rates of GQ20-Man2 were higher than that of GQ20-Man1, with a maximum value of 3.04 cm yr À1 and an average value of 0.72 G 0.65 cm yr À1 .At GQ20-GL1, the sedimentation rate was the lowest, with a mean value of 0.30 G 0.13 cm yr À1
The Pearson's correlation analysis showed that TOC content in mangrove sediment was significantly negatively correlated with DBD and positively correlated with SWC, TN and clay content (Figure 3).The correlation between TOC and TN was the strongest, followed by SWC and DBD, and the correlation with clay content was the weakest.No significant correlation was found between TOC concentration and sand  content.In GQ20-Man2, TOC also showed a (significant) positive and negative correlation with TP concentration and silt content, respectively.In GQ20-GL1, the concentration of TOC showed a significantly stronger correlation with TN content, DBD and SWC.
Overall, downcore profiles of d 13 C and C/N increased with depth, while d 15 N showed a decreased tendency in mangroves.Since 1950, the d 13 C values of mangroves have increased and have been depleted since 1980.On the contrary, the d 15 N values have increased since 1950, especially after 1980.Both d 13 C and d 15 N in GQ20-GL1 exhibited oscillations and became slightly enriched after 1950.As for C/N, the values after 1950 are relatively low compared to those observed in the bottom among three sites.
The source mixing model based on d 13 C revealed that the sedimentary OM is mainly derived from the mangrove production, followed by terrestrial sources, while marine algae contribution was the lowest (Figure 4).The contribution of mangrove production in GQ20-Man1 has increased from 53% at the bottom to 90% at the surface, with a decrease in OM from both terrestrial and marine sources, especially since 1980.The contribution of GQ20-Man2 to mangrove production showed a trend of first increasing and then decreasing over time.Taking 1950 as the breaking point, the contribution of OM from terrestrial sources has increased recently, reaching 30% in the early 1990s.OM from mangrovederived sources decreased between 1950 and 1980, followed by a slight increase.
C/N/P accumulation rates at three cores showed an increasing trend since 1950, particularly after 1980, which is consistent with the increase of population and fertilizer consumption of Liangjiang City (Figure 5).Kruskal-Wallis ANOVA analysis was used to determine C/N/P accumulation rates before 1950, between 1950 and 1980, and after 1980.Results indicated significant differences (p < 0.05): in fact, compared to the period 1900-1950, the C/N/P accumulation rates in the mangrove region increased by 5.3, 4.8, and 8.6 times, respectively, after 1980.In comparison, the OCAR of grassland areas increased by 1.4 times, TNAR by 2.9 times, while TPAR decreased (0.7 times), indicating that mangroves accumulate more organic C and nutrients than grassland ecosystems.There were significant differences in OCAR and TNAR among the three sites (p < 0.05), except TPAR since 1950.

Chronology and sedimentation rates
The values of 210 Pb ex in GQ20-Man1 and GQ20-GL1 varied exponentially with depth (Figure 2), indicating that the sediment environment was stable. 24Due to enhanced or dilutive effects from changes in sediment supply, and some biotic or physical disturbance, the constant rate of supply (CRS) model was used for calculation.This is more reasonable because the CRS model is calculated based on the total accumulation of 210 Pb ex and is not sensitive to the effects of vertical mixing. 25,26 20,23Todos os Santos Bay and Itapessoca estuarine, Brazil; 16,17 Qinglan Bay, China; 24 Bintuni Bay, Indonesia. 25 The average sedimentation rate of Gaoqiao mangrove area is higher than that at a global scale (0.28 cm yr À1 ), 27 and also higher than the average of mangroves affected by urban effluents (0.55 G 0.20 cm yr À1 ). 28Since 1950, the sedimentation rate of Gaoqiao mangrove wetland has been increasing.On the mesoscale, the sedimentation rate in this ecosystem is determined by hydrodynamics and sediment supply. 29and use changes in catchment areas can alter the transport of suspended sediment and particulate OM.The reduction of forest area upstream of rivers and the expansion of agriculture may lead to increased soil erosion, thereby increasing the sedimentation rate of mangroves. 30,31In addition, dense aquaculture ponds are distributed outside the Gaoqiao mangrove wetland; consequently, after the seasons shrimp harvest, organic materials such as fish bait and animal waste are discharged into the mangroves area, 32 thus contributing to the increase in sedimentation rate. 33Mangroves in Mexico showed an exponential increase in sediment mass accumulation rates due to the intensification of continental erosion caused by land use and population growth. 26In Puerto Rico, mangrove sedimentation rate at Can ˜o Martin Pen ˜a, the most urbanized area, has varied significantly in recent decades compared to historical decades. 34

