Hi Sarah,
We do something similar to this with our bird monitoring. We have a Survey polygon layer, and when the user creates features within the polygon (points, lines, polygons), we use an intersection function within a python script to generate keys for all the overlapping features. Relationship classes created in Pro before we published the feature service then automatically create relationships based on those keys. You could also create keys automatically using a GlobalID and GUID fields within your relationship classes origin and foreign keys. We like the python method because it skips the step of a user having to click on the Survey plot first in order to add related data to it. Our parent Survey layer is static and is created at the beginning of the field season, but you could have a parent layer that is created on the fly by your biologists as well. Once you have a relationship class for a feature, you can experiment with contingent values within Field Maps to populate other features or use python to populate attributes. Let me know if you need more detail on any of this.
Good luck!
Travis
