Reconstruction of 400 years of solar activity using probabilistic programming

The solar magnetic cycle  is one of the main drivers of solar variability.  During the solar cycle, the number of magnetic active regions ebbs and flows and this in turns determines the frequency and intensity of space weather events, the structure and composition of the solar corona and solar wind, and the diffusion of cosmic rays through the solar system.   Understanding the solar cycle is challenging because it evolves in decadal timescales, severely limiting the availability and quality of solar observations.   Thanks to an ongoing effort by science historians, we have been able to recover and digitize sunspot drawings made during the last 400 years.  At present we have drawings for almost every solar cycle in the last four centuries, but observational coverage is not uniform and a lot of data gaps exist.   In this talk we take a journey through history and the scientific understanding of the Sun, as well as using probabilistic programming frameworks to integrate the past and present of solar observations to obtain the most plausible reconstruction of historical solar activity given our current observations.   We will discuss the implications on how the Sun entered the Maunder Minimum (a period of extraordinarily low solar activity in the 18th century) and what this approach tells us about the future of the current solar cycle 25.