Site differences of organic C accumulation
The spatial variability of sediment organic C stocks and accumulation rates in mangroves is very large. 2,27][37] Sediment properties are considered the main factor controlling TOC content in mangrove sediment, as well as TN content, SWC and sediment texture. 38,39TN in sediments is closely coupled with TOC and influences the input of OM by affecting plant growth. 39,40An increase in SWC, which generally mirrors a lower DBD, 41 will inhibit the diffusion and penetration of oxygen and affect microbial metabolism. 38This leads to a decrease in the redox reaction potential and decomposition rate, resulting in an increase of TOC content in sediment. 42The physical preservation of TOC by clay particles can slow down the decomposition of microorganisms, 38 but the effect of clay content on TOC was weaker than that of SWC and TN content in Gaoqiao mangrove wetlands.The role of other sediment properties such as pH and salinity cannot be ignored and further exploration is needed. 43ften, elevated primary productivity can lead to higher debris input and fine root turnover rates. 44GQ20-Man1 is located in the high-value area of gross primary production (GPP), while GQ20-Man2 has a lower GPP. 44Primary productivity is often positively correlated with species age. 45Mangroves of GQ20-Man1 are older than GQ20-Man2, with longer-lasting inputs of litter and roots, which could promote organic C accumulation. 46nlike terrestrial ecosystems, the Gaoqiao mangrove wetland belongs to the estuarine type, with GQ20-Man1 and GQ20-Man2 distributed in the intertidal zone, receiving supply from tidal and river sediments (Figure 6).High OCARs are typically found in mature forests in river deltas and severely affected catchments 47 due to high primary productivity, allochthonous source inputs, and sedimentation rates. 48,49ecause of the higher sedimentation rates of GQ20-Man2, the OCAR was greater than GQ20-Man1, while the organic C stock was lower than GQ20-Man1 due to the dilution effect. 50A significant linear positive correlation between OCAR and mass accumulation rate (MAR) is found, whose correlation coefficient is greater than that between OCAR and TOC (Figure S1).It indicated that the sedimentation rate of the Gaoqiao mangrove is a more important factor affecting the OCAR than the TOC. 51

Identification of sediment organic C sources
The sources of OM in mangroves typically include local production and external inputs.The former mainly comes from the litter, roots, and benthic microalgae of mangroves, while the latter comes from the phytoplankton and suspended sediments carried by tidal and river inputs. 38,52In general, C3 plants (including those characterizing mangroves) have a d 13 C value from À32 to À24&, a d 15 N from À10& to 10&, and a C/N > 20, whereas marine algae have a d 13 C between À16 and À23&, a d 15 N in the range 6&-11&, and a C/N ratio between 4 and 10. 49,53,54 Therefore, the combination of d 13 C, d 15 N, and C/N can effectively identify the source of coastal sediments.The values we collected in Gaoqiao reflected that OM is mainly derived from C3 vegetation, consistent with previous reports. 43edimentary d 15 N has also been used to reconstruct the impact and historical changes of human activities on N accumulation. 17,55For example, human sewage input, fertilizers, feces, and other allochthonous OM sources are usually enriched in the heavier 15 N. 56 In addition to the effects of N source input, d 15 N signatures are often determined by complex biogeochemistry processes (including N fixation, remineralization, nitrification, volatilization, and denitrification). 57Therefore, the increase in d 15 N in recent sediments may be related to increased denitrification and inputs from aquaculture wastewater and farm runoff. 28,56,58,59Relative to mangroves, grasslands had lower d 15 N and stable vertical variation, reflecting the weaker influence of human activities on GQ20-GL1.
In recent sediments, C/N of GQ20-Man1 has begun to decrease, possibly due to N adsorption from human sources into the sediment. 60he low value of C/N may also be related to the transformation of OM sources, passive leaching and microbial decomposition during early diagenesis, resulting in a greater loss of C than N. 61,62 Afterward, there was almost no change in C/N, indicating that C and N degradation pathways were consistent at the same rate during this period. 63At the bottom, the increased C/N of GQ20-Man2 and GQ20-GL1 may be attributed to the preferential loss of N-rich OM. 60 To sum up, compared with C/N and d 15 N, the ''conservative'' of d 13 C in identifying sources is not inhibited by the decomposition of OM. 53,64 An endmember mixing model was established using d 13 C to quantitatively estimate the contributions of terrestrial, mangrove, and algae sources to mangrove sediments.The slight positive deviation of the d 13 C value between 1950 and 1980 indicates an increase in the contribution of terrestrial and algae sources, possibly induced by deforestation and agricultural expansion. 65,66After 1980, increased nutrients indicated by enriched d 15 N promoted litter and root deposition of mangroves, 40,67,68 resulting in lower d 13 C values.[71]

