Severe Weather 2023 (1 Viewer)

[staphory]

In my circles, we all think the forecast is a bit on the aggressive side but the low track and dynamics behind the system are strong.

What do you mean by this?

 

[nolaspe]

Welp, Jefferson Parish is under a tornado warning...

 

[Dan in Lafayette]

Some pretty strong thunderstorms just came thru over the last half hour.

 

[bclemms]

What do you mean by this?

That it is yet another squall line that is being presented as a tornado outbreak.

 

[ktulu909]

That it is yet another squall line that is being presented as a tornado outbreak.

Am I the only one that the last second tornado warning for jefferson parish at 10:43 was more of a "just in case something happens" warning? I didnt see jack on radar or velocity that resembled a rotating cell?

 

[gboudx]

Am I the only one that the last second tornado warning for jefferson parish at 10:43 was more of a "just in case something happens" warning? I didnt see jack on radar or velocity that resembled a rotating cell?

According to the Tornado Warning product around that time, it was a "radar indicated rotation". And if you were in Gretna, the threat wasn't near you, it was out towards Destrahan, Waggaman, Avondale and then across the river into Jefferson and Metairie.

 

[bclemms]

According to the Tornado Warning product around that time, it was a "radar indicated rotation". And if you were in Gretna, the threat wasn't near you, it was out towards Destrahan, Waggaman, Avondale and then across the river into Jefferson and Metairie.

Yeah but there was nothing on radar indicating a tornado was immenent or even likely. It was right on the wind shift, looked good on reflectivity but had zero rotation on velocity. We were laughing about it in our group chats, The storm to the north of that one actually looked better as it was coming into Laplace but still only showed weak broad rotation. Honestly, outside of Houston, Orange and a couple other brief spots nothing deserved to be warned. I think the tally is going to be a couple EF1 tornadoes, a few EF0 and maybe an EF2. Then again, when the forecast is for strong tornadoes across 5 states the ratings seem to skew way to the high side and vice versa.

 

[ktulu909]

Yeah but there was nothing on radar indicating a tornado was immenent or even likely. It was right on the wind shift, looked good on reflectivity but had zero rotation on velocity. We were laughing about it in our group chats, The storm to the north of that one actually looked better as it was coming into Laplace but still only showed weak broad rotation. Honestly, outside of Houston, Orange and a couple other brief spots nothing deserved to be warned. I think the tally is going to be a couple EF1 tornadoes, a few EF0 and maybe an EF2. Then again, when the forecast is for strong tornadoes across 5 states the ratings seem to skew way to the high side and vice versa.

Exactly this. I use radarscope as well and know what signature to look for on velocity. There was nothing rotating on the south shore as that line moved through the city.

 

[gboudx]

Interesting that LIX would issue a Tornado Warning, suggesting a radar indicated rotation which apparently didn't exist. So as ktulu mentioned, "just in case". I get better to be safe than sorry, but that's irresponsible to worry people needlessly. The Tornado Watch should cover, "in case it happens".

 

[staphory]

Interesting that LIX would issue a Tornado Warning, suggesting a radar indicated rotation which apparently didn't exist. So as ktulu mentioned, "just in case". I get better to be safe than sorry, but that's irresponsible to worry people needlessly. The Tornado Watch should cover, "in case it happens".

That happens here frequently. People are just starting to ignore it.

 

[bclemms]

That happens here frequently. People are just starting to ignore it.

That is the issue with overwarning. A comprehensive study after the 2011 super outbreak suggested that over warning past events led to a significant increase in fatalities. After that study, NWS made it a point to increase their warning criteria, focus more on false alarm rate and were warning much less.

Problem with that, several weak tornadoes went unwarned and average lead time went down a little and the TV interviews of tornado victims screaming, "we never even got a warning" put pressure on the NWS to not miss even the weakest tornadoes. Next, many local offices issue warnings and to keep their false alarm rate down, the same people issuing the warmings would go out and survey. Guess what? They suddenly started finding a lot more tornadoes. They classify a tree branch down as a weak tornado in order to verify the warning they issued.

Here we are more than a decade later and more stuff is getting warned than ever. We haven't had a real big outbreak in a long time and it's setting up for a disaster.

The same trends are happening with SPC forecasts. The average event seems to be at least 1 level higher on the forecast level despite a huge drop in significant tornadoes over the same period. I'm convinced the forecasters get caught up in the hype circle on social media.

