A Review of the 3-Day Tornado Forecast

Much of the eastern half of Kansas has been in a tornado drought.

So, I worried that people were out of practice when it came to dealing with major tornado threats. So, I made the first-ever (for this blog) 3-day advance forecast for the tornadoes that occurred Tuesday (12th). Because I urged readers to take preparatory action, it is only fair that I review the forecast as it was not a success.

The Forecast

The forecast was for "strong" tornadoes which meteorologists define as EF-2 or greater. Because the high risk area (red) has been nearly tornado-free the last five years, I wanted to give readers plenty of time to prepare. 

What Actually Happened

Red = Tornado. Blue = Damaging thunderstorm winds. Green = large hail.

There were strong tornadoes in both Iowa and Texas. The image below, from CBS Austin (via Twitter), was of the Salado, TX tornado's aftermath.

But, with the exception of the tornadoes in southeast Nebraska, the tornadoes were outside of my highest risk area. In that sense, the forecast was a bust. 

There were thunderstorms in Kansas but no tornadoes occurred.

What went wrong? 

The answer is that I am not entirely sure. While the proximate cause was that temperatures 7,000 feet above the ground were too warm, I'm not sure why temperatures aloft over Kansas didn't cool Tuesday afternoon as I forecast. While I make my own forecasts, what makes this especially frustrating is that the best model in the world, the ECMWF, and others showed thunderstorms developing. Had they developed, atmospheric conditions were such that I'm nearly certain that strong tornadoes would have developed. I did learn that I made a mistake of making the orange area too small, which is a mistake I have made before. 

How would I rate the original forecast?

• Strong tornadoes would occur?  "A"
• Timing? "A"
• Geographic accuracy?  "D"  The only thing that keeps it from being an "F" were the tornadoes in southeast Nebraska and near Shreveport. 

One thing that was correct was my prediction Tuesday morning. Unless it is absolutely necessary, the day of a tornado event, I try to stay consistent with the National Weather Service's Storm Prediction Center. Social science studies demonstrate that confusion makes people less likely to act. However, I felt strongly that if supercell thunderstorms developed, there would be a relatively high tornado risk.

Of course, strong tornadoes occurred in that area. 

I'll continue to make these forecasts. Hopefully, the quality will be better in the future. Thank you for reading.