November 1963 vs. November 2013: Is Weather History Repeating Itself?

By: By Nick Wiltgen

Published: November 20,2013

 

 

 

 

 

 

Decades ago, before the advent of modern supercomputers, one of the ways that meteorologists tried to forecast the weather was to find similar patterns in the past, and infer future outcomes based on how those past patterns played out. This method is known as "analog forecasting."
Today, we have powerful supercomputers that simulate the atmosphere using sophisticated mathematical formulas, and they are increasingly enabling us to forecast the weather accurately many days in advance.
(MORE: Sandy: Triumph of the Models)
Still, sometimes it's interesting to use the old method, if only as a curiosity. That's particularly true now, as the weather pattern heading toward Thanksgiving looks eerily similar to the pattern 50 years ago. We'll get to that in a bit – but first, another example of how "analog" forecasting hinted at the recent tornado outbreak in the Midwest.

Nov. 17 Outbreak: Weather Pattern Gave Early Hint

A forecast of upper-atmospheric weather patterns valid Nov. 14, 2013, made by the American GFS model nine days earlier on Nov. 5, 2013. By Nov. 5, all major computer models agreed that an unusually strong zone of high pressure would be near Alaska nine days later. (Source: NOAA)

• ENLARGE THIS IMAGE

Earlier this month, on Nov. 5, I noticed all of the major computer models were forecasting a massive ridge of high pressure in the upper atmosphere over the Aleutian Islands of Alaska on about Nov. 14 (see inset map). It so happens that the Climate Prediction Center (CPC), an arm of NOAA, publishes analog dates for upper-atmospheric patterns expected one to two weeks in advance. These analog dates correspond to times when similar atmospheric patterns were observed in the past.
So I looked at the CPC's 10 analog dates for the expected pattern and found that most of them coincided with tornadoes either on the same day or within about three days after the pattern occurred. In a few cases, the tornadoes were unusually far north, such as a tornado in southern Wisconsin on Nov. 1, 1971, and another even farther north in Shawano County, Wis., on Dec. 1, 1970.
As it turns out, three days after that unusually strong high-pressure pattern showed up over Alaska this year, a major tornado outbreak raked states as far north as northern Michigan on Nov. 17.
(MORE: Recap of Nov. 17 Outbreak)
Now, nobody could have predicted that specific outcome so far in advance. After all, some of the CPC analog dates showed tornadoes on the Gulf Coast and one or two had little or no severe weather at all. But it was at least an early hint that there could be a significant weather event sometime in mid-November.
It's also worth noting that this was an analog forecast based on the outcomes of very powerful computer models – not really the same kind of analog forecasting meteorologists did many decades ago.

Eerie Coincidence: Nov. 20, 1963 vs. Nov. 20, 2013

On Wednesday morning I noticed that Edmonton, Alberta, Canada was unusually cold. The temperature had plunged to -32ºC (-26ºF) as of 9 a.m. MST, which seemed unusually cold for this early in the season to me.
So I consulted Environment Canada's weather records for Edmonton and found that the all-time November record low at Edmonton International Airport was -35.6ºC (-32.1ºF) exactly 50 years ago on Nov. 20, 1963. What a coincidence!

Surface weather maps for Nov. 20, 1963 (top) and Nov. 20, 2013 (bottom). Except for a strong low pressure system near Canada's Hudson Bay on the 2013 map and a weak frontal system in the southern U.S. on the 1963 map, most of the high and low pressure centers and fronts are very similar in strength and position. (Source: NOAA)

• ENLARGE THIS IMAGE

That coincidence piqued my curiosity even further, so I looked up NOAA's daily weather map for that date. And that's when the coincidences started multiplying.
As the maps on the right show, most of the features present on the Nov. 20, 1963 weather chart were present again in the analysis on the morning of Nov. 20, 2013. In both cases, there was arctic high pressure of about 1030 millibars over northern Alberta (hence the similarly cold conditions in Edmonton); a low pressure of about 1004 millibars over northern Idaho; a weak low along a kink in a frontal boundary bringing rain to California; a weak high pressure center over the Four Corners region; a weak low over Kansas with precipitation occurring over parts of the Central Plains; and strong high pressure over or very near upstate New York.
The only differences were a weak frontal system analyzed across the South on the 1963 map, and a fairly strong low pressure center near Canada's Hudson Bay on the 2013 map.
Still, the similarities between the two maps are quite striking – and it turns out the upper-air pattern (not shown here) is also very, very similar, especially considering we're talking about two days separated by precisely half a century.
Given some of the talk of a Thanksgiving storm along the East Coast, I thought I'd take a look ahead and find out what came of the 1963 pattern.
First, similar to what is expected this weekend, the cold air in western Canada plunged south into west Texas within about three days, sending temperatures tumbling (though not as low as we're expecting this time around).
Then, on Nov. 28, 1963 – coincidentally, Nov. 28 was also Thanksgiving Day 50 years ago – weak low pressure started developing off the Gulf Coast of Texas, just as many computer models are forecasting to happen this year, albeit closer to Tuesday the 26th.

Surface weather map for 1 a.m. EST on Nov. 30, 1963. (Source: NOAA)

• ENLARGE THIS IMAGE

That low then intensified significantly as it moved into Georgia on Nov. 29, 1963, and became a powerful 975-millibar storm centered over New Jersey by early Nov. 30. It was accompanied by strong winds and rain across much of the Northeast; the snow was confined to places well inland such as Syracuse and Pittsburgh, and it wasn't particularly heavy. Boston surged into the 60s for a few hours overnight Nov. 29-30 as the storm pulled warm air north on a southerly wind.
Is this a sign of what we can expect for Thanksgiving week in 2013?
Well, computer models are in broad agreement on forecasting some sort of significant low pressure to come up the East Coast during Thanksgiving week, albeit a little earlier than the 1963 storm did.
While the strong similarities in the upper-atmospheric patterns now versus 50 years ago do suggest that the next several days should be broadly similar, it's not at all unreasonable to believe that the subtle differences between now and a half-century ago could play out to a meaningful difference between next week's weather and that of Nov. 24-30, 1963.
It certainly seems that there will be some sort of a storm near the East Coast – but whether it's mostly a wind and rain maker, or a more ominous snowstorm, or just a storm that quickly zips out to sea with less impact on holiday travel – well, those details have yet to be resolved, even by today's powerful supercomputers.
(WATCH: Change Your Thanksgiving Travel Plans?)
Stay with The Weather Channel and weather.com to see just how the holiday week's weather will play out.
MORE: Vintage Air Travel Photos

38 / 187

Circa 1956: American playwright Arthur Miller and his wife, actress Marilyn Monroe wave as they board a plane. (Hulton Archive/Getty Images)