That is an excellent observation about the use of horizontal or vertical polarization for geostationary satellites -- obviously that works well in that case (much higher frequencies and zero relative motion).
Perhaps it is a combination of relative motion and signals passing through different parts of the ionosphere that causes the problem. When talking about VHF signals coming from the ISS in particular there isn't any random attitude change (unlike a tumbling/spinning satellite). Certainly there is a slow and predictable change in attitude due to relative motion. If that is all it was, then a single RHCP or LHCP antenna would do the trick 100% of the time.
The ARISS crew highly recommends the use of CP antennas (certainly understandable) AND polarity switches. I pushed back on that requirement due to cost and availability issues, but went ahead and installed the polarity switch on our CP antennas. With certainty I can tell you that at random times it made a huge difference in downlink signal level.
The effect usually happened close to the horizon -- perhaps the VHF signals are undergoing polarity changes due to tropospheric ducting. But the use of a CP polarity switch proved to be very useful. Sometimes it might not happen at all. And at least once it happened near the point of closest approach.
it is an interesting issue for sure, and not much literature on it.
73, Bob, WB4SON