Microbial Decomposition of Lignin to Methane Reduces Net Blue Carbon Benefit Across China's Saltmarshes

ABSTRACT

CH₄ emissions from mangrove, saltmarsh, and seagrass ecosystems partially offset carbon sequestration, potentially diminishing the climate mitigation capacity of these blue carbon habitats. However, a mechanistic understanding of the processes governing CH₄ production potential across large spatial scales remains limited. By integrating incubation‐based measurements from 116 sites, we reveal significant ecosystem‐specific differences in CH₄ production potential, with saltmarshes emerging as CH₄ production hotspot relative to mangroves and seagrass meadows. Using an integrated analytical approach encompassing more than 30 environmental, biogeochemical, and microbial parameters, we demonstrate that CH₄ production potential converges on sediment organic carbon availability, particularly plant‐derived carbon, as a key regulatory axis. Additionally, metagenome‐assembled genomes (MAGs) recovered from saltmarshes show a functional bias toward lignin degradation, thereby fueling downstream CH₄ production via methylotrophic pathways. Lignin‐addition and stable carbon isotope experiments further provide supportive evidence that lignin decomposition enhances Chinese saltmarsh CH₄ production potential, revealing a pathway that may reduce net blue carbon benefit. Together, these findings underscore that saltmarsh plant‐derived lignin is less stable than conventionally assumed, as microbial processing redirects stored carbon toward CH₄ production, challenging current blue carbon accounting frameworks at a continental scale within China.

Across 116 coastal sites in China, saltmarsh sediments exhibited substantially higher CH₄ production potential than mangrove and seagrass meadow sediments. Integrated geochemical, metagenomic, incubation, and isotopic analyses revealed that plant‐derived lignin fuels CH₄ production in saltmarshes through microbial decomposition pathways, particularly those associated with lignin‐degrading Pseudomonadota and methylotrophic methanogenesis. These findings suggest that lignin, traditionally regarded as a stable blue carbon pool, can be transformed into CH₄, thereby reducing the net climate‐mitigation benefits of saltmarsh ecosystems.

Coastal blue carbon ecosystems, including mangroves, saltmarshes, and seagrass meadows, play a critical role in climate change mitigation (Mcleod et al. 2011). This role is primarily determined by their capacity to sequester and bury carbon over long timescales, thereby functioning as long‐term carbon sinks. However, the effectiveness of this function depends on the net balance between carbon accumulation and loss processes (Campbell et al. 2022; Spivak et al. 2019). Importantly, this balance can be substantially altered by coastal sedimentary greenhouse gas emissions (Zhu et al. 2025). Despite the typically high sulfate concentrations in blue carbon sediments, which energetically favor sulfate reduction and suppress methanogenesis through competitive substrate utilization by sulfate‐reducing bacteria, robust CH₄ production and release can still be sustained (Rosentreter et al. 2018). At the quantitative level, approximately 3%–4% of the organic carbon deposited on the seabed is microbially converted into CH₄ (Egger et al. 2018). The high global warming potential of CH₄ means that its emissions are estimated to offset 9%–12% of marine organic carbon burial when radiative forcing is considered (Egger et al. 2018). Localized studies suggest that CH₄ emissions may offset as much as 20%–50% of carbon sequestration in mangroves (Liu et al. 2020; Rosentreter et al. 2018). At the global scale, coastal environments account for approximately three‐quarters of total oceanic CH₄ emissions, with an assessment estimating emissions at up to 13 Tg year⁻¹ (Bange et al. 1994). Together, these observations underscore the importance of understanding the potential and mechanisms underlying CH₄ production, as these processes can complicate blue carbon accounting.

Among the most fundamental factors, carbon sources provide both the material and energy basis for microbial metabolism. Few local‐scale investigations indicate that coastal microbiomes decompose complex plant macromolecules and release CH₄ (Dai et al. 2025; Schorn et al. 2022; Yuan et al. 2019). However, large‐scale evaluations remain constrained by limited insight into microbial biogeography due to the lack of coordinated datasets or inconsistent methodological approaches and the coupled physical, chemical, and biological processes that together shape methanogenic ability across heterogeneous ecosystems. Additionally, sediment organic matter comprises a complex mixture of sources (Guigue et al. 2016), including plant‐derived and microbial necromass carbon (Li, Song, et al. 2023; Zhang et al. 2023; Xiao et al. 2026). Current assessments of greenhouse gas production from blue carbon ecosystems rarely distinguish among sediment carbon sources. This omission introduces substantial uncertainty into estimates of blue carbon storage capacity and undermines effective carbon‐sink management.

Numerous studies have documented a wide array of other abiotic and biotic factors influencing CH₄ emissions in coastal wetlands, including salinity, sulfate availability, nutrient status, redox conditions, vegetation type, and climate conditions (Al‐Haj and Fulweiler 2020; Yu et al. 2025). Most of these insights are derived from localized or ecosystem‐specific investigations, where site‐specific geochemical settings strongly modulate CH₄ production and emission processes (Roth et al. 2022). As a result, the relative importance of individual drivers often varies across studies, making it difficult to identify generalizable controls on CH₄ dynamics. The pronounced heterogeneity in microbial assemblages further complicates efforts to integrate these factors within a unified framework. Although meta‐analytical studies have advanced our understanding of the processes controlling CH₄ emissions in blue carbon ecosystems (Al‐Haj and Fulweiler 2020; Rosentreter et al. 2021), the limited availability of coordinated, multi‐parameter assessments across large spatial scales within a standardized framework makes it challenging to disentangle the relative contributions of carbon sources, microbial attributes, and physicochemical conditions in regulating CH₄ production and emissions.

Available information on CH₄ emissions from blue carbon ecosystems is primarily derived from in situ measurements using chamber‐ or tower‐based approaches (Al‐Haj and Fulweiler 2020; Rosentreter et al. 2021). While these data have substantially advanced our understanding of CH₄ dynamics, considerable uncertainty remains in attributing the underlying drivers of CH₄ release. This is mainly because CH₄ undergoes complex processes of production, transport, and consumption before it reaches the atmosphere. In this context, laboratory microcosm experiments serve as a valuable complement to in situ measurements, providing controlled conditions that facilitate the identification of fundamental drivers. Over the past 50 years, if not longer, such experiments have contributed substantially to mechanistic insights into CH₄ production processes (Oremland et al. 1982). Based on high‐resolution incubation measurements derived from a comprehensive dataset comprising 116 sites spanning mangrove, saltmarsh, and seagrass meadow ecosystems (Figure 1A), our study provides a large‐scale assessment of CH₄ production potential. Global syntheses of CH₄ fluxes indicate that mangroves and saltmarshes exhibit comparable emission rates, which are consistently higher than those observed in seagrass meadows (Al‐Haj and Fulweiler 2020; Rosentreter et al. 2021). Accordingly, we hypothesize that a similar pattern extends to CH₄ production potential across these three blue carbon ecosystems. Given the pronounced heterogeneity, we further hypothesize that the organizing principles of CH₄ production differ among mangroves, saltmarshes, and seagrass meadows, reflecting ecosystem‐specific integration of carbon processing and functional microbiomes. To test this hypothesis, we integrated more than 30 environmental, biogeochemical, and microbial parameters. Our findings demonstrate that the stability of plant‐derived lignin in Chinese saltmarshes is lower than traditionally assumed, as microbially mediated CH₄ production redirects stored carbon away from long‐term sequestration, thereby challenging prevailing assessments of saltmarsh carbon sink strength.

FIGURE 1.

CH₄ production potential across Chinese blue carbon ecosystems. (A) Sampling locations and corresponding ecosystem types. The inset highlights dense sampling in the mangroves of Southern China. (B) Mean CH₄ production rate of incubated sediments. Boxplots depict medians (line), first and third quartiles (box, interquartile range or IQR), and upper and lower extremes (whiskers). The whiskers are defined as ≤ 1.5 times IQR from the first and third quartiles, respectively. Different letters indicate statistical differences (Kruskal–Wallis test, p < 0.05).

1. Results and Discussion

1.1. Distinct CH₄ Production Potential Across Three Blue Carbon Ecosystems

Our large‐scale sampling revealed that CH₄ production potential was highly heterogeneous (Figure S1), with variations exceeding two orders of magnitude across 116 sites (range: 0.011–4.18 μmol kg⁻¹ sediment day⁻¹, Tables S1 and S2). This pattern is consistent with previous studies highlighting pronounced variability in CH₄ emissions across coastal ecosystems (Al‐Haj and Fulweiler 2020; Capooci et al. 2023; Roth et al. 2022). This heterogeneity likely reflects the combined influences of local hydrology, vegetation composition, and sediment biogeochemistry, which create distinct microenvironments for methanogenesis and CH₄ oxidation. Our results thus provide direct evidence linking the pronounced heterogeneity of CH₄ emissions to substantial variability in underlying CH₄ production.

At ecosystem scale, CH₄ production potential in mangroves (n = 34, median: 0.034 μmol kg⁻¹ sediment day⁻¹; interquartile range or IQR: 0.020–0.045) was comparable to those of seagrass meadows (n = 25, median: 0.024; IQR: 0.018–0.043), but significantly lower than saltmarshes (n = 57, median: 0.046; IQR: 0.027–0.252, Kruskal–Wallis test p = 0.0007, Figure 1B). Although synthesis studies reported comparable median CH₄ fluxes between mangroves and saltmarshes at the global scale (Al‐Haj and Fulweiler 2020; Rosentreter et al. 2021), the robust CH₄ production potential observed in Chinese saltmarsh sediments indicates that sedimentary processes can further differentiate CH₄ dynamics among coastal ecosystems. This discrepancy between global syntheses and our continental‐scale findings may also reflect regional differences in dominant plant species and environmental conditions; for instance, the widespread invasion of Spartina alterniflora in Chinese saltmarshes has been shown to substantially enhance CH₄ production potential by supplying abundant substrates and promoting methylotrophic methanogenesis (Yang et al. 2021; Yuan et al. 2019). Furthermore, our incubation‐based approach specifically quantified sedimentary CH₄ production potential while minimizing confounding influences from plant‐mediated transport and oxidative consumption processes (Bastviken et al. 2023), which can otherwise obscure underlying differences in production.

1.2. Sediment Organic Carbon as the Organizing Axis of Saltmarsh CH₄ Production Potential

In this section, we analyzed key factors driving CH₄ production. We classified the analyzed parameters, totaling over 30, into three categories: sediment organic carbon, microbial and genetic diversity, and geochemical characteristics. The accumulation of both plant‐derived carbon (Figure 2A) and microbial necromass carbon (Figure 2B) followed the order: mangroves > saltmarshes > seagrass meadows, also as their subunits (Figure S2). In addition to the widely known plant‐derived materials for blue carbon storage, we found that microbial necromass carbon is also an important source of sediment organic carbon, consistent with observations from global terrestrial ecosystems (Camenzind et al. 2023; Liang et al. 2019). Both plant‐derived and microbial necromass carbon appeared to not correlate with CH₄ production potential in mangroves and seagrass meadows (Spearman correlations, p > 0.05, Figure 2C). In contrast, CH₄ production potential from Chinese saltmarshes was likely regulated by both types (p < 0.05). This finding also helps explain the observed increase in CH₄ production and emissions following the invasion of Spartina alterniflora in Chinese saltmarshes, which may contribute plant‐derived organic carbon to sediments and methanogens (Gao et al. 2018; Yang et al. 2021). This pattern likely reflects fundamental differences in microbial community structure and carbon‐use efficiency. Mangrove and seagrass systems may fail to rapidly assimilate organic carbon into methanogenic pathways, whereas saltmarsh systems may facilitate more direct carbon‐to‐methanogen transfer. This divergent carbon processing pathway could imply that global carbon models assuming uniform substrate availability might have difficulty in predicting the observed coastal CH₄ emissions (Macreadie et al. 2025).

