RT Journal Article
T1 Deep Learning Event Detector from Long-term Signal Variation for Seismic Activity Warning out of Schumann Resonance
A1 Cano-Domingo, Carlos
A1 Stoean, Ruxandra
A1 Soler-Ortiz, Manuel
A1 Novas-Castellano, Nuria
A1 Fernández-Ros, Manuel
A1 Joya-Caparrós, Gonzalo
A1 Gázquez-Parra, José A.
K1 Sismología - Innovaciones tecnológicas
AB Deep Learning (DL) has shown capability in many areas of impact on everydaylife. The paper proposes a DL architecture tailored for event detectionfrom examining the time evolution of a signal. With temporal characteristicsextracted by a Convolutional Neural Network (CNN) encoder and fedas input to a recurrent neural network, the model targets the detection of apossibly occurring investigated event in the given time interval. The utilityof DL methodologies to solve physical problems is demonstrated for an applicationof the complex experimentally-studied existing interaction betweenSchumann Resonance (SR) and seismic activity. SR signals are electromagneticwaves propagating along the Earth-ionosphere cavity. Intense lightning activity is continuously present at the same locations around the world, beingsensitive to physical perturbation. Seismic activity modifies this steadylightning pattern. The new DL model is applied to answer the researchquestion of whether the variation of the SR signal is truly a verifiable forerunnerof seismic activity. Several parameter configurations are explored,either model-related or linked to criteria for selecting seismic events. Resultsshow preliminary evidence about the relation between distance-intensityspace and SR perturbation, and provide valuable corroboration about thesensitivity of the sensor to a specific azimuth between the observatory andthe Earthquake (EQ) epicenter, hence argumentatively supporting the SRtemporal characteristics as an early seismic warning. This is the first generalizationof seismic disturbance as a derivative of the SR, based only on itssignal time series variation, as a hypothesized precursor of the EQ event.
PB Elsevier
YR 2025
FD 2025-10
LK https://hdl.handle.net/10630/41392
UL https://hdl.handle.net/10630/41392
LA eng
NO Carlos Cano-Domingo, Ruxandra Stoean, Manuel Soler-Ortiz, Nuria Novas, Manuel Fernández-Ros, Gonzalo Joya, Jose A. Gázquez Parra, Deep learning event detector from long-term signal variation for seismic activity warning out of Schumann resonance, Knowledge-Based Systems, Volume 328, 2025, 114166, ISSN 0950-7051, https://doi.org/10.1016/j.knosys.2025.114166. (https://www.sciencedirect.com/science/article/pii/S0950705125012079)
NO CETPartnership
NO Innovation, Science and Enterprise, Andalusian Regional Government
NO Romanian Ministry of Research and Innovation
NO Electronics Technology Department of the University of Malaga
NO Ministry of Economics and Competitiveness of Spain
NO UAL18-TIC-A025-A
NO TEC2014-60132-P
NO COFUND-CETP 40/2024, UE-FISCDI PNCDI IV
NO 178PCE/2021, PN-III-P4-ID-PCE-2020-0788
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 20 ene 2026