RCNYILWX

The common miss northwest or the (less common this season) miss southeast special. Place your bets and run. Picking that this will be the event to jackpot the heart of the metro would be like those who put money on FDU and Princeton to win their first round games. Not smart bets, but felt good when they worked out.

We better not get another shot of -20C at 850 mb lol. Does unfortunately look like a chance for another decent cool shot by late March standards beyond the Thursday-Friday system next week. The ensemble mean snow output next weekend into the following work week hints that some threat for snow may accompany the cool down. Ensemble guidance and teleconnection suggests a relaxation of the predominantly -NAO of late early-mid next week should be followed by another round of -NAO blocking. Until we can fully kick that blocking tendency, we may see a continuation of fairly pronounced cold shots following warm ups. The good news is that there currently aren't any signs of April 2018-like nonsense heading into next month. Sent from my SM-G998U using Tapatalk

I wasn't even looking at specific timing too much since the ingredients were fairly similar on the overnight guidance. Considering the propensity of warm fronts to get hung up by the lake this time of year, the 00z Euro had a plausible look to it. Plus if the wave is lower amplitude, you'd be more likely to have a weaker sfc low initially and gradual intensification more conducive to slower northward movement of the warm front. That's probably even too specific at 6+ days out, though using typical early spring setups as a guide, it's certainly more common for warm fronts to get hung up farther south than for them to rapidly surge north.

Next Thursday checks pattern recognition boxes for a legit severe/tornado threat in the region given a juicy air mass (55-60+ Td) surging northward with a sharpening warm front, 70-90 kt flow at h5, and forecast steep mid-level lapse rates. The forecast lower amplitude 500 mb pattern results in a more favorable wind profile with height, WSW or SW at h5, vs. meridional south or SSW with a negatively tilted trough. In addition, there would be a chance for right moving storms to have residence time near the warm front as opposed to immediately surging north of the boundary. All the usual caveats apply, including meaningful shifts in the currently favorable looking setup, but at this lead time, one of the better late March synoptic patterns in recent years. March 28, 2020 comes to mind and that event not performing to its potential exemplifies the caveats for March setups. Signal is there for a flooding threat as well if the boundary can stay convectively active through the night. At least it's something to monitor and more interesting than the brutal CAD to start this weekend.

Ratios probably were close to 10:1 in the narrow corridor with the warning criteria totals. Certainly overperformed my expectations up there on amounts, but I'd say overall thoughts on how things would play out weren't bad. If I had confidence in the amounts that were observed, particularly in central and northern McHenry County, would have issued the warning. Every event is a learning experience when you work in this field. Sent from my SM-G998U using Tapatalk

It's always been something this season that's resulted in the snow hole from the city and back south and southwest. The issue this afternoon was the stubborn dry wedge. This evening, despite more than cold enough 850 and 925 mb temps, it's brisk east- southeast low level flow that's brought in just enough low level warmth to result in full melting below 1-1.5kft, if you've been following trends on correlation coefficient. Interestingly, the RAP and HRRR had that low level warmth on runs earlier today, but the problem is they don't consistently perform well enough to put as much stock in them vs the other guidance. So ultimately we did get the heavier precip rates here this evening, and similar reflectivity last Friday was ripping aggregates with TSSN, but alas, it just ain't our winter.

