I don't have the references handy, but IIRC "flu season" and winter diseases in general were essentially resolved by a model that takes absolute (not relative) humidity and gives the probability that aerosols (e.g., a sneeze or breath) will remain in air for a while, or fall to the ground.
IIRC, same computations correctly predict flu season (and other viral diseases) for the last 10 years or so, but fail to predict covid-19 dispersion (and in particular, predicted a drop during the summer of the northern hemisphere, which did not happen) - which may indicate either that the model is borked, or that the primary dispersion vector is NOT, as commonly assumed, aerosolized -- in fact, there are quite a few pieces of evidence pointing in that direction. (Not that it's impossible, just that it's not the primary one).
Given we know aerosolization is a vector I'd take humidity to be confounded with something else. Even then, unless experiment reveals it you can't take simulations as anything except hypothesis generators.
Serum vitamin d concentration vs virus damage would be an interesting trial.
We do not know that it is the primary vector, though. We know that the fecal-oral route is also a vector, one that’s historically a big problem in nursing homes, and plays a nontrivial part in spread through food preparation (unlike e.g. with flu which is essentially all aerosol with short life).
The spread data is not properly reflected in recommendations IMHO and likely for political and politically correct reasons.
IIRC, same computations correctly predict flu season (and other viral diseases) for the last 10 years or so, but fail to predict covid-19 dispersion (and in particular, predicted a drop during the summer of the northern hemisphere, which did not happen) - which may indicate either that the model is borked, or that the primary dispersion vector is NOT, as commonly assumed, aerosolized -- in fact, there are quite a few pieces of evidence pointing in that direction. (Not that it's impossible, just that it's not the primary one).