Mortality across borders

I was curious about where the yellow cedar mortality was distributed across the entire range - if the same patterns that we are seeing in AK are also found in Canada, across the entire 500,000+ ha range of disturbance.  That would imply that the same mechanism is causing mortality.  Since the mechanism implicated is a lack of snow leading to root mortality, we expect a few things:

Mortality should be found at higher elevations further south, because it should more or less track the snow-rain transition line.

Mortality should be found on all aspects everywhere, since that is  weak control on snow accumulations and persistence in these areas (given the climate).

Mortality should be found more or less at similar slopes, though there is anecdotal evidence that steeper slopes are dying in BC.  Thin soils are implicated (shallow rooting) due to the slope, but that's not entirely convincing, as there are thin soils in AK as well.  Perhaps thin soils + higher elevations mean longer cold snaps there than in AK, where the snow is deeper at higher elevations?

Mortality should be associated with winter temperatures hovering around zero.  Higher than that and the cedar likely died out earlier, or never dominated.  Lower than that and it should be safe (for now).  

The plots bear those predictions out, which is great to see, across a very wide range of latitudes:

Boxplots show distribution of decline as a function of elevation, aspect, slope, and mean winter temperature across the entire latitudinal band of this disturbance event.  Trend line is a simple loess smoothing operation.  No relationship is seen for aspect or winter temperature across latitudes; slope has a slight decline with increasing latitude and elevation has a strong decline with latitude.