Abstract

An environmental modeling system is being constructed around the EPIC model that will be interfaced with an economic component (RS-CRAM) within an integrated modeling system to analyze agricultural policies such as GRIP for western Canada. A description of the major EPIC subcomponents is provided, including the most important data inputs. An environmental database also has been constructed for EPIC. This database consists of: (1) soil layer and landform data in separate databases for each Prairie Province, (2) ARA 31-year daily historical weather data (precipitation and maximum and minimum temperature) in EPIC format, (3) EPIC weather generator tables for each ARA, and (4) EPIC wind arrays for selected climate stations in Alberta, Manitoba, and Saskatchewan.

Calibration and testing of the EPIC model for selected sites in the Prairies and U.S. Northern Great Plaines is reviewed. Comparisons of predicted yields with measured data show that the model is producing reasonable yield estimates for different crops that are grown in western Canada. Yield estimates by EPIC show that it is definitely better at predicting long-term average yields as compared with individual crop year yields. However, the model does accurately reflect key trends such as crop response to fertilizer and productivity response to soil erosion. The results affirm other findings that EPIC output be used to provide relative comparisons of environmental impacts as opposed to absolute values.