CALIBRATION OF THE DISTRIBUTED HYDROLOGY-SOILS-VEGETATION MODEL TO THE LAKE WHATCOM WATERSHED, WASHINGTON STATE

Lake Whatcom watershed occupies an area of 36,270 hectares in the North Cascades foothills in northwestern Washington. About 80% of the watershed is forestlands with pockets of urban development. Lake Whatcom is a 2,040 hectare lake in the watershed that provides drinking water for over 86,000 people in the region. The objective of local water managers is to preserve the lake as a long-term source of drinking water, both in terms of water quantity and quality. Our goal is to quantify surface and ground water inflow to the lake under varying climatic conditions using the Distributed Hydrology-Soils-Vegetation Model (DHSVM). The lake level and all other hydraulic inputs and outputs of the lake are measured. Three precipitation gauges and two weather stations exist in the watershed. Discharge is monitored on three perennial streams. There are about 36 small streams in the watershed.

DHSVM is a physically based, distributed hydrologic model that simulates a water and energy balance at the scale of a digital elevation model (DEM). The inputs required by DHSVM are GIS grids of the topography, watershed area, soils, and vegetation. The input grids were formatted using ArcInfo software. USGS 10 meter DEMs provided the topography of the watershed. The soil grid was created from the CONUS soil database which is formatted specifically for climate and hydrologic modeling. The USGS National Land Cover classification grid was used to define vegetation within the watershed. Simulations were performed using 30 meter square grids of the watershed. Required meteorological data for DHSVM include precipitation, air temperature, relative humidity, wind speed, shortwave radiation and long wave radiation.

The model was calibrated to the watershed using streamflow data from two of the gauged steams and meteorological data from two weather stations. Hydrologic conditions were simulated using one-hour time steps for two water years of available data. Preliminary results indicate that the model validates the timing of peak inflow rates into the lake based on water budget analyses.