Earthquake Early Warning
Development of Methods and Algorithms for Automatic Real Time Identification of Waveforms Introduction from Local Earthquakes in Increased Level of Man-induced Noises for the Purposes of Ultra-short-term Warning about an Occured Earthquake.
Tech Area / Field
- ENV-SEM/Seismic Monitoring/Environment
8 Project completed
Senior Project Manager
Russo P A
International Institute of Earthquake Prediction Theory and Mathematical Geophysics, Russia, Moscow
- Research Institute of Aviation Systems, Russia, Moscow
- Royal Observatory of Belgium, Belgium, Brussels\nNational Institute for Earth Science and Disaster Prevention, Japan, Ibaraki
Project summaryThe vulnerability of world civilization in the face of inevitable catastrophic earthquakes is rapidly increasing owing to urbanization, the development of big cities and the propagation of high risk facilities like nuclear power stations, major dams, strategic pipelines, oil refineries and so on. The problem of earthquake prediction and warning about large earthquakes that have occurred is thus comparable with disarmament, nuclear terrorism, and environmental contamination.
The chief danger lies in the fact that earthquakes occur suddenly. This can be illustrated by several recent major earthquakes that occurred in the Asian region. It should be noted that the significant loss of human lives is usually due to the suddenness of the earthquake, considering that no warning systems are available. However, there are certain ways and means to provide ultra-immediate warning about a local earthquake that has just occurred. Any earthquake that occurs in the Earth excites several types of waves that propagate in the crust at different group velocities. This difference in seismic wave velocities allows the development of a technique for recognition of a damaging earthquake in order to be able to develop a cheap seismic instrument that can identify the arrivals of fast, "low energy" P and S waves. A real time analysis of first arrivals allows calculation of the time that remains before the arrival of the main damaging surface wave. Rough calculations show that the time difference between P and surface waves must be about 30 seconds, when the distance between the earthquake epicenter and the monitoring site is of order 200 km. This time is enough for people to leave potentially hazardous buildings, to insert moderators of nuclear reactions at nuclear power stations, to close the pipelines, to stop traffic etc. One specially important point about the use of such algorithms is when they are to be used in areas of increased seismic alarm to warn the population about shocks about to occur in seismic disaster zones. Under these conditions such an instrument may happen to be virtually the only tool of information delivery, when the communication lines have been destroyed by the earthquake in question.
Taken by itself, the idea of ultra-immediate warning about an earthquake that has occurred (i.e., predicting the arrival at a given geographic point of damaging waves from an earthquake that has occurred) is not a novelty. In areas with a dense coverage of seismic stations (certain areas in Japan, California), earthquake waves are traced as they travel from station to station and their parameters are computed, including the time arrival at any point and expected intensity of shaking there. However, these systems require expensive installation of special seismic equipment (drilling of wells etc.), but the chief difficulty in these projects consists in transmitting information (alarm signals) to the final user (industry, transportation, population). It is this gap that our method is to fill, whose leading feature will be self-contained operation without expensive connection to existing complex earthquake prediction systems. The essential difference from the existing systems will be the possibility of installing the devices developed inside buildings and structures on carrying walls and structural members, not only in the basement, since our algorithms for seismic signal processing will incorporate the presence of strong residential and industrial noise using its spectrum.
The implementation of this project envisages both scientific and commercial results. In the scientific aspect, the important thing is the development of techniques and algorithms for automatic computation of earthquake parameters in real time based on observations of a few sensors under high seismic noise. The commercial effect will result from the manufacture and sale of the instrument giving ultra-immediate warning about local earthquakes.
The project envisages the development of technologic, technical and software tools, the creation of several models of the different specialized seismic sensors, as well as of an experimental mechanical installation to simulate ground motion (based on digital signals of near earthquakes previously recorded at seismic stations) for experimental testing an debugging of the algorithms for identification of first earthquake wave arrivals.
The following features assure scientific and technical advantages of the project:
- The technical and software tools to be developed for this project envisage automatic detection and estimation of parameters for large regional seismic signals in real time, i.e., before the arrival of damaging waves, based on data from a single seismic sensor or a local array of seismic sensors.
- The technique to be developed will allow the use of modernized seismometers with a large dynamic range designed for operation with near (local) earthquakes. These seismometers will provide for recording of low amplitude signals (P waves) as precursors of strong ground motion. This will enable the instruments to be used for warning about fore- and aftershock activity around the epicenter of a large earthquake. Also, the seismic signal detection algorithm will identify S waves in addition to P waves, thus substantially reducing the probability of false alarms and yielding a more accurate forecast of the time the main damaging earthquake waves are to arrive.
- It will be possible to use these techniques and algorithms to develop sufficiently cheap (for the end user) seismic instruments of small dimensions and capable of self-contained operation.
- The software support for processing of seismic data will be based on the optimal statistical time series processing algorithms. In addition, the implementation of P and S wave identification problem will use algorithms of frequency signature on base of widely used methods of allocation of a useful signal on a hum noise. In particular, these include:
- An adaptive algorithm that allows suppression of coherent noise and identification of low amplitude seismic phases in noise;
- A statistical algorithm based on the maximum likelihood method for accurate determination of arrival times for low amplitude seismic phases masked in the coda of earlier phases.
5. Taken on the whole, the project envisages the development of a technology for real time, self-contained, local seismic monitoring wholly based on Russian designs in the field of seismic instrumentation and software for seismic data processing. The technology will be a logical complement to global monitoring systems for seismic alarms under development abroad. It seems to provide a cheap alternative to complex monitoring systems for saving the lives of end users.
The investigations proposed for this project can be provisionally subpided into several main lines of research, each having certain methodological approaches of its own:
- Creation of a computer system for the acquisition and analysis of seismic data;
- Development of algorithms and software for fast automatic processing of seismic observations in real time;
- Development (modernization) of a test unit testing and adjustment of algorithms for identification of first earthquake wave phases.
- Problems arising in connection with modernization of strong motion seismometers designed to deal with a large input dynamic range of seismic motion.
The above problems will be dealt with, in addition to conventional techniques employed in the processing of seismic data and in the design of geophysical instruments by essentially similar techniques and technologies in use for developing aviation navigation systems, automatic decision-making systems, and techniques for carrying out telemetric tests.
The present project will be executed during three years, its implementation is guaranteed by already available developments concerning components of the seismicity monitoring system carried out at IIEPT RAS and FRIAS for previous ISTC projects #415 and #1539 (superbroadband seismometers, components of telemetric and satellite communication systems, algorithms and software for seismic data analysis).
The scientific results derived for this project are to be presented at international symposia and to be published in leading geophysical journals.
IIEPT RAS and FRIAS maintain close scientific cooperation with many research organizations abroad (in Japan, USA, Belgium, Norway, Germany, France, Italy and other countries). Our successful experience in carrying out joint international projects will be used in the implementation of the present one.