John I share some of your concerns and I thought about the possibility of electrical noise, in similar way that stray sun light was affecting the optical sensor in your case. One can always miss things, maybe something similar is happening in my case.
About what Q allows or not in terms of omega-dot I don't see any problem, think of a sudden small fall of the bob along the rod, energy doesn't change but the frequency changes in one period.
About the clock being a sensitive seismometer, well the clock is a seismometer, we can't certainly ask a pendulum not to measure an acceleration.
The picture which makes more sense to me now is that the sensor and the pivot are linked together well and both are mechanically attached to the building structure which vibrates statically, with wavelengths much longer than the rod. What originates the signal is the inertia of the suspended bob, which is doing its job.
The other interpretation on the table, the one of the "infrasonics", which I said, obviously joking, we could rule out taking the air out of my apartment, can't stand if the pendulum air interaction is the same as the one we see in stationary state. Some time ago I evaluated empirically the clock pressure sensitivity at .45 μs/mbar, you showed us recently a similar estimate, if the interaction would be similar in this case, the same at 1/2 Hz as at 10-5 Hz, the corresponding pressure variation at a SD of 45 μs would pop my ears all the time when it's a very windy day.
Two more considerations: the detector/pendulum could be undersampling and I overestimated the SD
Second: I haven't understood yet if one can use the usual first order perturbation theory conclusions (Ayrie's expressions to be practical), when there is so much of high frequency perturbations. Same thing which was bothering me with the, not by chance, aerodynamics of the open/close door effect.
I might have easily said too much, don't ever think I really understand what I say.--