I wouldn't pay too much attention on what other people said, not many have actually tried it in practice.
I suggest using port injection with manifold injection at the same time, until you get comfortable with it, you can decide. The effectiveness of temperature reduction by manifold injection is subject to how much water vapour the air can absorb at a given temperture and pressure. The humidity of the uncooled air has also effect of the degree of cooling.
Inlet tract cooling is only a small part of overall cooling process by water, journey after the intake valve is equally important but topic is rarely discussed - there were some discussion on this section earlier. Due to difficulties in measuring the temperature after the inlet valve, the discussion was not very detailed due of lack of evidence.
I don't have proof either but I have witnessed the effectiveness during many mapping sessions on rally cars, the onset knock event is more even amongst the cylinders compared to manifold injection. Knock events seemed to be more evident on one or two individual cylinders, I think this is due to uneven water distribution rather than hot spots or poor cyliderhead water jacket design.
You can acutally detect blocked port jet by using two flow sensors, just compare the signals between two branches. We can supply an internal filter for all jet sizes but only put them on the 0.3mm jets. The internal filter surface area is very small so it will defeat the object of preventing clogging when large flow nozzles tends to collect more debris in comparison.
I can supply you with sintered filters to be put into bigger jets but ensure the flow is PWM controlled so flow is limited.
Richard