Impact of 20 Years of Land-Cover Change on the Hydrology of Streams in the Southeastern United States

Land-cover changes for portions of Alabama, Georgia, and Tennessee were estimated for the years 1980, 1990, and 2000 using classified Landsat images, and associated with hydrologic indices for 12 watersheds in the region. Rainfall-adjusted mean annual streamflow, an ET proxy (precipitation minus runoff), frequency of inundation above thresholds, and duration of inundation were used to characterize the hydrologic response of the test basins over the two-decade study period. Results indicate that several of the watersheds had undergone significant (> 20%) reductions in agricultural land cover with a coincident increase in forested land. Attempts to correlate the hydrologic results with the land-cover changes were only partially successful. Watersheds with the largest land conversion from agriculture to forest (20% or more) did show significant trends in hydrologic indices indicating decreasing streamflow; however, other basins evidenced ambiguous results. The net conclusions of the study are that land-cover effects on hydrologic variables may be nuanced and can sometimes be only indirectly evident, and that a rigorous and detailed land-cover classification effort along with a battery of statistical tests with the same objective may be necessary to uncover these effects.