Human impacts on trends of C accumulation rates
The TPAR of Gaoqiao mangrove is higher than the global average of 0.5 g m À2 $yr À1 , while the TNAR is lower than the global average of 8.9 g m À2 $yr À1 . 72However, in recent decades, the TNAR of GQ20-Man1 reached 18.8 g m À2 $yr À1 .High nutrient fluxes are related to the discharge of farmland fertilizers and aquaculture wastewater. 23After 1980, the application of chemical fertilizers in Lianjiang City increased sharply, and the government advocated for aquaculture.Therefore, reclamation aquaculture became very popular in the 1990s.The sediment discharged from shrimp ponds carries unabsorbed nutrients into the mangrove forest, making it a sink of nutrients. 73he OCAR in mangroves affected by human activities reached 1,023 g m À2 $yr À1 in the Cubata ˜o mangroves in Brazil, 15 but there are also mangroves with much lower values worldwide (Table 1). 17,74This is because the input of nutrients simultaneously stimulates the production and decomposition of OM in the sediment, and the impact on mangroves depends on the degree of nutrient enrichment and the nature of the sediments themselves. 9However, some studies have shown that nutrient availability hampers C storage by increasing the availability of mineral nutrients for decomposers and altering litter composition. 75However, N addition can also induce changes in microbial communities and sediment acidification to inhibit the production of oxidase and microbial metabolism, thereby delaying sediment C decomposition. 76In addition, nutrient enrichment promotes root and branch growth and expansion in leaf area, 40,67 increases aboveground and belowground biomass, fixing more C from the atmosphere to the soil. 77,78Nutrients can also stimulate the growth of algae, which can be deposited into sediment along with the original mangrove litter. 15,79Thus, the increased OCAR in nutrient-limited Gaoqiao mangrove sediments can be partially attributed to the contribution of nutrient inputs to organic C sequestration, as indicated by the positive correlation among TOC, TN, and TP (Figure 2).The OM inputs, mainly plant or root biomass, to sediment are mostly composed of labile, particulate fractions that are more prone to a rapid decomposition.Thus, although the development of mangroves increased TOC, it also resulted in greater amounts of labile organic C fractions in mangroves 80 ; as a consequence, it is necessary to improve the research on the mechanism of sediment organic C stabilization, which is a key path to increase the blue C sinks in coastal zones.Eutrophication has also been proven to contribute to marsh loss, 19 in addition to focusing on the dynamics of organic C accumulation, the coastal landscape conversion cannot be ignored.
In conclusion, identifying the C sequestration potential of mangroves plays an important role in implementing ecological restoration, formulating management policies, and achieving the ''dual C'' goals.This study combined the 210 Pb and 137 Cs dating methods and then used d 13 C and d 15 N, to investigate the changes in the rate and source of organic C accumulation in the Zhanjiang mangrove wetland over the past century, as well as the underlying mechanisms.
The spatial distribution characteristics of organic C accumulation in mangrove wetlands are jointly determined by primary productivity, sedimentation rate, and sediment properties.After 1950, the OCAR initially increased and sharply increased after 1980, consistently with the variations in population and agricultural development in Zhanjiang City.At the same time, the agricultural expansion and reclamation in the adjacent areas had impacts on the source of OM.Since 1980, the input of nutrients has promoted the development of mangrove biomass, increasing the contribution of OM from mangrove sources to sediment.However, compared to organic C input, the high sedimentation rate caused by land use transformation had a more significant effect on organic C sequestration.Finally, investigating the stabilization mechanisms allowing a long-term preservation of OM in sediment is of great significance for increasing the C sequestration capacity of coastal wetlands.

