About the only advantage to knowing the effect on economy due to wind, would be to adjust the fueling/timing to maximize efficiency. But, you'd really be playing with the same set of parameters as you would for anything else that creates load on the drivetrain, such as a hill, increased rolling resistance and the like. So, there's no real reason to try to simulate a wind, when it's really engine load vs fuel used that you want to improve.
And, as long as you're operating in closed loop, you'd probably have to "bias" the O2 sensor values somehow to allow much of a change, which I imagine could be done.
Bill's already said he can set the A/F ratio leaner than stoich if the load is low and it's set richer than stoich when the load is high, so using a dyno just probably allows him to refine a "theoretically derived" tune for a particular vehicle. I imagine it's time for all of us to truck on over to Georgia for a dyno tune!
- Jack