FIGURE 2.

Carbon sources and the constraints on CH₄ production potential. (A) Lignin phenol content, represented as the sum of vanillyls (V), syringyls (S), and cinnamyls (C). Different lowercase letters indicate significant differences (Kruskal–Wallis test, p < 0.05). (B) Microbial necromass carbon content, calculated as the sum of bacterial and fungal necromass. (C) Potential drivers on CH₄ production potential (Spearman correlations). The red box indicates statistically significant results (p < 0.05) following adjustment for multiple‐testing effects. Sediment organic carbon resource includes lignin phenols, cinnamyl monomers, syringyl monomers, vanillyl monomers, microbial necromass carbon, bacterial and fungal necromass carbon. Microbial diversity includes the Shannon index of MAGs (Shannon‐MAGs), the inverse Simpson index of MAGs (Simpson‐MAGs), the first (PCoA1‐MAGs) and second components (PCoA2‐MAGs) of principal coordinate analysis (PCoA) of MAGs. Gene diversity is represented by the Shannon index of genes (Shannon‐genes), the inverse Simpson index of genes (Simpson‐genes), the first (PCoA1‐Genes) and second components (PCoA2‐Genes) of PCoA of genes. Physicochemical characteristics include total organic carbon (TOC), particulate organic carbon (POC), mineral‐associated organic carbon (MAOC), total nitrogen (TN), carbon to nitrogen ratio (C/N), total phosphorus (TP), carbon to phosphorus ratio (C/P), pH, electrical conductivity (EC), free iron compounds (Fed), amorphous Fe compounds (Feo), organically bound Fe (Fep), and particle diameter (ø, mean value; ø4, lower than 4 μm; ø4–63, between 4 and 63 μm; ø63, larger than 63 μm).

Despite the general understanding that mineral‐associated organic carbon (MAOC) is more stable than particulate organic carbon (POC) (Guo et al. 2022; Liu, Ji, et al. 2022; Liu, Qin, et al. 2022), a range of evidence also suggests that both MAOC and POC are highly sensitive to microbial utilization, with MAOC exhibiting even greater sensitivity (Díaz‐Martínez et al. 2024; Lugato et al. 2021). Our findings revealed that both MAOC and POC were significantly related to saltmarsh CH₄ production potential (Spearman correlations, p < 0.05, Figure 2C). This relationship may be attributed to the vulnerability of MAOC during anaerobic events characteristic of wet conditions, as demonstrated by year‐long anaerobic soil incubation experiments showing that MAOC can constitute a bioavailable carbon pool under anoxic conditions, thereby revealing a trade‐off between physicochemical stabilization and kinetic‐thermodynamic protection mechanisms (Huang et al. 2020). Additionally, minerals may promote methanogenesis by providing attachment sites for microbial communities and acting as catalysts for organic reactions. The greater specific surface area of particles can enhance mineral‐organic‐microbial interactions, potentially facilitating these processes (Kleber et al. 2021). Iron mineral composition and content showed complex correlation with CH₄ production potential. The inhibitory effect of oxidized iron oxides, coupled with the promoting role of reduced iron oxides, may introduce significant complexity to the relationship between CH₄ production capacity and iron oxides (Jäckel and Schnell 2000; Xiao et al. 2019). Our large‐scale, natural abundance approach may have obscured these mechanistic relationships because field‐collected sediments contain complex Fe mineral assemblages at varying redox states, whereas controlled laboratory experiments typically use pure mineral phases.

Salinity exhibited an inhibitory effect on CH₄ production potential in Chinese saltmarshes (Spearman correlations, p < 0.05, Figure 2C). This inverse relationship aligns with the well‐established paradigm of competitive exclusion (Bartlett et al. 1987; Soued et al. 2024), where sulfate‐reducing bacteria outcompete methanogens for substrates, such as acetate and H₂. Yet our data show that saltmarshes produced more CH₄ than other ecosystems, even though plant‐ and microbe‐derived carbon pools were lower than in mangroves, highlighting the importance of non‐competitive methanogenic pathways in saltmarsh sediments. This finding reconciles apparent contradictions in the literature where some high‐salinity marshes emit substantial CH₄ (Capooci et al. 2023), while others show strong suppression (Soued et al. 2024). The key distinction appears to be whether the system receives continuous plant inputs capable of fueling methylotrophic pathways that bypass competitive substrates, a mechanism that, as we demonstrate below, is widely prevalent in saltmarshes.

Additionally, Chinese saltmarsh with high total nitrogen and phosphorus levels tended to stimulate CH₄ production potential (Spearman correlations, p < 0.05, Figure 2C), likely because the enrichment of inorganic nutrients facilitated microbial activity related to organic carbon degradation and methanogenesis (Sosa et al. 2020; Xiao et al. 2017). The magnitude of nutrient effects in our study was comparable to the dominant effect of carbon sources, consistent with study in phosphorus‐limited seagrass meadows where phosphorus alone can double methanogenic rates (Sosa et al. 2020). However, we found no significant effect of phosphorus on CH₄ production potential in seagrass meadows (Spearman correlations, p > 0.05). This discrepancy may reflect the naturally high background nutrient levels across our coastal sites, many of which receive substantial anthropogenic inputs (Qu and Kroeze 2010), potentially creating nutrient saturation.

