Brackish Water Rewetting Enables Resilient Methane Suppression across Coastal Peatland Land Uses

Abstract

Peatland rewetting with freshwater reduces carbon dioxide (CO₂) emissions but often increases methane (CH₄), undermining its climate benefits. Brackish water, rich in sulfate (SO₄ ^2–), offers a potential alternative for CH₄ mitigation. However, its effectiveness and dominant mechanisms across different coastal land uses remain unclear, hindering broader application. Here, we conducted a four-month incubation experiment using intact soil cores from eight drained coastal peatlands representing four land-use types: grass-cut, grass-graze, arable, and unmanaged. We assessed CH₄ emissions from soils rewetted with either natural brackish water (SO₄ ^2– = 392 ± 2 mg L^–1) or synthetic water containing 0% (simulating freshwater), 50%, or 100% of the natural SO₄ ^2– concentration. Results showed that natural brackish water significantly suppressed CH₄ emissions by ∼90% compared to freshwater, with stronger suppression in grass-cut, arable, and unmanaged soils, but limited effects in grass-graze sites. Such differences were attributed to land-use-induced soil variations, where higher pH and lower methanogen relative abundance enhanced suppression. Mechanistically, natural brackish water primarily suppressed CH₄ via physiological inhibition of methanogens (likely through salt stress and hydrogen sulfide toxicity). This effect persisted after SO₄ ^2– reduction inhibition (day 90) and lasted until the end of incubation, reflecting resilience to short-term SO₄ ^2– depletion. Our study advances understanding of CH₄ dynamics and informs targeted use of brackish water rewetting for CH₄ mitigation in coastal peatlands.

1. Introduction

Pristine peatlands are vital carbon sinks, storing approximately one-third (∼600 ± 100 GtC) of the world’s soil carbon within just ∼3% of the global land area. , However, since the 20th century, drainage for agriculture and forestry has degraded at least 15% of global peatlands. This degradation has turned them from carbon sinks to sources, releasing about 2 Gt of carbon dioxide (CO₂) annually and thereby exacerbating global warming. Rewetting drained peatlands by raising the water table is a cost-effective CO₂ mitigation method, but rewetting with freshwater often triggers high methane (CH₄) emissions due to the reestablishment of anaerobic conditions that favor methanogen growth. , These CH₄ emissions can partially offset the climate benefits of reduced CO₂, given CH₄’s 28-fold higher global warming potential over 100 years. Minimizing CH₄ emissions is therefore crucial for enhancing the greenhouse gas mitigation potential of peatland rewetting.

Coastal peatlands, due to their unique geographical locations, have access to different water sources for rewetting, including freshwater from rivers or groundwater, as well as brackish water from coastal estuaries or lagoons. Studies suggest that rewetting with sulfate (SO₄ ^2–)-containing brackish water may help reduce CH₄ emissions, − making it a potentially more effective strategy for restoring the carbon sink function compared to freshwater rewetting. The suppression of CH₄ emissions by brackish water rewetting can be attributed to several key mechanisms. First, the elevated SO₄ ^2– concentration may promote the growth of sulfate-reducing bacteria (SRB), which outcompete methanogens for common substrates such as hydrogen (H₂) and acetate (CH₃COO^–) due to their higher substrate affinity , and greater energy yield during SO₄ ^2– reduction. , Additionally, the produced CH₄ may be oxidized by certain anaerobic methanotrophic archaea (ANME), which form symbiotic consortia with SRB, utilizing SO₄ ^2– as the terminal electron acceptor to drive anaerobic oxidation of methane (AOM). , Moreover, methanogen populations may be physiologically suppressed by hydrogen sulfide (H₂S) toxicity, , which is the primary product of SO₄ ^2– reduction, as well as by salt stress resulting from the high concentrations of sodium (Na⁺) and chloride (Cl^–) ions in brackish water.

Although different mechanisms may result in similar overall effects of CH₄ suppression when SO₄ ^2– is abundant, their short-term responses differ when SO₄ ^2– becomes depleted due to factors such as heavy rainfall, groundwater infiltration, or insufficient brackish water recharge. , Specifically, CH₄ suppression via mechanisms that physiologically inhibit methanogens (i.e., H₂S toxicity and salt stress) could persist in the short term even after SO₄ ^2– depletion, as methanogens require time to recover and cannot immediately respond to environmental changes. , In contrast, CH₄ suppression driven by SO₄ ^2–-derived pathways, including substrate competition and sulfate-dependent AOM (sulfate-AOM), may wane more rapidly due to a sharp decline in SO₄ ^2– reduction efficiency caused by SO₄ ^2– depletion, leading to a faster recovery of CH₄ emissions. , Therefore, the dominant suppression mechanisms significantly affect the short-term stability of CH₄ suppression following natural disturbances. Given that short-term emission fluctuations can accumulate over timeparticularly under increasingly variable climatic conditionsthese mechanisms also play a crucial role in shaping the long-term greenhouse gas balance and climate feedbacks of rewetted coastal peatlands.

Nevertheless, while field-based evidence of suppressed CH₄ emissions in brackish water-rewetted coastal peatlands has been reported, ,, current knowledge of the dominant suppression mechanisms remains limited. Additionally, drained coastal peatlands may have undergone different management practices, such as arable farming, grazing, mowing, or abandonment, which can lead to distinct soil physicochemical and microbial properties. , Given that CH₄ emissions are highly microbially driven and closely linked to soil characteristics, both the extent of CH₄ suppression and the dominant suppression mechanisms are expected to vary across land-use types. This potential variability may influence the broad applicability of brackish water rewetting as a mitigation strategy in coastal areas, necessitating further investigation.

