Thanks for the further clarification of purpose.?
It's not enough to get any real purchase on the problem, but OK that's your business not mine.
I'm just saying that the tolerances on real spark gaps, and the many unknowns that a good model might erm, model, do suggest that it's not the easiest of approaches.?
You might get 20% repeatability.??
There are many established methods for measuring very small currents at very high voltages -? stick to the path if you can, it avoids solving inessential problems.?
Bonkers