Timeliness of earthquake magnitude estimation from the prompt elasto-gravity signal  using Deep Learning

2021 ◽  
Author(s):  
Andrea Licciardi ◽  
Quentin Bletery ◽  
Bertrand Rouet-Leduc ◽  
Jean-Paul Ampuero ◽  
Kévin Juhel
Keyword(s):  
Deep Learning ◽  
Magnitude Estimation ◽  
Early Warning ◽  
Early Warning Systems ◽  
Earthquake Magnitude ◽  
Reliable Estimate ◽  
Origin Time ◽  
Tsunami Early Warning ◽  
Gravity Signal ◽  
Large Earthquakes

<p>Mass redistribution during large earthquakes produces a prompt elasto-gravity signal (PEGS) that travels at the speed of light and can be observed on seismograms before the arrival of P-waves. PEGS carries information about earthquake magnitude and the temporal evolution of seismic moment, therefore it could be used to both improve the accuracy of current early source estimation systems and speed-up early warning. However, PEGS has been detected for only a handful of very large earthquakes so far, and its potential use for operational early warning remains to be established. In this work, we study the timeliness of magnitude estimation for subduction earthquakes in Japan using PEGS waveforms by means of Deep Learning and Bayesian uncertainty analysis. Given the paucity of PEGS observations, we train the model on a database of synthetic seismograms augmented with empirical noise in order to simulate more realistic waveforms. We use about 80 stations from the Japanese F-Net network and from networks with data available through IRIS.</p><p>Under this experimental setup, we find that our model is able to track the moment release for earthquakes with a final Mw above 8.0, with a system latency that depends on the signal-to-noise ratio of PEGS. The application of our model to the Mw=9.1 Tohoku-Oki earthquake shows a latency of about 50 s after which the model is able to track well the evolving Mw of the earthquake. After about 2 minutes from the earthquake origin time, a reliable estimate of its final Mw is obtained. Similar performances in terms of timeliness of final Mw estimation are observed for the relatively smaller Hokkaido earthquake (Mw=8.1) although with higher uncertainty.</p><p>Our results highlight the potential of PEGS to enhance the performance of existing tsunami early warning systems where estimating the magnitude of very large earthquakes within few minutes is vital.</p><p> </p>

Performance of a Hybrid Demonstration Earthquake Early Warning System in the Sichuan–Yunnan Border Region

2020 ◽  
Vol 91 (2A) ◽  
pp. 835-846
Author(s):  
Chaoyong Peng ◽  
Qiang Ma ◽  
Peng Jiang ◽  
Wenhui Huang ◽  
Dake Yang ◽  
...  
Keyword(s):  
Magnitude Estimation ◽  
Early Warning ◽  
Risk Mitigation ◽  
Effective Means ◽  
Early Warning Systems ◽  
Test Period ◽  
Earthquake Early Warning ◽  
Border Region ◽  
Hybrid Network ◽  
Origin Time

Abstract Earthquake early warning systems (EEWSs) are considered to be one of the most effective means for seismic risk mitigation, in terms of both losses and societal resilience, by releasing an alarm immediately after an earthquake occurs and before strong ground shaking arrives the target sites to be protected. To gain experience for the National System for Fast Seismic Intensity Report and Earthquake Early Warning project, we deployed a hybrid demonstration EEWS in the Sichuan–Yunnan border region with micro-electro-mechanical system-based sensors and broadband seismographs and low-latency data transmission. In this study, we described the structure of this EEWS and analyzed its performance in the first 2 yr from January 2017 to December 2018. During this test period, the EEWS detected and processed a total of 126 ML 3.0+ earthquakes, with excellent epicentral location and magnitude estimation. The average location and magnitude estimation errors for the first alert were 4.2±7.1  km and 0.2±0.31, respectively. For the earthquakes that occurred inside and outside the hybrid network, the first alert was generated 13.4±5.1  s and 26.3±13.5  s after the origin time (OT), respectively. We analyzed the performance of the EEWS for the 31 October 2018 M 5.1 earthquake, because it was the largest event that occurred inside the hybrid network during the test period. The first alert was obtained at 7.5 s after the OT, with a magnitude error of 0.1 magnitude unit, a location error of about 1 km, and a depth error of 8 km. Finally, we discussed the main differences between the EEWS’s estimates and the catalogs obtained by the China Earthquake Network Center, and proposed improvements to reduce the reporting time. This study demonstrated that we constructed a reliable, effective hybrid EEWS for the test region, which can provide sufficient support for the design of the National EEWS project.

