Seasonal Precipitation Variability Associated with Climate Change and Impacts on Vermont Flood Risk

Abstract:

Concurrent with climate change, the risk of severe flooding has increased in Vermont. Several climatic factors can impact peak flows such as precipitation intensity and duration, snowpack, and air temperature. However, precipitation-return periods are poor predictors of flow-return periods. We show that a primary mode of precipitation climatic change during spring months trends toward more consecutive wet days rather than more precipitation per day. While high duration, depth and intensity precipitation are associated with high flows, the increased persistence associated with regional climate change is shown to be another dominant predictor of increased springtime flood risk. Moreover we find summertime precipitation changes dominated by increasingly intense rainfall events, as expected from higher latent heat following increasing temperature. We show that prediction of high flows must be performed on a seasonal or monthly basis due to seasonally variable connections between precipitation and runoff. Discriminant analysis is used to confirm the role of both persistence and intensity as predictors of seasonal flood risk in the Mad River watershed in Vermont, suggesting seasonally variable modes of climatic change at high temporal resolution as dominant drivers of observed flow non-stationarity.