Controls on spatial variation in porewater methane concentrations across United States tidal wetlands

Tidal wetlands can be a substantial sink of greenhouse gases, which can be offset by variable methane (CH₄) emissions under certain environmental conditions and anthropogenic interventions. Land managers and policymakers need maps of tidal wetland CH₄ properties to make restoration decisions and inventory greenhouse gases. However, there is a mismatch in spatial scale between point-based sampling of porewater CH₄ concentration and its predictors, and the coarser resolution mapping products used to upscale these data. We sampled porewater CH₄ concentrations, salinity, sulfate (SO₄²⁻), ammonium (NH₄⁺), and total Fe using a spatially stratified sampling at 27 tidal wetlands in the United States. We measured porewater CH₄ concentrations across four orders of magnitude (0.05 to 852.9 μM). The relative contribution of spatial scale to variance in CH₄ was highest between- and within-sites. Porewater CH₄ concentration was best explained by SO₄²⁻ concentration with segmented linear regression (p < 0.01, R² = 0.54) indicating lesser sensitivity of CH₄ to SO₄²⁻ below 0.62 mM SO₄²⁻. Salinity was a significant proxy for CH₄ concentration, because it was highly correlated with SO₄²⁻ (p < 0.01, R² = 0.909). However, salinity was less predictive of CH₄ with segmented linear regression (p < 0.01, R² = 0.319) relative to SO₄²⁻. Neither NH₄⁺, total Fe, nor relative tidal elevation correlated significantly with porewater CH₄; however, NH₄⁺ was positively and significantly correlated with SO₄²⁻ after detrending CH₄ for its relationship with SO₄²⁻ (p < 0.01, R² = 0.194). Future sampling should focus on within- and between-site environmental gradients to accurately map CH₄ variation. Mapping salinity at sub-watershed scales has some potential for mapping SO₄²⁻, and by proxy, constraining spatial variation in porewater CH₄ concentrations. Additional work is needed to explain site-level deviations from the salinity-sulfate relationship and elucidate other predictors of methanogenesis. This work demonstrates a unique approach to remote team science and the potential to strengthen collaborative research networks.