Effects of soil heterogeneity on the uncertainty in modelling the fate of urinary nitrogen deposited during winter forage grazing

A sensitivity analysis was performed to quantify how changes in the physical description of soil layers as model input parameters affect various outputs of the biophysical APSIM model regarding the fate of urinary nitrogen following winter grazing. The model was set up according to a field experiment on a poorly drained Flaxton soil in the Canterbury area of New Zealand. Soil profile descriptions were based on various measured datasets, and including +/- compacted soil surface layers due to treading during grazing of winter forage crops. The effect of the inclusion of more permeable sandy layers within the subsoil, as common within the study region, was also examined. By systematically replacing soil layers with different physical the effect of these on the N loss pathways was evaluated. The analysis identified that the bottom layer (1 to 1.5m depth) had the largest impact on the loss pathways of nitrogen, rather than the compaction of the top soil layers due to treading. Nitrate leaching was lower by 20%, and pasture N uptake, denitrification and nitrous oxide emissions were greater (by 5%, 58% and 43%) when the soil profile included a low permeability bottom layer. When replacing the bottom layer with much higher permeability, the description of the bottom layer remained the most important layer for nitrate leaching and plant uptake, but other layers became more important for denitrification and nitrous oxide emissions. The results highlight the need for accurate soil physical descriptions for adequate model parameterisation, and their use for assessing management practices.