Extended Abstract (956K)9A.2
Forecast challenges at the NWS Storm Prediction Center relating to the frequency of favorable severe storm environments
Andrew R. Dean, CIMMS/Univ. of Oklahoma & NOAA/NWS Storm Prediction Center, Norman, OK; and R. S. Schneider
One major challenge in forecast evaluation is the simple fact that some forecasts are more difficult than others. In the case of forecasting severe convection, forecast difficulty is closely related to the convective environment that is present or expected to develop over the area of interest. The frequency of favorable environments over a given time period must be taken into account in order to properly diagnose trends in forecast performance.
Efforts are underway to examine the frequency of important severe weather environments and the relationship of these environments to convective watches. The SPC has developed a database of 3-dimensional environmental parameters associated with severe convection. This database is derived from archived hourly SPC mesoscale analysis grids that are available from 2003 to present, and also includes severe weather reports (138,136 severe reports; 7224 tornadoes), gridded lightning data, and SPC forecast products. Using this database, we have calculated both standard and specialized forecast verification metrics as a function of environment. Basic analysis of the data is performed in “CAPE-Shear” parameter space. This type of analysis allows us to quantify low, moderate, and high values of severe parameters, to classify forecast performance based on environmental characteristics, and to identify the types of cases where there is the greatest potential for forecast improvement. The initial verification efforts focus on severe thunderstorm and tornado watch performance.
Preliminary results support the intuitive notion that watch forecast performance in very favorable (i.e., high values of CAPE and shear) or very unfavorable (i.e. low values of CAPE and shear) environments is better than performance in cases where one or more of the parameters appears to be marginal for supporting severe convection. Forecast performance is quite good in environments typified by high CAPE and moderate shear (often seen in the Plains in late spring), or low-moderate CAPE and high shear (often seen in the Southeast in the cool season). Cases where both CAPE and shear are marginally supportive of severe convection tend to limit forecast performance. In particular, the greatest negative contribution to the probability of detection (POD) for severe hail and tornado events occurs in environments where deep-layer shear is moderate (~ 15-20 m/s), but CAPE is very low (< 500 J/kg). The greatest negative POD contribution for severe wind events occurs in moderately unstable (CAPE ~ 1000 J/kg), weakly sheared (10 m/s or less) environments. Meanwhile, an environmental analysis of false alarm watches is underway in order to assess the potential negative effects of trying to raise POD in marginal cases.
Overall, the goal of this context-based verification analysis is to provide insight into the scientific and forecast challenges that must be confronted in order to improve severe storm prediction.
Session 9A, Forecasting, Warning, and/or Verification
Tuesday, 28 October 2008, 4:30 PM-6:00 PM, North & Center Ballroom
Previous paper Next paper
Recorded presentation