A 48X24 tank would have been more consitent as some of these lights are mean to cover more than 24 inches and it would allow you have have extra data points concerning spread. The guys at BRS use a grid. I think a grid made of acrylic would do the job well.
something like
YOu could probably get a large piece of plexi cut it down to the tanks dimensions, route a grid on to it with a router going a few millimeters into the surface. Sick some rare earth magnets on the intersections you want to sample and have a magnet on the probe that stick to that magnet. The magnets would help with alignment and keep the positions consistent. YOu could also buy four acrylic rulers to set at each corner to set the different heights with out the worry for a level since all corners should be at the same measurement. An acrylic rod with magnets on both sides held by magnets on the outside of the tank could be used to hold the height.
Unless you are going to go with static points of measurements. I think the guys over at BRS TV have a rig they use, maybe it's worth shooting them an email see if you can get your hands on it or plans to make one. They use a 24x24 cube except when they are measuring lights meant to be wider than that.