http://www.lpi.usra.edu/meetings/5thMars99/pdf/6127.pdf

The most interesting (and challenging) part of their observations is the large variation in brightness of the sky shown in figure one. It seems like a model of a 'perfect' illuminant would need to consider exactly what time of day it was, to get an individual contour map of the sky for that time. As the sky shows not insignificant brightness and its spectrum (at least at the times where dust content is high) varies than the direct solar spectrum, one needs to consider the total brightness imparted to the surface (especially to shadowed areas) from the sky's inherient brightness.

It is interesting to note the large variation they describe in the illuminant difference between direct sun and shadowed regions.

I've gotten the sense, through reading the papers associated with Pathfinder and other missions, that there is still much to be learned about the seasonal changes in atmospheric brightness. The difference between the sky brightness during dusty times and (as is true for MER now) relatively dust free times could be large. I'm sure there is still a great deal of research into these conditions which are ongoing now, probably through the subfield sky pans they have been taking every couple weeks on MER. Without calibration data, its hard for us to tell what they might be finding, but heres hoping the atmospherics group is getting good data which will help us with this question in the future.