$P = 2 \pi T \omega$, where $T$ is torque and $\omega$ is the angular velocity in radians per second. Assuming that you know the flow and RPM, you can solve for torque.
I must add a caveat -- the above assumes no losses. Any friction, or water leaking backward in the assembly after it's been pumped part way up is going to reduce the efficiency, and thus increase the torque for a given flow.
(Edit, after the fact -- sorry for not thinking to include this before):
I know almost nothing about these pumps. I could come up with the above equation purely by using conservation of energy. If you're adding potential energy to something at a given rate, then conservation of energy says that if the process is 100% efficient then you need to be expending that same amount of energy, be it electrical, mechanical, or whatever.
So all I needed to do was to (A) know the power vs. torque and speed relation, and (B) assert 100% efficiency. You can use this sort of thing everywhere -- motors, pumps, switching power supplies (that's kind of a specific example, but I'm an EE), cars, trains, whatever. If you know the power out, and you can reasonably guess at the efficiency (or look it up in a book), then you know the power in.