Did SPC F up ?

[andyhb]

duhh, but it's how it gets relayed to the public which is more important. The Media only catches on whenever Mod or High risks are issued.

18 reported tornadoes in a day within a 200 mile radius is more than a Slight Risk.

In a word, Stop.

[Hoosier]

I think there are times where it's good to ask what went wrong or what was missed (to help with recognizing similar setups in the future), but imo yesterday's case doesn't really fit. The threat was identified and handled relatively well.

Consider the surprise tornado outbreak of April 20, 2004. The 20z outlook, which was issued only about 2 hours before the outbreak got into full swing, didn't even have a 2% tornado contour in the area that was impacted, there was no mention of severe in the area and the tornado watch wasn't issued until after tornadoes were touching down. There were almost 30 tornadoes and 8 people died. THAT is a case where you ask what was missed.

[andyhb]

1/7/08, 12/31/10 and 5/12/04, granted those events did have 5% tornado risks, but it each was all the way to 20z and each event ended up far exceeding this.

Yesterday's event had a 10% at 1630z and produced less tornadoes than any of these events.

[Srain]

http://www.srh.noaa.gov/fwd/?n=april32012sum

[brettjrob]

The addition of a 10% contour over precisely the area to be affected at 1630z showed impressive skill. In my view, it was a very good forecast, all things considered. In hindsight, would hatching of the tornado probabilities have been warranted over the Metroplex? Probably, but that's a minor detail that goes unnoticed by 99.99% of the public and even the media.

I agree with Hoosier; something like the IL/IN event he referenced would be legitimate cause to ask, "what went wrong?" The question doesn't even seem relevant in this case, at least insofar as SPC is concerned.

[Hoosier]

Re: outlooks and the general public, I agree that there are some issues. I almost never use the slight, moderate, or high wording when I tell people about possible severe weather and certainly not probabilities. I might say "decent" or "heightened" chance on some moderate or high risk days.

[IMADreamer]

duhh, but it's how it gets relayed to the public which is more important. The Media only catches on whenever Mod or High risks are issued.

18 reported tornadoes in a day within a 200 mile radius is more than a Slight Risk.

You need to stop, your ignorance is pouring through the screen right now. By further arguing your point you are only proving your lack of understanding of SPC risk assestments.

[Chumpson]

It's always interesting seeing the different views from all the interested folks after an event like 4/3/12 - everything from a terrible forecast to an awesome forecast! This thread suggests a couple of things to me: 1) SPC forecasts are good enough now that they are expected to always be good, and 2) some people like to voice their opinions regardless of their experience (or lack thereof). Convective outlooks for the entire CONUS are no easy task, and they are composed within the contraints of time deadlines and occasional intra-agency collaboration. Most of the loudest critics have never issued a single public outlook of CONUS scale, with their name on the forecast. It can be a pretty thankless job when you receive muted praise for big events because "it was obvious" to everyone, and loud complaints when you miss one that nobody else happened to mention until after it occurred.

Even when a forecast doesn't work too well, it's information provided directly by SPC that allows many of you to judge the quality of the forecast, from archived products/reports to real-time environmental information. All of this comes from ~4 operational meteorologists at one time, with a handful of dedicated support staff in the background. For the record, I think the 10% upgrade was reasonable at 1630z. You can make a decent argument for MDT risk (maybe 15% SIG) after the fact, but the 10% area and upgrade to TOR watch were pretty good forecasts.

[OceanStWx]

I realize experts, mets and people on here have a standard by which they assess a 10% risk but if these statements are to be useful to the public then they need to be re-worked because the public interprets 10% as meaning "highly unlikely to occur"

Or if people don't understand what they are looking at they can search for more information. And it isn't very hard, seeing as it is linked right below the outlook probability image (http://spc.noaa.gov/products/outlook/probinfo.html).

As many have mentioned already, those outlooks and probabilities aren't necessarily meant for public consumption (just as products like the AFD weren't originally intended that way). However, meteorologists are expected to know what those probabilities mean, and are expected to convey that threat to the public. The link I provided gives a great example for late April/early May. That region of Texas on any given day that week can expect a tornado within 25 miles of a point between 1% and 1.25%. Meaning that a 10% probability in the 1630z outlook was around 10 times greater threat than a normal day. A significant increase.