In this study, we selected eight drained peatland sites representing four land-use types on the west coast of Jutland, Denmark. These land-use types included grass-cut (grassland managed for harvesting), grass-graze (grassland used for livestock grazing), arable (land actively cultivated for row-cropping), and unmanaged (land left to undergo natural succession without human intervention). We collected undisturbed soil cores and conducted a four-month incubation experiment, rewetting them with water from different sources containing varying SO₄ ^2– concentrations. Later in the incubation, we introduced a specific SO₄ ^2– reduction inhibitor to examine how CH₄ emissions respond to SO₄ ^2– depletion. We aimed to identify the key soil factors influencing the extent of CH₄ suppression and to determine the dominant suppression mechanisms and their variability across land-use types. This study will provide valuable insights into the biogeochemical processes governing CH₄ dynamics and support the implementation of brackish water rewetting in coastal peatlands as an effective climate mitigation strategy.

2. Materials and Methods

2.1. Site Description and Sampling

Denmark has a temperate to cold climate (Cfb/Dfb, Köppen–Geiger), with annual averages of 8.7 °C and 759 mm precipitation during 1991–2020. It contains approximately 291,000 ha of organic soils, 60% of which are used for agriculture, mainly as grassland or cropland. The country has a 1701 km general coastline (excluding most offshore islands and the 180 km-long Limfjord), with no location more than 52 km from the coast. Coastal areas are generally shallow with diverse physical, chemical, and biological conditions; salinity is higher on the western side facing the North Sea and lower on the eastern side toward the Baltic Sea. Soil sampling was conducted on the west coast of Jutland near Ringkøbing Fjord, a shallow coastal lagoon. The water temperatures in Ringkøbing Fjord, influenced by North Sea conditions, have a mean annual value of 10.3 °C, with a minimum of 3.2 °C in February and a maximum of 18.2 °C in August.

To capture variations across land uses, we selected two drained peatland sites from each of four categories: grass-cut, grass-graze, arable, and unmanaged (Figure ). These sites have generally remained in the same land use category without rotation for decades. Most sites were situated within 3 km of the coastline, except for one arable site located slightly farther but still within 5 km of the coast. Grass-cut sites were dominated by grass species, likely including Lolium perenne, Phleum pratense, Poa pratensis, and Festuca rubra, as well as clover species (Trifolium repens). Grass-graze sites featured similar grass dominance, primarily L. perenne and F. rubra. Arable sites were cultivated with either Pisum sativum or Hordeum vulgare during the sampling period. Unmanaged sites exhibited heterogeneous vegetation: one site was dominated by herbaceous species, including Urtica dioica, Cirsium arvense, and grasses (likely L. perenne and F. rubra), while the other was dominated by woody species, mainly Betula pubescens, alongside mixed Poaceae grass species. Detailed coordinates of the study sites and their distances from the sea are provided in Table S1.

1.

Locations of eight coastal sampling sites, categorized into four land-use types and indicated by different colors. Squares mark the I sites, while triangles mark the II sites. NBW, natural brackish water.

To reduce variability among replicate samples and target the most biologically active soil layer, we removed visible aboveground vegetation and extracted 18 surface soil cores from each site, using stainless steel rings (radius: 3 cm; height: 8.7 cm). Roots were left in place to avoid disturbing the natural soil structure. To support incubation and further analyses, we also collected natural brackish water from the Ringkøbing Fjord (Figure ) and additional fresh soil samples from the sites. The collected soil cores and natural brackish water were immediately transported and stored at 2 °C before incubation. Additional fresh soil samples were frozen at −70 °C prior to analysis. The sampling was completed in 1 day (June 12, 2024), and the incubation began 9 days later (June 21, 2024).

2.2. Rewetting Water Preparation

To characterize natural water chemistry and prepare for the rewetting experiment, we analyzed the collected natural brackish water for pH, electrical conductivity at 25 °C (EC₂₅), and concentrations of dissolved organic carbon (DOC), ammonium (NH₄ ⁺), nitrate (NO₃ ^–), sulfate (SO₄ ^2–), and chloride (Cl^–). Analyses were conducted directly after filtration and dilution using the same instruments employed for soil analysis, as detailed in the subsequent methodology section. The measured values (mean ± standard deviation) were as follows: pH (7.77 ± 0.01), EC₂₅ (19.05 ± 0.04 mS cm^–1), DOC (14.75 ± 2.71 mg L^–1), NH₄ ⁺ (0.09 ± 0.01 mg L^–1), NO₃ ^– (0.01 ± 0.02 mg L^–1), SO₄ ^2– (392 ± 2 mg L^–1), and Cl^– (4085 ± 16 mg L^–1). The natural brackish water SO₄ ^2– concentration of 392 mg L^–1 (C) was used as a reference to prepare synthetic brackish water with two SO₄ ^2– levels: 100% (C; 392 mg L^–1) and 50% (C/2; 196 mg L^–1), using sodium sulfate (Na₂SO₄) and deionized water. To simulate freshwater with limited SO₄ ^2–, pure deionized water (0% SO₄ ^2–) was used.

2.3. Experimental Incubations

To investigate the effects of different rewetting strategies on CH₄ emissions across land-use types, we established a laboratory incubation experiment. Specifically, 15 of the 18 soil cores from each site were randomly assigned into five groups, with three cores per group serving as replicates. One group remained drained throughout the incubation; three groups were rewetted with synthetic water at three SO₄ ^2– levels: 0% (Syn0), 50% (Syn50), and 100% (Syn100); and the final group was rewetted with natural brackish water (Nat100). The water level in drained treatments was maintained at −8 cm, while in rewetted treatments, it was adjusted to the soil surface (0 cm). Soil cores were incubated individually in sealed 2.41 L glass jars equipped with rubber septa for gas sampling, leaving a headspace of 2.16 L for drained treatments and 1.28 L for rewetted treatments. In total, this resulted in 120 incubation jars: 30 jars per land-use type, 24 jars per treatment, and 6 jars per land use–treatment combination (3 from each site). Additionally, four jars containing only natural brackish water were used as controls. All jars were initially incubated in the dark at 20 °C for three months (90 days).

