I was recently asked about the coauthorship policy for JGR Space Physics. AGU has a society-wide guideline and policy on authorship. About a decade ago, the AGU Council unveiled the latest incarnation of this policy in an Eos article. The second paragraph starts off with the policy statement, which I think is pretty clear: to deserve coauthorship, someone has to contribute “significantly” to either the work or the writing of the manuscript. In addition, the policy statement indicates that the corresponding author takes responsibility that not only should just those with significant contributions be coauthors, but also that all of those with significant contributions are listed as coauthors.
Even more information is provided at the AGU ethics-for-authors policy page (note that the link near the end of the Eos article is old and broken). I discussed this page in posts ~6 months ago and ~18 months ago, but I felt it was time to write about this topic again, because it is a regular question on people’s minds. Point #10 in the list of author responsibilities is about coauthorship. It’s pretty much the same as the AGU Council policy statement in the Eos article.
Near the top of the web site is a link to a PDF of Scientific Integrity and Professional Ethics. This is a ~20 page document of how to be an ethical scientist, a small bit of which is about coauthorship. It’s actually exactly the same text as on the website link above, but you can see where it came from and the broader context of ethical scientific behavior.
There is ambiguity in defining “significant scientific contributions” towards a paper. Adding people just because they are your friends is unethical but so is leaving off those that helped make the study possible. There is a gray area there in which feelings can be hurt.
The question arises of how to police coauthorship. It is rather difficult to stop the addition of gratuitous coauthorship, because informal conversations at group meetings or over lunch could be construed as justification for inclusion as a coauthor. I think that one or two informal conversations does not rise to the level of coauthorship, but in-depth conversations that lead to an impact on the direction, content, and findings of the paper are enough to warrant coauthorship.
The reverse case is perhaps easier to identify: an omitted contributor needs to request authorship through the author, editor, society, or publisher. The person should show convincing evidence that they “significantly contributed” to the work. If the corresponding author and editor agree, then a correction should be issued for the paper changing the author list. If not, then the editor/society/publisher needs to serve as judge about who is right.
At JGR Space Physics, we generally leave it up to the main author to decide the level of significance that deserves coauthorship and rarely question authorship decisions at the editorial level during review. It is up to your good judgment and the high ethical standards of the space physics research community.