You may not like it, but this is a scientific way to quite accurately portray a severe weather threat, rather than arbitrarily throwing numbers out there that people who might happen to click their way to the SPC website might understand. I mean what would have an acceptable tornado probability have been for this event? 50%? Equally as likely to occur as not occur? So would you need to be above 50%? Now you're talking equal risk between probability of precipitation for the next 6 hours and a tornado, but which of those is more likely?

[WEATHER53]

The two replys right here above are from people who know a lot more about weather than me.

I do know what 10% means and dont think it gets a different definition just for weather.

[andyhb]

It's always interesting seeing the different views from all the interested folks after an event like 4/3/12 - everything from a terrible forecast to an awesome forecast! This thread suggests a couple of things to me: 1) SPC forecasts are good enough now that they are expected to always be good, and 2) some people like to voice their opinions regardless of their experience (or lack thereof). Convective outlooks for the entire CONUS are no easy task, and they are composed within the contraints of time deadlines and occasional intra-agency collaboration. Most of the loudest critics have never issued a single public outlook of CONUS scale, with their name on the forecast. It can be a pretty thankless job when you receive muted praise for big events because "it was obvious" to everyone, and loud complaints when you miss one that nobody else happened to mention until after it occurred.

Even when a forecast doesn't work too well, it's information provided directly by SPC that allows many of you to judge the quality of the forecast, from archived products/reports to real-time environmental information. All of this comes from ~4 operational meteorologists at one time, with a handful of dedicated support staff in the background. For the record, I think the 10% upgrade was reasonable at 1630z. You can make a decent argument for MDT risk (maybe 15% SIG) after the fact, but the 10% area and upgrade to TOR watch were pretty good forecasts.

Well said, Mr. Thompson.

[OceanStWx]

I think there are times where it's good to ask what went wrong or what was missed (to help with recognizing similar setups in the future), but imo yesterday's case doesn't really fit. The threat was identified and handled relatively well.

Consider the surprise tornado outbreak of April 20, 2004. The 20z outlook, which was issued only about 2 hours before the outbreak got into full swing, didn't even have a 2% tornado contour in the area that was impacted, there was no mention of severe in the area and the tornado watch wasn't issued until after tornadoes were touching down. There were almost 30 tornadoes and 8 people died. THAT is a case where you ask what was missed.

Great example.

If I'm remembering the stories I heard about that day, it was either all or nothing (which does occur from time to time). It was low clouds and drizzle that was forecast to stay entrenched, only late in the day it cleared out and there was ample sunshine and moisture for CAPE, low LCLs, with shear in place, and a differential heating boundary (at least in the DVN portion of that outbreak, though I'm sure similar conditions existed eastward).

Those types of set ups do occur, and a forecast of no severe weather can quickly turn into locally widespread. This wasn't even the case in Dallas as they were outlooked for several days and had a watch in place. The question was whether there would be enough heating to break the cap and produce discrete cells ahead of the line. When it became obvious that this would occur, the outlook increased the tornado threat and the upgrade to a tornado watch happened as Chumpson just said.

That's really the usefulness of the 1630z outlook. It is the outlook that provides a look at where morning convection has left boundaries and where the greatest threat will be prior to diurnally driven initiation.

[weatherwiz]

It's always interesting seeing the different views from all the interested folks after an event like 4/3/12 - everything from a terrible forecast to an awesome forecast! This thread suggests a couple of things to me: 1) SPC forecasts are good enough now that they are expected to always be good, and 2) some people like to voice their opinions regardless of their experience (or lack thereof). Convective outlooks for the entire CONUS are no easy task, and they are composed within the contraints of time deadlines and occasional intra-agency collaboration. Most of the loudest critics have never issued a single public outlook of CONUS scale, with their name on the forecast. It can be a pretty thankless job when you receive muted praise for big events because "it was obvious" to everyone, and loud complaints when you miss one that nobody else happened to mention until after it occurred.

Even when a forecast doesn't work too well, it's information provided directly by SPC that allows many of you to judge the quality of the forecast, from archived products/reports to real-time environmental information. All of this comes from ~4 operational meteorologists at one time, with a handful of dedicated support staff in the background. For the record, I think the 10% upgrade was reasonable at 1630z. You can make a decent argument for MDT risk (maybe 15% SIG) after the fact, but the 10% area and upgrade to TOR watch were pretty good forecasts.

