James Spann speaks to the media and the NWS concerning

[OHSnow]

Another problem is inconsistency with the various NWS WFOs. Some will blanket tornado warn QLCSs showing any signs of rotation, but others will not. I remember at least three instances where a line of tornado warnings coming out of IWX's CWA magically converted to a line of severe thunderstorm warnings when it hit CLE's CWA.

DTX will also sometimes use these blanket warnings, so often times Lenawee and Monroe Counties might be tornado warned, while Lucas County, OH will have a severe thunderstorm warning for the same system.

[isohume]

I do say NWS going to FB is a good thing. I used it a lot in one recent evening when the threat for severe weather was high. Used it to talk about possible development areas and timing. I got many good responses back.

I think overall it's a good thing...but I've seen other office's FB pages with comments from folks which are a bit out there. Just stupid stuff like "why didn't I get rain?" type posts. Are we going to be obliged to respond to these sort of comments in a courteous manner? If so, who has time for this? With all the chat we are doing already with EMs, media, CWSUs, and other NWS offices do we really need to be including the general public? I know FB is different than our chat responsibilities...but it's not too far removed. To envision a public chat room in the future would be just too much and not needed. In my opinion we (the NWS) are transparent enough already.

[The Great Zo]

I think overall it's a good thing...but I've seen other office's FB pages with comments from folks which are a bit out there. Just stupid stuff like "why didn't I get rain?" type posts. Are we going to be obliged to respond to these sort of comments in a courteous manner? If so, who has time for this?

I don't think there's a large-scale expectation that all of these "low-level" comments are replied to (the documentation from NWSHQ says as such), but there are quite a few southern region offices who are replying to virtually every little request/comment/question -- and almost in real time. That's a rough precedent to be setting for everyone else.

[isohume]

I think the offices are supposed to have a policy on this subject. That was my understanding. They are not going to post every watch and warning - they are not going to answer every post - they will use the Facebook page for certain items and updates. Our local office is also using it after events - for people to send in damage reports - areas that might need to be surveyed - other.

Facebook is a learning process - for everyone. I was surprised to learn how many people are using it for information when their satellite dishes go out because of heavy storms - they can't tune in to the local media. They are using their cell phones to follow severe weather threads. Took me a couple of events to figure this out.

I laughed about Facebook a few years ago but now I realize just how big the media has become. Quite amazing.

Yeah I highly doubt every watch and warning would be posted...that'd be crazy. I think FB will be best served as a general informational portal. Just have the occaisional posts about spotter talks, tours, the latest technology. Stuff like that.

I don't think there's a large-scale expectation that all of these "low-level" comments are replied to (the documentation from NWSHQ says as such), but there are quite a few southern region offices who are replying to virtually every little request/comment/question -- and almost in real time. That's a rough precedent to be setting for everyone else.

I could see individual MICs setting an overzealous and casual office FB policy. All it would take is one office to look good in region's eyes and all of a sudden we're all doing the same thing. I just don't think we have the tme for an interactive FB page...a static one would be alright tho. We'll see. For some reason all this social networking is very sexy to region and hq. Hopfully it's just a popular fad, being so new and all.

[isohume]

From Dr. Doswell - one of his articles

Snip...

It’s my opinion that a forecaster should confine the content in a weather warning to the weather and not engage in telling warning recipients what to do. Those recipients find themselves in various kinds of situations and whatever a forecaster might have to offer to some segment of the recipients will be irrelevant or even possibly harmful to other segments. For private sector forecasters who have clients with very specifically-defined situations, of course, a warning could be very precise about including information about what its recipients should be doing. But a private sector weather broadcaster addressing warnings to "the public" (very close to the same audience as the public sector forecasters address, since the NWS does very little to disseminate its warnings, except to the very limited audience which happens to have "NOAA" weather radios) has the same problem as a public sector forecaster. Weather forecasters should forecast the weather, not make forecasts of societal impacts or tell their users what to do!

Lol...Doswell.

[AtticaFanatica]

Lol...Doswell.

Come again?

