We are about to enter a wild week of weather, and this blog will be my first attempt to break it all down for you. There is an insane amount of material to cover here, so please just bear with me as I try and lay it all out for you. I am going to provide a brief summary first of the snow threats that we have over the coming week, followed by a summary of the other threats I see from storms and the pattern, especially the record cold that we could be seeing. After I lay all my basic thoughts on the table and give you an idea of what we are looking at, I want to delve into the forecast much deeper. I am going to break down multiple model runs with pictures, etc. to show you exactly why I am favoring more snow with the upcoming coastal. This will be a fairly technical discussion, so if you are looking for the quick and simple forecast, just read the first few paragraphs and then I will highlight where the technical forecast will begin. You may also want to read the last few paragraphs, which I will label as “conclusion” and I will wrap everything together and say both what we know, what I expect, and what it is too early for me to say. Here we go…
First, I want to briefly mention that tomorrow we may have a few snow squalls or showers move through in the afternoon or evening. No accumulations expected, though a quick dusting is possible should the stronger squalls stay together longer than expected. Roads can briefly become icy in the squalls, but no major impacts expected. I won’t be referencing this again on the site, but make sure to follow along on Twitter as I’ll be covering the movement of snow squalls tomorrow if I expect any impacts. It’s very possible the entire region misses the snow altogether.
With that out of the way, I want to focus on the storm threat that I will be breaking down in the technical discussion later. A long-duration snow event looks to impact all of Southwestern Connecticut between Thursday and Friday and has the potential to impact schools either on one or both of those days. Closures are certainly not out of the question. Light snow will break out Thursday morning and continue through the day off and on as overrunning precipitation from a primary low pressure center moving through the Ohio River Valley moves through. A couple of inches of snow may fall on Thursday, but I don’t expect the snow to really be heavy. In fact, most weather models show that the snow on Thursday will generally be quite light. As the upper level energy in the large trough digs, deepens, and becomes negatively tilted (more on this later), energy from the inland low will quickly transfer to low pressure forming off the coast. What this will do is enhance snowfall and create a precipitation shield around this forming Nor’ Easter that will be to our south moving at a decent clip to the northeast. Enhanced snowfall is thus expected into Thursday night, and some models show enough blocking to slow down the low pressure resulting in significant snow lasting into the day on Friday. There still remains a lot of model disagreement regarding how heavy snow will get Thursday night into Friday and where the low pressure will be placed, and I’ll go more into that later. Snow will then move out later on Friday and lead to an absolutely frigid but sunny weekend.
That’s the other thing I am going to touch upon before I get into the technical discussion: the frigid air coming Friday night into Saturday. Weather models show that Friday night behind the storm temperatures will drop like rocks, especially as winds ease and skies clear. Low temperatures across the area will likely get down near 0. By 1 AM most weather models have the entire region in the single digits with heavy snowpack cover and still multiple more hours for temperatures to drop. This will likely be the coldest night of the Winter thus far, and though the cold air mass will not last very long it has the potential to break some daily records. Some areas could easily get below 0 away from the coast. Granted, next Monday night into Tuesday/Wednesday we could be looking at even COLDER air despite a brief moderation Sunday into Monday morning, but either way Friday night, Saturday, and Saturday night will be downright frigid. After that, I see yet another storm threat Sunday night into Monday, which is why there will be a temperature moderation. Because the cold air will move out very quickly, there is a chance that the low pressure Sunday night cuts to our west and gives us mainly rain Sunday night into Monday, and most models have that. However, a few still have enough cold air and blocking to force the low pressure to our east, which would again allow more snow. The main focus over the next few days will be on the January 2nd and 3rd storm, but the treat of this other storm a little further in the future will also be followed.
