Identification of Electromagnetic Pre-Earthquake Perturbations from the DEMETER Data by Machine Learning

The low-altitude satellite DEMETER recorded many cases of ionospheric perturbations observed on occasion of large seismic events. In this paper, we explore 16 spot-checking classification algorithms, among which, the top classifier with low-frequency power spectra of electric and magnetic fields was used for ionospheric perturbation analysis. This study included the analysis of satellite data spanning over six years, during which about 8760 earthquakes with magnitude greater than or equal to 5.0 occurred in the world. We discover that among these methods, a gradient boosting-based method called LightGBM outperforms others and achieves superior performance in a five-fold cross-validation test on the benchmarking datasets, which shows a strong capability in discriminating electromagnetic pre-earthquake perturbations. The results show that the electromagnetic pre-earthquake data within a circular region with its center at the epicenter and its radius given by the Dobrovolsky’s formula and the time window of about a few hours before shocks are much better at discriminating electromagnetic pre-earthquake perturbations. Moreover, by investigating different earthquake databases, we confirm that some low-frequency electric and magnetic fields’ frequency bands are the dominant features for electromagnetic pre-earthquake perturbations identification. We have also found that the choice of the geographical region used to simulate the training set of non-seismic data influences, to a certain extent, the performance of the LightGBM model, by reducing its capability in discriminating electromagnetic pre-earthquake perturbations.

[1]  Michel Parrot,et al.  Special issue of Planetary and Space Science ‘DEMETER’ , 2006 .

[2]  Xuhui Shen,et al.  Progress and development on multi-parameters remote sensing application in earthquake monitoring in China , 2013 .

[3]  A. Gwal,et al.  Satellite monitoring of anomalous effects in the ionosphere related to the great Wenchuan earthquake of May 12, 2008 , 2010 .

[4]  Michel Parrot,et al.  Statistical analysis of automatically detected ion density variations recorded by DEMETER and their relation to seismic activity , 2012 .

[5]  Xuhui Shen,et al.  Analysis of ionospheric plasma perturbations before Wenchuan earthquake , 2009 .

[6]  M. Parrot,et al.  "Real time analysis" of the ion density measured by the satellite DEMETER in relation with the seismic activity , 2012 .

[7]  M. Parrot,et al.  Ionospheric density variations recorded before the 2010 Mw 8.8 earthquake in Chile , 2011 .

[8]  D. Cox.  The Regression Analysis of Binary Sequences , 1958 .

[9]  Paul A. Johnson,et al.  Continuous chatter of the Cascadia subduction zone revealed by machine learning , 2018, Nature Geoscience.

[10]  M. Parrot,et al.  Statistical analysis of an ionospheric parameter as a base for earthquake prediction , 2013 .

[11]  D. Pi,et al.  Recognizing methods for epicenter-neighboring orbits with ionospheric information from DEMETER satellite data , 2017 .

[12]  Michel Parrot,et al.  Statistical Study on Variations of the Ionospheric Ion Density Observed by DEMETER and Related to Seismic Activities , 2017 .

[13]  I. P. Dobrovolsky,et al.  Estimation of the size of earthquake preparation zones , 1979 .

[14]  M. E. Maron,et al.  Automatic Indexing: An Experimental Inquiry , 1961, JACM.

[15]  M. Parrot,et al.  Temporal and spatial analyses on seismo-electric anomalies associated with the 27 February 2010 M = 8.8 Chile earthquake observed by DEMETER satellite , 2013 .

[16]  M. Parrot,et al.  Propagation of whistler mode chorus to low altitudes: Spacecraft observations of structured ELF hiss , 2006 .

[17]  M. Parrot,et al.  Attenuation of electromagnetic waves at the frequency ~1.7 kHz in the upper ionosphere observed by the DEMETER satellite in the vicinity of earthquakes , 2012 .

[18]  R. Fisher.  THE USE OF MULTIPLE MEASUREMENTS IN TAXONOMIC PROBLEMS , 1936 .

[19]  Koby Crammer,et al.  Online Passive-Aggressive Algorithms , 2003, J. Mach. Learn. Res..

[20]  Roman Wronowski,et al.  Studies of the electromagnetic field variations in ELF frequency range registered by DEMETER over the Sichuan region prior to the 12 May 2008 earthquake , 2010 .

[21]  S. Pulinets,et al.  Lithosphere-Atmosphere-Ionosphere Coupling (LAIC) Model - An Unified Concept for Earthquake Precursors Validation , 2011 .

[22]  Sergey Pulinets,et al.  Suspected seismo‐ionospheric coupling observed by satellite measurements and GPS TEC related to the M7.9 Wenchuan earthquake of 12 May 2008 , 2014 .

[23]  J. Friedman.  Greedy function approximation: A gradient boosting machine. , 2001 .

[24]  M. Parrot,et al.  Source of the low‐altitude hiss in the ionosphere , 2017 .

[25]  M. Hayakawa,et al.  Ionospheric perturbations observed by the low altitude satellite DEMETER and possible relation with seismicity , 2013 .

[26]  M. Parrot,et al.  Case studies on the wave propagation and polarization of ELF emissions observed by Freja around the local proton gyrofrequency , 1999 .

[27]  Jean-Pierre Lebreton,et al.  Examples of unusual ionospheric observations made by the DEMETER satellite over seismic regions , 2006 .

[28]  Cheng-Ling Kuo,et al.  An improved coupling model for the lithosphere‐atmosphere‐ionosphere system , 2014 .

[29]  Clara E Yoon,et al.  Earthquake detection through computationally efficient similarity search , 2015, Science Advances.

