Yeah, I'd say you're reaching daylight now.
First let's look at your 1000 vs 400 gph pump example. Simply because one pumps 200 gph while the other does 100 gph doesn't tell us too much. What you need to know is how much energy the motor uses. We can assume that the 1000 producing 200 gph is running at a lower efficiency since it is pumping 20% of "rated" flow vs 25% for the other. This may be a bad assumption, however, as flow vs head (and efficiency) curves differ for different pumps.
If you have a chart of efficiency vs flow (commonly efficiency gets graphed right along with head vs flow), then you can answer the question. If not, you must measure power consumed by the motor.
Regarding your "volume of water present" bit, you are entirely correct. Ultimately, volume has zero bearing if we run the system long enough to reach steady-state. All it does is factor into how long it takes to reach steady-state. If you want to pick nits, you can argue that more volume requires more surface area (tubing, reservoir, etc) and this added surface area aids the radiator in dispelling heat. Yeah, whatever. Unless the reservoir is relatively large, its convective heat loss will be peanuts next to a good radiator with sufficient flow of ambient air.
Finally, for your question of calculting "the power of moving 100 gph". Sure, power is nothing more than flow rate multiplied by delta-P. If you can measure delta-P (ala BillA's manometer explaination a few days back), then you can determine the useful work of the pump.
You also asked about density. Density alone doesn't have much bearing on flow power. In an open system where suction and discharge are at different elevations, sure, but not in a closed system. Viscosity really determines flow for a given pump and piping system. Lower viscosity = higher flowrate. As viscosity increases, pressure drop versus flowrate decreases. Total pressure drop tends should remain reasonably constant as the lower viscosity gets offset by higher flow.