Deficient Tornado Warning in Arkansas

I am a huge fan of the National Weather Service (NWS) and its core mission of collecting data and making it available, maintaining the meteorological infrastructure and its vital mission of forecasts and storm warnings for the public-at-large. I even wrote a book on the subject. I've called the NWS "one of the few essential federal agencies."

Unfortunately, in recent years, the NWS's tornado warnings have become less accurate both in terms of tornado detection (known as "probability of detection" or PoP) and "lead time" (the amount of time between the warning and the arrival of the tornado). The National Weather Service, and its parent agency NOAA, have a published lead time goal of 13 minutes. The NWS routinely achieved this from 2005 to 2011. 

Because the United States lacks a Natural Disaster Review Board (other nations have them), we don't know all of the reasons the tornado warning program has deteriorated. But, I believe it is valuable to document these cases so that these issues may eventually be solved. I published an analysis of a Georgia situation yesterday.

Another example of the deteriorated tornado warnings occurred last night in Arkansas. The tornado in the Hot Springs Village had a lead time of, at best, negative one minute. At left, is the NWS's survey done today and, at right, last night's tornado warning. The reason I stated "at best" is because at least two witnesses (one a meteorologist) said the tornado began to the west of the highway (see map on survey image). If correct, it would mean the tornado lead time was worse. 

To be clear, this was not a situation that lent itself to 13 to 15 minutes of lead time (the research shows that amount is ideal). But, at least some positive lead time was possible. 

Radar Review

At 7:51pm, there was rotation (right panel) but it did not match well with the reflectivity data (left). I would not have issued a warning at this point. 

However, at 7:53, an absolutely clear sign of a tornado presents itself. 

At right, the velocity data shows "gate-to-gate" rotation -- a clear tornado signal. At least, the circle shows the G2G rotation is now at the northeast tip of the developing hook echo. A tornado warning should have immediately been issued. Allowing a minute or so to get the warning out, that would have allowed a positive four minutes of lead time. That is enough time to dash into a shelter. 

The rotation continued on subsequent radar frames. 

News reports say there was one injury. That was lucky given it was a a strong tornado. 

There is legislation before both houses of Congress to create a Natural Disaster Review Board. If you support the concept, please contact your congressional delegation via their web sites. It is easy to do.