Americans getting less advance notice for tornadoes; researchers struggle to understand why

[Steve Miller]

Over the past decade, scientific understanding of tornadoes has improved, and the technology available to detect them has leapt forward. Yet recent statistics show National Weather Service forecasters are providing less advance notice for tornadoes than they did five years ago and more frequently failing to detect them.

Read more: https://www.washingtonpost.com/news...gle-to-understand-why/?utm_term=.8e64eb2a4cdb

 

[rdale]

A big part was the NWS decision to count an unwarned tornado as a warning tornado with 0 minutes of lead time. There's no scientific reason for that to have happened as all it does is skew the numbers.

 

[Greg Flint]

We should have two sets of statistics, in my opinion, because ACCURATE advance notice for an EF0 tornado should not be weighted equally as an EF3-EF4 or EF5 tornado. I wonder what two separate sets of data would look like if we looked at EF0/EF1 tornadoes and advance notice vs. advance notice for EF2 & greater.

 

[Randy Jennings]

What always bugs me when folks start talking about verification of warnings is that a warning that results in a funnel cloud or a rotating wall cloud about to touch the ground but never produces a tornado is a false positive (as does a real tornado that touches down in the middle of no where and no one reports the damage). Yet if NWS waits until it is a tornado - then it is zero lead time. They just can't win. I don't think anyone here would say that a funnel cloud should not be warned, but yet it counts against them if it doesn't touch the ground and do damage. Having said that, it is important to have discussions about the policies, training, an science (both meteorological and psychological) that goes into the warning progress.

 

[rdale]

The number of tornadoes that occur in tornado warnings, but are not reported, is likely a VERY small percentage. That doesn't mess with stats very much. It could occur once or twice in rare storm areas, but I can't for the life of me imagine a tornado touching down in KS that is too brief to be seen and/or too small to cause damage. And I'm not sure I understand your other point - a rotating wall cloud is not a tornado. It does no damage. Therefore it should not "qualify" as a tornado.

 

[Randy Jennings]

The point I was trying to make about the wall cloud was that if the goal is 13 mins lead time before tornado touch down, a forecaster can't wait until a funnel forms to issue the warning. The time from funnel to tornado is typically a very short time frame. At some point, a forecaster has to go with what they are seeing on radar and reports from spotters/chasers even before a funnel forms. Yet the majority of wall clouds never produce a tornado and thus count as a "false positive" against the office if it doesn't end up with a tornado. Sure the wall cloud causes no damage, but a forecaster can't wait until it is a tornado to issue a tornado warning if we want advanced warning.

 

[Randy Jennings]

I should probably add that I'm not saying all wall clouds should be warned. I'm simply saying that goals of 13 mins lead time and almost no false positives are seemingly in conflict with each other borrowing some major advancement it forecasting. It is a great goal we need to work towards, but it is an unrealistic goal at this time. Every time one of these stories come out the general public is left with the impression that NWS isn't doing a good job, and I think that is grossly unfair.

 

[Royce Sheibal]

This is just completely from my own experience here locally in the last couple of years, by all means not a scientific study. We've had 4 confirmed touchdowns in the last 12 months in the Omaha metro. None were particularly large, but a couple were EF2's that did some damage and a couple minor injuries. All four touchdowns were not tornado warned until they had been already confirmed on the ground, despite what appeared to be decent (however not significant) rotational structures for 15-30 minutes beforehand. Outside of those four touchdowns there have been at least a dozen other cells that *should* have been warned based on their structure, but were not warned, and did not produce. I would say that on average, less than half of cells with good structure, the kind that I, as a chaser, would go chase after, are warned in this area, and the warning that have occurred on actual touchdowns have been zero (or negative) lead-times.

 

[Taylor Wright]

Other chasers as well as I have reported tornadoes on the ground where NWS refused to warn the storm, whether it be because the "tornado was too weak" or because their velocity data didn't support a tornado forming. In this sense, an advance in technology might create a false sense of confidence to not warn a storm, because tornadoes often form in unfavorable velocity signatures. In one instance I had to provide video evidence of the tornado being on the ground for them to even "count" my report. The storm was never warned.