Great point! While we didn't show this here, we've done similar analysis where we transform the snowfall data to make it less skewed and use just linear regression to look at the ENSO/snowfall relationship. In that framework the main centers of action are retained.