Rick Thoman pointed out today on Twitter that this Sunday will - in all likelihood - be the 5th Halloween in the past 6 years to have less than an inch of snow on the ground in Fairbanks. For a span of 75 years - 1941 through 2015 - only one year managed the same feat. (Incidentally, that year was 1962, which bears more than a passing resemblance to this year in terms of alignment of various aspects of the global climate system. Interesting.)
The disappearance of Fairbanks snowcover at Halloween is consistent with a downward trend in October 31 snow depth that seems to have started more than two decades ago; it's been 24 years since more than 6 inches was on the ground, but prior to 2000 about a third of all years had at least that much snow.
Here's Rick's excellent graphic:
Total October snowfall in Fairbanks also shows the same kind of decline. Before 2009 it was quite normal to have 10" or more of snowfall in October; this happened in nearly 50% of all years. But amazingly it hasn't happened for the last 13 consecutive years. Based on the 1930-2008 climate, and assuming each year is independent of the last, the chance of less than 10" happening 13 straight times is less than 1 in 1000. So this appears to be a highly statistically significant change.
Interestingly the ERA5 reanalysis captures the preponderance of low snow accumulations in October over the last 12 years; the map below uses a 1951-2010 median as the baseline "normal". According to this data, the anomaly has been most significant over the Bering Sea, but it extends across southern Alaska to the Canada's west coast and northwestern regions.
The lack of October snow does not principally reflect dry weather, at least not in the Bering Sea region; ERA5 actually shows more wet than dry weather in this area, and since 2009 Fairbanks has seen above-median liquid-equivalent precipitation in 5 of 13 years, including 4 very wet Octobers (2012, 2017, 2019, and this year).
Rising temperatures are a much more plausible explanation for the loss of snow; the ERA5 map below uses a 1951-2010 average for the climatological baseline. In Fairbanks, all but 2 of the last 12 Octobers have been warmer than the 1930-2008 mean, and more than half have been over 5°F warmer than the old normal.
Alaska isn't the only place to have seen a paucity of October snow in the past decade or so. Here's the ERA5 map for a global domain: note the similar low-snow signal from eastern Canada across the northern North Atlantic and in much of Russia.