I'm working on a number of projects right now, some are near "completion" (as much as that is ever possible!). I've been thinking a lot about preprints lately. When would the appropriate time to upload one? At the same time as you submit a manuscript to a journal? My fear is that all my previous work has changed quite a bit in the review process (which seems normal enough), so if this is normal for my work should I avoid preprints at this stage? What if the results are modified by incorporating suggestions of reviewers?
I don't know about ecology specifically, but in biology in genearl (my experienece is molecular and computational biology and genomics), I always submit a preprint when I first submit the article to a journal for review. Yes, the paper sometimes changes between submission and final publication, but that just gives you something new to talk about in your tweet-torial about the final publication that was different to when you tweeted about the preprint!
Almost all preprint servers will allow you to upload new versions of papers after initial submisison, but be aware that some journals allow you to submit articles that have been preprinted, but then do not allow the upload of updated versions after that point.
As was said in a comment, some journals don't accept papers that are published as preprints (although it seems to become unusual these days). So if you have one or a few journals in mind, check their policy before putting out your preprint!
You don't want to have a preprint out there that is embarrassing, so wait with publishing a preprint until you're really convinced of it (and maybe have the material discussed with a few people).
If these issues are out of the way, I'd say publishing a preprint is fine and sometimes you want to do it early in case others work on the same thing and a time stamp can tell that you were there first. Also, if your preprint is out some time before submission, you may discuss it with some more people and improve your journal submission. (I do realise that one may sense a slight contradiction to the first issue, but everything can be improved, even if you are at some point convinced that it's good enough!) People also may cite your preprint in case the publication process takes so long that the journal version doesn't exist at the point where others want to cite you.