Vertical crustal motion of active plate convergence in Taiwan derived from tide gauge, altimetry, and GPS data

(1)

Vertical crustal motion of active plate convergence in Taiwan derived from tide

gauge, altimetry, and GPS data

Emmy T.Y. Chang

a,

⁎

, Benjamin F. Chao

b,c

, Chieh-Chung Chiang

a

, Cheinway Hwang

d

a

Institute of Oceanography, National Taiwan University, Taipei, Taiwan

b_{Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan} c

College of Earth Sciences, National Central University, Chungli, Taiwan

d

Dept. of Civil Engineering, National Chiao Tung University, Hsinchu, Taiwan

a b s t r a c t

a r t i c l e i n f o

Article history: Received 9 October 2010

Received in revised form 16 September 2011 Accepted 3 October 2011

Available online 12 October 2011 Keywords: Tide gauge Altimeter GPS Uplift Subsidence Convergence

Located at the converging junction between the Eurasian and Philippine Sea plates, the island of Taiwan is subject to an active lithospheric deformation as well as seismicity. Taking the difference between the satellite altimetry data (ALT) that give the absolute sea level variation and the tide gauge data (TG) that record the relative sea level variation, we obtain the absolute vertical crustal motion of the tide gauge sites. We use 20 TG stations along the west and east coasts of Taiwan along with the ALT measurements from the TOPEX/Poseidon–Jason satellites in the nearby waters. The ALT–TG results are compared with vertical GPS measurements in discussing vertical motion. Weﬁnd a general subsidence of the entire Taiwan coast during the past two decades. The west coast sees no prominent vertical motion but with a severe local subsidence due to the over-withdrawal of groundwater. On the east coast, the ALT–TG results in the northern section demonstrate a northward dipping motion. The elastic thickness of the neighboring oceanic lithosphere mod-eled as an elastic plate with theﬂexure of the subducting plate shows that the adjacent Philippine Sea plate should be an old, thick oceanic plate, which could drag the slab into the mantle as manifested in a gentle northward subsidence in the northeast Taiwan. In the southern section of the east coast, the ALT–TG results reveal a segmented or undulating pattern in the vertical-motion rates. Judging from the different behaviors between the co-seismic and interseismic vertical motions marked by the major earthquakes during the stud-ied period, we postulate a temporal saw-tooth scenario for the deformation in phases. It demonstrates the opposite motions under different mechanisms in the frontal sections of the subduction zone, which can be understood with lateral collision and slab dragging subject to varied temporal and spatial dependences.

1. Introduction

The crustal deformation has been measured by, among other geo-detic means, the Global Positioning System (GPS). Since the 1990s GPS measurements have demonstrated that the present-day horizon-tal crushorizon-tal motions conform largely to that predicted by plate tecton-ics based on geological evidences on million-year timescales. The discrepancies between the modern geodetic measurements and geo-logical evidences are then source observation for reﬁnement of the regional plate-motion models. Many such studies have been con-ducted to elucidate the Taiwan tectonics (e.g.Angelier et al., 2009; Chang et al., 2003; Rau et al., 2008; Wu et al., 2009; Yu and Kuo, 2001; Yu et al., 1997). On the other hand, for the vertical component of the crustal deformation, while gravimeters and other traditional geodetic sensors are routinely used to record the non-tectonic

vertical motion due to tides, atmospheric and oceanic mass loadings, and regional post-glacial rebound, the tectonics-related or orogenic vertical motions have received comparatively less attention. One rea-son is the generally small magnitude of such vertical deformations during relatively short timespans of observation, compounded by the well-known fact that GPS data have much larger errors in the ver-tical than the horizontal components (e.g.Dixon, 1991; Hager et al., 1991).

Spanning 400 km N–S and lying on the western boundary of the Paciﬁc–Eurasian plate convergence zone, the island of Taiwan is char-acterized by a double subduction system along two offshore trenches (Fig. 1), whose corresponding Wadati–Benioff zones are clearly delin-eated by the great number of subduction earthquakes recorded over the past several decades (Tsai, 1986). On the east, the Philippine Sea plate subducts under the Eurasian plate extending under northern Taiwan, which simultaneously upducts upon the Eurasian plate along the N_{–S trending Longitudinal Valley (LV) in eastern Taiwan,} forming the E–W trending Ryukyu Trench along the latitude around 23.5°N. The LV is the suture zone between the Eurasian continent on the west and the Philippine Sea plate on the east. The Coastal

⁎ Corresponding author at: Institute of Oceanography, National Taiwan University, No.1, Sec. 4, Roosevelt Rd. Taipei, Taiwan 106. Tel.: + 886 2 3366 1629; fax: + 886 2 2392 5294.

E-mail address:etychang@ntu.edu.tw(E.T.Y. Chang).

Contents lists available atSciVerse ScienceDirect

Tectonophysics

(2)

Range east of LV has been considered to be part of the Philippine Sea plate (Ernst, 1977; Hsu, 1956), corroborated by its distinct surface motion (Yu et al., 1997). On the south, the subduction system of Taiwan is where the Eurasian plate subducts towards the east beneath the Philippine Sea plate, forming the N–S trending Manila Trench (Fig. 1). The vigorous orogeny between these two oblique subduction systems thus generates a great complexity in surface deformation of the Taiwan island.

In this study, we adopt two types of measurements that are sensi-tive to the vertical crustal deformation, even though not originally designed to do so— namely the tide gauge and the satellite ocean altimeter. A tide gauge records the coastal sea surface height (SSH) in a continuous fashion in time but relative to the tide gauge location ﬁxed to the ground, while the ocean altimetry measures the “abso-lute” SSH in the geocentric terrestrial reference frame. Their difference, that is, the altimeter-determined SSH minus the tide gauge-determined SSH, thus signiﬁes the crustal vertical deforma-tion of the tide gauge locadeforma-tion in an absolute sense (Cazenave et al., 1999; Mazzotti et al., 2008; Nerem and Mitchum, 2002). Such approach has successfully resolved recent vertical crustal motion of, for example, post-glacial isostatic rebound (Kuo et al., 2004) and the subsidence related to lithospheric plate subduction (García et al., 2007).

Thus, making use of the satellite ocean altimetry data with the long-term recordings from tide gauges around the island residing upon both the Philippine Sea and the Eurasian plates, we shall deter-mine the vertical motion in the Taiwan coastal region. When compared against the independent, continuous GPS vertical measurements that

are neighboring to the tide gauge, these results elucidate the complexity of the vertical deformation of Taiwan coasts and its geophysical mech-anisms in the context of the plate convergence.

