Wednesday, November 28, 2018

A Multimodel Perspective on This Week's Mountain Snow

Although it is not unusual for the GFS to be wetter than the NAM, the difference between the two models over the next couple of days is exceptionally large.  For Alta from 5 AM this morning through 11 AM Saturday, the 0600 UTC initialized NAM generates 0.84 inches of water and 12 inches of snow.  In contrast, the 0600 UTC initialized GFS generates 2.84" of water and 39 inches of snow. 

There are several reasons why this is the case, but a major one is the track and characteristics of the storm that will move across the region Thursday night and Friday.  The NAM produces a trough that is displaced further south with precipitation occurring mainly near and south of the Salt Lake-Utah County line.


In contrast, the GFS trough is further north and precipitation occurs all the way up into southern Idaho.  There are also some details about the upper-level trough orientation that are different, but which I won't get into here. 


Although I do not have comparable 4-panels to the one above for the experimental FV3 model, which will eventually replace the current GFS, it too favors the more northern track with precipitation into southern Idaho. 


Another perspective is provided by the SREF plume for Alta.  There are 26 members in this ensemble and 24 of them produce at least 1.5 inches of water and fall in the range between 1.5 and 3.25 inches of water.  However, there are two members that lean toward a NAM-like solution, producing less than an inch of water.  Note how the main difference in these members is what happens after 0000 UTC 30 November, which would be that critical Thursday night and Friday period. 

Finally, the ECMWF model is producing 1.5 inches of water for the period at Alta and has a solution close to the GFS and FV3. 

The tendency in such situation is to toss the outlier solutions and focus on the meat of forecast distribution, which would probably yield a forecast calling for 1.5 to 2.5 inches of water at Alta from now to 11 AM Saturday (01/18Z in the plumes above), which would probably translate to 15 to 30 inches of snow given the anticipated snow-to-liquid ratios.  However, I've always felt that one should be cognizant of the full distribution, so a better interpretation is that 1.5 to 2.5 inches of water and 15 to 30 inches of snow is the most likely outcome, with lower odds that  the trough tracks to the south and we end up drier or that the wetter solutions verify.

My suggestion is that you all do snow dances to ensure that the NAM solution does not verify. 

1 comment: