Date and Time
Location
312 Ag Engineering Building
Large watersheds are a patchwork of differing land types, climate, and human influences, which store and export pollutants over time scales ranging from days to centuries. Scientists’ understanding of these watersheds’ response to changing conditions has historically been based on com-plex models with many built-in assumptions. This work describes a simpler alternative, based on a differential ex-tension of the USGS’ SPAtially Referenced Regressions On Watershed attributes (SPARROW) steady-state water-quality model. The extension, called Spatiotemporal Wa-tershed Accumulation of Net effects (SWAN), is designed to empirically link a regional pattern of observed changes in pollutant export, over a multi-year time scale, to maps of changes in suspected causes, while keeping “rules” to a minimum. In a pilot application, we modeled changes in the export of nitrogen, a nutrient whose excess is responsi-ble for decline in Chesapeake Bay fisheries, from 43 bay tributaries between 1990 and 2010. Our model explained 80% of the observed changes, linking overall reductions to changes in inputs from point, agricultural, atmospheric, and urban sources, coupled with increases in temperature and precipitation. Delayed export of nitrogen input before 1990 had a smaller influence, indicating that the water-shed’s “memory” of degradation that occurred long ago, although evident, does not overshadow actions taken to improve regional water quality at the 20-year time scale. By providing hard evidence of the effects of past chang-es, this approach can better inform today’s manage-ment choices and help set realistic expectations for their outcomes.