For large populations across the western U.S., water supply prediction relies centrally on knowledge of spring snow conditions, where snowpack can provide critical early warning of drought. As future temperatures rise and snowpack dwindles, the relationship between snow and streamflow is expected to shift. Recent research found that in the future, snowpack will be less predictive of drought in snowmelt-dominated systems in the western U.S.
In this three-year project, we will develop and evaluate new techniques for drought prediction in the Intermountain West and Pacific Northwest that go beyond exclusively snow-based prediction methods, harnessing alternative datasets to identify better ways to predict and respond to drought. A key innovation will be the use of Machine Learning tools to explore different models to improve current and future drought prediction.
This research seeks to incorporate the needs of western U.S. water management entities, considering regional characteristics and shifts to a warmer, less snow-dominated future climate. The team will collect direct input from regional water managers to help shape the modeling and Machine Learning work and assess the feasibility of alternative strategies.