Opening the Black Box of the Radiation Belt Machine Learning Model - Research Article

Aug 5, 2026

Upload by Jiven

A high-accuracy machine learning model for Earth's radiation belt electron flux is examined through interpretability methods. The ORIENT neural network uses only solar wind conditions and geomagnetic indices as inputs, and DeepSHAP is applied to explain the “black box” behavior. Two electron flux enhancement events from Van Allen Probes—one storm interval (17–18 March 2013) and one non-storm interval (19–20 September 2013)—are analyzed. Feature importance results highlight physically meaningful effects aligned with current radiation belt understanding and support general application to similar models.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Radiation belt dynamics and modeling challenges

Neural-network alternative using ORIENT

Model structure and energy coverage

Interpretability approach

RESEARCH ARTICLE 10.1029/2022SW003339 Special Section: Machine Learning in Heliophysics Key Points: • We demonstrate the feature attribution method for a machine learning model of electron flux • We quantify the effects of geomagnetic indices and solar wind parameters on electron flux during a storm time event and a non-storm event • Our feature importance results identify physical effects that are consistent with our current understanding Correspondence to: D. Ma, [dma96@atmos.ucla.edu](dma96@atmos.ucla.edu) [Citation:](Citation:) Ma, D., Bortnik, J., Chu, X., Claudepierre, S. G., Ma, Q., & Kellerman, A. (2023) . Opening the black box of the radiation belt machine learning model. Space Weather, 21, e2022SW003339. [https://doi](https://doi). org/10.1029/2022SW003339 Received 6 NOV 2022 Accepted 3 FEB 2023 © 2023. The Authors. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Opening the Black Box of the Radiation Belt Machine Learning Model Donglai Ma1 , Jacob Bortnik1 , Xiangning Chu2 , Seth G. Claudepierre1 , Qianli Ma1,3 , and Adam Kellerman1 1Department of Atmospheric and Oceanic Sciences, University of California, Los Angeles, Los Angeles, CA, USA, 2Laboratory for Atmospheric and Space Physics, University of Colorado Boulder, Boulder, CO, USA, 3Center for Space Physics, Boston University, Boston, MA, USA Abstract Many Machine Learning (ML) systems, especially deep neural networks, are fundamentally regarded as black boxes since it is difficult to fully grasp how they function once they have been trained. Here, we tackle the issue of the interpretability of a high-accuracy ML model created to model the flux of Earth's radiation belt electrons. The Outer RadIation belt Electron Neural net (ORIENT) model uses only solar wind conditions and geomagnetic indices as input features. Using the Deep SHAPley additive explanations (DeepSHAP) method, for the first time, we show that the “black box” ORIENT model can be successfully explained. Two significant electron flux enhancement events observed by Van Allen Probes during the storm interval of 17–18 March 2013 and non-storm interval of 19–20 September 2013 are investigated using the DeepSHAP method. The results show that the feature importance calculated from the purely data-driven ORIENT model identifies physically meaningful behavior consistent with current physical understanding. This work not only demonstrates that the physics of the radiation belt was captured in the training of our previous model, but that this method can also be applied generally to other similar models to better explain the results and to potentially discover new physical mechanisms. Plain Language Summary A neural network is regarded as a black box model since it can approximate any function but its structure won't give any insights on the nature of the function being approximated. A set of neural network models named Outer RadIation belt Electron Neural net have been developed previously to model the electron flux of the outer radiation belt. In this work, we demonstrate the general flow of explaining the machine learning (ML) model of radiation belts and investigate two typical events during the storm and non-storm times. The results identify physically meaningful behavior and are consistent with current physical understanding, additionally providing new insight into radiation belt dynamics. Furthermore, the proposed framework can be generalized for a variety of other ML models, including various plasma parameters in the Earth's magnetosphere. 1. Introduction The Earth's radiation belts consist of energetic charged particles trapped by the geomagnetic field into two regions, a relatively stable inner zone, and a more dynamic outer zone. These particles range in energy from tens ofkeV to multiple MeV (e.g.,

FAQs

Why is the radiation belt machine learning model considered a “black box”?

Deep neural networks are hard to interpret after training because their internal structure does not directly reveal how inputs drive outputs. In this work, the goal is to make that behavior understandable for a high-accuracy electron-flux model.

What model inputs does ORIENT use?

ORIENT uses solar wind conditions and geomagnetic indices as the input features, without directly requiring other physical parameters.

How are storm and non-storm events investigated and what is the key outcome?

Two Van Allen Probes electron-flux enhancement events are analyzed with the DeepSHAP method, covering a storm interval (17–18 March 2013) and a non-storm interval (19–20 September 2013). The resulting feature importance indicates physically meaningful effects consistent with current radiation belt physics.

Aug 5, 2026

Upload by Jiven