Friday, May 3, 2024

April Snow and Temperature

This year in Fairbanks, the date of snowpack meltout - the first day with less than 50% snow cover, or less than an inch depth reported - was April 23.  This is just a day or two ahead of the long-term normal, despite April being distinctly warmer than normal: the month was nearly 3°F warmer than the 1991-2020 average, and fell just inside the warmest 33% of the historical distribution.

However, the snow depth at the beginning of April wasn't greater than normal - in fact it was slightly less (15" versus a 1991-2020 median of 20") - so it's not immediately clear why meltout didn't occur sooner.  But of course there's a certain amount of randomness in the meltout dates at a single location, and the measuring site hasn't remained the same over the years.

One other note on the meltout date: two of the earliest meltout dates in Fairbanks history were in 2016 and 1998 (April 9 and 10 respectively), and those two years were coming out of intense El Niño episodes.  Given that we also had a very strong El Niño this winter - almost on a par with those earlier events - it's again interesting that meltout wasn't sooner.

Let's consider briefly the relationship between April snow cover and April temperatures.  Of course this is a chicken-and-egg problem: temperature affects snow cover, but snow cover also influences temperature.  The basic relationship is obvious, with a tendency for less snow on the ground in warmer Aprils.

Notice that there's a better correlation for daily minimum temperatures than for daily highs.  This makes sense, as nighttime temperatures are prone to dropping more sharply over snow-covered ground than over bare ground.

The major outlier with record April snow depth was 1991 (42" remaining on the 15th!), but the monthly mean temperature was only slightly cooler than this year.  The incredible snow pack that month lingered in spite of the monthly temperature, not because of it.

If we examine the 850mb temperature in relation to April snow depth, the correlation is weaker: see below.  This chart more cleanly illustrates the influence of air mass temperature on snow depth, rather than the other way around, because we wouldn't expect the snow cover to greatly affect the air temperature several thousand feet above the ground in the "free" atmosphere.

In an attempt to extract the reverse causation - the influence of snow depth on surface temperature - the following chart looks at the temperature difference from the surface to 850mb.

It's a weak relationship overall, but better for daily low temperatures than for high temperatures.  Note that 1991 is less of an outlier here.  In April 1991, particularly at night, the snow pack seems to have held back the surface temperatures compared to what the 850mb air mass temperatures would otherwise suggest.

I also started looking at daily temperatures in an effort to see the seasonality (calendar timing) of snow's influence on temperature in Fairbanks, but I'll discuss those results in another post.


  1. "so it's not immediately clear why meltout didn't occur sooner. But of course there's a certain amount of randomness in the meltout dates at a single location, and the measuring site hasn't remained the same over the years."
    I think this is a big factor. In my experience snow depth is one of the least consistently comparable data points, due to the factor you mentioned plus variability of human measurement techniques, and influence of non-meteorological factors, for instance dirt or dust on the snow from runway maintenance, or a new hanger built nearby.
    Nevertheless you show interesting correlations.

    1. Thanks for the comment, Jim. I agree, too many variables.