2. Data and analysis

The ocean radar altimetry data (ALT for short) used in this study are obtained from the satellite missions of TOPEX/Poseidon and its follow-on Jason-1, available from the DEOS/RADS data center (Delft Institute for Earth-Oriented Space research/Radar Altimeter Database System, also seehttp://rads.tudelft.nl/rads/). Satellite altimetry deter-mines the distance from the satellite to the nadir surface by measur-ing the satellite-to-surface round-trip time ofﬂight of a radar pulse. Knowing the satellite orbit (by ground or GPS tracking), one obtains SSH with respect to the reference ellipsoid for the given location at the given moment.

TOPEX/Poseidon was launched in 1992 into a 10-day repeat orbit of 66° inclination. The mission was followed by Jason-1 since 2001 (e.g.Schrama et al., 2000). The continuous ALT time series we use spans 16 years of 1992_{–2008 at 10-day intervals.} Fig. 2shows the TOPEX/Poseidon–Jason-1 altimetry 10-day repeat-orbit groundtracks near Taiwan (Passes # 051, 127, 164, and 240). Considering the re-gional variations of sea-level trend in the offshore area (Cazenave and Nerem, 2004; Church et al., 2004),ﬁve zones close to our tide gauge stations are selected in this study, indicated as A to E. Here the ALT time series is compiled from the sequence of spatially aver-aged SSH in the selected zone during each given satellite transect (zone B has two transects) (Figs. 2 and 3a). As is generally the case,

Fig. 1. Tectonic setting of Taiwan controlled by the two subduction systems at the Ryukyu and Manila trenches: The large black arrow at the lower-right gives the motion of the Philippine Sea plate relative to the Eurasian plate; the blue shadows show the surface projection of the two Wadati–Benioff zones identiﬁed from the seismicity of 50–150 km depth. Blue dashed lines (T1 and TE) indicate the proﬁles discussed in the text. (CoR: Coastal Range; CR: Central Range; HR: Hsuëshan Range; IP: Ilan Plain; LH: Lanyu islet; LT: Lutao islet; LV: Longitudinal Valley; PH: Penghu islet; TP: Taipei basin; WF: western foothill.)

(3)

the data too close to the coastlines are excluded because of the instru-mental land contamination (e.g.Hwang et al., 2006; Vignudelli et al., 2005). Standard corrections have been made with the ALT RADS data-base including geophysical corrections (dry and wet troposphere and ionosphere), sea state bias, and various tidal effects. The TOPEX/Po-seidon altimetry has achieved a general accuracy at 2–3 cm level, and Jason-1 approaches the 1 cm accuracy (Luthcke et al., 2003; Naeije et al., 2002; Schrama et al., 2000).

The tide gauge (TG) time series that we use are hourly records sustained by the Central Weather Bureau of Taiwan. There are 20 TG stations chosen along the coast of Taiwan and two off-shore islets (Stations LT and PH, seeFig. 2). Only well-maintained stations whose records show steady behavior are used, and we avoid records or re-cord sections containing erratic or discontinuous motions, which are likely the result of equipment maintenance or generated by seismic activities (see examples of TG time series inFig. 3(b)). We further re-move large spurious transients caused by typhoons or other short-term meteorological effects as“outliers” with respect to the 2.5-standard deviation. The distances between the TG stations and the corresponding ALT zones range from 35 to 150 km; the long-term sea-level variability on at least decadal time scales is presumably insensitive to this spatial separation.

We thus model the time series of either ALT (t) or TG (t) as: a0þ a1tþ

Xn i

AisinðωitÞ þ BicosðωitÞ

½ þ residual ð1Þ

where a0, a1, Ai, and Bi, are the coefﬁcients to be estimated by

least-squaresﬁtting. The bracket in Eq.(1)represents the major tidal and seasonal terms with their known periods, including semi-diurnal

tides (M2, S2, N2, K2, L2, T2, and 2N2), diurnal tides (K1, O1, P1, and

Q1), the long-period tides (Msm, Mm, MSf, and Mf), and the annual

and semi-annual terms (same as the Sa, Ssa long-period tides). These harmonic terms are then removed by subtraction as they are of no interest in this study. The coefﬁcient a1represents the optimal

estimate of the temporal rate, or linear trend, of the vertical motion, which is the parameter of interest here. The difference between the linear trends of ALT and TG should be able to give unequivocal in-formation about the long-term vertical ground motion of the TG site itself— a positive value of the linear slope of the time series means the TG site is uplifting, and a negative value subsiding. For-mal errors of the least-squares regression are evaluated by the general error propagation considering the standard deviation of recorded SSH and the data length used for regression (Taylor, 1997). The statistical uncertainty of the ALT–TG trend is determined from the variance-sum of the regression uncertainties of each ALT– TG pair, assuming non-correlation between the two.

We compare our ALT–TG results with the vertical motion inde-pendently measured by the continuous GPS stations along the coast of Taiwan. Here the GPS measurements are reported with respect to the International Terrestrial Reference Frame, the same reference for our ALT–TG values. The GPS time series are daily averaged values pro-cessed and compiled by the Data Center of the Taiwan Earthquake research Center (TECDC) (Hugentobler et al., 2001; Kuo, 2008; Yu et al., 1997) (for more information see TECDC website,http://tecdc. earth.sinica.edu.tw/, and the GPS service website:http://gps.earth. sinica.edu.tw/). The uncertainties of the GPS measurements are generally in the range of 0.2–0.6 mm/yr, depending partly on the time span of the given stations many of which were installed within the last decade.

Fig. 2. Map view of the ALT–TG estimates (circles) as well as the GPS measurements (squares) of the vertical crustal motion, magnitude expressed in color. Black spots indicate the locations of the TG stations. TOPEX/Poseidon–Jason-1 (for ALT) ground tracks and the ﬁve selected zones (A–E) are shown. The GPS measurements affected by co-seismic motions are not presented.

(4)

3. Results

Fig. 3(a) presents the monthly-averaged ALT time series (which have been removed of the high-frequency constituents as above) of the ﬁve selected zones. Undulating monthly at about 200 or 250 mm, the absolute sea level around Taiwan appears to rise at the rate of 4.0–5.0 mm/yr, or twice the rate of the global sea-level rise (Tseng et al., 2009), except for Zone E located at the northern margin of the South China Sea which is at a much slower but uncertain rate of 1.1 ± 1.4 mm/yr.