To examine the response of CH₄ emissions under different rewetting strategies to natural disturbances, we introduced a second incubation phase after the initial 90 day period. On day 90, one site from each land-use category with relatively higher CH₄ emissions, designated as the “I site,” was selected for altered incubation conditions, while the other site (the “II site”) remained unaltered. For the I sites, rewetted treatments were amended with either Na₂SO₄ or sodium molybdate (Na₂MoO₄). Specifically, 10 mL of 0.412 M Na₂SO₄ was added to the Syn0 treatments to achieve a final SO₄ ^2– concentration of approximately 392 mg L^–1, aiming to reestablish SO₄ ^2– reduction. For the Syn50, Syn100, and Nat100 treatments, 10 mL of 2 M Na₂MoO₄ was added, resulting in a final MoO₄ ^2– concentration of approximately 20 mM to inhibit SO₄ ^2– reduction. , The conditions for the drained treatments remained unchanged. All incubations continued in the dark at 20 °C for an additional month (30 days), resulting in a total incubation period of 120 days.

2.4. CH₄ Measurement

To monitor CH₄ dynamics throughout the incubation, we conducted gas sampling on days 1, 4, 8, 11, 15, 18, 22, 25, 29, 36, 43, 50, 60, 75, 90, 99, 109, and 120. During each sampling, 10 mL of headspace gas was withdrawn from each jar using a syringe fitted with a hypodermic needle, after pushing and pulling the syringe plunger several times to thoroughly mix the headspace gas. The gas samples were then transferred to evacuated 6 mL Exetainer vials (Labco Ltd., Lampeter, UK) for analyzing CH₄ concentrations using a gas chromatograph (8890 GC System, Agilent, Santa Clara, USA). Concentrations measured on day 1 were used as the initial baselines to correct the measurements on day 4. Starting from day 4, after each sampling, the jars were opened for 10 min to equilibrate with ambient air with the help of two small electric fans. After closing the jars, six 10 mL samples of ambient air were taken from a height of 10 cm above the jars to document a new baseline concentration for the next sampling. To minimize any influence from exhaled air, the experimenter wore a disposable face mask and sampled from the side of the incubation table, keeping their head outside the sampling zone and consciously avoiding breathing toward it. Baseline measurements showed no irregularities, indicating that contamination did not occur.

CH₄ emissions for each incubation interval ( $E_{{\mathrm{C}\mathrm{H}}_4,T}$ ) were calculated as follows

$$E_{{\mathrm{C}\mathrm{H}}_4,T}(\mu \mathrm{m}\mathrm{o}\mathrm{l})=\frac{(C_t-C_0)\times V_{\mathrm{h}}}{V_{\mathrm{m}}}$$ 1

where C _t is the headspace CH₄ concentration (μL L^–1) measured at each sampling, C ₀ is the corresponding baseline concentration (μL L^–1), V _h is the effective headspace volume (L), and V _m is the molar gas volume at 20 °C (24 L mol^–1).

The area-scaled emission rates (fluxes) for each incubation interval and the area-scaled cumulative emissions throughout the incubation were calculated as follows

$$\mathrm{F}\mathrm{l}\mathrm{u}\mathrm{x}(\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}{\mathrm{m}}^{-2}{\mathrm{d}}^{-1})=\frac{E_{{\mathrm{C}\mathrm{H}}_4,T}}{T\times S}\times {10}^4$$ 2

$$\mathrm{C}\mathrm{u}\mathrm{m}\mathrm{u}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}(\mathrm{n}\mathrm{m}\mathrm{o}\mathrm{l}{\mathrm{m}}^{-2})=\frac{\sum E_{{\mathrm{C}\mathrm{H}}_4,T}}{S}\times {10}^4$$ 3

where T is the duration (days) of each incubation interval, S is the surface area of the soil cores (28.27 cm²), and $\sum E_{{\mathrm{C}\mathrm{H}}_4,T}$ is the sum of the individual interval emissions.

To facilitate comparison between I and II sites, we calculated mean CH₄ fluxes over the initial incubation period as cumulative emissions divided by 90 days. To assess the immediate CH₄ response to altered incubation conditions, we calculated changes in CH₄ fluxes (ΔR) as the difference between the linear slope of cumulative emissions at days 90 and 99 and that fitted over days 60, 75, and 90.

As a key target gas in rewetting activities, CO₂ was measured simultaneously and processed in the same way, with results presented in Figures S1 and S2.

2.5. Soil Measurement

To determine soil physicochemical properties, the remaining three cores from each site were oven-dried at 60 °C for 1 week, then ground and sieved to <2 mm for the measurement of bulk density, gravimetric water content, pH, EC, total organic carbon (TOC), total nitrogen (TN), SO₄ ^2–, Cl^–. Bulk density was determined by dividing the dry soil weight by soil core volume. Water content was calculated by dividing the difference between wet and dry soil weight by dry weight. The pH and EC were measured at a soil-to-water ratio of 1:2.5 (w/v) using a multiparameter meter (Eutech PC 450, Thermo Fisher Scientific, Waltham, MA, USA). The measured EC was corrected to EC at 25 °C (EC₂₅) using the temperature correction equation: EC₂₅ = EC_t/[1 + 0.021 (t-25)]. Total carbon and TN were determined using an elemental analyzer (VarioMax, Elementar GmbH, Hanau, Germany). The presence of carbonate in the dry soil samples was tested by HCl effervescence. All soils were free of inorganic carbonates, and thus the total carbon content was considered TOC content. SO₄ ^2– and Cl^– were extracted with deionized water at a soil-to-water ratio of 1:5 (w/v) and determined using an ion chromatography (Eco IC, Metrohm, Herisau, Switzerland) after filtration through 0.45 μm membrane filters.

