Simulating the fate of subsurface-banded urea

Abstract

Compared to surface broadcasting of nitrogen (N), subsurface-banding increases crop yield and N removal, reducing the N amount available for loss into the environment. Subsurface banding results in two-dimensional (2-D) N movement and high localized N concentrations that reduce N transformation rates. A physically-based, field-scale model was developed and tested for its ability to simulate the fate of subsurface-banded N. The 1-D moisture sub-model simulates moisture redistribution, evapotranspiration, and percolation using the Richards equation. Using the convective-dispersive equation, the 2-D N sub-model uniquely simulates the fate of urea, ammonium, and nitrate by accounting for urea particle dissolution rate and substrate concentration effects. Over 325 d, the model displayed robustness with respect to selection of grid sizes and exhibited mathematical accuracy. Analyses highlighted the importance of incorporating substrate concentration effects on N transformations; disregarding substrate concentration effects could result in overestimation of crop N removal and N movement in the soil. Analyses also indicated the importance of including molecular diffusion in simulating the fate of subsurface-banded N. While evapotranspiration, crop N removal, and residual applied inorganic-N estimates appeared realistic, measured and simulated moisture and applied-N amounts and distributions differed greatly. Errors in measurements of moisture and N and uncertainty in parameter estimation likely increased the difference in simulated versus measured values. The model was highly sensitive to soil pH and Freundlich distribution coefficient. Additional model evaluation with field data, though required, could not be performed due to lack of data describing distribution of soil N resulting from subsurface-banding.