Great explanation AZD. I would add that although there may be some reduction in heat when throttling the discharge, there is also a corresponding loss in cooling flow. IME, the loss of cooling is greater than the heat reduction and when restricted more than a token amount (my pure guess is ~20%), results in the pump running warmer. A great example of this is the use of MJ1200s in powering media reactors. As they are often restricted to optimize the flow through the reactor - the loss of flow (cooling) results in the impeller and shaft running warmer and can dramatically increase the amount and rate of calcium build-up and result in increased maintenance. Another option is to by-pass some of the discharge water back to the sump. Although you don't gain the energy reduction, the pump runs cooler.
I mentioned the above cooling issue as I've been running a WaterBlaster HY7000 for about the past year as my return pump and have had problems with it seizing about every 6 weeks due to calcium deposits in the bushing. I currently have it restricted about 30%, but am going to work on a bypass too see if it extends the time between tear-downs and cleanings.