Taiwan is an exceptional area to investigate orogenic processes because of very rapid rates of deformation and erosion, as well as because of the wealth of data already available (geodynamic setting, structure of the crust, recent deformation documented by GPS and tectonic geomorphology…). The occurrence of the Mw=7.6 Chichi earthquake in 1999 allows for investigating how deformation resulting from earthquakes or from slower aseismic processes during postseismic relaxation, or during the interseismic period, contribute jointly to geological deformation over the long term. 

Geodynamic setting

Overview of Taiwan geology

Taiwan and the "critical-wedge” model.

How is deformation partitioned across Taiwan?

The 1999 Chi-Chi earthquake: insights into the seismic cycle of West Central Taiwan.

References

Geodynamic setting

Taiwan is located at the boundary between the Philippine Sea Plate to the East and the Eurasian Plate to the West, with a convergence rate of ~ 80 mm/yr in a ~N118E direction [Seno, 1977; Yu et al., 1997] (figure1). This plate boundary is rather complex since it comprises two subduction zones of reverse polarities (figure 1). To the southwest, the ongoing consumption of the oceanic crust of the South China Sea led to the collision between the Chinese Continental Margin and the Luzon Volcanic Arc ~ 6.5 Ma ago [e.g Lin et al., 2003], that resulted in the Taiwanese mountain range.

Figure 1: Geodynamic setting of the island of Taiwan..

Because of the obliquity of the convergence, the collision propagated to the south: the collision is mature and fading to the north of the island while to the south, the South China Sea is still subducting along the Manila Trench (figures 1 and 2). Consequently, Taiwan is the perfect location to better understand how the transition from subduction to collision occurs (figure 2).

Figure 2: Transition from subduction of the oceanic crust of the South China Sea beneath the Manila Trench to the south (C), to mature collision between the Luzon volcanic arc and the Chinese passive margin in Central Taiwan (A). Incipient collision (B) occurs in the area of the Hengchun Peninsula in southernmost Taiwan. Sketches are from [Malavieille et al., 2002].

Overall geology of Taiwan

The geology of Taiwan may be described in terms of different separated units (figure 3). From East to West:

• The Coastal Range which corresponds to the accreted Luzon volcanic arc.
• The Longitudinal Valley which is considered as the suture zone between the Luzon arc and the Chinese continental margin.
• The Tananao Schists are comprised of the metamorphic pre-Tertiary basement of the Eurasian passive margin. In addition to the most recent metamorphic event, this unit has also recorded past orogenic events.
• The Slate belt, composed of metamorphosed and deformed sediments of the Chinese passive margin, with from East to West:
  • The Backbone Range, with mostly Miocene to Eocene slates, corresponds to the area of highest altitudes in the island.
  • The Hsueshan Range, composed of mostly Eocene and Oligocene sediments.
• The Western Foothills, at lower altitudes, where syn-orogenic sediments of the foreland basin have been accreted and deformed.
• The Coastal Plain which is part of the present foreland basin of Taiwan.

Figure 3: Major geological units of the Island of Taiwan. CR: Coastal Range; LV: Longitudinal Valley; ECR: East Central Range or Tananao Schist complex; WCR: West Central Range or Backbone Range; HSR: Hsueshan Range; WF: Western Foothills; CP: Coastal Plain. From Central Geological Survey of Taiwan-MOEA.

Although, the degree of metamorphism in the Slate belt cannot be easily documented due to the poor mineralogy of these rocks, it is thought that metamorphism increases from west to east across the Taiwanese range. Within the TO, this overall idea should be tested and refined by applying new methods that have been recently developed to document maximum temperatures encountered by slates during their metamorphic cycle.

Taiwan and the “critical-wedge” model.

Since the early 80’s and 90’s [e.g. Barr and Dahlen, 1989; Barr et al., 1991; Dahlen and Barr, 1989; Dahlen et al., 1984; Davis et al., 1983; Suppe, 1981],  Taiwan has been referred to as the best example of the “critical wedge” type mountain belt (figure 4).

