Global correlation of Swarm satellites magnetic field and TEC data with M4+ earthquakes between 2014 and 2024

Between 50 and 80 earthquakes are recorded daily, resulting in over 20,000 yearly seismic events. Currently, no reliable earthquake precursors can provide an early warning. Still, research has inspected anomalies in the Magnetic Field Vector over the Y-axis (MFV-Y) and Total Electron Content (TEC) a...

Descripción completa

Detalles Bibliográficos
Autores: Boudriki Semlali, Badr Eddine|||0000-0003-0671-4808, Molina Ordóñez, Carlos|||0000-0003-0300-4106, Hyuk, Park|||0000-0003-0031-0802, Camps Carmona, Adriano José|||0000-0002-9514-4992
Tipo de recurso: artículo
Fecha de publicación:2025
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/428687
Acceso en línea:https://hdl.handle.net/2117/428687
https://dx.doi.org/10.1016/j.asr.2025.02.065
Access Level:acceso abierto
Palabra clave:Magnetic field vector
Total electron content
IGRF
IRI models
Earthquake precursors
Big data analytics
Descripción
Sumario:Between 50 and 80 earthquakes are recorded daily, resulting in over 20,000 yearly seismic events. Currently, no reliable earthquake precursors can provide an early warning. Still, research has inspected anomalies in the Magnetic Field Vector over the Y-axis (MFV-Y) and Total Electron Content (TEC) as possible precursors of upcoming seismic activity. This study has employed a large global dataset of MFV-Y and TEC data acquired from Swarm satellites between 2014 and 2024 to analyze ionospheric anomalies in earthquake-affected areas. More than 200,000 earthquakes with magnitudes M4+ and within ± 60° latitude have been studied. The Swarm data were compared with physical models, notably, the International Geomagnetic Reference Field (IGRF) and the International Reference Ionosphere (IRI), at the exact locations and times to pinpoint anomalies through the Root Mean Square Error Difference (RMSD) in the time series. This research marks the first use of the Confusion Matrix (CM), Receiver Operating Characteristic Curve (ROC), and various other Figures of Merit (FoM) to assess and try to improve the performance of the methodologies employed and to find the optimal configuration conditions for serving as proxies for earthquake forecasting. As a result, positive anomalies in MFV-Y and TEC were identified 1 to 7 days before the analyzed earthquakes.