Modelling nitrous oxide emissions from mown-grass and grain-cropping systems: Testing and sensitivity analysis of DailyDayCent using high frequency measurements

The DailyDayCent biogeochemical model was used to simulate nitrous oxide (N2O) emissions from two contrasting agro-ecosystems viz. a mown-grassland and a grain-cropping system in France. Model performance was tested using high frequency measurements over three years; additionally a local sensitivity analysis was performed. Annual N2O emissions of 1.97 and 1.24 kg N ha− 1 year− 1 were simulated from mown-grassland and grain-cropland, respectively. Measured and simulated water filled pore space (r = 0.86, ME = − 2.5%) and soil temperature (r = 0.96, ME = − 0.63 °C) at 10 cm soil depth matched well in mown-grassland. The model predicted cumulative hay and crop production effectively. The model simulated soil mineral nitrogen (N) concentrations, particularly ammonium (NH4+), reasonably, but the model significantly underestimated soil nitrate (NO3−) concentration under both systems. In general, the model effectively simulated the dynamics and the magnitude of daily N2O flux over the whole experimental period in grain-cropland (r = 0.16, ME = − 0.81 g N ha− 1 day− 1), with reasonable agreement between measured and modelled N2O fluxes for the mown-grassland (r = 0.63, ME = − 0.65 g N ha− 1 day− 1). Our results indicate that DailyDayCent has potential for use as a tool for predicting overall N2O emissions in the study region. However, in-depth analysis shows some systematic discrepancies between measured and simulated N2O fluxes on a daily basis. The current exercise suggests that the DailyDayCent may need improvement, particularly the sub-module responsible for N transformations, for better simulating soil mineral N, especially soil NO3− concentration, and N2O flux on a daily basis. The sensitivity analysis shows that many factors such as climate change, N-fertilizer use, input uncertainty and parameter value could influence the simulation of N2O emissions. Sensitivity estimation also helped to identify critical parameters, which need careful estimation or site-specific calibration for successful modelling of N2O emissions in the study region.