Fig. 3(b) demonstrates two TG time series and their nearby GPS observations selected for a parallel comparison of vertical motions along the coast of Taiwan. Even though the magnitudes of the tidal constituents vary considerably at different locations around Taiwan, the detided residual TG time series reveal a minor undulation of less than 100 mm. Note that in a relatively short time span the GPS obser-vations sometimes show opposite sense in the vertical motion from the neighboring stations, for example the stations BANP versus KASH (seeFig. 3(b)).Table 1lists all the ALT_{–TG values for the} verti-cal rate at the TG locations (seeFig. 2), and the corresponding nearby GPS measurements for comparison. The uncertainties in ALT–TG values are apparently much larger than those in GPS and seem to stem mainly from the relatively sparse sampling in the ALT series. Note that in general the ALT–TG results are derived from the data se-ries longer than ten years, whereas the GPS are considerably shorter in time span. We thereforeﬁnd it difﬁcult and undesirable to estimate the ALT–TG rates from the difference time series of the two. For the cases where this is possible, our trial results show little difference be-tween the two schemes.

Our comparison results are further summarized inFigs. 2 and 4. In Fig. 2, the vertical motion magnitude is indicated in shade of color with red meaning uplift (positive values) and blue subsidence (nega-tive values). It is evident that, despite the difference in time spans adopted, the ALT–TG and GPS vertical motion measurements are in general agreement with each other, except for sites in the northern end of Taiwan around KL, and JG and AP just south of the station BZL in the western middle Taiwan (groundwater over-withdrawal region, see below). It is interesting to note that BZL and KL are also where the largest subsidences are seen (the colored shade in Fig. 2is made to saturate, at ± 20 mm/yr, for subsidence at these stations).

It is evident that the vertical motions determined from both ALT– TG and GPS portray a general subsidence around most of the Taiwan coasts. ALT–TG and GPS results give predominantly negative values: the (unweighted) average ALT_{–TG rate is about −7.0 mm/yr and} that of GPS−4.0 mm/yr, respectively. If we exclude the large (non-tectonic) subsidence around BZL, the dominant subsidence motion is somewhat reduced to about−4.5 mm/yr and −2.6 mm/yr, respec-tively. On the other hand, in a consistent manner both ALT–TG and GPS values show considerable spatial variability, sometimes quite locally.

4. Discussions

An apparent general subsidence found around the island coast confronts the notion that Taiwan as an island owes its very exis-tence to the upward motion associated with a vigorous orogeny (e.g. Dadson et al., 2003; Ho, 1986). Furthermore, corroborating the subsiding motion detected at some local spots (e.g. Chang et al., 2004; Teng, 1996) and the strongly lateral motions reported in recent GPS observation (e.g. Angelier et al., 2009; Chiu et al., 2008), our complex spatial pattern of the vertical motion indicates local and varied responses to tectonic forcing, and would defy sim-plistic crustal deformation model of Taiwan. We shall discuss them in more detail below.

Fig. 3. (a) Time series from theﬁve ALT data zones as a function of calendar year. Thin gray lines are the raw ALT series; black lines the monthly mean series. Thick dashed lines show the linear uplift trends estimated by least-squares regression. (b) Two TG time series at KS and CK, and their neighboring GPS observations. Thin gray lines are the raw TG time series; black lines the detided section used to determine the vertical rate. Thick dashed lines show the linear LS trends. Units are in mm.

(5)

Table 1

Vertical motion rates obtained in this study. Positive values represent uplift, negative values subsidence. (a) Vertical motion rate determined from ALT–TG

ALT1

TG2 _ALT–TG

Zone LS trend (mm/yr)

station Data period

(yr)

LS trend (mm/yr)

Vertical rate (mm/yr) Name Long. (°E) Lat.(°N)

A + 4.9 ±1.2 KL 121.744 25.157 1992–2002 + 26.7 ± 0.5 −21.8±1.3 LD 121.950 25.120 2004–2007 + 1.5 ± 1.6 + 3.4 ± 2.0 GF 121.862 24.907 1998–2007 + 1.3 ± 0.3 + 3.6 ± 1.2 SA 121.869 24.586 1992–2005 + 22.1 ± 3.1 −17.2±3.3 B + 3.7 ± 1.7 HL 121.623 23.980 1997–2002 −5.2±0.2 + 8.9 ± 1.7 ST 121.500 23.486 2001–2007 + 15.8 ± 1.1 −12.1±2.0 C + 4.3 ± 1.5 CK 121.377 23.089 1993–2003 + 4.2 ± 0.4 + 0.1 ± 1.6 FG 121.191 22.784 1992–2003 + 12.9 ± 0.1 −8.6±1.5 LT 121.450 22.650 2001–2008 + 12.7 ± 0.4 −8.4±1.6 DW 120.896 22.337 2003–2007 + 6.2 ± 0.3 −1.9±1.5 D + 5.0 ± 1.7 DS 121.416 25.177 1994–2008 −3.2±0.2 + 8.2 ± 1.7 HC 120.912 24.850 1992–2008 + 6.2 ± 1.2 −1.4±2.1 TC 120.525 24.289 2004–2008 + 8.9 ± 0.5 −3.9±1.7 FY 120.283 23.900 1992–2008 + 19.8 ± 0.2 −14.8±1.7 BZL 120.138 23.619 1995–2008 + 57.4 ± 1.6 −52.4±2.3 JG 120.078 23.211 1992–2001 + 2.9 ± 0.4 + 2.1 ± 1.7 AP 120.128 22.973 2001–2007 + 16.0 ± 0.5 −11.0±1.8 PH 119.567 23.563 1998–2008 + 8.6 ± 0.7 −3.6±1.8 E + 1.1 ± 1.4 KS 120.281 22.619 1992–2007 + 1.1 ± 0.1 + 0.0 ± 1.4 DG 120.438 22.464 2004–2008 + 7.7 ± 0.4 −6.6±1.4 Note:1

ALT data series are compiled from the T/P and Jason-1 missions, from 1992 to 2008.

2

TG rates are interseismic-phase motion free from co-seismic effects (see text for detail). (b) Vertical motion rate determined by GPS

Station Long. (°E) Lat.(°N) Start End Vertical rate (mm/yr)

