We do have velocity variations (with reference to magnitude and direction) within an individual pore; as well as along the various (tortuous) pores (along the direction of fluid flow).
But still, all these microscopic variations in magnitude and velocity (somehow or magically) get cancelled or nullified; and what we have is the (mostly) well defined - only one - resultant direction of fluid flow - in a saturated groundwater system at the macroscopic scale.
Is there an effort to prove mathematically that this final - single - resultant fluid flow direction at the macroscopic-scale results from the conglomeration of velocity variations at the microscopic-scale - in a typical saturated groundwater system?