Could it be that I found the reason that my lightning detector rebooted on strong sparks? When I ran the project without the programming cable connected I stopped getting resets. Connected it would reboot. Perhaps some kind of ground loop through the laptop I am using for programming? In any case, I seem to escaped the executioner on this particular problem.
I also started to move the algorithm from the Python code running on my Linux box back into the micro. I seem to have the strike detection code working. Now I need to implement the warning, danger, and all-clear storm code. With those implemented, my Linux box will need to process very few events. Simply pending on a one byte notification of a storm warning, danger, or all clear. Perhaps I can get to it tonight.