Calibration of the Total Solar Irradiance Data Record

Leif Svalgaard
DOI: https://doi.org/10.48550/arXiv.1812.09301
2018-12-21
Solar and Stellar Astrophysics
Abstract:Solar surface magnetic field seems to be able to explain variations in Total Solar Irradiance on timescales from hours to decades. Using magnetograms from spacecraft (MDI and HMI) and ground-based observatories (MWO and WSO) I build a composite dataset of the Total Line-of-Sight Unsigned Magnetic Flux over the solar disk stretching back to 1976, validated by excellent correlations with the solar microwave flux (F10.7) and the Sunspot Group Number. Direct measurements of TSI by space borne sensors have been carried out since late 1978. The early instruments were plagued by scattered light entering the aperture, but this construction flaw can be corrected for. At the AGU 2018 meeting, a new TSI composite has been proposed based on a novel mathematical method vetted by representatives from all current and most past TSI instruments. Although an 'official' release of the dataset has not been offered yet, a preliminary version is available. Anticipating that any last-minute changes might be minor, I compare this new version with the magnetic flux composite. It is clear that we have two TSI populations: values before 1993 that are seriously too low and values from 1993 onwards. I elect to normalize the magnetic flux (the driver of variations of TSI) to a New TSI using the regression equation for the recent population with the smallest uncertainty. With this normalization, there is now total agreement between the variation of the magnetic flux and of the New TSI as well as with the F10.7 and Group Number proxies. We now have two choices: (1) the Sun underwent a dramatic change in how its magnetic field drives variation of TSI or (2) the New Consensus TSI reconstruction does not work and the new dataset is premature and not useful neither for solar nor for climate research. Following David Hume, we should always believe whatever would be the lesser miracle, which in our case would be choice (2).
What problem does this paper attempt to address?