Disentangling the driving mechanism of streamflow trends using runoff senstivity to land use and climate change.

Abstract:

The majority of watersheds within the United States have been disturbed by anthropogenic land use change. On top of this, there is strong evidence of (historic and projected) climatic changes that affect earth's hydrologic cycle. Streamflow measurements integrate the effects of land use and climate change on watershed hydrology. Therefore, when temporal trends are present, teasing out the cause is challenging due to the overlying climate and land use signals. In this study, we develop an analytical framework for distinguishing trends in streamflow that are driven by climate change from those that are driven by land use change. This framework is based on the theory that during wetter years runoff is affected more by changes in climate than during drier years. Whereas, the inverse is true for land use change. During wetter years runoff is affected less by land use change than during drier years. This difference can be seen in the quantile regression of the 75^th and 25^th percentile annual stream flows which represent wetter and drier years, respectively. This creates a defining characteristic in how these two forcing mechanisms manifest within the streamflow record. We empirically test this framework and show that the sensitivity of runoff to climate and land use change is uniquely dependent on the spatiotemporal water and energy limitations of a catchment. Finally we apply the framework using 1,566 watersheds across the contiguous United States. We use gages from the United States Geological Survey (USGS) National Water Information System (NWIS) network. The gages are selected because they have continuous and complete data from the years 1950 to 2009 and represent watersheds which are characterized by a range of disturbances. Our results show that the driving mechanisms of streamflow change across the U.S. are regionally coherent and correspond with land management activities and climate zones. This methodology provides a simple means of classifying watershed to regional scale hydroclimatic change without relying on reference stream gages, complex models, or observational climate networks.