2026 Mid-Atlantic Severe Storm General Discussion

[vortex95]

17 minutes ago, batmanbrad said:

CSU-MLP latest still painting the maximum TOR possibilities right over the general DC area:

That's a solid sign.  One thing that can misleading -- cool, dreary, and stable-looking can make ppl think, "how can this 'go away' so fast and big svr occur?"  Well, it does and can be quite dramatic.  Wichita Falls TX was socked in stable ahead of the warm front on 4/10/1979, and then an F4 cut directly through the city for the costliest tor up to this time.

Same thing in Windsor CO on 5/22/2008.  This was an odd one b/c it occurred the late morning, moved NNW, and the storm formed in a very narrow sweet spot of instability/shear ahead of a sfc low.  See here:
https://www.weather.gov/bou/TornadoMay22nd2008

Mean-looking!
https://www.youtube.com/watch?v=GfYuT0oalXQ
 

 

[wxmeddler]

2 minutes ago, vortex95 said:

It loves to make giant convective blobs on its composite reflectivity fcst.  Once it goes operational, and ppl use it more, I bet that will be a sig complaint, and they will address it.

A big issue w/ mesoscale models now is that the more hires they get and more directly simulate the atmosphere processes, the challenge to get everything in balance goes up in a non-linear fashion.  In other words, any small tweak to fix one issue can have big impacts to another part of the model.
 

The over-convection and "big blobs" on the RRFS is a known bias or "bug" with the FV3 Dynamical Core. This was explored recently by Lou Wicker at OU (See: https://doi.org/10.1175/MWR-D-25-0230.1). In short, it can't really be fixed without significant changes to the core. OMD hasn't said it outright publicly but the original plan was to turn off the HRRR and RAP at the same time as the RRFS Operationalization. But they've scrapped that specifically because of the FV3 core issues. Now the plan is to turn off the HRRR and RAP when the RRFS transitions to the MPAS Core in ~2028. (https://www.weather.gov/media/wrn/calendar/RRFS_webinar_20260625.pdf)

 

[vortex95]

30 minutes ago, Kmlwx said:

Good demonstration of how more resolution is not always automatically "better." 

There's probably a "happy medium" where errors can be smoothed while still having good resolution...but I imagine if we had like 0.25km resolution models it would be chaotic at best. Until better computing comes along I guess. 

Here is the best line I have seen that tempers the idea that the models are so good, they are better all the time by default, and by extension, human input is not needed.

"Models now can give very precise-looking forecasts/output that can be *precisely* wrong!
 

And going one step further:

"Precision does not mean accuracy!"

[Rhino16]

23 minutes ago, wxmeddler said:

The over-convection and "big blobs" on the RRFS is a known bias or "bug" with the FV3 Dynamical Core. This was explored recently by Lou Wicker at OU (See: https://doi.org/10.1175/MWR-D-25-0230.1). In short, it can't really be fixed without significant changes to the core. OMD hasn't said it outright publicly but the original plan was to turn off the HRRR and RAP at the same time as the RRFS Operationalization. But they've scrapped that specifically because of the FV3 core issues. Now the plan is to turn off the HRRR and RAP when the RRFS transitions to the MPAS Core in ~2028. (https://www.weather.gov/media/wrn/calendar/RRFS_webinar_20260625.pdf)

 

The part about the RAP is interesting to me because it seems like SPC Mesoanalysis is created using the RAP. I wonder how much work it is to change all that over unless something else replaces it like WoFS.

