Projected increases and shifts in rain-on-snow flood risk over western North America

Author: K.N. Musselman, F. Lehner, K. Ikeda, M.P. Clark, A.F. Prein, C. Liu, M. Barlage, R. Rasmussen

We present an analysis of daily ROS events with flood potential simulated at high resolution (4 km horizontal grid spacing; Supplementary Fig. 1) over western North America using the Weather Research and Forecasting (WRF) model. Two 13-year simulations were used: a historical reanalysis (2000–2013), verified against an observationally constrained snow model (Supplementary Fig. 2), and a pseudo-global warming climate simulation (see Methods and ref 24). In our analysis, we focus on three primary mechanisms that can affect ROS flood potential in a warmer climate: (1) reduced snowcover persistence25 decreasing ROS frequency and spatial extent, (2) a greater proportion of annual precipitation falling as rain on a pre-existing snowpack26 increasing ROS frequency and spatial extent, and (3) more intense rainfall27 and/or snowmelt increasing ROS intensity. We assess how these change mechanisms interact and combine over elevation gradients to impact future ROS water available for runoff. Finally, we assess projected changes in basinscale potential runoff from the ten largest ROS events simulated in the historical and warmer climate scenarios. We present the first high-resolution process representation necessary to understand how this major flood hazard may respond to climate change over large mountainous areas.

Our results suggest that increases in ROS water available for
runoff coupled with a greater proportion of precipitation falling
as rain, and increased rainfall intensity, should be considered in
flood control planning. The largest increases in ROS runoff are
projected for mountainous river basins that are historically prone
to flooding and/or in regions where the storage and transport
of snow water resources are paramount. For example, the water
volume produced by the most intense daily ROS events simulated
for the Sacramento River basin in California is projected
to increase by > 20% in the warmer climate scenario. This basin
contains the Oroville Dam, which was critically damaged by emergency
water releases during a 2017 ROS event. This river basin
and many others we highlight include some of the largest metropolitan
regions in western North America, underscoring the
potential societal and economic impacts of the projected changes.
Thus, flood control and reservoir management systems in these
mountainous regions must consider future changes in rain-onsnow
events to fully quantify changes in basin-scale flood risk
with anthropogenic warming.

Our results suggest that increases in ROS water available for
runoff coupled with a greater proportion of precipitation falling
as rain, and increased rainfall intensity, should be considered in
flood control planning. The largest increases in ROS runoff are
projected for mountainous river basins that are historically prone
to flooding and/or in regions where the storage and transport
of snow water resources are paramount. For example, the water
volume produced by the most intense daily ROS events simulated
for the Sacramento River basin in California is projected
to increase by > 20% in the warmer climate scenario. This basin
contains the Oroville Dam, which was critically damaged by emergency
water releases during a 2017 ROS event. This river basin
and many others we highlight include some of the largest metropolitan
regions in western North America, underscoring the
potential societal and economic impacts of the projected changes.
Thus, flood control and reservoir management systems in these
mountainous regions must consider future changes in rain-onsnow
events to fully quantify changes in basin-scale flood risk
with anthropogenic warming.