8.2 Probabilistic thunderstorm guidance from the High Resolution Rapid Refresh (HRRR) model: evaluation and evolution of a prototype system

Wednesday, 3 August 2011: 9:15 AM
Imperial Suite ABC (Los Angeles Airport Marriott)
Eric P. James, CIRES/Univ. of Colorado and NOAA/ESRL/GSD, Boulder, CO; and C. Alexander, S. Weygandt, S. Benjamin, and P. Hofmann

Handout (6.6 MB)

The High Resolution Rapid Refresh (HRRR) model, nested within the 13-km Rapid Refresh (RR), is a CONUS-domain convection-resolving configuration of the Advanced Research WRF (ARW) model run every hour at the Global Systems Division of the Earth Systems Research Laboratory. Utilizing a time-lagged ensemble approach applied to the hourly HRRR runs, we generate an hourly convective guidance product known as the HRRR Convective Probability Forecast (HCPF). The HCPF provides a forecast of the likelihood of convection around a given location and time, and has many potential applications.

In preparation for the 2011 convective season, some upgrades to the HCPF are being completed. First, we undertook work to fully incorporate the logistic regression. The logistic regression improves statistical reliability and allows for the computation of optimal (non-constant) weights of the various ensemble members. The coefficients for the regression calculation are derived from a training period in which both uncorrected HRRR forecasts and the verification analyses are available.

Second, work is ongoing to use climatology information in the HCPF (provided by historical verification data). Finally, code flexibility has been greatly improved to allow easy incorporation of future changes. Possible changes include the use of different predictors, use of input from multiple models, and the eventual merging of this algorithm into a more general aviation hazard likelihood guidance product (such as the Very Short Range Ensemble Forecast (VSREF), or a product based on the future HRRR Ensemble).

At the conference, we will describe these enhancements and present initial convective probability from the 2011 season.

This research is partially in response to requirements and funding by the Federal Aviation Administration (FAA). The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner