psuhoffman

Btw anyone want to chime in about something not getting much talk but that’s equally part of our recent snow issues. The nao since the 1960s is trending more positive. There are up and down cycles but since the 60s each positive is more positive and each negative is less negative. The fact is our winter months are now in a positive nao state about 80% of the time. That’s a problem for our snow. I have a theory but I kinda want to hear other hypotheses

@Terpeast I’m also glancing at pre 1950 data and it seems to support my idea of analogs working better if you can superimpose the larger hemisphere cycle. But it’s problematic because take a similar cycle in the early 1900s. First of all there is evidence the PDO is trending more negative during negative cycles lately. But also that was a much colder regime. So a 9 year period that produced a 17” avg at BWI seems ok but that’s coming off a period when BWIs avg was around 27”! That 17” would be more like 12-14” today. And if the PDO is in fact increasing its variance maybe that knocks it down to 10”? But in general I think the evidence is it’s really hard to get a truly snowy winter in this type cycle The only catch to this method is we can’t predict 100% when the cycle might flip and if it does quickly and unexpectedly obviously you bust on the edges of longer decadal cycle changes all that to say if the solar cycle peak lingers and the PDO cycle does flip (we are kinda due) suddenly a year like 2014 isn’t crazy

3 analogs were bunched between 10-14” with one high and low outlier (18” & 0”). I’m tempted to toss the outliers but the 0 is also the most recent analog (22-23). That’s troubling. Once we get to November I’ll try to discern and rank these analogs more accurately. This was just a prelim identification of where we are now. But unless I see signs we are breaking out of the larger long wave pacific cycle we’ve been in since 2016 our upside is severely limited. There have been 3 similar comp cycles (including now) to this one. Defined by me as deeply consistent -pdo overlapping a predominant +nao comprising 23 seasons. I tend to focus much more on the long wave pattern identified by seasonal mean h5 that indexes. But in all those years the largest snowfall seasons were 2 that made it into the 20s. But those both were Ninos. Among all the neutral and Nina seasons we’ve never cracked 20” at BWI during the 3 recorded similar cyles. So the only thing that would significantly alter this would be if there were signs the larger background state were about to change this winter. I don’t see signs of that yet. It could move some of the Nina looks more neutral and that brings some other analogs higher. But the only thing that could significantly improve our expectations would be this most recent northern hemisphere long wave cycle ending. It’s damn near impossible to get a truly snowy winter in this phase.

I have been experimenting with finding a way to superimpose the long term cyclical phases of the pacific and high latitudes over other variables in seasonal forecasting. My methods have boiled down to taking the long term mean 500mb phase of the central pacific, AO and NAO domains and using that to weight analog years produced by other variables. This method significantly improved the ability to forecast the coming winter. For example: last year this new methodology would have given me 4 different top analogs. At BWI the result of the snowfall for those 4 seasons after applying a sliding reduction to older analogs for warming would have given me a predicted BWI snowfall of 10.5". The actual total was 11.3". I found the results to be similar for most years, not to that extreme, but the prediction became much closer to reality when applying this method. However, there is one HUGE catch... we cannot always predict when a long term cycle is about to change. So every once in a while, when a phase change happens during a winter season, it would produce a huge error. As of right now this method has a predicted snowfall at BWI of 10.2". I will update this again sometime in November with an actual seasonal forecast and explanation. Some of the factors and variables could still change some.

Maybe my memory is bad but weren’t they really only about 30 miles west of the other guidance with Helene?

If this exact same look leading into winter was happening during a less hostile pacific and Atlantic cycle I’d be way more optimistic. But one been doing some work on incorporating the long term cycles to other more common analog metrics and it seems applying this method retroactively would have made my forecasts much more accurate. The same factors don’t lead to the same results when supplanted into different cycles. I’m not optimistic that the same set of factors that lead to 2014 would have nearly the same impact now given we were in a generally favorable pdo then and we are in about the worst possible now. That said I don’t expect a total dud either. I think we have enough in our favor (I do agree that the signs indicate a more poleward pac ridge) that we should get some snow. I expect more of a just typical crap v omg this is god awful crap. Something like 2018 and 2022 maybe.

Translated he basically said “we could get a pattern that could lead to either one of our snowiest or least snowy winters ever”. Thanks.

But did you read the thread? He includes 1989 and 2014 in the same analog group. What good is that. One was a dud way below avg year and the other one of the snowiest ever. That’s a completely useless use of analogs.

There was one other moderate storm that just missed also. Those both hit and we remember 2011 as another 1996 type Nina. It can happen. And both of those misses were flukes not typical Nina issues imo. The Dec storm was a gulf stj storm NOT a late miller b but it had a sloppy weird phase at the worst possible spot in its progression. But I am skeptical of getting that cold of a winter in the current pacific cycle.

