Specifying Properties of Dayside Magnetopause Reconnection from a Machine-Learning Model for the Earths Cusps

ROSES ID: NNH22ZDA001N-LWS      Selection Year: 2022      

Program Element: Focused Science Topic

Principal Investigator: Gonzalo A Cucho Padin

Affiliation(s): Catholic University Of America

Project Member(s):
Wang, Xueyi Co-I/Institutional PI AUBURN UNIVERSITY
Sibeck, David G Co-I/Institutional PI NASA/GODDARD SPACE FLIGHT CENTER
da Silva, Daniel E Co-I/Institutional PI UNIVERSITY OF MARYLAND BALTIMORE COUNTY

Summary:

This proposal quantifies the properties of dayside magnetopause reconnection via a novel machine-learning-based analysis of in-situ measurements of ion fluxes within the terrestrial cusps. It addresses the following science questions:_x000D_
_x000D_
How do solar wind plasma and magnetic field conditions control the spatiotemporal variability of the high/mid-altitude cusps? _x000D_
How do solar wind plasma and magnetic field conditions control the spatiotemporal variability of the low-altitude cusps? _x000D_
What can we infer about the location(s) of reconnection on the magnetopause from cusp properties?_x000D_
What can we infer about the time dependence of reconnection on the magnetopause from cusp properties?_x000D_
_x000D_
Magnetic reconnection on the dayside magnetopause represents the primary mechanism for transporting mass, momentum, and energy from the solar wind into the terrestrial magnetosphere. Despite its crucial role, accurate information regarding the location, extent, and time dependence of reconnection remains difficult to obtain. Several studies have demonstrated that the spatiotemporal dynamics of the dayside magnetic reconnection can be inferred remotely from the time-energy dispersion of ions in the cusps; however, it is still difficult to infer the overall cusp behavior from intermittent snapshots of isolated in-situ measurements._x000D_
_x000D_
This project aims to use novel machine learning techniques to leverage the immense amount of ion flux observations available from the cusps, provide reliable models for cusp dynamics, and, through these, improve our understanding of magnetic reconnection on the dayside magnetopause. First, we will implement regression models for the high/mid-altitude Northern cusp (beyond <2 Earth Radii) using artificial neural networks (ANNs) as well as solar wind/IMF data and ion flux spectra from the Cluster/CODIF instruments. This data-based model will yield a 3-D distribution of ion flux within the cusp region for each energetic channel in the CODIF instrument. Uncertainties of the model will be estimated through a robust statistical analysis of modeling results and observations. We will assess the modeled cusp features by comparing them with cusp simulations from the physics-based model Auburn global hybrid simulation code in 3D (ANGIE3D) and in-situ observations of the Polar/TIMAS instrument. Then, we will identify the relative role of each input solar wind parameter on the cusp structure using the DeepSHAP algorithm. Second, following a similar procedure, we will generate regression models for the low-altitude cusp using data from the SSJ5 ion spectrometers on board the DMSP-F17/F18 missions. Then, we will estimate their corresponding uncertainties and will assess the modeling results by comparing them with ANGIE3D model simulations and observations of the SSJ5 instrument onboard the DSMP-F16/F19 spacecrafts. Also, we will use DeepSHAP to evaluate the relationship between the low-altitude cusp ion structures and each input solar wind parameter. Third, we will conduct a correlation study between the spatial distribution of dayside magnetopause reconnection sites and the 3-D ion flux distributions in the cusp under steady solar wind conditions. For this, we will run ANGIE3D simulations of the dayside solar wind-magnetospheric interaction region and support the results with in-situ plasma observations of the MMS mission. Then, for a number of cases, we will analyze the relationship between the location and extent of reconnection entry points and the 3-D structure of the cusp. Finally, we will leverage the time-varying property of the data-based cusp models to investigate the relationship between the spatiotemporal properties of reconnection entry points estimated by ANGIE3D and the spatiotemporal ion flux distributions in the cusp during transient events such as solar wind tangential discontinuities (TD) and rotational discontinuities (RD).

---

Export to PDF