January 16-January 31, 2005 Pattern Discussion

[donsutherland1]

Last week’s discussion raised a number of issues of what one might expect for the January 9-22, 2005 period.

• In the East, the unseasonable warmth has defined the opening days of January. Even as New England has been able to experience brief periods of cold along with snow, readings have averaged above to much above normal along the East Coast. That situation is likely to continue into the second week of the month…

The East should see readings continue to average above normal from Richmond to Washington, DC through January 14. During that period, it is possible that the temperature might reach 60° on one additional day as far north as New York City. Philadelphia, Washington, DC and Richmond should see readings into the 60s on one or more occasions. A conservative estimate of the highest temperature in each of those cities for the January 9-14 period would be:

Boston: 57°
New York City: 60°
Philadelphia: 63°
Richmond: 71°
Washington, DC: 66°

All of the above cities saw the temperature peak above 60° during the height of the warmth ahead of the strong cold front that brought an end to the unseasonable warmth.

Actual Highest Temperatures:
Boston: 63°
New York City: 66°
Philadelphia: 66°
Richmond: 75°
Washington, DC: 71°

• [T]here are growing signs that a significant pattern change could be unfolding, especially around or just after mid-month… While there may be doubts about whether such a pattern will be sustained, a look at the historic experience given similar ENSO regional profiles argues that the pattern change will occur and be sustained… A pattern change is nearing and it should ensure that the second half of the month sees a predominant trough in the East/ridge in the West configuration. Also, from the December 31 discussion for the January 1-15 timeframe: given the changes occurring over the Pacific Ocean, there are some indications that the NAO could be trending downward toward the end of the first week in January and perhaps reach negative January 11 +/- 3 days.

Evolving changes in the global indices (AO: went negative on January 13; NAO went ever so slightly negative on January 14; PNA is rising) and model guidance now strongly support the significant pattern change with the idea that the cold front that blasted off the East Coast today marked the shift in the East. Longer-range modeling and the GFS ensembles concerning the global indices also suggest that much of the rest of January should average below normal but there has been a reversal of the trend and now the Ensembles point to a return to a negative PNA and are split on the sustainability of a negative NAO. This situation may yet have to do with a trough that might erode the positive PNA but not necessarily kill it.

Hints of retrogression both on the 5-day moving average of the 500 mb height anomalies and day-to-day averages suggest that the NAO might well be sustained at negative in the longer-term with the potential for considerable high latitude blocking down the road.

For now, the idea that the January 16-31 period will average below normal in temperatures in the East still looks good. This does not mean that for short periods of time the temperature cannot rise to somewhat above normal levels. However, the January has seen its highest temperatures.

• [T]he second week of January [in the East] should prove cooler than the first week. However, temperatures could spike after mid-week ahead of the advance of a potentially much colder air mass.

Mean Temperatures (January 1-7):
Boston: 37.7°
New York City: 44.6°
Philadelphia: 44.1°
Richmond: 56.0°
Washington, DC: 48.3°

Mean Temperatures (January 8-14 through January 14 4 pm):
Boston: 36.9°
New York City: 41.7°
Philadelphia: 44.1°
Richmond: 53.6°
Washington, DC: 47.7°

• The colder January 15-22 period might see the opportunity for accumulating snow even in such snow-deprived (to date) cities as Washington, DC and Philadelphia…

As the trough shifts toward the East, I expect that there will be additional accumulations of snow in such cities as Detroit and Chicago…

As this happens, cities such as Bismarck, Rapid City, Omaha, and Kansas City will likely see their coldest readings so far this season probably during the January 13-16 timeframe.

The latest computer guidance is hinting at the possibility for snow along at least parts of the East Coast for the latter part of this period.

So far, with lows of –27° in Bismarck and -8° in Omaha on January 14, both cities have reached new low temperatures for the season. It still appears reasonably likely that Rapid City could do the same. Kansas City might not.

• The latest model guidance suggests that a series of deep troughs should move through the Pacific Northwest. Today and tomorrow [January 7-8], some parts of Washington State could receive some snow. It remains to be seen whether the snow reaches Seattle.