West Proﬁle GS09 121.6519 25.2086 2005/08/16 2008/12/31 −0.8±0.3 TANS 121.4269 25.1815 2004/05/17 2008/12/31 −3.4±0.3 KYIN 121.0804 25.04105 2004/05/22 2009/06/30 + 0.4 ± 0.3 TWTF 121.1645 24.95356 2001/11/09 2009/06/30 + 1.2 ± 03 HCHM 120.9846 24.7925 2006/03/17 2008/12/31 + 2.0 ± 0.3 GS14 120.9595 24.8032 2005/09/13 2008/12/31 −1.0±0.2 GS02 120.9823 24.8097 2004/10/30 2008/12/31 −10.6±0.3 JUNA 120.8754 24.68395 2005/02/03 2008/12/31 + 0.4 ± 0.2 MIAO 120.8103 24.58345 2003/05/26 2008/12/31 −5.5±0.2 TASO 120.6951 24.46126 2007/11/08 2009/06/30 −4.0±0.5 GS36 120.6253 24.362 2007/01/01 2009/06/30 −1.0±0.3 CHIN 120.5822 24.271 2001/11/14 2009/06/30 + 2.6 ± 0.3 TACH 120.5351 24.2908 2005/09/30 2009/06/30 −3.1±0.2 TC12 120.5979 24.2075 2005/09/28 2009/06/30 −4.4±0.2 LUKN 120.4351 24.06001 2001/11/03 2008/12/31 −3.0±0.2 CHSG 120.2891 23.86034 2007/01/01 2009/06/30 −33.4±0.3 JIBE 119.6133 23.74142 2006/10/02 2009/06/30 −1.9±0.1 TASI 120.1888 23.72027 2007/10/31 2009/06/30 −17.8±0.8 WIAN 119.4808 23.56753 2006/10/02 2009/06/30 −2.1±0.1 PANG 119.5637 23.5652 1994/01/07 2005/12/26 −0.8±0.2 CHYI 120.1402 23.45076 2004/02/19 2008/12/20 −33±0.2 BDES 120.1719 23.38057 2007/06/12 2009/06/30 −56.6±0.4 PEIM 120.1686 23.2938 2004/04/04 2009/06/30 −19.5±0.4 HOKN 120.1349 23.1884 1998/11/03 2008/12/31 −12.6±0.2 YSAN 120.086 23.1466 2004/12/02 2009/06/30 −6.7±0.2 CKSV 120.22 22.9989 2007/01/30 2008/12/31 −3.2±0.3 CKGM 120.2201 22.9988 2006/01/24 2008/12/31 + 6.0 ± 0.3 GS34 120.2751 22.93922 2007/01/01 2008/12/31 −1.6±0.3 ZEND 120.2176 22.9433 2006/03/29 2008/12/31 + 3.4 ± 0.3 BANP 120.3054 22.6931 2007/04/03 2008/12/31 + 5.2 ± 0.4 KASH 120.2883 22.6145 2004/03/09 2009/06/30 −4.3±0.2 SGAN 120.3497 22.5813 2006/05/19 2009/06/30 −2.3±0.3 HENC 120.7465 22.0039 1996/03/17 2008/12/31 −4.4±0.3 East Proﬁle LNDO 121.9181 25.0974 2003/12/05 2008/12/20 −2.9±0.2 GOLI 121.9874 25.0204 2005/02/06 2008/12/20 −2.7±0.2 FLON 121.9375 25.0204 2004/04/15 2008/12/20 −2.7±0.3 SUAB 121.8679 24.5939 2005/06/29 2009/06/30 −5.1±0.5 SUAO 121.8671 24.5924 1999/10/01 2006/10/01 −5.7±0.6 NAAO 121.8102 24.4493 2004/04/06 2008/12/20 −8.9±0.3 HUAP 121.7494 24.3090 2004/05/24 2009/06/30 −3.4±0.4 PEPU 121.6103 24.0179 2001/11/01 2006/09/30 + 0.9 ± 0.3

(6)

4.1. West coast— stable tectonic basement?

On the west coast of Taiwan, the salient exception is the BZL station where occurred an extraordinary sinking as high as−52.2±2.3 mm/ yr, as corroborated by the GPS results showing a large land subsidence surpassing−50 mm/yr. This is due to over-withdrawal of groundwater for aquaculture, an important economic activity in west Taiwan in the past years, according to the reports of the Groundwater Monitoring Network of the Water Resources Agency (WRA) of Taiwan (WRA of ﬁ-cial website http://www.drc.ntu.edu.tw/gwater/applications-07.php) (e.g.Fan, 2001; Liew et al., 2004). Indeed, compared with the water-level monitored by borehole instruments reporting extreme subsidence as high as 116 mm/yr during 1975–1995, our ALT–TG and the GPS mea-surements show a relative reduction of this anthropogenic subsidence in the recent years.

In contrast, the ALT–TG estimate of the PH station, on the western off-shore islet of Penghu in the same latitude band but located off in Taiwan Strait, shows a slight subsidence at_{−3.6±1.8 mm/yr (}Fig. 4(a),Table 1). The face values of three continuous GPS there (JIBE, PANG, and WIAN) show similar subsidence (Fig. 4(a),Table 1). Located geologically on the continental basement, the Penghu islet is presumably little in ﬂu-enced by the Taiwan orogeny from the east and therefore tectonically inactive (e.g.Ho, 1986). Yet there seems to be an important continental margin opening in the southeast China, of potential of petroleum re-serves (e.g.Covey, 1984), while recent lithological studies stress the ex-istence of a lateral shear zone across the Taiwan Strait in the N–S direction (Chen et al., 2006; Yokoyama et al., 2007). These observations infer that the Taiwan Strait is not tectonically inert but relates to another deformation mechanism out of the Taiwan orogeny itself.

4.2. Northern section of east coast— a complex deformation pattern The east coast is juxtaposed next to the paralleling LV suture zone of convergence and perpendicular to the Ryukyu subduction trench at the middle. Tectonically, the Ryukyu Trench laterally collides into east Taiwan near Hualien, which is at the northern tip of the Coastal Range and LV (Fig. 1).

North of Hualien, both the ALT–TG and GPS measurements along the east coast reveal a non-uniform vertical motion in both spatial

Table 1 (continued)

(b) Vertical motion rate determined by GPS

Station Long. (°E) Lat.(°N) Start End Vertical rate (mm/yr)

