One of the most common rhetorical games creationists play, especially those influenced by those frauds at Answers in Genesis, is to erect a phony distinction between historical science, which they claim is not a science, and observational science, which they claim is the one true kind of science. It’s a way for them to deny claims of events in the past having any credibility unless there is a direct, eye-witness, personal, written account (a restriction they blithely ignore when it comes to things like the first five days of the creation week, or the life of Jesus, which is all by second-hand accounts).
So it’s always useful to collect good summaries of how you certainly can evaluate claims about the past, and how science can legitimately study historical processes. John Wilkins adds some more arguments.
To deny that we can know the past in any sense is not science. It is in effect an admission of failure, but we need not be so pessimistic. For example, we know Caesar crossed the Rubicon with his legions. We might not know how they were dressed or if it was raining that day, but we do know it happened. Likewise, we know the earth is 3.85 billion years old since the surface hardened. The evidence is there, supported by experimental observations.