It's not the bitemporality that helps, it's primary data immutability.

The event sourcing community has its own specificities (event sourcing advocates saving decisions, not outside data), but not on that aspect: if you store events immutably as this article describes, you are bound to benefit from read models that you can trash and rebuild at will.