Environment layer planning

I’m working on animations with camera moving through the scene from a car driving viewpoint and want to find an efficient way to separate the environment into layers so I am not re-rendering the whole scene when I add a bit of additional detail to only part of the environment.

An example…
My camera/car is driving on a residential street with houses, fences, trees, gardens, sidewalks, kerbs etc etc
I rendered several hundred frames and now want to add a few extra trees here and there. Render again. I just noticed I need letterboxes on the houses. Render again.
Now my environment is complete I want to animate pedestrians walking on the sidewalk and cars on the road. Render again!!!

Today I started separating my streetscape environment into left and right corridors on different layers. A simplified explanation that I have houses/front garden and fence in far left and right corridors. Sidewalk in mid-left/right corridors. Naturestrips/sidewalk in near left/right and road in the centre corridor. I render each layer/corridor separately with alpha backgrounds and later use each rendered layer in the video editing mode to compile the final video.

This does take longer to render the overall animation but allows me to re-render only the corridor that is modified and selecting a border allows only the minimum viewport to be reworked saving a lot of time on revisions. As I only just started using layers this way I’m yet to see the full benefit or suffered the unanticipated downsides to this method so I’m hoping there may be some systematic approach others use to break their environments up to be later compiled together.

Some uncertaintly I have so far is my “corridor” method may be problematic causing parallax issues-
-bends/corners in the moving camera/car path.
-Some objects such as powerlines and some trees cross through multiple corridors

Where can I learn more about planning ahead to separate environments for easier editing and compositing etc?