Model‐based estimation of long‐duration design precipitation for basins with large storage volumes of reservoirs and snowpacks

• Published in: Journal of flood risk management 2024-09, Vol.17 (3), p.n/a
• Main Authors: Hiraga, Yusuke, Iseri, Yoshihiko, Warner, Michael D., Duren, Angela M., England, John F., Frans, Chris D., Kavvas, M. Levent
• Format: Article
• Language: English
• Subjects: atmospheric rivers; Basins; Boundary conditions; Confidence intervals; Dams; Design; design flood; Drainage; Drainage area; Drainage basins; Flood forecasting; Floods; Moisture flux; Moisture transfer; Precipitation; probable maximum precipitation; Rain; Relative humidity; Reservoir storage; Reservoirs; River basins; snowmelt flood; Snowpack; storm sequence; Storms; Weather forecasting; Winter; WRF model
• ISSN: 1753-318X; 1753-318X
• DOI: 10.1111/jfr3.12992

This study proposes a model‐based methodology to estimate design precipitation for long durations during the winter and spring seasons (October to June) through its application to the drainage areas of two dams in the Columbia River Basin, United States. For basins with large reservoir storage or snowpack, design precipitation and floods need to be estimated based on long‐duration processes rather than focusing only on flood peaks or single storm durations. This study used the advanced research version of weather research and forecasting (WRF) model to maximize the target precipitation over the drainage areas by means of the Atmospheric Boundary Condition Shifting and Relative Humidity Perturbation with relaxed moisture flux thresholds. The greatest cumulative basin‐average precipitation depths during Oct–Jun were estimated to be 1220.5 and 1595.4 mm for the drainage areas of Bonneville and Libby Dams, respectively. The 95% confidence interval (CI) of the exceedance probabilities of the estimated design precipitation depths were found to range from 10−3 to 10−5 at Bonneville Dam's drainage area. Those orders were found to be comparable with the documented exceedance probabilities of PMP/PMF in the US. The estimated design precipitation and corresponding atmospheric/land‐surface fields together will drive a physical model to estimate the design flood.