Some snow fell in Seattle on January 9. There were no accumulations. On January 12, a new surge of colder air pushed into the region. The PNA had also begun to rise by that time.

• The first 10 days of January should see continued above normal rainfall on the West Coast. Rainfall should not be as extreme as the recent 3-day total of 6.84” in Los Angeles but it could still be significant. Cities where above normal precipitation should occur extend beyond the West Coast and include Los Angeles, San Francisco, Omaha, and Houston… After January 10, the water-logged Pacific West Coast (especially Los Angeles and San Francisco) should begin to dry out.

No 3-day period produced as much rainfall at Los Angeles as the 3-day total in December but the January 7-9 period came close with 5.25”.

Precipitation Amounts through January 10:

Houston: 0.10” Much Below Normal
Los Angeles: 8.80” Very Much Above Normal
Omaha: 0.49” Above Normal (14.1” snowfall January 5-6)
San Francisco: 3.34” Above Normal

• Cold air will continue to make its presence felt in the Pacific Northwest and there might yet be another snowfall opportunity around the January 11-13 timeframe. After the 14th, a significant warming trend could evolve.

The latest computer guidance supports a warming trend beginning this weekend. Overall, that’s a day later than the idea stated above. Seattle did not see any additional snow during the January 11-13 period.

The January 16-31 Ideas:

January’s Spring fling has ended. Per analogs such a pattern could last 2-3 weeks before possibly breaking and then reloading after a pause.

Instead of sunshine, warm breezes that remind one of Southern climes or Spring Break, winter will be seducing the East with opportunities for snowfall. Whether such flirtation proves serious and snowstorms result remains to be seen, but there does seem to be good potential. However, in my view, all of the big cities from Richmond to Boston should see some accumulations of snow during the second half of the month. Possible amplification in the closing 5-7 days of the month might heighten the risk of a significant snowfall.

• In the East, Philadelphia to Boston will likely see some light snow and flurries from a clipper system late Sunday into Monday. Some accumulations are possible. Later, Philadelphia to Boston should see additional accumulating snow in the January 19-21 timeframe. Some areas may see an appreciable snowfall, especially in New England where the potential might exist for at least several inches or more. Washington, DC and Baltimore might also see a small accumulation from this system. However, from Richmond southward, flurries might be the worst of the system.

Lowest temperatures during the January 16-22 period may fall below 10° as far south as Philadelphia (the January 14 12z ECMWF takes the 850 mb temperature below –21°C for 1/18 12z in each of these cities). At warmest, look for low teens in New York City and Philadelphia. Boston might see the temperature drop into the lower single digits.

The January 23-31 period has the potential to bring another strong shot of cold that might send readings into the single digits to Philadelphia. There should be one or two opportunities for accumulating snow in both the Mid-Atlantic and Northeast. I fully expect that Washington, DC through Boston will add to their snow accumulations during this period.

• Clipper systems will likely add to seasonal snowfall totals in Cleveland, Detroit, Chicago, and Pittsburgh during the January 16-31 period. For the most part, readings there should average below normal and it is possible that Detroit, Chicago, Cleveland, and perhaps Pittsburgh might see the temperature fall below zero one at least one day during the January 21-28 period.

• Waterlogged California should enjoy relatively benign weather for the most part through at least January 25.

• The January 16-22 period will likely see warmer readings return to much of the West. It is possible that this period (perhaps Tuesday or Wednesday) could see the temperature reach 50° in Seattle.

• While the East is experiencing the return of winter, warmth should overspread the Plains States by the end of the week. Thus Thursday or Friday could see the temperature exceed 40° in Omaha, Kansas City, and Bismarck. The potential exists for the thermometer to climb into the 50s in both Omaha and Kansas City. The following week should remain relatively mild there, though a brief shot of colder air is not out of the question.

ENSO Developments:

The data for the week of January 5 revealed:

• Region 1+2: 23.7°C –0.2°C
• Region 3.4: 27.1°C +0.6°C

The cool anomaly is back in Region 1+2. Moreover, Region 3.4 has seen its temperature fall for the second consecutive week. This raises the issue as to whether the weak El Niño event is nearing an end.