scholarly journals Rapid Estimation of Earthquake Magnitude using GNSS Data

2021 ◽  
Vol 873 (1) ◽  
pp. 012063
Author(s):  
Susilo ◽  
Irwan Meilano ◽  
Thomas Hardy ◽  
Muhammad Al Kautsar ◽  
Dina A. Sarsito ◽  
...  
Keyword(s):  
Early Warning ◽  
Warning System ◽  
Earthquake Magnitude ◽  
High Rate ◽  
Origin Time ◽  
Earthquake Parameters ◽  
Tsunami Early Warning ◽  
Scaling Properties ◽  
Earthquake Magnitude Determination ◽  
Gnss Data

Abstract Earthquake parameters such as the hypocenter location and the magnitude size are important in the development of reliable Earthquake Early Warning (EEW) and Tsunami Early Warning System (TEWS). The Global Navigation Satellitte System (GNSS) data have been used to estimate the earthquake parameters rapidly over the last 15 years. In this study, we present the result and analysis of the scaling properties of Peak Ground Displacement (PGD) as measured by high-rate (sampled at 1 Hz or higher) GPS recordings from Lombok earthquake on August 05 th , 2018. The earthquake magnitude from the kinematic solution of CMAT GNSS station is equal to Mw 6.8. This value is achieved between 15 - 20 seconds after the origin time of the earthquake. Our result shows that the displacements from kinematic GNSS data can be used to rapidly determine the earthquake magnitude, typically within the first minute of rupture initiation. Rapid earthquake magnitude determination will be very useful to support EEW and TEWS.

Rapid earthquake source characterization in the context of tsunami early warning

2021 ◽  
Author(s):  
Alberto Armigliato ◽  
Martina Zanetti ◽  
Stefano Tinti ◽  
Filippo Zaniboni ◽  
Glauco Gallotti ◽  
...  
Keyword(s):  
Focal Mechanism ◽  
Early Warning ◽  
Fault Plane ◽  
Early Warning Systems ◽  
Slip Distribution ◽  
Generation Process ◽  
Tsunami Early Warning ◽  
Earthquake Focal Mechanism ◽  
Run Up ◽  
Seismic Networks

<p>It is well known that for earthquake-generated tsunamis impacting near-field coastlines the focal mechanism, the position of the fault with respect to the coastline and the on fault slip distribution are key factors in determining the efficiency of the generation process and the distribution of the maximum run-up and inundation along the nearby coasts. The time needed to obtain the aforementioned information from the analysis of seismic records is usually too long compared to the time required to issue a timely tsunami warning/alert to the nearest coastlines. In the context of tsunami early warning systems, a big challenge is hence to be able to define 1) the relative position of the hypocenter and of the fault and 2) the earthquake focal mechanism, based only on the preliminary earthquake localization and magnitude estimation, which are made available by seismic networks soon after the earthquake occurs.</p><p>In this study, the intrinsic unpredictability of the position of the hypocenter on the fault plane is studied through a probabilistic approach based on the analysis of two finite fault model datasets (SRCMOD and USGS) and by limiting the analysis to moderate-to-large shallow earthquakes (Mw  6 and depth  50 km). After a proper homogenization procedure needed to define a common geometry for all samples in the two datasets, the hypocentral positions are fitted with different probability density functions (PDFs) separately in the along-dip and along-strike directions.</p><p>Regarding the focal mechanism determination, different approaches have been tested: the most successful is restricted to subduction-type earthquakes. It defines average values and uncertainties for strike, dip and rake angles based on a combination of a proper zonation of the main tsunamigenic subduction areas worldwide and of subduction zone geometries available from publicdatabases.</p><p>The general workflow that we propose can be schematically outlined as follows. Once an earthquake occurs and the magnitude and hypocentral solutions are made available by seismic networks, it is possible to assign the focal mechanism by selecting the characteristic values for strike, dip and rake of the zone where the hypocenter falls into. Fault length and width, as well as the slip distribution on the fault plane, are computed through regression laws against magnitude proposed by previous studies. The resulting rectangular fault plane can be discretized into a matrix of subfaults: the position of the center of each subfault can be considered as a “realization” of the hypocenter position, which can then be assigned a probability. In this way, we can define a number of earthquake fault scenarios, each of which is assigned a probability, and we can run tsunami numerical simulations for each scenario to quantify the classical observables, such as water elevation time series in selected offshore/coastal tide-gauges, flow depth, run-up, inundation distance. The final results can be provided as probabilistic distributions of the different observables.</p><p>The general approach, which is still in a proof-of-concept stage, is applied to the 16 September 2015 Illapel (Chile) tsunamigenic earthquake (Mw = 8.2). The comparison with the available tsunami observations is discussed with special attention devoted to the early-warning perspective.</p>

scholarly journals Near real-time GPS applications for tsunami early warning systems

