Application of the Distributed Hydrology Soil Vegetation Model for Predicting Snowmelt Runoff in a Small, Steep Watershed in British Columbia

The Distributed Hydrology-Soil-Vegetation Model (DHSVM) is a physically based, distributed hydrology model that has been applied to snowmelt-fed mountain catchments in the Pacific Northwest for predicting the effects of forestry operations. Previous applications have focused on moderate to large catchments (from 26 km2 to about 3000 km2), but there appear to have been no applications of DHSVM to small mountain streams (watershed area < 1 km2) controlled by snow accumulation and melt. Such streams are important for domestic water supply in the interior of British Columbia, and the potential hydrologic impacts of forestry in these catchments generate significant concern and controversy. This study evaluates the performance of DHSVM during spring snowmelt in the Gurn Brook watershed located in the Slocan Valley, northwest of Nelson, BC. The watershed has a drainage area of 58 ha and ranges in elevation from 650 to 1600 m. Hillslopes mostly face west, with slope gradients exceeding 20 degrees over 77 percent of the catchment. Slopes are dominantly mantled by thin soils (typically less than 1 m deep) developed in colluvium overlying bedrock. The upper part of the watershed drains a concave (planform) slope that emerges into two ephemeral channels, which join and then flow through a narrow gully to the valley bottom. The digital elevation model (DEM) derived from a 1:5000 contour map with contour interval of 5 m formed the basic topographic data, which was also used to generate stream network. Preliminary evaluation of model performance suggests that explicit spatial quantification of soil, climatic and vegetation parameters appear to be much more important in smaller watersheds for accurate predictions of stream flow and snow cover distribution compared to larger watersheds, such as the nearby Redfish Creek, where DHSVM has previously been applied.