This is a very good question, and what follows is just my understanding (so not guaranteed correct).
First, let’s exclude images and sounds, which observers can license independently. The licensing we’re concerned about here is only for the text and data in an observation.
As others have noted, copyright law in several major countries clearly does not regard facts as subject to copyright. (In the United States this was most clearly affirmed in the 1991 case Feist v. Rural).
But there are enough ambiguities at a global level that it doesn’t make much sense for a small non-profit such as iNat to attempt to enforce a particular legal point of view on every type of information added by every observer in every jurisdiction.
Even a U.S.-based user might conceivably add original copyrightable material to an observation—let’s say I see a particularly beautiful limpet and I’m inspired to write a haiku in the Notes field. The easiest way for iNat to allow each user to preserve (or release) their rights in these cases is to give them a choice of licenses in the same way as for photos and sounds.
Alternatively, iNat could make users accept an agreement to surrender their rights to such data, in the same way as you might see on many commercial websites, but it’s a core part of the iNat ethos that users own their observations, so allowing them to choose their preferred license is a far better fit.
As a researcher you should be able to use all factual iNat data with little concern. It certainly is good practice to provide citations for specific observations: that’s courteous to the observer, demonstrates that you’re not claiming credit for the observations of others, and allows readers to follow up on your sources. But none of that is required by copyright.
In some circumstances you can even reproduce pieces of original, copyrighted content. For example, the “fair use” provisions in U.S. law would allow you to quote small portions of the Notes or Comments text from an observation, even when these clearly reflect the individual expression of the author.
But if you public a book of copyrighted haiku that you extracted from the Notes fields of litigious and deep-pocketed iNat users, then you should expect to find yourself as the defendant in a copyright lawsuit. The ability for users to choose how they license their observations is the easiest way for iNaturalist to avoid becoming your co-defendant.