Guangxi has a typical subtropical marine monsoon climate. In this area, the mean annual air temperature is 22.5°C, the coldest month is January, with an average temperature of 14.3°C (Li et al., 2015), the annual precipitation is 1,658 mm, and the mean sunshine duration is 1,673 h (Tian et al., 2021). Guangxi has approximately 8,715 ha of mangrove forests, and the main mangrove species are Sonneratia apetala, Aegiceras corniculatum, Kandelia obovata, Avicennia marina, Rhizophora stylosa, and Bruguiera gymnorrhiza (Zheng and Takeuchi, 2020). Pearl Bay is a semi-enclosed sea bay located in the city of Fangcheng Harbor, Guangxi, China. Pearl Bay covers an area of 94 km² and has a mouth length of 3.5 km. Two freshwater-dominated rivers flow into Pearl Bay: the Huangzhu and Xinlu Rivers (Fan et al., 2017). The diurnal average tide height in Pearl Bay is 2.24 m, the highest tide is 5.00 m (Chen et al., 2021), with an average salinity of 23.1‰ (Xu et al., 2020). The study area was a creek discharging into Pearl Bay ( Figure 1A ).

The natural mangrove forest in the sampling site is distributed in the intertidal zone, with an average tree age of 50 years. Halophila beccarii (Seagrass bed) are found on tidal flats at the edge of the mangrove forest. Some pioneering Avicennia trees grow in the mudflats in the low tide area. Mangrove species in the sampling site were A. corniculatum, B. gymnorrhiza, A. marina, and K. obovata. Underground aquaculture systems were built near the study area by burying plastic pipes in mud (Chen et al., 2021).

Two sediment cores were collected from a creek in a natural mangrove forest on the northeast side of Pearl Bay ( Figure 1A ) at GX1 (108°14′28″E, 21°37′2″N) and GX2 (108°14′21″E, 21°37′3″N) ( Figure 1B ) on June 30, 2021 using a gouge auger sampler (diameter: 6 cm; length: 100 cm) with minimal compaction. The sediment cores GX1 and GX2 were 88 and 96 cm long, respectively. At the sampling sites, the layers of sediment cores were sliced at every 2 cm, and the sub-samples were stored in polyethylene bags at low temperature (4°C). Drying of sediment samples at 45°C for two days (until constant weight) was completed within one week after sampling. The difference in mass of the sediment before and after drying in a 105°C oven was divided by the dry weight of the sediment to obtain the sediment moisture content, expressed as percentage dry weight.

Each sub-sample was treated with 20 ml H₂O₂ (40%) for 24 h to remove OM, followed by thorough rinsing with deionized water. Then, 10 ml of 1 M HCl was added to the residue to remove inorganic matter, this was allowed to react for 2 h until no more bubbles were produced. After the reaction was complete, the samples were rinsed with deionized water until the pH of the wash water was neutral. Tetrasodium pyrophosphate (2%) was used to disperse the sub-samples, and grain size was then determined using the particle size analyzer Mastersizer 2,000 (Malvern, England), which measures particle size ranging from 0.02 μm to 2,000 μm. Grain size was divided into three categories, namely, clay (< 4 μm), silt (4–63 μm), and sand (63–2000 μm). The lithology of the sub-sample was determined using the Shepard triangle method (Shepard, 1954). The mean grain size, standard deviation, skewness, and kurtosis of the sub-samples were calculated using the calculation procedure created by Blott and Pye (2001).

The freeze-dried sub-samples were treated with HCl (1 M) to eliminate inorganic carbon until no bubbles were produced. The sediments were then dried in an oven, homogenized, and analyzed for TN and TOC using an elemental analyzer (EA3000, EuroVector, Italy). The molar ratio of TOC and TN was calculated and henceforth known as C/N ratio.

Stable carbon isotope ratios were analyzed using a stable isotope mass spectrometer (Delta V + EA Isolink, Thermo Fisher Scientific, USA). The C/N ratios and δ¹³C of sediments in mangrove ecosystems have been extensively used to identify the sources of OM in mangrove sediments (Xiong et al., 2018b; Sasmito et al., 2020). OM from terrestrial C₃ plants, such as mangroves, with abundant cellulose exhibit a high C/N ratio (> 12) and low δ¹³C values (< -21‰), whereas OM from marine end-members, such as phytoplankton, with abundant protein have lower C/N ratio (5–7) and higher δ¹³C values (> -23‰) (Meyers, 1994; Lamb et al., 2006).

