H205-02
Linking Watershed Characteristics to Runoff Generation Processes over Large Scales

Wednesday, 16 December 2020: 08:34
Virtual
Hilary K McMillan, San Diego State University, Department of Geography, San Diego, CA, United States and Sebastian Gnann, University of Bristol, Bristol, BS8, United Kingdom
Abstract:
Runoff generation processes are often studied at the hillslope or watershed scale, however there would be many advantages to mapping watershed processes across the landscape at national or continental scales. Such an advance would enable us to tailor continental-scale hydrologic models to represent processes in diverse hydro-climates, and enable watershed interventions such as interception of floodwater or contaminants to be targeted towards dominant flow paths.

In this presentation, we propose a new conceptual framework to map hydrologic processes at large scales. This is achieved by linking physical watershed attributes to streamflow signatures, and in turn, linking streamflow signatures to hydrologic processes. Streamflow signatures are metrics that quantify streamflow dynamics, and provide a useful way to quantify runoff generation mechanisms. Extensive work in experimental watersheds has demonstrated the link between watershed processes and the dynamics observed in the stream.

Our work shows two major advances: (1) To strengthen our ability to predict streamflow signatures from physical watershed attributes by redesigning watershed attributes and signatures to target specific processes (2) To make increased use of regional process knowledge in large-scale predictions of streamflow signatures.

We demonstrate the framework with an application to predicting baseflow signatures in the U.S. Using examples from several U.S. regions, we show that region-specific knowledge is essential to quantifying how watershed attributes control baseflow index and recession function, but is currently underutilized in large sample studies. We show how a combination of baseflow signatures is needed to distinguish between different baseflow sources, and thus between hydrologically different catchments. The results demonstrate how selecting watershed attributes and streamflow signatures based on process knowledge improves our ability to link the form and function of a watershed.