2024-05-25 EVENT: OK/KS/TX/MO/AR

[Brian McKibben]

Going to start a thread for Saturday.

Euro & GFS showing neg tilt 500 mb trough across the southern plains are 00z. NAM also has speed max over southern plains but not neg tilt.

There looks to be a well time 700 mb shortwave move through Oklahoma in the afternoon. This coupled with >70 dewpoints should result is a favorable environment for supercells. LLJ gets cranking in the evening and we see some pretty decent hodographs per the 12z NAM (this morning 5/22).

 

[Jeff Duda]

Yep. Pretty much all medium range models aligning on a lot to like about this setup. I'm seeing no S-shape/VB in the forecast wind profiles, and it looks likely that there will be 70s dews ahead of the dryline throughout most of OK. The Canadian model was the one outlier so far keeping the 70-F isodrosotherm south of I40. I am seeing the signature of something that is becoming increasingly common in my experience regarding a northbound moisture surge that occurs late in the afternoon/early evening immediately against the dryline that fills in the warm sector that is otherwise a little "dry" east of the dryline earlier in the day.

Here's a depiction of what I mean:


Note that moisture drops in SW OK from 18-21Z but then resurges by 00Z. I've seen this happen before, both in the presence of and without the presence of existing convection. It's almost like a localized dry push above the PBL moves over the DL during the day, and the vertical mixing causes a temporary dropout in surface moisture quality, only for it to recover once the dry push moves on past. I'm not seeing as strong an indication in the GFS as in the ECMWF (depicted above), but it leads me to believe such a potential X factor could be in play.

How does this impact events? Not sure - it probably goes both ways. It could cause storms not to fire along the dryline in that portion of reduced moisture if capping is strong enough (the GFS shows a warm nose over the PBL this day). But it could also reduce storm competition and favor better spaced and discrete convection. There's probably also other impacts I'm not currently thinking of.

Anyway, we have two medium range CAM ensembles being run by NCAR until the end of the month, and at least from a terminal product point of view, they have somewhat differing views of this event:
C-SHiELD (FV3-based):

This one tends to omit most of OK from the fun and instead gives KS/MO a better show.

The MPAS-based ensemble is somewhat more favorable for the southern Plains, but there's still a lot of spatial spread:



Althouhh when you combine the ensemble information together, neither ensemble seems to think particularly highly of the setup, especially OK:


But given the forecast range, this may not be all that unexpected. Things will come into better focus over the next day or two.

CSU's ML outlook product appears to show a small moderate at this point:


We shall see. It didn't do half bad at the day 5 range with yesterday's event.

 

[Mike Smith]

I had some time to take a look at this and -- if forced to make a forecast for Saturday at this moment -- I'd probably be going (SPC terminology) moderate risk with at least a 10% hatched area in parts of KS and OK with giant hail to accompany the storms.

Of course, I greatly over-forecast May 6, so take this with a grain of salt.

 

[Jason N]

I've seen a lot of consistency in the longer-range models for this day over KS/OK and the mid/upper dynamics have improved over the past 3 days, but I have also noticed some models postponing convection til after sunset, but that's been inconsistent. Hopefully the CAMS will close the gap on timing, because this could be a pretty sizeable event, which I am still surprised but will guess the SPC will upgrade at least to 3 out of 5 soon, and possibly 4 out of 5 if some of those smaller scale items get resolved with the right timing.

 

[Brian McKibben]

new 12z data (5/23) is arriving for the NAM. I am curious if storms get going too early? 700mb height falls begin spreading over the warm sector around 18z. Plenty of cape in place at that time and winds are already good enough for supercells with possibility of tornadoes. By 00z LLJ really ramps up and we get some pretty extreme forecasted soundings.

The first look at the 4km NAM shows convection holding off until 22-00z timeframe along the DL.

I am curious if we will see some open warm sector storms. It doesn't look like there will be too much of a CAP in place but 3 days out we really don't have those details yet.

So far I am not seeing too much different from previous comments.

WIth this being a potentially high end setup, I am curious what everyone else is looking for in terms of possible failure modes? I know it could be a big day, but what are the models showing us that could potentially limit an outbreak? Any flies is the ointment?

 

[Jason N]

This morning's 12Z mid-range NAM/ENS guidance keep creeping upwards in terms of a potentially significant setup. I think the one thing I keep seeing some continued wavering on is the CI timing and WHEN/IF capping breaks. seems like run to run(00/12Z) there has been some wavering on CI timing, and I noticed some runs (12Z today) where even after 03Z there still wasn't any convection over OK but yes in KS closer to the SFC low. So, for me some challenges and thoughts to current runs.

Besides inherent model bias issues:
- 500/700mb shortwave placement/timing / alignment to dryline kick timing
- capping? (Big win or Big Bust scenario??)
- trafficability south of OKC to red river (recent heavy precip), main/secondary roads fine, tertiary dirtroad problems in some areas.

