Evaluation of the Environmental Policy Integrated Climate (EPIC) Model on Predicting Wheat, Barley, and Canola Yield on a 19-Year Field Study in the Semi-Arid Canadian Prairie
Lychuk, T.E., Moulin, A.P, Johnson, E.N., Olfert, O.O., Brandt, S.A., Izaurralde, R.C. 2015. Evaluation of the environmental policy integrated climate (EPIC) model on predicting wheat, barley, and canola yield on a 19-year field study in the semi-arid Canadian Prairie. Soil Science Society of America. Minneapolis, MN, Nov 15-18, 2015.
Résumé en langage clair
Producers are modifying cropping systems to improve environmental and economic sustainability in the Canadian Prairies. Analysis of the impact of management of crop inputs and diversity on yield and soil properties is complex, but can be done with computer models. This study assessed the Environmental Policy Integrated Climate Model with respect to prediction of crop yield. The model predicted long-term changes but not in the short term.
Farmers in Canadian Prairies have the option to move from conventional high-input production for one or two annual grain crops to diversified and extended cropping systems with reduced inputs and organic management. These cropping systems diversify crop production and reduce reliance on fertilizers and pesticides and improve economic and environmental sustainability. The Alternative Cropping System (ACS) field experiment was conducted from 1994 to 2013 in Scott, SK to assess the potential impact of agriculture on sustainable production and soil and environmental quality in the region. Modeling allows researchers and scientists to evaluate short- and long-term management decisions with respect to crop production and environmental quality. The Environmental Policy Integrated Climate (EPIC) model was updated with relevant weather, tillage, and crop management operations from the ACS study at Scott. The model was validated with annual and long-term yield data for wheat (Triticum aestivum L.), barley (Hordeum vulgare L.), and canola (Brassica napus L.). Regression results based on EPIC simulations indicated that model captured long-term yield trends for all three crops, but was weak at predicting annual yield variations. This was primarily due to (1) soil variability and micro-topography at the research site, (2) significant variation in precipitation rates and temperatures relative to region normals, (3) excessive damage due to extreme weather events such as hail and (4) model’s overestimation of mineralized N under low-nitrogen input systems. Correlations (R2) varied between > 0.60 for long-term yield predictions, and < 0.50 on an annual basis. The EPIC model should be adjusted with respect to N cycling and parameters that control soil hydrology and water use by plants. Overall, EPIC replicated effects of agricultural input systems and rotations on long-term crop yield. The model may be used by researchers and scientists as a long-term decision tool for agricultural productivity and sustainability in the Canadian Prairie.