I ran some numbers to see what kind of heat transfer we could possibly get from this thing. The calculations are pretty simple. This is based off the heat differential between the air and the coolant. Air (assumed dry, moist air will hold more heat) holds X amount of energy per cubic foot. If you move X number of cubic feet that is Y degrees above coolant temperature you have Z amount of energy. This is the chart I came up with. The chart units are degrees F (horizontal), CFM (vertical), and watts (chart body) to compare to a block heater for example.
When I googled 'car blower fan cfm' it seemed like 200-300 seemed normal. I have an extra blower fan kicking around that I plan on using should testing continue to prove this worthwhile.
Manifold temperature will be a future test. I know my Paseo's manifold got to 580F after 5 minutes of idling. Thus I know a higher temp is possible, so the chart goes to 700F.
One last but very important note, these numbers do not take into consideration the efficiency of the air to coolant heat exchanger (I'm thinking a heater core or two will do the job). The heat exchanger will definitely create a efficiency hit. Just off the top of my head, I'd estimate 50%.
The numbers aren't super amazing, but they are substantial I think. I think its reasonable to assume we can get 400°F temp air with 200 CFM giving us a possible 5600W, but at 50% heat exchanger efficiency we're looking at 2800W. If we compare this to MetroMPG's block heater vs time chart (test on his Metro), that would give us a 70F coolant temp rise in 12 minutes. Obviously, this also isn't realistic because everything has a warm up time. The manifold isn't going to instantly be making 400F air, and certainly not at 200 CFM.
Moving forward, testing will refine these numbers and show what is really possible out of this setup.