July 26. min. max. avg
Bitt 2nw valley 58.1 77.6 67.8
Garrett College 61.0 78.4 69.7
Can-Heights 59.5 74.2 66.8
CRN-Canaan 57.6 74.1 65.8
Cabin Mt 57.7 71.6 64.6
Cabin Mt north 58.6 72.7 65.6
Spruce Knob 56.5 70.5 63.5
Snowshoe 58.8 70.2 64.5
Can-Valley Floor 48.2 77.0 62.6
7Springs 61.5 75.0 68.2
Dreary looking start, some fog, low clouds and that then burnt off for a mix of clouds and sun of varying amounts from late morning, through evening with the sunny side of things increasing as the day went by.
Temp profile this afternoon
Satellite this afternoon
Radar, mostly last evening in where it did rain, with 24 hour estimates running back to 7pm 7/25
Flow
surface features and 500mb height anomalies
500mb flow
SST(Sea Surface Temperature) anomalies and 7 day change
Off The Models
Meteogram off the 12z ECMWF for Garrett County
12Z ECMWF
3 hourly precip
Total Precipitation
18z GFS
6 hourly precipitation
Total Precipitation
18Z 3km Nam
2m Temperatures
2m Dew Points
Simulated Radar
Total Precipitation
10:30pm reality, model comparison
The new 0z 3km Nam run is in. So thats the go to choice. The formula I’v been watching, to get a good feel for what the cold frost pockets can achieve is
1- 10m wind off the 3km Nam calm. In this case, we can check that off. At least for the valley spots. While 3km is high res, its still not tight enough to pinpoint details, but gives great hints.
2- Cloud cover, are we seeing clear skies. The 3km nam does very well at this. In this case, checked this off. 
3- 2m Dewpoints. Pretty big component. Again, 3km nam is detailed, but not tight enough for minute details. The observations and comparisons has led to the conclusion on many many occasions, that, with a dry air mass incoming, or in place, calm winds, clear skies, subtract 4-8° from the 3km nam modeled dewpoint and that is generally where the 2m temp comes in at in the cold pockets. Placement of the modeled 2m Dewpoints is not important. If conditions above(1&2) check off the list. Ex. If the models lowest dewpoint is in Elkins at 45 and Canaan reads 55 on the model, Canaans frost pocket actual dewpoint, air temp, if conditions(calm, clear)of 1 and 2 are equal to Elkins, (on the model) reality should run 38-42° 2m air temp. In the case below off tonights 0z 3km nam, 56.2° dewpoint is the min at the example time. All in all for this example, lowest dewpoints are in the ballpark area of the conversation, despite it doesn’t have to be. 
4- While in cases of east flows, CAD events, even daytime maxes with a breeze I love the 3km nam, (best there is bar none, beats HRRR easily) it is not good at hitting cold pockets. Again, not in tight enough for small details even at 3km, but I would think as in the current case, and many other times observed, it should do better. With 0 wind, 56° dewpoint, clear skies, why is it not modeling pockets of cooler temps? 3km should be enough to do that. This aspect of the model is one of its problems on 2m temps. Its not always great at convective storms and will falsely on the model cool locations that do not occur. That’s another downfall. It is though great for frontal timing however. Currently, thumbs down for tonights example on temps. Way to warm. 
Reality views. While the valley frost pocket has no anemometer, its evident conditions are calm, while at 4000′ a 10mph exist. The model, looks decent. At 4000 its 63°, not bad on the model, 51° in the colder valley area, no clue on the model, except using the 1,2,3 formula from observations of the model to reality.
Reality at a little past 10:30pm
(To note, Hrrr last night showed clear skies through daybreak, 3km nam showed increasing cloud cover and temps creeping up, 3km Nam did well.)
Satellite this afternoon
Allegheny Mountains Weather 