Direct observations over the past four centuries show that the number of sunspots observed on the Sun's surface varies periodically, going through successive maxima and minima. Following sunspot cycle 23, the Sun went into a prolonged minimum characterized by a very weak polar magnetic field and an unusually large number of days without sunspots. Sunspots are strongly magnetized regions generated by a dynamo mechanism that recreates the solar polar field mediated through plasma flows. Here we report results from kinematic dynamo simulations which demonstrate that a fast meridional flow in the first half of a cycle, followed by a slower flow in the second half, reproduces both characteristics of the minimum of sunspot cycle 23. Our model predicts that, in general, very deep minima are associated with weak polar fields. Sunspots govern the solar radiative energy and radio flux, and, in conjunction with the polar field, modulate the solar wind, the heliospheric open flux and, consequently, the cosmic ray flux at Earth.

The solar cycle and its associated magnetic activity are the main drivers behind changes in the interplanetary environment and the Earth's upper atmosphere (commonly referred to as space weather). These changes have a direct impact on the lifetime of space-based assets and can create hazards to astronauts in space. In recent years there has been an effort to develop accurate solar cycle predictions (with aims at predicting the long-term evolution of space weather), leading to nearly a hundred widely spread predictions for the amplitude of solar cycle 24. A major contributor to the disagreement is the lack of direct long-term databases covering different components of the solar magnetic field (toroidal vs. poloidal). Here we use sunspot area and polar faculae measurements spanning a full century (as our toroidal and poloidal field proxies), to study solar cycle propagation, memory, and prediction. Our results substantiate predictions based on the polar magnetic fields, whereas we find sunspot area to be uncorrelated to cycle amplitude unless multiplied by area-weighted average tilt. This suggests that the joint assimilation of tilt and sunspot area is a better choice (with aims to cycle prediction) than sunspot area alone, and adds to the evidence in favor of active region emergence and decay as the main mechanism of poloidal field generation (i.e. the Babcock-Leighton mechanism). Finally, by looking at the correlation between our poloidal and toroidal proxies across multiple cycles, we find solar cycle memory to be limited to only one cycle.