Not only was the forecast fantastic but the discussions in the outlooks prior to the 1630z outlook were fantastic as well with tremendous detail highlighting the setup, the threat for some significant severe, and how there certainly was a threat for tornadoes and mentioning the potential for outflow boundaries being a major player. Even the questions were highlighted...how much solar heating would there exactly be, where would the enhanced low-shear be located, what would the exact storm mode be? During the late morning/afternoon when these questions answered the necessary changes to the forecast were made. That 10% area could not be more perfectly placed.

The OP just really doesn't have understanding of the situation or the SPC products and how they work.

[OceanStWx]

The two replys right here above are from people who know a lot more about weather than me.

I do know what 10% means and dont think it gets a different definition just for weather.

My point is that you are misinformed, because this 10% risk is defined as being within 25 miles of any given point (which is a small probability because tornadoes are rare). It is a small probability, but relative to every day it is at least 10 times greater (a very large risk).

My point is that small probability can equal a large risk and not be wrong or confusing.

This isn't a case of it having a different definition just for weather.

[Hoosier]

Great example.

If I'm remembering the stories I heard about that day, it was either all or nothing (which does occur from time to time). It was low clouds and drizzle that was forecast to stay entrenched, only late in the day it cleared out and there was ample sunshine and moisture for CAPE, low LCLs, with shear in place, and a differential heating boundary (at least in the DVN portion of that outbreak, though I'm sure similar conditions existed eastward).

Those types of set ups do occur, and a forecast of no severe weather can quickly turn into locally widespread. This wasn't even the case in Dallas as they were outlooked for several days and had a watch in place. The question was whether there would be enough heating to break the cap and produce discrete cells ahead of the line. When it became obvious that this would occur, the outlook increased the tornado threat and the upgrade to a tornado watch happened as Chumpson just said.

That's really the usefulness of the 1630z outlook. It is the outlook that provides a look at where morning convection has left boundaries and where the greatest threat will be prior to diurnally driven initiation.

Jon Davies has a nice page on the 4/20/04 event:

http://www.jondavies.net/042004ilin/042004ilin.htm

[somethingfunny]

duhh, but it's how it gets relayed to the public which is more important. The Media only catches on whenever Mod or High risks are issued.

18 reported tornadoes in a day within a 200 mile radius is more than a Slight Risk.

A separate bubble indicating increased tornadic threat might work too.

Lets face it, most of the media was caught off-guard during this event. It seems they need the etch-a-sketch drawings to make it a big deal..

Their own wording hints at an 'unexpected' event.

Maybe 10% tornado risk should be included as a Mod risk from now on. why ? because a lot of TV personalities these days are hired for their looks and not their knowledge. They'll quickly glance at the SPC Convective Outlook and see Slight Risk...and be like "ohh no big deal". "few storms here & there"

Just tellin it how it is..

I was going to say some things about how the hell you would know what the DFW media knows about meteorology and interprets model data, SPC discussions/graphics, and local synoptic scale features from your location in Tobyhanna PA, but instead i'm going to call you an idiot and ask you why you're trolling this particular weather board. Maybe your TV meteorologists are just unqualified pretty faces in Philadelphia, but that **** will get you fired fast in Dallas.

I thought our media did an outstanding job yesterday, as did the Fort Worth WFO and the SPC up in Norman. You know how I know our meteorologists made a great forecast and communicated it effectively to a pre-emptively well educated population? Nobody died yesterday.

I think this should be the real story of this outbreak. Instead of wondering why so many died, like we were forced to do too many times in 2011, here we are with the luxury of wondering why so many survived. The strongest tornado was "just" EF-3 (a photo I snapped of it is my avatar now) but even when you look at some of the EF-2 damage, it's amazing that we had the outcome we had.

[Cory]

I don't think they should even include that slight risk language in the hazardous weather outlook that the public can see in their point and click forecast. I've noticed our local mets, if they mention the SPC outlook, will not say the word "slight" but instead say only "risk". Smart move.

That said, the watch they put out, along with the local media, gave people enough warning. I don't think the SPC messed this up.

I'd rather they have only two risk categories anyway: risk and high risk. Slight/moderate wording tells the general public that things may be iffy or not worth paying attention to. They just don't handle statistical threats or in-between wording very well.

