Soil fluxes of carbon dioxide (CO2) and methane (CH4) are often linked through heterotrophic dependence on fixed carbon (C) sources for energy, and the same soil profile may alternate between being a net source or a net sink for CH4 depending on the concentration of oxygen (O2) at microbial microsites. Heterotrophic respiration dominantly controls the rate of O2 consumption at soil microsites, with the resulting O2 concentrations strongly affecting the net flux of ch4 as it plays a dual role of being an essential substrate of methanotrophy or an inhibitor of methanogenesis. Most biogeochemical models generally simulate heterotrophic respiration and CH4 emission separately, but we contend that emissions of multiple GHGs can be simultaneously simulated using the same biophysical framework for estimating the consumption of soil c and o2 and for diffusion of gases through the soil profile. Moreover, considerable spatial variability in measured soil GHG fluxes have been frequently reported in the literature and our modeling approach also investigated the underlying mechanisms responsible for this heterogeneity by evaluating the probability distribution function of microsites with variable soil C and O2 concentrations. The skewness in the density of soil microsites was constrained with the variance of measured soil GHG fluxes.
We have obtained sufficient data from the howland forest site, central maine to enable integration of both CO2 and CH4 flux data with a revised Dual Arrhenius and Michaelis–Menten (DAMM) model that is parameterized and tested within a data rich environment. Soil CO2 efflux served as a reasonable proxy for O2 demand within the soil, where simulation of O2 consumption within the soil is based on measured CO2 efflux. The total observed CO2 efflux and the accompanying O2 consumption was partitioned according to a simulated log-normal probability distribution function of soil C. Chamber-based measurements of CO2 efflux provided a good constraint for the sum of the simulated CO2 production/O2 consumption rates across the simulated distribution of microsites. The resulting probability distribution function of O2 concentrations simulated methanogenesis and methanotrophy at the microsite scale. The net CH4 flux summed across simulated distributions of microsites was similar to observed chamber-based CH4 fluxes. The modeled variation in GHG fluxes was consistent with spatial heterogeneity measured among chambers. Our modeling exercise demonstrate that constraining a model with multiple streams of data, including both CO2 and CH4 fluxes, enhances the chances of getting skillful simulations resulting from improved parameterization and good process representation.