High Risk: NSSL WRF

Here is a graphic representing an attempt to extract hourly reports from local maxima from the NSSL WRF model for the UH, wind, and graupel variables we are outputting. Graupel is totally uncalibrated, meaning I chose random values that make sense but that is about all I can promise. Wind satisfies the 25.7 m/s criteria for severe, and UH follows a more systematic approach*. The algorithm I use to generate this is a double area, double threshold object identification scheme. It is performed on the hourly fields.

Black dots represent the individual "model reports". The shaded field is a gaussian smoothed (sigma of 160), neighborhood (roi = 100km) approach to represent the spatial extent of the reports for each respective "threat". As reports come in I will update the graphic with red dots indicating the observed storm reports from the SPC web site. There will also be an observed hail and tornado probability contour in blue perhaps overlaid on the UH and Hail graphic. This is a 24 hour period graphic.

The usual caveats apply. This is where the NSSL-WRF generates reports that meet my criteria. I estimate that there is a 1-4% chance that some "model reports" are missing or incorrectly identified. This is EXPERIMENTAL.

The areas identified for the main threat stretch from TX through IN and another corridor in NE.

UPDATE 1:
As far as the timing goes from NSSL-WRF here is how the above UH reports stack up vs the hail reports in time. I am using the Mondrian software as I have detailed in previous posts. The technique used here is called color brushing. I have given each time bin its own color and applied that color to the UH histogram. So prior to 1700 UTC (greenish hues; now) has quite a few weak UH reports. The highest UH occurs in reports after 0000 UTC (bluish hues).