Approximately 5 g of the freeze-dried sub-samples were weighed and placed into sample tubes, which were then compacted and sealed with paraffin film for 3 weeks to establish equilibrium for ²¹⁰Pb activities. Radioisotope analysis of the sub-samples was performed using an HPGe Gamma Spectrometer (ORTEC-GWL-120-15, EG&G, USA), with IAEA-RGU ore powder as the standard reference material. The specific activity of ²¹⁰Pb was calculated using the energy peak at 46.5 keV, and the specific activity of ²²⁶Ra was calculated using the energy peaks of its progeny ²¹⁴Pb (351.9 keV) and ²¹⁴Bi (609.2 keV). The difference in specific activity between ²¹⁰Pb and ²²⁶Ra represents the excess ²¹⁰Pb_ex specific activity. The statistical errors for ²¹⁰Pb and ²²⁶Ra were <20% and <10%, respectively (at a 95% confidence level). The vertical distribution of excess ²¹⁰Pb specific activities was used to calculate the sedimentation rate and establish an age model for each core.

The sediment samples were subjected to the CuO oxidation method by Hedges and Ertel (1982) to extract lignin-derived phenol. First, 0.5 g of dry sediment and 1.0 g of CuO were weighed into a 25 mL Polytetrafluoroethylene (PTFE) reaction minibomb. In a glove bag filled with N₂, 10 mL of 2 M oxygen-free NaOH (N₂ was blown into the solution for 5 min) was added into the PTFE vessels. The PTFE vessels were placed in a muffle furnace and heated at 170°C for 3 h. The reaction vessels were then cooled. The solution was washed with 1 M NaOH and then placed in a 50 ml centrifuge tube and centrifuged. This step was repeated three times. The lignin in the supernatant was extracted three times using C₄H₈O₂ in a separator funnel. Anhydrous Na₂SO₄ was added to the extract to remove water, and the extract was filtered using filter paper. The extract solution was added, with 100 μL of ethyl vanillin as the internal standard, and then concentrated to 1–2 ml using a rotary evaporator. N₂ was blown into the remaining concentrated solution to dry the solution and obtain solid oxidation products. An equal volume of C₅H₅N and C₈H₁₈F₃NOSi₂ with 10% C₃H₉ClSi were added into the oxidation products and heated at 90°C for derivation purposes. Gas chromatography (GC-2010 Plus, SHIMADZU, Japan) was used to elucidate the lignin monomers. The heating procedure of the GC column was from 100°C to 300°C at a rate of 4°C min^-1. The concentration of each lignin phenol was determined based on the retention time of each lignin and the internal standard.

Lignin is a good indicator of terrigenous OM and is widely utilized in the environment of continental shelf, estuarine, lake, peatland, and salt marsh (Lourençato et al., 2019). Lignin phenol is one of the main components of plant-derived OM, which can indicate the vegetation composition of mangroves due to its unique biogeochemical characteristics (Lowman et al., 2021; Xia et al., 2023). The value of Λ6 is the sum of syringyl and vanillin phenols, whereas the value of Λ8 is the sum of Λ6 and cinnamyl phenols. The ratio of syringyl to vanillyl phenols (S/V) can be used to distinguish gymnosperms (S/V is approximately 0) and angiosperms (S/V > 0), and the ratio of cinnamyl to vanillyl phenols (C/V) can be used to distinguish woody (C/V is approximately 0) and non-woody (C/V > 0) tissues (Hedges and Mann, 1979). Furthermore, Tareq et al. (2004) proposed the lignin–phenol vegetation index (LPVI) to accurately determine the source of terrestrial OM; the ranges of LPVI values of the woody and non-woody tissues of gymnosperms are 1 and 12–17, respectively, and the ranges of LPVI values of woody and non-woody tissues of angiosperms are 67–415 and 378–2782, respectively. In addition, the acid–aldehyde ratios of syringyl [(Ad/Al)s] and vanillyl [(Ad/Al)s] phenols can be used to indicate the degree of lignin degradation under microbial action. In general, the (Ad/Al)s and (Ad/Al)_V of fresh plant tissues were less than 0.14 and 0.3, respectively. If (Ad/Al)s is greater than 0.16 and (Ad/Al)v is greater than 0.6, lignin is considered to have undergone high degradation (Hedges et al., 1988).

The OC storage in the sediment was calculated as follows: (1) dry bulk density (D_dry ) is equal to the ratio of dry subsample mass (M_dry ) to wet sample volume (V_wet ) (Equation 1), (2) calculation of the amount of TOC based on TOC percentage and D_dry of each layer (Equation 2), and (3) SOC storage is the ratio of the TOC_layer of each layer to the cross-sectional area of each layer (Equation 3).

where M_dry is the mass of 2 cm^-3 sub-sample after drying, V_wet is 2 cm^-3, H is the thickness of each sediment layer (2 cm), and A is the cross-sectional area of the sample column (28.3 cm²). TOC_layer is the TOC content of each sediment layer ( Supplementary Table S2 ).

The sources of OM fractions in the forms of mangrove and terrestrial and marine fractions in the mangrove sediment were calculated using the SIMMR software package. The detailed calculation method and process of the SIMMR software package were described by Parnell and Inger (2016). The δ¹³C and C/N values of mangrove leaves of different tree species in Qinzhou City (Guangdong Province) were used to represent the end-members of mangrove plants, in which the average value of δ¹³C is -28.72 ± 0.78‰, and the average value of C/N is 37.41 ± 9.65 (Xia et al., 2015; Meng et al., 2016). As the collected data were exclusively obtained from mangrove forests in Guangxi, characterized by the same mangrove species, these data adequately represent the end-member value of mangrove OM. Xia et al. (2015) found that in areas with similar drainage and material sources, the C/N and δ¹³C values of river sediments are generally same. Therefore, this study used the river δ¹³C (-24.1 ± 0.6‰) and C/N (12.8 ± 2.1) data provided by Xia et al. (2015) as the end-member values for river OM. The average values of δ¹³C and C/N of phytoplankton in the northern South China Sea of -16.10 ± 0.80‰ and 6.50 ± 0.10, respectively (Hou, 2009), were used as the end-members for phytoplankton OM fraction. The specific data and sources of each end-member value are shown in Supplementary Table S1 .

Given that not all parameters conform to the normal distribution, the Mann–Whitney U test was used to compare the differences in measured parameters among different cores. The correlation between different parameters along the cores was determined using Spearman’s correlation analysis. IBM SPSS Statistics software was used for statistical analysis of the data, and a p value of 0.05 was used to determine its significance. Assessment of the significance of changes in C/N and δ¹³C values in sediments were based on variance changes using the F-test (Sánchez et al., 2023). Origin Pro 2022b was used to draw and analyze the data. The SIMMR software package was used to establish an end-member mixing model based on C/N values and δ¹³C value to analyze the relative contributions of different OM sources (Parnell et al., 2013).

Based on the SERAC package provided by Bruel and Sabatier (2020), three models, namely, the constant flux constant sedimentation (CFCS), the constant initial concentration (CIC), and the constant rate of supply (CRS), were used to calculate the sedimentation rate and age patterns along the GX1 and GX2 cores. The total ²¹⁰Pb specific activity (75.82–179.67 Bq kg^-1) along the two cores is greater than that of ²²⁶Ra (32.77–60.36 Bq kg^-1). The depth of the two cores, GX1 and GX2, has not reached the background value of ²¹⁰Pb in this area; thus, the CRS model is not applicable. In addition, the reliability of the calculated results of the CIC model is poor. Thus, the ages along the GX1 and GX2 cores were calculated using the CFCS model. The ²¹⁰Pb activity of GX1 has a high degree of linear fit (r² = 0.87) to the depth profile ( Figure 2A ). The pattern observed along GX2 is consistent with a stepped pattern of ²¹⁰Pb activity ( Figure 2B ); therefore, the age model along GX2 was divided into two parts, 0–71 and 71–95 cm, with a fit of 0.67 and 0.89, respectively. The mean sedimentation rate along GX1 was 17.7 ± 2.9 mm yr^-1, and the ages ranged between 1972 and 2021 ( Figure 2A ). The deposition rates along GX2 were 80 ± 22.4 and 17.3 ± 5.4 mm yr^-1 at 95–71 and 71–0 cm, and the corresponding ages were 1977–1980 and 1980–2021, respectively ( Figure 2B ).

