Wednesday, 10 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
This study is aiming to evaluate the skill of short-term WRF quantitative precipitation forecasts (QPFs) for real-time hydrological stream flows and flooding preparedness of the Santa Clara Valley (SCV) in Northern California. The real-time WRF 1km resolution simulations were conducted four times per day to provide short-term (0-36h) QPF for specific watersheds within the SCV. The model’s 12, 24, and 36-h QPF performance is evaluated during the extraordinary high-impact rain season of 2016-2017, which caused the runoff water in the nearby Anderson Dam to overtop its banks and cause record flooding for the Coyote Creek in the inner city of San Jose. Widely-used QPF skill scores (bias, threat score, Equitable Threat Score, Mean Absolute Error, etc.) are calculated for all available rain gauges as well as the Santa Clara Valley Water District Alert rain gauges. Model QPF results show a general wet bias for the high-impact rainfall episodes with larger timing errors of QPFs for the 12-h rainfall forecasts. The ultimate goal of this project is to be able to produce probabilistic forecasts to improve rainfall prediction during high-impact weather events for enhancing flooding preparedness for the greater San Jose area.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner