As a follow-up to Saturday's post, I put together an experimental graphic in another attempt to visualize how the frequency of warm and cold extremes in Fairbanks depends on the phases of ENSO, the PDO, and the North Pacific Mode (NPM). The results provide an additional perspective and are quite informative, I think.
The charts below show the number of days each winter (65 winters, 1950-2014) in which warm (red) or cold (blue) extremes were observed in Fairbanks, with the positions of the markers indicating the winter mean phases of the oceanic temperature patterns. As in the previous post, I used the 2nd and 98th percentiles of the daily temperature distribution as the thresholds for defining "extremes". Note that 12 of the winters had no warm extremes, and 23 of the winters had no cold extremes; these winters are not plotted on the charts.
Some interesting aspects of the PDO-ENSO charts are that La Niña favors warm extremes (more than El Niño) when the PDO is negative, as I noted before; this continues to be a surprise to me. Note the high overall frequency of extremes (cold or warm) when La Niña accompanies a negative PDO. It's also interesting to see just how rare cold extremes are when the PDO is positive.
The PDO-NPM charts below show that a negative PDO with positive NPM is very favorable for cold extremes, but some of these winters also have a good number of warm extremes. The most favorable phases for warmth are positive PDO with negative NPM; this differs from the positive PDO - positive NPM setup of the past couple of winters, which suggests that the PDO phase is a better explanation for the recent unusual warmth than the NPM.