ESTIMATES OF GROUND-WATER RECHARGE IN MINNESOTA

Regional-scale ground-water recharge in Minnesota was estimated using a generalized least-squares regression analysis. The regression analysis related precipitation, potential evapotranspiration, and soil characteristics to basin-scale recharge estimated by applying the Rorabough method to daily streamflow. Specific yield was modeled as the primary landscape characteristic that affects recharge from precipitation. Specific yield was computed from soil texture data in the State Soil Geographic Data Base (known as STATSGO) using the method published by Rawls and others, which developed a linear regression model that estimates specific yield from soil texture properties. Thirty eight minimally-disturbed basins distributed throughout the state having greater than 10 years of stream-flow data were used in the final Rorabough recharge regression analysis. Greatest recharge occurs in sandplain and outwash areas, but recharge in these areas is sensitive to the amount of annual precipitation and potential evapotranspiration. The greatest estimated recharge was about 32 cm per year in areas of high specific yield (greater than 0.19) and the difference between precipitation and potential evapotranspiration is about 25 cm per year. The lowest estimated recharge as computed by the regression analysis was near 0 cm per year in areas of low specific yield (less than 0.15) and less than 52 cm per year precipitation and generally occurred in the northwestern part where glacial Lake Agassiz sediments are present. The Rorabough recharge regression rates are consistent with ground-water fluctuation methods for estimating recharge, suggesting that this method could be a useful tool for regionalizing recharge estimates from long-term streamflow, specific yield, and precipitation data in other subhumid and humid areas of the world.