head img  
NTU Taiwan University img


Effective in Reducing Life and Property Loss - Department of Geosciences Professor Yih-min Wu's Research on Earthquake Early Warning System Received International Attention

If scientists could issue earthquake warnings 10 seconds to 30 seconds before the shockwaves of the earthquake spread, the human life and property losses in the disaster area should be effectively reduced. NTU Department of Geosciences Professor Dr. Yih-min Wu utilized the primary waves of an earthquake to conduct earthquake early warning systems research. His research findings have received international attention. The Pacific Tsunami Warning Center in Hawaii adopted Dr. Wu's programs last year to conduct early warning and monitoring of earthquakes.

Professor Wu's research findings have been published in succession in internationally renowned journals such as "Geophysical Research Letters." In addition to attracting attention from the seismological society of the United States, the Korea Meteorological Administration invited him to give a lecture at the end of last year (see Figure 1); whereas the Kyoto University of Japan decided to use Dr. Wu's concept to develop an earthquake early warning system.

Dr. Wu is currently developing a portable hand held earthquake warning device which takes advantage of modern day MEMS technology. When a disastrous earthquake takes place, people who live in the disaster area are able to receive warnings from the Central Weather Bureau, and take contingency measures on a real time basis right on the spot.

Professor Wu points out that, Taiwan is situated at the Circum-Pacific seismic belt wherein seismic activities abound, and disastrous earthquakes occur quite often. Instances such as the 1906 earthquake in Meishan, Chiayi, the 1935 strong earthquake in Hsinchu and Taichung, and the 1999 giant earthquake in Chichi all left indelible memories in people's minds. Therefore, early detection of earthquake happenings is an area of research that must be continually strengthened.

Earthquake prediction has always been a hot topic in disaster prevention. Although many precursory phenomena have been confirmed (1), so far they have not been proven to be practical. As the level of applicability of earthquake prediction remains low, many countries in the world divert their resources to the development of earthquake early warning systems (2). Rapid earthquake information not only serves as important indicators for disaster relief measures, but also meets the expectations of the general public and the media. Information provided by earthquake early warning systems can be the key to the emergency contingency measures taken on the part of major construction projects and civilian facilities.

Earthquake early warning refers to warnings issued about several seconds or tens of seconds before the disastrous shock waves strike. This short period of time is critical in directly reducing the harms inflicted by the earthquake. The scope of applications can be as follows:

  1. Schoolchildren can hide under their desks to seek protection and psychological adjustment. Research results of Mexico City's early warning system have shown that schoolchildren who received warning messages in advance greatly reduced their fears of earthquakes.
  2. Workers are given enough time to leave their dangerous work sites.
  3. The ongoing surgeries in a hospital can be stopped temporarily or make minute and sophisticated adjustments in their operation. For instance, eye surgeries.
  4. Transportation systems can be stopped automatically or reduce their running speed. For instance: high speed railways can reduce their running speed to lower the danger of derailment.
  5. Life-supporting pipelines and communication networks can be automatically adjusted, reorganized or shutdown. For instance: gas pipes and water hoses can be turned off to reduce the chances of fire caused by earthquake or other disasters.
  6. Factories can activate emergency response measures to protect vibration sensitive equipments. For instance: wafer manufacturing plants.

After careful reevaluation, earthquake early warning system is the most effective disaster reduction method to date. The United States, Japan, Mexico and Taiwan have all invested efforts in the development of early warning systems. The incentive for Taiwan's involvement in the development of earthquake early warning system stemmed from the lessons learned from the Hualien earthquake which took place on November 15th, 1986 and registered 7.8 on the moment magnitude scale. Although the epicenter of this earthquake was in Hualien area, Taipei City, which is about 120 kilometers from Hua-lien, was the area that was struck the most. According to the data of the travel time of seismic waves, it took at least 30 seconds for the shear waves to travel from Hualien to Taipei. Therefore, if an earthquake monitoring system could detect the geographical location of the earthquake's epicenter and the scale of the earthquake within 30 seconds, it could gain about several seconds to more than ten seconds warning time, which could be used in the employment of contingency measures to mitigate the disaster of the earthquake. Having learned this hard lesson, the Central Weather Bureau began to get involved in the development of earthquake warning systems in 1994.

Starting from the year 1995, the Central Weather Bureau installed real time strong earthquake monitoring systems to report on incoming earthquakes. In order to make use of the real time strong earthquake signals, early warning systems were also actively developed. By adopting the MLIO earthquake scale measuring method and the design of regional earthquake subnet and virtual subnets, the Central Weather Bureau was able to shorten the time needed to figure out the parameters of an earthquake to 20 seconds. Consequently, for metropolitan areas more than 70 kilometers away from the epicenter, the Central Weather Bureau was able to provide different amounts of early warning time.

