Representing microbial activity in a soil decomposition model

Complex biogeochemical processes involving vegetation, microbial communities, symbiotic relationships and nutrient cycling determines the rates at which carbon is transferred between the atmosphere and the soils, and eventually the terrestrial carbon storage. Because of these complexities, there are still large uncertainties connected to the representation of the terrestrial carbon cycling in Earth System Models.

An emerging approach to deal with these problems is to explicitly represent microbial pools as well as physically and chemically protected soil organic matter in the model. Based on this, we have developed a soil decomposition process model designed to capture and quantify relationships between soil microorganisms and their environment, focusing on high latitude systems. Our aim with this approach is dual: 1) Testing hypotheses in the model before designing field experiments will help to set up experiments that benefits both understanding of the ecology and improving the model and, 2) Incorporating such a model into an ESM will make it possible to validate the model with observations, and to identify possible climate feedbacks related to the soil dynamics.

The model includes the decomposer activity of saprotrophic fungal and bacterial communities and the symbiotic relationship between mycorrhizal fungi and vegetation. We include separate carbon and nitrogen reservoirs for these microbes, as well as for plant litter and soil organic matter. The transfer of C and N between the reservoirs is based on rate equations using various parameterizations found in literature, and is also transported vertically following a diffusion equation.

The model is forced with litter input and climatic variables from the Community Land Model (CLM5). For calibration and validation we use subsets from a large dataset containing soil profile data for ~1000 forested sites in Norway (Strand et al. 2016). Since the sites are distributed over a large area, they cover climatic gradients in both temperature and precipitation.

Comparisons of C and N content between simulations from the new decomposition model, the standard CLM5 and the observations will be presented, as well as sensitivity tests of different parameter choices and impacts of changes in climate forcing.