Re-evaluating solar irradiance reconstructions: No evidence for large secular trends

Abstract

Most current irradiance reconstructions suggest only a modest secular increase in TSI between the 1700 and 1986 activity minima ($<1$ Wm$^{-2}$). However, 2 models, the CHRONOS and Penza et al. (2024; PEA24) models, suggest substantially larger changes of 2.1-5.9 Wm$^{-2}$. Although the two models differ in their architecture, both use cosmogenic isotope records combined with neutron monitor data to describe the long-term irradiance variability. However, cross-calibrating modulation potential reconstructions from cosmogenic isotopes against neutron monitor data remains highly uncertain. We reassess the origin and the magnitude of the large secular trends in the CHRONOS and PEA24 models. We update the CHRONOS and PEA24 models using recent heliospheric modulation potential and open solar flux reconstructions based on geomagnetic data and neutron monitor measurements, which provide a more reliable connection between cosmogenic isotope and neutron monitor records. These allow the secular component of the reconstructions to be extended consistently to the satellite era and compare the resulting TSI reconstructions with direct TSI measurements. We find that the original CHRONOS and PEA24 reconstructions substantially overestimated the secular variability in irradiance. When constrained by direct TSI measurements, CHRONOS yields a TSI increase of about 0.1-0.44 Wm$^{-2}$ between the 1700 and 1986 minima, while PEA24 returns about 0.2-0.25 Wm$^{-2}$. Our analysis indicates that the previously inferred large secular trends arose primarily from improper linking cosmogenic isotope and neutron monitor records, together with issues in the adopted smoothing approach. The updated reconstructions presented here point toward a relatively modest secular increase in TSI since the Maunder Minimum, likely below 1 Wm$^{-2}$, consistent with the majority of current irradiance models.