Summer forecast for North America -- severe heat waves ahead

[ptb127]

I hope you are right, I love hot and humid summers. When should the heat "lock" in? no sign of that on any model yet...

[Roger Smith]

Second half of June to first half of September with a few brief interruptions.

[KokomoWX]

Second half of June to first half of September with a few brief interruptions.

I seriously hope your forecast for a miserable summer is dead wrong.

[Sunny and Warm]

Thanks for the forecast, but without any concrete basis for this, how will we know whether you nailed it or got it right for the wrong reasons? Or if you fail, why you failed. I mean, heat from mid June to Mid September with a few brief interruptions is a given in summer, and reads like a Farmer's Almanac prediction. Somewhere, sometime someone is going to have a very hot spell between June and September, so there is little predictive value in your forecast. Severe heat means what exactly? +2F? +4F? More? Less? Over what time span? Where in the East? Everywhere? Just the NE?

The forecast is too vague for me, and will likely verify on that basis, just like the Almanac.

[JBG]

Somewhere, sometime someone is going to have a very hot spell between June and September, so there is little predictive value in your forecast. Severe heat means what exactly? +2F? +4F? More? Less? Over what time span? Where in the East? Everywhere? Just the NE?

When were those hot spells during 1992, 1996, 2000 or 2004?

[KokomoWX]

Roger, it seems you might have nailed another forecast. Congrats.

[Ellinwood]

Roger, it seems you might have nailed another forecast. Congrats.

He's doing good so far in June and the foreseeable future. Remember though that it's still only June.

[Ottawa Blizzard]

Roger is definately doing better than JB who was calling for a cool summer along the lines of 1976 and what the Japanese model was forecasting . lol!

[Roger Smith]

Thanks, this is actually looking even more intense than any year except 1936, as I just posted in another thread, the Kansas-Nebraska heat wave is setting early sesaon and even monthly records for June, of course July has been 5 F higher at most places in their all-time records, but years that are mentioned for daily records past four days are mainly 1911 and 1954, two very severe heat summers. As for 1936 coming into play, that apparently started up around this date and started to spread east around July 4-7. The 1911 major heat wave built up around now and spread east June 29-July 5, then backed off slightly to frequent 95-100 F type heat after a few days closer to 105 F.

In other words, the signal is there for this extreme heat to spread out and move east in waves. I thunk the odds are pretty high that this summer will join the top ten in the northeast and almost 100% for the central plains states, so that says about 80% likely for the Midwest. Whether you want to use mean July-August temps or extreme highs or numbers of 90+ or 100+ days, I imagine it will rank top 10 in all those indices.

As to the northern storm track, that also seems to be locking in, there have been frequent storms in western Canada that may not have made the news in the U.S., for example in central B.C. major river and stream flooding and mudflows from melting alpine snow, that have displaced hundreds of people in the past week and threatened levees on the lower Fraser River (not too serious a situation there yet, but crests have not arrived from the central regions where rivers are at 40-year highs).

Looking at the tropical forecast, we are obviously building to a high total unless the season goes dormant soon, if it even hits average numbers for July and August, then an active September would see totals of at least 18-20 named storms. I have the feeling the El nino suppression concept may begin to appear in October-November preventing the use of Greek letters again, but would not bet much on that.

[carley5]

looks like you did well. looks like a long stretch of 90's here starting thursday.

[easternsnowman]

I truly believe we better get used to this kind of heat, it is becoming the norm!

[carley5]

Roger is definately doing better than JB who was calling for a cool summer along the lines of 1976 and what the Japanese model was forecasting . lol!

wasnt JB basing his bitter cold 12/13 winter forecast on this summer being well below normal?

[JamieOber]

Thanks, this is actually looking even more intense than any year except 1936, as I just posted in another thread, the Kansas-Nebraska heat wave is setting early sesaon and even monthly records for June, of course July has been 5 F higher at most places in their all-time records, but years that are mentioned for daily records past four days are mainly 1911 and 1954, two very severe heat summers. As for 1936 coming into play, that apparently started up around this date and started to spread east around July 4-7. The 1911 major heat wave built up around now and spread east June 29-July 5, then backed off slightly to frequent 95-100 F type heat after a few days closer to 105 F.

