The interface between RS-CRAM and the environmental component of the integrated modeling system is described for crops, crop sequences, and management systems representative of western Canada. An experimentally designed set of EPIC simulations were performed to generate erosion output that was used to construct wind and water erosion metamodels (response functions). The results of the EPIC simulations indicate that wind and water erosion would be the dominant erosion problem over most of Saskatchewan and Manitoba. For Alberta, water erosion was predicted to be the dominant problem, except for the southern portion of the province. Erosion impacts were sensitive to tillage and cropping patterns. EPIC-predicted yields did not vary much across tillage, a result consistent with measured observations. However, the EPIC yield estimates tend to be higher than previous average census yields used in CRAM. The model also appears to underpredict yields of crops grown on fallow in the more humid regions of the Prairies. The wind and water erosion metamodels estimated for the three provinces are very robust, with the majority possessing R2 values in the range of 0.80 to 0.97. The predictive power of the metamodels is confirmed in validation tests comparing metamodel output with the original simulation data. The efficiency of the metamodels in facilitating the integration of the complete policy modeling system is described. Finally, the process of aggregating environmental indicators, estimated with the metamodels, from the landscape polygon level to ARAs and ultimately to the CRAM region level is also described.