Quantifying the relative impact of climate and human activities on streamflow

Abstract

The objective of this study is to quantify the role of climate and human impacts on streamflow conditions by using historical streamflow records, in conjunction with trend analysis and hydrologic modeling. Four U.S. states, including Indiana, New York, Arizona and Georgia area used to represent various level of human activity based on population change and diverse climate conditions. The Mann–Kendall trend analysis is first used to examine the magnitude changes in precipitation, streamflow and potential evapotranspiration for the four states. Four hydrologic modeling methods, including linear regression, hydrologic simulation, annual balance, and Budyko analysis are then used to quantify the amount of climate and human impacts on streamflow. All four methods show that the human impact is higher on streamflow at most gauging stations in all four states compared to climate impact. Among the four methods used, the linear regression approach produced the best hydrologic output in terms of higher Nash–Sutcliffe coefficient. The methodology used in this study is also able to correctly highlight the areas with higher human impact such as the modified channelized reaches in the northwestern part of Indiana. The results from this study show that population alone cannot capture all the changes caused by human activities in a region. However, this approach provides a starting point towards understanding the role of individual human activities on streamflow changes.