No, it can not. The problem is science long ago became a buzzword and has since been used for many things which is not science at all. The entire scientific process is about observations and experimentation and developing repeatable and predictable experiments which can be used to prove or disprove theories which are used to explain the behavior. If you can't create a repeatable and predictable experiment then it's not really science.
This isn't to say that theories can not be used to potentially explain past events and much of science is done trying to do just that, but as soon as you make a claim that some past event *MUST* have been caused by some previous event you have left the realm of science. It may be the best theory and there may not be any other understood cause but unless you have a reliable observation you can never be certain. Therefore much of science is also based on assumptions which is fine as long as you understand that they are just that... they are unproven assumptions.
It gets more complicated when theories build on each other because while it can be very helpful it is often easy to lose sight of the base assumptions or worse get into cases where your basis of support is a circular argument that theory A proves Theory B which proves theory A though almost never as simple as two theories.