And Ya'll thought I was depressing

Yup, but while its true, that is kind of a typical issue, and has been, when the mean trough axis is always a little too far to the NW, which even in the "better" looks is still true in this base state. Periods of high amplitude likely lead to storms cutting way to our west...and in lower amplitude we are dealing with a bunch of weaker system running the boundary interfering with one another...but that at least gives us a chance for some snow. Big storms are going to be very hard to come by given the larger scale issues with this base state. Best chance might be getting really lucky with a progressive wave that is simply ridiculously juiced up. We saw that just south of DC in the Jan 2022 storm.

For reference...there is only so much variability to the pacific under this current base state and its bad, so really we are talking about just "kinda bad" v "awful" here...but this was the mean longwave pattern during the 3 snowiest winters of the last 8 years... This was the 3 least snowy winters... Neither of those looks good, which is why we've had no truly snowy winters lately, but the key difference in the "snowier" years was the more poleward extent of the pacific ridge shifting the trough out of western Canada some which allows more cold air transport into our area during transient periods. The less snowy look there is just the total kiss of death with the flat pacific ridge allowing the trough to really get stuck out west and the ridge to go ape in the east and its game over on a seasonal scale!

I mean what are historically "better patterns" are still producing more snow. What little snow we've had over the last 8 years did tend to come during the periods when we had a less hostile pattern. We have not had much luck during the "shit the blinds" patterns. If anything we have had even less luck during those than normal...its almost impossible to luck into snow in a bad pattern now because they are so warm now. The problem imo is we've spend less time than normal in a "good" pattern over the last 8 years, and our results in a bad pattern went from bad to god awful no hope at snow at all due to warming so... Also perhaps the results in the good aren't "as good" but I would want to see more data on that, we've had so few legit good long lasting patterns recently its hard to say that with any confidence. In the end we still want some combo of -AO,-NAO,-EPO,+PNA and we really need 3 out of those 4 to have a good chance at snow. And even if we get 3/4 the 4th cant be like historically god awful or it can still offset the other 3. That has always been true. We just haven't had that much lately because the pacific has been stuck in a god awful phase where the PNA is killing us most of the time.

It's too small a sample size to draw statistically significant conclusions, but just looking at the set of neutral or extremely weak nina winters over the years...it seems that perhaps they take on the flavor of the overall longer term cycles we are in. When we were in a more snowy cold period in general they tended to be better...and when we were in a warmer/less snowy period they sucked. If this is true...it would explain why the last few sucked.

That is totally logical and why most assume a stronger nina is worse, but the data just doesn't support that conclusion. Even if we go back 75 years there is almost no difference wrt snowfall between a weak and a strong nina. And recently weak nina's have been worse. Also, the other popular myth that a fading nina will save February/March is not true, the data indicates winters where a nina fades to neutral actually has less snow in Feb/March. I know you are asking why, but I just don't know. We are still talking about a small sample size, so many it is just plain luck and over a thousand years it would balance out and in the end a weaker nina is better...Kind of like we just got 5 coin flips in a row recently odds...but no one can say that for sure. On the flip side maybe there are subtle variables that are changed by a stronger nina. Maybe a stronger faster pacific jet is really bad at some level but if strong enough causes enough variability such that it becomes less hostile at times? I know later in winter nina forcing actually becomes more favorable once wavelenghts shorten, so are we are less likely to get a nice ending like 1999 and 2018 in a weaker nina possibly. I honestly don't know, I am just going by the data.

I wonder if we actually want a stronger nina given the options. There have been 2 examples of an enso neutral DJF following a nina during this recent pacific cycle, 2017 and 2020, and both were total dud winters. And the weakest actual Nina DJF we've had was the other total dud winter 2023. The stronger Ninas were actually significantly better. I see a lot of people that seem to be rooting against the nina in the enso thread but if we want snow I'm not sure that's really wise.

That checks out, if I get to 30" up here I will do cartwheels, and that's probably about the same % of "climo" as 15-20" for your area. Frankly if you offered me 25" right now, take it or leave it, I would take it and run. That would be unthinkable 10 years ago, but I am scarred and shell shocked by the frequency of total dud winters even up here recently.

@Eskimo Joe One last thing, more than once it's been asked "are big years like 2010 masking the true extent of the degradation some" and that is a subjective thing...depends on perspective. But variance is increasing, and the probabilities of a single digit snowfall season have gone up exponentially in the last 30 years. But maybe the best way to look at it is since 2000 IAD's average snowfall is 19". That's bad, and well below the normal of around 24" before that. But if we take out just 3 seasons, 2003, 2010, and 2014, the average is only 13" which is must more indicative of what we are likely to get in any given year that the 19" which is incredibly skewed by those 3 huge years in the last 23 seasons. What happened in the 20/23 is much more "what we should expect" in any given year. In past periods there wasn't as big a gap between the mean and median snowfall. But it could be argued that we are using both outdated and unrealistic expectations for what "normal" snowfall is around here now. Truth is "normal" is probably closer to 8-15" area wide than the numbers most of us have in our head.

