I see this more as a statistical question than a GIS one. Or maybe it's just my training. Assuming there is data representing each tree (labelled something like Tree1, Tree2, etc. in the database) and there are also fields for each of the states (survived, born, died) in each layer. I would write up a research question for each, something like this:
What is the difference in survival status for each tree between layer 1 and 2?
What is the difference in survival status for each tree between layer 2 and 3?
What is the difference in survival status for each tree between layer 3 and 4?
What is the difference in survival status for each tree between layer 4 and 1?
Swap out "survival" for "born" then "died" for your remaining research questions. Then query the database for each combination and permutation and dump the answer into a new field with a code depending on the status outcome. For example, for the new field representing survival status between layer 1 and 2 a tree survival could be a 1, no survival would be a 2. Repeat for the other two states and the difference between the layers.
Others with more GIS experience can tell you how to render these new fields on a map such that all four layers (do they represent time?) and all three conditions (survival, born and died) for each tree can be best visually represented.