Carbon and nutrient sequestration are among the key ecological co-benefits of wetlands. Methane (CH₄) emissions reduce the overall benefit of greenhouse-gas regulation by wetlands. To maximize the overall climate-regulation co-benefits, a wetland design should strive for reducing methane emissions, while maximizing carbon (CO₂) uptake. CH₄ emissions and CO₂ sequestration rates have very high spatial variability and vary strongly between ecological patches and hydrological conditions. This high variability makes the carbon fluxes of wetlands hard to predict.

We used an eddy covariance (EC) system (Ameriflux site ID: US-OWC) to observe CH₄ and CO₂ fluxes over a large footprint at Old Woman Creek (OWC NERR), a Lake Erie coastal, estuarine marsh. Our site-level EC observations show that inter-annual changes of wetland water depth, which are driven by increasing Lake Erie water elevation, combine with short-term riverine hydrological events are affecting CO₂ and CH₄ fluxes at the whole-wetland scale. At the patch scale, chamber measurements combined with porewater samplers show that both hydrology and vegetation type control CO₂ and CH₄ fluxes. We developed a classification approach for the Hybrid Landsat-Sentinel NDVI remote sensing images to determine the changes of vegetation patch type distribution and location in the wetland over the last decade. We found that vegetation distributions changed as Lake Erie water elevation increased.

We dated soil cores taken at different hydrological positions and depths to determine the long-term sequestration rates. Accumulation rates at different locations vary with patch-level residence time. The temporal dynamics of N, P, and C accumulation at these cores correlate with the nitrate load in the river, with additional effects of flow rate and nitrate concentration for C and N, or ammonium load for P. Our observations allow the direct calculation of parameters that are characteristic of patch types, such as aerenchyma conductivity to CH₄. We used these observations in a new, patch-level version of E3SM-ELM for predicting CH₄ and CO₂ fluxes. This model can be used to optimize the design of wetlands with regards to future climate regulation services.