Winter 2013-14 medium range discussion

[donsutherland1]

Don,

 

That was a fairly close to normal Dec for temps in northern Appalachians with several ice events. This includes a significant ice storm on Christmas Eve. Interesting that such storms are showing up on the models.

And consistent with those cases, I wouldn't be surprised if there were not at least one or two wintry events in central PA in December. It will be interesting to see how things finally evolve.

[donsutherland1]

Brief evening update...

 

The MEX MOS from the 11/30 12z run of the GFS illustrated the magnitude of the cold likely to impact the Pacific Northwest during the 12/3-10 timeframe. During that period, it appears likely that Seattle and Portland could see one or more low temperatures in the lower or middle 20s. One of those two cities could see a reading below 20°. Spokane will likely see one day with a high temperature in the upper teens or low 20s and a minimum temperature in the single digits.

 

The MEX MOS shows a high of 33° in Seattle and 18° at Spokane for 12/5. The lowest readings shown on the MEX MOS are currently 23° at Portland, 25° at Seattle, and 4° at Spokane. It should be noted that the MEX MOS does not extend through the entire period and it still gives some weight to climatology. The 11/30 12z run of the ECMWF was even colder for all three cities.

 

More details concerning those cities' records and data can be found in Messages #181 and #195.
 

[SnoSki14]

Great analysis but a huge bummer as well.

[WidreMann]

Great analysis but a huge bummer as well.

Not really. Don, and the models, have been trending away from the wall-to-wall warm we were expecting for December earlier in the season.

[tacoman25]

With the OPI so far looking good.  We better hope the EPO pulls it off or...

 

 

Sadly it's not winter for 2 more days. 

 

And technically 23 more. 

 

This negative EPO has been completely wasted.

 

In what sense?

 

 

 

And this upcoming blocking episode looks like it will deliver some very impressive cold anomalies...right on the heels of some impressive ones for the East.

 

[donsutherland1]

In what sense? And this upcoming blocking episode looks like it will deliver some very impressive cold anomalies...right on the heels of some impressive ones for the East.

 

12zecmwfCumulative850mbTempAnomalyNA240.gif

I suspect that the upcoming cold outbreak for the Pacific Northwest, Idaho, Montana, Wyoming, parts of the Dakotas, parts of BC, Alberta, and Saskatchewan will be among the highlights of winter 2013-14.

[tacoman25]

As a poster mentioned earlier, even though it is highly unusual for the EPO to remain negative for longer than 20-30 days at a time, there have definitely been winters where it kept coming back after short breaks: 1949-50 (between 12/10 and 2/6 only saw 3 +EPO days), 1955-56 (from Nov-Feb had about 85% of the days -EPO), 1956-57 (had reoccurring major -EPO periods, including events reaching -3 or lower in early December, mid January, early February, and late February), 1958-59, 1962-63 (-EPO dominated the entire Dec-Feb period), 1967-68, 1977-78 (Nov-Jan were predominantly -EPO), 1979-80 had reoccurring major -EPO episodes from Dec-Feb, Dec 1983's record-breaking -EPO event was followed by a major -3 event in January, mid Dec 1984-mid Feb 1985 had persistent -EPO periods, 1990-91, 1993-94, 1995-96 had major -EPO events that kept coming back, as did 2002-03.

 

There is no reason to think that just because we have seen a ton of -EPO so far it will disappear for the rest of the winter or even just December. It could just as easily go to +EPO for a week or less and then return to -EPO blocking again.

[donsutherland1]

As a poster mentioned earlier, even though it is highly unusual for the EPO to remain negative for longer than 20-30 days at a time, there have definitely been winters where it kept coming back after short breaks: 1949-50 (between 12/10 and 2/6 only saw 3 +EPO days), 1955-56 (from Nov-Feb had about 85% of the days -EPO), 1956-57 (had reoccurring major -EPO periods, including events reaching -3 or lower in early December, mid January, early February, and late February), 1958-59, 1962-63 (-EPO dominated the entire Dec-Feb period), 1967-68, 1977-78 (Nov-Jan were predominantly -EPO), 1979-80 had reoccurring major -EPO episodes from Dec-Feb, Dec 1983's record-breaking -EPO event was followed by a major -3 event in January, mid Dec 1984-mid Feb 1985 had persistent -EPO periods, 1990-91, 1993-94, 1995-96 had major -EPO events that kept coming back, as did 2002-03.