Additional fresh soil samples were used for measuring DOC, NH₄ ⁺, NO₃ ^–, and microbial community structure. DOC was extracted using 0.5 M K₂SO₄ at a soil-to-solution ratio of 1:5 (w/v) and determined using a TOC analyzer (TOC-VCPH, Shimadzu, Kyoto, Japan) after filtration (0.45 μm). NH₄ ⁺ and NO₃ ^– were extracted with 1 M KCl at a soil-to-solution ratio of 1:10 (w/v) and determined using a segmented flow analyzer (AA500 AutoAnalyzer, SEAL Analytical, Southampton, UK) after filtration through glass fiber filters. DOC, NH₄ ⁺, and NO₃ ^– concentrations were corrected to dry-weight values based on the water content.

To characterize bacterial and archaeal community composition, total genomic DNA was extracted and purified from 0.25 g of fresh soil and eluted to a final volume of 100 μL using the TGuide S96 Automated Nucleic Acid Extractor and the TGuide S96 Magnetic Soil/Stool DNA Kit (Tiangen Biotech, Beijing, China), following manufacturer instructions. The DNA concentration was measured using the Qubit dsDNA HS Assay Kit and Qubit 4.0 Fluorometer (Invitrogen, Thermo Fisher Scientific, Oregon, USA). The 16S rRNA gene in bacteria (V3–V4 region) and archaea (V4–V5 region) was amplified using the primer sets 338F/806R and Arc519F/Arc915R, respectively. High-throughput sequencing for 16S rRNA genes and subsequent data processing were performed at Biomarker Technologies (BMK) GmbH, Münster, Germany. Nonmetric multidimensional scaling (NMDS) of archaeal and bacterial communities was performed using the BMKCloud platform (www.biocloud.net) based on Bray–Curtis distances at the species level. All sequence reads are available in the National Center for Biotechnology Information (NCBI) Sequence Read Archive under BioProject number PRJNA1244897 (bacteria) and PRJNA1244914 (archaea). Detailed method for 16S rRNA gene sequencing is provided in Supporting Information Method 1.

After the four-month incubation period, soil cores from the II sites were removed from the jars to measure the same parameters, excluding bulk density. Additionally, the SO₄ ^2– concentration of the rewetting water in the jars was measured on day 90 for I sites and on day 120 for II sites. All measurements were performed in triplicate.

2.6. Statistical Analysis

To assess the effects of land-use types and rewetting treatments on mean CH₄ fluxes, we performed linear mixed-effects models (LMMs) in R (version 4.4.2). Land use and treatment, along with their interaction, were included as fixed effects, while site was included as a random effect. LMMs were applied using the “lmer” function from the “lme4” package. Residual normality and homoscedasticity were visually assessed, and data were log-transformed if assumptions were violated. When significant effects were detected through Type III analysis of variance (Type-III ANOVA), pairwise comparisons of least-squares means (computed with the “emmeans” package) were conducted using Tukey’s adjustment for multiple comparisons.

To evaluate the effects of initial soil properties (i.e., preincubation conditions) and rewetting strategies on CH₄ suppression, we conducted Spearman correlation analysis and multiple linear regression (MLR). Spearman correlation analysis was performed in OriginPro 2024 (version 10.1.0.170), and MLR was conducted in R (version 4.4.2). Rewetting strategies were represented by two variables: SO₄ ^2– concentration in rewetting water, and natural brackish water rewetting (coded as 1 for Nat100 and 0 for Syn50 and Syn100). CH₄ suppression was quantified as the difference in mean CH₄ fluxes between the SO₄ ^2–-free treatment (Syn0) and the SO₄ ^2–-containing treatments (Syn50, Syn100, Nat100).

To address multicollinearity observed in the correlation analysis, we implemented a two-step preselection of explanatory variables prior to MLR. In Step 1, variables exhibiting perfect collinearity were identified using the “alias” function in R and iteratively removed based on their average Spearman correlation coefficient with other variables until perfect collinearity was eliminated. In Step 2, the remaining variables from Step 1 were further filtered by iteratively removing those with the highest variance inflation factor (VIF) until all remaining variables had VIF <10. The final set of explanatory variables was used for MLR. MLR was performed by initially including all main effects and two-way interaction terms between soil properties and rewetting strategies, followed by backward selection with P < 0.05 as the stopping criterion. If more than one variable remained after backward selection, their relative importance was evaluated by comparing the change in the model’s Akaike Information Criterion (AIC) when each variable was individually removed. Moreover, the key soil properties identified through MLR were incorporated into simple linear regressions, conducted separately for SO₄ ^2–-free and SO₄ ^2–-containing treatments, to further elucidate their roles in regulating CH₄ fluxes.

A significance level of 0.05 was applied to all statistical tests.

3. Results and Discussion

3.1. Brackish Water Rewetting Suppressed CH₄ Emissions, but Varied across Land-Use Types

CH₄ fluxes from the two-phase incubation experiment (a 90 day initial period followed by a 30 day altered phase at I sites) are presented in Figure , with cumulative emissions shown in Figure S3. During the initial period, CH₄ fluxes in the drained treatments generally remained low or even negative, aligning with expectations. However, elevated fluxes were observed at two grass-graze sites (Figure b), possibly due to a higher initial relative abundance of methanogens, as detailed in Section . In the rewetted treatments, CH₄ fluxes exhibited a lag phase of 1–2 months before activation and gradual increase, with SO₄ ^2–-free treatments (Syn0) generally showing an earlier and faster increase than SO₄ ^2–-containing treatments (Syn50, Syn100, and Nat100).

