eduggs

That said, I could see the early next week wave turning into a minor event. The wave spacing at least on the GFS and NAM has improved. I'm leaning against it because there haven't even been many individual ensemble members that have bitten, much less an OP run. But we do have a snow supporting atmosphere plus an upper level wave.

I'm not sure I've ever seen the 10 day charts look this terrible this time of year on every model. 573dm at 500mb for days and days and days. It looks like June. There's not really even anything to root for except maybe record breaking warmth. We'd need some kind of massive multimodel error just to make it interesting.

The problem is using analogs to predict local weather. Yes the climate is steadily changing, but that's not what happened here.

Seconded. A lot of really smart people have researched extensively and continue to research all the gotcha questions in addition to legitimately interesting climate change issues.

Proximity to the Ocean definitely enhances our snowstorms. The midwest is a little moisture starved, but the ratios will be better than 10-1 in places like Wisconsin. Michigan should do well, especially with some lake enhancement, and even northern IL could do pretty well.

Anybody with access to the GEFS individual members care to share the approximate southeastern extent of the ULL amongst the individuals at 18z and 0z. The past few runs have shifted this feature slightly further SE.

The Canadian has the day 8 threat.

I think we just need a slightly different orientation of the height field. If that ULL is slightly less suppressive or a lobe retrogrades west and dives into the trof, that would work.

GFS still shifting ever so slightly southeastward at h5. We just need 3 or 4 more like that with no steps backward to make things interesting. The ICON shifted a little as well but the RGEM appears not to have.

Great post Don. You bring so much to these discussions. I think we mostly fail to meet you on your level. My only quibble with your excellent analysis is the your handling of sample size and its impact on confidence and uncertainty. But I don't even have a good solution for dealing with small sample sizes in weather-related statistics other than a big shrug. Sometimes there just might be too much uncertainty to draw conclusions.

I'm with you. It's just funny to look at the raw hourly GFS output and see that unusual weather reversal.

It's going to be an interesting and historical moment on Friday when it's 20F and snowing in DCA while it's raining on the summits of Mt. Washington, Mansfield, and Marcy. I think it's happened before, but still rare.

I was taking another look at the GFS and CMC 500mb charts for Thursday night into Friday. Yes it's a massive cutter with rain to interior Quebec. But if that strengthening and tilting ULL somehow manages to dive 200 miles further SE, the upper level divergence would start interacting with the coastline and we might start talking about a discrete secondary SLP. Any kind of moisture wrapped back into that arctic air would do wonders for the holiday spirit. Most likely any wraparound snow will remain far to the north and west. But the GFS has been gradually shifting the ULL SE the past few runs. Grasping at straws...

The raw model output is/was not showing snow locally. 3rd party vendors process the raw data to generate clown maps. Some do a very crude job at it, but it's not the fault of the GFS in this case - it's tropicaltidbits etc.

I like cold weather for its capacity to facilitate snowstorms. Since this cold likely won't be associated with snow locally, it's the worst kind of cold.

I was wondering the same last night. Hopefully someone with access to the individuals answered it. By now most of those wacky members have probably been weeded out of the suite.

Misleading graphic. There's little, if any actual snow for our area there. I think the rapid movement and extreme temp gradient of the front is playing tricks on the ptype algorithm.

The aftermath of next week's massive vortex looks miserable. Windy and bitterly cold with the continental height field entirely dominated for days.

March snow melts fast. It drips off trees, melting the snow beneath. It becomes dirty snow and mud. December snow lasts. It can form a snowpack. It's festive.

Explain what a meteorologist interprets with respect to model output? What is the reference point for that interpretation? And what does that have to do with wishing the word "pattern" went out of use? You might say, for example, that a model outcome doesn't fit a "pattern." But how is a future model outcome separate from a future modeled "pattern." They are both unknown and both depicted by modeling.

Humans cannot accurately predict future weather outcomes without models. We are entirely dependent on them. Meteorologists are biased by gut instinct, wish-casting, and false pattern recognition. It's why long-term forecasting has such a low success rate. Most meteorologists are not scientists. The constant mistake that is made is assuming that a modeled future "pattern" will come to exist. And then lots of effort is spent analyzing likely weather outcomes based on that assumed pattern. This is illogical and highly problematic for forecasting. "Patterns" aren't specific enough or stable enough to be useful in most mid-latitude regions. And we can't accurately see particular combinations of features, e.g., "patterns", coming very far out. Meteorologists should stop trying to be experts in things they don't understand.

All of that is very non scientific. The use of the word "pattern" should be banned from the lexicon.

The ULL/PV over Hudson Bay with soaring 500mb heights to Northern Quebec and a strong shortwave diving south through the Pac NW just looks so ugly. I guess we can hope that the weak preceding coastal precipitation scrapes us before the cutter, but it feels like a depressing consolation.

Disagree. 1) Our ability to recognize a future "pattern" is completely dependent on models. 2) "Patterns" only exist in a broad sense. They are not intrinsic nor defined to a fine scale. They are just descriptive and statistical human inventions. Our feeble brains try to simplify things into "patterns" for better understanding. But in truth, the weather, characterized by height fields, temperature, pressure, and all the other parameters is unique. Relying on general pattern recognition to very generally predict synoptic scale weather probabilities might be plausible. But anticipating specific weather in a local region based on "patterns" is fruitless.

So by now it seems likely that we'll be dealing with a cutter next week. But we can't forget the obvious lesson of uncertainty, which cuts both ways. The final outcome is still highly sensitive to the height field in key regions. Especially with an amplifying and anomalous trof, guidance could come storming back towards an east coast low. Chances have dwindled as we've used up lead time, but useful uncertainty remains.