Image-based optimization of coronal magnetic field models for improved space weather forecasting

The existing space weather forecasting frameworks show a significant dependence on the accuracy of the photospheric magnetograms and the extrapolation models used to reconstruct the magnetic filed in the solar corona. Minor uncertainties in the magnetic field magnitude and direction near the Sun, when propagated through the heliosphere, can lead to unacceptible prediction errors at 1 AU. We argue that ground based and satellite coronagraph images can provide valid geometric constraints that could be used for improving coronal magnetic field extrapolation results, enabling more reliable forecasts of extreme space weather events such as major CMEs. In contrast to the previously developed loop segmentation codes designed for detecting compact closed-field structures above solar active regions, we focus on the large-scale geometry of the open-field coronal regions up to 1-2 solar radii above the photosphere. By applying the developed image processing techniques to high-resolution Mauna Loa Solar Observatory images, we perform an optimized 3D B-line tracing for a full Carrington rotation using the magnetic field extrapolation code developed S. Jones at al. (ApJ 2016, 2017). Our tracing results are shown to be in a good qualitative agreement with the large-scale configuration of the optical corona, and lead to a more consistent reconstruction of the large-scale coronal magnetic field geometry, and potentially more accurate global heliospheric simulation results. Several upcoming data products for the space weather forecasting community will be also discussed.