2.

CH₄ fluxes during incubation. (a–d) CH₄ fluxes from different land-use types. The I sites were subjected to altered incubation conditions on day 90. Shaded area represents mean ± standard error of three replicates. (e–h) Mean CH₄ fluxes from different land-use types, calculated based on 90 day cumulative emissions. Error bars represent mean ± standard error of three replicates. Different letters next to the legend indicate statistically significant differences among treatments (averaged across land-use types), while different letters above the bars indicate significant differences among treatments within each land-use type, demonstrating a land use × treatment interaction effect (P < 0.05, LMMs). Syn0, Syn50, and Syn100 refer to treatments rewetted with synthetic water containing 0%, 50%, and 100% of the natural sulfate concentration, respectively; Nat100 was rewetted with natural brackish water.

LMMs were applied to evaluate the main effects of land use and rewetting treatment, as well as their interaction, on mean CH₄ fluxes (Figure e–h). Regarding the main effects, land use had no significant influence (P > 0.05), whereas rewetting treatment had a strong impact (F _4,96 = 15.76, P < 0.001). Specifically, rewetting with SO₄ ^2–-free water (Syn0) significantly increased mean CH₄ fluxes compared to the drained treatments (estimated means: 89.54 ± 45.00 vs 6.60 ± 6.32 nmol m^–2 d^–1; t ₉₆ = 6.93, P < 0.001). Nevertheless, rewetting with SO₄ ^2–-containing water resulted in a smaller increase in mean CH₄ fluxes. Compared to Syn0, the 50% SO₄ ^2– treatments (Syn50) exhibited reduced CH₄ fluxes (37.26 ± 20.60 nmol m^–2 d^–1), but the reduction was not significant. In contrast, the 100% SO₄ ^2– treatments, whether rewetted with synthetic (Syn100) or natural (Nat100) brackish water, showed significantly lower CH₄ fluxes than Syn0 (Syn100:20.66 ± 12.90 nmol m^–2 d^–1, t ₉₆ = −4.42, P < 0.001; Nat100:9.28 ± 7.57 nmol m^–2 d^–1, t ₉₆ = −6.29, P < 0.001). This demonstrates that rewetting coastal peatlands with brackish water containing higher SO₄ ^2– concentrations is more effective in suppressing CH₄ emissions, in line with previous research demonstrating a negative relationship between CH₄ emissions and SO₄ ^2– levels in aquatic ecosystems. −

Beyond the main effects, a significant interaction between land use and rewetting treatment was detected (F _12,96 = 3.59, P < 0.001), demonstrating that the effect of rewetting treatment on mean CH₄ fluxes was dependent on land use. Specifically, treatment effects were significant in grass-cut, arable, and unmanaged sites, with differences (P < 0.05) observed between Syn0 and drained treatments, as well as between Syn0 and Nat100 (Figure e,g,h). In contrast, no significant treatment effects were observed in grass-graze sites (Figure f). This indicates that, when different water sources are available, natural brackish water should be preferentially used for rewetting in grass-cut, arable, and unmanaged sites to mitigate CH₄ emissions, while its use in grass-graze sites may provide limited benefit.

3.2. Soil pH and Methanogen Abundance Influenced CH₄ Suppression Effectiveness

To better understand the observed variation in CH₄ suppression across land-use types and rewetting treatments, Spearman correlation analysis and MLR were used to examine associations with initial soil properties (measured prior to incubation; Figures S4–S6) and rewetting strategies (represented by SO₄ ^2– concentration in rewetting water and natural brackish water rewetting). The correlation analysis revealed significant relations between CH₄ suppression and multiple explanatory variables (Figure ). However, substantial intercorrelations among these variables were also observed, indicating potential multicollinearity. To mitigate this issue, explanatory variables were preselected using a two-step approach before being integrated into a backward stepwise MLR analysis (refer to Methods for details).

3.

Correlation analysis of initial soil properties, rewetting strategies, and CH₄ suppression. Correlation coefficients between initial soil properties and rewetting strategies are not shown due to their lack of physical significance. Larger circle size indicates stronger correlation. Significance levels: *P < 0.05, **P < 0.01, ***P < 0.001. BD, bulk density; WC, water content; EC₂₅, electrical conductivity at 25 °C; TOC, total organic carbon; TN, total nitrogen; DOC, dissolved organic carbon; NH₄ ⁺, ammonium; NO₃ ^–, nitrate; SO₄ ^2–, sulfate; Cl^–, chloride; Chao1_a, archaeal Chao1 index; Shannon_a, archaeal Shannon index; Chao1_b, bacterial Chao1 index; Shannon_b, bacterial Shannon index; SRB, sulfate-reducing bacteria; FRB, iron-reducing bacteria; SBM, syntrophic bacteria with methanogens; Water-SO₄ ^2–, sulfate concentration in rewetting water; NBW rewet, natural brackish water rewetting.