[isohume]

Come again?

Lol...Doswell.

[AtticaFanatica]

Lol...Doswell.

hmmmm

[AtticaFanatica]

I realize I'm about 24 hours late to the dual-Pol stuff, but I just wanted to add that I agree it doesn't do much of anything in increasing warning lead times. The debris signatures would be in addition to other already-existing signatures that should almost always be sufficient. There are dual-Pol signatures in tornadoes without debris as well, but you'd need great spatial resolution and those would probably be in addition to a tornado weak-echo hole.

[phil882]

I think the big debate we are having right now underlines the fact that we need more advanced radar technology in order to usher in tornado forecasting into the 21st century.

I know phase-array technology keeps getting pushed back farther and farther in time for public implementation, but it has the potential to make QLCS tornadoes much easier to forecast due to higher spacial and temporal resolution. Just look at some of the images produced by DOW's, you can easily identify tornadoes and other phenomenon even within squall lines. An upgrade in technology might even be able to help reduce the size of the warning polygon so that less people are warned necessarily. I think the NWS does the best job they can with the current technology available, but we simply need better resolution to make these forecast more accurate.

I think the main problem right now is that even in cases where storms verify a tornado warning, the tornado is such a small entity compared to the size of the polygon used to forecast it. Hence, even in a tornado warning verification, many people are often not affected at all, and thus will likely become less likely to heed the next warning were it to occur. However, you can't reduce the size of the polygon with today's current radar technology because we simply don't have the resolution to know specifically within the polygon where the tornado will develop. If you get better tools, you can pinpoint the location, and thus the need for such a large polygon disappears.

The debate about what the NWS should and should not warn is stupid and the focus should really be on making better radar technology more available so that we can improve these forecasts further. I think we are reaching the pinnacle of what the WSR-88D can give us in lead time and accuracy for tornadogenesis.

[AtticaFanatica]

I think the big debate we are having right now underlines the fact that we need more advanced radar technology in order to usher in tornado forecasting into the 21st century.

I know phase-array technology keeps getting pushed back farther and farther in time for public implementation, but it has the potential to make QLCS tornadoes much easier to forecast due to higher spacial and temporal resolution. Just look at some of the images produced by DOW's, you can easily identify tornadoes and other phenomenon even within squall lines. An upgrade in technology might even be able to help reduce the size of the warning polygon so that less people are warned necessarily. I think the NWS does the best job they can with the current technology available, but we simply need better resolution to make these forecast more accurate.

I think the main problem right now is that even in cases where storms verify a tornado warning, the tornado is such a small entity compared to the size of the polygon used to forecast it. Hence, even in a tornado warning verification, many people are often not affected at all, and thus will likely become less likely to heed the next warning were it to occur. However, you can't reduce the size of the polygon with today's current radar technology because we simply don't have the resolution to know specifically within the polygon where the tornado will develop. If you get better tools, you can pinpoint the location, and thus the need for such a large polygon disappears.

The debate about what the NWS should and should not warn is stupid and the focus should really be on making better radar technology more available so that we can improve these forecasts further. I think we are reaching the pinnacle of what the WSR-88D can give us in lead time and accuracy for tornadogenesis.

The prototype PAR has worse spatial resolution than the current WSR-88D, and depending on the angle to the antenna, the azimuthal resolution is much worse. There's also ongoing difficulties in melding dual-Pol with phased array.

[radarman]

I realize I'm about 24 hours late to the dual-Pol stuff, but I just wanted to add that I agree it doesn't do much of anything in increasing warning lead times. The debris signatures would be in addition to other already-existing signatures that should almost always be sufficient. There are dual-Pol signatures in tornadoes without debris as well, but you'd need great spatial resolution and those would probably be in addition to a tornado weak-echo hole.

I agree with you about warning times and dual pol.

Reading the testimonies of forecasters the last few weeks it's clear everyone is very sensitive to FAR, and the difficulty in obtaining ground truth factors into that.