Edit: I guess they did it in 2012 with the RUC

[yoda]

Any idea why 1630z OTLK from SPC is late?  I'm checking kamala and SPC... I don't see it

[vortex95]

18 minutes ago, wxmeddler said:

The over-convection and "big blobs" on the RRFS is a known bias or "bug" with the FV3 Dynamical Core. This was explored recently by Lou Wicker at OU (See: https://doi.org/10.1175/MWR-D-25-0230.1). In short, it can't really be fixed without significant changes to the core. OMD hasn't said it outright publicly but the original plan was to turn off the HRRR and RAP at the same time as the RRFS Operationalization. But they've scrapped that specifically because of the FV3 core issues. Now the plan is to turn off the HRRR and RAP when the RRFS transitions to the MPAS Core in ~2028. (https://www.weather.gov/media/wrn/calendar/RRFS_webinar_20260625.pdf)

 

Thanks!  Great to know and I will pass along.

Turning off the NAM I have no issue w/, but when you roll out a new model, you should always keep its predecessor for a time for comparison.  The NAM is severely outdated, but the HRRR is not and still does a solid job, so kept it UFN!

They were supposed to stop the HWRF and HMON last year, as the HAFS has shown marked improvement overall (I was sold when it nailed MIlton's RI, then weakening, then another RI, and then its rapid weakening leading to landfall).  The HMON?. recall when that was first rolled out in 2017 I think.  It was showing crazy intense TCs, so much the model fields would "turn inside-out" b/c the intensity was nuts.  But the HWRF and HMON continue.

I prefer avoid the notion "more is better."  Yes, second options and consensus have value, but only to a point.  It can easily become TMI and hurt the fcst process.  Also, you have those that always look for the one model that shows the worst case scenario, and they run w/ that for clicks/likes on social media, when it is well known X model is terrible or has bias for Y phenomena.  It's like the GFS for spinning up bogus TCs beyond day 7 or the ICON in just how it handles TCs (much too intense forecasts for current TCs).

[yoda]

1630z OTLK from SPC disco for our area

   While storms have weakened, showers/cloud cover remain prevalent
   near/east of the Appalachians at midday across the Mid-Atlantic
   region. This will tend to limit/delay destabilization in many areas
   (especially with north-northeastward extent). That said, air mass
   recovery/advection is expected from the west, with ample upstream
   clearing per visible satellite imagery across the Cumberland Plateau
   into parts of the Allegheny Plateau. Pending the timing/location of
   destabilization, some terrain/lee trough-related development may
   diurnally occur, although the most prevalent/more certain scenario
   may be for potentially severe storms to spread into the region this
   evening. Damaging winds will be the most common hazard regionally,
   although a tornado could occur, attributable to MCV-enhanced flow
   and/or with any storms interacting with the effective warm front
   across the Mid-Atlantic region.

[vortex95]

52 minutes ago, wxmeddler said:

The over-convection and "big blobs" on the RRFS is a known bias or "bug" with the FV3 Dynamical Core. This was explored recently by Lou Wicker at OU (See: https://doi.org/10.1175/MWR-D-25-0230.1). In short, it can't really be fixed without significant changes to the core. OMD hasn't said it outright publicly but the original plan was to turn off the HRRR and RAP at the same time as the RRFS Operationalization. But they've scrapped that specifically because of the FV3 core issues. Now the plan is to turn off the HRRR and RAP when the RRFS transitions to the MPAS Core in ~2028. (https://www.weather.gov/media/wrn/calendar/RRFS_webinar_20260625.pdf)

 

Along the same lines, how does the RRFS do w/ TCs?  And for that matter, the HRRR?  We know the NAM (3 and 12 km) is terrible and was never designed for TCs, but the newer CAMs?

[midatlanticweather]

76/76 now but breezes from the south have been more steady and the lower foggy haze is lifting. Also seing very bright patches as the clouds seem to want to break up. Little patches of thinner clouds are showing up on satellite. We shall see if this enhances anything. Air is thick 

 

[jewell2188]

Seems as though on visible, clearing should progress west to east very shortly!

[Eskimo Joe]

Latest HRRR puts a legit bow through Frederick/Loudoun/Montgomery/Carroll counties right at sunset.

[SnowenOutThere]

1 hour ago, vortex95 said:

Turning off the NAM I have no issue w/, but when you roll out a new model, you should always keep its predecessor for a time for comparison.  The NAM is severely outdated, but the HRRR is not and still does a solid job, so kept it UFN!

