I’ve been asked by a few people to share the weather forecasting sources I use. Before you bother looking at what I do, I encourage you to read Morey’s post on the topic, as he has a great deal more experience with all this than I do:
viewtopic.php?t=1193&highlight=links.

For surface forecasts I use Weather Underground,
http://www.wunderground.com/
and US Air Net,
http://usairnet.com/cgi-bin/launch/code ... t+Forecast.
The former has a wealth of information including real-time and historical data from many personal weather stations. I encourage you to spend some time exploring this site. (Thanks to Al for teaching me about it).

For winds aloft in the distant future (a few days), I use ADDS,
http://adds.aviationweather.noaa.gov/winds/
as it provides the big picture of what is predicted to happen in the country. For winds aloft in the immediate future (tomorrow and today), I use US Air Net’s rendition of NOAA’s aviation winds aloft forecast,
http://www.usairnet.com/cgi-bin/Winds/A ... se=azimuth
because it’s easier to read than the raw data.

I also look at the forecast for the position of the jet stream,
http://www.weatherimages.org/data/imag192.html
because one of my instructors pointed out that conditions can be rather less predictable when the jet stream is overhead, especially if you’re flying high mountain sites.

While I don’t yet have a subscription to XC Skies, the thing I use that is perhaps closest to it we refer to as “FSL” or “Soundings”,
http://www-frd.fsl.noaa.gov/mab/soundings/java/
This one provides winds aloft forecasts as well an interesting graph of temperature vs. altitude. Among other things this graph allows you to plot the rise of a pocket of air (say a thermal) starting at a particular temperature (say the ground temperature) and altitude (say the valley floor). Mostly though, I use it for winds aloft predictions the day before and the day of flight.

I look at the Bak40 source the day before I’m planning to fly and the Op40 on the day I’m planning to fly. This site produces loads of precision that should not be confused with accuracy. The Op40 is usually pretty good, but the Bak40 seems much less accurate.

For “Site” (i.e., location) you can enter airport codes like TUS, PHX, and FHU. I tend to use latitude/longitude coordinates:
Miller Canyon: 31.424047, -110.283942
Box Canyon: 31.821175, -110.796917
Mustang Mountain: 31.703075, -110.491989
South Mountain: 33.330906, -112.061175

I usually specify 5.0 hours of data, pick the starting time of interest, and then poke the Java-based plots button. This will produce the skew-T plot (temperature vs. altitude, skewed in some way that is not entirely clear to me). If there is an inversion in the forecast, it will be immediately obvious in the resulting graph.

Below the graph there will be a button for each of the hours of the forecast you requested. Poke the button for the time of interest, and the graph for that time is displayed. If you move your cursor around on the graph you will see numbers appear next to it. Move your cursor to a point where the temperature matches that predicted for the valley floor at the time in question and the altitude matches that of the valley floor, then left click. A line will then be drawn on the graph indicating the rise of a pocket of air given the starting temperature and altitude you specified. The point where this line intersects the red line is perhaps roughly equivalent to top of lift.

For any given time, you can then poke the “Get Text” button to produce a textual list of the winds and temperatures aloft forecast for that hour. It’s this piece that I use the most. Unfortunately, I think it’s probably much less convenient than the matrix view I’ve seen posted from XC Skies.

I also have bookmarks in my CrackBerry for most of the sites I fly so I can get an updated forecast while I’m kicking rocks on launch. For example, the one I use for South Mountain is:
http://www-frd.fsl.noaa.gov/mab/soundin ... %28none%29