1.3. Bacterial Processing of Lignin as a Key Pathway in Chinese Saltmarsh Carbon Decomposition

Across Chinese blue carbon ecosystems, we recovered 1954 medium‐ to high‐quality unique MAGs. This recovery is comparable to data from permafrost (Woodcroft et al. 2018), glacier (Liu, Ji, et al. 2022; Liu, Qin, et al. 2022) and lake (Garner et al. 2023) ecosystems. Among these, we identified 1481 MAGs in saltmarshes, which were higher than those found in mangroves (1307) and seagrass meadows (1171). This dataset reflected the inherent microbial diversity and complexity across blue carbon ecosystems, as demonstrated by analyses at levels of MAGs, genes, and reads (Figure 3; Figures S3 and S4). No significant correlation was observed between microbial diversity at both MAG and gene levels and CH₄ production potential (Spearman correlations, p > 0.05, Figure 2C); this lack of correlation aligns with functional redundancy theory and empirical observations from other ecosystems where methanogenic potential is decoupled from overall diversity metrics (McCalley et al. 2014). However, samples with large differences in CH₄ production potential (from high to low) in saltmarshes and seagrass meadows showed distinct community distributions (See Section 2, Adonis test, p = 0.15, 0.024, and 0.024 for mangroves, saltmarshes and seagrass meadows, respectively). That is, significant differences in community composition were observed when samples were regrouped according to CH₄ production potential (Figure 3; Figure S3). This pattern, where community composition correlates with function despite diversity metrics being uninformative, has been observed in other anaerobic system (McCalley et al. 2014) and suggests that specific functional guilds rather than community‐wide diversity drive CH₄ production. Therefore, certain microbial taxa associated with CH₄ production may cluster together. In Chinese saltmarsh, for example, a total of 25 MAGs were statistically positively correlated with CH₄ production potential, 7 of which belong to the phylum Pseudomonadota (Figure S5). This suggests that these clustered microorganisms may actively influence CH₄ production.

FIGURE 3.

Genome‐resolved view of the microbial communities across CH₄ production potential groups in three ecosystems. (a) Microbial community profile based on MAGs abundances (rows) from mangroves (violet), saltmarshes (blue) and seagrass meadows (green). Black lines delineate sites according to CH₄ production potential (see Section 2). H, high potential; L, low potential; M, median potential. The phylogenetic tree appears on the left. Only phyla with more than 10 MAGs recovered were noted. Pseudomonadota (in red font, 757 MAGs) is the most abundant phylum. (b) Shannon and Inverse Simpson diversity indices for each sample (filled circles). The p‐value from the Kruskal–Wallis test for the diversity among three coastal wetland types is displayed on the left.

Regarding the breakdown of macromolecular carbon to provide substrates for CH₄ production, the supply of labile carbon from macromolecular sources to support acetoclastic and hydrogenotrophic methanogens seemed to be lower in saltmarshes than in mangroves and seagrass meadows (Figure S6). For example, the abundance of MAGs corresponding to some steps in cellulose and xylan degradation pathways was lower in saltmarshes. On the contrary, bacterial metabolic pathways, analyzed at gene abundance level and associated with degradation of lignin and by‐products, as indicated by recent studies (Grevesse et al. 2022; Liu et al. 2025; Peng et al. 2025), were significantly enriched in saltmarshes (Figure S7). In brief, 11 representative pathways in saltmarshes were markedly more abundant than those in the other two ecosystems, with only one pathway found in mangroves and none in seagrass meadows. The relationship analysis between these functional genes and CH₄ production potential identified 25 genes positively correlated with CH₄ production, with 8 more enriched in saltmarshes, 1 gene in mangroves, and no gene in seagrass meadows (Figure S8). Notably, three of the top five representative genes–k00449, k00481, and K04101–were enriched in saltmarshes (Kruskal–Wallis test, p < 0.05). The degradation of lignin‐like compound by Pseudomonadota, also corroborated by numerous pure culture studies (Kumar et al. 2020; Salvachúa et al. 2020), was prominently demonstrated in this work (Figure S8), with 757 MAGs accounting for 38.8% of the total (Figure 3a). Additionally, the top four metabolic pathways involved in lignin degradation were predominantly driven by Pseudomonadota and were most abundant in saltmarshes (Figure 4A). The prominent role of Pseudomonadota in lignin degradation is consistent with genome‐resolved analyses of Arctic coastal sediments (Grevesse et al. 2022), where 59% of lignin‐degrading MAGs in terrestrially influenced habitats are affiliated with Pseudomonadota, many of which encode complete catechol and protocatechuate metabolic pathways. This cross‐biome consistency supports the contention that Pseudomonadota serves as keystone taxa processing plant‐derived carbon in saltmarshes. These results help explain why mangroves have higher levels of plant‐derived carbon than saltmarshes in China, but with an even weaker CH₄ production potential.

FIGURE 4.

Key bacterial phylum and methanogenesis‐related genes promote the breakdown of lignin‐like compounds for CH₄ production. (A) The microbial functional groups of four major metabolic pathways for lignin mineralization are shown, with Pseudomonadota being the dominant group (red font). This analysis underscores the key role of this bacterial phylum in saltmarsh lignin mineralization. Other bacterial phyla include Acidobacteriota (Acid), Actinomycetota (Acti), Armatimonadota (Arma), Bacillota (Baci), Bacteroidota (Bact), Bdellovibrionota (Bdel), Chloroflexota (Chlo), Campylobacterota (Camp), Desulfobacterota (Desu), Deinococcota (Dein), Gemmatimonadota (Gemm), Myxococcota (Myxo), Pseudomonadota (Pseu), Patescibacteria (Pate), Planctomycetota (Plan), Thermoproteota (Ther), Verrucomicrobiota (Verr). (B) Nine genes for key enzymes involved in methanogenic pathways significantly enriched in saltmarshes. The list of these genes includes acetyl‐CoA synthase (acs), acetate kinase (ack), formyl‐methanofuran dehydrogenase subunit B (fwdB), formyl‐methanofuran dehydrogenase subunit C (fwdC), formyl‐methanofuran dehydrogenase subunit D (fwdD), methyl‐coenzyme M reductase subunit C (mcrC), F₄₂₀‐dependent methylene‐H₄MPT reductase (mer), methanol‐specific corrinoid protein coenzyme M methyltransferase subunit A (MtaA), coenzyme M methyltransferase subunit A (mtrA). Boxplots show medians (line), 1st and 3rd quartiles (box, interquartile range or IQR), upper and lower extremes (whiskers). The whiskers were determined as ≤ 1.5 times IQR against 1st and 3rd quartiles, respectively. The different lowercase letters represent significant differences (Kruskal–Wallis test, p < 0.05).

