Back about 1976... I found out that a funnel showed a sharp increase in electrical activity, and somewhere I actually have the original plans and specs for a tornado detector. The activity is NOT always in the form of "lightning', but emf pulses generated from the rotation.... they have significant predictable signatures, but I just do not remember the exact deal. I think the Satellites are using similar, + visual, + data from a couple of other orbiting sensors
these figures below are incorrect, and I don't remember the exact research... if I stumble upon the stuff again I'll forward it...
Normally, these erratic pulses/pulse groups would be 'random', or
pulse groups less than 10pps or so within 5 seconds, less than 4 (5 second groups/minute) using one minute window.
If the cloud had rotation, other criteria, the pulse rate would increase to
Pulse trains >10pps occurring within 3 seconds, more than 12 (3 second groups/minute) in a one minute window
As I say, those paradigms aren't accurate.. I just don't remember... but you get the idea.
The device used a small AM radio, off channel, audio out to simple counting circuits. (555 driven, I think) It worked.
I sat on Skywarn mobile location several times and watched it trigger, saw a wall or rotation, and a couple of times a funnel. Lot of m home-made stuff and schematics etc got lost in the move in 2002...
It seems that that particular project might been derived have come from an article/project page in "Radio-Electronics" magazine.. I was heavy into such stuff during my career as an Electronics Tech.
Interesting thing.. the Blitzortung system, especially BLUE, lends itself well to this, and at least once a month, I wonder what happened to my long-lost project and data. I've got enough component inventory to build one up in an afternoon if I could locate the data. it was pretty simple, in a way.
Mike