The average grain size, sorting coefficient, kurtosis, and skewness of the sediments in the study area are 4.93±0.49, 2.28±0.18, 0.95±0.12, and 0.37±0.12, respectively, with the variation trends shown in Figures 3A–D . The following formula is used to calculate particle size. The larger the calculated φ value, the smaller the particle diameter D.

where X is the normalized particle size, and D is the particle diameter. φ values decreased from 5.33 φ (in 1964) to 4.74 φ (in 2001) along GX1 and decreased from 5.99 φ (in 1963) to 4.64 φ (in 2001) along GX2 ( Figure 3A ), indicating increased particle size from around 1963 to 2001. No evident change was observed in the trend of average particle size along both cores from 2000 to 2020. According to the Shepard method (Shepard, 1954), three lithologies exist in the GX1 and GX2 sites ( Figure 4A ): silty-sand, sandy-silt, and clayey-silt. Both GX1 and GX2 showed a silty-sand type (more sand) in the surface layer and a sandy-silt type (more silt) in the deeper layer. Averages of 39.31% ± 14.32% sand, 47.43% ± 12.25% silt, and 13.27% ± 2.87% clay were found throughout 1963–2001 along both sediment cores, and 50.87% ± 5.46% sand, 36.39% ± 4.04% silt, and 12.74% ± 1.76% clay were observed throughout 2001–2020 ( Figure 4B ).

The percentages of TOC and TN values along both cores ranged from 0.39% to 2.05% and 0.026% to 0.093%, respectively. The mean content of TN along GX1 and GX2 was the same at 0.05%; however, the average content of TOC in GX1 (1.23%) was slightly higher than that of GX2 (1.15%). The molar ratio of TOC/TN ranged from 15 to 30 ( Supplementary Table S2 ). The C/N values of GX1 remained at approximately 25 from 1963 to 2001, but decreased from 25 to 15 from 2001 to 2020, which was significant (F_{8,14, α=0.05} = 3.49). The C/N ratio along GX2 also showed a noticeable decrease, but this change is not significant. Stable carbon isotope ratios (δ¹³C) ranged from -27.21‰ to -25.65‰ ( Supplementary Table S2 ).

TOC content along GX1 and GX2 showed a slight decreasing trend from 1963 to 2020, whereas TN content remained stable from 1963 to 2010 and increased slightly from 2010 to 2020 ( Figures 5A, B ). From 1963 to 2001, the mean δ¹³C values of GX1 and GX2 were -26.50‰ and -26.68‰, respectively. From 2000 to 2020, the δ¹³C values of GX1 and GX2 increased from -26.50‰ to -25.65‰ and from -26.68‰ to -25.82‰, respectively ( Figures 5C, D ). The variation in δ¹³C values along GX1 was significant (F_{8,14, α=0.05} = 3.32), whereas there is no significant variation along GX2.

The proportions of MOM, river OM and marine OM along the GX1 sediment core were 63.08% ± 6.73%, 26.36% ± 6.12% and 10.58% ± 0.95%, respectively. The proportions of MOM, river OM and marine OM along the GX2 sediment core were 63.99% ± 7.30%, 25.64% ± 6.15% and 10.36% ± 1.27%, respectively. In summary, the proportions of MOM, river OM and marine OM in GX1 and GX2 sediment accounted for 63.56% ± 7.05%, 25.98% ± 6.14% and 10.46% ± 1.13% of the total OM, respectively ( Supplementary Table S3 ). In the early stages of sediment deposition (1963–2001), MOM, river OM and marine OM accounted for approximately 65.73% ± 3.04%, 23.76% ± 2.35% and 10.52% ± 0.88% of total OM, respectively. In the following stage of sediment deposition (2001–2020), MOM, river OM and marine OM accounted for approximately 57.34% ± 7.52%, 31.50% ± 6.66% and 11.16% ± 1.21% of the total OM, respectively.

The value of Λ6 is the sum of syringyl and vanillin phenols, whereas the value of Λ8 is the sum of Λ6 and cinnamyl phenols. The average values of Λ6 and Λ8 in GX1 were 4.60 ± 1.00 and 4.94 ± 1.07 mg/100 mg OC, whereas in GX2, the values were 5.20 ± 1.31 and 5.66 ± 1.43 mg/100 mg OC. The trend of total lignin content Λ6 and Λ8 remains basically the same ( Figures 6A, B ; Spearman’s rho = 0.997, p < 0.001). The S/V and C/V ratios of GX1 ranged from 0.82–1.03 (mean 0.92 ± 0.06) and 0.09–0.21 (mean 0.14 ± 0.02), respectively. The S/V and C/V ratios of GX2 ranged from 0.86–1.21 (mean 0.99 ± 0.09) and 0.11–0.26 (mean 0.17 ± 0.04), respectively. The LPVI values of GX1 and GX2 were 876–3279 and 1091–8038, with an average of 1727 and 2771, respectively. The (Ad/Al)s and (Ad/Al)v were 0.14–0.21 (mean 0.19 ± 0.02) and 0.21–0.27 (mean 0.26 ± 0.02) along GX1 and 0.17–0.48 (mean 0.33 ± 0.10) and 0.22–0.30 (mean 0.27 ± 0.02) along GX2, respectively ( Supplementary Table S4 ). The Mann–Whitney U test results showed significant differences in C/V values (p = 0.003) and S/V values (p = 0.003) between GX1 and GX2 sites. The results of the Mann–Whitney U test showed significant differences in (Ad/Al)s (p < 0.001) and (Ad/Al)v (p = 0.043) values between the GX1 and GX2 samples.

