Modelling nitrogen transport in sugar cane from soil to runoff from banded surface and buried fertiliser using hydrus2d and a post-processing algorithm

Abstract

Simulated rainfall was applied to sugar cane at the Macknade Research Station, Herbert region North Queensland, to determine the loss of nitrogen to runoff for fertiliser placed in a 100 mm wide strip on raised beds (spaced 1.8 m apart and a bed width of 1.2 m) applied either on the surface or buried at a depth of 50-150 mm. The sites were covered with shelters between rainfall simulations to avoid rainfall ingress, but evaporation could still take place. Modelling of these experiments was untaken to estimate nitrogen losses from recently fertilised soils in wet tropical catchments. The model parameters can be used to model other possible runoff scenarios. The modelling required a two-dimensional model with multiple solute species transport and chain reaction processes. The HYDRUS2D model (Simunek et al., 2012) was chosen; however, it cannot model the transfer of solutes to the runoff water. An analytical model, which can transfer of solutes to runoff was developed by Wallach and van Genuchten (1990): [[ (0,) () r s J t k c t c θ =--(1) where J(0,t) is the flux density of solute transferred from the soil surface to the runoff water [M L-3 T-1[ at time t [T], θs is the saturated volumetric water content at the soil surface [L 3 L-3[, k is the mass transfer coefficient [L T-1[, c(t) is the concentration at the soil surface with time [M L-3[ and cr is the concentration in the runoff water [M L-3[, which we set to zero as Wallach and van Genuchten (1990) did. A post-processing method to compute the solute transport of nitrate, nitrite and ammonium to the runoff using eqn (1) was developed. This required surface concentrations of the solutes with time in a cross-section across the bed (c(x,t) where x is the cross-section distance) and integrating both with space to give the mass transfer at time ti, M(s,ti), using: ().