When it's a constant battle of "warning fatigue" vs. "late warning for Springfield" you've probably got to grade out the success/failure ratio, but it's not very clear how to improve that scoring. I'm wondering if the dual-pol sigs can maybe be proxies for ground truth for the weak spinups that often go unsurveyed. Of course, a funnel cloud probably deserves a TOR even if it never touches down and therefore has no debris sig. It's a very difficult problem to address...

[phil882]

The prototype PAR has worse spatial resolution than the current WSR-88D, and depending on the angle to the antenna, the azimuthal resolution is much worse. There's also ongoing difficulties in melding dual-Pol with phased array.

Well there goes my hope... although my basic premise to applies. We need better technology to usher in the next generation of tornado detection, as we have reached pretty much the extent of our forecasting ability using the WSR-88D.

[tornadotony]

I agree with you about warning times and dual pol.

Reading the testimonies of forecasters the last few weeks it's clear everyone is very sensitive to FAR, and the difficulty in obtaining ground truth factors into that.

When it's a constant battle of "warning fatigue" vs. "late warning for Springfield" you've probably got to grade out the success/failure ratio, but it's not very clear how to improve that scoring. I'm wondering if the dual-pol sigs can maybe be proxies for ground truth for the weak spinups that often go unsurveyed. Of course, a funnel cloud probably deserves a TOR even if it never touches down and therefore has no debris sig. It's a very difficult problem to address...

I very much agree with this. I didn't mean to imply that dual-pol will help improve lead time. Obviously, if a dual-pol radar is detecting debris and there's no tornado warning, then it's a bit late to be joining the ballgame. My implication is exactly what you stated, that dual-pol could help serve as a proxy for ground truth.

Also, since it was brought up, does lead time really need to increase, at least on average? Yeah, we have instances, such as Springfield, where lead time definitely needs to/needed to be better. But how much is too much? We've seen in the past few weeks TORs that go out for 60-70 mins. If you're at the end of that TOR, are you going to feel immediately threatened for 70 straight minutes? The NWS criteria is that warnings should only last 15-45 minutes, and I think cracking down on that might be a good, albeit small, first step in addressing the false alarm perception problem.

[The Great Zo]

I'm not sure that a radar with 5-meter/hundredth-of-a-degree resolution and 10-second timesteps would help that much with spin-ups. There have been fortunate occurrences in which they have occurred very close to a TDWR, allowing for as good of sampling as is possible right now -- and they still show up from absolutely nowhere in a matter of a couple minutes. As for the processes that are going on in the 5-10 minutes before the spin-up develops? That's probably over my head in science, but I would wager it's rather chaotic and not something one could really look for on radar.

I'm wondering if the dual-pol sigs can maybe be proxies for ground truth for the weak spinups that often go unsurveyed. Of course, a funnel cloud probably deserves a TOR even if it never touches down and therefore has no debris sig. It's a very difficult problem to address...

By their nature, debris signatures require debris to be lofted to the level of the beam (refraction issues aside). Unless you're really, really close to a radar, weak/brief spin-ups are unlikely to do that.

[The Great Zo]

We've seen in the past few weeks TORs that go out for 60-70 mins. If you're at the end of that TOR, are you going to feel immediately threatened for 70 straight minutes? The NWS criteria is that warnings should only last 15-45 minutes, and I think cracking down on that might be a good, albeit small, first step in addressing the false alarm perception problem.

Indeed. At one time, I heard of a study that actually described an "optimum" lead time -- at which point, any longer lead times led to a diminishing return on people taking appropriate action. I can't recall where I heard this, so if anyone else knows where this study could be found, post it up!

Edit: This might be it: http://www.caps.ou.edu/reu/reu09/papers/Hoekstra.pdf

[AtticaFanatica]

I'm wondering if the dual-pol sigs can maybe be proxies for ground truth for the weak spinups that often go unsurveyed.