East Proﬁle HUAL 121.6135 23.9754 1994/01/04 2006/09/30 −6.6±0.4 HUAL* 121.6135 23.9754 1994/01/04 2008/12/20 + 5.8 ± 0.9 YENL 121.6018 23.9035 2003/01/01 2006/09/30 −12.1±0.3 SOFN 121.5982 23.8703 2005/12/10 2006/09/30 −12.6±0.3 SHUL 121.5627 23.7876 2006/07/01 2008/12/20 −15.6±0.3 JSUI 121.4239 23.4920 2003/05/13 2008/12/08 −2.7±0.4 KNKO 121.5057 23.4722 2002/04/21 2008/12/08 −3.7±0.3 NHSI 121.4530 23.4062 2005/08/13 2008/12/08 −4.7±0.3 PING 121.4543 23.3195 2002/01/22 2008/12/08 + 4.5 ± 0.3 CHEN 121.3736 23.0974 1995/01/31 2003/12/09 −1.0±0.4 CHEN* 121.3736 23.0974 1994/01/14 2008/12/31 + 56.1 ± 1.0 T101 121.3236 23.0203 2004/04/17 2008/12/08 + 1.2 ± 0.3 T102 121.2768 23.0160 2003/12/18 2008/12/08 + 3.2 ± 0.3 SINL 121.2546 22.9083 2005/06/08 2008/12/08 + 3.1 ± 0.3 S104 121.1894 22.8208 1994/03/13 2008/12/08 + 2.1 ± 0.3 FUGN 121.1922 22.7907 2003/12/10 2008/12/08 −6.7±0.3 PEIN 121.1231 22.8011 2005/08/26 2009/06/30 −9.4±0.3 LUDA 121.4759 22.6581 2003/12/15 2006/12/31 −4.0±0.3 TMAM 121.0075 22.6161 1995/03/01 2008/12/31 −5.1±0.2 DAWU 120.8900 22.3406 2005/07/27 2008/12/31 −4.1±0.3 DAJN 120.8650 22.3113 2004/06/04 2008/12/31 −4.1±0.3 LANY 121.5581 22.0373 1994/01/24 2009/06/30 + 10.1 ± 0.4 KDNM 120.7820 21.9494 1994/01/02 2009/06/30 + 0.1 ± 0.2

Note: The GPS time series are adopted from the Taiwan Earthquake research Center, for interseismic-phase motion free from co-seismic effects in recent years (mainly since 2004), except *CHEN and *HUAL, which include the 2003 Chengkung earthquake and the 2006 Hualien near-shore earthquake, respectively.

Fig. 4. Vertical crustal motion rate as functions of latitude along the two separate N–S proﬁles of Taiwan: (a) the west coast, and (b) the east coast. Two earthquakes (near CK and HL stations) with focal mechanism solutions are indicated in (b); their aftershock extents are highlighted in green shade.

(7)

pattern and time span, as implied in the discrepancies between the short-term GPS and the decadal ALT–TG determinations (Fig. 4, Table 1). It may be generated by local adjustment of complex defor-mation combining convergence with extension, an issue that can be further assured by a long-term monitoring in the northern Taiwan. It is worth noting that, barring some local effect (for example at GF) and excluding the very-short-term measurements at the LD and HL, the east ALT–TG results illustrate the northward subsidence (Fig. 4(b)). This feature can also be seen from the ground relief in east-ern Taiwan. InFig. 5a, a northward topographic dipping with a steady ~0.2° with respect to the sea level can be seen from south of Hualien (HL station) all the way to the northern offshore area of Taiwan (in a length about 170 km). This interesting inter-relationship unfortunately cannot be further verified by the relatively short spans of the GPS face values (cf.Dixon, 1991; Feigl et al., 1990), which show a clear segment-ed pattern well consistent with the recent co-seismic motions in a few significant earthquakes occurring in the near field along the east coast (Fig. 4b).

This northward dipping of topography may be a surface manifes-tation of northward subduction. Physically, an old, cold and heavy plate can drive a significant mantle flow, further inducing rapid trench rollback and regional depression. The driven convection flows can induce suction, dragging plate downward and causing the topographic subsidence near the conjunction area (Conrad and Lithgow Bertelloni, 2004).

To shed light on the possible mechanism of the slab suction in the northeast Taiwan, we shall scrutinize the thickness and age of the neighboring Philippine Sea Plate. It is common to check the elastic thickness of the oceanic lithosphere with theﬂexure or bending of the subducting plate, modeled as an elastic plate, at trenches bounded by the overriding plate which exerts a bending moment generated from its weight. Ultimately, the topographic feature of the forebulge at subducting plate becomes a proxy for the elastic plate thickness,

which is proportional to the plate's age. The height of the deflection profile of the plate (H(x)) can be described as (McNutt and Menard, 1982): H xð Þ ¼ wb ffiffiffi 2 p eπ=4exp −π 4 x_−x0 xb−x0 sin π 4 x_−x0 xb−x0 ð2Þ where x and ware both referencing to the location where the topo-graphic relief reaches its most bottom site along the bending plate; xbandwbrespectively indicate the distance and amplitude of the

fore-bulge accounted from the reference sitex0. The“zero” values of the

topographic relief and the distance are defined at the point where the slope of the topographic pro_{file is “zero” along the bending} ocean-ic plate.Fig. 5(b) demonstrates the topographic relief profile across the Ryukyu trench to the east of Taiwan (Profile TE inFig. 1), against various synthetic lithospheric de_{flection profiles depending on x}b. In

making the comparison, we take wb= 0.8 km and xb= 70 km from

the DEM curves inFig. 5(b). The elastic thickness (Te) of the plate

can then be derived (McNutt and Menard, 1982; also seeTurcotte and Schubert, 2002, Equations 3–72, 3–127) to be 28–31 km, assum-ing the mantle density of 3300 kg/m3_{, the Poisson ratio ofv = 0.25,}

and the Young's modulus E of the oceanic lithosphere of 70– 100 GPa. The derived Tecorresponds to the age of 90 Ma or older for

the western Philippine Sea plate next to Taiwan, according to the thickness–age relationship proposed by Stein and Stein (1992). Note that previously the age determination of the western Philippine Sea Plate has been uncertain, ranging from a few to one hundred mil-lion years (e.g.Deschamps et al., 2000; Richard et al., 1986; Sibuet et al., 1999). This is consistent with the assertion that the western Philippine Sea basin is an old, heavy oceanic lithosphere (Deschamps et al., 2000), which can induce a signiﬁcant topographic depression adjacent to the Ryukyu Trench as required in our scenario. This appears to be a plausible mechanism for the general subsiding back-ground observed in the northern section ofFig. 4(b).

4.3. Southern section of the east coast — a saw-tooth deformation scenario

At the northern tip of the southern section of the east coast, the city of Hualien itself is seismically active, suffering from many middle-to-large earthquakes (for example the 1951 earthquake se-quence and the 2006 earthquake in this study). This leaves its_“stable” TG series in short segments, hence the ALT–TG face values at the HL station may not be representative of the long-term value. Indeed, the face value determined from 1997_{–2002 (5 years) in} Table 1 shows an apparent uplift of + 8.9 ± 1.7 mm/yr. On the other hand, the nearby GPS shows an interseismic-phase subsidence at rate of about −6.6±0.4 mm/yr during 1994–2006 excluding the large thrusting event of Oct. 1st, 2006, but an uplift at the rate of about +5.8 ± 0.9 mm/yr when the latter is included (indicated as HUAL* inFig. 4(b) andTable 1). This has actual bearings on theﬁnding of the permanent uplift left in the Hualien TG by the 1951 Hualien earth-quake sequence (e.g.Ota et al., 2005). Note that this behavior is in contrast to the northward dipping trend for the other east-coast TG stations, indicating the extreme complicity of the tectonic“corner” setting of Hualien. South to the Hualien city and the Ryukyu Trench, both ALT–TG and GPS results for the vertical-motion rates quickly turn to a segmented, undulating pattern. Note that the value at LT, a small volcanic islet offshore of southeast Taiwan on the Philippine Sea plate, rides along the same trend with the other eastern stations, indicating that the vertical motion along the eastern coast of Taiwan couples with the eastern oceanic plate.

