There are 24 years with annual snow totals for both Val d'Isere and Avoriaz. Positive correlation means they tend to have good, average or bad seasons at the same time. 91% is a very high number, comparable to the correlations between Mammoth and June, Alta and Solitude, Killington and Sugarbush, or Mt. Rainier to Washington State ski areas.
The caveat is that 24 observations is not as many as I would like to see. In North America, where nearly all the data is month-by-month, there are on average 120 common month observations with which to correlate areas and project annual snowfall.
The other difficulty in the Alps is the sharp divide between north and south. Val d'Isere and Isola are less than 1.5 degrees of latitude apart at similar longitude and have zero correlation. By contrast Oregon and California areas 6 degrees apart are still ~50% correlated and results at comparable distance are similar in the Rockies. Zermatt/Cervinia gets both northern and southern influences, doesn't correlate well with anywhere else in Fraser's data. So I had to live with the skimpy raw data that exists for my feature article. Fraser thinks those recent years are overall close enough to the long term climate there to be a reasonable representation.
