Optimizing Coronal and Solar Wind Model Inputs with Data Assimilation

The Wang-Sheeley-Arge (WSA) model uses line-of-sight observations of the Sun's surface magnetic field as input to a magnetostatic potential field source surface (PFSS) model of coronal expansion, which is in turn used to determine the ambient solar wind speed and IMF polarity. Despite its relative simplicity, the WSA does a very reasonable job of predicting large-scale, persistent structures in the solar wind. In fact, in the absence of transient events like coronal mass ejections (CMEs), the primary error source for WSA predictions often appears to be related to errors in its solar magnetogram inputs. We employ a data assimilation scheme that augments the usual model state vector with variables that modify the WSA's magnetogram inputs in a manner that minimizes WSA prediction errors when compared to in situ measurements.