Figure 4: Critical wedge model of mountain belts. A: Basics of such model as proposed by [Davis et al., 1983]; it was refined later by [Barr and Dahlen, 1989; Barr et al., 1991; Dahlen and Barr, 1989; Suppe, 1981] for Taiwan. Rocks get to be deformed as sand would be in front of a rigid buttress overriding a rigid footwall. B: Sand-box experiment image (courtesy of J. Malavieille) illustrating this model.

In this model, a mountain belt is viewed as a deforming wedge whose geometry reflects the balance between frictional stresses at the bottom, and gravitational loads induced by the topography. The mountain belt mostly grows by frontal accretion and internal thickening. Since the wedge has to maintain its equilibrium geometry, horizontal shortening is expected to be distributed, such as the active faults that accommodate this deformation. The model successfully accounts for some aspects of the Taiwan orogen, such as its gross topography and metamorphic structure; it had been tested, given the data that were available by then. However, new constraints documenting mountain-building processes in Taiwan had been published since, and even more should result from the several TO investigations, so that this model needs be re-evaluated and tested in light of these new data.

How is deformation partitioned across Taiwan?

Interseismic  deformation (prior to the 1999 ChiChi earthquake) recorded by GPS [Yu et al., 1997] allows for gaining some insights into how the ~ 80mm/yr of total convergence are partitioned across the island (figure 5).

Figure 5: Interseismic deformation recorded by GPS relatively to the Penghu Islands [Yu et al., 1997]. The Chichi earthquake surface rupture has also been reported. Figure from [Dominguez et al., 2003].

Some portion of the shortening is consumed along the Longitudinal Valley (figure 5); in addition to that, elastic dislocation modeling of these GPS data [Dominguez et al., 2003; Hsu et al., 2003; Loevenbruck et al., 2001] suggests that ~30 to 45 mm/yr would be transferred to the Western Foothills through a sub-horizontal decollement beneath the Central Range (figure 6).However, how deformation in the interseismic period relates to long term slip rates on faults is poorly understood because these slip rates are only crudely estimated; it is unclear whether the shortening rate is absorbed by a great number of small faults, resulting in distributed deformation alike that predicted from critical wedge models, or whether it is localized on a few large faults. Within the TO, several projects aim at solving these important issues.

The 1999 Chi-Chi earthquake:
insights into the seismic cycle of West Central Taiwan.

The Mw=7.6 ChiChi earthquake in central Taiwan that ruptured the Chelungpu fault over ~ 80 km in September 1999 [Ma et al., 1999] (figure 5) is to date the most extensively documented earthquake. This allows for investigating the different stages of the seismic cycle, and to define their relative contribution to the long term deformation and mountain building process.

The detail of the seismic slip history and the co-seismic static slip has been constrained from the inversion of seismic records [Ji et al., 2001], field [Lee et al., 2001] and GPS measurements  [Yu et al., 2001] as well as SPOT images [Dominguez et al., 2003] (figure 6). These studies show that coseismic slip was mainly confined to a ramp dipping about 30°E at depths less than 5-7km, and reached a maximum value over 12m at a very shallow depth in the northern portion of the fault.

Figure 6: The ChiChi earthquake and the seismic cycle in Central Taiwan. Top panel: Horizontal coseismic displacements from SPOT images; interseismic and coseismic displacements have been added together so as to calculate a recurrence time for such earthquakes. Middle panel: GPS displacements prior to the ChiChi earthquake and modeled interseismic deformation assuming that 40 mm/yr are transferred to the foothills through a sub-horizontal decollement beneath the Central Range. Bottom panel: Geometry of the system and how it compares to the thermal structure proposed by [Lin, 2000]. Figures from [Dominguez et al., 2003].

Modeling of interseismic GPS data show that the Chelungpu fault was previously locked to a depth where it would root to some sub-horizontal decollement [Dominguez et al., 2003; Hsu et al., 2003; Loevenbruck et al., 2001 ] (figure 6). If coseismic displacements are compared to interseismic displacements, the recurrence interval can be estimated to 150 to 250 years. This estimate assumes that all the shortening stored during the interseismic period would be taken up by coseismic deformation; also this assumes that the shortening rate measured over the few years before the ChiChi earthquake by GPS would be representative of the long term rate. These assumptions might both be questioned and need further investigation.

• home
• regional tectonics
• research
• collaborators
• publications
• data