After an examination of weekly data, I believe Region 3.4’s temperatures will likely fade only very slowly. Overall through January, Region 3.4’s temperatures will likely average 27.0°C +/- 0.1°C. I don’t believe that Region 3.4’s temperature will fall below 26.8°C during January.

Finally, the idea of a generally stable ENSO situation through January is strongly supported by the ENSO models. Consequently, the potent combination of a weak El Niño-weak West/later weak East QBO should exist as the calendar progresses into and through February. Put simply, consistent with past analogs, February has reasonable potential to prove colder than normal and snowier than normal, especially from the northern Mid-Atlantic northward.

Seasonal Snowfall Update:

Through January 13, the select cities highlighted in the Winter 2004-05 forecast were faring as follows:

• Have Received More than 30% of the Minimum Estimate for Seasonal Snowfall:
Boston, Chicago, Cleveland, Detroit, Providence, and St. Louis
• Have Received Less than 10% of the Minimum Estimate for Seasonal Snowfall:
Baltimore, Philadelphia, and Washington, DC

The cities in the latter category should be seeing increased snowfall in the second half of January and also during February.

Totals to Date (through January 13):
Albany: 21.6”
Baltimore: Trace
Boston: 19.9”
Chicago: 16.5”
Cleveland: 46.2”
Detroit: 24.8”
New York: 3.0”
Philadelphia: 0.4”
Pittsburgh: 7.4”
Providence: 20.0”
St. Louis: 5.5”
Washington, DC: 0.1”

Overall, so far, as anticipated, the Great Lakes region and Ohio Valley are doing very well in the snowfall department. Pittsburgh is lagging. New England is also doing very well. The Middle Atlantic region is the main laggard at this point in time but, as noted above, things should begin to improve there.

Analogs: Added Value or Useless “Data Mining?”

Recently, at another message board, there appeared to be some misunderstanding about the concept and use of analogs in seasonal forecasting with analogs being criticized as “essentially data-mining expeditions looking for past massively snowy NE winters that have the slightest resemblance in even the most trivial or even imagined aspect to current conditions.”

Little could be farther from the truth. To be sure, some might well toss around the term or search for analogs with a predetermined outcome in mind. But that would be a misuse of analogs and not necessarily evidence that analogs are useless. In fact, against the development and application of analogs in an objective fashion, that overgeneralization has no more to do with reality than if one were to take an extreme outlier among the GFS or ECMWF ensembles and use that outlier as a basis for charging that ensembles are used to predict extreme weather that has nothing to do with current conditions.

Let’s take a closer look.

First, for those who are not fully familiar with analogs, an analog is “something that is analogous or similar to something else” (Webster’s Ninth New Collegiate Dictionary). In the case of meteorology, an analog concerns a synoptic setup that is reasonably similar to the current one in a bid to better understand how a given pattern might evolve, especially in a timeframe that is beyond what is available from model guidance or when such guidance is least reliable.

Second, and this is crucial, analogs are not used to make date specific, temperature-specific, or precipitation-specific predictions. They are intended to offer broad guidance e.g., how patterns might unfold. For example, with regard to the pattern change that is now evolving, a strong cluster of analog seasons suggested that such a change had occurred in the past (70%). Thus, given the strength of analogs, it was not unreasonable to argue that a pattern change would occur around mid-month (give or take a few days).

As a general rule, in times of changing patterns where uncertainty is highest, analogs can provide invaluable guidance.

Third, the criticism of analogs is inconsistent with the general premise of synoptic meteorology. In general, Ridge-Trough placement (i.e., patterns such as a PNA+/NAO-, PNA-/NAO+, etc.) offer reasonably defined outcomes for specific parts of North America over a given period of time. SSTAs, among other factors, can help point to such placement. Hence, if one were to assert that an examination of analogs is useless, one would be implying that the idea that Ridge-Trough placements have reasonably-defined outcomes is similarly useless.

Fourth, when one is analogging, some critical issues do arise. Often, these pertain to either very small sample size or difficulty in identifying applicable analogs.

