Paper No. 28-18
Presentation Time: 8:00 AM-5:30 PM
APPLICATION OF THE USGS SOIL-WATER-BALANCE (SWB) MODEL TO ESTIMATE GROUNDWATER RECHARGE IN CENTRAL IOWA
DAY, Erik, Department of Geological and Atmospheric Sciences, Iowa State University, 2237 Osborn Drive, Ames, IA 50011 and SIMPKINS, William W., Département de géologie et de génie géologique, Université Laval, Pavillon Adrien-Pouliot, 3748, avenue de la médecine Local 4309, Quebec, QC G1V 0A6, Canada
Groundwater recharge is an important parameter for groundwater flow and contaminant transport-modeling, estimating aquifer sustainability and assessing climate and land use/land cover change. Methods to estimate recharge include assuming some percentage of precipitation, baseflow separation from streamflow, rainfall-runoff modeling, using a groundwater model to estimate recharge via model calibration, and the 1957 Thornthwaite-Mather soil-water-balance approach. The latter approach was the subject of a dissertation by Weston Dripps at the University of Wisconsin-Madison in 2003. He incorporated GIS coverages and lookup tables into the model and, for the first time, was able to estimate the temporal and spatial distribution of recharge at the watershed scale. In 2010, Stephen Westenbroek, with the USGS Water Science Center in Wisconsin, modified the Dripps code to be more user-friendly and the model is known as SWB.
In this study, the SWB model was used to estimate annual groundwater recharge rates for central Iowa during a 20-year period (1996-2015) at a spatial resolution of 100 meters. The mean annual recharge rate for the 20-year period is 6.3 inches. Recharge values are closely related to surficial geology and precipitation. The lowest annual recharge was 0.3 inches in 2000 during a drought year with only 27.6 inches of precipitation. The highest annual recharge was 14.3 inches in 2008 during the second wettest year with 47.8 inches of precipitation. A pattern of low recharge is present north of the Altamont moraine, shadowing areas of lacustrine sediments. Sensitivity analysis and calibration were performed by manually adjusting soil-curve-runoff numbers at fixed intervals. Regression was used to compare annual recharge to baseflow estimates from four watersheds. The model shows good correlation with an R2=0.76. This model accounts for spatial variability while still accurately accounting for the watershed scale resolution.