Seasonal water level strongly affects CH4 emissions in a natural estuarine wetland: Current and future predictions using a mechanistic model

Wetlands are the largest natural source of methane (CH₄) worldwide yet have large uncertainty in models. This uncertainty comes from a number of sources such as poor global representation of inundated area, poor characterization of spatial heterogeneity, and poor representation of existing knowledge in models. Recent advances in modeling attempt to address the representation of underlying processes and the influence of environmental conditions. Observational studies have shown that soil temperature and water level are often the most influential factors on wetland emissions. Temperature directly affects CH₄ by affecting microbial process rates and is relatively simple to represent in models. Water level, on the other hand, affects CH₄ indirectly by restricting oxygen (O₂) transport to the soil. A reduction of soil O₂ increases the thermodynamic favorability of methanogenesis and suppresses methanotrophy. O₂ availability also affects plant growth and deposition of root exudates which fuel the soil microbial community. Empirical models have provided valuable information about this linkage, but their ability to extrapolate is limited, especially under non-stationary climate conditions. Mechanistic models can be used to bridge this gap and improve predictions of future wetland CH₄ emissions.

We used a mechanistic model (ecosys) to simulate the GHG budget of the Old Woman Creek in Huron Ohio. Ecosys includes fully coupled hydrology, vegetation, thermodynamic, and biogeochemistry modules through which it explicitly represents O₂ and CH₄ transport and transformations. We created two-column model simulations for the dominant patch types at the site and mixed them to match the tower footprint and successfully tested the model against CO₂ and CH₄ eddy covariance, peepers, and chamber observations. The site's seasonally variable water level was an important factor affecting seasonal and annual CH₄ budgets. We predict that changes in precipitation and temperature over the coming century will lead to significant changes in CH₄ and CO₂ emissions, respectively and ran simulations with future climate scenario data for Lake Erie wetlands. Finally, we show that CH₄ and CO₂ emissions predicted from traditional empirical approaches are unable to capture these complex interdependencies.