A New 3D Asymmetric Magnetopause Model with New Consistent Magnetopause Crossing Database and Machine Learning Techniques

As the inner boundary of the magnetosheath, the magnetopause is the location for energy, mass, and momentum transfer between the solar wind and the magnetosphere. The location of the magnetopause as well as its dependence on solar wind and geophysical conditions is critical for solar wind-magnetosphere interaction and global magnetosphere dynamics. Since the first theoretical solution of the shape and size of the magnetopause, many models have been proposed for magnetopause location. However, none of these models make use of new magnetopause crossings with much better coverage of high-latitude magnetopause and the cusps. Previous studies also usually mandated analytical descriptions of magnetopause shape, which were then fit to subsets of crossings. In this study, we constructed a new database of magnetopause crossings using magnetic field and plasma observations from Cluster, Geotail, Polar, Themis, and Wind. The latest observations of these missions have been included and the same criteria are used to achieve a consistent database. Magnetopause crossings inside the cusp are specially taken care of in this new database, which is very important to construct clear funnel-shape cusp structure. Advanced machine learning technique, Support Vector Regression Machine (SVRM), and machine learning tool,MineTool, are used in this study to explore this big database for the control of the magnetopause locations by various solar wind and geophysical factors, including Earth’s dipole tilt, solar wind velocity, IMF, solar wind dynamic pressure, Beta, and Alfven Mach number. Comparison of our magnetopause model with some leading earlier models shows that the new model has small error, very well captures the subsolar magnetopause, high-latitude magnetopause, and cusps. Behavior of the new magnetopause model, including its asymmetry, magnetopause and cusp locations, is further studied under various typical solar wind conditions.