He's making a stupid argument.
Turbulent flow is chaotic and extremely difficult to accurately model. Laminar flow is much easier to model and understand and close enough to reality in most important cases. (Bikers aren't bumblebees.) Wind tunnels are designed to test and improve models so have to be carefully controlled and reproducible. That's why they (generally) have laminar flow.
If one wants to study turbulent flow (high Reynolds numbers), one can certainly do so with various tricks.https://en.wikipedia.org/wiki/Reynolds_numberhttps://en.wikipedia.org/wiki/Wind_tunnel
The aerodynamic properties of an object can not all remain the same for a scaled model. However, by observing certain similarity rules, a very satisfactory correspondence between the aerodynamic properties of a scaled model and a full-size object can be achieved. The choice of similarity parameters depends on the purpose of the test, but the most important conditions to satisfy are usually:
* Geometric similarity: all dimensions of the object must be proportionally scaled;
* Mach number: the ratio of the airspeed to the speed of sound should be identical for the scaled model and the actual object (having identical Mach number in a wind tunnel and around the actual object is -not- equal to having identical airspeeds)
* Reynolds number: the ratio of inertial forces to viscous forces should be kept. This parameter is difficult to satisfy with a scaled model and has led to development of pressurized and cryogenic wind tunnels in which the viscosity of the working fluid can be greatly changed to compensate for the reduced scale of the model.
In certain particular test cases, other similarity parameters must be satisfied, such as e.g. Froude number.
The Coefficient of Drag has a definition. If he wants to redefine the term, then he's talking about something else.https://en.wikipedia.org/wiki/Drag_coefficient
tl;dr - Everything's complicated when you look at the details. But it seems to me that he's throwing up smoke.
HTH a little. Good luck!