The hard part is that turning the raw data into actual forecasts requires a lot of processing power.

Hmm, I hadn't really thought about that; I was just thinking about stations reporting current conditions. But yeah, you're right that that's the important part. Is weather modeling software another one of those areas like CAD where the state-of-the-art is locked up in proprietary shit, or is it government/scientific enough that the software is public? If one were to start building a distributed weather prediction system, are we talking about refactoring existing software to be distributed or reading research papers and implementing algorithms from scratch?