The lower‐than‐expected stability of lignin in saltmarshes challenges long‐standing assumptions about the recalcitrant aromatic carbon preservation in natural ecosystems. Emerging evidence from coastal marine systems indicates that anaerobic lignin degradation can occur widely (Grevesse et al. 2022; Ortega‐Arbulú et al. 2019). This rapid turnover has profound implications for blue carbon accounting, as current IPCC methodologies assume that lignin constitutes a stable, long‐term carbon pool. Our data suggest that saltmarshes receiving substantial terrestrial plant inputs (Li, Fu, et al. 2023) and harboring lignin‐degrading microbial communities may support more vulnerable carbon stocks than previously recognized, with important implications for revising persistence factors in carbon offset protocols.

1.4. Lignin‐Fueled Methylotrophic Methanogenesis in Chinese Saltmarshes

Methermicoccus has been found to degrade aromatic compounds to release CH₄ (Cornelia 2016; Mayumi et al. 2016) demonstrating its metabolic potential to directly utilize plant‐derived lignin. Despite our large‐scale microbial sequencing efforts, Methermicoccus remained undetectable, possibly due to its preference for coal and oil fields (Cheng et al. 2007; Mayumi et al. 2016) or challenges in assembling genomes in coastal wetlands. In addition to Methermicoccus, naturally low abundance of archaea resulted in only a limited number of methanogens being classified into Methanotrichales, Methanomicrobiales, Methanofastidiosales and Methanomassiliicoccales (Figure S9). Direct use of methyl compounds by the first two groups has rarely been reported, likely due to a lack of corresponding metabolic pathways in their genomes. Methanofastidiosales and Methanomassiliicoccales are shown to possess the capacity to complete the methylotrophic pathway (Figure S9; Nobu et al. 2016; Zhang et al. 2023) suggesting their potential contribution to CH₄ production. We further analyzed 30 key genes involved in the three methanogenic pathways and found that 9 of these genes were significantly enriched in Chinese saltmarshes (Figure 4B; Figure S10). In contrast, none was statistically enriched in mangroves or seagrass meadows. Notably, mtrA, a marker gene for the methylotrophic pathway, was significantly enriched in saltmarshes (Kruskal–Wallis test, p < 0.001). This methyl‐substrate‐based methanogenic metabolism, which largely circumvents competition with sulfate‐reducing microorganisms, likely helps explain the robust CH₄ production potential observed in Chinese saltmarshes.

Additionally, an incubation experiment involving lignin addition demonstrated an obvious increase in saltmarsh CH₄ production potential (Figure 5), providing supportive evidence for the CH₄ production resulting from lignin mineralization. Concurrently, the δ¹³C‐CH₄ values were more negative than those from samples without lignin addition, further suggesting that a substantial CH₄ source may derive from lignin decomposition‐fueled methylotrophic methanogenesis (Krzycki et al. 1987; Zhuang et al. 2016). However, this is not robust proof of the methylotrophic methanogenic pathway, as the lignin was not isotopically labeled, and the lignin dose used (0.05 g lignin to 1 g sediment) appears high and may not accurately reflect natural substrate availability. In support of our findings, a recent incubation experiment conducted under anoxic conditions demonstrated that lignin served as a direct substrate for CH₄ production in peatland soils (Liu et al. 2025). However, lignin or plant‐derived phenolics were also proposed to inhibit methanogenesis in lakes (Emilson et al. 2017) and peatlands (Medvedeff et al. 2015), highlighting a context‐dependent role of lignin in regulating methanogenic potential. Overall, the effect of lignin degradation on CH₄ production, whether it promotes (Medvedeff et al. 2015) or inhibits (Emilson et al. 2017; Medvedeff et al. 2015), appears to be strongly ecosystem dependent and may even vary across micro‐ecological niches, as contrasting outcomes were reported within peatland systems. These divergent responses likely reflect differences in the composition and functional potential of key microbial taxa mediating lignin transformation, as exemplified by the contrasting microbial configurations observed in blue carbon ecosystems.

FIGURE 5.

Enhanced CH₄ production potential from methyl compounds with lignin as a substrate in Chinese saltmarshes. In unsterilized samples (A, C), the addition of lignin significantly increased CH₄ production potential in saltmarshes, indicating that lignin is an important methanogenic substrate. In treatments with added lignin (B, C), unsterilized samples exhibited higher CH₄ production potential, suggesting that living microbiota play a key role in catalyzing lignin degradation. Microbial communities in saltmarshes are more capable of utilizing lignin compared to those in mangroves and seagrass meadows. Additionally, lower δ¹³C values with lignin addition further indicate that saltmarshes may preferentially utilize downstream products from lignin degradation, such as methyl compounds, for CH₄ production.

1.5. The Role of Carbon Pools in Blue Carbon Ecosystem Offset Potential

The limited incorporation of CH₄‐generating processes into blue carbon frameworks introduces substantial uncertainty in assessing the net climate mitigation potential of coastal restoration (Jones et al. 2024). Our high‐resolution measurements improve reliability to confine key factors contributing to continental‐scale CH₄ production within China. Geochemical properties contribute only 3% to CH₄ production potential in saltmarshes (Figure 6a), slightly higher than the 2% attributed to microbial diversity but significantly lower than the 53% from carbon inputs. Furthermore, plant sources are more effective in explaining CH₄ production potential compared to microbial necromass carbon (32.4 vs. 13.8%, Figure S11). These results underscore the importance of explicitly distinguishing among carbon inputs when assessing the carbon sequestration potential of coastal ecosystems, particularly in saltmarshes. While bulk carbon stocks are often treated as functionally equivalent in blue carbon assessments (Fu et al. 2021; Wang et al. 2023), our findings demonstrate that different carbon pools exert markedly different controls on CH₄ production potential. Accordingly, not all stored carbon confers the same degree of climatic benefit. This source‐specific effect is particularly relevant for saltmarshes, where high inputs of vascular plant material coincide with strong microbial coupling between lignin mineralization and CH₄ production (Figure 6b). Collectively, these patterns highlight that accurate evaluation of coastal carbon sinks requires not only quantification of total carbon stocks, but also explicit consideration of carbon source identity and its biogeochemical fate. Importantly, lignin can be actively mineralized under anoxic conditions, challenging its long‐assumed role as a stable blue carbon pool. Previous estimates of SOC loss from vegetated coastal ecosystems (Fu et al. 2021) may have been underestimated due to the omission of lignin mineralization, particularly in areas where terrestrial plant carbon constitutes more than half of the blue carbon sink (Li, Fu, et al. 2023). Given the likely importance of plant species, regional biogeography, salinity regimes, the invasion by Spartina alterniflora , and sediment geochemistry, the conclusions of the current study can only be framed more explicitly as applicable to the sampled systems in China. Further field and indoor studies on a global scale would help elucidate the lignin‐driven saltmarsh CH₄ production mechanisms in general.