Sorry you feel the way you do about our efforts. It's cliche, but in weather forecasting, there's always going to be a bit of 'you win some, you lose some.' However, that being said, I can assure you the headlines are not based "off feeling." When there's a watch in effect and potential to upgrade to a warning, it's a collaborative intra office discussion, in addition to the inter-office collaboration that goes on. Last Thursday, we had about 3 hours worth of discussions on how to handle the existing watch, not exaggerating. Since we inherited a watch on the (Wednesday) day shift yesterday, and it's within a timeframe when we usually make a decision, we're left with a) upgrade the watch or a part of it to a warning; b ) issue a WWA; c) hold onto the watch Options a or b are the typical courses of action. Option C is what we went with for last Friday's event given the unusually large uncertainty up to go time. Preference today (including input up to local management level), was to go with the WWA as our highest confidence option. In that, we explicitly forecast up to 7" amounts in spots for the state line counties. That shows it's not as simple as 6" = definite warning. My own thinking on the decision process was informed on how last Friday played out, with a similar expected 2m temp starting point at precip onset, and the 1-2 hour lag until progressively worse impacts, amidst intense snow rates and TSSN. If part of the evening commute is spent in a lagged ground and pavement response prior to worsening impacts, and 12-hour amounts end up at 6" or less due to shorter duration of sustained heavy rates, plus lighter snow rates Friday AM with temps near to slightly above 32F, that's part of how we felt comfortable going with a WWA. Confidence was not high in widespread 6-7" amounts per non-NCEP guidance being drier, and with much of the accums occurring at a less impactful time of day, we didn't think it necessitated a warning issuance. This of course doesn't imply that we will be right, but shows the process that goes into headline decisions. There's also collaboration between WPC and the WFOs regarding QPF, SLRs and snow and ice amounts. With the milder winter we've had, and no real classic winter storm setups this winter, we've been left with marginal and thus more uncertain synoptic evolutions. There is consistent post event evaluation going on that we use to build mental models of future events. But it doesn't ever make forecasts and headline decisions easy, and there's opportunities for future improvement that come about, from forecasts and headlines to our graphics and DSS.

With every event, the NAM gives more evidence that it should already have been retired. Sent from my SM-G998U using Tapatalk

Obviously intense rates overcome even April sun, but nighttime will definitely help since you don't need sustained heavy rates to accumulate efficiently. Sent from my SM-G998U using Tapatalk

Good call on the slightly farther southeast track. Interestingly, the models that showed mid 970s mb pressures did well. It was a fascinating system overall - like a tropical system with snow, and an incredible amount of TSSN. Without a strong high pressure system to the north or northwest, proximity to the center of the low was the biggest driver in getting into that core of damaging northeast winds. If the only detail you gave for a forecast for the Chicago area was that a 976 mb low would track north of the Ohio River, you'd say most of the metro would have been a lock for heavy snow. The fact that it was a nearly vertically stacked low meant that the mid level lows were more tucked in than a classic winter storm, keeping the deformation area/fronto banding also more tucked in. The radar for a time was very reminiscent of 2/24/16 and the intensity of the banding resulted in strong low level subsidence and drying that overcame good mid-upper level lift and steep lapse rates in place across the central and northern metro. Just not our winter. Sent from my SM-G998U using Tapatalk

Other unfortunate aspect is that all the model run total maps shown by the TV mets are algorithm driven too, and we all own that forecast. My office didn't put out an explicit snow map on more public facing WxStory graphics (it was available on the probabilistic page) on our home page and social media until Thursday PM, but yet "we" forecasted 10" in the metro. QPF and snow depth change maps, plus forecast soundings with omega are the way to go, plus its important to consider whether the DGZ is supersaturated. Looking at the Cobb output page helps too, although even that scientifically more rigorous methodology can have overdone ratios too. Better to play it conservative (ie. pos depth change including var density HRRR product OHweather mentioned) than to get burned with 10:1 and Kuchera ratios. Sent from my SM-G998U using Tapatalk

The CMC was not good, right for the wrong reasons in the Chicago metro. It was too weak and too far south with the surface low, and didn't capture the dynamic cooling that occurred to support the accums of 2-5" of paste in the southeast 1/3 of the CWA. It only gets credit for having the farther south precip swath. GFS operational did fairly well and then once the ECMWF adjusted, it did pretty well, not a great performance overall though because it was too far north until pretty late in the game. We only have the UKMET via Pivotal Wx, but it did perform pretty well at a longer lead time. NAMs performed the worst of the models we commonly utilize. HRRR and other CAMs did well with capturing the high winds in Central IL and IN. This event was also a good case for the algorithm snow maps to be banned. Forecast positive snow depth change was the best option for snow amounts.