- Thermodynamics / Shear looks impressive 4K+/ 1km 250-300ms2+ north of the Red to KC border
- Hodo Shapes from red river up to KS border look photo genic, a little higher LCL in SRN KS, maybe a little less Scud.

- SRN KS looks somewhat better SE of Dodge City to the OK border to me in terms of being closer to the SFC low with slightly more backed flow/convergence, and a "little" higher LCL for potentially better viewing, but I think for me, since I will be out there, I will position myself around Enid to Kiowa and play on the Nose of the DL push and hope there is a little better backed flow there. Also, I'd prefer to stay out of the OKC area, plus this position makes a north or south shot easier if needed.

 

[Jason N]

to illustrate the difference between 00 and 12Z NAM, both valid 03Z, while the ST's first stab at the 00Z point shows CI but further west on DL and closer to the ejecting SFC low. OfCourse, it's still early and CAM solutions are coming. The Euro on the other hand is convective bullish Valid 00Z 26th.

 

[Patrick K]

WIth this being a potentially high end setup, I am curious what everyone else is looking for in terms of possible failure modes?

To paraphrase Jeff Goldblum from Jurassic Park: "Busts, uh, find a way."

In all seriousness, we just had a bust-like high risk setup on May 6th where the forecast was simply wrong in a lot of respects. That's always possible and how are you even going to know until the real-time observations confirm? IMO this setup certainly has some possibility of BSB, at least in one's chosen target area. Some of the CAMs aren't initiating anything over wide areas, showing lingering cloud cover and decent CIN still in place at 00Z. That said, I don't know how much I believe the CAMs in this situation with sparse dryline initiation. SPC clearly believes there will be storms, but where? Choosing the wrong target area would be a failure mode, wouldn't it?

 

[Jason N]

Times like these Big win/busts illicit 2 results, stick with your guns and go for broke.. or speed/curse relocate. with CI timing being so critical, if its past 03Z it makes moving further than an hour away useless for those trying to catch magic hour and not chase at night. I think the point about SPC clearly believing is true. GFS/EURO have all trended more on "something happening" over not, but solutions have been mixed. The CAMS are in the early stages still. So, for now, planning to hold position when I get there around the KS/OK border somewhere between Enid to Perryton, keep eyes on institu/radar/satellite and be ready to push north or south. If convective coverage remains limited, it's going to be a massive hoard on whatever cell(s) do develop. So hopefully the coverage will be SCT or far enough apart to limit significant convergence.

 

[Jeff Duda]

To paraphrase Jeff Goldblum from Jurassic Park: "Busts, uh, find a way."

In all seriousness, we just had a bust-like high risk setup on May 6th where the forecast was simply wrong in a lot of respects. That's always possible and how are you even going to know until the real-time observations confirm? IMO this setup certainly has some possibility of BSB, at least in one's chosen target area. Some of the CAMs aren't initiating anything over wide areas, showing lingering cloud cover and decent CIN still in place at 00Z. That said, I don't know how much I believe the CAMs in this situation with sparse dryline initiation. SPC clearly believes there will be storms, but where? Choosing the wrong target area would be a failure mode, wouldn't it?

I disagree that the forecasts were wrong for May 6th. I think the forecasts were pretty good. I just think most people don't know how to properly interpret the CAM forecasts. The CAM forecasts suggested possible issues with storm mode and coverage that most people seemed to simply ignore, relying instead on synoptic scale pattern recognition to dictate their expectation of the event.

That doesn't work these days. We have tools that can dive deeper into the details. CAMs.

In furtherance of the discussion of this particular event, here are last night's forecast UH fields for this day for the two experimental medium-range CAM ensembles. First, the ensemble max UH tracks (MPAS left, FV3 right):


Now the neighborhood ensemble probabilities of UH exceeding 75 m2/s2 (MPAS left, FV3 right):

If you're truly paying attention to the state-of-the-art advancements in forecast models, then right away you can see that there is likely more than meets the eye for this setup given that both ensembles are pretty scant with rotating storms south of the KS-OK border. And it's really easy for people to get caught up in the idea that Oklahoma will be the best place to be based on the synoptic scale pattern.

Not according to these models! And I bet they're not horribly wrong.

Pay attention to the details; fickle as they may be at this range, they should tell you something about the model's larger-scale pattern supporting this event. It's not a slam dunk for tornadoes in Oklahoma by any means.

 

[Jason N]

I

It's not a slam dunk for tornadoes in Oklahoma by any means

I agree Jeff and you're totally right about CAM solutions over Synoptic ones. Synoptically, The Target seems more in line with Northern KS, with SFC low center/DL warm frontal surge area and park it around Salina and see what happens. but the big fish lay to the south, whereas you said, nothing is for certain right now. another reason I want to play the border to maybe have the time to push north if need be.

There are so many CAMS now, I don't have a solid idea on all of them bias wise tbh. and not to take it too far off topic, but if you have a good link to where there is some good training for the cams, I would appreciate it. I feel comfy with HRRR, the NESTED and WRF-ARW. I have gotten some recent indoctrination with MPAS as well, but that's still new for me and I am still learning about its Voroni mesh pluses and minuses.