Limitations of the study
Although stable isotopes have been widely used to quantify the relative contribution of external and local organic C sources in coastal ecosystems, there is a cross overlap between the d 13 C values of mangrove and terrestrial plants.In subsequent research, it would be useful to separate multiple inputs from mangrove sediments by combining specific taxonomic biomarkers, such as lipids and lignin, unique to mangroves.Moreover, collecting more samples and exploring the effects of other soil properties, climate, and human activities on organic C accumulation by combining random forest model, structural equation model, and redundancy analysis would contribute to better understand the role of mangrove environments to (C, N, and P) global biogeochemical cycles.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:  End-member mixture models are used to quantitatively calculate the relative contributions of different potential end members to the sample.IsoSource can calculate n isotopes and > n+1 end members.It is based on a stable isotope mass conservation model; when the source is > n+1, all possible combinations of each source are calculated by superposition in a given increment (1% in this study).The weighted average of each group and the measured value of the mixture are compared.If it is within the tolerance, it is considered a possible solution.1][92] The algae source used the average d 13 C value of Gaoqiao phytoplankton and benthic microalgae (À20.3&).The contribution percentage of each C source was calculated as follows: 1) where M (Mangrove), T (Terrestrial), and A (Algae) are the potential OM end-members and f represents the contributory percentage of each endmember.

Age dating and chronology
210 Pb and 137 Cs age dating was determined by a g spectrum analysis system, mainly composed of high-purity germanium well detector (Ortec HPGe GWL) and Ortec 919 spectrum controller (EG&G Ortec G Company).The standard samples of 210 Pb were compared by the University of Liverpool, and the standard samples of 137 Cs and 226 Ra were provided by the China Institute of Atomic Energy.The CRS model was used for the calculations, according to the following equations: t = 1 = l 3 lnðC 0 = C Z Þ (Equation 3) SR = Dz=Dt (Equation 4) Then, the organic C stock and OCAR were calculated by the following equations: OCAR À g , m À 2 , yr À 1 Á = TOC À g , kg À 1 Á 3 DBD À g , cm À 3 Á 3 SR À cm , yr À 1 Á 3 10 (Equation 5) 3 HðcmÞ 3 100 (Equation 6) where t is the age of deposition, l is the decay constant (0.03114 a À1 ), C 0 is the total 210 Pb ex input (Bq$cm À2 ) in the sediment cores, C z is the cumulative 210 Pb ex input (Bq$cm À2 ) in the core from Z cm depth to the bottom layer, SR is the sedimentation rate in each layer, H is the sediment thickness.Finally, 10 and 100 are both unit conversion factors.

QUANTIFICATION AND STATISTICAL ANALYSIS
The results obtained were tested for normality using the Shapiro-Wilk test.Differences in sediment physical and chemical properties between mangroves and grasslands were determined using the independent samples T-test for indicators that were approximately normally distributed; otherwise, the Mann-Whitney U Test was used.The correlation between TOC and other properties was analyzed using Person linear correlation.These analyses were conducted at a 95% confidence level.

Figure 1 .
Figure 1.Dating results of the sediment cores from wetlands and grassland 210 Pb ex (green circles) and 137 Cs (purple circles) activities depth profiles, age-depth model (gray circle) and sedimentation rates (blue circles) in GQ20-Man1 (A), GQ20-Man2 (B), and GQ20-GL1 (C).Chronologies were calculated using the constant rate of supply (CRS) model.