Unfortunately, most of the regularly-reported global data only goes back to 1950. However, one can reasonably overcome these issues by developing broad-based analogs (e.g., given ENSO analogs) and then tweaking them for different scenarios e.g., the QBO, etc. That way, one comes up with a variety of scenarios and much as one uses ensembles to reach judgments, one can do the same with the variety of scenarios that result from such sensitivity analysis. If, for example, 80% of the analogs point to the idea that Boston could well see 50” or more seasonal snowfall, that’s probably a pretty strong indication that a snowy season lies ahead. As no specific model currently exists that even attempts to predict seasonal snowfall, analogs can play an invaluable role, especially as such information could be critical e.g., to municipalities when making decisions as to how much rock salt to purchase for a coming season.

Another problem that arises concerns the lack of highly specific global data. Instead, the data is for large regions of the world and this creates some inherent limitations. Nevertheless, use of some of the major indices can create reasonable representations and such representations can add to understanding.

Perhaps well into the future, modeling will make the use of analogs largely irrelevant. But for now, that’s simply not where things stand. At this point in time, analogs are not the whole answer to seasonal forecasting but they are a part of the solution.

[TT-SEA]

Excellent post. Extremely informative and meaningful.

I did notice the forecasted trend downward in the PNA toward the end of the month. Do you have any inclination whether it will go sharply negative (as it was recently) in early Febraury?

[AnthonyC]

I'm surprised with the extremely negative PNA phase we were just in, that Western Washington wasn't entrenched in arctic air. Doesn't a very negative PNA phase ALWAYS mean brutally cold air for Western Washington? We had glancing blows of arctic air...about three to be exact...but no direct hit. Current indications suggest another extreme negative phase late January, early February.

[donsutherland1]

TT-SEA and AnthonyC,

The latest GFS ensembles marked the first step for the return of a negative PNA. It's too soon to suggest that this is a trend at this point in time.

If the PDO were to hold negative or go more strongly negative, that would be favorable for a more negative PNA. Past ENSO analogs (top 15) are not highly enthusiastic about a predominant negative PNA during the last week in the month. But there is one case where it did go strongly negative (1969 and that was a negative PDO season).

Right now, I believe the PNA will peak (as shown on the GFS Ensembles) and then trend downward. I don't believe the ridge in the means will be fully eroded and drops in the PNA could be transient during the rest of this month. Thus, for now, I'd suggest that the PNA could go neutral late in January and perhaps negative for a short time. I just don't know whether it would be sustained negative. I don't believe it will go as strongly negative as happened recently (though the trough didn't align itself in a fashion that would have brought severe cold to Seattle itself).

[R-Dub]

I have heard that if both the AO and the PNA were to be both fairly negitive, that is the best setup for us in the pac northwest for winter weather? Just what I have heard, I am no expert.

[deguy50]

great post

[Anonymous]

Don, when you say that the east will see snowfall you didnt mention any big systems. You only mentioned clipper systems. Is there a chance that places like philly and the east coast could see a SECS or a MECS?

[donsutherland1]

Jrodd321,

In my discussion, I had stated, "Possible amplification in the closing 5-7 days of the month might heighten the risk of a significant snowfall." Occasionally, a clipper system can blow up into a major snowstorm. Moreover, before the pattern shifts, I believe the southern stream will have had a chance to become involved.

Right now, it is too soon to try to pin down any specific threat for a significant snowfall. This does not mean that there won't be such a storm.

[Chris the Weather Man]

Jrodd321,

In my discussion, I had stated, "Possible amplification in the closing 5-7 days of the month might heighten the risk of a significant snowfall." Occasionally, a clipper system can blow up into a major snowstorm. Moreover, before the pattern shifts, I believe the southern stream will have had a chance to become involved.

Right now, it is too soon to try to pin down any specific threat for a significant snowfall. This does not mean that there won't be such a storm.

Ah yes, Don, I will never forget the Mauler from Canada, on Dec 30th 2000.....

