Winter of 1905-06

"The days were much lengthened now, and the worst of the winter was done.  There would still be cold and storm, but hardly again of the same intensity and duration."

So wrote the worthy "archdeacon of the Yukon", Hudson Stuck, in his 1914 account of "Ten Thousand Miles with a Dog Sled".  I recently acquired a copy and would highly recommend its splendid descriptions of the winter environment of interior Alaska as well as its admirable study of both native and non-native inhabitants of the time.

 

The first section of the book gives a detailed account of a 2200 mile circuit in the winter of 1905-06, taking four and a half months, from Fairbanks to the Kobuk River and Kotzebue, then to Nome, and then back to the Yukon River via Unalakleet and finally back up to the Tanana River and Fairbanks.  Stuck's account of the weather is remarkable reading, because it just so happens that the winter of 1905-06 produced some of the most severe cold in the recorded history of interior and northern Alaska.  He got into trouble quite early on the trip, unwisely starting out on the leg from Circle to Fort Yukon with an inexperienced guide and the thermometer at -52°F.  Of course we know Stuck was a man of remarkable strength, as he co-led the first group to summit Denali in 1913, but nevertheless his experience of the episode near Circle is almost beyond belief:

"We had been out about twenty hours in a temperature ranging from 52 to 60 below zero, had walked about forty-four miles, labouring incessantly as well as walking, what time we were with the sled, with nothing to eat - it was too cold to stop for eating - and, in addition to this, one of us had been in water to the waist, yet none of us took any harm."

It's interesting to compare Stuck's observations on the trail to the temperature observations that were made at the three central interior sites that were operative that winter: Tanana, Rampart (about 60 miles up the Yukon from Tanana), and Fairbanks.  It was the first winter of mostly complete temperature measurements from both Rampart and Fairbanks.

Right off the bat we notice that Tanana reported -70°F on three days in January 1906, and only three such days have ever been recorded there subsequently (two in 1947, and the famous -76°F in 1989).  As for Fairbanks, the -64°F on January 19, 1906 has only been exceeded by the -66°F in 1934.  And the rather short period of record from Rampart (1900-1933) has no other winter with as many days reaching -60°F (8 days in January 1906).

The charts below show the reported daily high and low temperature for the three sites.  Stuck reached Circle on December 6, and made the decision to leave on the 7th despite the cold.  Clearly Stuck's thermometer was more or less correct, as Fairbanks reached -60°F on the 7th; this is tied for earliest on record for such cold in Fairbanks.

 

Stuck had intended to spend Christmas at Bettles, 250 miles up the trail from Fort Yukon, but ended up spending the holiday in a village on the Chandalar River to provide assistance during a diphtheria outbreak.  He departed on the 27th: impeccable timing with respect to the next cold snap at the end of December.  Temperatures of -40° to -50° were seen for much of the time until his arrival in Bettles on January 12.

The archdeacon then reports that the temperature never rose above -50° for two weeks in Bettles, delaying any further progress in that time.  Looking at the charts above, it's clear that there was in fact a brief interlude in mid-January when the central interior saw warmer conditions, but Bettles is quite a long way from Tanana, Rampart, and Fairbanks.  It's possible that a few days of cloud cover produced the warm-up farther south, while Bettles stayed locked in the cold.  Stuck describes the weather as "of quite exceptional severity" and even reports a rare occurrence of wind blowing at -58°F.

Finally, on January 26, Stuck judged that the cold snap was over, based on a downward trend of the barometer: "the barometer had been falling slowly for a couple of days, and I was convinced the cold spell was over".  This proved to be correct, as there was no more severe cold that winter, and indeed March turned out to be extremely warm, causing difficulty of another kind for Stuck's return journey.

What was going on with the weather patterns in January 1906?  It seems there was an abundance of high pressure over Arctic Russia, and we know that there was persistent low pressure near Iceland (positive NAO phase).  Here's the estimated MSLP anomaly from the 20th Century Reanalysis:

Monthly Weather Review reports that "Temperatures below the average were reported from Yukon Territory and the extreme northern districts of British Columbia, but in all other portions of Canada the normal was exceeded, and in many localities to a very marked extent."

My interpretation is that the pressure gradient across the Arctic basin transported cold Siberian air masses into Alaska, and there was a persistent but localized trough over western Canada.  The 20th Century Reanalysis indeed shows this pattern, which is pretty classic for winter cold in Alaska.

However, the model doesn't quite capture the magnitude of the cold for interior and northern Alaska - see below.  This isn't surprising, as to my knowledge the model only uses surface pressure and sea surface temperature measurements to construct its estimate of past conditions; the surface temperature data are not fed into the model.  Evidently the model doesn't quite have the trough in the right position, or perhaps its physics are inadequate to produce the surface-level cold that actually occurred.

 

Could something like this happen again?  Yes and no.  Given the right seasonal climate influences and random chance, a similar circulation pattern could probably recur in the modern climate, bringing an intense focus of extreme cold to Alaska for weeks at a time.

But I think that even the coldest sites are very unlikely to see cold as extreme as January 1906 in the coming centuries, because of the increase in high-emissivity gases aloft, primarily CO₂ and water vapor.  Extreme cold in interior Alaska occurs in a very thin layer of air at the surface, and so a "small" change in the composition of the air aloft makes a big difference.  Taking Tanana as an example, there hasn't been a single instance of -65°F since 1991, whereas it apparently occurred 22 times from 1903-1947, and 11 times from 1948-1991.

So - hats off to Hudson Stuck, who seems to have come through it all unscathed: a remarkable man, in remarkable circumstances.