5B.5 Radar Based Quantitative Precipitation Nowcasting Using a Non-steady Radar Extrapolation Scheme

Monday, 28 August 2017: 11:30 AM
St. Gallen (Swissotel Chicago)
Yadong Wang, Southern Illinois University, Edwardsville, IL; and J. Zhang
Manuscript (1.3 MB)

Short term quantitative precipitation forecast (QPF) plays an important role in flash flood warnings. Many existing storm nowcasting algorithms focus on predictions of individual storm locations through extrapolation of radar observations. The extrapolated storm locations have shown to be relatively accurate and are often superior to numerical model forecasts in very short-term (e.g., 0~3hr). The current study tries to generate quantitative forecast of precipitation through an extrapolation that combines numerical weather prediction model wind fields and a non-steady radar-based storm tracking. The extrapolation tries to segregate a radar precipitation rate or accumulation field into individual cells of different intensities. Motion vectors are estimated by tracking the precipitation cells in time, and temporal variations of the area, total liquid water content and spatial distribution of intensity are calculated for each cell. A background wind field from numerical weather prediction models is adjusted by the estimated motion vectors. The adjusted wind field is used to extrapolate the radar precipitation estimation field to generate 0-3hr precipitation forecasts assuming the same temporal variations of the precipitation cells’ areas and intensity as estimated in the tracking. The non-steady extrapolation QPF scheme is evaluated with a large storm outbreak in Oklahoma and Texas during May 25 to 26, 2015.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner