Why Snowfall is so Tough to Forecast

One of my favorite things to do after a snowstorm is to look and see where the most challenging gradient of snowfall occurs…in other words, the greatest change in snowfall totals over the shortest distance. This storm provided another good lesson in that. Check out the map below. I’ve plotted some select snow totals from the Mount Holly and Upton CWA’s from the event for Orange County, NY, and Sussex, Warren, Hunterdon, Somerset, Middlesex, Union, Essex, Morris, and Passaic Counties in NJ. Location on the map is approximate, and you can enlarge the map by clicking on it:

Check out some of the disparities. In one instance you go from 22″ to about 11″ over the course of *maybe* 10 miles. In the most extreme instance, Pottersville, NJ in extreme NW Somerset County reported 4.5″ (NWS Co-Op site). Elizabeth, NJ near Newark Airport in Union County reported 32.0″ (trained spotter). The distance between both locations is approximately 30 miles as the crow flies. That’s incredible. It’s common though. In snowstorms, especially ones of this intensity, you get intense areas of what we call mesoscale banding…bands of extremely heavy snow (sometimes with thunder) that impact small areas. To take this to another level, which I won’t do, you can also have something known intriguingly enough as “CSI,” or conditional symmetric instability…or slantwise convection. You often get this in intense snowstorms. I won’t bore you with details right now, but if you’re interested, a whole website is devoted to the topic here. It will be interesting to see the number of papers that I’m sure will come out in a year or three about this storm and the ones of last winter. Anyway, this basically means, some areas get absolutely walloped, while others get shafted.

Anyway, back to the topic at hand, you could probably replicate a map like this across southwest NJ and in parts of New England too. But this just goes to show you, for meteorologists making a snowfall forecast map, these are some of the complexities that we all have to deal with.

Lastly, the poll question continues. Do you all prefer the current system of occasional blog posts, or would you rather have occasional blog posts and a Facebook “like” page with occasional links/discussion/question areas? Vote below, and we’ll see where this goes.

Advertisements

One for the Haters!

So not everyone loves snow. As a meteorologist, even if I dislike dealing with it (working four years in lake effect country will do that to you), I love the unique quality of events like this and appreciate what has to come together atmospherically for something like this to happen. It’s not THE perfect storm, but it’s one of the most perfect storms. But some people can’t deal with it, hate it, curse it, etc. And that’s fine. So I dedicate most of this post to you.

6-10 Day Average 850mb Temperature Anomalies from the GFS - http://raleighwx.americanwx.com

At least temporarily, there’s going to be a pattern change in the East. Once the energy and upper level trough from this storm lifts away, we’re finally going to see  some changes to allow the seemingly endless cold in the East to abate. Strong troughing is going to build into the West as the pattern sort of blocks up out here (a couple cutoff lows to allow for some of the coldest dry weather we’ve had here in some time). This will help allow a ridge to pump up in the East. There are still some questions as to how strong the ridge in the East will get and how long it will last, but at least as we go into the New Year, things begin to quiet down and thaw out a bit…and a lot of the snow that’s falling will be long gone. Enjoy it while you have it. The map to the left above, from http://raleighwx.americanwx.com, shows the average 850 mb temperature anomalies off the GFS model 6-10 days from now. It’s not a perfect representation, but you get the idea…cold West, warm East. It’s somewhat of an odd pattern, and truthfully, given the fact that blocking in the Arctic and Greenland wants to hold on, I don’t know how long it will last. But it is more “normal” for a La Nina pattern, at least temperature wise.

So truly an awesome storm unfolding. Storm totals are outrageous, 12-20″ widespread from Norfolk, VA-New England, and 6-12″ in parts of interior VA/NC, and even 1-3″ or so on the Outer Banks! It snowed as far south as Savannah, GA and Jacksonville, FL this morning. Pretty remarkable. I may do a wrap up post later, but pass on your totals if you’ve got them in your backyard!

Lastly, I’m curious if people are content with this format of either RSS/Google Reader, getting a link on Facebook, or email subscribing…or if you’d also like see a Facebook “Like” page for more short-form updates before and during events and occasional links of interest. Vote below and we’ll see what sorts of changes we can make.

Also, enjoy this video of the blizzard from NYC, including thundersnow around 1:45-1:50:

The Boxing Day Blizzard

Well, it’s underway, and miraculously, there haven’t been any dramatic changes in the forecast in the last 12 hours. The main areas of concern to me seem to be:

  • Cape Cod/Providence/Boston/Long Island: How much falls as sleet/rain vs. snow, thus cutting down snow totals there. The low should ride the coast from about Montauk to Martha’s Vineyard, so this will both lead to a mix/changeover scenario on Cape Cod/southeast Connecticut/Rhode Island and may even have to a dry slot cutting down totals in general. Will this make it as far north as Boston? Doubtful, but we’ll see.
  • Jersey Shore from LBI south: There could also be some mixing issues here. In my experience there are almost always mixing issues in this region when a low passes this close to the coast. So this may cut down totals, and I’ve lowered the range a little to 7-14 (down from 8-16). It’ll still be a good storm, but the combination of a little less precip and this mixing scenario has me cutting back a smidge.
  • Philly-Morristown-Albany: The biggest question of all is regarding how far west the heavy snow will get. And this corridor stands the most to gain or lose depending on the exact amount the precip shield barges inland. I could see this being as little as 6″ in Philly to as much as 12-16″ there depending on the exact track (same goes for the I-287 corridor in Jersey and I-87 from Newburgh to Albany.
  • NYC-Hartford-Interior New England: This could be an absolute wallop for places like Worcester, the Berkshires, the Litchfield Hills, etc. A lot of potential for a lot of snow in these areas, and there could be a cluster of significant 20″+ amounts anywhere between NYC and Boston depending on where the best mesoscale banding sets up.

Lastly, the wind impact from this cannot be stated enough. This is a storm which is going to see its barometric pressure fall from about 996mb off Cape Hatteras to perhaps 970 mb or lower south of Long Island. In meteorological terms, we call this an atmospheric “bomb.” In other words, it’s going to be rapidly deepening, and it will be progressing along close to shore. This is going to create absolutely ferocious winds all the way up the coast and rather far inland. With everyone’s Christmas lights up, coupled with heavy snow, the wind has the potential to cause a LOT of damage/mess. It may not be a bad idea to secure some of the holiday displays Sunday morning. There will also be the potential for significant power outages as well, so be prepared, especially because temperatures will drop into the low 20s to teens Monday night and Tuesday morning, with wind chill on top of that.

Travel is going to be next to impossible and is not recommended, especially from Northeast NJ into New England after about 9 AM-Noon tomorrow. And it may be that this is the case throughout most of eastern PA/NJ/DE/eastern NY as well. This really to me seems like last February, shifted a couple hundred miles further north. So please use caution if you have to travel, and be prepared for numerous road closures/states of emergency by Sunday evening.

7:30 AM Sunday Update: I would likely trim some of the snow totals back east a bit in the 3-7″ range, esp in NY/PA. The storm is moving along at a decent clip. The real intensification takes place this evening…that’s when the “blizzard” aspect will begin to take shape. In terms of the actual amounts, I might be apt to trim back south Jersey/Delaware a bit (more like 6-12 now to me). Overnight models (not going to read into them too much at this point) did cut back snow totals there to the 6-10″ range). Still looks good that the bullseye is between NYC and Boston..but not IN NYC or Boston. Boston may end up on the lower end of my forecast range, but inland areas will push the higher end. The western extent of the snow between the Delaware River and the Hudson River is going to be EXTREMELY sharp now it appears. Could easily go from 2-4″ far west to 12″+ over a very short distance. Whether that gradient is centered over the Delaware Water Gap or Denville in North Jersey, I’m still not sure. Still a tough forecast. Keep an eye on things throughout the day. Also, from northeast NJ into New England, the bad travel may not start until more like 11-2, rather than 9-12.

1:30 PM EST Update:  Everything still on track. Morning model guidance and current radar trends validating a major storm that’s underway. Latest SPC Mesoanalysis showing that the storm is rapidly deepening off the NC coast as it gradually pulls north. This will bring blizzard conditions to most of coastal NJ/Delmarva and inland to about Rt. 206. The blizzard conditions will spread north into NYC and New England this evening. The worst of the weather will be bounded on a line south of I-90 in New England and east of I-287 in NJ-I-87 in NY.

If you live in Southeast Jersey, snowfall rates of at least 1″/hour are occurring and those heavier bands will continue to push onshore from time to time this afternoon and evening. Heavier snow continues to gradually push inland, and to be honest, the snowfall map from last evening doesn’t look too bad at this point (though I’d still shift the 4-8″ area a little further east). Some totals so far include Somers Point at 6″, ACY at 4.3″, a little over 3″ at Ocean City, MD, and 6-10″ in the Hampton Roads area of Southeast VA and parts of Northeast NC, back through Raleigh/Durham. The DC/Baltimore areas have been for the most part “shut out” of the storm so far, and it’s not looking likely that those areas will be hit nearly as hard as others north and east.

Travel and conditions in general will continue to deteriorate this afternoon. Be safe!

Here’s your map for new snow tonight-Monday (click to make it bigger). Feel free to add any questions in the comments, and if you like the updates or have any comments/ideas for future items, let me know! And Happy Boxing Day!

Final Snowfall forecast for Boxing Day Blizzard (click to enlarge)
The storm is moving along at a decent clip. The real intensification takes place this evening…that’s when the “blizzard” aspect will begin to take shape. In terms of the actual amounts, I might be apt to trim back south Jersey/Delaware a bit (more like 6-12 now to me). Overnight models (not going to read into them too much at this point) did cut back snow totals there to the 6-10″ range). Still looks good that the bullseye is between NYC and Boston..but not IN NYC or Boston. Boston may end up on the lower end of my forecast range, but inland areas will push the higher end. The western extent of the snow between the Delaware River and the Hudson River is going to be EXTREMELY sharp now it appears. Could easily go from 2-4″ far west to 12″+ over a very short distance. Whether that gradient is centered over the Delaware Water Gap or Denville in North Jersey, I’m still not sure. Still a tough forecast. Keep an eye on things throughout the day.

Yes, This is a Christmas Miracle

Irony of ironies, almost all the models are converging on a solution. There are still uncertainties on this storm, especially regarding how far inland the significant snow goes. One model is suggesting it goes almost all the way to Pittsburgh. Another says it barely crosses the Delaware River. So that will be what needs to be worked out over the next 12-18 hours. Not going to be a very relaxing Christmas for many meteorologists in the East.

This storm has truly been remarkable to watch. It’s been all over the place, and there are a number of very well seasoned meteorologists I’ve heard from that say they have never seen anything even remotely this poor in their careers. This just serves to remind everyone that weather is a VERY inexact science…and it will always throw us curve balls. I’ve included an updated snowfall forecast map below, and like I would encourage you with any snow map you see today and tonight, take it with a grain of salt. There’s uncertainties that still cloud this forecast. The bottom line: There will be a high impact storm for most areas north and east of Baltimore. Regardless of specifics, this will be significant for a number of areas, and travel within and to the Northeast and Mid-Atlantic is going to be difficult to perhaps impossible on Sunday afternoon through Monday afternoon. There will be mixing issues along the coast of NJ, Long Island and Southeast Massachusetts. Those mixing issues could even extend inland a bit too, depending on the exact track of the storm. Any mixing will cut significantly into snowfall totals in those areas. There is the potential for higher amounts in the eastern areas of the blue color (4-8) region, again, depending on the exact track. There will be blizzard conditions when it’s snowing on the coast and possibly a bit inland.

This is a big storm, so despite the uncertainties, doubts, and changes you’ve heard of, and the jokes you may be inclined to make about “those weather people,” do take it seriously, as it will be a high impact storm that could still throw curve balls. And if you hate snow, it looks to warm up a bit with an early January thaw soon after this storm.

Click the map to enlarge.

Revised Snowfall Forecast for Christmas Weekend-Monday

It’s a Christmas Miracle? Flip a Coin.

Just a quick update. I won’t be drawing a snow map this evening for a couple of reasons: One, it’s Christmas Eve, two, the Euro won’t be out for another hour+, and three, the amount of inconsistency and volatility regarding this storm is something I’ve heard people with years of forecasting experience beyond my own, say they have never ever witnessed. In other words: You’re looking at one of the most complex storm systems we’ve ever seen…moreso because of the location it may or may not impact rather than because of the storm itself.

But I will provide you with a couple of major teases. The two maps I place at the end of this entry are raw snowfall forecasts from the GFS and NAM computer models from this evening. Now, don’t go shouting from the rooftops. I STILL am not behind this storm. I still see flaws. I still think the models are suffering problems (given their wild swings and disagreements over the last 2-4 days or so, this may seem obvious). And I still think we’re in for changes here in the final 36-48 hours before gametime. The Hydrometeorological Prediction Center’s model discussion this evening is once again hinting at the potential that the models are erring, and that could be leading to solutions later in their runs that are, well, incorrect.

