• A friendly and periodic reminder of the rules we use for fostering high SNR and quality conversation and interaction at Stormtrack: Forum rules

    P.S. - Nothing specific happened to prompt this message! No one is in trouble, there are no flame wars in effect, nor any inappropriate conversation ongoing. This is being posted sitewide as a casual refresher.

Evidently, At Least One NWS Office Has Decided to Get Out of the Tornado Warning Business

It just keeps occurring.

Please note the comments from the Star-Telegram article. My gosh, why did we taxpayers fund the NEXRAD network if they are going to wait around for multiple visual sitings and/or two consecutive volume scans ( = 9 minutes since the radar was set at 4.5 minute data intervals)?

“We like to have at least two volume scans in our vision, in our pocket, in our queue, before jumping on the warning message,” he said. “By the time we had that second scan that really made it clear we had a tornado, that’s when the warning decision was made.”
Read more at: https://www.star-telegram.com/news/weather-news/article276975443.html#storylink=cpy

Have you written your congresspeople?
 
I saw that one about the 1400 lb disc when looking through storm reports. .wow.

As far as having more than one scan before posting a warning, to some degree I can't blame a NWS office for wanting to reduce the number of false alarms. (though they deff should be increasing the scan rate!)

Oh and now I see where the jokes/comments about DVD size hail came from!

As far as writing my congressmen. Nope. Ain't gonna do that
 
I'm not into writing to congressmen/etc. Not something I've ever done. Its just not my thing.

And while the local NWS office may not be perfect, I don't think its so broken that intervention is needed.

Oh and maybe selfish/maybe-not ... but where I am, I personally have basically zero worry that I'll ever be hit by a tornado.
(though "never say 'never' " does apply since I was hit by that storm that dropped a tor in Highlands Ranch - before it got to that neighborhood/before it became tornadic)
 
There was an EF1, that to my knowledge, went unwarned in Delaware on the 9th. I don't believe it's on the level of your example above, as it was a different set of circumstances (QLCS), but is in-line with some of your previous posts.

I'm actually kinda irked with myself as I had monitored the situation prior to leaving for work, and was basically looking right at it (albeit from 10 miles) without even realizing it. I saw an interesting lowering while on the road, but chalked it up to SLC, as the shear was nonexistent. Looking at radar after the fact, it seemed to be the product of a cell merger, and a weak velocity couplet was only evident for 2 or 3 scans. Given the environmental parameters and brevity of the radar indications, I'm not sure it would be fair to lay blame. There was some homes damaged, but thankfully, I don't believe anyone was injured.
 
There was an EF1, that to my knowledge, went unwarned in Delaware on the 9th. I don't believe it's on the level of your example above, as it was a different set of circumstances (QLCS), but is in-line with some of your previous posts.

Hi Mark,

Thanks for chiming in.

There are lots of unwarned tornadoes for various reasons. Some are just not-warnable. Others' have durations that are just too brief. I generally don't get into those nor do I fault the NWS when they occur.

My concern is that obvious, strong tornadoes are going unwarned -- which puts the public at extreme risk. While I've written most recently about Texas and Colorado, earlier this season there were major tornadoes in Florida and Virginia that went unwarned.
This doesn't include all of the major, unwarned tornadoes before spring, 2023.

The situation is only getting worse. I don't see what is going to change without pressure from Congress. Dr. Spinrad at NOAA doesn't seem to care and the NWS management is in denial (yes, I've communicated with them).

Plus, one person pointing these out doesn't have a lot of credibility. It is going to take many of us to make a difference.

Those are my 2¢. Thanks again.

Mike
 
I haven't looked at this topic in detail enough to have a strong opinion on it, but one thing sticks out to me in the data that Mike presents is the dates. Mike is comparing "2005-2011 13.3 min lead time" and "2012-2020 8.4 min lead time". I asked myself what changed during that period and the first thought was WSR-88D went dual-pol (which Mike also noted along with GOES lighting detection advances). I went and checked, and yes - 2011 to 2012 is when nearly all the WSR-88Ds were converted to dual-pol. I know with that change the list of things to consider when issuing a tornado warning got longer. @Mike Smith - Do you have any year-by-year data you can share or point us to?

Earlier in my career I spent a great deal of time working on computer malware detection (if you look me up I have a number of patents in that area). One of the things I learned in that work is there are things we could do that would detect every malware sample we sent at the system, but those things also flagged an unacceptable number of things that were not malware. It is very hard to balance false positive and false negatives. I suspect the same is true when it comes to radar indicators of tornados. While a false negative in tornado warnings is much worse than a false positive, the false positive has a cost of time that is hard to measure - people start ignoring tornado warnings when they believe them to be inaccurate.

I can't help but wonder if we aren't drowning in a sea of data. As a chaser I know that having more data doesn't always help me make better chase decisions, in fact sometimes I make worse decisions. Some of my best chases are when I rely less on technology and get back to the basics.
 
I can't help but wonder if we aren't drowning in a sea of data. As a chaser I know that having more data doesn't always help me make better chase decisions, in fact sometimes I make worse decisions. Some of my best chases are when I rely less on technology and get back to the basics.

This is highly insightful. Dr. Tom Stewart of SUNY-Albany did studies into this and found that meteorologists love more data because it makes them feel good about the forecast but it does not make the forecast more accurate.

I've had an epiphany w/r/t models. I've pretty much down to the HRRR, 3 km NAM, and FV3 CAMs and ECMWF and Canadian beyond 48hr. For hurricane season, I'll add the UKMET. The rest are just noise.

Going back to radar, D-P was sold, primarily, because it was going to radically improve flash flood warnings. There may be some improvement, I don't know (the numbers are behind a login and password) but it certainly doesn't seem like there has been a radical improvement. It may also increase side lobes in velocity data.

I had an incident in my own shop when a meteorologist didn't want to issue a flash flood warning (it would have been catastrophic if she hadn't) because the D-P rainfall numbers were so high she didn't trust them! I told her there was an easy way to check which was the ground truth of the OK Mesonet. She did and they matched. Out when the flash flood warning -- which stopped a major train derailment. There was just so much supposedly sophisticated data to wade through, the basic didn't occur to her.

Last Saturday, we had no rain in the Wichita forecast for that evening. By 6p, the sky was filled with ACCAS. Sure sign of thunderstorms developing in the area. Yet, the forecasts weren't changed until after the thunderstorms formed. We don't even seem to be able to relate cloud formations to near-term weather anymore.

If I woke up as king of weather tomorrow, I would completely change how we train meteorologists. If we can't relate clouds to future weather, why do we continue to need humans?
 
Dr. Judith Curry, was nice enough to request a post on this topic for her blog. It has a partial list of cases for 2021-present.

Mike, in this article you cite a 24% decline in PoD, but it’s actually 14%-points (from 73% to 59%) or a 19% decline (14/73). Doesn’t change the spirit of the argument at all, and it’s a stark decline regardless, but just think it’s important to be as accurate as possible for credibility with the readers.
 
Back
Top