20% chance of rain = "Crap...it's going to rain today".

20% chance of tornadoes = "Only 20%? Nothing to worry about".

[SnowGoose69]

They had it captured in the tornado risk, N Texas is very tough to predict because most of the historic monster days over S KS and OK which hit ICT/TUL/OKC often times do nothing in DFW because it remains too capped and they are too far from the surface low forcing. The days which do have potential for them often have the risk of going linear too quickly because you usually need a strong upper trough or surface low to induce tornado outbreaks the more south you go. It seems to me the worst days in DFW usually occur when the cap breaks farther south on what is forecast to be more of an event for Oklahoma or a linear/MCS scenario is more discrete than expected.

[mike2010]

The OP just really doesn't have understanding of the situation or the SPC products and how they work.

I've been analyzing SPC products before you've even been born.

Several here agree an upgrade to Mod risk should of been done.

And to make everyone happy, i'm droppin it now..

[BeauDodson]

To appreciate how difficult some of these events can be -

I can't even imagine the pressure they are under during these big events. I think they do an excellent job - yes there are days where an event will over/under achieve - thus is the nature of meteorology. But - compared to years ago - there has certainly been a tremendous amount of improvement.

http://www.weatherbrains.com/audio/wb032612.mp3

[isohume]

[somethingfunny]

I see absolutely no reason to have upgraded to a Moderate Risk at 11:30 AM. The Tornado Watch was issued by then and the tornadoes were on the ground by noon. What would this have accomplished? Hell once they're on the ground and tracking into Dallas County SPC should have drawn a big pink High Risk hatched area just over Dallas County by your suggestion. What's the point, once the system's already doing things in real time? Our media already knew about the tornado threat.... the helicopters had fuel in them and pilots on standby to capture the storms before they entered the Metro, did they not?

[mike2010]

[somethingfunny]

No, I really want to know which Dallas area TV meteorologists you believe are easily confused by SPC graphics, not some meme implying you're surprised that people are arguing with a provocative thread you started.

[knitwx]

It's always interesting seeing the different views from all the interested folks after an event like 4/3/12 - everything from a terrible forecast to an awesome forecast! This thread suggests a couple of things to me: 1) SPC forecasts are good enough now that they are expected to always be good, and 2) some people like to voice their opinions regardless of their experience (or lack thereof). Convective outlooks for the entire CONUS are no easy task, and they are composed within the contraints of time deadlines and occasional intra-agency collaboration. Most of the loudest critics have never issued a single public outlook of CONUS scale, with their name on the forecast. It can be a pretty thankless job when you receive muted praise for big events because "it was obvious" to everyone, and loud complaints when you miss one that nobody else happened to mention until after it occurred.

Even when a forecast doesn't work too well, it's information provided directly by SPC that allows many of you to judge the quality of the forecast, from archived products/reports to real-time environmental information. All of this comes from ~4 operational meteorologists at one time, with a handful of dedicated support staff in the background. For the record, I think the 10% upgrade was reasonable at 1630z. You can make a decent argument for MDT risk (maybe 15% SIG) after the fact, but the 10% area and upgrade to TOR watch were pretty good forecasts.

Well spoken.

[weatherwiz]

I've been analyzing SPC products before you've even been born.

Several here agree an upgrade to Mod risk should of been done.

And to make everyone happy, i'm droppin it now..

Well than that's sad you still don't have an understanding of how they work or what they mean.

[arizonasooner]

Something very similar to this (although less severe) happened maybe 30 or so years ago in DFW. I remember a KXAS (maybe WBAP then) newscaster joking to meteorologist Ron Godbey (Harold Taft's #2) about some "pretty nasty looking partly cloudy skies" or soimething to that effect.

Godbey was not amused.

Seems that a very similar mesoscale boundary affected the conditions significantly between morning and noon (or noon and 5 pm, I forget) such that somewhat unexpected severe weather had occurred. And I also seem to remember that there were no injuries or deaths to be reported from the event, so the newscaster probably felt justified in joking about it.

To me, it is more of a testament to the weather-awareness of folks in the metroplex. Stuff happens and people pay attention when they need to!

[David Reimer]

Considering I was chasing and well in the loop about the behind the scenes events on Tuesday, let me go ahead and brief you on what exactly came into play.

