high risk

Certainly a synoptic wind threat later tomorrow, especially along and behind the front. The question is whether there are severe gusts with a line of showers just ahead of the front - the NAM is impressive but is somewhat of an outlier with that feature for now.

Interesting thought. In my mind, the severe season here doesn't really kick in on average until around May 10 (yes, there are exceptions), so moving that start up to late March would be a big jump. How about if we try a smaller increment and the severe season in mid to late April instead? I could also see an introduction of a fall season (aside from tropical events). We usually end up just a bit cool for those dynamic autumn systems.

Who wants to drive to Williamsport on Wednesday to experience a squall line of heavy snow?

Always always always use the NAM Nest for temperatures in a cold air damming situation when other guidance is faster trying to erode the wedge.

Most of Friday looks nasty for sure. That said, we'll end up with a high in the 61-65 range due to the temperature at 12:01 AM.

Definitely some sleet pellets mixed in here in southern Howard County. Models are pretty insistent that the heaviest downpours will be after dark - some elevated CAPE suggests that a few rumbles of thunder aren't out of the question.

The NAM Nest is pretty clearly a dry outlier; it looks like every other model (including the NAM parent) has snow here tomorrow. You're spot on, though, that the temperatures are super tricky. It's tough to imagine a scenario in which the roads get messy (unless the rates are *really* impressive), but I can definitely envision some accumulation on grassy surfaces and decks.

Sort of. The FV3 is the model "core". The GFS already uses the FV3 core. The NAM uses the NMMB core, and the RAP/HRRR use the ARW core. The NAM/RAP are technically to be "replaced" by the GFS. All of the hi-res models (including the NAM Nest and HRRR) will be replaced by the Rapid Refresh Forecast System (RRFS) ensemble system which will use the FV3 core.

A couple of corrections here: 1) The SREF is not truly connected to the NAM. Half of the SREF members use the same model core as the NAM, but half use the same core that is used by the HRRR. 2) Neither the NAM nor the SREF has been upgraded in many years, and both systems are due for retirement within the next 2-3 years.

While we're enjoying an amazing afternoon, the 18Z NAM just happy hour'd us for Saturday snow.

All of the CAMs nicely depict cooler air to our east coming westward towards our area, but they really don't have it arriving until after dark. By then, nocturnal cooling will make it less apparent that a backdoor front is arriving.

If one loops back through the older NAM Nest forecasts for this morning, its 00Z cycle last night picked up on the idea that it would take time this morning to bust into the warmer air. 80 is likely now off of the table for most, but at least the low 70s seems likely for much of the DC area with 75 still possible on the south side of town.

The best signal is for those north of the DC Beltway, but most of the CAMs do show convection rolling through during the early to mid afternoon hours. The soundings to me support small hail more than they support strong winds, so I'm not sure about the 5% wind threat, but I'll take it for late February.

Nice to see the entire 00Z suite of CAMs show a "non-shutout" (with some much better) for most of us.

agreed. The forecasted soundings show some elevated CAPE, and there were a few lightning strikes earlier in the storms that fired just east of DC.

Actually, it warms my heart to see so many people posting the accumulated snow depth maps, as the 10:1 maps are for sure going to be inflated. That said, while I strongly support the snow depth products, they can run low in events in which the soil is warm. If we can hold off the snow until dinner time and get some decent rates, it might stick a bit ore efficiently than the snow depth maps show. Not a fan of the Kuchera.

The primary low transfers to a new low near.... check notes.... Winchester.

Radar looking impressive!

Pretty solid signal in the CAMs this evening for low-topped convection later Thursday for north-central MD. Soundings might support some very small hail if the updrafts are sufficiently strong.

I should have clarified that the model struggles with the temperature drop were east of the mountains.... which makes sense.

The models have done a lot of aspects of this frontal passage well, but one struggle has been how quickly the temps drop behind the front, and that had huge implications for potential for the flash freeze. Even all of the 12Z models today show me in the low 20s by 11AM, but it's still 28 here. If you look at HRRR trends, it took a long time to catch on to the idea that the big drop would be delayed.

Snowing in southern Howard County. 37 degrees

Verbatim, the CAMs have consistently been showing that if there is a threat of snow developing immediately behind the front, it won't be until the boundary crosses the Potomac. Doesn't mean it will happen, but the lack of snow in northern VA should not be a surprise.

Here is how it works. Inside the GFS, there is a bucket for snow accumulation. Anything from the direct model integration of the microphysics scheme that reaches the ground as snow or sleet goes into the snow accumulation bucket as a liquid equivalent. This liquid equivalent is output as a snowfall product. The snow (+ sleet) is also passed into the land-surface part of the model where an SLR is applied. This determines how much snow is on the ground in the model. The disconnect has 2 sources: 1) users apply their own ratios. Most use 10:1, but some use the generous Kuchera. Kuchera for this case will love the crashing temps and put some weenie ratios and end up with big accumulations 2) the snow depth is instantaneous, and some melting (not in the Friday example) or compacting may occur by the time shown The model is absolutely "handling the fast-moving front". Right or wrong, it has snow falling on the cold side of the front. The products showing "snowfall" do have an element of post-processing artifacts, especially the Kuchera.