 

[Jeff Duda]

There are so many CAMS now, I don't have a solid idea on all of them bias wise tbh. and not to take it too far off topic, but if you have a good link to where there is some good training for the cams, I would appreciate it. I feel comfy with HRRR, the NESTED and WRF-ARW. I have gotten some recent indoctrination with MPAS as well, but that's still new for me and I am still learning about its Voroni mesh pluses and minuses.

Yes, and I can certainly see how some would be disinclined to examine these other model solutions without better knowledge of their forecast tendencies. It doesn't help that many of these particular model configurations will likely change, and the forecasts themselves will not run year-round. Thus, we are only getting an ephemeral glimpse at their capabilities and their contributions to the forecast uncertainty envelope.

But therein lies their utility - they sample a different portion of the uncertainty envelope. Other established models have known biases (ex: the HRRR tends to run dry and has less convection than other models; 3 km NAM tends to be the most moist at the surface and favors linear storm modes), but that doesn't mean that those models are all you ever need to make an informed decision. These other models give different depictions of events, but depictions that are still in-line with the possibilities given the current known/measured state of the atmosphere. These models are sufficiently well constructed and use the types of model physics that have been known to produce skillful forecasts in research applications, so they should be regarded as legitimate inputs to the forecast process. In my (limited) experience comparing the NSSL MPAS forecasts to reality, for example, they have tended to perform pleasingly well compared to operational CAMs in terms of their reflectivity pattern and storm evolution. So IMO they are worth adding to your regular repertoire of models to examine.

For this particular event and range, none of the above MPAS, or the operational, CAMs are yet in range, so we are forced to examine lesser tested systems such as the NCAR CAM ensembles. It is unknown how well they'll fare, but my argument is that since their largest source of error is not likely to be in their physics configuration - it's likely to be in the initial condition error that would be present in all models kicked off at this time - that they are still worth heeding.

To answer your main question, Jason, there are no training modules for these experimental CAMs. It takes a mind that is used to examining NWP model output and understanding what goes on within them that is most useful for understanding their capabilities when they're in an experimental state like this. But as such a person who possesses that experience and understanding, I advise those forecasting to take heed of what these particular models depict, because again, I doubt they're that different from what we would see if HRRR or the 3-km NAM went out that far.

 

[Brian McKibben]

I

I agree Jeff and you're totally right about CAM solutions over Synoptic ones. Synoptically, The Target seems more in line with Northern KS, with SFC low center/DL warm frontal surge area and park it around Salina and see what happens. but the big fish lay to the south, whereas you said, nothing is for certain right now. another reason I want to play the border to maybe have the time to push north if need be.

There are so many CAMS now, I don't have a solid idea on all of them bias wise tbh. and not to take it too far off topic, but if you have a good link to where there is some good training for the cams, I would appreciate it. I feel comfy with HRRR, the NESTED and WRF-ARW. I have gotten some recent indoctrination with MPAS as well, but that's still new for me and I am still learning about its Voroni mesh pluses and minuses.

One thing Iike to do after each event is re-examine all the prior CAM model runs available and see how well each model performed leading up to the event.

A lot of times people get mad that the models didn't perform like they thought. But I have found CAMS actually do a pretty good job if you step back and look at it.

Maybe they had initiation wro g by 2 hours but did good after that.

Maybe they were off in location by 100 miles.

Doesn't mean they failed.

Too many young chasers take a CAM reflective or UH swaths as exact gospel. You gotta be smart and recognize the evolution. Throughout the day and adjust based upon real-time data and the expectation of what the CAMs were forecasting.

I hope that makes sense.

 

[Brett Nickeson]

Maybe it's just my pessimistic tendencies, but I'm not convinced there will be much during daylight hours on Saturday. As some of the CAMs roll in, they seem to be confirming what I thought last night and this morning. That moisture appears to be advecting a bit too late for a Kansas daylight play, with many models barely showing the 60 isodrosotherm reaching I-70 by 0Z and the 70 around the Oklahoma border. There also seems to be a bit of a warm EML at around the 2km layer, though the NAM may be overdoing the cap due to its cold bias at the surface.

I hope I'm wrong, but this could be one of those spring-to-summer tweeners with high CAPE and EHI bullseyes that end up as a daytime bust. I'd wager that there will very likely be twilight and after-dark supercells after that better moisture gets north, but I think I'll hold off on calling for an outbreak for now. Plus, the target right now doesn't quite seem to be Iowa or southwest Oklahoma (yet), so it's got it's work cut out for it.

As an aside, I'll note that there is still a fairly large spread in the models for only 48 hours out. They seem to be struggling a bit with this one.

 

[Matt Hunt]

After my string of failures this week, and several uncertainties for Saturday, I think I'm going to call it a wrap and start my 19+ hour drive home tomorrow. Have also noticed the CAM's not breaking out convection in daylight hours, and I'm also not seeing a sharp dryline until after dark. Perhaps it's not only the cap, but also a lack of adequate convergence to initiate convection.