Living and Learning

Our first storm of the year is in the books.  Alta-Collins came in with about 1.46" of water and, per resort reports this morning, 8" of snow. Snow depth on the ground at Alta-Collins sits at 6", although that is a point measurement and I suspect there is a good deal of variability.  Due to the warm nature of the first part of the storm, snow cover at the base is lower and perhaps an inch or two based on web cams.  

I like to look back at our model-derived forecasts so we can identify capabilities and limitations for future improvements.  Let's first look at the forecast I used for my post these used from my two posts covering the storm.  From Wednesday (A Storm System with Lots of Moving Parts), the 06 UTC GFS forecast was going for a storm total of about 1.3" of water and just over 12" of snow.  I've reproduced that forecast below.  

And below is the 00 UTC Utah Snow Ensemble forecast showing a range for water equivalent of about 0.75" to 3" with means around 1.5" and 1.75" for the ENS and GEFS, respectively, and for snow a range of 6 to just over 30" with means of about 15" and 18", respectively.  Note that I am ignoring the precipitation produce by some ensemble members next week which is well after this storm period.  

So, the GFS water equivalent was just a bit below and, while we can't verify the ensemble spread and probabilities with one storm, the ensemble means were just a bit higher than observed. 

The real trouble in these forecast was the conversion of water equivalent to snow amount.  The forecasts of snow-to-liquid ratio in this instance were simply too high.  This is something that we have seen previously in warm storms in the eastern US, especially those beginning with no preexisting snow cover and warm, unfrozen ground.  This leads to melting and densification of snow on the ground, which we have not explicitly accounted for. 

Such situations don't occur much at Alta-Collins except in the early season, we we actually noticed this looking at storms in the eastern United States.  Behind the scenes we have been developing a system to forecast snow-to-liquid ratio across the continental US and testing it in the experimental Rapid Refresh Forecast System (RRFS) Ensemble.  Below is a forecast for the storm that affected northern New York and New England last week.  I'll focus on northern New York where snow amounts are annotated for Whiteface Mountain.  This forecast called for 14.6", and the NWS Burlington forecast office that forecasts for northern New York reported that the received 15". That sounds like a great forecast, but that 15" is for the summit (4800+ feet) and our RRFS-derived forecast is for about 2500 feet where accumulations were lower.  In addition, a look at the forecast below shows widespread accumulating snow, whereas weather cams showed primarily snowfalling but no accumulations on the ground.   

So, we are living, learning, and taking our medicine during these warm storms.  Often, snow accumulations during such storms can vary depending on the surface and this is something we have also been thinking about, such as providing two or three different accumulation forecasts depending on whether or not the surface is grass, pavement, etc.  

Concerning this most recent Utah storm, perhaps we did a bit better at upper elevations?  If you are insane enough to venture high, let me know.  Note that I do not encourage skiing.  Not only are these scant amounts, but the flow direction was probably not optimal for wind transport into the main chute.  Someone probably skied it anyway.