There are no ZFC or arithmetic axioms, for example second-order arithmetic, in the physical world. Why can I solve differential equations in the electricity subject?
We may also assume the world has discrete time and space, or just we don't care about the physical rule in the small scale. Then the differential equation is something approximate to a difference equation, and we don't need real numbers of infinite digits, and its existence in real world.
Assuming the consistency of ZFC, we could prove there are provable error bound between the solution of difference equation and the solution of differential equation.