Training on IFR cold weather decision making?

[iamtheari]

For those who have experience and talent reading Skew-T diagrams, where do you get the diagrams from (I only know of Unisys's site) and how far from a sounding station is the information at all useful? I'm 140nm east of one station, 120nm northwest of another, and 230nm north of a third. Is a Skew-T going to help me out at all other than when I fly over a sounding station?

 

[Clark1961]

http://rucsoundings.noaa.gov/

The link above can generate skew-ts from the ruc model. I think it is what most folks rely on. Obviously not perfect but then a sounding observation is only valid for that moment in time. Any forecast is just that, a forecast.

 

[RotorDude]

For those who have experience and talent reading Skew-T diagrams, where do you get the diagrams from (I only know of Unisys's site) and how far from a sounding station is the information at all useful? I'm 140nm east of one station, 120nm northwest of another, and 230nm north of a third. Is a Skew-T going to help me out at all other than when I fly over a sounding station?

The beauty of Skew-T is that although the actual sounding grid is coarse in both space and time, the processed results are of infinite resolution, allowing you to get the "virtual sounding" at any point in space and time, so you can pretty much forget the original coarse sounding grid. Of course, like any forecast tool, the further into the future you go (up to 2 weeks or so), the less reliable the prediction, and there are several (normally three) different prediction models to choose from, so you need to learn to use this tool and correlate it to your own experience to get maximum utility from it.
When used properly, Skew-T rocks, outdoing pretty much all other forecast tools (although the smart pilot will use all of them). I personally rely on Skew-T for soaring, as well as icing prediction. I wouldn't launch non-FIKI IFR/IMC into near-icing conditions without this tool, to complement the others.
Here is the desktop tool I currently use for Skew-T, which has tons of other amazing stuff in it for the weather addicts. There is also a great Android App, though limited to 24 hours or so.

 

[RotorDude]

Keep in mind the "virtual" soundings (forecasts and analyses) have a limited resolution. They are from a forecast model that has a 13 km native resolution (25 km practical resolution). And the data is interpolated to a 40 km grid (Op40). So when you choose an airport or lat/long you are getting the sounding at the closest grid point.

Also they are not the same as a radiosonde observation. If a balloon ascends through a cumulus cloud that will be represented directly on the sounding. However the forecast sounding will only represent the environment that the cloud will grow within. You have to lift a parcel to determine the potential for cumuliform clouds (bases and tops).

Makes sense. The site I linked above (twisterdata) has "infinite resolution" input, in that you can move the cursor on the map arbitrarily, but I can see that the output is interpolated into a coarser grid, as you say. Although, for practical purposes, any finer resolution would probably be useless as the prediction error would be greater than any interpolation benefit.
Also, as far as the "virtual" vs. actual radiosonde, I think for our purposes it's better to have a more averaged environmental result than the actual radiosonde since (unless things are very homogeneous) those specific clouds it's rising through will be gone by the time we get there.

 

[RotorDude]

Yes, it's still a point forecast.
Be careful here. A model will smooth out important details in the atmosphere that you can see very clearly in a RAOB. Weather prediction models are not capable of modeling every cloud, especially when it comes to moist convection. [snip] It's also useful for understanding how the atmosphere is poised (high RH in the surface layer where radiation fog may form).

Are you saying that relatively thin but high RH near the surface will be averaged out by the models? I certainly see those very often, so what am I seeing?