Extended Abstract (420K)6B.12
The National Severe Storms Laboratory's QPE verification system
Steven V. Vasiloff, NOAA/NSSL, Norman, OK; and B. Kaney
A significant challenge in moving research results into operations is robust validation of new applications. Efforts at the NSSL to improve operational flash flood warnings and river forecasts through improved quantitative precipitation estimation (QPE) have included the development of the QPE Verification System, or QVS. The philosophy behind the development of this project is that near-realtime testing across all meteorological and geographical regimes is essential for thorough testing of new methods. Historically, studies have focused on isolated cases, thus limiting the applicability of the results to other situations. The QVS has all the necessary components for a viable community-wide testbed. For instance, satellite, weather radar, numerical model and gauge data are all routinely ingested in their native formats (e.g., Level 2 for WSR-88D). These data are available for various algorithm applications with many experimental NSSL reflectivity mosaic and derived precipitation products output every 5 min at 1x1 km resolution for the continental U.S. The products are then automatically evaluated against operational gauge data, which includes standard statistical analyses (e.g., Heidke Skill Score). The QVS also ingests several NWS operational products for evaluation, including the “gold standard” Stage 4 product. In addition to automatic evaluation metrics, the QVS contains many tools for case diagnosis including time series plotting of up to 5 variables. Data will be available online for up to 6 months.
Session 6B, Radar IIPS and Applications, Part I (Continued)
Wednesday, 23 January 2008, 10:30 AM-12:00 PM, 207
Previous paper
Recorded presentation