GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process

Abstract Accurately predicting the horizontal component of ground magnetic field perturbation (dBH), a key quantity for calculating the geomagnetically induced currents (GICs), is crucial for assessing the space weather impact of geomagnetic disturbances. The current operational first‐principles Mic...

Full description

Saved in:
Bibliographic Details
Main Authors: Hongfan Chen, Gabor Toth, Yang Chen, Shasha Zou, Zhenguang Huang, Xun Huan
Format: Article
Language:English
Published: Wiley 2025-06-01
Series:Space Weather
Subjects:
Online Access:https://doi.org/10.1029/2024SW004301
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849415812861919232
author Hongfan Chen
Gabor Toth
Yang Chen
Shasha Zou
Zhenguang Huang
Xun Huan
author_facet Hongfan Chen
Gabor Toth
Yang Chen
Shasha Zou
Zhenguang Huang
Xun Huan
author_sort Hongfan Chen
collection DOAJ
description Abstract Accurately predicting the horizontal component of ground magnetic field perturbation (dBH), a key quantity for calculating the geomagnetically induced currents (GICs), is crucial for assessing the space weather impact of geomagnetic disturbances. The current operational first‐principles Michigan Geospace model provides effective forecasts of dBH, but requires significant computational resources to achieve real‐time speeds. Existing data‐driven methods tend to underpredict dBH and lack uncertainty quantification, which is either overlooked or treated as secondary. In this work, we introduce GeoDGP, a novel and efficient data‐driven model based on the deep Gaussian process. GeoDGP provides global probabilistic forecasts of dBH with a lead time of at least 1 hr, at 1‐min time cadence, and at arbitrary spatial locations. The model takes solar wind measurements, the Dst index, and the prediction location in solar magnetic coordinate system as inputs, and is trained on 28 years of data from SuperMAG global magnetometer stations. Additionally, GeoDGP is also trained to predict the north (dBN) and east (dBE) components of perturbations. We evaluate GeoDGP's performance at over 200 stations worldwide during 24 geomagnetic storms, including the Gannon extreme storm of May 2024. Comparisons with the first‐principles Michigan Geospace model and the data‐driven DAGGER model revealed that GeoDGP significantly outperforms both across multiple performance metrics.
format Article
id doaj-art-8c6f69a3417545fd97a8b328b9b436cb
institution Kabale University
issn 1542-7390
language English
publishDate 2025-06-01
publisher Wiley
record_format Article
series Space Weather
spelling doaj-art-8c6f69a3417545fd97a8b328b9b436cb2025-08-20T03:33:25ZengWileySpace Weather1542-73902025-06-01236n/an/a10.1029/2024SW004301GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian ProcessHongfan Chen0Gabor Toth1Yang Chen2Shasha Zou3Zhenguang Huang4Xun Huan5Department of Mechanical Engineering University of Michigan Ann Arbor MI USADepartment of Climate and Space Sciences and Engineering University of Michigan Ann Arbor MI USADepartment of Statistics University of Michigan Ann Arbor MI USADepartment of Climate and Space Sciences and Engineering University of Michigan Ann Arbor MI USADepartment of Climate and Space Sciences and Engineering University of Michigan Ann Arbor MI USADepartment of Mechanical Engineering University of Michigan Ann Arbor MI USAAbstract Accurately predicting the horizontal component of ground magnetic field perturbation (dBH), a key quantity for calculating the geomagnetically induced currents (GICs), is crucial for assessing the space weather impact of geomagnetic disturbances. The current operational first‐principles Michigan Geospace model provides effective forecasts of dBH, but requires significant computational resources to achieve real‐time speeds. Existing data‐driven methods tend to underpredict dBH and lack uncertainty quantification, which is either overlooked or treated as secondary. In this work, we introduce GeoDGP, a novel and efficient data‐driven model based on the deep Gaussian process. GeoDGP provides global probabilistic forecasts of dBH with a lead time of at least 1 hr, at 1‐min time cadence, and at arbitrary spatial locations. The model takes solar wind measurements, the Dst index, and the prediction location in solar magnetic coordinate system as inputs, and is trained on 28 years of data from SuperMAG global magnetometer stations. Additionally, GeoDGP is also trained to predict the north (dBN) and east (dBE) components of perturbations. We evaluate GeoDGP's performance at over 200 stations worldwide during 24 geomagnetic storms, including the Gannon extreme storm of May 2024. Comparisons with the first‐principles Michigan Geospace model and the data‐driven DAGGER model revealed that GeoDGP significantly outperforms both across multiple performance metrics.https://doi.org/10.1029/2024SW004301space weatheruncertainty quantificationmachine learning
spellingShingle Hongfan Chen
Gabor Toth
Yang Chen
Shasha Zou
Zhenguang Huang
Xun Huan
GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
Space Weather
space weather
uncertainty quantification
machine learning
title GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
title_full GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
title_fullStr GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
title_full_unstemmed GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
title_short GeoDGP: One‐Hour Ahead Global Probabilistic Geomagnetic Perturbation Forecasting Using Deep Gaussian Process
title_sort geodgp one hour ahead global probabilistic geomagnetic perturbation forecasting using deep gaussian process
topic space weather
uncertainty quantification
machine learning
url https://doi.org/10.1029/2024SW004301
work_keys_str_mv AT hongfanchen geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess
AT gabortoth geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess
AT yangchen geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess
AT shashazou geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess
AT zhenguanghuang geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess
AT xunhuan geodgponehouraheadglobalprobabilisticgeomagneticperturbationforecastingusingdeepgaussianprocess