Estimating hydrologic alteration from basin characteristics in Massachusetts

It is clear that humans are impacting the water cycle. There is interest in both determining where and how aquatic systems are most impacted by human development, and in determining the types and locations of basin modifications that are having the most impact. Instead of complex physical models of individual basins, we propose the use of a statistical approach to look at the relationship between human basin modifications and the resulting impacts on streamflow. We develop a set of multiple linear regression models, using principal component analysis to guide independent variable selection, to estimate current, altered streamflow from a full range of both natural and anthropogenic basin characteristics. Natural streamflow is then estimated by simulating basins with no alterations, and the difference between the natural and altered streamflow are summarized by use of the ecochange percent metric. The model suggests that dam storage, water withdrawals and discharges, and land use all impact stream flow and non-point source land use modifications such as impervious cover are potentially increasing low flows. The approach provides an opportunity to increase our understanding of the relation between human basin modifications and changes in streamflow. The model developed could potentially be used to estimate streamflow alteration at ungaged sites.