Field evaluation of models of denitrification linked to nitrate leaching for aggregated soil

Abstract
Summary: Dynamic simulation models of the nitrogen cycle are an important aid for predicting the effect of land management on nitrate concentration in ground and surface waters. These models often neglect, or treat in a highly empirical way, the process of denitrification. The solute leaching model SLIM, suitable for use in aggregated soil, was linked to two alternative denitrification submodels ('simple structure'and ‘aggregate assembly’ treatments). The linked models can be used in dynamic simulation of the fate of nitrate in soil. The models were calibrated using water release and soil respiration data from two field sites (a clay loam and a sandy loam). The fate of Br−1and 99 atom percentage15NO3tracers on drained plots was measured at these sites. Both models predicted less denitrification from plots receiving small inputs of fresh organic residue than was actually lost; both did better where residue inputs were large. The ‘aggregate assembly’ model predicted short‐term losses immediately after incorporation of residues more successfully than did the ‘simple structured’ model. The simulated dynamics of denitrification during the autumn and winter in the two models were also very different. Overall N balances were better predicted in the clay loam than the sandy loam. Possible reasons for discrepancies between model output and real data include immobilization (especially when large inputs of residues were incorporated), denitrification in the subsoil, and an inadequate description of the way in which degradation of organic matter depends on temperature.