CONSOLIDATING UPPER LOW OVER THE EAST...
LOW PRESSURE FORMING UP THE EAST COAST...PREFERENCE FOR A 00Z
NAM/15Z SREF MEAN COMPROMISE
THERE ARE SMALL SCALE DETAILS THE NAM/GFS/ECWMF ARE SHOWING ISSUES
WITH WHICH MAY IMPACT ITS FORECAST LATER ON.  THEIR SOUTHERN
STREAM SHORTWAVE IN TEXAS ARE 30-40 METERS TOO WEAK AT THE 500 HPA
LEVEL BASED ON RAOB REPORTS FROM TEXAS.  THE NAM DOES NOT CAPTURE
THE ONGOING CONVECTION IN EASTERN TEXAS.  THE 18Z AND 00Z GFS RUNS
ARE SHOWING SIGNS OF CONVECTIVE/GRIDSCALE FEEDBACK AS THE SYSTEM
MOVES ALONG THE GULF COAST WHICH COULD BE MAKING ITS SOLUTION AT
THE SURFACE AND ALOFT TOO STRONG.  THERE ARE HINTS THAT THE RIDGE
INITIALLY ALONG THE EAST COAST IS MORE AMPLIFIED THAN THE
NAM...WHICH WOULD BE CLOSEST TO THE 12Z ECMWF/12Z ECMWF ENSEMBLE
MEAN FORECASTS FOR 00Z.  ANY OF THESE DETAILS COULD
DEGRADE/NEGATIVELY AFFECT THE SHORT RANGE FORECAST.

So there’s still much uncertainty. I mean, honestly, there’s NO forecast out there now that stands much chance of being right on, including my own. This is simply one of those storms where you sit back, roll the dice, and pray you come close.

Consider a compromise of the maps below as the potential for the maximum impact of this storm. Click on either map to enlarge it. Consider my forecast map from earlier this morning the potential for minimum impact of this storm. So enjoy the tease, because tomorrow is going to tell us that either a hefty dose of reality is going to kick in, or you really will believe in miracles. Merry Christmas!

00Z NAM Forecast Snow: http://wxcaster4.com (Worth bookmarking)
00Z GFS Forecast Snowfall: http://wxcaster4.com

Boxing Day Storm: TKO

As if dealing with some of the intricacies and inherent uncertainties of the weather wasn’t enough, you occasionally have days like today, where everything suddenly shifts and it looks like that monster storm you discounted yesterday may not be dead after all. And you start to wonder what has happened in 12 hours to completely change your thinking. And then the Hydrometeorological Prediction Center issues its discussion on the models and says this:

PREFERENCE: EQUAL BLEND OF THE 06Z GFS/00Z ECMWF OR THE ENSEMBLE
MEANS

INITIALIZATION ERRORS IN NUMEROUS DIAGNOSTIC
QUANTITIES...INCLUDING HEIGHT/VORTICITY FIELDS/RH...ARE EVIDENT IN
BOTH THE 12Z NAM/GFS WITH SMALL BUT LIKELY SIGNIFICANT SHORTWAVE
TROUGHS OVER SOUTH DAKOTA/NEBRASKA ALONG WITH
SASKATCHEWAN/MANITOBA...WITH THESE AREAS ALSO NOT PARTICULARLY
RESOLVED OR PREDICTED WELL BY THE 00Z ECMWF. THUS...THE SPECIFIC
PREDICTIONS BY ALL DETERMINISTIC GUIDANCE ARE IN QUESTION...WITH
THE RECOMMENDATION TO FOLLOW CONTINUITY...WITH THE FINAL OUTCOME
MOST BELIEVED TO LIE BETWEEN THE 06Z GFS AND 00Z ECMWF...WITH ALL
ENSEMBLE GUIDANCE INCLUDING THE SREF MEAN/GEFS MEAN/ECMWF ENSEMBLE
MEAN ALSO CONSIDERED USEFUL TO ADDRESS THE CONTINUED UNCERTAINTY.
THIS APPROACH DISREGARDS THE SUBSTANTIALLY DEEPER AND WESTWARD
SHIFT OF THE 12Z GFS REGARDING THE POWERFUL LOW TRACKING UP THE
EASTERN SEABOARD...AND TO A LESSER EXTENT THE 12Z NAM WHICH LIES
NEAR THE FAST EDGE OF THE GUIDANCE WITH THE DEVELOPING LOW

 

In other words, trust NOTHING. In reality, it’s more or less that there was a problem initializing both the NAM (short range model) and GFS model that was serious enough to make them believe that their solutions showing a much further west and stronger storm should be discounted completely. Basically, when the model is run, it’s run off of a snapshot of the atmosphere taken at the beginning…the initialization. If that’s significantly incorrect, you better believe the rest of the run of the model is also incorrect. So you literally trash it.

