Manifold interpretation of Navier-Stokes equations
Solution 1:
For the Navier Stokes equation, I am rather doubtful, because the presence of the viscosity term makes the equation not time-symmetric. And the geodesic flow is necessarily time-symmetric.
If you remove the viscosity term, you are down to the Euler equation, for which there is a well-known characterisation due to V I Arnold.
Solution 2:
Maybe Lagrangian coordinates is what you are after.