Forecasting Winter Weather

[Jonathan Whitehead]

This post is for those without meteorological forecasting jobs:

How in Gods name do you accurately make forecast for winter weather, specifically more than 3 days out, when all u really have to go on is the GFS and u have no way to look at cross-sections?

 

[Robert Dewey]

At >24 hours out, there really isn't a need to look at cross sections... at least IMO. Things will probably change so much that you wouldn't even recognize a 72hr cross section compared to what actually verifies. Forecast soundings actually work just as well for diagnosing the DGZ, moisture, potential instability, etc..

In the medium range, I usually keep things simple. I look for the -4C to -6C isotherm at 850mb as a first guess for the rain/snow line. The surface map with isobars and wind, QPF, and 2M temps is also a must for me. I also keep an eye on the 500mb and 300mb levels, looking at vorticity and jet structure respectively. No need to go deeper than that when so much can change with one run...

What catches my attention in the medium range is strong baroclinic zones & negatively tilted 500mb troughs with associated vorticity... and agreement among the operational models in regards to the placement of the surface features and mid level trough.

The medium range doesn't entail too much for me... just looking at the several models (GFS / GEM / UKMET / ECMWF). I'll admit that I don't use ensembles all that much, other than to check the means and spreads.

 

[rdale]

when all u really have to go on is the GFS and u have no way to look at cross-sections?

The GFS is not the only model available to "non-forecasters." ECWMF, UKMET, CMC, NOGAPS -- you name it, they are available. GFS ensembles, CMC ensembles, etc. Free websites have them, GEMPAK has them, cheap websites have them, and so on.

And Robert's point is great - you should not be looking at cross sections at day 6. No value.