Not the thread for this question, but is the conversation for it. I understand turning the NAM off for convective season, however, I don’t understand turning it off for winter short term stuff. I would love to know why, I assume it’s something in how it handles thermodynamics and WAA, but the NAM is almost always the most correct model when it comes to refining the mixing line in synoptic scale winter storms. It just seems odd to turn off a model which has a very useful edge on all others, even if it sucks at a lot of other aspects; could at least compromise and run it before major storms on the east coast. Would love to know your thoughts on both why it surpasses the other mesoscale models in the mixing line placement and how to compensate for its disappearance come winter time. 

[vortex95]

1 minute ago, SnowenOutThere said:

Not the thread for this question, but is the conversation for it. I understand turning the NAM off for convective season, however, I don’t understand turning it off for winter short term stuff. I would love to know why, I assume it’s something in how it handles thermodynamics and WAA, but the NAM is almost always the most correct model when it comes to refining the mixing line in synoptic scale winter storms. It just seems odd to turn off a model which has a very useful edge on all others, even if it sucks at a lot of other aspects; could at least compromise and run it before major storms on the east coast. Would love to know your thoughts on both why it surpasses the other mesoscale models in the mixing line placement and how to compensate for its disappearance come winter time. 

Yes, the NAM has some uses.  It handles CAD much better than the GFS, as one example and its MOS reflects that.  But the HRRR handles CAD well, so no loss there.

Since it has more vertical levels than all the global models, its should handle mix zone better and we are used to that.  But again, the HRRR and now RRFS should effectively replace that skill (eventually).

Losing some for other gain happens.   I recall NHC objected to an upgrade of the GFS some years ago b/c it was clearly shown that the upgrade made TC skill less for the EPAC basin.  The upgrade was done regardless.  So it's hard to please everyone!

 

[Eskimo Joe]

1 minute ago, vortex95 said:

Yes, the NAM has some uses.  It handles CAD much better than the GFS, as one example and its MOS reflects that.  But the HRRR handles CAD well, so no loss there.

Since it has more vertical levels than all the global models, its should handle mix zone better and we are used to that.  But again, the HRRR and now RRFS should effectively replace that skill (eventually).

Losing some for other gain happens.   I recall NHC objected to an upgrade of the GFS some years ago b/c it was clearly shown that the upgrade made TC skill less for the EPAC basin.  The upgrade was done regardless.  So it's hard to please everyone!

 

Use NAM MOS and HRRR together and you knew when the flip from SN -> IP -> RA would occur 90% of the time. That combination consistently schooled the GFS and Euro inside 30 hrs.

[vortex95]

1 hour ago, midatlanticweather said:

76/76 now but breezes from the south have been more steady and the lower foggy haze is lifting. Also seing very bright patches as the clouds seem to want to break up. Little patches of thinner clouds are showing up on satellite. We shall see if this enhances anything. Air is thick 

 

In tropical situations, it can be useful to use the virtual temp at the sfc instead of the dry bulb temp.  Using it can represent CAPE better in such environments (virtual temp decreases CIN  and increases CAPE).

76 dew point is nothing to shake a stick at!  Don't need too much heating for things to go and go well w/ that kind of juice.

[Eskimo Joe]

Per SPC mesoanalysis, there's a pretty solid are of instability cresting the Allegheny Front.

[vortex95]

41 minutes ago, Eskimo Joe said:

Latest HRRR puts a legit bow through Frederick/Loudoun/Montgomery/Carroll counties right at sunset.

The line is already there in ern OH (looks like a LEWP), so the HRRR is not creating the mini-bow echo from nowhere.

HRRR has a couple of cells before the bow, and then clusters after it!  My kind of tstm day!

[yoda]

Watch coming shortly... looks like for i95 corridor westward

https://www.spc.noaa.gov/products/md/md2000.html