The issue is visibility. Just like self-publishing a book, you’re in an uphill battle to even get your work seen. If you ever want to be cited by future publications, your research needs to be easy for future researchers to find and access. And traditionally, that has been done by consolidating studies via scientific journals.
Sure, you could try to start your own FOSS version of a journal. But that will require web developers, hosting costs, and the enthusiastic support from (and adoption by) other researchers. Because if nobody uses your system, it’s worthless for it’s intended purpose.