In other words, the signal is there for this extreme heat to spread out and move east in waves. I thunk the odds are pretty high that this summer will join the top ten in the northeast and almost 100% for the central plains states, so that says about 80% likely for the Midwest. Whether you want to use mean July-August temps or extreme highs or numbers of 90+ or 100+ days, I imagine it will rank top 10 in all those indices.

As to the northern storm track, that also seems to be locking in, there have been frequent storms in western Canada that may not have made the news in the U.S., for example in central B.C. major river and stream flooding and mudflows from melting alpine snow, that have displaced hundreds of people in the past week and threatened levees on the lower Fraser River (not too serious a situation there yet, but crests have not arrived from the central regions where rivers are at 40-year highs).

Looking at the tropical forecast, we are obviously building to a high total unless the season goes dormant soon, if it even hits average numbers for July and August, then an active September would see totals of at least 18-20 named storms. I have the feeling the El nino suppression concept may begin to appear in October-November preventing the use of Greek letters again, but would not bet much on that.

Are you thinking dry for mid-Atlantic/Northeast?

[Roger Smith]

Well, summer rainfall can all come in a few days but I think the anomalies for rainfall will be generally 50-75 per cent of normal in a broad zone from Kansas east to Virginia and 125-200 per cent from Minnesota to northern New York and north/central New England, so this leaves states from about Iowa to New Jersey in a near normal position but with scorching heat and lots of sunshine, some places may find that 80-90 per cent of normal rains leave the ground as dry as 50-60 per cent might do in a less extreme heat situation. Of course, one 1955 style hurricane could overturn all those concepts in statistical terms although you still get to experience weeks of heat and drought before the heavy rains, so in terms of a forecast, it depends on whether you are talking impact or experience. Summer tourism is not dependent on one large rainfall in an otherwise dry season. By mentioning 1955 you might guess that I have my eye on the potential for the hurricane season to become hyperactive near the east coast at some point. I wonder if this might be a blend of many different extreme heat summers because many of them are in the analogue set.

[Ottawa Blizzard]

wasnt JB basing his bitter cold 12/13 winter forecast on this summer being well below normal?

He sure was! It wouldn't surprise me in the least if that forecast goes down the drain as well.

[Roger Smith]

For a cold winter, there's always that 1933-34 analogy, I guess.

FWIW the extension of my data set for this forecast goes quite cold at times in the winter months of 2012-13, although the average is a slight positive anomaly, but I reboot the model in terms of weighting analogues before final release. Let's say the preliminary estimate from my method would rule out another non-stop torch winter.

[BeauDodson]

For a cold winter, there's always that 1933-34 analogy, I guess.

FWIW the extension of my data set for this forecast goes quite cold at times in the winter months of 2012-13, although the average is a slight positive anomaly, but I reboot the model in terms of weighting analogues before final release. Let's say the preliminary estimate from my method would rule out another non-stop torch winter.

Roger

Nice call on the summer heat

We have had a lot of extremes in this region over the past decade. Some of the wilder swings of my lifetime. Last year we had the Flood of 2011 - the flood was comparable to the 1937 flood. This year we are having an incredible drought. Many farmers are not going to have a crop - very little crop to harvest.

This is my region

We are now in a drought that is likely going to rival 1988 - and you may have to go back to the 1954 event - then the 1930s. We saw the largest 1 year drop in the PDSI reading - that I could find at least.

We are baking in the 100s. Some areas have only experienced 4" of rain this year - that is simply stunning and incredible. Check out some of the KPAH (Paducah, KY NWS) rainfall data for the last few months. Little or no rain.

One item that I have noticed, with long range forecasting, is that forecasters can not predict the big events months in advance. Nobody predicted the greatest ice storm of this regions history back in 2009 - nobody predicted the greatest flood since 1937 in 2011 - nobody predicted the incredible/widespread drought of 2012. At least nobody that I know of - nobody can consistently predict these extreme events.