That said, there still will be a storm, and it still could be significant for a couple places, primarily southeast Massachusetts again, as well as potentially southeast Virginia and northeast North Carolina. There’s really not much else to say, except you’ll just need to keep an eye on things to make sure they don’t change…a 50-100 mile shift either way will shift those already tight banded snowfall amounts in one direction or another. But frankly, I’m not too worried about this one. So your post-Christmas return rush and Monday for the most part look okay, unless you live on Cape Cod. Forecast map is below (the Comic Sans font for the title was done intentionally to emphasize how much of a joke this storm has been)…not planning to do a new one unless there’s a significant change. I was a little generous with how far inland the “coating-inch” zone was. That’s mainly to account for remaining uncertainty. Many areas may just see flurries that don’t really accumulate. Only area I didn’t really look too terribly closely at is SE Virginia. Really seems to be a toss up there as to how much will accumulate depending on timing and track…could be more or could be less. We’ll see. Merry Christmas!

Snowfall Forecast Map for the Boxing Day Coastal Storm
 

PREFERENCE: EQUAL BLEND OF THE 06Z GFS/00Z ECMWF OR THE ENSEMBLE
MEANS

INITIALIZATION ERRORS IN NUMEROUS DIAGNOSTIC
QUANTITIES...INCLUDING HEIGHT/VORTICITY FIELDS/RH...ARE EVIDENT IN
BOTH THE 12Z NAM/GFS WITH SMALL BUT LIKELY SIGNIFICANT SHORTWAVE
TROUGHS OVER SOUTH DAKOTA/NEBRASKA ALONG WITH
SASKATCHEWAN/MANITOBA...WITH THESE AREAS ALSO NOT PARTICULARLY
RESOLVED OR PREDICTED WELL BY THE 00Z ECMWF. THUS...THE SPECIFIC
PREDICTIONS BY ALL DETERMINISTIC GUIDANCE ARE IN QUESTION...WITH
THE RECOMMENDATION TO FOLLOW CONTINUITY...WITH THE FINAL OUTCOME
MOST BELIEVED TO LIE BETWEEN THE 06Z GFS AND 00Z ECMWF...WITH ALL
ENSEMBLE GUIDANCE INCLUDING THE SREF MEAN/GEFS MEAN/ECMWF ENSEMBLE
MEAN ALSO CONSIDERED USEFUL TO ADDRESS THE CONTINUED UNCERTAINTY.
THIS APPROACH DISREGARDS THE SUBSTANTIALLY DEEPER AND WESTWARD
SHIFT OF THE 12Z GFS REGARDING THE POWERFUL LOW TRACKING UP THE
EASTERN SEABOARD...AND TO A LESSER EXTENT THE 12Z NAM WHICH LIES
NEAR THE FAST EDGE OF THE GUIDANCE WITH THE DEVELOPING LOW.

 


Models Throwing Punches for Boxing Day Storm

Maps from Sunday and Monday Morning from this morning's European Model: http://raleighwx.americanwx.com

I’m still not ready to commit entirely to a major storm yet for the I-95 cities or the coast, but this morning’s Euro run is holding serve awfully steadily. The Euro, as you can see to the left (click to enlarge), is an utter monster. Keep in mind, this model has little to no support for a solution like this, but it has also been quite consistent. Its ensemble mean is a bit further away from shore, but this would still imply a good hit for a lot of locations. The GFS is a big storm, but just grazes the coast and hits far Eastern New England hard. The Canadian model is somewhere in between the two models.

Again, I will emphasize that the pattern out west with a ridge axis centered over CO/WY is not exactly ideal for a major East Coast storm for the Megalopolis. Yet, something is going on here to allow it to happen on the Euro. The Euro model solution is absolutely incredible. It implies the potential for heavy snow as far south as Jacksonville, FL and on up the East Coast through New England, with the heaviest snow likely in interior VA-NJ/PA and southern New England.

My take:

  • Odds for a storm to impact the coastal areas are equal or greater than it was this time yesterday.
  • Odds for a storm to impact the interior/Megalopolis are about the same as it was yesterday; maybe slightly higher.
  • Odds for an epic blizzard like the Euro model is showing are very low to near zero in my opinion.
  • But…the odds for a significant plowable snow in a lot of areas are slightly higher than yesterday.

Tonight’s model runs will be the first real key, as the main energy involved in blowing this storm up is now onshore here in California. If tonight’s Euro shows something similar to this morning, and/or the GFS model trends closer to the Euro model, we have to think odds are increasing for a big Boxing Day Knockout. If the Euro trends toward the GFS tonight, I think the chances of a major, high impact storm drops substantially. I can’t emphasize this enough: The Euro stands alone right now…it CAN score a coup; it has before, but it also has not been its usual stellar self this winter, which is why I sit here still skeptical. If the models trend toward any given solution on consecutive runs both tonight and Thursday morning, I think the odds of whatever that solution is increase multi-fold. So the next two main model runs (7-10 PM tonight & 7-10 AM tomorrow) are absolutely key in all this. Stay tuned.