Evaluation of the Spatially Varying Water Balance Processes in a Semi-Arid Mountainous Watershed of Idaho Using the Soil Water Assessment Tool (SWAT) Model

Publication Date


Type of Culminating Activity


Degree Title

Master of Science in Hydrologic Sciences



Major Advisor

Venkataramana Sridhar


The distributed Soil Water Assessment Tool (SWAT) hydrologic model was applied to a research watershed, the Dry Creek Experimental Watershed (DCEW), near Boise Idaho to investigate its water balance components both temporally and spatially. Natural variability in hydroclimatology of the basin (with average monthly precipitation ranging between 1.3 mm and 92.3 mm) was large even during the short span of study period (2001-2007). Daily streamflow data from four stream gages were used for calibration and validation of the model. SWAT performed very well during calibration to streamflow predictions with monthly discharge values for the three year calibration-phase fitting observations with a Nash Sutcliffe correlation coefficient of 0.79. The SWAT model proved to do an adequate job of estimating the timing, distribution, and magnitude of the partitioning of precipitation through the water balance components of the DCEW. The three year average (during the validation time period) of the water balance components were: precipitation of 554 mm, ET of 329 mm, drainage of 215 mm, streamflow of 131 mm and soil moisture flux of 73 mm.

Model predicted soil moisture values were compared with data from two monitoring pits for both dry and wet moisture states. Predicted moisture values were 125 mm and 25 mm while observations were about 175 mm and 50 mm, respectively. Monthly spatially distributed soil moisture and evapotranspiration (ET) maps were produced for the growing season of 2007 from the SWAT model outputs. These maps showed significant spatiotemporal variation of ET (varying from 0-350 mm) and soil moisture storage (varying from 0-160 mm). These maps appropriately match expected seasonal trends of ET and soil moisture varying with elevation and aspect throughout the basin, from April through September. First, ET was dominated by the lower elevation grasses, peaking in May. Then, as the lower elevation grasses began to dry out during June, the higher elevation vegetation, dominated by forested lands, contributed the greater amount of the total ET. Parameters pertaining to soil (e.g., available water content and saturated hydraulic conductivity), groundwater recharge (e.g., deep groundwater recharge), and vegetation (e.g., leaf area index and maximum canopy index) processes were estimated matching the overall basin hydrologic conditions during model calibration. This study highlights the necessity for better techniques to precisely identify and drive the model with commonly observed climatic inversion-related snowmelt or rain-on-snow weather events as well as to prescribe soil hydraulic properties and vegetation conditions that more closely represent the conditions in this complex terrain.

Files over 30MB may be slow to open. For best results, right-click and select "save as..."