 

There is no reason to think that just because we have seen a ton of -EPO so far it will disappear for the rest of the winter or even just December. It could just as easily go to +EPO for a week or less and then return to -EPO blocking again.

No disagreement from me. As I noted in the New England subforum, some winters following a November with an EPO of -1 or below wound up with a predominantly positive  EPO (e.g., 1966-67) and others remained predominantly negative (e.g., 1978-79).

 

[donsutherland1]

The EPO-/PNA- pattern is highly conducive for cold weather in the Pacific Northwest. For December, that pattern accounts for more than three quarters of days on which the temperature fell to 25° or below in Seattle. Some charts to illustrate the EPO/PNA patterns and frequency of December minimum readings of 25° or below. From 1950 through 2012, there were 101 such days in December.

 

[Harry]

No disagreement from me. As I noted in the New England subforum, some winters following a November with an EPO of -1 or below wound up with a predominantly positive  EPO (e.g., 1966-67) and others remained predominantly negative (e.g., 1978-79).

 

 

And both 66-67/78-79 were decent winters here ( 66-67 is 3rd snowiest ever and featured the biggest snowstorm ever for this location with 28-31" via Jan 67 snowstorm that also hit Chicago )  and 79 had a big 18+ event here/Chicago and along i95 ala PD1 as well. And yeah the -EPO waited till March to show back up in 66-67. 66-67 was a very hyperactive winter with all the storms hitting this region and i-95 corridor and it did not let up till late April atleast around here with the snow storms.

 

Wonder if the big storm threat is by chance or something tied into the EPO somehow? Even the winter of 78 which featured the ne blizz and OV/GL blizz saw the EPO swing a bit near the time of both blizzards? In all 3 cases though the QBO was +westerly.. Even March 93 fits the profile? EDIT.. Can add Jan 99 MW/GL and GHD in 2011.. Interesting..

[Absolute Humidity]

Thanks Wes. Monthly ideas are sort of like throwing a long bomb. The risk of error is always great. I don't think there's any merit in standing with an idea when the data seems to have turned overwhelmingly unsupportive. In fact, doing so would probably be more foolish than anything, as the idea would depend strictly on the hope that something changes rather than any concrete premise.

 

Even as the EPO is difficult to forecast and it is tempting to go with persistence, the idea that the EPO- will be approaching historical benchmarks by mid-December argues for some relaxation/retrogression of the Alaska ridge. That could be all that's needed for the AO+ to exert a larger influence than it has to date.

 

Also, the magnitude of cold likely to impact the Pacific Northwest and Northern Plains in the near- and medium-term tends to support the composite maps that come up with the recent data. Unfortunately, it also strengthens my overall winter idea of a milder winter in much of the East.

Hi Don and thanks for the great write ups. Do you think we here in east, NYC PHL area are looking at another 2012 like snowless winter of under 10"?

[donsutherland1]

There had been a posted message in this thread that raised some issues about longer-range forecasting. I will respond to the issues that I recalled.

 

First issue: Skill. 

 

Weekly, monthly, and seasonal forecasts are less skillful than those over the very near-term. Complexity and chaos all play a role. However, chaos is not equally distributed e.g., recent research (SAI and OPI) indicate that the Arctic Oscillation may be reasonably predictable for the predominant winter state. No such work has yet been revealed with respect to the EPO. The PNA also has a relationship to the PDO.

 

It is unreasonable to expect longer-range forecast to be anything close to as skillful as shorter-range ones. Indeed, I've referred to such forecasts as amounting to a "long bomg" in football (high risk of interception/incomplete passes and high risk of error with regard to the forecasts).

 

Second Issue: Natural Gas prices as a predictor.

 

The hypothesis is that natural gas traders, because they are placing financial bets, would demonstrate more skill. The literature suggests no conclusive outcome in this favor. Cognitive biases including but not limited to sunk cost fallacy and confirmation bias represent barriers to objectivity. Moreover, such trades are, in effect, reactions to models and other forecasts, all of which face the same limits any long-range forecasts possess.

 

Out of curiousity, I ran the numbers for 2013. The idea was that the change in natural gas prices over the last five trading days prior to a month offered insight into the monthly outcome in the eastern half of the CONUS (highest energy utilization region). The prices are available at the Energy Information Administration's website at:

 

http://www.eia.gov/dnav/ng/hist/rngc1d.htm

 

The table below shows the outcomes. If one gives half credit for a "near normal" outcome, one is still dealing with a 50-50 outcome. If one goes by the stricter definition that "near normal" should be treated as an error against a "colder" or "warmer" forecast, the error rate this year is 55%.

