Tuesday

Accurate prediction of solar activity calls for precise calibration of solar cycle models. Consequently we aim to find optimal parameters for models which describe the physical processes on the solar surface. We use a genetic algorithm to optimize surface flux transport models for Cycle 23 using NSO magnetogram data. This is applied to both a one-dimensional model that inserts new magnetic flux in the form of idealized bipolar magnetic regions, and also to a two-dimensional model that assimilates specific shapes of real active regions. The genetic algorithm searches for parameter sets that produce the best fit between observed and simulated butterfly diagrams, weighted by a latitude-dependent error structure which reflects uncertainty in observations. Due to the easily-adaptable nature of the 2D model, the optimization process is repeated for Cycles 21, 22 and 24 in order to analyse cycle-to-cycle variation of the optimal solution. We find that the ranges and optimal solutions for the various regimes are in reasonable agreement with results from the literature, both theoretical and observational. The optimal meridional flow profiles for each regime are almost entirely within observational bounds determined by magnetic feature tracking, with the 2D model being able to accommodate the mean observed profile more successfully. Differences between models appear to be important in deciding values for the diffusive and decay terms. In like fashion, differences in the behaviours of different solar cycles lead to contrasts in parameters defining the meridional flow and initial field strength.