Anybody else seeing computer model waffling where they are?

[elagache]

Dear WeatherCat climate watchers,

In my last posting on the  how dry is California? thread - just 2 days ago - the forecast discussion was dogged that there wasn't any chance of rain for Northern California for at least 10 days.  This morning, there is now a chance of rain next week.  2-days ago the rain chances would have been 7-days out.  Two days later, now the computer models have reversed.  The GFS brings rain as early as Wednesday.  The ECMWF and DGEX models wait until Thursday before bringing rain.  To me at least this is quite a surprise.

This to me is another indication that the experts forecasting the weather are seeing phenomena that they can't compensate for.  Once more until recently, the short and medium term computer models appeared to be doing better (at least as I recall.)  Am I the only one noticing the forecast waffling more than it has done in the past?

Cheers, Edouard

[Steve]

This is the norm for us, and has been as long as meteorologists have talked about forecast models. We constantly get changes in forecasts, and the local guys pretty much say to take anything more than 3 days out as an educated guess. And models will change drastically overnight. Irks people when the NWS forecasts something severe and it doesn't happen, then they get irked when the NWS takes more conservative approach and people get dumped on. It is a thankless job.

[HantaYo]

In Colorado the models can very.   An echo anything beyond 3 days is a guess.  A lot of times if they are forecasting lots of snow it snows very little.  If small amount of snow are in the forecast this is when we usually get our heavy snows.  Sometimes even a day is a guess.  It all depends on where the L pressure goes.  This depends on the season though and if a H pressure has establised itself  to the SW of Colorado the models can be pretty accurate for more than 3 days.

[elagache]

Dear Steve, Jeff, and WeatherCat climate watchers,

I didn't realize that forecasts were even more variable further east of California.  Given the density of population, I assumed (naively) that the focus of research would have been on getting those forecasts correct.  The rule of thumb I had heard was that out to 5-days the forecast accuracy was "reasonable."  I suppose this is a case of some "reasonable things" being more "reasonable" than others! 

In the meantime, the forecast has waffled twice since I wrote my last posting.  Yesterday's afternoon discussion was gloom and doom - no rain next week.  This morning's forecast the rain is back for Wednesday.

Oddly enough, I have noticed this pattern.  The afternoon forecast discussion is frequently more pessimistic for rain than the morning discussion.  It would be easily to blame the forecaster's personality, but they do explicitly cite the latest model runs in the discussion.  If true, the models are picking up on conditions during the day that would move the jet stream one way.  Then overnight the models would pick up phenomena that brings them back to the previous morning's forecast.

  . . . This is really weird, but it has a nice upside.  You can choose your forecast!  If you want rain, only read the morning discussion.  If you want sun, only read the afternoon discussion! 

That makes perfect sense - doesn't it? 

Cheers, Edouard