2010 ◽  
Vol 10 (2) ◽  
pp. 181-189 ◽  
Author(s):  
C. Falck ◽  
M. Ramatschi ◽  
C. Subarya ◽  
M. Bartsch ◽  
A. Merx ◽  
...  
Keyword(s):  
Real Time ◽  
Early Warning ◽  
Early Warning System ◽  
Tide Gauge ◽  
Warning System ◽  
Early Warning Systems ◽  
Time Data ◽  
Essential Information ◽  
Tsunami Early Warning ◽  
Land Mass

Abstract. GPS (Global Positioning System) technology is widely used for positioning applications. Many of them have high requirements with respect to precision, reliability or fast product delivery, but usually not all at the same time as it is the case for early warning applications. The tasks for the GPS-based components within the GITEWS project (German Indonesian Tsunami Early Warning System, Rudloff et al., 2009) are to support the determination of sea levels (measured onshore and offshore) and to detect co-seismic land mass displacements with the lowest possible latency (design goal: first reliable results after 5 min). The completed system was designed to fulfil these tasks in near real-time, rather than for scientific research requirements. The obtained data products (movements of GPS antennas) are supporting the warning process in different ways. The measurements from GPS instruments on buoys allow the earliest possible detection or confirmation of tsunami waves on the ocean. Onshore GPS measurements are made collocated with tide gauges or seismological stations and give information about co-seismic land mass movements as recorded, e.g., during the great Sumatra-Andaman earthquake of 2004 (Subarya et al., 2006). This information is important to separate tsunami-caused sea height movements from apparent sea height changes at tide gauge locations (sensor station movement) and also as additional information about earthquakes' mechanisms, as this is an essential information to predict a tsunami (Sobolev et al., 2007). This article gives an end-to-end overview of the GITEWS GPS-component system, from the GPS sensors (GPS receiver with GPS antenna and auxiliary systems, either onshore or offshore) to the early warning centre displays. We describe how the GPS sensors have been installed, how they are operated and the methods used to collect, transfer and process the GPS data in near real-time. This includes the sensor system design, the communication system layout with real-time data streaming, the data processing strategy and the final products of the GPS-based early warning system components.

scholarly journals Tsunami Generation Potential of a Strike-slip Fault Tip in the Westernmost Mediterranean

2021 ◽  
Author(s):  
F. Estrada ◽  
J. M. González-Vida ◽  
J. A. Peláez ◽  
J. Galindo-Zaldívar ◽  
S. Ortega ◽  
...  
Keyword(s):  
Early Warning ◽  
Seismic Activity ◽  
Early Warning Systems ◽  
Strike Slip ◽  
Strike Slip Fault ◽  
Tsunami Generation ◽  
Warning Systems ◽  
Tsunami Early Warning ◽  
Seafloor Deformation ◽  
Tsunami Early Warning Systems

Abstract Tsunamis are triggered by sudden seafloor displacements, and usually originate from seismic activity at faults. Nevertheless, strike-slip faults are usually disregarded as major triggers, as they are thought to be capable of generating only moderate seafloor deformation; accordingly, the tsunamigenic potential of the vertical throw at the tips of strike-slip faults is not thought to be significant. We found the active dextral NW-SE Averroes Fault in the central Alboran Sea (westernmost Mediterranean) has a historical vertical throw of up to 5.4 m at its northwestern tip corresponding to an earthquake of Mw 7.0. We modelled the tsunamigenic potential of this seafloor deformation by Tsunami-HySEA software using the Coulomb 3.3 code. Waves propagating on two main branches reach highly populated sectors of the Iberian coast with maximum arrival heights of 6 m within 21 and 35 min, which is too quick for current early-warning systems to operate successfully. These findings suggest that the tsunamigenic potential of strike-slip faults is more important than previously thought, and justify the re-evaluation of tsunami early-warning systems worldwide.

Sign in / Sign up

Export Citation Format

Share Document