You should at monthly nao, ao, enso values by month for those 20 years you targeted. See the progressions and do a 3-6 month trailing evaluation to find a correlation coefficient to the previous 6 months. Even than I think the correlation to snow on a local basis is incredibly low. If you used snowfall on a much larger geographic basis it would be higher, but with how localized storms can be it's like a monkey throwing darts. We can but root for chances Sent from my SM-G970U using Tapatalk