It seems like we can predict that a pattern favors above or below normal temperatures/precipitation. This information isn't very useful for the public.

Hopefully the science will get us there - one day - for these expensive/historic events.

The big question for our region, at this time, is are we facing a multi-year drought. You can see from the graph above that often times when you have a severe drought that it is followed by one or two years with dry conditions.

[Ottawa Blizzard]

For a cold winter, there's always that 1933-34 analogy, I guess.

FWIW the extension of my data set for this forecast goes quite cold at times in the winter months of 2012-13, although the average is a slight positive anomaly, but I reboot the model in terms of weighting analogues before final release. Let's say the preliminary estimate from my method would rule out another non-stop torch winter.

How about 1935-36?

1933-34 was very cold in eastern Ontario and northern new England, but not as cold further southwest, I believe.

I think JB is out to lunch with his hinting at an upcoming winter similar to the late 70s, unless he means 1979-80.

[Roger Smith]

First of all, since other forecasters have posted evaluations, I'm just going to say this was not the worst long-range forecast ever made by any human being or by this forecaster. As to grades, well, I'll start the bidding with around a B+ to A- and if you want to go down from there, I won't be ultra-surprised.

The main weakness was a lack of maps and maybe the tropical outlook is a bit robust although hey, 21 named storms, want to bet against it still? I figure the current likely outcome is 19/11/2 or something like that. So the tropical outlook is more like a B- than the overall forecast.

The process I had in mind of gradually building severe heat from overlapping retrograde and prograde signals is the thing I would flag as the best feature of this approach and that comes from a combination of studying the numerical analogue output and some sort of theoretical analysis based on assumptions about cause and effect which I already know are about as popular around here as a Mitt Romney speech at a poverty conference, so I'm just going to say take it or leave it on my theoretical stuff, soon enough I shall be six feet under and some other lucky dude can have a go at this theory. I've had my fun with it and except for two or three reasonably accurate seasonal outlooks, and whatever, then it's basically whatever you want to say (thought control is not in my skill set).

Hope I'm bringing you a few laughs along the way.

OUTPUT FOR AUTUMN AND WINTER

____________________________________

Oh good, you're saying, something new that you can get badly wrong so we can forget about those earlier forecasts.

Well here's the bottom line, October looks to turn warm again across large portions of the central and eastern U.S. and November looks to stay rather mild. December looks highly variable but still with a mild theme dominant. January looks "all over the place" in layman's terms and so I would say highly variable near normal outcome. This might be good for one big east coast storm but not all analogues suggest any storminess just a fast flow with frequent variations between very mild and very cold. The February analogues I have are mostly on the mild to very mild side with a brief return to more wintry weather in March before a warm spell like the one seen last March, only later in the month (towards the end).

The numerical output was extended into summer 2013 and that looks a lot closer to average than the last three, in fact June comes out cooler by 2-3 degrees.

SPECIFIC THOUGHTS ABOUT THIS PARTICULAR CASE 2012-13

________________________________________________________

The numerical analogues are one thing, but we are dealing with the specific factors of a rather obvious potential for a singularity based on the 2007-like arctic ice anomaly (there being almost none, in other words). I would note that the 2007-08 rebound concept was widely discussed and the results appeared most valid in the northeast U.S. and eastern Canada, especially Quebec-New England-New Brunswick where snowfall anomalies were locally 3-5 times normal values. Is something like that going to appear again this time? Or will the rebound effect fire off in a different direction perhaps? I think this is almost too complex to model, from any starting point, to be honest ... but since my other numerical data are showing a high-variability outcome with the implication of a very strong Pacific ridge oscillating east-west enough to provide a sharp variability of WNW to NNW flow types into central and then eastern regions, the implication of this rebound effect factored into that might be the highest variability we can imagine. So in other words, figure on a wild ride through the winter season and that should have the following impacts on eastern U.S. in particular:

** despite high variability, a rather mild theme in the background, occasional record highs, these most likely around lunar declination maxima from a study of other high variability winters. Dec 10-12, Dec 25-28, and possibly further east Jan 8-11 would be these mildest signals.

