[supercat]

Head loss (friction in the piping and the energy required to change the direction of water in horizontal turns of the pipe) is still the controlling factor. The power needed is still proportional to the cube of the volumetric flow rate.

There are two forces one has to work against: head loss and gravity. In any given system, I would expect the power requirement to be approximately k1*V^3 + k2*V where V is the volumetric flow rate and k1 and k2 are proportionality constants. If V is much less than sqrt(k2/k1), the linear term will dominate. If V is much greater than sqrt(k2/k1), the cubic term will dominate. If V equal or nearly equal to sqrt(k2/k1), both terms will contribute significantly.

[burzum]

I agree. There are ways to add head to the pump to ensure that you don't have to compensate for vertical lifts in pipe. You can do this by attaching a surge tank above it (and the surge tank has the same level in it as the area to which you are pumping). Then, when you pump, you only have to compensate for head loss. This is very simple, and I would be surprised if they didn't do it.