The segmented behavior of vertical crustal deformation can be recognized in the 2003 Mw 6.5 Chengkung earthquake and the recent 2006 Mw 4.4 Hualien earthquake. The green shades (Fig. 4(b)) high-light the extent of the respective coseismic slip determined by

Fig. 5. (a) N–S topographic proﬁle (T1 inFig. 1) through the east Taiwan. (b) Theﬂexures of oceanic lithosphere as a function of distance from forebulge (xb), for various lithospheric

elastic thicknesses. The thick curve gives the observed topographic relief along Proﬁle TE (Fig. 1).

(8)

aftershocks. The Chengkung earthquake occurred in the eastern near-shore on Dec. 10, 2003, which is apparently a reactivation of the LV fault but in a distinct segment with mainly a thrusting plus a strong lateral motion (Chen et al., 2004; Lee et al., 2006). Locations of the tide gauge station CK and the GPS station CHEN are at its hanging wall. As mentioned above, our reported ALT_{–TG rate value at the CK} station is near zero, 0.1 ± 1.6 mm/yr, pertaining only to the interseis-mic phase prior to the Chengkung earthquake and excludes the in ﬂu-ence of the earthquake. The corresponding GPS station CHEN's steady records also showed a slight subsidence at−1.0±0.4 mm/yr (see Fig. 4(b)). However, taking the longer record of 1996–2008 including the coseismic thrusting of the earthquake, GPS gives an average ver-tical uplift motion of +56.1 ± 1.0 mm/yr (indicated as CHEN* in Fig. 4(b) andTable 1), opposite in sign to the interseismic phase.

Thus, the alternating sense and amplitude of motion between the coseismic and interseismic phases can be modeled by a saw-tooth type of motion as depicted inFig. 6. That is, slow subsidence of the re-gion is intermittently interrupted and overtaken by a series of co-seismic (plus possible short-term post-co-seismic) thrustings that more than reverse the interseismic subsidence trend. The net, very long-term effect then is an overall uplift of the east Taiwan (especially the coast along with the Coastal Range). The topography and geolog-ical data record the average long-term uplift shown by the envelope of the saw-tooth curve, constituting the orogeny of the Coastal Ranges of eastern Taiwan.

Considering the northward-subsiding pattern, the observed subsi-dence from ALT–TG and GPS in the southern section of east Taiwan (in the last decade or two) may simply be the depression of the plate with respect to the sustained subduction at the Ryukyu trench, as observed in the northern section. From this point of view, the so-called interseismic-phase motion is actually the aseismic subsidence consistent to the bending of the subducting plate. In the southern sec-tion of east Taiwan, the subducsec-tion of the Philippine Sea plate can drag down the east Taiwan (which is identiﬁed as parts of the Philippine Sea plate) and cause the regional subsidence; in the meantime the con-siderable thrusting in earthquakes occurs as subject to the lateral convergence of plates. Both the mechanisms of subduction and lateral collision should simultaneously affect the ground motion but in differ-ent temporal and spatial scales.

The saw-tooth behavior can accommodate the interseismic pat-tern in the easpat-tern Taiwan detected by the recent leveling measure-ments (Chen et al., 2011), where the vertical motion exhibits a segmented pattern in the Coastal Range. This deformation model is also consistent with the geological observations in the region: the vertical motion derived from samples from Holocene coral terraces with radiocarbon and Uranium-series dating and corrected of the sea-level rise history gave uplift rates of about 5 mm/yr and

3–5 mm/yr respectively for the eastern limb of Coastal Range and the southernmost Taiwan (Peng et al., 1977; Wang and Burnett, 1990). Lacking suf_{ﬁcient temporal resolution that could identify any} coseismic and interseismic behavior, these measurements naturally detect only the long-term uplift.

On the other hand, such saw-tooth tectonic behavior is interesting in its own right in reﬂecting the discordance of transient-release of strain with long-term tectonic motion. Similar behavior has also been observed in other subduction areas, where the opposite motions correspond to different seismic phases, e.g. the Cascadia subduction near the junction of the southwestern Canada and northwestern USA (Dragert et al., 2001), the Sunda megathrust in West Sumatra (Sieh et al., 2008). All these cases elucidate on the deformation process in subduction seismogenic zone, and further reveal the fact that the subducting plate would play a decisive role in the convergence not only affecting local or regional topography but also the earthquake cycle.

5. Conclusions

Using the decadal ALT–TG measurements and the more recent GPS data, we obtain a comprehensive determination of vertical mo-tion around the coast of the island of Taiwan controlled by an active tectonic convergent complex. Weﬁnd the following:

(1) We confirm for the first time the general pattern of subsidence of the entire Taiwan coast during our studied period of the past 1–2 decades. From the values inTable 1(andFigs. 2 and 4), we see that overall the TG rates are mostly strongly positive while the ALT rates are also positive but significantly smaller (although at 4–5 mm/yr the latter is still twice the global average.) The net result then is a general negative ALT_{–TG, giving the overall} pat-tern of subsidence along most of the Taiwan coast. The vertical motion rates from the ALT–TG and GPS measurements can be represented with a rough (unweighted) mean value of about −4.5 mm/yr and −2.6 mm/yr, respectively (excluding the mea-surements around the BZL station).

(2) The west coast sees no prominent vertical motion, with only a small overall subsidence of land, suggesting weak orogenic buckling consistent with a foreland basin away from the active orogeny along with sediment compactions. The only exception is an extraordinary local subsidence seen in the middle section, which is known for years to be a consequence of the over-withdrawal of groundwater for aquaculture. Our ALT–TG and the GPS measurements show a slight reduction of this anthro-pogenic subsidence in the most recent years.

(3) Both the ALT–TG and GPS estimates of the Penghu islet show a slight subsidence. This suggests that the Taiwan Strait is not tectonically inert but relates to some distinct deformation mechanism out of the Taiwan orogeny.

