> I doubt any extant or near-future forecasting system is able to tell you that a specific storm cell will be in a specific spot at an exact minute all that far in advance.

Well, this is exactly my plan, and I'm going to do it ASAP. Prototype should be done by summer. There are millions of little weather stations all over Florida that nobody is using (smartphones). By connecting those stations together and pulling the live sensor data into forecasts, I do in fact believe that these small thunderstorm cells should be very well-forecasted if we take in this dense dataset.

This forecast system doesn't exist yet, you're right. But it should exist in the next couple months/years, and when it does, it should be able to provide this kind of forecast.

BUT only if that is worth money to a group like SpaceX - is it? That's really my question. I know I can build the technology, but will anyone like SpaceX buy, and at what price?

How accurate would it be? Even if you can predict with, say, 95% accuracy that a thunderstorm will cause a launch scrub, SpaceX might want to give it a try anyway in case that 5% comes true. They want to launch ASAP, and they'd look dumb if they called it off only to find that the weather was fine.

Accurate forecasts are no doubt extremely valuable in many cases, but for this specific case I don't see it changing much.

What do cell phones have that are particularly useful to detect lightning? Or look at other sources for lightning, then thousands of barometers to predict movement?

Cell phones aren't needed to detect lightning. It's the pressure measurements that are potentially valuable in making a more accurate map than can be gathered by scattered weather stations.

http://www.lightningmaps.org/realtime shows lightning in realtime as they're detected by stations across the planet. If you turn on the "stations" display, you can see detectors in Utah spotting lightning in the Gulf of Mexico.

To cryptoz (can't reply, probably too deep?): I was going to argue that it would be terrible difficult to build up that smartphone network. But then I saw that you do PressureNet (awesome tool btw!) and, oh well... looks like you've definitely got the network part done already.

OTOH, why not just ping @elonmusk on twitter?

Well thanks! But actually, building the network is indeed a big challenge. We've spent the better part of our 3 years just thinking and implementing methods to do it. We've arrived at an SDK model [1] where we work with app partners to integrate simple sensor code and contribute to our network. So we're able to grow quite quickly compared to a single app to download, but it's still not easy getting the necessary scale. We're aiming for about 10 million sensors, currently have 0.5 million. So we're getting there.

I've sent @elonmusk a few messages over the years but no replies so far. I am @vallesmarineris, so I think one day he'll notice that I already live on Mars and we'll chat.

[1]: https://github.com/Cbsoftware/pressureNET-SDK

preassurenet is cool!

It always reminds me about this interchange: https://news.ycombinator.com/item?id=8292091

Probably easier to google up someone on the launch team and ask them.