Solar cycle prediction using precursors and flux transport models

We study the origin of the predictive skill of some methods to forecast the strength of solar activity cycles. A simple flux transport model for the azimuthally averaged radial magnetic field at the solar surface is used, which contains a source term describing the emergence of new flux based on observational sunspot data. We consider the magnetic flux diffusing over the equator as a predictor, since this quantity is directly related to the global dipole field from which a Babcock-Leighton dynamo generates the toroidal field for the next activity cycle. If the source is represented schematically by a narrow activity belt drifting with constant speed over a fixed range of latitudes between activity minima, our predictor shows considerable predictive skill, with correlation coefficients up to 0.95 for past cycles. However, the predictive skill is completely lost when the actually observed emergence latitudes are used. This result originates from the fact that the precursor amplitude is determined by the sunspot activity a few years before solar minimum. Since stronger cycles tend to rise faster to their maximum activity (known as the Waldmeier effect), the temporal overlapping of cycles leads to a shift of the minimum epochs that depends on the strength of the following cycle. This information is picked up by precursor methods and also by our flux transport model with a schematic source. Therefore, their predictive skill does not require a memory, i.e., a physical connection between the surface manifestations of subsequent activity cycles.