Modelling belowground processes to explain field-scale emissions of nitrous oxide

Globally, soils are the most important source of nitrous oxide (N2O). In several recent studies, N2O fluxes from soils have been monitored. Regression relationships between N2O fluxes and variables like average soil moisture content, nitrogen content and temperature did in general not explain more than 50% of the flux variance. The quantitative insight into the relation between measured fluxes on the one hand, and the underlying basic processes and their determining factors on the other, is still limited. This study aims to contribute to this insight by modelling underground processes to explain N2O profiles in the soil. On sandy grassland soils in the Wageningen Rhizolab, we monitored aboveground N2O fluxes as well as underground profiles of several variables. A process-based model was adapted to simulate the relevant processes. In the model we assumed that the sandy soil profile consisted of homogeneous soil layers. The modelled processes included biological and physical processes. Nitrous oxide is produced during denitrification and is transported via diffusion. The simulated O2 and CO2 profiles were satisfactory, indicating that the (bulk) respiration rates used in the model were realistic. The N2O profiles, however, were less well simulated. With respect to N2O production, the assumption of homogeneity within the soil layers was probably not correct for the sandy soil. We discuss modelling approaches to explain observed N2O profiles by describing partial anaerobicity in soil layers. Of these, the so-called ‘randomly distributed-pore model’ [Eur. J. Soil Sci. 46 (1995) 507] and an adapted version for regularly distributed pores appear good candidates for incorporation in field-scale models for a layered sandy soil. Such models would describe processes like bulk respiration and the transport and plant uptake of water and nutrients. The randomly distributed-pore model is a relatively simple model for a steady state situation. For sandy soil layers, the model predicts that a substantial part of the volume can be anoxic under a range of conditions including conditions that often occurred in our experiment: a ratio (volumetric moisture content)/(saturated volumetric moisture content)≥0.7 and oxygen percentages in the air-filled pores between 5 and 21%. The regularly distributed-pore model also predicts the occurrence of anoxic fractions under common experimental conditions, but these fractions are significantly smaller than those according to the randomly distributed-pore model.