However, the current method cannot provide early warnings to areas which are within 70 kilometers from the epicenter, mainly because the MLIO estimation method is a traditional one which relies on the earth movement signals which are discovered 10 seconds after these signals are detected. As a result, the MLIO method cannot provide early warning within 15 seconds after the earthquake takes place. So in recent years the earthquake early warning research utilizes the analysis of primary waves of an earthquake to issue early warnings in the hope to shorten the warning time to within 10 seconds after the earthquake takes place.

Estimation of the primary waves: In seismic early warning research, utilizing the primary waves to determine the scale of an earthquake is the most important and the most difficult technology. The traditional ML estimation is arrived at by measuring the maximum amplitude of a particular band modified by distance. However, when we observe the maximum amplitude, we are left with no time to issue early warnings. Therefore, the traditional ML estimation is not suitable for early warning systems. Fortunately, past researches conduced by many people(6,7,8) indicate that the bigger an earthquake, the longer the vibration cycle of its signals. We are thus able to project the scale of an earthquake by measuring the vibration cycle of its primary waves.

Utilizing the primary waves to predict the intensity of an earthquake: In general the primary waves of an earthquake carry information about the earthquake, whereas the shear waves carry the bulk of the earthquake's energy which could cause great damage. If we could analyze the primary waves to determine the strength of the shear waves, we can then apply our analysis to the early warning systems directly. So we analyzed the seismic records of Taiwan (12) and Southern California (13) and found that, 3 seconds after the primary waves arrived, the peak amplitudes of displacement(Pd) were in a linear logarithmic relationship with the peak ground motion velocity (PGV). Therefore, we can estimate the intensity of an earthquake by the primary waves(13), and the margin of error can be less than 1 Richter scale. This capability works well for targets which are less sensitive to inaccurate information such as elevators. When an earthquake is being told to take place, the elevator needs only to stop at the adjacent floor and keep its door open. Even if the intensity of the earthquake is overestimated, the elevator will not cause any harm.

Rapid identification of catastrophic earthquakes: The identification of catastrophic earthquakes is an important key to efficient disaster relief. Therefore, we analyzed the records of shallow crustal earthquakes which took place in Taiwan. Our research results indicate that by combing tc and Pd we are able to identify catastrophic earthquakes (9). By the same principle we are able to shorten the time needed for judging whether a earthquake will cause tremendous damage to less than 10 seconds, thus buying us more time to take contingency measures for disaster relief.

Insofar as disaster reduction is concerned, our current technology in earthquake prediction leaves a lot to be desired. So earthquake early warning system is the most practical method at present. In order to gain more warning time before strong shock waves attack us, we use the signals obtained 3 seconds after the arrival of primary waves to determine the scale of the earthquake, to predict its intensity and to ascertain whether the earthquake is a catastrophic one . By adopting the tc method we may be able to shorten the scale estimation time to less than 10 seconds, and provide warnings to metropolitan areas 30 kilometers away from the epicenter. By measuring the peak amplitude of displacement we are able to estimate the degree of the impact, ]which can be directly applied to current early warning system. When the value of tc times Pd is larger than 1, we may conclude that a catastrophic earthquake is taking place.


  1. Y. M. Wu and L. Y. Chiao, Bull. Seism. Soc. Am., 96, in press (2006).
  2. H. Kanamori, E. Hauksson, and T. Heaton, Nature, 390, 461-464 (1997).
  3. Y. M. Wu, T. C. Shin, and Y. B. Tsai, Bull. Seism. Soc. Am., 88, 1254-1259 (1998).
  4. Y. M. Wu, J. K. Chung, T. C. Shin, N. C. Hsiao, Y. B. Tsai, W. H. K. Lee, and T. L. Teng, Terrestrial, Atmospheric and Oceanic Sciences, 10, 719-736 (1999).
  5. Y. M. Wu and T. L. Teng, Bull. Seism. Soc. Am., 92, 2008-2018 (2002).
  6. Y. Nakamura, Proceeding of 9th world conference on earthquake engineering, Tokyo-Kyoto, Japan, (1988).
  7. R. M. Allen and H. Kanamori, Science 300, 685-848 (2003).
  8. H. Kanamori, Annual Review of Earth and Planetary Sciences, 33, 5.1-5.20 (2005).
  9. Y. M. Wu and H. Kanamori, Bull. Seism. Soc. Am., 95, 347-353 (2005).
  10. Y. M. Wu , H. Kanamori, R. M. Allen, and E. G. Hauksson, Bull. Seism. Soc. Am., submitted (2006).
  11. Y. M. Wu, H. Y. Yen, L. Zhao, B. S. Huang, and W. T. Liang, Geophysical Research Letters, submitted (2006).
  12. Y. M. Wu and H. Kanamori, Bull. Seism. Soc. Am., 95, 1181-1185 (2005).
  13. Y. M. Wu, T. L. Teng, T. C. Shin, and N. C. Hsiao, Bull. Seism. Soc. Am., 93, 386-396 (2003).

Chinese version