Approaches and concepts of modelling denitrification: increased process understanding using observational data can reduce uncertainties

Denitrification is a key but poorly quantified component of the Ncycle. Because it is difficult to measure the gaseous (NO$_{x}$, N$_{2}$O, N$_{2}$)and soluble (NO$_{3}$) components of denitrification with sufficientintensity, models of varying scope and complexity have beendeveloped and applied to estimate how vegetation cover, landmanagement and environmental factors such as soil type andweather interact to control these variables. In this paper we assessthe strengths and limitations of different modeling approaches,highlight major uncertainties, and suggest how differentobservational methods and process-based understanding can becombined to better quantify N cycling. Representation of howbiogeochemical (e.g. org. C., pH) and physical (e.g. soil structure)factors influence denitrification rates and product ratios combinedwith ensemble approaches may increase accuracy withoutrequiring additional site level model inputs.