Total lignin content increased slightly from 1964 to 2000 but decreased from 2000 to 2020 ( Figures 6C, D ). No significant change is observed in the (Ad/Al)s of GX1 from the bottom to the surface sediment; however, a significant decrease in (Ad/Al)s from the bottom to surface of core GX2 is observed. The (Ad/Al)v values of GX1 and GX2 did not change significantly from the bottom to the top of the sediment ( Figures 6E, F ). The values of C/V (Spearman’s rho = 0.662 to 0.704, p < 0.001) and S/V (Spearman’s rho = 0.482 to 0.633, p < 0.015) showed a significant positive correlation with age at GX1 and GX2. (Ad/Al)s value (Spearman’s rho = 0.893, p < 0.001) was significantly negatively correlated with age at GX2, whereas the (Ad/Al)s value of GX1 did not change significantly with age.

Mangrove forests store a remarkable amount of OC in the sediments. In the current study, the OC stock in the one-meter-top sediments is estimated to be an average of 144.5 Mg C_org ha^-1. The estimation aligns closely with the estimated SOC stock in the 8-year-old restored mangrove forests at the Shijiao station of Pearl Bay; however, this value is lower than the natural mangrove forests (Figure 6.5 in Chen et al., 2021). Integrating these results indicates that the average SOC stock in the mangrove forests of Pearl Bay is approximately 150 Mg C_org ha^-1. This figure is notably lower than the national average for China, which is reported as 270.39 ± 76.25 Mg C_org ha^-1 (Liu et al., 2014) and the global average of 334 Mg C_org ha^-1 (Kauffman et al., 2020). Such comparisons suggest that the mangrove forests in Pearl Bay possess a comparatively diminished capacity for carbon storage. Various factors, both biotic and abiotic, influence the carbon storage capacity of mangrove forests. Notably, temperature always emerges as a primary abiotic factor governing the spatial pattern of SOC stock (Kauffman et al., 2020; Wang et al., 2021). Recently, tidal range has been identified as a critical driver of carbon storage (Wang et al., 2021). Compared to previously published results, the average SOC stock in our study is similar to the area with a similar temperature condition within the meso-tidal range of 2–4 m, which is lower than the area with a micro-tidal range of 0–2 m ( Supplementary Figure S1 ). The average SOC value of six micro-tide areas was 304.68 Mg C_org ha^-1 (Cai et al., 2013; Feng et al., 2019; Gao et al., 2019; Wang et al., 2019; Peneva-Reed et al., 2021; Rovai et al., 2021), and the average SOC value of six meso-tide areas was 137.23 Mg C_org ha^-1 (Bhomia et al., 2016; Palacios Peñaranda et al., 2019; Wang et al., 2019; Su et al., 2020; Bacar et al., 2023). Lower tidal ranges can reduce mangrove contact time with the tide, thus reducing the likelihood of SOC being washed away and retaining more SOC (Wang et al., 2021). Perera and Amarasinghe (2019) also found that mangrove sediment in Sri Lanka have a better carbon sequestration function under a micro-tide environment, and the SOC sequestration capacity is significantly negatively correlated with the average tidal amplitude. Furthermore, lower SOC may have attributed to a faster rate of SOC decomposition in areas with larger tidal ranges and more frequent tidal inundation due to the frequent renewal of electron acceptors, which may lead to a lower degree of OC preservation and lower OC stock in sediments (Marchand et al., 2004, 2006).

