this is good stuff. i think the R^2 would probably stay low regardless of the variance since we're not talking about huge shifts in the average (though enough to be obvious). what's more concerning is the median which is lower since '84 even with the increase in average during that time span (i just quickly ran some numbers). the variability is definitely still there, but what may happen if temps continue to trend upward is less snow days, but when it does, larger totals (which for now, seems to be offsetting things when the data is spread out over several winters).