FIGURE 6.

Ecosystem‐specific role of lignin in regulating CH₄ production potential across blue carbon habitats. (a) Partitioning potential drivers on CH₄ production potential across mangroves, saltmarshes, and seagrass meadows using variation partitioning analysis. “residuals” means random variation, that cannot be explained by the selected parameters. (b) A conceptual model illustrating CH₄ production potential and their drivers. Microbial diversity had a minor impact on CH₄ production potential; however, key microbial taxa, such as Pseudomonadota in saltmarshes, facilitated methanogenic progress by accessing functional carbon substrate (i.e., lignin‐like compounds and their downstream products). The unit for CH₄ production potential is expressed in nmol kg⁻¹ sediment day⁻¹.

2. Methods

2.1. Field Sampling, Sediment Incubation and CH₄ Analysis

We collected 181 surface sediments (0–20 cm) along the coastline of China (not including Hong Kong, Macao, and Taiwan because of access difficulties in these regions), representing the three primary blue carbon ecosystems (Fu et al. 2021): mangroves, saltmarshes, and seagrass meadows. Surface sediments were targeted because they represent the most dynamic biogeochemical interface, where microbial communities actively process newly deposited terrestrial and autochthonous organic matter and regulate its transformation and stabilization before burial (Chen et al. 2022). From these, we selected 116 intact sediment samples (34 from mangroves, 57 from saltmarshes, and 25 from seagrass meadows; see Figure 1A; Table S1) to test CH₄ production potential through microcosm incubation experiments. The relative proportions of mangroves (27.6% vs. 29.3%), saltmarshes (51.4% vs. 49.1%), and seagrass meadows (21.0 vs. 21.6%) were preserved before and after filtering, with no statistically significant difference (χ ² = 0.15, p = 0.93, df = 2). Additionally, these samples were specifically selected because they represent a well‐paired dataset with corresponding sediment geochemical properties, carbon composition characteristics, and microbial community information, allowing an integrated assessment of the biogeochemical and microbial controls on CH₄ production.

For the microcosm incubation experiments, seawater was filtered using a 0.22 μm filter membrane and sterilized at 121°C for 30 min before being mixed with 3 g of sediment at a 1:1 ratio (w/v) in a 10 mL anaerobic tube. To create anaerobic conditions, four cycles of vacuum and charging high‐purity N₂ for durations of 10, 5, 3 and 2 min were performed. The anaerobic tubes were then incubated at 25°C in the dark. While this temperature is commonly used in laboratory incubations, it should be noted that in situ temperatures rarely remain stable at this set point. CH₄ accumulation was measured using a gas chromatograph (Agilent 7820A, USA) every 2 to 3 days. Given the high spatiotemporal variability of CH₄ production potential that complicates the estimates in vegetated coastal ecosystems (Roth et al. 2022), selecting a single time point to define the CH₄ production potential for all 116 samples seems arbitrary. Consequently, a dataset comprising over 2000 measurements was obtained after 51 days of incubation (Figure S1). Over 7‐week incubation period ensured that all samples reached or approached their maximum production potential (Figure S1). The CH₄ production potential was calculated for each sampling time point by dividing the accumulated CH₄ amount by the incubation duration. Both the mean and median values were similar (Table S2, rows 2 and 3), indicating that the data were not strongly skewed; thus, the mean value was used for subsequent analyses and statistical evaluations. Ancillary parameters used to elucidate the controls on CH₄ production potential included total organic carbon (TOC), particulate organic carbon (POC), mineral‐associated organic carbon (MAOC), pH, electrical conductivity (EC), total nitrogen (TN), total phosphorus (TP), iron oxides (Fep, organically bound Fe compounds; Feo, amorphous Fe compounds; Fed, free iron compounds), and particle diameter. Detailed analytical methods are provided in Supporting Information.

2.2. Metagenomic Analysis

All 116 samples of natural sediments were used for metagenomic analysis. DNA was extracted and used to construct Paired‐end (PE) libraries, which were sequenced using the Illumina Hiseq Xten platform. The resulting sequence data was deposited in the National Microbiology Data Centre of China under accession number NMDC10019198. Sequences were trimmed, decontaminated and assembled de novo on a Linux server; further details are provided in the Supporting Information. Open reading frames of the contigs were predicted, dereplicated and annotated by the eggNOG database for function and KEGG assignments. The quantification of genes was normalized to copies per million reads for downstream analysis. Metagenome‐assembled genomes (MAGs) were obtained and quantified in the same manner using MetaWRAP. Taxonomic classifications and gene functions were annotated with pre‐built databases.

To determine whether lineage‐closed microbial taxa related to CH₄ production potential can cluster, the 116 samples were classified into three categories: high (38 samples), medium (40 samples), and low (38 samples), based on their CH₄ production potential, from maximum to minimum. The first third of the samples, comprising 8, 26 and 4 samples for mangroves, saltmarshes, and seagrass meadows respectively, were classified as “high”. In contrast, the latter third, consisting of 12, 13 and 13 samples for mangroves, saltmarshes, and seagrass meadows respectively, were classified as “low”. The medium category, which includes 14, 18 and 8 samples from mangroves, saltmarshes, and seagrass meadows, respectively, is situated between the “high” and “low” classifications.