In areas with similar temperatures and tidal ranges (approximately 2.4 m), we found that the SOC of mangroves in Pearl Bay is also significantly lower than that of Gaoqiao (260.3 Mg C_org ha^-1), Yingluo Bay (237.68 Mg C_org ha^-1) and Zhanjiang (202.7 Mg C_org ha^-1) mangroves (Wang et al., 2013; Gao et al., 2019; Yan et al., 2024). We hypothesize that the rate of sediment increase is another key factor and that faster sedimentation rates lead to lower SOC content. Higher sedimentation rates suggest that mangroves may have received more input from inland sediments, which tend to contain more mineral sediment, thus diluting the OC content (Kusumaningtyas et al., 2019). Therefore, when deposited to the same depth, a high deposition rate corresponds to low carbon storage, which is confirmed by the fact that the deposition rate of Yingluo Bay and Gaoqiao (6.5–11 mm a^-1) sediments is lower than that of Pearl Bay (17.5 mm a^-1) (Zhu et al., 2016).

Previous studies have shown that OC decomposition is constrained by the composition of OC because the residues of different plants have different resistance to decomposition (Hemminga and Buth, 1991; Valéry et al., 2004). Sedimentary OC in mangrove forests is a composite of autochthonous inputs, including above ground biomass (e.g., leaves and branches) and below ground biomass (e.g., fine roots), as well as allochthonous inputs from river and marine environments (Krauss et al., 2014; Stringer et al., 2016; Xiong et al., 2018b; Sasmito et al., 2020). The SOC in salt marsh ecosystems primarily relies on its combination with fine-grained minerals to form mineral OC, whereas SOC in mangrove forests mainly depends on high aromaticity and low mineral OC (Sun et al., 2019). Therefore, we consider that the decomposition rate of MOM is lower than that of allochthonous OM in mangrove sediments. Under such a mechanism for carbon fixation, nitrogen content can significantly affect the decomposition rate of OM, and OM with a higher C/N value has stronger resistance to biodegradation (Jones et al., 2016). MOM from mangrove vegetation has a higher C/N value and a high refractory compound content, such as tannins, lignin, and cellulose (Holmer and Olsen, 2002; Friesen et al., 2018). In contrast, allochthonous OM from river or marine environments are relatively more susceptible to degradation due to their high nitrogen content. The high C/N (23.74±4.06‰), C/V (0.96 ± 0.09), S/V (0.16 ± 0.03), and lower δ¹³C value (-26.45±0.35‰) in the study area indicate that the organic matter is primarily derived from local C₃ mangrove angiosperms ( Figures 7A, B ). This finding was consistent with the output of the SIMMR model showing that MOM accounted for 63.56% ± 7.05%, river OM accounted for 25.98 ± 6.14% and that marine OM accounted for 10.46% ± 1.13% of the total sedimentary OM. Based on Figures 8A, B , the decrease in river and marine OM with increasing depth aligns with the hypothesis of preferential degradation of river and marine OM. In regions with similar temperatures, we also observed a high proportion of MOM in the sediments. For instance, in Yingluo Bay, the C/N ratios varied from 12.3 to 25.0, with the δ¹³C values from -28.8‰ to -26.6‰ and 67.0%–88.1% autochthonous input (Xia et al., 2015). On Qi’ao Island, the C/N ratios ranged from 12.032 to 26.690, with a 70.21% autochthonous input (Jiang et al., 2021). Among the three investigated regions, Pearl Bay demonstrated the lowest concentration of MOM, thereby leading to lower SOC levels (150 Mg C_org ha^-1) relative to Yingluo Bay (237.68 Mg C_org ha^-1) and Qi’ao Island (210.62 Mg C_org ha^-1) (Xia et al., 2015; Jiang et al., 2021). Similar findings were reported by Suello et al. (2022) in the mangroves of Ecuador, where mangrove forests with high inputs of MOM exhibit relatively higher SOC content.