But what do you consider "climo"? When I moved to northern VA in 1993 IAD's "climo" was 24". If we continue to get similar results the rest of this decade to what we've had recently the new "climo" would be around 17". The problem is during a shifting climate a 30 year mean is always behind the curve. Same with temps...at the rate we are warming it is incredibly difficult to get an extended stretch of "below normal" because the normals we are using are already out dated and the new baseline status quo has warmed.

I believe so. Precipitation events are also increasing so when we do get a rare snowy year they can be big. Not only is our mean and median dropping but variance is increasing. From 1996 to 2016 we were getting more 10”+ snowstorms even while getting less snow overall! We were getting some of our snowiest winters while getting less snow overall. The big red flag to me was that places that used to be correlated to our snow decoupled. NYC started getting a lot of above avg snow seasons while we were below. There started to be a very sharp cutoff just to our north. Kinda like used to happen just to our south! I mean even in a snowy winter at some point you’re just too far south for it to matter. IMO that line was shifting north on average. Because we use current averages as a baseline it masks this some because we’ve already lost a significant % of our snow climo. I think this is the best way to illustrate it. From 1996-2016 the mean pattern was this! look at that. Over 21 years that was the average base pattern. But at DCA only 6 years of the 21 were above avg snow! And this was the averages at the 3 reporting stations DCA: 15.9” IAD: 23.7”. BWI: 22.9” Now if we compare that to what’s happened since that looks like snow Valhalla. But the 30 year mean at all 3 airports going into that period was 18.5”, 23.6” and 22.3” The mean avg across all 3 stations going into 1996 was 21.5” but from 1996 to 2016 with the best possible long term cycle we could hope for the seasonal avg was 20.9”. The mean avg snowfall across the area was actually slightly less during those 21 years than the 30 years before despite being what should have been a very snowy period which was very snowy…just to our north! I think the “you’re too far south for the pattern to matter” line has been shifting north for a while and even once the pattern gets better there is a chance on some years it’s still to our north.

Except from 2000-2020 while north of us was setting records left and right we were actually slightly below avg (old averages anyways) for the period. Our snow climo was already degrading even during the favorable pattern we were in during that period.

I don't know about the bad luck part. Sure we've had opportunities that didn't work out...but I don't feel like the fail rate has been higher than is typical. What if some of the few snow events we did get had NOT worked out! Numerically I think we've had about what we should expect from the status quo longwave pattern of the last 8 years. I say that because there have been 2 comparable historical periods with a very similar predominant longwave pattern. The 1950s and 1970s. And while the current snowfall results have been worse than both, they have only been about 15-20% worse depending on the location and that is what several regression studies indicate can be expected due to warming since those two periods. So IMO our snowfall results have been what they should be if you factor in the pattern and adjust expectations down from past similar periods due to the warmer climate. See below, if you adjust for the colder climo of the 1950's and 1970s...the current ass pattern matches pretty closely with the ass patterns of those two prior periods, and the results when adjusted for warming are exactly what should be expected. Past Ass Current Ass Until this current pac dominant pattern breaks we probably shouldn't expect much deviation from what we've experience since 2016. In our specific location that is below normal snowfall...with the only debate being how much below.

That pattern is better than 2020 or 2023. It’s not a total snowless dud type pattern. But it’s not a snowy pattern either. It’s likely something like 2018/2021/2022. Maybe in the new normal that is snowy lol.

Sorry if I didn’t articulate it well. I agreed with your analysis of that prog. It wasn’t awful, about as good as we could expect in a Nina. I’d take it. I was making a 30,000 ft view comment not specific to that plot. And yea in theory the flow should matter more. Unless it’s just warm all over which has actually been the case too often. This has been the reality the last 9 years. and our snow had been what you would expect in those temps! Sure we could get lucky and someday we get a cycle where the pattern puts that little tiny blue dot that’s over BC over us instead. But the probabilities of that when 90% of the hemisphere has seasonal temp anomalies that wouldn’t support a snowy winter here isn’t very good regardless of the pattern. Just for reminders this is what a snowy winter looks like. The last 5 above normal snowfall winters at BWI That’s what snowy looks like. It’s been a long arse time since we saw those kind of seasonal negative anomalies anywhere in the mid latitudes. But that’s what we need! Now this doesn’t mean we can’t get any snow. And I’ll be tracking as always looking for that fluke that might happen on an otherwise hostile regime. I’ve recalibrated. But I doubt if we can get a seasonal anomaly that looks like our “snowy winters” one with current SST profiles no matter what the specific long wave pattern details are. We’re rooting for one off fluke events imo.