Simulation of water and nitrogen balances in a perennial forage system using the STICS model

Soil-crop models can be effective tools for calculating the water and nitrogen (N) balances of agricultural systems, because such models are able to simulate the complex interactions between crop, soil water, and soil N. However, model performance needs to be evaluated against field measurements. Our objectives were to evaluate the STICS model for simulating inorganic N and water fluxes in a perennial timothy (Phleum pratense L.) forage system in Eastern Canada and use the model for estimating water and N balances under three N treatments. Experimental data were collected for three years on a sandy loam soil with three N treatments (0 kg N ha−1, mineral fertilizer at 140 kg N ha−1, and raw liquid swine manure at 140 kg total N ha−1). The model had good performance for predicting harvested biomass and soil moisture (relative root mean square error below 30%). Total soil mineral N (nitrate + ammonium) was simulated reasonably well, but soil nitrate content was overestimated during the spring growth of timothy. The model simulated N emissions efficiently and reproduced the tendency toward higher emissions in the liquid swine manure treatment. Despite these discrepancies, general trends of soil mineral N and gaseous N emissions were well reproduced. The annual simulated plant-soil N balance varied with fertilizer types, N rates, and weather conditions. Gaseous N emissions were greater with liquid swine manure than with mineral N fertilizer because of higher ammonia volatilization. Nitrate leaching represented a very small component of the N balance. This study is a significant step toward the validation of the STICS model for simulating N and water cycles in perennial forage systems in cold, humid continental climates.