Is it possible that the tertiary mirror accumulates heat from all the starlight VIS and IR , that the large primary is collecting? How is the effective temperature of the mirrors monitored? have you considered FLIR cameras looking at the tertiary from an off lite path angle? Watt per steradian per second ^2 estimates my prove useful to determine scale of possible solid state cooling devises applied to the rear of the mirrors.
This is a possible cause and imagined solution. like CCDs that are electrically cooled, and sometimes Liquid cooled, the Optical elements may need to be cooled to improve the quality of the light reaching the LSST camera. Are you having trouble cooling the detector in crowded fields?