In light of recent posts on winter rain in Bethel and elsewhere, I thought it would be worthwhile to see what the new ERA5 reanalysis data has to say about winter precipitation type in Alaska. I've mentioned the ERA5 data on previous occasions, for example here:
The ERA5 dataset includes a precipitation type indicator, which makes it a simple task to assess the frequency of each category across the 0.25-degree latitude/longitude grid. The six categories are: Snow, Wet Snow, Rain/Snow Mix, Rain, Freezing Rain, and Ice Pellets; the determination is based on the modeled surface temperature and the temperature profile aloft. Here is some more information on the model algorithm:
Based on the hourly ERA5 data for 40 winters since 1979-1980, I simply added up the number of occurrences of each category of precipitation type for the months of November through March. Below are maps for the four simple categories, i.e. snow, rain, and in-between. Note that a precipitation type is assigned whenever the model has a non-zero precipitation rate, no matter how small; so for example the model indicates snow falling more than 90% of the time over the Arctic waters, but rates would be extremely low much of the time. Click to enlarge the maps.
Here are the two other categories, freezing rain and ice pellets, with a different scale to reveal areas of low frequency. For reference, a frequency of 1% corresponds to 36 hours per winter.
And here are the rain, rain/snow mix, and "wet snow" maps on the same "zoomed in" color scale:
It's interesting to see relatively high frequencies extending around the north side of the Alaska Range to the middle Tanana River valley (e.g. Nenana and Fairbanks); the spatial distribution looks realistic to me, although the absolute values of the frequencies might be biased high or low (more on that in another post).
Finally, here's the sum of frequencies for everything other than dry snow. I'd be interested to hear readers' perceptions of these results - feel free to leave a comment below.