Title: Hemispheric Coupling: Comparing Dynamo Simulations and Observations

Abstract: Numerical simulations that reproduce solar-like magnetic cycles can be used to generate long-term statistics. The variations in N-S hemispheric cycle synchronicity and amplitude produced in simulations has not been widely compared to observations. The observed limits on asymmetry show that hemispheric sunspot area production is no more than 20% asymmetric for cycles 12-23 and phase lags do not exceed 20% (2 yrs) of the total cycle period. Independent studies have found a long-term trend in phase values as one hemisphere leads the other for ~four cycles. Such persistence in phase is not indicative of a stochastic phenomenon. We compare the findings to results from a numerical simulation of solar convection recently produced with the EULAG-MHD model. This simulation spans 1600 yrs and generated 40 regular, sunspot-like cycles. While the simulated cycle length is too long and the toroidal bands remain at too high of latitudes, some solar-like aspects of hemispheric asymmetry are reproduced. The model reproduces the synchrony of polarity inversions and onset of cycle as the simulated phase lags do not exceed 20% of the cycle period. Simulated amplitude variations between the N and S hemispheres are larger than observed in the Sun. The simulations show one hemisphere persistently leads the other for several successive cycles, placing an upper bound on the efficiency of transequatorial magnetic coupling mechanisms. These include magnetic diffusion, cross-equatorial mixing within elongated convective rolls and transequatorial meridional flow cells. One or more of these processes may lead to magnetic flux cancellation whereby the oppositely directed fields come in close proximity and cancel each other across the magnetic equator late in the solar cycle. We discuss the discrepancies between model and observations and the constraints they pose on possible mechanisms of hemispheric coupling.