Biases in Rain-on-Snow Days in Climate Models

When rain falls on a snowpack, a Rain-on-Snow (ROS) event, the combination of rain and snowmelt can lead to severe hazards such as flooding, avalanches, and landslides. ROS events are a forecasting challenge as they often produce larger flooding events than expected. "Why, how, and when do rain-on-snow events produce exceptional runoff?" is one of the unsolved questions in hydrology (Blochel et al. 2019).

How ROS events may change in the future has been explored in regional climate simulations (e.g. Jeong and Sushama 2018), but our confidence in these projections depends on how ROS events are represented in climate models during the historical period. In this study we use observational data from 1982 to 2011 to characterize ROS days over North America. We then use the same methodology to detect ROS days during the same time period for simulations in the Coupled Model Intercomparison Project Phase 6 (CMIP6). For the region east of the Rocky Mountains, the CMIP6 simulations tend to produce fewer ROS days than the observational data and produce a different trends in the number of ROS days.

We also consider simulations with fixed sea surface temperatures, higher resolutions, and regional domains to explore which aspect of these models results in these biases. Finally, we explore the relationship between ROS days and precipitation and snowpack characteristics in the models. Models that produce the fewest number of ROS days tend to have the lowest fraction of days where snow melts after precipitation, however, when a ROS day occurs these models also tend to produce a larger decrease in snow water equivalent.