Since you mention that, what do plate boundaries do to groundwave? I recall hearing that the San Andreas fault attenuated AM groundwave propagation. Michael posted right after you, mentioning his history of hearing numerous San Francisco stations in Sacramento, but without knowing which side of the fault their sticks are on, that doesn't tell me a whole lot. (Also, I would assume that if it did have an attenuation effect, it wouldn't necessarily be so severe as to stop propagation to such a short distance away as Sacramento.)
If this is a factor in reception, more important than the San Andreas would be the Hayward Fault, which runs under the Warren Freeway (Highway 13) and under the University of California football stadium. Mike would be on the other side of the Hayward Fault from those stations. I actually was, too, when I lived in Oakland because I was on the east side of 13. The Hayward Fault hasn’t had a major earthquake since 1868; experts have repeatedly said it’s overdue for one, and there are fairly frequent small quakes along it that tend to come at 3 or 4 o’clock in the morning.
I’m a bit dubious about faults being a factor in AM ground wave radio reception but I don’t think a study has ever been done on it, at least not I’m aware of.
I agree with David, a very interesting topic. I'd add one other consideration: is a fault that's got a lot of pent-up unreleased force (i.e. a fault like the Hayward, where after ~150 years there's tons of pressure just begging to be released) different that a fault that's had a temblor event in recent years and released whatever forces had built up? Would one or the other inhibit signals more or less, or would that be a non-issue? I wonder if that's ever been studied, and if not, might the Hayward not be a great place to try observing pre- and post-quake changes? Maybe with stations like KNBR (super-strong low-band omni), KGO (mid-band, figure 8, moderately strong to the east in its side lobe) or KZDG (high-band, limited to the east)? All three of those have their transmitters and towers within a few miles of each other, between the San Mateo and Dumbarton bridges, so at least that aspect would be an apples-to-apples comparison.This is a truly interesting subject, and has not been well-studied. Our points of reference are generally AM stations that have randomly moved to the "other side" of a major fault line, and there is usually no comparative operation to contrast data with.