 

 

Third Issue: Random chance.

 

Synoptic patterns have tendencies to produce certain outcomes, though there is variability. For example, the scatter diagram and numbers I provided for Seattle in Message #219 clearly demonstrates how a given synoptic pattern skews the outcomes. It is not an example of random chance.

 

Fourth Issue: Analogs alone.

 

Analogs represent some cases where teleconnections, for example, were similar. They are not a substitute for all other tools. One should use the full range of tools in trying to gain insight into the longer-range. Analogs can provide some insight. But one should always be aware of the limitations of all of the tools and the reality that long-range forecasting is subject to much greater uncertainty and inherent difficulty than shorter range forecasting.

 

Fifth Issue: Given low skill, why bother making such forecasts.

 

This thread is for discussion of medium and longer-term patterns. Most people who participate on weather boards such as AmericanWx have a degree of understanding that long-range forecasts are far more difficult and far more prone to error than short-range ones. There is no implicit assumption that they have the same degree of accuracy.

 

Members and meteorologists here and elsewhere share their thoughts with full understanding that the participants on such boards are generally more literate when it comes to weather than the general public. Hence they know that their thoughts won't be misconstrued as high probability outcomes.

[donsutherland1]

Hi Don and thanks for the great write ups. Do you think we here in east, NYC PHL area are looking at another 2012 like snowless winter of under 10"?

I'm not yet sure. A lot will depend on whether the lack of precipitation in recent months (with a few exceptions) gives way to more frequent and meaningful precipitation.

[GaWx]

 

Second Issue: Natural Gas prices as a predictor.

 

The hypothesis is that natural gas traders, because they are placing financial bets, would demonstrate more skill. The literature suggests no conclusive outcome in this favor. Cognitive biases including but not limited to sunk cost fallacy and confirmation bias represent barriers to objectivity. Moreover, such trades are, in effect, reactions to models and other forecasts, all of which face the same limits any long-range forecasts possess.

 

Out of curiousity, I ran the numbers for 2013. The idea was that the change in natural gas prices over the last five trading days prior to a month offered insight into the monthly outcome in the eastern half of the CONUS (highest energy utilization region). The prices are available at the Energy Information Administration's website at:

 

http://www.eia.gov/dnav/ng/hist/rngc1d.htm

 

The table below shows the outcomes. If one gives half credit for a "near normal" outcome, one is still dealing with a 50-50 outcome. If one goes by the stricter definition that "near normal" should be treated as an error against a "colder" or "warmer" forecast, the error rate this year is 55%.

 

 

 

 

Don,

That's a very interesting analysis of natural gas price movements. Thanks for posting that. However, I'd like to counter your conclusion somewhat:

1. The end of August prediction for September should have been for cooling rather than heating as September is more of a cooling month than heating, especially early to mid. Overall US heating demand doesn't exceed cooling demand til quite late in Sep. on average. If one assumes it should be cooling, it is actually a correct prediction per your criteria.

2. The months with the largest absolute value of movement actually did very well. Regarding the six months with absolute value of 0.168+, five were "correct", one gets half credit, and none were erroneous after making the above referenced adjustment for September. So, 5.5 out of 6 total possible points or ~92% if it is looked at that way.  Regarding the five months with the least movement prior to their start (0.126 and lower), only one out of five was "correct" or only 20%. The point is that one could look at this as saying that the months with the greatest "confidence" of the traders at the end of the prior month, which averaged 0.19 (confidence measured by degree of price movement), were actually predicted quite well and the ones with less price movement, which averaged only 0.08, were predicted poorly.

 

3. One could argue that this isn't a fair way to analyze the traders because the month ahead starts getting analyzed well before the last five days of the prior month since even just two calendar week forecasts start going into the next month about 10 trading days prior. Traders don't just suddenly get a first forecast for the next month only within five days before that month starts. If one looks at it that way, it could be that the last five trading days, alone, were during a period for which the change in the forecasted anomaly for the upcoming month went in the opposite direction of that month's actual anomaly. For example, let's look at movement just prior to January. The movement was +.005,which is basically no movement (neutral) to be fair about it. However, even if it is counted as a prediction of a cold Jan., it could be that the predictions for Jan. actually were too warm five days prior and were cooled very slightly during the last five days to something that verified as being more accurate. So,IF that were actually the case, the traders would have gotten Jan. "correct" during the last five days prior to it even with a still very warm E CONUS verifying.

