Improving the Simulation of Soil Water and Nitrogen Dynamic in the Epic Model.

Soil-water and nutrients dynamics are key elements in simulating agro-ecosystems because these are fundamental to other biophysical processes such as runoff, infiltration, percolation, evapotranspiration, nutrient cycling, and crop growth. Accurate simulation of soil water content is important for estimating soil nitrogen balances in cropping systems simulation models. In particular, denitrification occurs in anoxic condition when soil is saturated with water; thus, inaccurate estimation of soil water content often leads to unreliable simulation of denitrification rates. Considering these critical aspects, a new subroutine was developed to improve the simulation of soil water percolation in root zone soils in the Environmental Policy Integrated Climate (EPIC). The original EPIC percolation method uses a linear function to estimate the daily amount of soil water percolation using saturated hydraulic conductivity (SHC), which often results in overestimating downward drainage of soil water to subsoil layers. In the new method, percolation is estimated iteratively using a non-linear function based on effective hydraulic conductivity (EHC) where EHC is continuously updated while iterating the daily percolation processes to reflect the nonlinearity between soil water content and the mobility of soil water. Compared to the original method, this EHC-based approach allows EPIC to estimate the end-of-day soil water contents higher than field capacity, especially during wet weather periods with prolonged rainfall. The improvement in soil hydraulics simulation influences the long-term calculation of denitrification in soils. In this paper, we use the EPIC model (v1102) to evaluate simulated soil water content, soil nitrogen content, and soil nitrous oxide emission rate measured at multiple experimental sites (France, Australia, and Brazil). These sites are characterized by different climate conditions, soil types (silty-loam, clay, and sandy-loam texture respectively) and cropping systems (maize-wheat rotation in France and Australia, maize-oat-maize-wheat in Brazil). Significance of the improved soil water simulation to calculating denitrification in shallow soils in croplands and rangelands at these sites will be presented.