Modeling methane fluxes in wetlands with gas‐transporting plants: 3. Plot scale

Abstract

A process model based on kinetic principles was developed for methane fluxes from wetlands with gas‐transporting plants and a fluctuating water table. Water dynamics are modeled with the 1‐D Richards equation. For temperature a standard diffusion equation is used. The depth‐dependent dynamics of methane, oxygen, molecular nitrogen, carbon dioxide, soil carbon, electron acceptors in oxidized and in reduced form are affected by transport processes and kinetic processes. Modeled transport processes are convection and diffusion in the soil matrix, ebullition, and plant‐mediated gas transport. Modeled kinetic processes are carbon mineralization, aerobic respiration, methane production, methane oxidation, electron acceptor reduction, and electron acceptor reoxidation. Concentration gradients around gas‐transporting roots in water‐saturated soil are accounted for by the models from the two previous papers, ensuring an explicit connection between process knowledge at the kinetic level (millimeter scale) and methane fluxes at the plot scale. We applied the model to a fen, and without any fitting, simulated methane fluxes are within 1 order of magnitude of measured methane fluxes. The seasonal variations however, are much weaker in the simulations compared to the measurements. Simulated methane fluxes are sensitive to several uncertain parameters such as the distribution over depth of carbon mineralization, the total pool size of reduced and oxidized electron acceptors, and the root‐shoot ratio. Because of the process‐based character of the model it is probable that these sensitivities are present in reality as well, which explains why the measured variability is usually very high. Interestingly, heterogeneities within a rooted soil layer seem to be less important than heterogeneities between different soil layers. This is due to the strong influence of the interaction between water table and profile scale processes on the oxygen input to the system and hence on net methane production. Other existing process models are discussed and compared with the presented model.