I'm aware that the coherence length of sunlight can be observed with a Michelson-Morley interferometer. I haven't seen any experiment that separated the different colors within sunlight to determine if the coherence length varied with frequency.
My thoughts are that the sun is emitting bursts of electromagnetic radiation that vary in frequency, polarization, and phase. The only way I can see fringes forming is if a single burst is split by the mirrors producing two waves that are perfectly correlated in frequency, phase, and polarization. Once one arm of the interferometer is lengthened so that the two paths exceed the length of the EM bursts then fringes would be impossible. How could selecting specific frequencies cause the EM bursts to be correlated, especially in polarization?