Monday, October 21, 2013

Will Warm Weather Continue?

In view of the exceptionally warm weather of the past week or so across interior Alaska, I thought it would be interesting to look at similar past events to see if the history reveals any consistent patterns in the subsequent weeks.  In other words, how does the weather typically evolve after unusual warmth at this time of year?  I looked up the warmest weeks ending between October 10th and 31st in Fairbanks, based on the mean daily temperature departure from the 1981-2010 daily climate normals.  Here's the list of weeks (only one per year allowed) along with the temperature anomaly in terms of standard deviations:

Oct 18-24, 1938      +2.86
Oct 7-13, 1969        +2.04
Oct 13-19, 2013      +2.01
Oct 18-24, 1980      +1.89
Oct 25-31, 2002      +1.75
Oct 21-27, 1981      +1.70
Oct 17-23, 1957      +1.68
Oct 22-28, 1936      +1.66
Oct 25-31, 1954      +1.61
Oct 23-29, 1942      +1.53
Oct 19-25, 1987      +1.49

The chart below shows the temperature anomaly over the 60 days following the end of each warm week, with each line representing a different year.  Admittedly it's a bit difficult to follow each individual year with so many lines, but focus on the thick black line which shows the median of the 10 years.  It's no surprise to see that temperatures tend to stay above normal for several weeks after a very warm October week in Fairbanks.  However, it's interesting that the median reaches its highest point at four weeks (28-29 days) after the end of the warm October week; indeed, several of the historical events (e.g. 2002, 1936, 1957, 1981, 1938) produced very warm conditions again (relative to normal of course) in the period from about 3 to 5 weeks later.  In fact, only 3 of the 10 years were not significantly above-normal in that period (1942, 1969, 1987).  This could be a statistical quirk, but it's not unreasonable to imagine that a natural periodicity in the global circulation might favor the reintensification of a pronounced feature such as the ridge over the northeast Pacific.

The other interesting aspect of the data is that half of the years experienced significantly below-normal temperatures around 8 weeks after the end of the warm October week; so there is no suggestion that unusual warmth is likely to linger into a second month.

In summary, the history of similar past events appears to suggest that relatively warm conditions are quite likely to continue in Fairbanks for another 4-5 weeks, but by mid-December it's entirely possible that below-normal temperatures will emerge.


  1. Very interesting analysis Richard. Perhaps the CFSv2 is tapped into this?

  2. By now you know I have to does it look after a normal, or unusually cold October, over the same period? Do the data respond in an opposite or a similar trend?


    1. Here are the numbers for Anchorage (since they are handy). They are departures from the 1981-2010 climate normal period – not actual temperatures.. Warm Octobers are nearly always followed by warm Novembers (R-squared of 0.14) but Decembers are typically colder than normal.

      Year Oct Nov Dec
      2013 10.7 ??? ???
      1936 7.7 8.0 -7.3
      1925 7.4 6.6 -1.1
      1923 7.3 4.7 -7.4
      1937 7.2 4.3 -7.1
      2002 6.8 12.7 4.9
      1938 6.8 -0.3 0.1
      1969 6.7 1.5 9.5
      1979 6.7 11.4 -8.9
      2009 6.0 -1.9 0.8

    2. Gary,

      The results for very cold spells in the same period are not quite as pronounced. Unusually cold conditions are likely again around 15-18 days after the end of the October cold week, and then again (and especially so) around 36-45 days. But other than that, there are mixed signals, with cold and warm about equally likely.

    3. Thank you folks for the very interesting analyses. This Blog is a daily morning read of fascinating info. Now to see what 2013 brings under this scrutiny.