[30]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[31]  Jing Liu,et al.  Analysis of the enhanced negative correlation between electron density and electron temperature related to earthquakes , 2015 .

[32]  Michael J. Rycroft,et al.  Additional attenuation of natural VLF electromagnetic waves observed by the DEMETER spacecraft resulting from preseismic activity , 2013 .

[33]  Pier Francesco Biagi,et al.  Investigation of TEC and VLF space measurements associated to L'Aquila (Italy) earthquakes , 2010 .

[34]  Sergey Pulinets,et al.  Seismo‐ionospheric coupling appearing as equatorial electron density enhancements observed via DEMETER electron density measurements , 2014 .

[35]  M. Parrot,et al.  Application of wave distribution function methods to an ELF hiss event at high latitudes , 2000 .

[36]  Ondrej Santolik,et al.  Decrease of intensity of ELF/VLF waves observed in the upper ionosphere close to earthquakes: A statistical study , 2009 .

[37]  Jiadong Qian,et al.  Anomaly of the ionospheric electron density close to earthquakes: Case studies of Pu’er and Wenchuan earthquakes , 2011 .

[38]  Ondrej Santolik,et al.  Spacecraft observations of electromagnetic perturbations connected with seismic activity , 2008 .

[39]  José Antonio Lozano,et al.  Sensitivity Analysis of k-Fold Cross Validation in Prediction Error Estimation , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[40]  Paul A. Johnson,et al.  Similarity of fast and slow earthquakes illuminated by machine learning , 2018, Nature Geoscience.

[41]  Sergey Pulinets,et al.  Synchronization of atmospheric indicators at the last stage of earthquake preparation cycle , 2015 .

[42]  F. Leblanc,et al.  ICE, the electric field experiment on DEMETER , 2006 .

[43]  Xuemin Zhang,et al.  Ionospheric electromagnetic perturbations observed on DEMETER satellite before Chile M7.9 earthquake , 2009 .

[44]  M. Parrot,et al.  Statistical Analysis of the Ionospheric Density Recorded by the DEMETER Satellite During Seismic Activity , 2018, Pre-Earthquake Processes.

[45]  S. Pulinets,et al.  A spatial analysis on seismo-ionospheric anomalies observed by DEMETER during the 2008 M8.0 Wenchuan earthquake , 2015 .

[46]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[47]  M. Rycroft,et al.  Atmospheric electricity coupling between earthquake regions and the ionosphere , 2010 .

[48]  Irregularities of the ionospheric plasma and the ULF electric components obtained from DEMETER satellite experiments above Chile earthquake (27 February 2010) , 2015, Arabian Journal of Geosciences.

[49]  Jann‐Yenq Liu,et al.  Seismo-ionospheric anomalies in total electron content of the GIM and electron density of DEMETER before the 27 February 2010 M8.8 Chile earthquake , 2013 .

[50]  J. Lebreton,et al.  New observations of electromagnetic harmonic ELF emissions in the ionosphere by the DEMETER satellite during large magnetic storms , 2006 .

[51]  Michel Parrot,et al.  Electric and magnetic field perturbations recorded by DEMETER satellite before seismic events of the 17th July 2006 M 7.7 earthquake in Indonesia , 2009 .

[52]  Mirko Piersanti,et al.  Electromagnetic field observations by the DEMETER satellite in connection with the 2009 L'Aquila earthquake , 2018, Annales Geophysicae.

[53]  M. Parrot,et al.  Statistical Characteristics of Ionospheric Hiss Waves , 2019, Geophysical Research Letters.

[54]  Enhanced ULF radiation observed by DEMETER two months around the strong 2010 Haiti earthquake , 2010, 1012.1533.

[55]  Feng Zhang,et al.  CloudNet: Ground‐Based Cloud Classification With Deep Convolutional Neural Network , 2018, Geophysical Research Letters.

[56]  Dechang Pi,et al.  Seismo-ionospheric precursory anomalies detection from DEMETER satellite data based on data mining , 2014, Natural Hazards.

[57]  Dechang Pi,et al.  Seismic classification-based method for recognizing epicenter-neighboring orbits , 2017 .

[58]  Pierre Geurts,et al.  Extremely randomized trees , 2006, Machine Learning.

[59]  F. Lefeuvre,et al.  The magnetic field experiment IMSC and its data processing onboard DEMETER: Scientific objectives, description and first results , 2006 .

[60]  V. Kadirkamanathan,et al.  Changes in the ultra-low frequency wave field during the precursor phase to the Sichuan earthquake: DEMETER observations , 2013 .

[61]  Maarten V. de Hoop,et al.  Machine learning for data-driven discovery in solid Earth geoscience , 2019, Science.

[62]  Thomas Lengauer,et al.  ROCR: visualizing classifier performance in R , 2005, Bioinform..

[63]  Sergey Pulinets,et al.  Ionospheric precursors of earthquakes and Global Electric Circuit , 2014 .

[64]  Jiadong Qian,et al.  Response of the ionospheric electron density to different types of seismic events , 2011 .

[65]  M. Parrot,et al.  Variations of electron density and temperature in ionosphere based on the DEMETER ISL data , 2010 .

[66]  M. Parrot.  Statistical analysis of the ion density measured by the satellite DEMETER in relation with the seismic activity , 2011 .

[67]  Michaël Gharbi,et al.  Convolutional neural network for earthquake detection and location , 2017, Science Advances.

[68]  J. Sauvaud,et al.  Unexpected Very Low Frequency (VLF) Radio Events Recorded by the Ionospheric Satellite DEMETER , 2015, Surveys in Geophysics.