 

4. Of course, there are significant market moving factors during the last five days of a month in addition to the change in the upcoming month's forecast that could easily skew things. One of those is the weekly inventory report, which can itself cause sharp movement on the day when it is released if the actual inventory change is quite a bit different from what was expected. So, that could skew everything.. Another factor is that just being near the end of the calendar month may cause unusual price movement due to people wanting to get out of their positions before the end of the month for whatever reason. That could skew things somewhat for a short time.

[donsutherland1]

Don,

That's a very interesting analysis of natural gas price movements. Thanks for posting that. However, I'd like to counter your conclusion somewhat:

1. The end of August prediction for September should have been for cooling rather than heating as September is more of a cooling month than heating, especially early to mid. Overall US heating demand doesn't exceed cooling demand til quite late in Sep. on average. If one assumes it should be cooling, it is actually a correct prediction per your criteria.

2. The months with the largest absolute value of movement actually did very well. Regarding the six months with absolute value of 0.168+, five were "correct", one gets half credit, and none were erroneous after making the above referenced adjustment for September. So, 5.5 out of 6 total possible points or ~92% if it is looked at that way.  Regarding the five months with the least movement prior to their start (0.126 and lower), only one out of five was "correct" or only 20%. The point is that one could look at this as saying that the months with the greatest "confidence" of the traders at the end of the prior month, which averaged 0.19 (confidence measured by degree of price movement), were actually predicted quite well and the ones with less price movement, which averaged only 0.08, were predicted poorly.

 

3. One could argue that this isn't a fair way to analyze the traders because the month ahead starts getting analyzed well before the last five days of the prior month since even just two calendar week forecasts start going into the next month about 10 trading days prior. Traders don't just suddenly get a first forecast for the next month only within five days before that month starts. If one looks at it that way, it could be that the last five trading days, alone, were during a period for which the change in the forecasted anomaly for the upcoming month went in the opposite direction of that month's actual anomaly. For example, let's look at movement just prior to January. The movement was +.005,which is basically no movement (neutral) to be fair about it. However, even if it is counted as a prediction of a cold Jan., it could be that the predictions for Jan. actually were too warm five days prior and were cooled very slightly during the last five days to something that verified as being more accurate. So,IF that were actually the case, the traders would have gotten Jan. "correct" during the last five days prior to it even with a still very warm E CONUS verifying.

 

4. Of course, there are significant market moving factors during the last five days of a month in addition to the change in the upcoming month's forecast that could easily skew things. One of those is the weekly inventory report, which can itself cause sharp movement on the day when it is released if the actual inventory change is quite a bit different from what was expected. So, that could skew everything.. Another factor is that just being near the end of the calendar month may cause unusual price movement due to people wanting to get out of their positions before the end of the month for whatever reason. That could skew things somewhat for a short time.

I don't disagree that there can be some indications and my illustration was a simple one. On a monthly scale, one is dealing with similar issues as any forecaster e.g., asymmetric information. Moreover, prices reflect more than must weather-related information. Indeed, in most cases in recent years, one finds that the average price in the last 5 days of the month is lower than the average price for the preceding 5 days. Technical issues regarding the industry are at play. September is interesting. In parts of the East, it's still the cooling season. In parts of the Midwest, it flips to the heating season. Even if one includes September in the cooling season, one is still dealing with 55% correct so far this year.

 

The big movement proves more correct, but that should be the case when there is higher confidence (e.g., a much stronger signal on the ensembles). One could take a similar approach with respect to confidence, calling for warm or cold anomalies only in areas of high confidence e.g., for the upcoming December, the Southeast might be designated warm, Northern Plains/Pacific Northwest cold and the rest equal chances (inconclusive signal that would be outweighed by variability). That's not a bad approach.  

 

My overall point was that long-range forecasting is inherently difficult and there is no way to avoid that difficulty. Markets do well in the short-term in discounting on a range of factors. But in the long-run, they also hit barriers ranging from information asymmetry to cognitive biases that impede long-range forecasting. The posted message which is not here seemed to imply that the natural gas market provides superior long-range weather forecasts (despite being subject to the same information limitations, cognitive biases, group dynamics e.g., herding at times, and responding to factors beyond expected weather). The fact is that there is no empirical evidence in the economic literature to support that claim. Long-range forecasting is difficult.