TECHNICAL: With the summary of the weather pattern now set, I want to take some time to give everyone who is more interested in the meteorology behind the upcoming storm (or people who just want to know what my detailed thoughts are regarding what will happen with the storm) a breakdown of what models show what and what I believe will end up happening based off of model trends. The reason everyone is so hyped up about the storm and you may have heard of a “massive snowstorm” coming is because one of the most accurate models, the ECMWF, is very aggressive, showing around a foot or more of snow across the entire region. Its ensemble members, 51 similar models run with the same algorithms under slightly different conditions, seem to show a similar solution, meaning that there is support for this model run. The main issue is that numerous other weather models disagree with the solution that the ECMWF shows, hence why one cannot forecast a large snow storm at this time, and for me it is just too early to forecast something like that. Other weather models are not as significant, but still generally seem to give the area 3-6 or 4-8 inches of snow, even without an extremely strong coastal low pressure that the ECMWF has had. The key that I will be focusing on, though, and the reason why my initial forecast for all of Southwestern Connecticut is for 8-16 inches of snow with this storm, is because of both the upper level trends and the surface trends that other models have been exhibiting that seem to point towards a solution similar to the ECMWF, though not quite as extreme. In fact, I think something like 8-12 inches is more realistic with the solution I want, but because of both the potential for this low pressure to really bomb out and because of the really cold air on the back end that could easily bump up snowfall ratios I am keeping it as 8-16 inches, especially because the storm is so far out right now. This mainly means that I expect the National Weather Service to be issuing Winter Storm Warnings for all of Southwestern Connecticut due to the upcoming storm, though I still am not totally sold on a storm as slow nor as extreme as the ECMWF. It could happen, but the ECMWF has a model bias to over-amplify sometimes (though strangely the GGEM has a worse bias to do that and so far it does not seem to have over-amplified the upper levels). What this means is that instead of a due west to east flow, the blocks in the jet stream pushing it more north/south are sometimes made out to be too strong and thus the ECMWF and sometimes the GGEM can forecast the jet stream to bend and buckle more than it actually will. This could be at play here, but we are also dealing with the opposite bias in American models, that typically keep the upper level air patterns too “zonal” in that they have them going too west-east and not having them bend or buckle enough to support large storms on the east coast. Let’s look at a few examples below.
First we look at the European model, the ECMWF. This image is a standard setup for a Nor’ Easter. The top left frame is the 500mb pattern, whereas the top right is the surface pressure frame. Those are the two we will be focusing in on this analysis. What we see in the top left is a trough of low pressure, or a U shaped dip in the jet stream that allows a direct discharge of upper level energy downward. This happens most when the bottom of the U points in an easterly direction, making the trough “negatively tilted” like we see here. Though the trough is actually close to neutral (due north/south orientation) over the next 12 hours, which are not publicly available, the trough turns negative and the already strong surface low pressure center, seen on the top right off Virginia, moves off to the northeast and strengthens even more rapidly, throwing back a lot of precipitation and hitting the Northeast with strong winds. I am going to compare this now with the Canadian weather model, the GGEM, at the same time. That image is here. As you can see, the storm is gone. Literally, the low pressure center is off the surface map on the top right, and the trough on the top left is neutrally tilted but over New England, having pushed the storm away. Thus you can see that the GGEM is much quicker with the storm, which is weird because as I mentioned it is typically one of the slower models. Back up 12 hours and look here. This looks more like the ECMWF with a defined low pressure center off the east coast (so you know, this is 7 PM Thursday now, whereas we were looking at 7 AM Friday before). Still, the surface low pressure center is much further east, and the trough on the top left is much further east than the ECMWF showed and does not extend as far to the south. These are the two most accurate models and yet they show radically different solutions with the storm. Typically, I would favor the GGEM, as I like the way it has been performing since a major upgrade last season, but I’m favoring the ECMWF solution more, just not as slow/extreme, mainly because of other model trends that I have been noticing. So far this year, low pressure centers seem to have ticked northwest near the end of the run and formed slightly earlier than guidance would have suggested, a trend that favors the ECMWF. Along with that, ridging out west supports placement of low pressure slightly further west than the GGEM shows when it tries to skirt it out to sea. But even in the GGEM scenario, enough precipitation from the upper level energy and from remaining low pressure near the Northeast remains to drop significant snow, mainly because the distance of the main low pressure allows so much cold air to remain in the region that snowfall ratios will be really high and snowfall would be quite fluffy. Thus the GGEM still drops around 8 inches of snow, even having the main low pressure miss us. That’s why I am fairly confident in the 8-16 inch forecast I had before.
But why do I think that models will slowly trend towards the ECMWF? A couple of reasons. First, the GFS does not align well with its own GEFS members when looking at the 500mb pattern, with the 500mb average among the ensemble members showing that the GFS should trend back towards the ECMWF. This image shows the average 500mb pattern from the GFS ensembles on the left and each individual member on the right. The white line is the operational individual GFS weather model. You see at this time, again 7 AM on Friday, that the GEFS ensembles (on the left) show a slightly neutral trough moving near New England, though it has not yet moved through, a timing in between the ECMWF and the GGEM. I expect a similar solution to what the GEFS shows actually, with a major snow storm strengthening past the region and throwing back enough snow for significant snowfall even though SWCT will not be in the bullseye. The operational GFS shows something completely different on the right though. Look at the 528 DM blue lines near New England. Almost all of them have dips that are pointed to the east, signifying a negatively tilted trough allowing the surface low to rapidly strengthen. The operational GFS, which again is the white line inside all the blue ones, seems to be pointed fairly significantly in a westerly direction, shunting the potential strengthening of any low pressure. There is not much ensemble support that, and that is a known bias of the GFS to string out low pressure and remain too progressive along the east coast. The ensembles are trying to show that the storm should strengthen faster than the main model shows, meaning that the 12z GFS which only gave the region 4-6 inches is likely inaccurate. And in fact while I was typing and researching all of this, the 18z GFS came out, and while the surface low was only a little stronger, it followed the trend I had laid forecast and now showed anywhere from 8-10 inches across the region, more in line with the forecast. I think this trend will continue with more precipitation on the GFS being around the low pressure center that moves up the coast. In this way, the GFS is already starting to look more like the ECMWF weather model and I expect this trend to continue.