Social media AI is a massive problem. If there are two articles, one with the heading of "Partly cloudy with a chance of thunderstorms in the afternoon, some could be severe" or "Millions under threat for violent tornadoes" which one are you going to click on? So the algos are seeing that people are clicking on the 2nd title and the only metric in which the algos grade quality is clicks. So then it recommends that article and buries the other. Now, everything is calling for the end of the world. SPC is put in the position that if an event overperforms everyone else knew it. So then when nothing happens it doesn't seem to matter because there are 26 tornadoes that were confirmed (most weren't ever a tornado to begin with) and suddenly a system that would normally have been a threat level of 1 was hyped to a level 3 and it is now confirmed by the NWS who called a downed limb a tornado to keep their false alarm rate up.

It's a broken system that is breaking more with each event because now the historical tornado data is getting skewed which will lead to even worse forecasts in the future.

 

[gboudx]

That is the issue with overwarning. A comprehensive study after the 2011 super outbreak suggested that over warning past events led to a significant increase in fatalities. After that study, NWS made it a point to increase their warning criteria, focus more on false alarm rate and were warning much less.

Problem with that, several weak tornadoes went unwarned and average lead time went down a little and the TV interviews of tornado victims screaming, "we never even got a warning" put pressure on the NWS to not miss even the weakest tornadoes. Next, many local offices issue warnings and to keep their false alarm rate down, the same people issuing the warmings would go out and survey. Guess what? They suddenly started finding a lot more tornadoes. They classify a tree branch down as a weak tornado in order to verify the warning they issued.

Here we are more than a decade later and more stuff is getting warned than ever. We haven't had a real big outbreak in a long time and it's setting up for a disaster.

The same trends are happening with SPC forecasts. The average event seems to be at least 1 level higher on the forecast level despite a huge drop in significant tornadoes over the same period. I'm convinced the forecasters get caught up in the hype circle on social media.

Social media AI is a massive problem. If there are two articles, one with the heading of "Partly cloudy with a chance of thunderstorms in the afternoon, some could be severe" or "Millions under threat for violent tornadoes" which one are you going to click on? So the algos are seeing that people are clicking on the 2nd title and the only metric in which the algos grade quality is clicks. So then it recommends that article and buries the other. Now, everything is calling for the end of the world. SPC is put in the position that if an event overperforms everyone else knew it. So then when nothing happens it doesn't seem to matter because there are 26 tornadoes that were confirmed (most weren't ever a tornado to begin with) and suddenly a system that would normally have been a threat level of 1 was hyped to a level 3 and it is now confirmed by the NWS who called a downed limb a tornado to keep their false alarm rate up.

It's a broken system that is breaking more with each event because now the historical tornado data is getting skewed which will lead to even worse forecasts in the future.

Wow, I didn't realize this. Sounds like a recipe for a preventable disaster resulting from a natural disaster. I've been following the FWD office for 20+ years and I don't get this position from them, either from their Watch/Warn products, or their area forecast discussions. If anything, they seem to tend more conservative and patient. It's obvious they monitor the socials because they will include graphics on their home page warning about possible misinformation and/or hype. Their message is usually to set a tone of "use caution" on the social media.

But in regards to tornados, maybe they don't have to "fake it" because tornados are more prevalent in this part of the world, especially the most destructive.

 

[bclemms]

Wow, I didn't realize this. Sounds like a recipe for a preventable disaster resulting from a natural disaster. I've been following the FWD office for 20+ years and I don't get this position from them, either from their Watch/Warn products, or their area forecast discussions. If anything, they seem to tend more conservative and patient. It's obvious they monitor the socials because they will include graphics on their home page warning about possible misinformation and/or hype. Their message is usually to set a tone of "use caution" on the social media.

But in regards to tornados, maybe they don't have to "fake it" because tornados are more prevalent in this part of the world, especially the most destructive.

We have a system in which a person's job performance, promotional path and pay is impacted by these decisions. Then those same people are being asked to grade their test. Next, we have an increase in tornado warnings and tornadoes but only in areas in which tornadoes are typically not visible while the areas like the plains are seeing a significant decrease. Despite all the extra tornadoes in populated areas tornado injuries, fatalities, and financial costs have been decreasing. Many times, it is in places where I am sitting where these "tornadoes" occur and they come back after the fact and call a storm a tornado that wasn't even close to producing. Even when I've presented the video evidence after the fact it gets ignored.

I see it first hand. I'm not saying they are all like that. There are forecast offices that do much better than others. I'm not saying the forecasters have this intention to do harm because they do not. It's just a really poorly designed system that encourages this behavior. There is also bias for tornadoes to be rated higher on days where the risk is deemed higher. It may have always been this way but I've probably seen more tornado damage paths than anyone on the planet at this point and it is painfully obvious to me.

Again, it's not all. I have a lot of friends in NWS offices and they even recognize the inconsistencies. Some are even trying to slowly make changes but the red tape is absurd.

Ok, I'm done discussing it because I could suffer blow back just from discussing these on a public forum.