The MLR results showed that a final model including pH and methanogens as predictors significantly explained the variation in CH₄ suppression (adjusted R ² = 0.561, F _2,21 = 15.71, P < 0.001; Table ), highlighting the key influence of initial soil properties. Partial regression coefficients (b) revealed a significant positive effect of pH (b = 198 ± 37, t ₂₁ = 5.38, P < 0.001) and a significant negative effect of methanogens (b = −56 ± 24, t ₂₁ = −2.39, P = 0.026) on CH₄ suppression. This demonstrates that CH₄ suppression through brackish water rewetting is more effective in soils with higher pH and lower relative abundance of methanogens, while its effectiveness declines under lower pH and higher methanogen abundance. To further assess the relative contributions of these two soil factors, changes in AIC were compared after their removal from the model. Excluding pH or methanogens increased AIC by 18.8 and 3.78, respectively, indicating that pH exerted a stronger influence on CH₄ suppression. Therefore, pH may serve as a primary indicator for identifying coastal peat soils where brackish water rewetting can deliver greater CH₄ mitigation benefits. While this finding offers a promising direction, we acknowledge that it is based on a limited number of sampling sites, and further research across a broader range of coastal peatlands is needed to validate its general applicability.

1. Multiple Linear Regression Analysis of CH₄ Suppression with Potential Influencing Factors.

response variable	VIF-filtered explanatory variables	model	Adj. R 2	Sig.	AIC	key explanatory variables	b	SE	Sig.
CH4 suppression	pH, DOC, NH4 +, NO3 –, Cl–, Methanogens, Water-SO4 2–, NBW rewet	1	0.561	<0.001	298.00	(intercept)	–820.48	189.51	<0.001
						pH	198.08	36.81	<0.001
						methanogens	–56.47	23.62	0.026
		2	0.467	<0.001	301.78	(intercept)	–798.75	208.60	<0.001
						pH	184.31	40.07	<0.001
		3	0.004	0.309	316.80	(intercept)	186.26	45.46	<0.001
						methanogens	–36.60	35.15	0.309

^aDOC, dissolved organic carbon; NH₄ ⁺, ammonium; NO₃ ^–, nitrate; Cl^–, chloride; Methanogens, relative abundance of methanogens; Water-SO₄ ^2–, sulfate concentration in rewetting water; NBW rewet, natural brackish water rewetting.

^bModels 2 and 3 were compared with model 1 to evaluate the relative contributions of individual variables to model performance.

The positive effect of initial soil pH on CH₄ suppression in rewetted coastal peat soils was driven by a stronger increase in CH₄ emissions with rising pH under freshwater rewetting compared to brackish water rewetting (Figure S7a,c). In acidic peat soils, CH₄ emissions generally increase with pH, as near-neutral conditions enhance both methanogenic activity and upstream substrate availability. , This relationship holds in both freshwater , and brackish water , systems. However, as demonstrated in our results, the pH sensitivity is more pronounced under freshwater conditions, where favorable environmental conditions allow elevated pH to markedly increase CH₄ production. In contrast, methanogens in brackish systems experience stronger competition from SRB and additional stressors such as salinity, , which can inhibit CH₄ emissions even at higher pH. This difference in pH sensitivity may have contributed to a larger gap in CH₄ emissions between freshwater and brackish conditions at higher soil pH, thereby enhancing the CH₄ mitigation potential of brackish water rewetting.

Similarly, the negative effect of the initial relative abundance of methanogens on CH₄ suppression can be attributed to differences in CH₄ response sensitivity under freshwater and brackish water rewetting. Specifically, CH₄ emissions exhibited no clear correlation with methanogen abundance in freshwater treatments (Figure S7b), whereas a weak positive correlation was observed under brackish conditions (Figure S7d). This may help explain the nonsignificant treatment effects observed at grass-graze sites (Figure f), where relatively high initial methanogen abundance (Figure S6), particularly Methanobacteriaceae (Figure S8), could have contributed to elevated CH₄ emissions under brackish water rewetting, thereby diminishing its effectiveness in suppressing CH₄ emissions. The relatively high methanogen abundance in grass-graze sites may result from the combined effects of dung application, which could introduce exogenous methanogens and organic substrates, as well as trampling-induced soil compaction, which may reduce aeration and promote anaerobic conditions favorable to methanogen growth. These findings highlight the lasting impact of historical land uses of drained coastal peatlands on soil biogeochemistry, and consequently on the effectiveness of CH₄ suppression via brackish water rewetting.

3.3. Brackish Water Rewetting Facilitates Resilient CH₄ Suppression in Coastal Peat Soils

To examine the short-term response of CH₄ emissions to natural disturbances, we amended the rewetted treatments of I sites with either Na₂SO₄ or Na₂MoO₄ on day 90 and assessed the resulting immediate changes in CH₄ fluxes (ΔR). As shown in Figure a, similar response patterns were observed across land-use types. In the Syn0 treatments, the addition of Na₂SO₄ had minimal immediate effect on CH₄ fluxes (ΔR close to zero), likely due to the delayed response of SRB to the sudden increase in SO₄ ^2– concentration. In the Syn50 treatments, the addition of Na₂MoO₄ as a specific inhibitor of SO₄ ^2– reduction caused a rapid increase in CH₄ fluxes, with ΔR exceeding 600 nmol m^–2 d^–1. A similar but less pronounced increase was observed in the Syn100 treatments. In contrast, Nat100 treatments showed little to no change in CH₄ fluxes, except at the grass-graze site, where a modest increase was observed; however, this increase remained lower than those observed in the Syn50 and Syn100 treatments. These results demonstrate that rewetting with natural brackish water can not only reduce CH₄ emissions but also provide prolonged suppression effects in the event of SO₄ ^2– depletion due to natural disturbances. For II sites that did not undergo alteration, most treatments showed minimal changes in CH₄ fluxes after day 90 (Figure b), aligning with our expectations.

4.

