Automatic Detection and Tracking of of Far-side Active Regions in Helioseismic Observations for the Advective Flux Transport Model

The Advective Flux Transport (AFT) model, which uses the data assimilation of near-side, has been exceptionally successful in modelling the global solar magnetic field as well as the evolution of Active Regions (ARs) over their lifetime. One of the pitfalls of the AFT model is that it can not predict the emergence of ARs in the farside, some of which may give rise to flares. These far-side activities can have severe space weather impacts and significantly affect our satellites and communication systems. To overcome this limitation, the ARs observed from the Extreme UltraViolet Imager (EUVI) onboard Solar TErrestrial RElations Observatory (STEREO), on the far side of the Sun, have been incorporated in AFT and demonstrated a significant improvement in the model. Unfortunately, STEREO has now returned to the Earth side and it is not available to serve this purpose. Hence, here, we utilise the far-side unsigned magnetic flux maps generated from the helioseismic acoustic maps. These acoustic maps are derived from the Helioseismic and Magnetic Imager (HMI) Doppler measurements and then converted into the magnetic maps using a Machine Learning approach. In contrast with the EUVI observations, these ML acoustic maps have considerable uncertainties and suffer from false AR detections. In this work we have developed an automatic AR detection and tracking algorithm to assess the accuracy of fars-side ARs. We will present some preliminary results based on our tracking algorithm that will give insight on how to effectively use the far-side magnetic field map in the AFT model.