Evaluation of Precipitation Forecast Uncertainty During Extreme Atmospheric River Events

Atmospheric rivers are meteorological phenomena of a low-level horizontal moisture transport that can produce extreme precipitation over the Western United States and elsewhere. The extreme precipitation also impacts runoff generation and often leads to significant flooding in flood-prone watersheds in the West. A detailed atmospheric forecast that can improve the representation of these weather systems will help to ensure reliable water supply and mitigate flood risk. The Center for Western Weather and Water Extremes (CW3E) has developed a customized version of the Weather Research and Forecasting model (called West-WRF), to better forecast atmospheric rivers that make landfall in northern California. However, a deterministic forecast suffers a lack of uncertainty information that helps risk-based decision-makers. Therefore, different ensemble forecast approaches are used to quantify the uncertainty prediction. In this study, we will examine a 1-km nested precipitation forecast ensemble during an extreme atmospheric river that occurred between February 6-9, 2017 and had a significant impact on the Russian River Watershed and contributed to the Oroville Dam spillway incident. This case study ensemble is analyzed to evaluate West-WRF uncertainties by performing a model-data comparison of precipitation and other meteorological variables critical for the representation of atmospheric rivers. Two different techniques were used to build this ensemble: (1) different physics parameterization and stochastics kinetic energy backscatter schemes (SKEBS). Statistics (e.g., rank histograms, spread-error correlation) are used to test the spread of the ensemble precipitation and to estimate how well are we representing the uncertainty in precipitation. This study allows us to apply this ensemble in hydrological models to estimate streamflow sensitivity.