(4) The different regimes of tectonic influence along the east coast manifest themselves in the complex behavior in their vertical crustal motion. In the northern section, we raise the mechanism of mantleflow suction for the regional subsidence response. The plateflexure indicates that the adjacent Philippine Sea plate is an old, thick oceanic plate, which can drive slab dragging in the east Taiwan, manifesting in a gentle northward down-dip of the topography along this section. In the southern section of the east coast, judging from the different behavior between the co-seismic and interco-seismic vertical motions marked by the major earthquakes during the studied period, we postulate a temporal saw-tooth scenario for the deformation in different phases. It dem-onstrates the opposite motions affected by the different mecha-nisms in the frontal sections of the subduction zone. These different motions can be understood with lateral collision and slab dragging subject to varied temporal and spatial dependence.

Fig. 6. Schematic model of a vertical displacement scenario for the southern segment of the east coast of Taiwan, where successive coseismic slip (Δu) and the opposite interseismic motion form a saw-tooth history for the local deformation.

(9)

Acknowledgments

We thank the Central Weather Bureau of Taiwan for the tide gauge data, and the Taiwan Earthquake research Center (TEC) and the Academia Sinica for the GPS data (courtesy of Drs. Suei-Bei Yu and Long-Chen Kuo). Assistance from Dr. Chung-Yen Kuo on altimeter data processing is also acknowledged. The reviewers helped greatly in improving the manu-script. This study is supported by the National Science Council of Taiwan, under Grants NSC 97-2745-M-002-012, 3113-M-002-002, and 100-2611-M-002-008.

References

Angelier, J., Chang, E.T.Y., Hu, J.C., Chang, C.P., Siame, L., Lee, J.C., Deffontaines, B., Chu, H.T., Lu, C.Y., 2009. Does extrusion occurs at both tips of the Taiwan collision belt? Insights from active deformation studies in the Ilan Plain and Pingtung Plain regions. Tectonophysics 466, 356–376.

Cazenave, A., Nerem, R.S., 2004. Present-day sea level change: observations and causes. Reviews of Geophysics 42, RG3001. doi:10.1029/2003RG000139.

Cazenave, A.K. Dominh, Ponchaut, F., Soudarin, L., Cretaux, J.F., Le Provost, C., 1999. Sea level changes from Topex–Poseidon altimetry and tide gauges, and vertical crustal motions from DORIS. Geophysical Research Letters 26, 2077–2080.

Chang, C.P., Chang, E.T.Y., Angelier, J., Kao, H., Lee, J.C., Yu, S.B., 2003. Stress and strain ﬁeld in Taiwan oblique convergent system: constrains from GPS observations and tectonic data. Earth and Planetary Science Letters 214, 115–127.

Chang, C.P., Chen, K.S., Wang, C.T., Yen, J.Y., Chang, T.Y., Lin, C.W., 2004. Application of space-borne radar interferometry on crustal deformations in Taiwan; a perspective from the nature of events. TAO 15 (3), 523–543.

Chen, H.Y., Yu, S.B., Kuo, L.C., Liu, C.C., 2004. Coseimic and postseismic displacements of the 10 December 2003, Mw 6.5 Chengkung eastern Taiwan, earthquake. The 2004 International Symposium on GNSS/GPS, Dec. 6th -8th, Sydney, Australia. Chen, C.H., Lu, H.Y., Lin, W., Lee, C.Y., 2006. Thermal event records in SE China coastal

areas: constraints from Monazite ages of beach sands from two sides of the Taiwan Strait. Chemical Geology 231, 118–134.

Chen, K.-H., Yang, M., Huang, Y.-T., Ching, K.-E., Rau, R.-J., 2011. Vertical displacement rateﬁeld of Taiwan from geodetic leveling data 2000–2008. Survey Review 43 (321), 296–302.

Chiu, Y.T., Ching, K.E., Hou, C.S., Hu, J.C., Rau, R.J., 2008. Crustal deformation of Ilan plain from GPS observations, 2002–2006. Central Geological Survey Special Publication 20, 111–124.

Church, J.A., White, N.J., Coleman, R., Lambeck, K., Mitrovica, J., 2004. Estimates of the regional distribution of sea level rise over the 1950–2000 period. Journal of Climate 17, 2609–2625.

Conrad, C.P., Lithgow Bertelloni, C., 2004. The temporal evolution of plate driving forces: importance of“slab suction” versus “slab pull” during the Cenozoic. Journal of Geophysical Research 109, B10407. doi:10.1029/2004JB002991.

Covey, M., 1984. Lithofacies analysis and basin reconstruction, Plio-Pleistocene western Taiwan foredeep. Petroleum Geology of Taiwan 20, 53–83.

Dadson, S.J., Hovius, N., Chen, H., Bade, W., Hsieh, M.L., Willet, S., Hu, J.C., Horng, M.J., Chen, M.C., Stark, C.P., Lague, D., Lin, J.C., 2003. Links between erosion, runoff variability and seismicity in the Taiwan orogen. Nature 426, 648–651.

Deschamps, A., Monié, P., Lallemand, S., Hsu, S.K., Yeh, K.Y., 2000. Evidence for early Cretaceous oceanic crust trapped in the Philippine Sea plate. Earth and Planetary Science Letters 179, 503–516.

Dixon, T.H., 1991. An introduction to the Global Positional System and some geological applications. Reviews of Geophysics 29, 249–276.

Dragert, H., Wang, K., James, T.S., 2001. A silent slip event on the deeper Cascadia subduction interface. Science 292, 1525–1528.

Ernst, W.G., 1977. Olistostromes and included ophiolite debris from the Coastal Range of eastern Taiwan. Memoirs: Geological Society of China 2, 97–114.

Fan, K.L., 2001. Some coastal environmental problems in Taiwan. Acta Oceanographica Taiwanica 39, 1–10.

Feigl, K.L., King, E.W., Jordan, T.H., 1990. Geodetic measurement of tectonic deformation in Santa Maria fold and thrust belt, California. Journal of Geophysical Research 95 (B3), 2679–2699.

García, D., Vigo, I., Chao, B.F., Martínez, M.C., 2007. Vertical crustal motion along the Mediterranean and Black Sea coast derived from ocean altimetry and tide gauge data. Pure and Applied Geophysics 164, 851–853.

Hager, B.H., King, R.W., Murray, M.H., 1991. Measurement of crustal deformation using the Global Positioning System. Annual Review of Earth and Planetary Sciences 19, 351–382.

Ho, C.S., 1986. A synthesis of the geologic evolution of Taiwan. Tectonophysics 125, 1–16.

Hsu, T.L., 1956. Geology of the coastal range, eastern Taiwan. Bulletin. Geological Survey, Taiwan 8, 39–63.

Hugentobler, U., Schar, S., Fridez, P., Beutler, E., 2001. Bernese GPS Software: Version 4.2, Astronomical Institute. Astronomical Institute, University of Bern.

