One way to think about this kind of fluids problem: Information about
obstacles in the flow tends to propagate as waves. Depending on the
nature of the waves involved, each will travel at a certain velocity
*with respect to the fluid*. This means that the disturbances will, on
the average, flow downstream of the object causing the disturbance. If
the flow is slow enough, some information about the disturbance does
travel upstream of the object causing the disturbance. For example, you
might see a rise in water level just in front of the obstacle.
A book you might want to take a look at:
An Album of Fluid Motion, by Milton Van Dyke. Lots of good pictures
of flow around obstacles.