As one of the main components of plant-derived OM, lignin is also widely used to trace OM in continental shelves, estuaries, lakes, and other environments (Lourençato et al., 2019; Lowman et al., 2021; Xia et al., 2023). The Λ8 value of this study is 5.31 mg/100 mg TOC, whereas the Λ8 value of general mangrove sediment is 1.05–6.78 mg/100 mg TOC, and the Λ8 value of estuarine sediment is 0.36–1.87 mg/100 mg TOC, which provides additional confirmation of the relatively high MOM input in mangroves (Li et al., 2017). The S/V and C/V ratios range from 0.82 to 1.21 (Avg. 0.96 ± 0.09) and 0.09 to 0.26 (Avg. 0.16 ± 0.03), respectively. These values indicate that OM in the sediments primarily originated from the non-woody tissues of angiosperms, which is consistent with the LPVI values, ranging from 876 to 8083, with an average of 2271 ± 1308. Considering that only mangrove plants exist in the vicinity of the sampling points, these non-woody tissues of angiosperms mainly derive from mangrove leaves. Similar mixed signals were found in the surrounding area. For example, the C/V and S/V values range from 0.09–0.18 and 0.82–1.02 in Bamen Bay, Hainan (Bao et al., 2013b). Compared with Ad/Al values of fresh plants, only syringyl phenols have degraded a small part in mangrove sediments, whereas vanillyl phenols have almost not degraded, which indicates that lignin in mangrove sediments in Pearl Bay is well preserved (Hedges et al., 1988). This finding also confirms the hypothesis mentioned earlier that a large input of MOM leads to slower decomposition and enhances its preservation, thereby increasing SOC content. The same situation was found in Guangxi, Guangdong, and Hainan, where lignin in sediments was basically not degraded. In the Pearl River Estuary and Shenzhen Bay, the (Ad/Al)s value is 0.16, and the (Ad/Al)v value ranges from 0.16 to s0.22 (Zhao et al., 2023). Another possible reason for the excellent preservation of lignin could be attributed to the comparatively higher rate of sedimentation. The half-life of lignin in mangrove leaf litter is 150–300 years (Dittmar and Lara, 2001). In addition, the high deposition rate causes the surface sediments to be rapidly buried in an anaerobic environment, and the lignin is more difficult to decompose.

Although the decomposition rate of OM in sediments is relatively slow in mangrove ecosystems, it still causes changes in C/N and δ¹³C values in sediment cores (Dittmar and Lara, 2001). Sampling sites GX1 and GX2 are located near the estuary, where sediment distribution is influenced by multiple disturbances such as sea level rise and human activities. These factors may enhance the degradation potential of OM in the sediments (Jonge et al., 2015). As shown in Figures 3A and 4B , the sediment particle size gradually increases from the bottom to the top of the core, which indicates that the sedimentary environment of the sampling point has changed from a silt sedimentary environment (weak tidal action) to a sand sedimentary environment (strong tidal action). In a sand sedimentary environment, the inability of coarse particles to effectively encapsulate OM and prevent its decomposition, combined with the aforementioned strong tidal action, leads to a faster rate of OM degradation (Sanders et al., 2012). However, if the OM input source remains unchanged, Figure 9A indicates that the carbon stock increases with sediment age, which contradicts the theory of organic OM degradation. Cui et al. (2021) noted that due to the high content of recalcitrant carbon, such as lignin, in MOM, the decomposition of fresh litter leads to an initial increase followed by a decrease in the proportion of labile OC in the sediment. Figure 5A shows that the OC content continues to increase with age, which does not conform to the theory proposed by Cui et al., indicating that the degradation of OM in the sediments cannot explain this trend. Yan et al. (2023) pointed out that the degradation of sediment OM leads to an increase in δ¹³C values and a decrease in the C/N ratio. However, the trends shown in Figures 5C, D are contrary to this. Moreover, if significant degradation of OM were occurring, TOC and TN values would decrease with the increasing age of the sediments. However, Figures 5A, B clearly do not show this trend. Correlation tests showed that the total lignin content is significantly correlated with TOC and TN (p < 0.001), which indicated that the degree of degradation of OC may be the same as that of lignin. However, the (Ad/Al)s and (Ad/Al)v values showed that lignin is basically not degraded and that the total lignin content in the surface layer of the sediment increases with increasing depth. This trend further confirms that OM has not undergone significant degradation. Under the condition of constant sedimentation rates and well-preserved OM in surface sediments, an increase in SOC with depth indicates an increase in the OC content of the OM. Therefore, the degradation of OM in the surface sediment of the sampling point is not the main reason for the change of C/N and δ¹³C values, but is more likely to be caused by the change in OM input source. Because OM from riverine and marine environments has a lower C/N ratio and higher δ¹³C values ( Figure 7A ), whereas mangrove OM has a higher C/N ratio and lower δ¹³C values (Lamb et al., 2006), and considering the trends of the lignin parameters, the trends observed in Figures 5C, D may be attributed to a decrease in MOM input.

