According to Schlichting for high Reynolds number the boundary layer is approximately 5 * Re ^ (-1/2).
So lets choose the chassis height as a length scale, say 25cm and a velocity of about 200kph or 50 m/s.
Then the Reynolds number is about 1250000, so it's pretty high.
The boundary layer thickness is then going to be approximately 4.5mm.
That duct really isn't much deeper than that.
Adrian Newby wrote:I proposed that Newey was trying to peel off any turbulent air in that area, and then use that low-energy air to cool the KERS.
Yeah, I think the above is the calculation for laminar boundary layer. I just don't think that there will be a great deal of air induced by such an intake - perhaps they don't need much?