not really... all you want is a signal out that is twice whatever your noise floor is... if you're running 40mv basement, just don't trigger that low.. trigger at a level at least 85 and set to 'ingnore' 10% below....the last paradigm Egon mentioned regarding this was a server algorithm working " if signal NOT at least twice the noise level, ignore it"
This is NOT about pure audio... this is Burst Impulse Broad band energy... the 'frequencies' energy is still there... the noise is just that.. noise. I'd suggest any basement around 20-60 mv is fine. If it's much above 40 average, reduce your gains! Or downsize antennas! We're talking about the signal "grass' here, not spurious signals... The Noise is NOT disturber pulses coming from motors etc, those are 'signals' but you could count 'em as 'basement'level if they're consistently there, and set stuff to match their strength.
To state the obvious... if your basement is at 40 db and you continuously send signals that are only 70 db, you're wasting time, bandwidth, server time for 'useless' data. Which is why folks will complain that 'I registered the stroke, but the server didn't recognized it and I didn't get credit'.... or similar queries........
That 'response' is based on the 'discharge' impulse strength... and I believe it relates to the 83% level of the signal max. That is, your '83%" peak energy must be "Twice" the noise basement. Once the server has allowed what's supposed to be the discharge pulse into time-stamp processing, the remaining pulse groups of that sferic will be analyzed for 4 or so iterations of a "Zero Crossing" algorithm for TOGA computation. even though they may be significantly lower than the Discharge impulse. even if 'amplitude wise' they're buried in the nose basement.... the frequencies involved are still retrievable. So, if you've triggered on a signal further away than about 50 miles, you didn't trigger on the true discharge impulse, but a reflection of it.... and at that distance the group crossing pulses may be too weak and distorted to be processed cleanly, though the server will try manfully to 'filter' them out of the noise. So your signal isn't necessarly computed with TOGA, just the TOA of the reflected pulse,... which leads to big deviation circles...
Normal to have DC offsets, and variations in noise levels between channels... but not significant unless totally ridiculous.
In other words, don't out-think or over analyze 'the grass'...