Maybe if there were a national network of CASA radars

[radarman]

I'm not sure that a radar with 5-meter/hundredth-of-a-degree resolution and 10-second timesteps would help that much with spin-ups. There have been fortunate occurrences in which they have occurred very close to a TDWR, allowing for as good of sampling as is possible right now -- and they still show up from absolutely nowhere in a matter of a couple minutes. As for the processes that are going on in the 5-10 minutes before the spin-up develops? That's probably over my head in science, but I would wager it's rather chaotic and not something one could really look for on radar.

By their nature, debris signatures require debris to be lofted to the level of the beam (refraction issues aside). Unless you're really, really close to a radar, weak/brief spin-ups are unlikely to do that.

Well you're right about the earth curvature problem of course.

Disagree a little about this though: "I'm not sure that a radar with 5-meter/hundredth-of-a-degree resolution and 10-second timesteps would help that much with spin-ups."

The images on page 3 of this thread show one of those unwarned spinups that caused a lot of damage and the specs of that radar are nowhere close to what you've mentioned.

It would be great to see if KOUN was running that day. (apr 2, 2010) I wonder if it saw anything in the dual pol?

[The Great Zo]

The images on page 3 of this thread show one of those unwarned spinups that caused a lot of damage and the specs of that radar are nowhere close to what you've mentioned.

Oh I know. I was exaggerating, of course. I guess my question is less one of "can we see the spin-up as it's occurring" (because, depending on the storm's location relative to a radar, we know we can). I'd ask what that same radar was showing five minutes (even ten minutes) before touchdown -- which would be the key to actually getting a meaningful tornado warning out in time for people to take action.

[radarman]

Oh I know. I was exaggerating, of course. I guess my question is less one of "can we see the spin-up as it's occurring" (because, depending on the storm's location relative to a radar, we know we can). I'd ask what that same radar was showing five minutes (even ten minutes) before touchdown -- which would be the key to actually getting a meaningful tornado warning out in time for people to take action.

Well, I mentioned that I am a little skeptical about much additional warning time time for these as well... At least in the short term.

An OU grad student wrote up a case-study for that Apr. 2 case that's worth checking out. (no mention of KOUN though, which might indicate it wasn't running/available)

http://ams.confex.com/ams/pdfpapers/176200.pdf

[Memphis Weather]

I very much agree with this. I didn't mean to imply that dual-pol will help improve lead time. Obviously, if a dual-pol radar is detecting debris and there's no tornado warning, then it's a bit late to be joining the ballgame. My implication is exactly what you stated, that dual-pol could help serve as a proxy for ground truth.

Also, since it was brought up, does lead time really need to increase, at least on average? Yeah, we have instances, such as Springfield, where lead time definitely needs to/needed to be better. But how much is too much? We've seen in the past few weeks TORs that go out for 60-70 mins. If you're at the end of that TOR, are you going to feel immediately threatened for 70 straight minutes? The NWS criteria is that warnings should only last 15-45 minutes, and I think cracking down on that might be a good, albeit small, first step in addressing the false alarm perception problem.

Yeah...besides my feelings on the QLCS issue as I've outlined in previous posts...this is probably my other biggest issue with the warning process. The NWS directives still state that warnings (SVR or TOR) should not exceed 45 minutes...but so many times recently we've seen situations where warnings are going out for 60 minutes...I've even seen a few 90 minute warnings before. At some point...increasing lead time just doesn't do any more good.

After the April 2, 2006 tornado outbreak in the MEG CWA...they went out and found out that several people did not take appropriate shelter because...although they heard the warning...after 30 minutes or so with nothing happening (lead times were up to one hour in some cases)...people assumed there was either no threat or that the threat had passed...and that's when the tornado struck. Generally...MEG keeps TORs going at 45 minutes since...though I've seen a few 60 minute warnings here before still.

Personally...I would like to see warnings no longer than 30 minutes...and keep 45 minute warnings for extreme cases or something like a very fast moving storm. Though this is not related to the tornado issue...here in the South...every summer we see SVRs issued (almost daily) for pulse storm/wet microburst occurrences...and these warnings are usually 45 to 60 minutes in length...when the threat for any severe weather probably doesn't last for more than 15 or 30 at most before they begin to collapse. You'll usually see them cancelled before expiration...but still doesn't make sense to issue a long-term warning for something that will almost certainly be a short-term threat.

