Prediction of Energetic Electrons in the Inner Radiation Belt and Slot Region With a Double‐Layer LSTM Neural Network Model

Abstract The prediction of high‐energy radiation belt electrons is vital for preventing their damage to satellites. Previous machine learning models mostly predict the fluxes of high‐energy electrons (hundreds of keV to MeV) in the outer radiation belt and slot region (L > 2.6). Here, we trained...

Full description

Saved in:
Bibliographic Details
Main Authors: Ling Yang, Liuyuan Li, Jinbin Cao
Format: Article
Language:English
Published: Wiley 2025-02-01
Series:Space Weather
Subjects:
Online Access:https://doi.org/10.1029/2024SW004141
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract The prediction of high‐energy radiation belt electrons is vital for preventing their damage to satellites. Previous machine learning models mostly predict the fluxes of high‐energy electrons (hundreds of keV to MeV) in the outer radiation belt and slot region (L > 2.6). Here, we trained a double‐layer long short‐term memory (LSTM) neural network model and successfully predicted the spatial and temporal variations of the 108–749 keV electrons in the inner radiation belt (L ∼ 1.2–2.2) and slot region (L ∼ 2.2–3.2). Under different solar or geomagnetic conditions, the prediction efficiency of the present model maintains 0.6–0.99 in the inner belt and slot region, and its prediction error is less than 0.48. The high‐resolution (∼11 s) LSTM model could predict the rapid injection events of high‐energy electrons within several minutes in the radiation belts.
ISSN:1542-7390