[donsutherland1]

Some quick afternoon thoughts...

 

1. The big story over the next 7-10 days will be the cold that will invade the Pacific Northwest and Northern Plains. GFS ensemble forecasts continue to show some temperatures 3 standard deviations below normal at the height of the cold.

 

2. Some of that cold will come eastward during the second week of December, though it does not appear to be a direct shot.

 

3. Some storminess could coincide with the cold's approach late in the second week of December.

 

4. The PNA is forecast to remain negative through mid-month. Odds of a colder than normal December in the Pacific Northwest are fairly high given that pattern and the cold anomalies that will develop during the first half of December.

[GaWx]

Don,

I agree that there is no empirical evidence that the NG market provides superior long range weather forecasts. That would be very difficult to determine, regardless. However, my point was that I thought that the posted NG monthly price movement table made the NG market look more clueless than reality for the reasons I gave. I think that the market wasn't as clueless as was implied. That's all I was trying to say.

[donsutherland1]

Don,

I agree that there is no empirical evidence that the NG market provides superior long range weather forecasts. That would be very difficult to determine, regardless. However, my point was that I thought that the posted NG monthly price movement table made the NG market look more clueless than reality for the reasons I gave. I think that the market wasn't as clueless as was implied. That's all I was trying to say.

 

I didn't mean to imply that the market is clueless and regret that my chart might have created that impression, as that wasn't intended to be my point. It isn't clueless. I think it probably does as good a job as one can expect given information limitations, etc.  

 

I do believe some recent progress e.g., the work associated with the Arctic Oscillation, holds promise in leading to better long-range forecasts down the road. Nevertheless, there will always be a higher degree of uncertainty than is the case with short-term forecasts.

[bluewave]

Don,

I agree that there is no empirical evidence that the NG market provides superior long range weather forecasts. That would be very difficult to determine, regardless. However, my point was that I thought that the posted NG monthly price movement table made the NG market look more clueless than reality for the reasons I gave. I think that the market wasn't as clueless as was implied. That's all I was trying to say.

 

The NG market prices track the CPC 6-10 and 8-14 day forecasts fairly closely as we head into the heating season.

 

http://www.cpc.ncep.noaa.gov/products/predictions/814day/index.php

 

http://www.bloomberg.com/news/2013-11-29/natural-gas-futures-gain-most-since-march-on-colder-u-s-

weather.html

 

Natural gas jumped to a five-month high in New York, capping its largest monthly gain since March, on forecasts for colder weather that would spur demand for heating homes and businesses.

 

Futures rose for a seventh day, the longest streak of gains since January 2011, as most of the U.S. will be blanketed with below-average temperatures from Dec. 6 through Dec. 12, according to the Climate Prediction Center in College Park, Maryland. About 49 percent of U.S. households use natural gas for heating, government data show.

[donsutherland1]

The NG market prices track the CPC 6-10 and 8-14 day forecasts fairly closely as we head into the heating season.

 

http://www.cpc.ncep.noaa.gov/products/predictions/814day/index.php

 

http://www.bloomberg.com/news/2013-11-29/natural-gas-futures-gain-most-since-march-on-colder-u-s-

weather.html

 

Natural gas jumped to a five-month high in New York, capping its largest monthly gain since March, on forecasts for colder weather that would spur demand for heating homes and businesses.

 

Futures rose for a seventh day, the longest streak of gains since January 2011, as most of the U.S. will be blanketed with below-average temperatures from Dec. 6 through Dec. 12, according to the Climate Prediction Center in College Park, Maryland. About 49 percent of U.S. households use natural gas for heating, government data show.

I don't disagree. The markets are skilled users of forecast information and CPC is as good a source of forecast information as there is. There are also many good private sector forecasters who also add value.

[donsutherland1]

Continuing the focus on the Pacific Northwest, below are the GFS ensembles and also forecast PNA.

 

 

Ahead of the possible press of cold air toward the East, there could be a day or two of much above normal temperatures in the East, particularly the Southeast. The above frame from the GFS ensembles shows an area of  2σ anomalies. For reference, high temperatures that are 1σ-2σ above normal for the December 1-10 period (1981-2010 period) range from 67°-76°. Low temperatures with similar departures range from 47°-56°.

