[SteveP]

Guys,

If I may, I'd like to inject some numbers here so as to demonstrate the error that is involved.

As we know, the formula for drag force is F = .5 rho v^2 Cd A where rho is the density of air. If we do a coastdown test and measure the force used to overcome air drag at a known speed, we can rearrange the formula to calc Cd (assuming we know the area of our car):

Cd = 2 F / (rho V^2 A)

Suppose we do a coastdown test and determine that at 70mph (31 m/sec) we have an aerodynamic drag force of 78 lbf (347 N). Air density varies with temperature as follows: at 0C, density is 1.292 kg/m^3, at 10C it is 1.247kg/m^3, at 20C it is 1.204kg/m^3, and at 30C it is 1.165 kg/m^3.

If we do the calc using a density of 1.292 kg/m^3, and assuming an area of 20 sq ft (1.86 m^2), we get (if I haven't botched the arithmetic) a Cd of .3005; if use do the calc using 1.165 kg/m^3, we get a Cd of .333. That's a big enough difference to swamp the effect of a typical aero mod, so more accuracy in determining air density would seem to be in order (as would similar increases in the precision of measurement of pressure, temperature, and humidity).

As to the effect of altitude, rising from sea level to 5,000' (1524m) changes the density by about 16% which amounts to about .2 kg/m^3--in other words, more than a 30 degree (C) rise in temperature.

Water vapor also changes the density, but I don't have the numbers offhand.

--Steve