I don't think there's a business model there at all. Once anyone sells a forecast to anyone it's OUT THERE and hard to sell again, and this is a HARD problem, basically the definition of chaos, severely limited by supercomputer time and raw data from dedicated satellites and thousands if not millions of instruments around the world and subject to exponentially diminishing returns. Weather forecasts basically perfectly fit the definition of a public good.
Now, if you were talking establishing a new network of weather sensors and selling that DATA to institutions...
Hi, I'm an entrepreneur looking for a startup idea.
In my experience, the reason most startups fail is because they never actually solve anyone's problem. So I'm cheating and starting out by identifying a specific person with a specific problem.
So I'm asking you, what's the most annoying part of your life/job? Also, how much would you pay for a solution?