Using Bayesian hierarchical modeling of total solar irradiance to estimate uncertainty in solar forcing since the pre-industrial

Abstract:

Estimates of changes in total solar irradiance (TSI) since the pre-industrial period vary widely due to uncertainty in inter-cycle TSI variability, sparse and inaccurate sunspot records before the 20th century, and uncertainty in the relationship between TSI and sunspot proxies. Because these sources of error in TSI reconstruction are interdependent, a Bayesian hierarchical model is developed to jointly estimate the true values of annual sunspot counts, TSI, and hyperparameters describing measurement errors from satellite TSI observations as well as the nonlinear coupling of TSI and sunspots. Daily TSI observations from satellite observation platforms, individual sunspot observations, and magnetic activity proxies, including solar radio flux and the Mg II index, are used to generate the posterior distribution. This analysis provides a range of magnitudes for solar forcing since 1610, which is of critical importance in determining the extent to which a grand minima of solar activity, such as the Maunder Minimum, may have contributed a large negative radiative forcing effect on Earth. An improved record of long-term solar variability and forcing will help distinguish between internal variability and external forcing in generating climate variability.