Changes in CH₄ fluxes (ΔR) after day 90 for (a) I sites subjected to altered incubation conditions and (b) II sites maintained under unchanged incubation conditions. Shaded areas around the symbols represent the standard error of ΔR. Syn0, Syn50, and Syn100 refer to treatments rewetted with synthetic water containing 0%, 50%, and 100% of the natural sulfate concentration, respectively; Nat100 was rewetted with natural brackish water. (c) Conceptual model illustrating the mechanisms by which brackish water rewetting suppresses CH₄ emissions. Numbers ① to ④ indicate different suppression mechanisms: ① sulfate-reducing bacteria (SRB) outcompete methanogens for common substrates (electron donors); ②: sulfate-dependent anaerobic oxidation of methane (sulfate-AOM); ③: hydrogen sulfide (H₂S) toxicity to methanogens; ④: salt stress on methanogens.

To elucidate the microbial mechanisms of CH₄ suppression under brackish water rewetting, archaeal and bacterial communities were characterized after incubation, with similar trends observed across land-use types (Figure ). The relative abundance of methanogens was markedly lower in Nat100 than in Syn50 and Syn100 treatments across grass-cut, grass-graze, and arable sites (Figure c). In the unmanaged site, methanogens were more abundant overall, yet their levels in Nat100 remained comparable to or lower than those in Syn50 and Syn100 treatments. NMDS analysis (Figure e) and archaeal class composition (Figure f) further revealed distinct community shifts in Nat100 relative to Syn50 and Syn100 treatments. Despite having the same SO₄ ^2– concentration in rewetting water, Syn100 and Nat100 treatments differed in methanogen abundance and archaeal community structure. This suggests that factors beyond SO₄ ^2–, most likely the high salt content in natural brackish water, played a role. Additionally, the relative abundance of SRB increased progressively from Syn50 to Syn100 to Nat100 treatments (Figure d). This trend corresponded with enhanced SO₄ ^2– reduction efficiency (Table S2), potentially resulting in higher H₂S production in the Nat100 treatments. Compared to Syn50 and Syn100, the Nat100 treatments exhibited similar or reduced relative abundances of iron-reducing bacteria (FRB) and methanotrophs, suggesting a generally lower potential contribution of these bacterial groups to CH₄ suppression.

5.

Variations in microbial communities under different treatments after incubation at II sites. (a,b) Alpha diversity indices (Chao1 and Shannon) of archaeal (a) and bacterial (b) communities. Error bars represent mean ± standard deviation of three replicates. (c,d) Relative abundance of key CH₄-related microbial families (see details in Figure S8), including methanogens, sulfate-reducing bacteria (SRB), iron-reducing bacteria (FRB), methanotrophs, and syntrophic bacteria with methanogens (SBM). (e) Nonmetric multidimensional scaling (NMDS) of archaeal communities based on Bray–Curtis distances at the species level (NMDS of bacterial communities is presented in Figure S9). Ellipses represent the 95% confidence interval. (f,g) Relative abundance of the top 10 most abundant classes in archaeal (f) and bacterial (g) communities. Sample labels denote treatments: D, drained; S₀–S100, rewetted with synthetic water containing 0%, 50%, or 100% of the natural sulfate concentration; N100, rewetted with natural brackish water.

These findings demonstrate that the dominant mechanisms of CH₄ suppression differed substantially between natural and synthetic brackish water rewetting. In the Syn50 and Syn100 treatments, CH₄ suppression can be mainly attributed to substrate competition and/or sulfate-AOM (Figure c, panel (i). These mechanisms rely strongly on SO₄ ^2–-reducing activity and have a less direct influence on methanogens, thus allowing for quicker recovery of CH₄ fluxes once SO₄ ^2– reduction is inhibited. In contrast, CH₄ suppression in the Nat100 treatments (regardless of land use) can be largely explained by H₂S toxicity and salt stress (Figure c, panel (ii). these mechanisms act directly on methanogens rather than depending on active SO₄ ^2– reduction, allowing CH₄ inhibition to persist even after SO₄ ^2– reduction ceases. H₂S toxicity primarily arises from its ability to penetrate cell membranes and inhibit enzyme activity by denaturing proteins. It also contributes to the formation of insoluble metal sulfides, thus reducing the bioavailability of essential metals required by methanogens. Salt stress mainly acts by increasing osmotic pressure and reducing microbial enzyme activity and gene expression. Methanogens are generally more sensitive to salt stress than SRB. , This may have led to a preferential inhibition of methanogens in the Nat100 treatments, giving SRB a stronger competitive advantage compared to those in the Syn100 treatments (Figure d). The enhanced competitiveness of SRB could, in turn, increase the reduction efficiency of SO₄ ^2– to H₂S (Table S2), thus further intensifying the inhibition of methanogens.

In our experiment, CH₄ suppression following SO₄ ^2– reduction inhibition in the Nat100 treatments persisted throughout the one-month observation period; however, in natural systems, this effect is likely to last longer. Based on the identified mechanisms, the duration of suppression largely depends on how long H₂S and salts remain in the peat. The removal of salts from peat is a slow process, occurring mainly through leaching and plant uptake. Moreover, although the cessation of SO₄ ^2– reduction limits new H₂S production, minor sources from organic sulfur decomposition may sustain low H₂S levels, further prolonging unfavorable conditions for methanogens. Therefore, CH₄ suppression is expected to last considerably beyond the experimental time scale.

