Richard from Georgia asked about the correlation of temperatures in October to the following winter. As a first crack at answering this, to the left is a plot of the October mean temperature vs. the mean temperature the following mid-winter, December through February. I've divided up the scatter plot in four quadrants based on the mean temperatures for the 81 years of data used here. As you can see, there is not much correlation (R=0.13). Some cold Octobers are followed by very mild winters, and vice versa. For example, Oct 2000 had an average temperature of 22.3F, about 3 degrees below the long term average, but was followed by the warmest mid-winter in the past 82 years. In contrast, last October was very mild (mean temperature 28.9F), but was followed by a mid-winter with an average temperature of -11.0F, the coldest in years.
However, this is not the whole story. We know that the warm phase of the Pacific Decadal Oscillation (1976-2007) correlated with much warmer winters in Interior Alaska than previous decades, but October (and the autumn in general) did not warm very much at all during the same time period. I'll take at look at this complication in another post.