 

One should note that the GFS ensembles are forecasting 850 mb temperatures. If there is precipitation and no sunshine, highs could be closer to 1.5σ. Right now, that's still 5-6 days away, but the idea that Atlanta might see a reading or two in the 70s (lower to possibly middle) during the first week of December and one or two nights with temperatures in at least the middle 50s or possibly a little warmer is something to watch for based on the ensemble forecast.

 

For purposes of reference, the 12z ECMWF forecasts a temperature of 21.5°C (71°F) and the 12z GFS forecasts a temperature of 20.9°C (70°F) on December 6.

[bluewave]

I don't disagree. The markets are skilled users of forecast information and CPC is as good a source of forecast information as there is. There are also many good private sector forecasters who also add value.

 

Yeah, you could watch the turn in the market after the later October warmer forecasts gave way to cooler ones in early November. 

 

 

 

[GaWx]

I didn't mean to imply that the market is clueless and regret that my chart might have created that impression, as that wasn't intended to be my point. It isn't clueless. I think it probably does as good a job as one can expect given information limitations, etc.  

 

I do believe some recent progress e.g., the work associated with the Arctic Oscillation, holds promise in leading to better long-range forecasts down the road. Nevertheless, there will always be a higher degree of uncertainty than is the case with short-term forecasts.

 

 Fair enough. Regardless of how good or not is the NG market overall with respect to the longer range, I enjoy your forecast analyses very much as I like your index/teleconnection/analog/statistical approach to forecasting keeping in mind I have a statistical background.

[bluewave]

We are right now in the most negative state of the EPO for the 2013-2014 winter.

2013 10 25 -100.402013 10 26 -163.842013 10 27 -170.982013 10 28 -101.832013 10 29  -20.892013 10 30   47.692013 10 31   55.422013 11 01    4.102013 11 02  -14.102013 11 03   -5.722013 11 04    8.272013 11 05   -1.182013 11 06   -9.092013 11 07   17.862013 11 08    3.682013 11 09  -77.792013 11 10 -110.792013 11 11 -174.312013 11 12 -234.392013 11 13 -254.582013 11 14 -248.572013 11 15 -216.572013 11 16 -185.432013 11 17 -136.342013 11 18  -95.952013 11 19  -58.372013 11 20  -44.782013 11 21  -70.822013 11 22  -62.682013 11 23  -51.762013 11 24 -125.672013 11 25 -178.142013 11 26 -211.102013 11 27 -221.95

can't last much longer, but who knows. We can be in a once in a hundred year event.

 

January through November 2013 just set the record for the strongest 500 mb

blocking event in that region since at least 1950 surpassing 1957 and 1989.

The very impressive 2006 blocking episode was centered a bit further west

closer to the Aleutians.

 

 

 

 

[donsutherland1]

 Fair enough. Regardless of how good or not is the NG market overall with respect to the longer range, I enjoy your forecast analyses very much as I like your index/teleconnection/analog/statistical approach to forecasting keeping in mind I have a statistical background.

Thank you for the kind words, Larry. I enjoy reading your thoughts as well, especially those on ENSO.

[donsutherland1]

Evening thoughts...

 

1. The cold air mass that will be surging into the Pacific Northwest and Northern Plains in coming days is now showing up on the global temperature anomalies chart.

 

 

 

2. There are growing signs on the ensembles that the strong EPO- will either relax or give way to an EPO+ regime just after mid-month. The duration of such a change, if it occurs, remains uncertain. But at least a brief regime change would be likely against historical parameters.

 

3. If that change occurs when the AO is positive or strongly positive, much milder weather could overspread the CONUS, though some lag between the date the EPO went positive and the onset of the milder pattern would be plausible.

[Isotherm]

January through November 2013 just set the record for the strongest 500 mb

blocking event in that region since at least 1950 surpassing 1957 and 1989.

The very impressive 2006 blocking episode was centered a bit further west

closer to the Aleutians.

 

 

 

2006.png

 

 

 

It's also interesting to note the resultant pattern from that west-centered NPAC ridge in 2006. Strong positive height anomaly near the Aleutians meant lowering of heights near the British Columbia coast, and that is a furnace regime for most of the Eastern US. The placement of the north Pacific mid level ridge and its poleward extent are absolutely crucial for the downstream temperature pattern in the northern tier. The strong mid level ridge in 2013 is/has been located in a much more conducive geographical area for cross polar flow directed into North America.

