Methane and nitrous oxide fluxes from reference, restored, and disturbed estuarine wetlands in Pacific Northwest, USA

Abstract There is substantial interest in restoring tidal wetlands because of their high rates of long‐term soil carbon sequestration and other valued ecosystem services. However, these wetlands are sometimes net sources of greenhouse gases (GHG) that may offset their climate cooling potential. GHG fluxes vary widely within and across tidal wetlands, so it is essential to better understand how key environmental drivers, and importantly, land management, affect GHG dynamics. We measured methane (CH 4 ) and nitrous oxide (N 2 O) fluxes at 26 reference and restored tidal wetland sites and eight nontidal pastures (mostly diked former tidal wetlands) in five estuaries in the Pacific Northwest (PNW), USA. We measured fluxes 7–8 times over one year to assess the effects of environmental drivers, wetland type, and land management on CH 4 and N 2 O fluxes. Linear relationships between CH 4 fluxes and environmental drivers were poor, but a machine‐learning approach with boosted regression trees provided strong predictability for fluxes based upon wetland surface elevation, water‐table level, and salinity. Less important variables were groundwater pH, wetland type, and temperature. Under oligohaline conditions, CH 4 fluxes were variable and sometimes very high, but fluxes at salinities above 2 ppt were relatively low on an annual basis. Fluxes of CH 4 were higher in restored tidal marshes and wet pastures than in reference tidal marshes, tidal swamps, and dry pastures. The N 2 O model had lower predictive power than the CH 4 model, with wetland type as the most important factor, although N 2 O fluxes across all wetland types were low (median of zero). Our results indicate that estuarine hydrologic gradients are a key driver of CH 4 fluxes and that wetland land use impacts on CH 4 fluxes are largely mediated by their varying environmental conditions. In the PNW, estuarine wetlands that have low salinity, lower elevation, and have high water tables are more likely to have increased CH 4 emissions that may offset their carbon sequestration benefits until they gain enough elevation through accretion. This study also provides a transferrable modeling approach to predict the consequences of coastal wetland management on GHG fluxes using monitoring data from a limited set of key environmental drivers.