Meridional circulation dynamics: comparing observations and results from EULAG global 3D MHD simulations

As time progresses, Helioseismology is being able to peer into the deeper layers of the solar convection almost to tachocline depth. While differential rotation is well mapped in radius and latitude, the mapping of the meridional circulation (MC) profile (a much slower large scale flow) still presents a challenge to current measurement methodologies. A clear sign of these difficulties is the different profiles obtained by different groups (especially in depth). This weak large scale flow is one of the key ingredients in current solar dynamo theories and its temporal behavior and spatial morphology has a large impact in the modeling of the large scale magnetic field responsible for the solar cycle. Dynamo modelers try their best to incorporate the constraints imposed by observations but the shape of the MC has been a subject of debate (given the lack of observational data). Besides that, only in the last 5 years or so, the temporal variation of the amplitude of the MC has been incorporated into mean-field dynamo models. The complete temporal and spatial profile of the MC is therefore something very important for mean-field dynamo modeling.
A complementary (and more recent) way of modeling the solar dynamo is through global 3D MHD simulations of solar convection. These type of simulations are now starting to provide some answers about how the MC forms and how it is influenced by magnetic fields. In this talk I will present a comparison between the observational characteristics of the MC and a recent study based on a global MHD simulation of solar convection performed with the EUALG-MHD code. This study shows that the amplitude modulation of the MC (at certain depths) can be explained by angular momentum transfers mediated by the magnetic field. This result, when (if) confirmed by other groups, should have a profound impact in the way current mean-field dynamo models are implemented, i.e. the ubiquitous kinematic regime used can no longer be considered the best approximation to run these models since it is skipping important dynamic behaviors.