This suggestion is cool, and something I would be all for if this was a game still in development (for the first time). But what you are hoping for is a
massive time and money investment for - from the company's point of view - a minimal return. Each season would constitute another 40-60% time per zone (rough estimates, since to be honest, while I've done some 3D modeling, I'm pretty much just an amateur) to re-skin and mutate the various models. For four seasons, that would be nearly triple the amount of time it would take to make one zone.
Organic growth is a cool idea. Though to do it right, SE would have to abandon their age old modus opperendi, which is to individually model every part of their game. The easiest way to accomplish this would be to write an algorithm (probably based on
SpeedTree) that would slowly randomly mutate the various organic environmental elements over time. SE doesn't seem to like giving up that kind of control though.
Unfortunately, it comes down to a return on investment. All of these ideas are
possible, but are they worth the amount of time it would take to create them? Most developers seem to think not. Personally, I think if they spent the first year of the game working on extra terrain models for each zone per season, rather than working on the first expansion right off the bat, I don't think the backlash would be too bad and it would make the environments really interesting. The problem for us though, is that FFXIV has already been delayed years, they can't afford any more.