At the very end of this article which I'm not bothering to read because I've seen nothing but trash (if not outright AI slop) from earth.com in the past, they write:
The study is published in the journal bioRxiv.
bioRxiv is not a journal. It's a preprint server, and calling it a "journal" either means they have an inflexible template to slot this into reading "This study is published in the journal [blah]" or less generously that they genuinely don't know this, which is concerning as fuck.
Here's the actual preprint, but I want to emphasize that preprints like this haven't been peer reviewed. You should really wait until the paper is released in a peer reviewed journal to get the best, most accurate version of it (which earth.com couldn't be bothered to do, instead choosing the sweet, sweet ad revenue knowing 0.01% of their readership would actually read the original).