Converged ensemble simulations of climate: Potential trends in total solar irradiance cannot explain global warming alone Provisionally Accepted

Gábor Drótos^{1, 2*} Mátyás Herein^{1, 3, 4} Timea Haszpra¹ Imre M. Jánosi^{5, 6}

• ¹ELKH-ELTE Theoretical Physics Research Group, Budapest, Hungary, Hungary
• ²Institute for Nuclear Research, Debrecen, Hungary., Hungary
• ³Eötvös Loránd University, Hungary
• ⁴Institute of Earth Physics and Space Science (EPSS), Hungary
• ⁵Max Planck Institute for the Physics of Complex Systems, Germany
• ⁶Department of Water and Environmental Policy, Faculty of Water Sciences, National Public Service University, Hungary

We address the hypothetical question of whether an increasing total solar irradiance (TSI) trend, without anthropogenic contributions, could be sufficient to explain the ongoing global warming. To this end, the intermediate-complexity climate model PlaSim is used. To consider the total internal variability, we present a set of ensemble simulations, with different forcing histories in TSI and CO$_2$ concentration, that have sufficiently converged to the relevant probability distributions to provide a satisfactory bound on any spurious trends possibly arising from a sampling bias; similar bounds on any other unforced contributions to ensemble mean trends are also estimated. A key point is the consideration, among the forcing histories, the steepest increasing trend in TSI that is still consistent with observations according to a recent study; thereby, we essentially revisit corresponding TSI reconstructions, more than twenty years after their last modeling-based evaluation, by improving the analysis through taking care of all possible sources of error or uncertainty and incorporating data that have become available since then. Without any change in CO$_2$ concentration, our TSI trend (an upper bound on actual TSI trends) is found to be insufficient to produce outcomes compatible with the observational record in global mean surface temperature (GMST) with a nonnegligible probability. We formalize our statement for quantifiers of GMST trends through evaluating their distributions over the ensemble, and we speculate that the hypothesis about the exclusive role of an increasing TSI remains implausible even beyond our particular model setup. At the same time, if we consider a constant TSI, and the observational record in CO$_2$ concentration is applied as forcing, the simulation results and the recorded GMST match well. While we currently need to leave the question of a precise attribution open, we conclude by pointing out that an attribution of the ongoing global warming to an increasing TSI alone could be made plausible only if the temporal increase in a bias in the set of land-based instrumental temperature measurements were commonly underestimated; an assessment of the latter possibility is out of the scope of our study, as well as addressing solar forcing mechanisms beyond the effect of TSI.