Evaluation of the environmental policy integrated climate model on predicting crop yield in the canadian prairies: A case study
Citation
Lychuk, T.E., Moulin, A.P., Johnson, E.N., Olfert, O.O., Brandt, S.A., Izaurralde, R.C. (2017). Evaluation of the environmental policy integrated climate model on predicting crop yield in the canadian prairies: A case study. Canadian Journal of Soil Science, [online] 97(4), 692-702. http://dx.doi.org/10.1139/cjss-2017-0044
Plain language summary
Modeling is a useful tool that allows researchers and scientists make important short- and long-term management decisions concerning crop production and environmental quality. The Environmental Policy Integrated Climate model was updated with relevant weather, tillage, and crop management operations from the long-term Alternative Cropping Systems (ACS) field experiment, which was conducted in 1994 - 2013 at Scott, Saskatchewan, Canada. The goal of the field experiment was to assess the interaction of cropping diversity and agricultural inputs on crop production in the region. Agricultural inputs were organic, minimum, and conventional tillage. Cropping diversities were wheat-fallow, diversified annual grains, and diversified annual perennials. The objective of this modeling study was to evaluate model’s accuracy in simulating annual and long-term yield of wheat, barley, and canola from the ACS study. Analysis of model output indicated that the model captured well long-term yield trends, but was less accurate in predicting annual yield dynamics. This was due to influence of soil properties, field topography, extreme weather events, and the model’s overestimation of available nitrogen under low-nitrogen input systems. We concluded that overall, model captured well effects of tillage and cropping diversity on simulated yield, thereby it may be used to aid with future cropping decisions and agronomic planning.
Abstract
The Environmental Policy Integrated Climate (EPIC) model was updated with relevant weather, tillage, and crop management operations from the 1994 to 2013 Alternative Cropping Systems study to assess simulations of annual and long-term yield of wheat, barley, and canola. Linear regression and coefficients of determination (R2), root mean square error of prediction (RMSE), the d index, and paired sample t-test were used to assess the relationship between simulated and experimental values. Simulations indicated that the model captured long-term yield trends but was less accurate at predicting annual variations. These variations were due to variability of soil properties at the research field, terrain attributes, extreme weather events, and the model’s overestimation of available nitrogen (N) under low-N input systems. The R2, RMSE, and the d index values on long-term yield were R2 = 0.74, RMSE = 205 kg ha-1, and d = 0.75 for wheat; R2 = 0.90, RMSE = 226 kg ha-1, and d = 0.73 for barley; R2 = 0.98, RMSE = 238 kg ha-1, and d = 0.76 for canola, indicating good model performance. The EPIC model effectively simulated crop yields affected by agricultural inputs and cropping diversity, and may be used to assess future cropping decisions and agronomic management.