[bluewave]

It's also interesting to note the resultant pattern from that west-centered NPAC ridge in 2006. Strong positive height anomaly near the Aleutians meant lowering of heights near the British Columbia coast, and that is a furnace regime for most of the Eastern US. The placement of the north Pacific mid level ridge and its poleward extent are absolutely crucial for the downstream temperature pattern in the northern tier. The strong mid level ridge in 2013 is/has been located in a much more conducive geographical area for cross polar flow directed into North America.

 

True. We are seeing a much different pattern this year as the ridge in that location is keeping the temperatures cooler.

The raging +EPO in 2012 just flooded the lower 48 with record warmth.

 

 

 

[donsutherland1]

Strong Cold Shot Imminent in Pacific Northwest & Northern Plains:

 

This morning, the PNA was -2.055, its lowest value since September 28, 2013 when the PNA stood at -2.228. The current EPO-/PNA- pattern is one that favors cold in the Pacific Northwest, as well as parts of the Northern Plains. Not too surprisingly, the cold that has been showing up on the operational guidance and ensembles for some time is now imminent for that part of North America.

 

The latest values from the MEX MOS, GFS, and ECWMF for Portland, Seattle, and Spokane are below:

 

 

In addition, even as the ensemble guidance continues to show cold anomalies of 2-3 standard deviations below the norm in that part of North America, they show a brief shot of warmth across the Gulf States and Southeastern U.S. of around 2 sigma. Even more impressive, parts of Mexico are forecast to have readings around 3 standard deviations below normal during the December 4-6 timeframe.

 

Following the first week in December, some of the cold remains likely to push south and eastward with cold anomalies encompassing a large part of the CONUS, in addition to Canada at times. Some storminess could coincide with the arrival of the cold air producing at least some opportunities for snow. The greatest chance for mostly or all snow would probably exist in interior sections of the northern Middle Atlantic region, across central and upstate New York and into interior and northern sections of New England. For now, it is too soon to worry about details.

 

In the longer-range, the all-important question concerns whether the EPO will go positive, possibly in the face of a predominantly positive AO. The outcome has yet to be resolved, but that outcome could have an important influence on how things evolve after mid-December in the CONUS. Canada would likely hold onto the cold for a longer period of time.

 

Finally, when it comes to November, the 2008 teleconnection case proved to be the closest match for the actual outcome.

 

 

From reading messages on various message boards, it seems that there is some misconception, at least among the general public, when it comes to cases or "analogs." Such cases or analogs do not, and the emphasis is on "not," seek to provide perfect or near-perfect forecasts. The role of analogs is to provide some measure of insight rather than exact details. Hence, one could reasonably state that from given patterns similar to the expected one, the eastern U.S. proved colder than normal or much of Europe proved warmer than normal. One should not state that Philadelphia or Paris would be colder/warmer than normal, but that probabilities might favor certain outcomes. Perhaps a reasonable analogy is a legal case. One might draw upon a landmark decision to gain insight into how a court might rule in framing one's arguments. Yet, the exact nuances of the case (or a judge's subjective evaluation e.g., judicial philosophy) might lead to a different outcome. The same holds true with respect to analogs in forecasting. Hence, they are not a subsitute for ensembles or other guidance. They complement it.

 

Going back to the above maps, one can see that the 2008 case provided a lot of insight into how things actually worked out. Nevertheless, they were far from perfect.

 

Another note is also important. Developing analogs requires assumptions about the variables one uses. If the assumptions are reasonable, one might gain decent insight. But there is always uncertainty whether it comes to ENSO, the teleconnections, or other variables one incorporates. Hence, the assumptions might be off leading to a bad outcome. The same also holds true with modeling. As data becomes clearer, model outcomes may change markedly, sometimes on a run-to-run basis (often a signal of high uncertainty). That aspect of modeling is one of the reasons ensembles have become so important. Ensembles allow for tweaking of variables (a form of sensitivity analysis), hence one can analyze alternative outcomes and also take a look at the ensemble means (a composite based on the various scenarios). The spread in ensemble solutions can offer insight into the degree of uncertainty.

[tacoman25]

12z Euro shows wave after wave of extreme cold invading the West/Central U.S. over the next week or so. Day 7 is particularly extreme, with sub -30C into Montana. Widespread daily anomalies of -25 to -30F are likely with this, with possible weekly anomalies in some areas of -15 to -20F.