Just after 11 AM CDT, at least 30 minutes before the 1630Z outlook came out with the 10% tornado probability, FWD mentioned an increased tornado risk that afternoon due to an outflow boundary and very high 3 KM CAPE values in NWS Chat. All local media members are in the chat and use it extensively. There was no "surprise" that the tornado potential was going to increase into the early afternoon. Local media members began disseminating that information and began relaying that a Tornado Watch would likely be issued once the mesoscale discussion was put out by SPC. While North Texas has been relatively quiet in terms of major tornadoes during the past few years, the D/FW market has really taken hold of social media. The National Weather Service office in Fort Worth was the first office to have a facebook page and they have close to 15,000 likes. In addition, I put out the word on the Texas Storm Chasers social media platforms (which have over 100,000 followers) that the tornado threat was more elevated then earlier expected and that everyone should have a way to receive weather warnings in the early afternoon hours. The Storm Prediction Center recognized the potential and increased the tornado probabilities and issued a mesoscale discussion highlighting that potential and the likelihood of an upgrade to a tornado watch. We were already experiencing severe weather with the elevated supercells in Hood/Parker/Wise county beginning to become surface based, so all media outlets had their staffing increased. The National Weather Service was no exception and ample man power in place to support warning operations. I headed down to Burleson to intercept the developing supercell that was near Cleburne. While I honestly had not had time to do an in-depth analysis of conditions that morning, I knew that there was a boundary in place over the metroplex and 3 KM CAPE values were in excess of 350 joules per kilogram. That in itself supported a threat for brief spinup tornadoes.

I gained a visual of the supercell's base about the time the first report of a tornado came in near Joshua, TX. I had a visual of a wall cloud with some minor rotation, but I was still a good distance away. As the storm began moving closer to my location, it underwent a cycle and looked like it was weakening for about two minutes. I thought the storm was a candidate for a brief tornado at that point. After about two minutes, something occurred which changed my entire perception of the threat. Suddenly and without warning, scud rapidly began rising into the base of the storm. Now mind you, this wasn't my first chase and I've seen plenty of tornadoes. The overall velocity of the rising scud was very fast and wasn't the kind of motion you see on a storm capable of producing "marginal" tornadoes. At that point, I forwarded that information onto FWD, in addition to streaming the entire event with audio. At that same time, a second report came in of a tornado on the ground near a high school in Joshua. I did notice a darker area of contrast to my northwest, but could not confirm visually that a tornado was down. At this point the storm was beginning to move away from my location and began moving north on Interstate 35W. A rotating wall cloud crossed the highway over my location and looked like it was lowering and very well have been in the funnel stage as it crossed just to my north, I'll have to sit down and review the dash-cam video to confirm. Once the circulation passed east of the highway, I moved north to Interstate 20 and blasted east when I began hearing multiple reports of a large tornado on the ground. I ended up just west of Highway 820/US-287 North when I pulled over since I was getting slammed by golfball size hail and was beginning to see a feature to my south. I knew there was a tornado on the ground and that it would pass fairly close to my location, but I was unsure if it was maintaining a more northward movement or had in-fact turned more to the right. I got a frantic call from another chaser about three miles to my house that the tornado was moving at my location and that I needed to go south on Highway 287 to get out of the way, fast. I did just that and had a fairly violent multiple-vortex tornado form about 100 yards to my west as I blasted (by blasted, I mean BLASTED) south on 287. It was in an impressive multiple-vortex stage and was ripping trees of the ground. I got a mile south of the circulation, jumped out of the car and looked north just as the tornado became a large multiple-vortex tornado on Highway 287. This was around the same time that the large tornado was moving through Lancaster, which mind you is pretty close to my house. The rest of the storm's tale is history...

