high risk

Same here! As for the overnight rainfall, a bunch of HRRR cycles late yesterday showed a significant minimum in the precip for most of the area. I dismissed it, because it really didn't fit the narrative, but it was spot on.

It's remarkable that every CAM has a forced line of convection tomorrow morning arriving on the west side of DC around 7AM and hitting the eastern side around 8. I'm also very intrigued by all CAMs showing a broken line on the secondary front in the early afternoon. This environment will have some CAPE and some low freezing levels, so I suspect that these could be accompanied by small hail or graupel.

It is indeed remarkable how fast these AI-based models run. It really opens the door to running larger ensembles than ever thought possible, although getting meaningful spread still needs research

All of the new AI NOAA global model products are using Graphcast. Tropical Tidbits is displaying the deterministic GFS, called the AIGFS. There is also an AIGEFS, but it's not yet displayed here. The HGEFS is the hybrid GEFS, which combines the 31 members of the "regular" GEFS with the 31 members of the AIGEFS. Note that the AIGEFS has been shown to be somewhat underdispersive, so treat high probabilities with caution. https://www.noaa.gov/news-release/noaa-deploys-new-generation-of-ai-driven-global-weather-models

This is the Round 2 for midday / early afternoon tomorrow that I referenced in the above post. Interesting to see some impressive modeled reflectivities with that.

Every CAM has a nice forced line just ahead of the front, with consensus showing it passing through DC around 6 or maybe 7AM. Most of these events end up being just heavy rain, but they do occasionally surprise us with more. Another thing to watch is that all of the guidance has another round of scattered heavy showers around midday in a well-mixed environment with a bit of low-level CAPE and low freezing levels. Would not be surprised to see some graupel out of those.

The NAM Nest and HiRes Window FV3 both have a similar idea of initial rain turning to snow, with a nice band moving to the southeast in the late night hours. Surface temperatures are VERY marginal, but most of the column is "cold" other than right near the ground, so it should be mostly snow after initial rain for some. It's a good thing that this will fall at night, or else stickage would be completely dependent on rates.

Probably this spring. But what about this run is so clearly awful?

Actually, there are a few SREF and GEFS members that (no surprise) are super wet. Those inputs are driving those high percentile values.

Honestly, I usually start with the 10:1 maps and make some modest mental adjustments (up or down, depending on whether temperatures are marginal or solidly cold). The concept of Kuchera is good, but it's not as simple as a colder profile means better ratios, and I find that Kuchera runs outrageously hot in many cases when the snow falls at temperatures well below 32.

This thread is reminding me why I love summer: the long hours of daylight, the warm breezes, and not have to look at a stupidly-inflated Kuchera plot for 8 months.

I think this is a great point - social science is really going to be a factor here. Tomorrow's decision will be made with Tuesday's decision fresh in their minds, and they'll absolutely be scared to death of another delay or closing that ends up not needed. And it seems like a 2h delay wouldn't accomplish much tomorrow, as it will be cold and snowing 2 hours later. It's probably either a "normal day" or "closure" decision.

What stands out to me is that even if the totals are very light, temperatures at onset are in the mid to upper 20s for most of us (except maybe for the far southern crew). With the earlier onset, as others have noted, the morning rush hour could be very tricky, and school systems are going to have some very challenging decisions to make.

This is bothering me a lot, as the NAMs are usually the colder solutions. I think we're still in the range where the NAM could be having synoptic issues, but if it continues on Monday to keep the accumulations further northwest of the I-95 corridor, I'm going to guess it's right. The faster onset, shown by the GFS and its hi-res HiResW FV3, would help a ton here, although even where it shows snow, the thermal profile has pretty much zero margin for error.

NAMs will around through at least the rest of this winter.

Incredibly tricky temperature forecasts this weekend. Warm air will be lurking just our south and southwest, but there is increasing concern now that we'll end up on the cooler side of the boundary each day.

While we'll be freezing tomorrow, there is very good model agreement that we get quite warm on Sunday.

These low reflectivity/high wind event fascinate me.

Personally, I look at both the HRRR and NAM Nest as my go-to CAMs. The 3 Hi-Res Windows are lower on my list. The HRRR and NAM Nest both have their strengths and weaknesses. The HRRR has a big advantage in running hourly, and there are things it does to create a more accurate starting point (a sophisticated assimilation of radar data, heavy usage of surface obs....) The NAM Nest hasn't been updated in 8 years. The HRRR hasn't had a recent update but was last updated 4 years ago. In terms of biases with rainfall, NAM Nest tends to run wet, but the HRRR sometimes runs dry - might be microphysics, but tough to confirm. And the HRRR is very sensitive to the hourly updates, so its solutions bounce around a lot. The NAM Nest tends to be too cool with sfc temps, but it handles low-level inversions way better than the HRRR due to a better PBL scheme. And as noted earlier, the NAM Nest tends to overdo orographic precip due to a very sharp topography field that was put in to enhance winds over terrain. That's just a few things off of the top of my head.

2% TOR and 5% WIND for much of the area in the new Day 2 outlook. Probably not enough sfc-based instability to take advantage of the intense shear, but the threats are certainly non-zero.

We might actually get a good amount of rain (Wednesday night into early Thursday) AND have it hit 70 around midday Thursday.

As I noted in the other thread, we need to keep an eye on Thursday. There is growing model agreement that we'll enter the warm sector in the morning and clear out a bit, after the widespread warm advection rains and convection exit to the northeast. There will likely be some scattered convection from late morning into the early afternoon along the occluded front, and it will be in a strongly-sheared environment. The question is how warm we get. It's tough to destablize here in late October, and we'll need really warm surface temps to overcome lousy lapse rates. It will get very warm in southeast VA and the southern Eastern Shore, but will that really warm air be advected this far north? So far, the experimental RRFS is the most aggressive, with temperatures reaching 70 in DC and an area of 750-1250 surface-based CAPE covering much of the area.

Some sneaky severe potential Thursday midday or early afternoon as we enter the warm sector. At the least, some elevated convection late Wednesday night into Thursday morning seems likely.

Pleasantly surprised to see some lighting strikes over northern Virginia

MRGL for most of the area which seems fair. Guidance is now in very good agreement that a strongly-forced, narrow line of heavy showers will race across the area during the evening: 9pm or so in the western burbs, 10pm in the DC-Baltimore area, and over the Bay by around 11. Wind fields are intense with some low-level shear, but instability will be negligible. Lightning therefore seems unlikely, and while winds will be gusty, severe gusts will be difficult to achieve. That said, I still can't completely rule out lightning or a few higher end gusts, and it's also possible that we end up with a tornado warning for someone for a brief couplet.