It's not "heat haze". That appears to be an amateur enthusiast composition so he probably used raw images instead of calibrated data. There is CCD readout smear visible when a bright object is in the frame and the raw images don't have it removed so it has to be done manually. That looks like a manual effort to remove it.This is not a theory of mine, I've seen the raw data and played around with it, that explanation is from the MSL team member who works at MSSS and who posts at UMSF.com.