In the previous section, we excluded the factor of OM degradation. The damage to mangroves caused by climate change-induced sea level rise or increased intensity of human activities may also lead to the loss of OC in sediment (Bozi et al., 2021; Wang and Gu, 2021). Sea level rise increases tidal height and the frequency of seawater intrusion into the mangrove, which affects the sources of OM and carbon storage in mangrove sediments (Perera et al., 2018; Wang et al., 2019). As shown in Figures 8A–C a gradual decrease in MOM and total lignin content is observed from the bottom to the top of the core. The lack of a clear pattern in local precipitation and temperature variations suggests that these factors may not be the primary influences ( Figures 8D, E ). While the enhanced tidal effects resulting from sea level rise may contribute to this trend, we consider that this is not the primary reason. This is because the average sediment deposition rate in the study area is 17.5 mm a^-1, which is much higher than the sea level rise rate of Pearl Bay (2 mm aa^-1), indicating that the salinity and tidal inundation frequency of mangrove habitats remain largely unaffected, as also found in Ellison (2008) and Ma et al. (2022). Xia et al. (2019) also proposed that when the sediment deposition rate (4.4 mm a^-1) is higher than the sea level rise rate, the change of sediment OM source exhibits a minimal relationship with sea level rise. Li et al. (2015) established a vulnerability assessment index system for coastal mangroves in response to sea level rise and indicated that the mangroves in Pearl Bay, Guangxi, have shown non-vulnerability in the face of sea level rise. This also supports the opinions of our study. Since sea level rise is not the primary factor threatening mangroves, we hypothesize that the changes in parameters are due to mangrove degradation caused by human activities, leading to a decrease in the input of native OM. Based on the changing trends in the sources of OM, the anthropogenic activities in the sedimentary record can be divided into two distinct stages: 1963–2001 and 2001–2020. In stage 1 (1963–2001), mangroves were less affected by human activities, thereby ensuring a consistent supply of MOM. However, in stage 2 (2001–2020), the mangroves may have experienced a certain degree of degradation due to frequent human activities, leading to a gradual reduction in the provision of MOM. The decline in the rate of SOC accumulation ( Figures 9A , D ) from 2.57 Mg C_org ha^-1 yr^-1 (stage 1) to 1.97 Mg C_org ha^-1 yr^-1 (stage 2) implies mangrove forest degradation. As the lignin in the sampling locations is exclusively derived from fresh mangrove tissues (Bao et al., 2013a), the declining trend in total lignin content from 2001 to 2020 provides additional support for our hypothesis ( Figure 8C ). In pursuit of economic growth, coastal regions have extensively converted mangrove wetlands into shrimp farming ponds. As a result, fishery-related economic indicators are employed as proxies for the intensity of human activities. Based on the variations in gross fishery value ( Figure 9B ) and the yield of cultured marine products ( Figure 9C ), during the second phase, a substantial increase in human activity intensity is the primary cause of mangrove degradation. In addition, in 2011, a marine aquaculture system was established underneath the mangrove forest in the study area. Ouyang and Guo (2016) proposed that establishing a mangrove-aquaculture coupling system in mangrove areas can effectively treat wastewater generated by aquaculture, while the wastewater can provide essential nutrients for the growth and development of mangroves. However, this requires strict regulation of the scale of aquaculture. If wastewater discharge exceeds the limit, it can lead to eutrophication of the mangrove area, adversely affecting the growth of mangroves. While this aquaculture facility does not directly occupy the land within the mangrove forest, the discharge of aquaculture wastewater may induce changes in the physical and chemical properties of the surface sediment, ultimately leading to a loss of sediment carbon storage in the mangrove forest (Fan et al., 2013; Santos-Andrade et al., 2021).

The degradation of mangroves must be taken seriously, as the destruction of mangroves not only leads to a decrease in the OC deposition rate but also accelerates the decomposition of stored carbon in sediments, releasing greenhouse gases into the atmosphere. Ultimately, this could transform mangroves from a carbon sink into a carbon source (Senger et al., 2021). The effects of human activities on the mangroves in the study area are expected to gradually decrease in the future. This is evident from the inclusion of blue carbon conservation in the national strategy since 2015 and the introduction of the “Special Action Plan for Mangrove Conservation and Restoration (2020–2025)” in 2020 (Feng et al., 2024). However, whether the changes in temperature and precipitation patterns caused by climate change will pose a threat to the development of Pearl Bay mangroves remains uncertain. Currently, the observed changes in temperature and precipitation do not correspond well to the variations in parameters within mangrove sediments ( Figures 8D, E ). Therefore, these relationships must be investigated in future studies.