Modelling 3D urine patch spread in grazed pasture soils to determine potential inhibitor effectiveness

Abstract

Urine patches represent hot-spots of nitrogen (N) loss in dairy-grazed soils. Targeted application of urease and nitrification inhibitors that slow down certain N transformations in the urine patches is a potential method to reduce N losses. However, for optimum effectiveness the inhibitors need to be in close physical contact with the urine in the soil under urine patches. In practice, there will always be some time delay between urine deposition and application of inhibitors. It is therefore important to understand how the urine is transported in the soil following deposition. In this study, we developed an empirical model of urine patch area from thermal images of urine patches applied on two different soil types, at two different initial moisture contents, and with three different applied urine volumes. Spatial measurements using Spikey®-R (a mobile device that measures soil surface layer electrical conductivity) were used to test the model. A linear regression model of the ratio (urine volume)/(patch area) against the soil air-filled pore space explained 45 % of the variation in the ratio and had a Nash-Sutcliffe efficiency of +0.74 in predicting the mean patch area. This regression model was then used to define the boundary conditions for HYDRUS2D/3D simulations of urine movement through the soil after application. These simulations reasonably predicted the amount of urine-N in the top 50 and 100 mm of the soil 4 h after application (model efficiencies +0.38 and +0.42, respectively), but the model efficiencies were only −0.18 and +0.14 after 24 h. The measurements also had a high degree of spatial variability.After 24 h 44–78 % of the urine-N measured in the profile was within 50 mm of the surface. This represents a limit on the proportion of urine-N that could be physically intercepted by a post-grazing inhibitor application.