Active Tectonics and Seismic Hazard in Taiwan – ACTS TAIWAN

Earthquakes are among the most destructive of natural disasters. We have chosen Taiwan as our
study area because of the high intensity of seismic hazard and the possibility of hazard mitigation.
Active faults can be studied on land and on sea, related to the overall subduction and collisional
setting, and also include extensional zones. The deformation and erosion rates are among the
highest on earth in this area which is particularly vulnerable (due to its high population density).
Furthermore, this region has been closely monitored by a network of geophysical instruments during
more than a decade, and the surrounding waters have been object of numerous oceanographic
cruises. An associated international laboratory (LIA) focussed on the subject of this proposal provides
the legal basis for an ongoing scientific cooperation.
During the last century, about a dozen earthquakes of magnitude 7 or greater have struck the island
of Taiwan, of which the earthquake of ChiChi (on land, 1999, M=7.6, 2500 death) and Hualien (at sea,
1920, M>= 8.0) are related to the oblique subduction of the Eurasian plate (continental China plate)
and the oceanic Philippine Sea plate, upon which rests the Luzon volcanic arc. The northern
termination of the Luzon arc acts as a deforming body which indents the accretionary wedge of the
Ryukyu subduction. Thus, Taiwan presents two primary zones of interest, an oceanic subduction fault
plane (Ryukyu arc) which may rupture during a major earthquake, and an orogenic wedge, accessible
to land studies (field geology, GPS, etc.). We have selected three key study areas in order to
characterize the ongoing deformation and estimate the associated seimic risk.
• E-Taiwan : Here the deformation rates are the highest, near the indenter formed by the
northern limit of the Coastal Range. The main objective is to estimate the seismic hazard on
land (posed mainly by the Longitudinal Valley fault) and at sea (fragmentation of the volcanic
arc and the Ryukyu subduction fault plane) on the basis of distribution of active faults and
recent earthquakes. The geophysical database includes recently acquired marine seismic
data sets acquired in collaboration with the Taiwanese and American teams. They
demonstrate the intense nature of deformation along the Ryukyu margin (faults,
microseismicity). An out-of-sequence thrust seems to be the source of a concentration of
seismicity in the fore-arc region. It is important to properly image this structure and to explain
its relation to the interplate seismogenic zone. We propose to characterise the active faults
which cause this anomalous concentration of earthquakes close to Taiwan, in order to develop
a deformation model in order to estimate the risk of a strong earthquake along these faults. In
order to address this problem, propose conducting a combined active and passive seismic
experiment using ocean-bottom seismometers (OBS) in the seismically active region and
using an airgun array towed by an oceanographic vessel (TAIGER project funded by the NSF
and NSC-Taiwan, Spring 2008). This will allow us to precisely locate active faults and to
construct a high resolution tomographic model. Once the rupture zones have been
characterized we will investigate the mechanics of fault segment interaction using numerical
modelling (eg. ADELI finite-element code), in order to determine the likelihood that rupture
along one fore-arc segment could sufficiently load the subduction fault plane and trigger a
major earthquake of magnitude 8.0 or higher.
• N-Taiwan : the deformation is currently extensional in this region with two densely populated
sedimentary basins (Capital City Taipei and Ilan plain). To properly assess and mitigate the
seismic hazard in these basins we will adopt a four part strategy: (1) define the seismotectonic
evolution of the normal faults and constrain their geometry and 3-D kinematics using
seismological data; (2) study the strong ground motions which depend on the seismic source
and on site-effects in the basins bounded by these faults; (3) analyse the vulnerability of
buildings and their resistance to strong ground motion using the method of ambient vibration
analysis (4) the interaction between earthquakes and their triggering is an important
component of the seismic risk which will be characterized by the analysis of Taiwanese
earthquake catalogues using a new numerical model based on a cinematic friction law (nonstatic
coefficient of friction). Finally, a probabilistic study of the seismic catalogues coupled
with the knowledge of site-effects will provide a test of our new methods of the quantification
of the seismic risk, which can then be applied to other less well-studied areas in the world.
This zone of deformation extends offshore into the Okinawa back-arc basin, where we
propose to acquire a wide-angle seismic profile.
4
• W-Taiwan : deformation is concentrated along frontal thrust faults in the orogenic wedge
(ChelungPu Fault, which has been penetrated during an international drilling project and which
will allow us to study the deformation mechanisms within the gouge zone of the fault) and
north-eastwards to the termination of the indenter formed by the Coastal Range. The seismic
potential of the frontal faults (densely populated region) will be determined. The risk of
landslides, which may be induced by earthquakes and the exceptionally rapid uplift rates
associated with the construction of the Central Range will also be estimated.
Our project is focussed on the characterisation, the functioning and the relations between the active
faults on land and offshore in the vicinity of the indenter. One fundamental question is the
accommodation of oblique convergence close to Hualien between the northernmost segment of the
Longitudinal Valley fault and the oppositely vergent thrust faults at sea. The results obtained will be
used to constrain numerical modelling studies of deformation, to characterise the sources and siteeffects
using strong ground motion analysis, to classify the vulnerability of existing buildings and
propose new para-seismic building codes, to determine laws of segment interaction and their impact
on seismic hazard assessment and finally to estimate the seismic hazard using a probabilistic
analysis. The risk of landslides will be studied in relation to the landscape evolution induced by
vertical motions in the Central Range and their triggering by strong earthquakes.