You can have two stage objects, but only the top one is going to be in effect. So if you swap them you’ll get different sets of cameras.
Or you can create your animations at vastly different points on the timeline.
Or you can do what most people end up doing, and creating different scenes and keeping careful control over the assets that go into each scene. That means keeping things in self-contained, carefully named nulls, such as ‘Cams, ‘Trees,’ ‘Bushes’–and I say carefully because sometimes you have to number the nulls and note down which is the most current set of each section of your scene. Also it sometimes is a good idea to keep your lighting self-contained inside those sets using light restrictions. The idea is that you can swap in fixed geometry to multiple camera takes without accidentally rendering with the wrong geometry or the wrong cameras.
This is all a dumbed-down version of the asset management systems that VFX shops spend so much time getting right so as not to make a huge versioning error in a setting where a lot of artists are sharing the same assets.
But we don’t have xrefs like they do, which may is a mixed blessing.