Hacker News Reader

Comment by cbfrench | Hacker News Reader

While that can still be true, lead times have gotten much better in general, and Mayfield actually had over 2.5 hours of warning from KPAH[0], which is remarkable. But 45 mins or more seems to be the norm these days, at least in my general observations. Local authorities choosing to sound the alarms as soon as the warnings are issued is perhaps another matter.

Edit: Someone else commented elsewhere that the current average lead time on TOR warnings is 9 minutes, so I stand corrected. Some of this probably depends on the storm mode associated with the tornado.

For a supercell, the radar signatures are pretty clear and distinct, so it’s easier to provide earlier warnings. A tornado embedded in a QLCS is more difficult to detect, so I’d expect advanced notification to be more difficult.

[0] See, e.g., https://cliffmass.blogspot.com/2021/12/a-critical-gap-in-tor...

View on HN

Tornado sirens are generally triggered by tornado warnings, which require confirmation of a tornado (either by visual confirmation of tornado, or more frequently, the telltale signs on radar). The blog post you cite talks about sufficient evidence to drive severe storm warnings or tornado watches, which are events that can happen close to weekly for some locations--too frequent to be used for tornado sirens.

It should be obvious that the text from the KPAH tweet at 7:03 is not your run-of-the-mill SVR warning or TOR watch (which aren’t issued by tweet, anyway). It was a very clear alert that a storm that had already produced a visibly-confirmed and dangerous tornado was heading toward that location. NWS offices don’t tweet out that kind of language that far in advance unless they are certain about it. They weren’t going to trigger an actual warning at that point, as there was still time for the storm to lift or dissipate, but when they finally did trigger the PDS warning for Mayfield[0] at 9:05, it had been clearly signaled beforehand that that storm was on the way and that it was serious.

So, no, local authorities aren’t going to sound the sirens based on a tweet over two hours in advance of an event (nor was anyone suggesting they should have done), but emergency managers should have been well aware in advance of the first warning issuance that a very dangerous tornado was on its way.

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

It was pointed out to me above that the 2.5 hour figure is spurious. (Cliff Mass apparently didn’t account for the fact that time stamps on tweets display the time of the tweet in the reader’s local time zone, not in the OP’s—or OT’s in this case.) So, NWS Paducah tweeted that alert at 9:03 CST, effectively contemporaneous with their issuance of the PDS warning. Still a pretty good lead time for a TOR warning, but nothing like hours of notice. In any case, I stand corrected on that.

That much warning time is an eternity and people who may be in the target zone may end up getting nothing.

When the Joplin tornado hit, we lived "in the path" of where it was expected to go (tracking WSW)/ We hunkered down... but the storm ended up going 15 miles south of us. This was just in a short period of time (maybe 20 minutes).

A lot can change in hours. Storms can dissipate or strengthen with frightening speed. Directions can change.

Right, which is why KPAH didn’t issue an actual warning until 9:05. Giving people a strongly-worded heads-up a couple of hours in advance via tweet gets the word out to emergency managers and other interested parties who can then disseminate that information as the situation develops. Once it becomes clear the tornadic cell hasn’t weakened or deviated, then you can trigger the actual alert systems. But I don’t think anyone can argue that no one had any idea what was coming until minutes in advance.

They had 2 days of warning, not 2.5 hours.

Reply on news.ycombinator.com