[sertorius]

Don:

Thanks again for the great info!! I really learn alot from your posts!! I have to admit that I'm kind of sad to see us return to a benign pattern for the next couple of weeks. Of course I kind of thought it was comming-the models showed it the pna was rising and the fact is, it is the unique winter where we have over 2 weeks of sustained cold-I got nearly 2 weeks of sub freezing temps. and have had two arctic out breaks-I really can't complain (but man I love the cold!!!) I just wish the precip and cold air could have lined up better-we sure missed some chances. While we are not in the sweet spot, we are still, at least it looks that way to me, on the western side of the trough and could very well see some storminess at times (maybe the ridge in the west will not move too far east???) I'll just sit back, watch the models, listen to everyone in the East enjoy their snow and wait for February!!! By the way, this pattern does not look to bad for St. Louis-I still think your forcast will work out there. Our forcast of 29 inches by one met is in jepordy-his forcast was not bad-we've had almost 3 inches of precip (another reason I'm worried about much more snow here-we've already got our total precip for Jan. and Feb) this month-the upper temps. just blew it. Thanks again-you are a true joy to read and I hope some day I can write my posts as logically as you do!!!

[donsutherland1]

Sertorius,

Thanks for the very kind words.

I don't believe winter is finished in Kansas though a milder period is looking likely beginning late next week. There may be some attempts to dent the positive PNA as the month nears an end and some retrogression may occur. I believe that there will still be at least one additional period where the trough is centered in the Plains states and opportunities for snowfall should be present at that point in time.

As for St. Louis, the city finally caught a break and received some accumulating snow. I'm expecting 18"-28" there and now it is among the cities that is within 30% of the minimum seasonal amount that I had expected. The Mid-Atlantic states have fared worse but I expect progress to be made there.

[donsutherland1]

Chris,

That's one of the few storms I actually missed. I was in Detroit at the time and only experienced lighter snows from the system that eventually blew up and brought very heavy snows to parts of NJ and SE New York.

[AnthonyC]

donsutherland1,

What's the news on the latest PNA forecast for the end of January. Early indications suggest another significant negative phase? Could this happen again?

Anthony

[donsutherland1]

AnthonyC,

The GFS ensembles have trended more in that direction over the past two days. I'll probably go into that further in my pattern discussion for next Friday, as I'll have had more time to look at possible analogs, SSTAs, and seen more of the model/ensemble guidance.

[donsutherland1]

Further addressing the analogs issue at the beginning of this thread, it should be noted that to date, Washington, DC has received 0.1" of snow. The overwhelming weight of analogs (almost 85%) saw DCA receive more than 1" in January with approximately 1-in-4 seeing 6" or more for the month.

I went back to the analogs on which I relied to make an estimate that DCA should see close to normal snowfall by the time the season is finished and here's how the analogging fared for December with regard to Boston, New York City, and Washington, DC:

Boston:
Average snowfall with outliers removed: 6.3"
Analogs with 4" or more: 70%
Analogs with 6" or more: 36%
Actual snowfall: 7.2"

New York City:
Average snowfall with outliers removed: 4.7"
Analogs with 2" or more: 70%
Analogs with 6" or more: 43%
Actual snowfall: 3.0"

DCA:
Average snowfall with outliers removed: 1.4"
Analogs with less than 1": 64%
Actual snowfall: 0.1"

January Information:

Boston:
Average snowfall with outliers removed: 11.1"
Analogs with 10" or more: 58%
Analogs with 15" or more: 32%
Snowfall to date: 8.8"

Note: Will exceed 10" and a good chance of exceeding 15"; analog idea working out extremely well.

New York City:
Average snowfall with outliers removed: 8.1"
Analogs with 2" or more: 74%
Analogs with 10" or more: 47%
Snowfall to date: Trace

Note: Slow start, but I would be surprised if NYC does not reach 4"-8" for January

DCA:
Average snowfall with outliers removed: 6.3"
Analogs with 1" or more: 84%
Analogs with 6" or more: 26%
Snowfall to date: None

Note: DCA will receive accumulating snow. Conservatively, the month will likely finish with 3"-6". Modeling is looking good and per analogs, the modeled threat is real.

My point is not that the analogs should be used to specific amounts of snowfall for given cities but should be used for general guidance. Overall, the broad idea of the analogs worked out reasonably well in December and they give added reason for confidence in the modeling that points to the return of snow to the Mid-Atlantic region.

[BritBob]

Excellent stuff Don. Always a pleasure to read your posts

[donsutherland1]

Thanks very much, Bob, for the kind words. I hope you're enjoying the winter.

Best wishes.