I'll be the first to say I did not expect the significant outbreak we received on Tuesday. I expected a fairly widespread event of very large hail, damaging winds, and maybe a brief tornado. My awareness level went up with that 10% TOR, but I still only called for a "few tornadoes" on the blog. What happened on Tuesday was the perfect storm of boundaries, relative atmospheric conditions, and timing. Do I think anyone "busted" or messed up? Absolutely not! The outflow boundary was a mesoscale feature that could not be resolved by model guidance, even high resolution models. The folks at FWD noticed this boundary at 11 AM and as a result beefed up their awareness level for possible tornadoes. SPC upgraded the severe watch to a tornado watch nearly an hour before the first tornado touched down in Johnson county. All media outlets were already aware of the severe weather potential and had helicopters ready to go, which for the first time, proved to be a major asset with a tornado event in North Texas. As a result of live video and reports, the National Weather Service in Fort Worth was able to confidently increase their warning urgency. ALL local media stations (Television, Radio, and Internet) were on continuous tornado coverage and social media played a major role in providing warning data and relaying reports.

In terms of Texas Storm Chasers, we have multiple individuals that are available to post information on our social media feeds during events. While my business partner Connor and myself were out chasing and not able to post frequent updates, Chelsea Bennett, Jenny Brown, John Burgdorf, and Ryan Sheff (a member of this forum under SmokeEater) were posting constant updates that were shared with tens of thousands of individuals. The reason we do social media is for events like Tuesday, because many people receives updates on their phones and computers. Social media provides an outlet to receive this information almost instantly. While I was viewing a tornado in Kennedale, Texas, those guys were posting the information about where it would be in five minutes.

To those who say that forecasters messed up big time on Tuesday in regards to downplaying (or under-forecasting) the tornado potential, I say go ahead and take a look at model data from the night before and even six hours before the evening. North Texas has to be one of the most prepared regions in the country for a major event. I'm not talking about just Fort Worth and Dallas, but the region as a whole. When we have a major disaster like a tornado outbreak, emergency crews don't operate as a city, but as a region. I arrived in Lancaster, Texas about 90 minutes after the tornado struck and there were emergency crews from cities over 25 miles away. It reminded me of the wildfire events last year. There is a reason no one died on Tuesday. Sure, some luck plays into it, but with so many populated areas being hit by strong tornadoes within such a concentrated zone, it's amazing that no one was killed. Weather education plays a big role here in North Texas. Complacency due to the lack of tornadoes over the past ten years may have been an issue before in D/FW, that complacency was shot to hell on Tuesday. Anyone who says the weather service did a bad job obviously doesn't know what they're talking about because FWD performed amazingly. Warning lead-times, warning urgency, severe weather reports, and the overall use of social media during and after the event shows why FWD is one of the best WFOs in the United States.

Sorry for the late response, but I flew up early yesterday to attend my grandmother's funeral and have been a little busy. With that in mind, I'll try to keep this as nice as I can and direct it to one specific individual in this thread, who I won't name for the sake of being "nice." Those who are trying to judge an event that are what I would consider more or less educated in meteorology as the general public, I have this to say to you. Learning more about weather is great and I promote that 100 percent on my website. Heck, I write all my discussions in a way the general public can understand and learn. However, there is a fine line between asking a question and trying to blame someone or something when you don't know what you're talking about. You know nothing about how complex forecasting mesoscale features in a severe weather setup can end up being. You know nothing about the differences in the probabilities in the convective outlook. You know nothing about how the Fort Worth NWS and their local relationships with the media and local spotters. Don't you dare pretend to understand. Suggesting that something be changed to help improve the warning response in the future is one thing, trying to be a back-seat blowhard and act like a fool is quite another.

[The Dude]

I don't think they should even include that slight risk language in the hazardous weather outlook that the public can see in their point and click forecast. I've noticed our local mets, if they mention the SPC outlook, will not say the word "slight" but instead say only "risk". Smart move.

That said, the watch they put out, along with the local media, gave people enough warning. I don't think the SPC messed this up.

I'd rather they have only two risk categories anyway: risk and high risk. Slight/moderate wording tells the general public that things may be iffy or not worth paying attention to. They just don't handle statistical threats or in-between wording very well.

20% chance of rain = "Crap...it's going to rain today".

20% chance of tornadoes = "Only 20%? Nothing to worry about".

The slight risk terminology drives me crazy. If you tell someone that there is a slight risk that something may happen, they are programmed to understand that as meaning there's a low probability of that thing occurring, and they won't worry about it.

I understand the SPC uses slight risk to differentiate from days when there is no risk, but I doubt the general public picks up on that nuance. I prefer going with something like:

Elevated Risk

High Risk

Extreme Risk

[weatherwiz]

That was a great account! Thanks for sharing.