NEMIS, a predictive model of denitrification on the field scale

We have developed a predictive model of denitrification in soils, NEMIS, based on data from a cultivated field composed of denitrification rates measured on undisturbed soil cores and the corresponding soil water-filled pore space, nitrate content and respiration. This model is a multiplicative combination of a potential denitrification rate measured under standard conditions defined in this paper, and dimensionless functions of nitrate content, water-filled pore space and temperature. The model was applied to two further databases obtained independently of its definition. Significant correlation was observed between the simulated and measured denitrification rates. The sensitivity of NEMIS to systematic variations of the input variables and of the coefficients included in the dimensionless functions was tested using these databases. NEMIS seems to be a very useful tool particularly adapted for interpolating a small number of denitrification measurements from a simple database composed of soil potential denitrification rate and the evolution of soil water-filled pore space, nitrate content and temperature. Its main limitation is its high sensitivity to the water-filled pore space from which denitrification starts to occur and which is included in the water function.