2.3. Lignin Phenols Analysis

Extraction of lignin phenols was conducted with a copper oxidation procedure, collecting vanillyls (V; vanillin, acetovanillone, vanillic acid), syringyls (S; syringaldehyde, acetosyringone, syringic acid), and cinnamyls (C; p‐coumaric acid, ferulic acid) units (VSC). Approximately 1 g of sediment was oxidized with 1 g of CuO, 100 mg (NH₄)₂Fe(SO₄)₂·6H₂O and 15 mL of nitrogen‐purged NaOH (2.0 M) in teflon lined bombs. All bombs were flushed with high purity nitrogen for 15 min and heated at 170°C for 2.5 h. The resulting extracts (lignin oxidation products, LOPs) were spiked with a surrogate standard (ethyl vanillin), acidified to pH 1 using 6 M HCl solution, and maintained in the dark (1 h, 25°C). LOPs were subjected to liquid–liquid extracted with ethyl acetate (10 mL), repeated three times for a combined extract, then concentrated under a constant nitrogen‐blowing. The LOPs were derivatized with N,O‐bis‐(trimethylsilyl) trifluoroacetamide (BSTFA, 400 μL) and pyridine (100 μL) for 3 h at 70°C to yield derivatives. These derivatives were analyzed using gas chromatography‐mass spectrometer (GC–MS). In brief, biomarkers of interest were quantified using internal standards (8 kinds of VSC units and ethyl vanillin) on a Trace 1310 gas chromatograph coupled to a mass spectrometric detectors using a TGSil‐5MS column (30 m × 0.25 mm × 0.25 μm). We used a constant current mode with a flow rate of carrier gas (high‐purity He) at 1.0 mL/min. The oven temperature was maintained at 65° for 2 min, then increased by 6°C per minute from 65°C to 300°C, followed by a final isothermal period at 300°C for 20 min. The mass spectrometer operated in electron impact mode (EI) at 70 eV and scanned over a mass‐to‐charge ratio (m/z) range of 35–500. Individual compounds were identified by comparing their mass spectra. In calculating the total lignin phenols, it is assumed that the release efficiencies for V, S, and C phenols are 33.3%, 90%, and 100%, respectively (Hautala et al. 1997).

2.4. Amino Sugar Analysis

Initially, a sediment sample weighing 1.0 g was hydrolyzed with 5 mL of 6 M HCl. After 2 min of nitrogen‐blowing, the samples were kept at 105°C for 8 h. The internal standard, myo‐inositol (250 μg), was added to quantify amino sugars. Following vortex shock for 30 s, the hydrolysate was centrifuged at 8000 rpm for 1 min, and 1 mL of the supernatant was collected and dried under nitrogen. The residue was re‐dissolved in 20 mL of MilliQ water, with the pH adjusted to 6.6–6.8 using KOH (0.4 M) and HCl (0.01 M). After centrifugation at 4000 rpm for 10 min, the supernatant was rotary‐dried at 65°C. Amino sugars were re‐dissolved in methanol and separated from salts by centrifugation at 4000 rpm for 10 min. The supernatant was mixed with adonitol (100 μg) and MilliQ water (1 mL) and then freeze‐dried. A mixed quantitative standard was prepared: including glucosamine (1 mg/mL), galactosamine (1 mg/mL), mannosamine (0.5 mg/mL) and muramic acid (0.25 mg/mL). The derivatization reagent (0.3 mL) was added to a 5 mL vial containing either the dry sample or the mixture standards. After shaking for several seconds, the solution was heated for 30 min at 80°C. The derivatives were further acetylated with acetic anhydride (1 mL) at 80°C for 25 min and mixed with dichloromethane (1.5 mL) after cooling. Excessive derivatization reagents were removed by extracting with HCl (1 M, 1 mL) and MilliQ water (1 mL), repeating three times, while the dichloromethane phase containing amino sugar derivatives was dried under a stream of high‐purity nitrogen condition. The resulting derivatives were analyzed using a TRACE 1300 gas chromatography. The oven temperature was held at 120°C for 1 min, increased to 230°C at a rate of 10°C per min, further increased to 250°C at a rate of 3°C per min, and finally increased to 300°C at a rate of 40°C per min, with a final isothermal hold at 300°C for 5 min. The mass spectrometer operated in the electron impact mode (EI) at 70 eV and scanned from 35 to 500 m/z. Finally, amino sugar data were used to calculate microbial necromass carbon (Liang et al. 2019). In brief, bacterial and fungal necromass carbon was quantified by determining specific and total amino sugars, including muramic acid [MurN], glucosamine [GluN], galactosamine [GalN], and mannosamine [ManN]. Bacterial residual carbon was estimated by multiplying the content of muramic acid by 45. Fungal residual carbon was calculated by subtracting bacterial‐derived glucosamine from total glucosamine, as follows:

$$\text{Bacterial necromass carbon}=45\times m$$ (1)

$$\text{Fungal necromass carbon}=\left(n/179.17-2\times m/251.23\right)\times 179.17\times 9$$ (2)

Here, m and n represent MurN and GluN, respectively; in Equation (1), 45 represents the conversion factor from MurA to bacterial necromass carbon; in Equation (2), 179.17 and 251.23 are the molecular weights of GlcN and MurA, respectively, while 9 is the conversion factor from fungal GlcN to fungal necromass carbon. Microbial necromass carbon was estimated as the sum of fungal and bacterial necromass carbon.

