Dynamics of the Martian bow shock location and machine learning detection

Abstract:

The Martian interaction with the solar wind is unique due to the influence of remanent crustal magnetic fields. The recent studies by the Mars Express (MEX) and Mars Atm-­osphere and Volatile Evolution (MAVEN) missions underline the strong and complex influence of the crustal magnetic fields on the Martian environment and its interaction with the solar wind. Among them is the influence on the dynamic plasma boundaries that shape this interaction and on the bow shock in particular. Here we analyze in detail the influence of the crustal fields on the Martian shock location by combining datasets from multiple spacecraft (MAVEN/MEX). It is confirmed that the crustal fields have an important impact on the Martian plasma boundaries including the shock, although the solar extreme ultraviolet fluxes and the upstream solar wind magnetosonic Mach number are the main drivers of the average shock location. We use for this study a detailed partial correlation analysis to remove statistical biases and identify fine relationships between parameters. A machine learning algorithm (specifically, Multilayer Perceptron) is developed to automatically detect boundary locations and categorize plasma regions, which has been trained successfully using MEX and MAVEN observations. The newly-developed machine learning algorithm will greatly enhance our ability of fast detection of bow shock locations, facilitating the study of the interaction between Mars and the solar wind.