Temps and dew points are solid above the colder guidance. Looking problematic for any accums in our IL counties and even NW IN will probably underperform unless it absolutely rips this afternoon following changeover. Strengthening east northeast flow may help bring in slightly lower dew points and help there, but especially IWX CWA. Up here in the metro, there's forecast to be strong lift in the DGZ, though strong subsidence and some drying in the lower levels northwest of the f-gen banding, per forecast soundings. Sent from my SM-G998U using Tapatalk

Cook County does have 3 zones, as does Will County. The evening shift opted to leave southern Cook out of the WSW, but did put in eastern Will. Sent from my SM-G998U using Tapatalk

They sent the watch at 957 PM EST, and then the warning upgrade at 1021 PM. Interesting. Sent from my SM-G998U using Tapatalk

Surprised IWX didn't join in the upgrade of the watch to present a unified message with us to the west and GRR to their north and DTX to their northeast. I see they did put some counties into a WWA. Sent from my SM-G998U using Tapatalk

18z EPS bumped north, more than a noise level change on the 10:1 snow swath. The "fun" continues lol. 18z vs 12z comparison attached.

Just sent an AFD update. Basically, it's hard to say we'll be 80% confident in warning snow amounts/impacts anywhere in the CWA. Assuming the NAMs are too far NW, general idea is southeast of I-55, *if* the banding is intense enough for dynamic cooling to overcome the marginal BL. Throw convection into the mix and the fact that we're basically trying to precisely place a mesoscale feature (response to f-gen circulation) and say we're confident the dynamic cooling will be enough, it's a tough call. Even with the GFS, we're talking pretty high end impacts for the populous south suburbs, Kankakee county, and NW Indiana corridor. But the Canadians (not that they're great models) have me concerned and getting vibes of the Christmas Eve 2014 bust. If the GFS is right, that's a similar look to what happened on 2/24/16. And again, there's error bars on the exact placement of banding northwest of the compact vertically stacked low. Sent from my SM-G998U using Tapatalk

The 00z Euro bumped back northwest some from the 18z run. It remains a very tough forecast. Impacts will be higher end in the heart of the band within the CWA, barring a UKMET or GEM like outcome. Lapse rates in the DGZ are progged to exceed 7C/km and there's quite a bit of -epv above strong fgen. Sent from my SM-G998U using Tapatalk

Last 3 runs of the GEFS 24-hour 10:1 snowfall mean. Edit: Added 00z GEFS members. Pretty large spread this close to the event.

We're actually technically not supposed to put the totals in the watch text products, according to NWS directives. However, the simplified formatter we went to 4-5 years ago spits out the deterministic ranges and most offices leave them. I think we (at LOT) usually change it to snow totals of 6 inches or more are possible. The unusual watch and warning issuance on the same calendar day is looking likely for this event. Sent from my SM-G998U using Tapatalk

100% agree, we're not talking huge differences in features anymore, but the difference in sensible weather impacts is dramatic. Gonna be a fun day at the office tomorrow lol. Sent from my SM-G998U using Tapatalk

I deleted my post since you beat me to it haha. I think this sort of variability closer in underscores the complex nature of this forecast. Sent from my SM-G998U using Tapatalk

I try to avoid looking at the HRRR at this lead time because it's a hot start model and frequently all over the map for winter systems. I really don't think it was meant to be used for winter forecasting outside 18-24 hours. It's more useful inside 12-18 hours. While it doesn't mean a particular solution in its extended runs can't have the right idea, it's much more often wrong than right this far out and therefore hard to put any stock in it. It was very bad for the ice storm (too warm) and also the event before that (too cold over northern IL). It may have performed decently for the pre Christmas storm at this lead time, but that's the only one that sticks out to me.

Very good agreement between the 18z EPS and GEFS actually with that shift south of the EPS. But that also means there's still a decent # of members north of the operational, similar to the 18z GFS/GEFS. When the 18z Euro did shift south, it was a good reminder to me to not necessarily put too much stock even into a model I tend to trust more, particularly in this type of forecast that is sensitive to relatively small changes in the setup and the mass fields. @OHweather, great post btw. Would you put more trust in the ensemble means at this point? Wondering if that may be the way to go since they've been generally more stable than the operational runs. Sent from my SM-G998U using Tapatalk