2.5. Lignin Addition and Carbon Isotopic Fractionation

To further investigate the contribution of lignin as substrates for CH₄ production, we conducted an additional experiment involving lignin addition. A mixture of 1 g of sediment and 0.05 g of lignin (Shanghai Aladdin Biochemical Technology Co. Ltd.) was incubated in a 10 mL anaerobic tube, with a sediment‐to‐water ratio of 1:1. Two control experiments were conducted simultaneously: (i) a mixture without lignin addition and (ii) a sterilized mixture. To create anaerobic conditions, four cycles of vacuum and high‐purity N₂ charging (10, 5, 3 and 2 min) were conducted. After placing the anaerobic tubes at 25°C in the dark for 1 month, the produced gases were collected and injected into a gas chromatography (Agilent 7820A, USA) for analysis. CH₄ stable carbon isotope composition was tested using a gas chromatograph combustion isotope ratio mass spectrometer (GC‐C‐IRMS) system (details see Xiao et al. 2020). In brief, approximately 1 mL of mixed gas was injected into a 100 mL sample container filled with helium gas (99.999% purity). Helium served as the carrier gas, transporting the mixed gases into a chemical trap that can remove CO₂ from the mixture. CH₄ was oxidized in a combustion reactor at 960°C, converting it into CO₂ and water. The combusted CO₂ was subsequently purified by two liquid nitrogen cold traps with internal filling of Ni wires before being transferred to the IRMS for determination. The abundance of ¹³C in a sample is given relative to a standard using the δ notation:

where PDB refers to the Pee Dee Belemnite carbonate that is used as standard which has a ¹³C/¹²C ratio of 0.0112372.

2.6. Statistical Analysis

The Shannon index, Inverse Simpson index, NMDS and PCoA analysis were conducted in R using the package vegan (v2.6.4) based on the relative abundance of MAGs. The similarity among different ecosystems was assessed by PERMANOVA with the adonis function in vegan (v2.6.4). The normality of the parameters was assessed using the shapiro.test function in R (version 4.3.2). Of the 33 parameters analyzed, only 2 were normally distributed (Table S3). Therefore, non‐parametric tests were used for this study. The kruskal.test function in R was used to test whether significant differences exist among the ecosystems. Pairwise comparisons were performed with the kwAllPairsNemenyiTest function in the PMCMRplus package (1.9.10) with the Bonferroni method to adjust p‐values. A p‐value < 0.05 was considered statistically significant. Spearman rank correlations were conducted to compare the influence of abiotic and biotic factors on CH₄ production potential. The p‐values were adjusted using the Benjamini‐Hochberg procedure to control the false discovery rate. Variation partitioning analyses were performed with the vegan package in R to compare the variance explained by linear models that incorporated all possible combinations of variables serving as proxies for carbon resources, microbial composition, and sediment physiochemical properties. The carbon resource variables included the contents of lignin phenols, cinnamyl monomers, syringyl monomers, vanillyl monomers, microbial necromass carbon, bacterial and fungal necromass carbon. The microbial community variables comprised the Shannon and Inverse Simpson index of the MAGs, as well as the first two components of PCoA based on Bray‐Curtis distances among the MAGs. The sediment physiochemical property variables included TOC, MAOC, POC, pH, electric conductance, total nitrogen, C/N ratio, total phosphorus, C/P ratio, Fed, Feo, Fep, ø, ø4, ø4–63, ø63.

Author Contributions

Isaac R. Santos: writing – original draft, writing – review and editing, methodology. Jian Liu: visualization, writing – review and editing, software, data curation. Jinkuo Lin: software, writing – review and editing. Leilei Xiao: conceptualization, methodology, data curation, investigation, validation, formal analysis, funding acquisition, visualization, writing – original draft, writing – review and editing. Carlos M. Duarte: writing – original draft, writing – review and editing, methodology. Kai Xiao: writing – review and editing, methodology. Chuancheng Fu: methodology, data curation, investigation, validation, writing – review and editing. Yongming Luo: conceptualization, writing – review and editing, project administration, supervision, resources, funding acquisition. Meng Zhou: investigation, writing – review and editing. Lifeng Zhou: investigation, writing – review and editing, data curation, visualization. Run Dang: investigation, writing – review and editing, data curation. Guangxuan Han: conceptualization, resources, supervision, writing – review and editing, project administration, funding acquisition.

Funding

This work was supported by National Key Research and Development Program in China (2022YFF0802101), National Natural Science Foundation of China (U2106209, 42077025, 42277236, 41991330), Youth Innovation Promotion Association of the Chinese Academy of Sciences (2021213), Yantai Institute of Coastal Zone Research, Chinese Academy of Sciences (YICE3510303), and the Ocean Negative Carbon Emissions (ONCE) Program.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Location of the sites and mean CH₄ production rates.

Table S2: CH₄ production rates calculated from sediment incubation experiments across three blue carbon ecosystems.

Table S3: Normal distribution test.

Figure S1: CH₄ production potential based on sediment incubation experiments conducted across three blue carbon ecosystems.

Figure S2: The carbon content derived from subunits of plant residues and microbial necromass.

Figure S3: A genome‐resolved view of microbial communities based on geographical distribution.

Figure S4: Microbial diversity across three ecosystems at three levels: genes, MAGs and reads.

Figure S5: MAGs in saltmarshes positively correlated with CH₄ production potential.

Figure S6: The microbial pathway associated with polysaccharide metabolism and CH₄ production.

Figure S7: Potential pathways for the decomposition of lignin‐like compounds.

Figure S8: Spearman rank correlations between functional genes (shown in Figure S7) and CH₄ production potential.

Figure S9: Potential methanogenic archaea and associated pathways.

Figure S10: Genes for key enzymes involved in methanogenic pathways.

Figure S11: Variation partitioning analysis (VPA) used to distinguish the relative contributions of plant‐derived residue carbon and microbial necromass carbon to CH₄ production.

Data Availability Statement

All data supporting the findings are available in the Figshare data repository (https://figshare.com/projects/Coastal/226524). Raw sequencing data that support the findings of this study have been deposited in the National Microbiology Data Centre of China (https://nmdc.cn/resource/genomics/project/detail/NMDC10019199) with the accession code: NMDC10019199.