** occasional outbreaks of near-record cold modified only by the ongoing warming and possible snow cover issues upstream (prairies and northern plains could run dry most of the winter). Specific output shows last 2-3 days of December into first week of January and mid-January as best possible times for severe cold.

** intervals with severe lake effect when the needle swings to cold NW flow especially as the upper Great Lakes are likely to maintain ice-free surfaces well into later January and February due to current anomalies and projected mild autumn. Would have to imagine that the severe cold periods indicated above might be massive lake effect scenarios.

** east coast storminess index not especially high in this sort of monsoonal flow, although Newfoundland likely to get repeatedly pummelled by severe snowstorms, but there is that later seasonal rebound effect from 2008 to consider as possible source of eventual heavy snowfalls inland New England, and overall, in this sort of high-variability likely to be high amplitude flow pattern, you could probably expect one big severe east coast blizzard, if I had to guess when, would say mid-late-January, as temperature signals have greatest frequency of cold over mild analogues then. Northern max around January 24, full moon January 27 could be the signal for the winter's most intense storm(s) -- February new moon overlaps a milder signal but if closer to the time we're into a prolonged cold spell then another snowstorm possible then (would say this one more likely to run east into NS-NB). Some analogues have almost spring-like warmth in February so flooding rainfalls might have to be considered highly possible if heavy snow comes in January.

** further afield from the northeast and Great Lakes, would be looking for a rather dry winter from southern CA across to TX and across much of the central and western U.S. and nearby southern Canada, and a wet pattern in Alaska, northern prairies and NWT-Yukon, streaming southeast to produce a few local max of near normal in Great Lakes and interior New England.

To show how variable the analogue set is, the research model produces the following set from past 170 years (weighting those marked with **) ...

1834-35, 1846-47, 1854-55, 1858-59, 1865-66, 1870-71, 1881-82, 1886-87, 1887-88, 1893-94, 1894-95, 1905-06, 1917-18, 1921-22, 1924-25, 1929-30**, 1933-34, 1936-37, 1940-41, 1941-42, 1952-53, 1953-54**, 1965-66**, 1966-67, 1976-77, 1979-80, 1980-81, 1983-84, 1988-89, 1996-97, 2000-01, 2004-05, 2005-06, 2007-08.

I would be grateful if UncleW could produce some maps for that analogue set and I will comment on those maps. I suspect they will have some rather odd looking details. But they contain a large variety of severe cold snaps and record warm spells. This is why I suspect the theme will be variability more than trend. However, the weighted analogues in particular favour a mild February, both 1930 and 1954 had long record warm spells. Other analogues in the set (1977, 1981, 1984) also go from cold in January to mild in February.

Realizing that this is quite at variance with another forecast on display here, it sets up an interest "compare and contrast" situation.

[PSUBlizzicane2007]

The main weakness was a lack of maps and maybe the tropical outlook is a bit robust although hey, 21 named storms, want to bet against it still? I figure the current likely outcome is 19/11/2 or something like that. So the tropical outlook is more like a B- than the overall forecast.

This is starting to look like a major stretch considering we are over halfway through September with only two storms for the month (compared to eight for August) with no discernable systems or signals appearing on medium to long range models. Considering that El Niño should start tightening its grip in October/November, we may only be looking at an additional 1-2 storms.

[Toronto4]

Great post Roger and I appreciate your thoughts for this upcoming winter. Your description for this winter is more in tune with a La Nina event. With some saying the atmosphere is struggling to react to the current El Nino, your forecast could very well pan out. A few weeks ago, David Phillips of Environment Canada was quoted as saying this upcoming winter for Toronto will be mild in general, but it will be harsher than last winter (which was a non-winter). It will be interesting to see what transpires this winter. Should be lots of fun!