Evaluation of EPIC's snowmelt and water erosion submodels using data from the Peace River region of Alberta

Water erosion due to snowmelt is a major form of erosion in boreal regions of the Canadian Prairie. Evaluation of erosion models is an essential step before recommending their use in local or regional assessments of erosion rates and control methods. Using inputs fi om a runoff study conducted at La Glace, Alberta(55 degrees 25'N, 119 degrees 10'W) from 1984 to 1986, we evaluated the Erosion-Productivity Impact Calculator (EPIC) for its ability to simulate runoff and sediment yield from snowmelt events. The model was initialized with soil profile data acquired at the study site and complemented with data from standard soil databases (Albright series; loam, Dark Gray Chernozem). Daily weather data were acquired from the nearest climatological station (annual precipitation = 475 mm). Management data were as reported and included combinations of conventional and reduced tillage, annual and perennial, and fallow cropping. Mean runoff volume measured in 1985 was 57 mm while in 1986 it was 76 mm. EPIC over-predicted runoff volume by 25% in 1985 but under-predicted it by 7% in 1986. The period in which snowmelt occurred (mid-March - beginning of April) was predicted correctly. Under the conditions of this study, with many cropping inputs obtained from different sources, the model was unable to reproduce the reported management effects on runoff and sediment yield. EPIC simulated springmelt soil temperature trends at 9-cm depth, although the predicted temperatures in 1985 were generally underestimated. Our results suggest that the EPIC model calculates adequate values of runoff volumes and sediment yields during snowmelt.