Don't Bet on the GFS (Or the "Euro" for that Matter)

I frequently see people on social media talking about a medium- or extended-range forecast produced by the GFS.  Alternatively, they might talk about a forecast produced by the ECMWF Integrated Forecast System, the so called "gold standard" of medium- and extended-range forecasting and commonly referred to as the "Euro" or ECMWF.

Don't take the bait.  

It's not that those forecasts aren't useful.  They can be, but only when viewed with the range of possibilities produced by an ensemble forecast system.  

The reality is that a GFS forecast (or a "Euro" forecast) is only one solution.  It represents one of many possible outcomes based on what we know about the state of the atmosphere, ocean, land-surface, and ice cover when we start the model.  

As a result, the GFS produces a credible forecast, one that is physically realistic and believable.  It just might not verify with great veracity.

Let's take a look at the GFS forecast initialized at 1200 UTC Sunday.  The 159 hour forecast valid 0300 UTC 25 October (9 PM MDT Saturday) shows a deep trough over the interior western United States with a potent cold front draped across Northern Utah.  Freshies for Sunday.  Hooray! 

However, the GFS forecast initialized at 0600 UTC Monday is less exciting.  the trough is much weaker and over Wyoming and Montana.  The front is less potent and northern Utah is in more of a brush-by situation.  

When it comes to precipitation and snowfall, the details matter, and no single model forecast can reliably predict those details at such lead times.  Not the GFS.  Not the Euro.  Additionally, at such lead times, it is not uncommon for GFS forecasts to flop around a bit, forecasting something snowy one run, drier the next, and snowier again the next after that.  Such forecast whiplash is a reality of the chaotic nature of the atmosphere.

With that in mind, let's take a look at the downscaled North American Ensemble Forecast System guidance for Alta.  At lower left, one sees that the 52 members produce a wide range of outcomes.  Eight members produce more than 20" through 0000 UTC 26 October (1800 MDT Sunday), 9 members 10-20", and the remaining 35 members less than 10 inches, some less than 5.  

Such an ensemble forecast is not unusual for Alta.  There are no sure bets in this part of the world and situations like this with a digging trough in northwesterly flow are especially uncertain at these longer lead times.  

We've actually seen this movie before.  Remember last weekend's trough?  Very similar.  That doesn't mean it's going to play out the same way.  What it means is that we need to recognize the range of possible outcomes and be cautious about betting on one possible outcome at these long lead times.