TESTING EFFECTIVENESS OF 4 SIMULATION-MODELS FOR ESTIMATING NITRATES AND WATER IN 2 SOILS

Information is required on the usefulness and limitations of simulation models for estimating nitrate-nitrogen (NO3-N) and water status of soil in different agroecological zones of western Canada, Therefore, four simulation models (CERES, EPIC, NLEAP and NTRM) were used to predict distribution of NO3-N and water in the soil profile from long-term spring wheat (Triticum aestivum L.) rotations for the 1990 growing season at Melfort and Scott, Saskatchewan. The models estimated NO3-N and water levels in the crop rooting zone (0-1.5 m) of the soil profiles generally well. However, EPIC underestimated NO3-N levels in the top 0.3 m for fallow (F) rotations at Melfort, whereas NLEAP overestimated amounts in the same depth increment of Melfort soil for the majority of rotations. In addition, CERES and NTRM markedly overestimated NO3-N levels in the Melfort soil solum (0-0.6 m) for the unfertilized continuous wheat (W) rotation. Soil water content was overestimated in the top 0.3-m layer of Melfort soil for the two W rotations by CERES and EPIC and for all rotations by NTRM. As well, EPIC markedly overestimated water content in the crop rooting zone for both rotations at Scott. There were differences among models in accurately simulating NO3-N and water status of the soil profile, but these differences were not great. Overall, these models should be effective in estimating soil NO3-N and water content for wheat cropping systems, thereby aiding agronomic decisions and environmental impact assessments.