I believe the small rear spoilers trip the laminar flow to induce turbulance. Turbulent flow reduces separation of flow drag at the rear, at least to my foggy recollection.
Tom & Pnieder,
I remember GM's comments regarding Cd to have been that they wished meet or exced the Prius Cd. Looking at the prototype and the EV1, I think your original assumption of 0.25 is more likely. The 7MPH headwind doesn't make sense to me. I can see this to be the case for two lane roadway situations, but most two lane roadway driving is at low city speeds. Most higher speed driving is on divided highway driving, and I'ld bet that if there is a divided roadway only study it will indicate that there is tailwind. Not that the following has a place in your calculation but should be considered: in reality, much commuting for work is done in heavy traffic which has even better aerdynamics. The value of plug-in EV's will be amplified in this type of driving, mostly from regen and no ICE idling.
I have look closer at the study, but if it calculates 29KWh draw with 500W accessory load and 60mph cruise as Tom posted in the public blog then there must be something off. Tesla with a bit higher Cd only draws about 14KWh at 60mph. At first glance, my guess is that the following assumptions are the problem: "The Traction Inverter, DC-DC Converter, and gear power efficiency are each 90%." This would be a combined efficiency of 72.9% for these components. Tesla claims their battery to wheel efficiency is better than 80% and this includes rolling resistance.