In the MP Low Light Shootout, there was a lot of testing below 1 Lux. Although this is a great analysis of the extreme end of the "environment harshness" scale I would think many/most installation will add IR illumination if the scene ever dropped to these levels anf a usable image was required at <1 Lux conditions. When IR is on the effective level of illumination increase greatly and the results of the shootout become less useful when comparing the different sensor pixel counts (5MP, 2/3MP, 1MP).
From the images of a indoor camcorder test below (Sony flagship SR1 with 1/3" CMOS):
One can see that usability of low light (read evening/early morning/during storms) color images rapidly deteriorates below about 5-10 Lux.
Therefore, IMO, an given that IR will be used at night, and given the decreasing price gap between 1/2/3/5MP cameras, an important decision making criteria will be the performance of the camera in 5-10 Lux lighting conditions.
So the question goes: has anyone done comparisons between between 5MP vs 2/3MP vs 1MP cameras at lighting levels around 5-10 Lux where color images are still usable but have begun to deteriorate significantly? Did you find that in these situations that the 5MP unit was not much worse then 3MP and even 1MP models?