An alternative modelling approach to predict emissions of N2O and NO from forest soils

Abstract

Emissions of N2O from forest soils in Europe are an important source of global greenhouse gas emissions. However, influencing the emission rates by forest management is difficult because the relations and feedbacks between forest and soils are complex. Process-based models covering both vegetation and soil biogeochemical processes are frequently used to analyse emission patterns. Particularly, the simulation of soil C and N turnover processes driving N2O production, consumption and emission from forest soils requires highly specific input data which renders their regional application difficult since at this scale, soil conditions are often not well understood. Therefore, a soil C and N model (DecoNit) has been developed which describes biogeochemical processes with a simplified structure compared to existing carbon/nitrogen models that nevertheless follows the basic physical and chemical laws involved and which allows to simulate N trace gas emissions. The DecoNit model was previously calibrated using an extensive dataset on decomposition rates of incubated plant materials, microbial dynamics and nitrification. The DecoNit model has now been embedded in a modular simulation environment (MoBiLE) where it is combined with soil water balance and forest process sub-modules. Here, we present the evaluation of MoBiLE-DecoNit with emission data of N2O and NO from forest soils of 15 European sites and compare simulation results with a previous study in which a more complex model (PnET-N-DNDC) was used. Evaluation criteria were as follows: (1) precision of modelled annual average emission rates; (2) coherence of modelled and measured annual average and daily emissions; (3) a dynamic representation of emission rates that correspond with the observed variance of fluxes. The results show that MoBiLE-DecoNit captures average annual emission rates more precisely than the more complex model PnET-N-DNDC. Also the structural underestimation of N trace gas fluxes from forest soils was resolved. Moreover, we present evidence that the new modelling approach is also somewhat more adequate for describing inter-daily emission dynamics. The combined MoBiLE-DecoNit is therefore thought to be a promising approach to simulate forest development and greenhouse gas balances on site and regional scales.