Figure 3 .
Figure 3. Relationships among sediment properties in mangrove and grassland (A) GQ20-Man1, (B) GQ20-Man2 and (C) GQ20-GL1.*, **, and *** represent significance of p < 0.05, p < 0.01, and p < 0.001, respectively.The size and color of the circles, as well as the number in the cells, represent the values of the correlation coefficients.

Figure 4 .
Figure 4. Carbon (d 13 C) and nitrogen (d 15 N) stable isotope signatures and C/N ratio in the sediment cores (A) Vertical profiles of d 13 C, d 15 N and C/N in sediment cores.(B) Comparison of d 13 C and d 15 N values found in the Gaoqiao area and other mangrove sites: Cochin estuary, India; 20,23 Todos os Santos Bay and Itapessoca estuarine, Brazil; 16,17 Qinglan Bay, China; 24 Bintuni Bay, Indonesia. 25Data are reported as average G SD.The ranges of the d 13 C and d 15 N values for different OM sources were established based on literature. 23,26C3 plants have a d 13 C value from À32 to À24&, a d 15 N from À10& to 10&, whereas marine algae have a d 13 C between À16 and À23&, a d 15 N in the range 6&-11&.(C) Variations in the contributions of mangrove-derived OM, algae-derived OM, and terrestrial OM in mangrove sediment cores.
Figure 4. Carbon (d 13 C) and nitrogen (d 15 N) stable isotope signatures and C/N ratio in the sediment cores (A) Vertical profiles of d 13 C, d 15 N and C/N in sediment cores.(B) Comparison of d 13 C and d 15 N values found in the Gaoqiao area and other mangrove sites: Cochin estuary, India; 20,23 Todos os Santos Bay and Itapessoca estuarine, Brazil; 16,17 Qinglan Bay, China; 24 Bintuni Bay, Indonesia. 25Data are reported as average G SD.The ranges of the d 13 C and d 15 N values for different OM sources were established based on literature. 23,26C3 plants have a d 13 C value from À32 to À24&, a d 15 N from À10& to 10&, whereas marine algae have a d 13 C between À16 and À23&, a d 15 N in the range 6&-11&.(C) Variations in the contributions of mangrove-derived OM, algae-derived OM, and terrestrial OM in mangrove sediment cores.

Figure 5 .
Figure 5. Changes in C/N/P accumulation rates (g$m À2 $yr À1 ) of sediment cores The straight and dotted lines represent the population and consumption of chemical fertilizers changes in Lianjiang City, respectively (see also Figure S1).

Figure 6 .
Figure 6.Map of the core area of Zhanjiang Mangrove Nature Reserve (ZMNR), Guangdong Province, China The dots are the sampling locations.

Table 1 .
C/N/P accumulation rates in mangrove wetlands around the worldRegionOCAR (g$m À2 $yr À1 ) TNAR (g$m À2 $yr À1 ) TPAR (g$m À2 $yr À1 ) Year Reference ZMNR, Zhanjiang Mangrove National Nature Reserve; NA, not available; NDL, non-anthropogenically disturbed locations; IDL, included anthropogenic disturbed locations.Data are reported as average G SD.TOC and TN concentration was measured using a multi N/C 3100 and an Elementar Vario Macro cube (Hesse, Germany), respectively.Samples were treated with 2 mol/L HCl for full reaction to remove inorganic C. The values of C/N ratios were calculated as the atomic ratio of the TOC to TN. TP was measured after a mixture-acid digestion (HNO 3 /H 2 O 2 /HClO 4 , 5/1/1, v/v) using an Atomic Emission Spectroscopy with Inductively Coupled Plasma (ICP-AES Prodigy 7, Teledune Leeman Labs, USA).The d 13 C and d 15 N values were measured using MAT 253 isotope ratio mass spectrometer (Thermo Electron, Bremen, Germany) following samples acidification.Stable isotope ratios were calculated as d 13 C = [(d 13 C sample /d 13 C standard )-1] 3 1000 and d 15 N = [(d 15 N sample /d 15 N standard )-1] 3 1000.Vienna Pee Dee Belemnite and atmospheric air (Air) were used as a reference for d 13 C and d 15 N, respectively.The analytical precision for both d 13 C and d 15 N was G0.10&.