[wxlady]

It had never occurred to me (until recently) that a lot of people do not know what severe means.

On social media I discovered quite a few people commenting on WPSD threads saying that meteorologists use the word severe way too often. It dawned on me that the posters (public) did not know what severe meant or that it had a specific definition. I had become so used to hearing it myself (and knowing what it meant) that I had forgotten the majority of the public may have no idea. To them the word severe means just that - SEVERE. When in actuality most severe thunderstorms are not that impressive. who cares about 58 mph winds? We have that quite often here just from gradient winds. To most people a wind gust of 58 mph is not severe. A wind gust of 70 mph or 80 mph is severe.

I think a power company that experienced nearly 60k power outages from downed trees in their service territory cares. Just had an event where there were about 12 wind gusts of 40+ mph, about 5 of those were 45-50 mph, and only 2 gusts qualified as officially "Severe" based on measured wind gust. No gusts measured greater than 70 mph.

[The Great Zo]

I agree.

People take shelter - they stay in shelter for a certain amount of time. If nothing happens then they come out of shelter.

This has been a big topic of discussion at some of the conferences over the last few years.

Right -- and as you mention, this isn't a brand new concept. As far as I know, it has yet to make its way into the NWS operationally, where the official statistics encourage lead times to be as long as possible.

[Fred Gossage]

I agree.

People take shelter - they stay in shelter for a certain amount of time. If nothing happens then they come out of shelter.

This has been a big topic of discussion at some of the conferences over the last few years.

I would be curious to know how the public perceives all of the severe thunderstorm warnings for cells like you mentioned - isolated - pulse up storms that drop a little bit of hail and/or gusty winds. Hundreds of severe thunderstorm warnings are issued each year for these type of storms. Does this cheapen the warnings? Would people take severe thunderstorm warnings more seriously if they were only used for the higher end events? They have already increased hail size (thankfully). I wish they would do the same with the winds.

Of course now they have significant weather advisories. I find a lot of people don't understand what those mean. I guess because they are new.

I wonder if it would help that.... if NWS went back to the term they were using to test them back in the early 2000s (the Southern Region offices were anyway)... "Thunderstorm Advisory". It instantly sounds more benign than a Severe Thunderstorm Warning... but if a storm is getting an advisory when some of the others around it aren't, people might get the idea that it's a little frisky. Or maybe... like the terminology that the OKC stations and FOX 6 use... "Heavy T-Storm Advisory" and "Strong T-Storm Alert".

[Memphis Weather]

I agree.

People take shelter - they stay in shelter for a certain amount of time. If nothing happens then they come out of shelter.

This has been a big topic of discussion at some of the conferences over the last few years.

I would be curious to know how the public perceives all of the severe thunderstorm warnings for cells like you mentioned - isolated - pulse up storms that drop a little bit of hail and/or gusty winds. Hundreds of severe thunderstorm warnings are issued each year for these type of storms. Does this cheapen the warnings? Would people take severe thunderstorm warnings more seriously if they were only used for the higher end events? They have already increased hail size (thankfully). I wish they would do the same with the winds.

Of course now they have significant weather advisories. I find a lot of people don't understand what those mean. I guess because they are new.

Given that so many people don't even take tornado warnings seriously...severe thunderstorm warnings are probably hardly taken seriously at all. Especially around here...you can easily get dozens of them every year in a specific area...and the vast majority will be at the low-end of the criteria. I've gotten myself to where I just tend to ignore them unless its a higher-end situation but I also am obviously pretty weather aware and will usually be pretty situationally aware to the goings-on of the moment...where most people won't. I don't know whether the criteria should be raised further...I don't mind the criteria as it is now since the 1" limit was employed...its just the way it is. I know I've heard some talk of doing something like the "Tornado Emergency" for Severe Thunderstorms...but that also brings about a lot of questions.

