Forecast Ensembles For Discussion

August 12, 2010

It’s time for some meteorological fun…let’s talk ensembles…good times.  You DEFINITELY need to click on the image to see what it is, and even then it’s a bit hard to read, but good stuff for the weather techie.  The top three images are the 12z GFS, ECMWF, and CANADIAN (GEM) models for 12z Tuesday…or 5am Tuesday morning.  These models ran this morning.  Those are 500mb heights (around 18,000′).  Notice the dramatic difference between models.  The ECMWF rams an upper-level low right into our region Monday night and Tuesday, putting a quick end to the heat wave.  The GEM is similar to the GFS, keeping it hot through at least Tuesday.  But there’s a bit more to the story.  Look at the “ENSEMBLES” from each model.  As mentioned in a previous post here, ensembles are multiple versions of the same model run…they often give us an idea if the “real operational” run of a model is an outlier. 

Looking at those ensembles something interesting shows up.  The operational runs of both the GFS and ECMWF (especially this one) both move the upper-low much closer and have it deeper than the “mean” of the ensembles.  It really sticks out for the ECMWF.  That casts some doubt on whether there will be a big cool-down Tuesday and Wednesday as it shows.  Due to this we’ve stuck with the 100 degree temp on Monday and put in a slower cool down Tuesday-Thursday of next week than what the ECMWF or some other runs of the GFS show.  We’re really trying to avoid model-riding, but that pony just keeps beckoning…can’t have much more fun than this with a mostly sunny 7 Day forecast!

Obviously the details of the 2nd half of our heat wave have very low confidence right now.  Those upper-level lows are a real pain! 

We DO know that hot weather is still on the way, and we probably won’t break any records.  The record highs at PDX for Saturday through Tuesday are 102, 100, 102, 105.  Those are from a heat wave in 2008 and another hot spell in 1977.  I think Monday (102) could be close.

Chief Meteorologist Mark Nelsen