Predicting the data structure prior to extreme events from passive observables using echo state network
Extreme events are defined as events that largely deviate from the nominal state of the system as observed in a time series. Due to the rarity and uncertainty of their occurrence, predicting extreme events has been challenging. In real life, some variables (passive variables) often encode significan...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-10-01
|
| Series: | Frontiers in Applied Mathematics and Statistics |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fams.2022.955044/full |
| _version_ | 1811340656679845888 |
|---|---|
| author | Abhirup Banerjee Abhirup Banerjee Arindam Mishra Arindam Mishra Syamal K. Dana Syamal K. Dana Chittaranjan Hens Chittaranjan Hens Tomasz Kapitaniak Jürgen Kurths Jürgen Kurths Jürgen Kurths Norbert Marwan Norbert Marwan |
| author_facet | Abhirup Banerjee Abhirup Banerjee Arindam Mishra Arindam Mishra Syamal K. Dana Syamal K. Dana Chittaranjan Hens Chittaranjan Hens Tomasz Kapitaniak Jürgen Kurths Jürgen Kurths Jürgen Kurths Norbert Marwan Norbert Marwan |
| author_sort | Abhirup Banerjee |
| collection | DOAJ |
| description | Extreme events are defined as events that largely deviate from the nominal state of the system as observed in a time series. Due to the rarity and uncertainty of their occurrence, predicting extreme events has been challenging. In real life, some variables (passive variables) often encode significant information about the occurrence of extreme events manifested in another variable (active variable). For example, observables such as temperature, pressure, etc., act as passive variables in case of extreme precipitation events. These passive variables do not show any large excursion from the nominal condition yet carry the fingerprint of the extreme events. In this study, we propose a reservoir computation-based framework that can predict the preceding structure or pattern in the time evolution of the active variable that leads to an extreme event using information from the passive variable. An appropriate threshold height of events is a prerequisite for detecting extreme events and improving the skill of their prediction. We demonstrate that the magnitude of extreme events and the appearance of a coherent pattern before the arrival of the extreme event in a time series affect the prediction skill. Quantitatively, we confirm this using a metric describing the mean phase difference between the input time signals, which decreases when the magnitude of the extreme event is relatively higher, thereby increasing the predictability skill. |
| first_indexed | 2024-04-13T18:45:13Z |
| format | Article |
| id | doaj.art-582b3e54fc1a43469cb292344169f3a7 |
| institution | Directory Open Access Journal |
| issn | 2297-4687 |
| language | English |
| last_indexed | 2024-04-13T18:45:13Z |
| publishDate | 2022-10-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Applied Mathematics and Statistics |
| spelling | doaj.art-582b3e54fc1a43469cb292344169f3a72022-12-22T02:34:36ZengFrontiers Media S.A.Frontiers in Applied Mathematics and Statistics2297-46872022-10-01810.3389/fams.2022.955044955044Predicting the data structure prior to extreme events from passive observables using echo state networkAbhirup Banerjee0Abhirup Banerjee1Arindam Mishra2Arindam Mishra3Syamal K. Dana4Syamal K. Dana5Chittaranjan Hens6Chittaranjan Hens7Tomasz Kapitaniak8Jürgen Kurths9Jürgen Kurths10Jürgen Kurths11Norbert Marwan12Norbert Marwan13Complexity Science, Potsdam Institute for Climate Impact Research, Potsdam, GermanyInstitute for Physics and Astronomy, University of Potsdam, Potsdam, GermanyDivision of Dynamics, Lodz University of Technology, Łódź, PolandDepartment of Physics, National University of Singapore, Singapore, SingaporeDivision of Dynamics, Lodz University of Technology, Łódź, PolandDepartment of Mathematics, National Institute of Technology, Durgapur, IndiaPhysics and Applied Mathematics Unit, Indian Statistical Institute, Kolkata, IndiaCenter for Computational Natural Sciences and Bioinformatics, International Institute of Information Technology Gachibowli, Hyderabad, IndiaDivision of Dynamics, Lodz University of Technology, Łódź, PolandComplexity Science, Potsdam Institute for Climate Impact Research, Potsdam, GermanyDivision of Dynamics, Lodz University of Technology, Łódź, PolandInstitute of Physics, Humboldt Universität zu Berlin, Berlin, GermanyComplexity Science, Potsdam Institute for Climate Impact Research, Potsdam, GermanyInstitute of Geoscience, University of Potsdam, Potsdam, GermanyExtreme events are defined as events that largely deviate from the nominal state of the system as observed in a time series. Due to the rarity and uncertainty of their occurrence, predicting extreme events has been challenging. In real life, some variables (passive variables) often encode significant information about the occurrence of extreme events manifested in another variable (active variable). For example, observables such as temperature, pressure, etc., act as passive variables in case of extreme precipitation events. These passive variables do not show any large excursion from the nominal condition yet carry the fingerprint of the extreme events. In this study, we propose a reservoir computation-based framework that can predict the preceding structure or pattern in the time evolution of the active variable that leads to an extreme event using information from the passive variable. An appropriate threshold height of events is a prerequisite for detecting extreme events and improving the skill of their prediction. We demonstrate that the magnitude of extreme events and the appearance of a coherent pattern before the arrival of the extreme event in a time series affect the prediction skill. Quantitatively, we confirm this using a metric describing the mean phase difference between the input time signals, which decreases when the magnitude of the extreme event is relatively higher, thereby increasing the predictability skill.https://www.frontiersin.org/articles/10.3389/fams.2022.955044/fullextreme eventscoupled neuron modelactive and passive variableprecursory structureecho state networkphase coherence |
| spellingShingle | Abhirup Banerjee Abhirup Banerjee Arindam Mishra Arindam Mishra Syamal K. Dana Syamal K. Dana Chittaranjan Hens Chittaranjan Hens Tomasz Kapitaniak Jürgen Kurths Jürgen Kurths Jürgen Kurths Norbert Marwan Norbert Marwan Predicting the data structure prior to extreme events from passive observables using echo state network Frontiers in Applied Mathematics and Statistics extreme events coupled neuron model active and passive variable precursory structure echo state network phase coherence |
| title | Predicting the data structure prior to extreme events from passive observables using echo state network |
| title_full | Predicting the data structure prior to extreme events from passive observables using echo state network |
| title_fullStr | Predicting the data structure prior to extreme events from passive observables using echo state network |
| title_full_unstemmed | Predicting the data structure prior to extreme events from passive observables using echo state network |
| title_short | Predicting the data structure prior to extreme events from passive observables using echo state network |
| title_sort | predicting the data structure prior to extreme events from passive observables using echo state network |
| topic | extreme events coupled neuron model active and passive variable precursory structure echo state network phase coherence |
| url | https://www.frontiersin.org/articles/10.3389/fams.2022.955044/full |
| work_keys_str_mv | AT abhirupbanerjee predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT abhirupbanerjee predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT arindammishra predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT arindammishra predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT syamalkdana predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT syamalkdana predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT chittaranjanhens predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT chittaranjanhens predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT tomaszkapitaniak predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT jurgenkurths predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT jurgenkurths predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT jurgenkurths predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT norbertmarwan predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork AT norbertmarwan predictingthedatastructurepriortoextremeeventsfrompassiveobservablesusingechostatenetwork |