I've always been rather sceptical of the need to keep the tank full, as I suspect that the number of air changes in the ullage over the course of a year is fairly small. The driving force for exchange is the slow variation in external air pressure as weather fronts go past, and probably more importantly the rapid fluctuation in pressure due to wind. A few years ago I did a study on the exchange rate in a vented nuclear reactor containment. I can't remember the details but the annual number of volume exchanges was quite small. In a fuel tank, the smaller the ullage the less water vapour will be drawn in. As it doesn't hurt to keep the tank full I try do so, but don't get obsessive about it.