While studies have repeatedly reported that brackish water rewetting reduces CH₄ emissions, ,,− research on CH₄ recovery following SO₄ ^2– depletion remains limited. A study by Capooci et al. (2019) in a tidal saltmarsh found a significant increase in CH₄ emissions during simulated storm surges, which caused a sudden decrease in salinity and SO₄ ^2– concentrations. However, this response may be attributed to regular low-magnitude saltwater intrusions, which could select for methanogen communities adapted to salinity and SRB influence. These tolerant methanogens may rapidly respond once conditions become more favorable (e.g., lower salinity and SO₄ ^2– levels). In contrast, studies on freshwater ecosystems generally yield findings consistent with ours. For instance, Berrier et al. (2022) observed that CH₄ production in freshwater wetland soils was suppressed in both Na₂SO₄- and NaCl-amended treatments compared to untreated controls, with only partial recovery in Na₂SO₄ treatments 28 days after SO₄ ^2– reduction was inhibited. Similarly, Gauci et al. (2005) reported incomplete CH₄ recovery in a pristine peatland two years after the cessation of SO₄ ^2– exposure, with recovery decreasing as SO₄ ^2– treatment intensity increased. Dowrick et al. (2006) also found that increased SO₄ ^2– availability during a drought suppressed CH₄ emissions in peat soils for 4 weeks following freshwater rewetting, with subsequent inhibition of SO₄ ^2– reduction resulting in only partial CH₄ recovery. Although these studies do not specifically examine brackish water-rewetted peatlands, they support our findings that factors beyond SO₄ ^2– reduction, such as NaCl-induced salt stress and H₂S toxicity, can contribute to the persistent suppression of CH₄ emissions even after SO₄ ^2– depletion.

3.4. Limitations and Perspectives

Our study revealed that natural brackish water rewetting can enable resilient and land-use-dependent suppression of CH₄ emissions in coastal peatlands, highlighting its potential as a robust and targeted strategy for climate change mitigation in these carbon-rich ecosystems. Despite yielding valuable findings, certain limitations remain and warrant consideration in future research.

Our experiment was limited by a relatively short incubation period and restricted sampling coverage, which focused solely on the temperate climate zone. Given the potential disparity between short- and long-term CH₄ dynamics, as well as regional variability in land-use legacies, future studies would benefit from longer-term experiments and broader sampling to better understand the temporal and spatial variations in CH₄ responses to brackish water rewetting. Second, our experiment did not account for tidal effects that may occur following the reconnection of diked peatlands to natural coastal dynamics. Tidal effects can cause periodic fluctuations in redox conditions, salinity, SO₄ ^2– concentration, and gas transport pathways, , potentially altering CH₄ responses and thus warranting deeper investigation in the future.

In addition, vegetation was excluded from the experiment, although its development after rewetting has been shown to strongly influence CH₄ dynamics in peatlands. Following rewetting, the rapid die-back of former agricultural vegetation may induce a short CH₄ pulse, as decomposing plant residues release substantial amounts of labile carbon while brackish water has not fully infiltrated the peat pores. As these substrates become depleted and brackish water infiltrates more uniformly, CH₄ production is expected to decline to very low levels. With the subsequent establishment of salt-tolerant emergent vascular plants such as Bolboschoenus maritimus, CH₄ fluxes may rise slightly due to renewed inputs of root exudates and enhanced plant-mediated gas transport, though root-zone oxygenation can foster limited oxidation. Overall, CH₄ emissions under brackish water rewetting are expected to remain low compared to freshwater systems, with magnitudes depending on the emerging vegetation composition. These dynamics suggest possible temporal variability in CH₄ emissions in the presence of vegetation, emphasizing the need for long-term observations to assess how vegetation succession under brackish conditions affects the effectiveness and persistence of CH₄ suppression.

Moreover, it is important to recognize that our laboratory incubations cannot fully capture the complexity of field conditions. Beyond tidal effects and vegetation dynamics, factors such as temperature fluctuations, atmospheric turbulence, and lateral hydrological flows can also interact to shape CH₄ emissions in situ. Our findings should therefore be viewed as mechanistic insights rather than predictions of field-scale outcomes, particularly with respect to the exact magnitudes of CH₄ responses and certain land-use-specific patterns. This underscores the need for future field studies to evaluate how well these findings translate to real coastal peatland settings.

3.5. Key Findings

We found that rewetting coastal peatlands with SO₄ ^2–-rich natural brackish water significantly suppressed CH₄ emissions (80 nmol m^–2 d^–1, 90%) compared to freshwater, with suppression more effective in grass-cut, arable, and unmanaged sites, but relatively limited in grass-graze sites. These differences were closely linked to land-use-induced variations in soil properties, with higher pH and lower relative abundance of methanogens facilitating stronger suppression. Mechanistically, natural brackish water primarily suppressed CH₄ through physiological inhibition of methanogens, likely driven by salt stress and H₂S toxicity. The suppressed methanogenic activity could not quickly recover following SO₄ ^2– reduction inhibition, indicating that CH₄ suppression can persist during short-term natural disturbances that cause a sudden decline in SO₄ ^2– concentrations. These findings underscore the importance of land-use legacies and water sources in regulating CH₄ dynamics, providing valuable insights for developing targeted rewetting strategies in coastal peatlands.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c16715.

• CO₂ fluxes and cumulative emissions during incubation; cumulative CH₄ emissions during incubation; soil characterization before and after incubation; alpha diversity indices of archaeal and bacterial communities before incubation; relative abundance of key CH₄-related microbial families before incubation; linear regressions showing the influence of key initial soil properties on mean CH₄ fluxes in sulfate-free and sulfate-containing treatments; detailed relative abundance of key CH₄-related microbial families; nonmetric multidimensional scaling of bacterial communities under different treatments after incubation at II sites; detailed description of studied sites; sulfate concentration in the rewetting water before and after incubation; detailed methods for 16S rRNA gene sequencing (PDF)

H.G.: Conceptualization, Methodology, Formal analysis, Investigation, Data Curation, WritingOriginal Draft, WritingReview & Editing, Visualization; S.C.: WritingReview & Editing; L.P.: WritingReview & Editing, Supervision; S.W.: Conceptualization, Methodology, WritingReview & Editing, Supervision, Project administration, Funding acquisition.

The authors declare no competing financial interest.