I would say 99% of the general public doesn't even know about significant weather advisories. MEG has been issuing them since 2002...but no TV station here crawls them...they only get a mention if perhaps a TV station is already on the air doing news or severe weather coverage and even then its not a guarantee. Plus...its even more inconsistent across WFOs how they're used...sometimes even within a WFO. Some cases you can get dozens of them in a day...another day that seems similar...you may only get a handful. And in big outbreak situations...they're usually completely forgotten about. Its a good concept product...I like the idea...but I'm not sure if its particularly effectively used yet...at least for the general public's purposes.

I know around here when you see the crawls on TV...for a Severe Thunderstorm Warning...they almost always say "T'storm Warning"...never with "Severe"...which I think makes the general public even more confused. I'm sure some people just think they're being told about a thunderstorm that may have some lightning or heavy rain...and nothing else. And of course many people think a thunderstorm with a lot of lightning is severe...or should be severe...when that's not a criteria at all.

[Baroclinicity]

The Wilmington, Ohio office seems to have discontinued short term forecasts in favor of special weather statements (same principle as the sig wx advisory). These are issued when thunderstorms are near severe intensity or are expected to reach it in the near term (also in other weather situations where the relevant advisory/warning isn't necessary). All four TV stations with news/weather departments here crawl watches and warnings with an map depiction in the corner of the screen, at least two of them crawl the entire warning text. They have also switched from the old county based depiction to the polygon based system. None of them display the special weather statement text, as far as I know (but at least one station does update the warning text with severe weather statements).

Occasionally I hear "severe" dropped in favor of "thunderstorm warning", usually by non weather people in the local media. "Watch" and "Warning" is sometimes used interchangeably by the same people. Several weeks ago a regular tornado watch somehow become a "severe tornado watch" on the radio.

[on_wx]

Some Environment Canada offices do not issue tornado warnings on lines of severe thunderstorms capable of brief rain-wrapped tornadoes.

They put up a severe thunderstorm warning and in the warning text it states that 'isolated brief tornadoes are possible in this line of thunderstorms'. Sometimes even when a brief and weak tornado is spotted, they'll issue or update the severe thunderstorm warning to include that a weak tornado was reported there is the possibility another one will form so be on alert.

They tend to reserve tornado warnings for dire situations.But, I guess it depends on the forecaster on shift. Sometimes they'll issue a tornado warning for a landspout or cold core funnel, other times they won't.

[OHweather]

Some Environment Canada offices do not issue tornado warnings on lines of severe thunderstorms capable of brief rain-wrapped tornadoes.

They put up a severe thunderstorm warning and in the warning text it states that 'isolated brief tornadoes are possible in this line of thunderstorms'. Sometimes even when a brief and weak tornado is spotted, they'll issue or update the severe thunderstorm warning to include that a weak tornado was reported there is the possibility another one will form so be on alert.

They tend to reserve tornado warnings for dire situations.But, I guess it depends on the forecaster on shift. Sometimes they'll issue a tornado warning for a landspout or cold core funnel, other times they won't.

The one thing I seem to notice about Environment Canada is that there are not many standards and a lot of inconsistencies between offices and even between forecasters in the same office, much more so than the NWS. Must make it frustrating.

[JoMo]

Just an FYI, Spann posted a 2nd part:

http://www.alabamawx.com/?p=48880

[isohume]

Just an FYI, Spann posted a 2nd part:

http://www.alabamawx.com/?p=48880

I don't believe Spann is barking up the right tree. He believes that by discontinuing the blanket TORs issued for "low end" spin-ups associated with squall lines...that the FAR will come down. The FAR is actually helped by these types of warnings because somewhere in that polygon there will most likely be a tornado. So, no false alarm. Spann also fails to recognize the tornado FAR taken on a national average is still very high which includes the majority of offices that dont issue those blanket type warnings.

What Spann needs to focus on and begin dialog on is...improving the science of tornadogenesis. There is still so much we don't understand about how and why tornados develop with one storm and do not develop with very similar storms. Once we have a better undertanding of these small-scale processes...the FAR for tornado warnings will improve.