Three-dimensional artificial neural network model of the dayside magnetopause

TL;DR

This paper develops a 3D artificial neural network model to simulate the shape and dynamics of the dayside magnetopause based on solar wind and magnetic field data, revealing three distinct regimes of behavior.

Contribution

It introduces a novel ANN-based model that captures the magnetopause's shape, size, and regime-dependent dynamics under varying solar wind conditions.

Findings

Model accurately describes magnetopause shape and size.

Identifies three regimes of magnetopause dynamics controlled by Bz.

Provides analytical expressions for size variation with By.

Abstract

Artificial Neural Networks (ANN) from package NeuroShell 2 are applied for modeling of dayside magnetopause. The model data set is based on the magnetopause crossings and on the corresponding hour-averaged measurements of solar wind plasma and interplanetary magnetic field. ANN model represents 3D shape and size of the dayside magnetopause in dependence on solar wind dynamic pressure, Bz and By components of interplanetary magnetic field. The model permits firstly to describe dynamics of the magnetic cusp and global asymmetry of the dayside magnetopause under wide range of the external conditions. Slightly change of the magnetopause size with change of By absolute value is presented quantitatively in the form of analytical expression. ANN model shows three regimes of the magnetopause dynamics that controlled by Bz component: pressure balance regime (Bz>0 nT), reconnection regime (0>Bz>-10 nT) and regime of reconnection saturation (Bz<-10 nT). Three different regimes of magnetopause dynamics at the subsolar point are described by modified pressure balance equation obtained from ANN model.