Twenty-five years after the first useful space-borne radiometry from the ERB and ACRIM experiments, a reliable record has been compiled of small (0.1%) irradiance fluctuations, caused mainly by photospheric magnetic structures, over time scales from days to the 11-yr activity cycle. Their existence is more remarkable than it might seem – why doesn’t heat simply re-adjust quickly to flow around the localized spots and faculae, with no effect on luminosity? The answer seems to lie in the enormously longer time scale for radiative relaxation, compared to convective diffusion, exhibited by a stellar convection zone.

This surprising “thermal superconductivity” and storage efficiency of the HCZ probably also explain why efforts to detect large- scale variations in heat flux outside of the magnetic structures (by radiometry, photometry, spectrophotometry, and diameter measurements) have produced no convincing detections. Previous evidence for larger luminosity variations from stellar photometry has recently been invalidated by the difficulty in finding true solar analog stars.

This stellar evidence provided the main scientific basis for multi-decadal solar irradiance trends widely used in climate attribution studies, including the IPCC, over the past decade. These “speculative” large – amplitude irradiance trends have recently been retracted. At present, there is no evidence for solar irradiance variations other than those generated by surface magnetism, which are too small to drive present – day climate models.

Nevertheless, recent correlations with deep-sea sediments indicate a positive sun-climate correlation extending back to the last Ice Age. Forcing by solar UV and by heliospheric modulation of GCR’s may play a role, although the correlation of the UV with global temperature since 1915 is surprisingly low. Feedbacks such as clouds may provide climate sensitivity to even the small luminosity variations presently observed. But some interesting new techniques I will discuss might still reveal important irradiance trends below our present detection threshold