Back to search

Boston University

Modeling whistler-mode waves and their effects on radiation belt electron dynamics: integrating physical and deep learning approaches

Abstract

dc:description.abstract

The Earth’s radiation belts, populated by high-energy particles, play a crucial role in space weather dynamics. Despite extensive satellite observations over decades, accurately simulating high-energy particle dynamics remains challenging due to the complex physical processes occurring across different scales. This dissertation advances our understanding of radiation belt dynamics by integrating physical simulations with deep learning modeling techniques, enabling the identification of patterns and processes not readily apparent from observations alone. The core of this research lies in the innovative application of deep learning to predict key parameters influencing radiation belt electron dynamics, and combining it with physical simulations. We developed a recurrent neural network-based model to construct the global distribution of electron density, which successfully reproduces observed electron density profiles. This model disentangles the time evolution of geomagnetic indices from satellite location, allowing for a novel representation of theinner magnetospheric state. Furthermore, we employed a probabilistic deep learning model to simulate the spatiotemporal evolution of hiss waves and their impact on particle dynamics. Our results reveal a close spatial correlation between plume and plume hiss, as well as rapid variations in hiss wave intensity during geomagnetic storms, leading to significant reductions in energetic electron fluxes over timescales of a few days. We quantified the effects of both plasmaspheric and plume hiss, demonstrating that plume hisscan account for up to an 80% decrease in electron fluxes at hundreds of keV at L greater than 4 over 3 days under relatively quiet geomagnetic conditions. This finding highlights the importance of including dynamic hiss distribution in future simulations of radiation belt electron dynamics. Additionally, we extended the framework to model trapped electron flux at different energies. By combining this enhanced framework with simulations of wave-particle interaction, we constructed global maps of energetic electron precipitation into Earth’s atmosphere driven by whistler-mode waves. Our two-stage method, integrating neural networks and quasi-linear theory, successfully reproduced rapidly varying precipitation patterns both spatially and temporally over various storm phases, consistent with satellite observations. This dissertation provides a novel approach to quantifying and predicting energetic electron dynamics influenced by whistler-mode waves in Earth’s magnetosphere. By leveraging deep learning with physical modeling techniques, it lays the groundwork for improved operational models capable of predicting space weather globally with high spatial and temporal resolution. Our research offers valuable insights into largescale variations of plasma waves and particles in space, often impossible to obtain from in-situ observations alone, contributing to a more comprehensive understanding of radiation belt, ring current, and their effect on atmospheric dynamics.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Astronomy
Grantor
Boston University
Year dc:date.issued
2025

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Huang, Sheng
Advisor dc:contributor.advisor
  • Li, Wen
ORCID dc:identifier.orcid

Subjects

dc:subject × 1

Rights

dc:rights
Statement dc:rights
  • Attribution 4.0 International
Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/2144/53252

Chain of custody

source
Harvested from
Boston University
Base URL
open.bu.edu/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
related terms
citation

Huang, Sheng. Modeling whistler-mode waves and their effects on radiation belt electron dynamics: integrating physical and deep learning approaches. doctoral thesis, Boston University, 2025. https://hdl.handle.net/2144/53252