The fourth Workshop


The 2015 NIED-TEM workshop of Seismic Hazard Assessment, August 25-28, 2015, Tsukuba, Japan

The TEM-NIED bilateral workshops have been carried on more than three years. The interests of both sides are far beyond a general information and presentation. With developing of Taiwan PSHA map, different seismogenic source models have appeared such as CGS version, and TEM version. Moreover, various PSHA parameters and models in practice have been argued, and some of them with dilemma are even difficulty in theoretical. These important issues were summarized in five categories: (1) Active fault parameters, (2) Velocity structure, (3) Scenario EQ waveform simulation, (4) GMPE, and (5) Seismic Hazard Assessment. Over 30 questions and arguments dig deeper to geological, seismology, strong motion simulation, surface geological survey, and modelings. Questions also went to organization, mechanism, and authorizations.

Therefore, a PSHA practice-oriented meeting was planed. In response the five categories, professional groups were formed respectively ahead of three months, to deal with the many issues and questions. In the considerations of bilateral discussions have to use the non-native English, and with unfamiliar scientific words, the organizers of the NIED-TEM workshop encouraged preparation in writing and email communication before the bilateral meeting.

In the result, non-stop discussions focus on the given questions during the entire workshop. Many questions and arguments recall the similar discussions in J-SHIS, some even back to 2001 HERP reports. On the other hand, many complexed questions are ongoing studies without consensus achieved yet. The TEM delegation with 17 professional are very satisfactory with discussions, even the special guest Dr. Matt Gerstemberger, from GNS, NZ. In the same time, people from NIED and GSJ feel fruitfully discussions and pleasures. The discussions benefit not only bilateral participants, also usefully for other seismic countries and regions.

Acknowledgement: The NIED-TEM organizers would like to thank all of contributions, discussions, and comments. This is rare for the participants of Taiwanese and Japanese, who the English is the 2nd language. All these are because of the well preparation on the homework before the workshop. Especially thank to supports from GSJ, AIST, for very constructive discussions for active faults, and introduce the works of ERC, HERP.

Questions raised from five categories

1. Active fault parameters

Interested by J. Bruce H. Shyu, Jian-Cheng Lee, Kuo-En Ching, Ray Chuang, Yuan-Hsi Lee, Ling-Ho Chung, and Ruey-Juin Rau, 粟田泰夫 (Yasuo Awata), 吾妻崇 (Takashi Azuma), 武田哲也 (Tetsuya Takeda)

  1. How to deal with fault segmentation and connection (e.g., the 1935 Taichung-Hsinchu event ruptured two adjacent fault segments)? Do we follow UCERF3 and propose ‘Plausible Rupture Criteria’?
  2. How to deal with Mu (the maximum magnitude of the fault) for active fault sources? The magnitude estimated from fault source is smaller than the reported magnitude of the historical events (e.g., 1906 Meishan).
  3. How to determine the fault slip rate without trenching data? How to deal with their uncertainty?
  4. How to determine the slip rate deficit properly on a creeping fault?
  5. How to estimate the earthquake hazard for a creeping fault?

2. Velocity structure

Interested by Kuo-Liang Wen and Chun-Te Chen, 先名重樹 (Shigeki Senna), 木村克己 (Katsumi Kimura), 松山尚典 (Hisanori Matsuyama)

  1. How to obtain the shallow and deep structure? What is the Suggested resolution? What information and data we need so as to construct the 3D velocity model?
  2. Can we combine shallow and deep structure with Tomography model?
  3. We have about 128 microtremor array measurements (interval 10~20km) and 3917 microtremor single station measurements (interval 1~2km) in Taiwan.
  4. Is these data enough for mapping the subsurface structure?
  5. How to plan appropriate fieldwork survey and how about the budget you need for the area like Taiwan? What is the current situation in Japan?
  6. How to do model modification compared to observed waveform data with numerical modeling?

3. Scenario EQ waveform simulation

Interested by Yi-Wen Liao and Yin-Tung Yen, 前田宜浩 (Takahiro Maeda), 青井真 (Shin Aoi), 岩城麻子 (Asako Iwaki)

  1. How to incorporate shallow structure into scenario earthquake simulation?
  2. How to simulate high frequency?
  3. How to create more reliable source models for scenario earthquake simulation?
  4. How to determine suitable match filter in broadband simulation?
  5. How to quantify sensitivity of velocity structure in various frequency band of ground motion simulation?


Interested by Po-Shen Lin, 森川信之 (Nobuyuki Morikawa), 藤原広行 (Hiroyuki Fujiwara)

  1. How to incorporate shallow structure (e.g., for response spectra)?
  2. How to minimize deviation?
  3. GMPEs in intensity?

5. Seismic Hazard Assessment

Interested by Yating Lee, Yu-Ju Wang, Chung-Han Chan, and Kuo-Fong Ma, 井元政二郎 (Masajiro Imoto), 森川信之, 河合伸一 (Shinichi Kawai), 郝憲生 (Ken XS. Hao), 藤原広行

  1. Is fault source independent of background area source? How to deal with Mu (the maximum magnitude of the area seismicity) for a- and b-values of the area sources with and without faults? TEM makes the M>Mu to be fault related.
  2. Use logic tree to implement area source and active fault source independently? Can area source parameters be modeled by earthquake catalog, whereas active fault ones by geological, geomorphological and geodetic evidence? How to evaluate the potential maximum magnitude of background area source? How to deal with fault segmentation for PSHA? How to deal with the slip rate for one segment or two?
  3. Hazard map in intensity? GMPE in intensity? Through a relation between PGA/PGV/SA and intensity?
  4. How to deal with unexpected extreme events (with extra-large magnitudes)?
  5. How to determine the b-values in characteristic earthquake model for faults which are spanned two or more areas?
  6. Should time-dependency be implemented? And how to implement it?
    Not a well-developed approach.
    Long-term (e.g., BPT, Lognormal etc.), how to treat non-characteristic events?
    How to determine the elapse time for those faults without historical/paleo-seismic (trenching data) rupture records?
    Short-term (e.g., ETAS, Coulomb etc.)
    Should time-dependency be magnitude-dependent? And how to quantify magnitude-dependent term?

For the detail, please to see the agenda of the 2015 NIED-TEM workshop of Seismic Hazard Assessment.