Perhaps the most striking comparison for me has been the NAM weather model, though. The 12z run of the model this morning showed almost no coastal low impacting us. The 18z model run this afternoon showed a major shift where a coastal low strengthens rapidly and just misses impacting us with strong snow. Here is the morning run at 7 PM Thursday. Looking at the top right you see a surface low pretty far off Cape Cod and it is fairly weak, with the central pressure only down to around 100mb. On the top left you see energy finally entering the storm through a trough that has just tilted neutral and is just east of the Ohio River Valley, actually a little further west than you would expect it to be given the surface low placement. But regardless, the surface low is weak and on the bottom right you can see most of the precipitation associated with the surface low pressure center is missing the Northeast. Compare that with the 18z run for the exact same time here. The trough on the top left is broader but is more negatively tilted, helping to strengthen the storm earlier. The low pressure is significantly stronger, down below 996mb. The low pressure is also slightly closer to the Northeast. So stronger and closer to the coast is the trend. You can especially see that with the precipitation on the bottom right, where the heaviest precipitation with the low is just off the coast, and some of it is actually hitting Cape Cod. In one model run, the model made a noticeable jog to the west and was stronger with the low pressure center after not making a defined low pressure center for awhile on previous runs. This is an important development that furthers the trend towards a stronger low pressure closer to the coast. Again, I am not sure it will be as strong, as slow, or as close to the coast as the ECMWF shows, but the storm will likely be more significant than models like the NAM, SREFs, and GFS currently show. If I were to show you the Short Range Ensemble SREFs I would basically be showing you the exact same thing as the NAM, as though they do not show a major storm for Southwestern Connecticut right now, I have a feeling that trends over the next 24-36 hours will have them showing it.
Ironically, one of the most inaccurate models, the NAVGEM, was the one to show the storm as a significant coastal, even before the ECMWF printed it out verbatim. The only reason no one is praising it is because no one was paying enough attention to it yesterday and last night to see that it was, as I typically just look at it for confirmation. The model has huge biases (typically it is too progressive like the GFS), but the fact that it has been consistent with a very strong low pressure center closer to the coast like the ECMWF is yet another signal that other models will be trending that way. I will keep an eye on that. The UKMET weather model, typically similar to the GGEM and ECMWF, is a little further southeast but still shows some solid snow across the region. I expect it to come in line as well.
CONCLUSION: Because of these model trends, I am forecasting a first take of 8-16 inches of snow between midnight/1 AM Thursday morning when the first light flakes may fly until Friday afternoon, which is the latest possible time the back end of the coastal storm may wind down. I believe the National Weather Service will have to issue Winter Storm Warnings because the storm will exceed the 8 inches in 24 hours across their forecasting area that is required for them to issue said warnings. Again, this is a complex storm, and there very well could be a lull in snow during the day on Thursday or Thursday night before the coastal gets going. In terms of impacts, I see the potential for an Early Dismissal or school impacts on Thursday, but impacts on Friday I think will be more likely. Snow will be heaviest Thursday night into Friday morning, so schools have a good chance of closing Friday if my forecast comes to fruition. As for Thursday, there will be on and off light to moderate snow throughout the day, so schools could close as well if overrunning snow looks to accumulate to a couple inches by the afternoon, which is possible. Some districts may also choose to tough it out knowing it is likely they close the following day. I will be following the timing and accumulation forecasts of both of these closely and factoring them into closing/delay/early dismissal forecasts.
With that, I want to end with a disclaimer that none of this is set in stone. I’ve spent the past 6-8 hours looking over model data and trying to discern what I can from the mountain of data that is provided to meteorologists, professional and amateur. This is a highly volatile forecast, like many are, because we have one low pressure trying to transfer energy to another, and we have overrunning precipitation that then turns into a precipitation shield from an approaching coastal low. A jog in a few hundred miles by the coastal low could result in us getting significantly less snow, though as I have said ratios also could be high enough for us to still do well. I do my best not to hype storms, in fact I have been known to downplay a few, but from what I see here there is good evidence that a large coastal storm will likely significantly impact Southwestern Connecticut. Again, by no means do I think we will be the bullseye, but I do think we will see significant accumulations over the 36 hours of the storm. Stay tuned as in future blog posts I will break down more trends, hone in on my accumulations forecast, and focus on other impacts from the storm that could significantly impact the area, such as winds and coastal flooding.