Hacker News Reader

Comment by syshum | Hacker News Reader

years ago they would only call a warning when they had a confirmed visual on a touch down, I am not sure when they made the change but for at least the last few years (and probably longer) I have noticed them calling more and more warning based on radar signature alone

View on HN

In part because the radar tech has advanced so much. Based on reflectivity, velocity, and correlation coefficient alone, they are able to make a pretty sound guess about whether or not a supercell or QLCS is capable of producing a tornado. There’s no reason why you would want to wait for visual confirmation of the radar output when it could mean the difference (potentially) between giving people in the path a long or a short lead time on the warning.

There is a reason, which is that the new criteria for warning trains people to not take them seriously as they don't necessarily mean acute danger from tornado anymore.

As our radar technology has advanced, it's really a shame that warnings are less indicative of a tornado than they were a decade ago.

A better solution would be three categories of warning (watch, new possibly fake warning, and the old real warning) as each should prompt more response than the lower tier.

The NWS does distinguish between warnings for radar-indicated tornadoes and for confirmed tornadoes on the ground (usually confirmed by spotters), but you have to read the warning text to tell (you'll see phrasing like "particularly dangerous situation" and "confirmed tornado" in the warning when they're working based on something that's been visually seen versus radar indicated). For an example of what this looks like, see the Text Data tab here [1]; this warning (from last night in Nashville) was originally issued for a radar-indicated tornado, which was subsequently observed on the ground and the warning was updated accordingly in Update 1.

But you should take any tornado warning seriously. Yes, they come more frequently than they used to, but that's because we didn't have any advance warning before-- if you wait to get to your safe place until a tornado's been confirmed on the ground, you might not be able to get to safety in time. Or you might not get to safety at all; just because someone hasn't seen a tornado and reported it to the NWS doesn't mean it's not on the ground and deadly.

That's why you have to read into the text to see this distinction: the actions you should take if you're in the path of a PDS (observed) tornado warning and the actions you should take if you're in the path of a radar-indicated tornado warning are the same.

[1] https://mesonet.agron.iastate.edu/vtec/#2021-O-NEW-KOHX-TO-W...

Excellent information; thank you. Communicating all this to the public is a significant challenge. You can't ask people to take shelter more often, and you also can't ask them to read more detailed reports. Compliance rate saves lives, for better or worse.

Also worth noting that “more frequently” for any given location is likely somewhere on the order of 1-2 warnings per year maximum.

The general public can’t grasp the distinction between watches and warnings, and your solution is to divide warnings into subcategories?

I suspect the confusion is the specific but uncommon word choice. Perhaps if they were called something like "tornado likely" and "active tornado", respectively, it would be more widely understood.

Now, that’s a reasonable solution. My only concern is that people won’t take “tornado likely” seriously. But I guess, based on OP, people don’t take “tornado warning” seriously, either.

My proposed solution is to continue to reserve the word warning for its old use, to keep people safe by preventing the cry-wolf effect and to introduce an intermediate term to appease the people who want to ring the alarm more often and communicate with locals who want more information.

I agree that the older, simpler watch and warning system was better, if that's what you're saying.

We agree that the existing watch/warning system as it stands is highly effective. Sure, there are more warnings than there used to be, but that is largely based on the fact that the radar data available to make warning decisions is much more reliable than it used to be. There are potentially more warnings issued that don’t verify, but they’re also based on data that is much likelier to verify. (As opposed to relying solely on spotters to make warning decisions.)

Adding complications to the existing watch/warning system just seems unnecessary. What people don’t grasp is that a tornado warning may not verify for their exact location, while it may (unfortunately) verify for a location two miles away. The fact that some warnings don’t verify doesn’t mean we should change our warning system. What we have in place remains highly accurate. It might be better if people understood how to interpret radar data and to evaluate warnings based on that.

Yes, the bottom line is we have to hope for more sophisticated interaction with weather alerts over time. Events like this weekend's tornado could have that silver lining as the timeline becomes more widely known as it currently appears to have involved tornado(s) that moved too quickly once on the ground for lives to be saved. I'm not optimistic given the careless behavior I've observed living in tornado-prevalent areas--but this was a bad one.

Yeah, there is a reason, after so many times of dropping what you are doing and rushing to the basement, only to have nothing happen, it becomes an annoyance and is not taken seriously.

I mean, frankly, how often are you under a legitimate TOR warning that you’re having to do this? I live in the Midwest—in tornado country, no less—and the number of TOR warnings that disrupt my life in a given year I can count on one hand (assuming there are any).

Also this just misunderstands the threat of tornadoes. A warning is issued for a county-wide area, even though a tornado threatens (at most) a couple-mile wide swath of land. That seems ridiculous, except for the fact that even small deviations in the storm’s path can disrupt lives significantly. The NWS isn’t wrong.

Tornado and other storm warnings haven't been issued by county for several years now-- warnings are issued for a polygon where the storm may impact (plus some margin of error). If you're in a tornado warning polygon now, you'll want to take shelter. If you're not in the polygon, you don't need to (depending on how you get your storm alerts, you might still be alerted by county, though-- that's my big complaint about the current state of weather radio).

https://www.weather.gov/media/pah/WeatherEducation/stormbase...

>>>Tornado and other storm warnings haven't been issued by county for several years now... depending on how you get your storm alerts, you might still be alerted by county, though-- that's my big complaint about the current state of weather radio

To a normal person this is perceived as as a warning being issued for them, even if you are technically correct that the NWS only target a narrow polygon, if a person phone lights up with an EAS saying "Tornado warning", or their TV, or weather radio, that is perceived as them personally being issued a tornado warning.

  > after so many times of dropping what you are doing and rushing to the basement, only to have nothing happen, it becomes an annoyance and is not taken seriously.

Is that feeling widespread? In my area we have sirens for different a type of threat, but everybody that I know drops what they are doing and rushes to the shelter, even if it happens two dozen times day and night.

To what would you attribute the lack of compliance? Attrition? How often is "so many times"? I'm seriously interested because it seems to different - and indifferent.

I have never once dropped what I was doing an rushed to a shelter unless forced to do so by overriding company policy while at my employers place of business.

If I am operating under my own accord if a warning is issued I normally find the nearest weather resource with active tracking of the cell to see if I should even care as warning are no where near targeted enough

If I am in path of the storm I then look to see if this is a confirmed or radar based warning, I then look to other sources (including visual outside) to my my own determination on if it is proper to shelter. 80% of the time I make the choice of Nope do not need to bother

I see, so at least you do check if the warning is pertinent to you. Our warnings cover relatively small geographical areas, so if we hear it then it sure is pertinent to us. But I understand that maybe your warnings cover larger areas.

Thanks for the detail. Here's to a siren-free week!

I'd venture a guess that's due to better radar and better spotting intelligence. When you can't physically see tornadoes when they're rain-wrapped, but have high confidence that they're there, you're putting lives at risk by not activating the sirens.

Reply on news.ycombinator.com