Hwang, C., Guo, J., Deng, X., Hsu, H.Y., Liu, Y., 2006. Coastal gravity anomaly from retracked Geosat/GM altimetry: improvement, limitation and the role of airborne gravity data. Journal of Geodesy. doi:10.1007/s00190-062-0052-x.

Kuo, L.C., 2008. The auto-processing system for Taiwan GPS and earth science researches, The technique Report of the Institute of Earth Sciences, Academia Sinica, Origin in Chinese.

Kuo, C.Y., Shum, C.K., Braun, A., Mitrovica, J.X., 2004. Vertical crustal motion determined by satellite altimetry and tide gauge data in Fennoscandia. Geophysical Research Letters 31, L01608. doi:10.1029/2003GL019106.

Lee, J.C., Chu, H.C., Angelier, J., Hu, J.C., Chen, H.Y., Yu, S.B., 2006. Quantitative analysis of surface coseismic faulting and postseismic creep accompanying the 2003, Mw = 6.5, Chengkung earthquake in eastern Taiwan. Journal of Geophysical Research 111. doi:10.1029/2005JB003612.

Liew, P.M., Hsieh, M.L., Shyu, B.H., 2004. An overview of coastal development in a young mountain belt—Taiwan. Quaternary International 115–116, 39–45. Luthcke, S.B., Zelensky, N.P., Rowlands, D.D., Lemoine, F.G., Williams, T.A., 2003. The

1-centimeter orbit: Jason-1 precision orbit determination using GPS, SLR, DORIS, and altimetry data. Marine Geodesy 26, 399–421.

Mazzotti, S., Jones, C., Thomson, R.E., 2008. Relative and absolute sea level rise in western Canada and northwestern United States from a combined tide gauge–GPS analysis. Journal of Geophysical Research 113, C11019. doi:10.1029/2008JC004835. McNutt, M.K., Menard, H.W., 1982. Constraints on yield strength in the oceanic lithosphere

derived from observations ofﬂexure. Geophysical Journal of the Royal Astronomical Society 71, 363–394.

Naeije, M., Doornbos, E., Mathers, L., Scharroo, R., Schrama, E., Visser, P., 2002. Radar altimeter database system: exploitation and extension, RADSxx. Final Report. SRON/NIVR/DEOS publ., NUSP-2 report 02–06, NUSP-2 project 6.3/IS-66. ISBN: 90-5623-077-8.

Nerem, R.S., Mitchum, G.T., 2002. Estimates of vertical crustal motion derived from differences of TOPEX/Poseidon and tide gauge sea level measurements. Geophysical Research Letters 29, 1934. doi:10.1029/2002GL015037.

Ota, Y., Chen, Y.G., Chen, W.S., 2005. Review of paleoseismological and active fault stud-ies in Taiwan in the light of the Chichi earthquake of September 21, 1999. Tectono-physics 408, 63–77.

Peng, T.H., Li, Y.H., Li, F.T., 1977. Tectonic uplift rates for the Taiwan island since the early Holocene. Memoirs: Geological Society of China 2, 57–69.

Rau, R.J., Ching, K.E., Hu, J.C., Lee, J.C., 2008. Crustal deformation and block kinematics in transition from collision to subduction: GPS measurements in northern Taiwan. Journal of Geophysical Research 1995–2005. doi:10.1029/2007JB005414. Richard, M.A., Bellon, H., Maury, R.C., Barrier, E., Juang, W.S., 1986. Miocene to recent

volcanism in eastern Taiwan K–Ar ages and petrography. Tectonophysics 125, 87–102.

Schrama, E., Scharroo, R., Naeije, M., 2000. Radar altimeter database system, RADS: towards a generic multi-satellite altimeter database system.ﬁnal report, 88, SRON/BCRS publ., USP-2 report 00–11.

Sibuet, J.C., Hsu, S.K., Le Pichon, X., Le Formal, J.P., Reed, D., Moore, G., Liu, C.S., 1999. Philippine Sea Plate and Taiwan Sea Plate: formation of Taiwan. Abstract in SEASI Conference, 21, May 9–12th, Montpellier, France.

Sieh, K., Natawidjaja, D.H., Meltzner, A.J., Shen, C.C., Cheng, H., Li, K.S., Suwargadi, B.W., Galetzka, J., Philibosian, B., Edwards, R.L., 2008. Earthquake supercycles inferred from sea-level changes recorded in the corals of west Sumatra. Science 322, 1674–1678.

Stein, C.A., Stein, S., 1992. A model for the global variation in oceanic depth and heat ﬂow with lithospheric age. Nature 359, 123–129.

Taylor, J.R., 1997. Chapter 8: Least-squaresﬁtting. An introduction to error analysis. University Science Books (181–193).

Teng, L.S., 1996. Extensional collapse of the northern Taiwan mountain belt. Geology 24 (10), 949–952.

Tsai, Y.B., 1986. Seismotectonics of Taiwan. Tectonophysics 125, 17–37.

Tseng, Y.H., Breaker, L.C., Chang, E.T.Y., 2009. Sea level variations in the regional seas around Taiwan. Journal of the Oceanography 66, 27–39.

Turcotte, D.L., Schubert, G., 2002. Geodynamics, 2nd ed. Cambridge University Press, Cambridge, New York, Melbourne (456).

Vignudelli, S., Cipollini, P., Roblou, L., Lyard, F., Gasparini, G.P., Manzella, G.M.R., Astraldi, M., 2005. Improved satellite altimetry in coastal systems: case study of the Corsica Channel, Mediterranean Sea. Geophysical Research Letters 32, L07608. doi:10.1029/2005GL22602.

Wang, C.H., Burnett, W.C., 1990. Holocene mean uplift rates across an active plate-collision boundary in Taiwan. Science 248, 204–206.

Wu, F.T., Liang, W.T., Lee, J.C., Benz, H., 2009. A model for the termination of the Ryukyu subduction zone against Taiwan: a junction of collision, subduction/separation, and subduction boundaries. Journal of Geophysical Research 114. doi:10.1029/ 2008JB005950.

Yokoyama, K., Tsutsumi, Y., Lee, C.S., Shen, J.J.S., Lan, C.Y., Zhao, L., 2007. Provenance study of tertiary sandstones from the Western foothills and Hsuehshan Range, Taiwan. Bulletin of the National Museum Natural Science, Series C 33, 7–26.

Yu, S.B., Kuo, L.C., 2001. Present-day crustal motion along the Longitudinal Valley fault, eastern Taiwan. Tectonophysics 333, 199–217.

Yu, S.B., Chen, H.Y., Kuo, L.C., 1997. Velocityﬁeld of GPS stations in the Taiwan area. Tectonophysics 274, 41–59.