Segmentation of the Manila Subduction System from Migrated Multichannel Seismics and Wedge Taper Analysis

Home , Benham Bank, Benham Rise, Manila Trench

Mar Geophys Res DOI 10.1007/s11001-013-9175-7

ORIGINAL RESEARCH PAPER

Segmentation of the Manila subduction system from migrated multichannel seismics and wedge taper analysis

Junjiang Zhu • Zongxun Sun • Heidrun Kopp • Xuelin Qiu • Huilong Xu • Sanzhong Li • Wenhuan Zhan

Received: 12 December 2012 / Accepted: 25 April 2013 Ó Springer Science+Business Media Dordrecht 2013

Abstract Based on bathymetric data and multichannel It suggests that subduction accretion dominates the north seismic data, the Manila subduction system is divided into Luzon and seamount chain segment, but the steep slope three segments, the North Luzon segment, the seamount indicates in the West Luzon segment and implies that chain segment and the West Luzon segment starts in tectonic erosion could dominate the West Luzon segment. Southwest Taiwan and runs as far as Mindoro. The volume variations of the accretionary prism, the forearc slope Keywords Manila subduction system Accretion and angle, taper angle variations support the segmentation of erosion Accretionary/frontal prism Forearc slope angle the Manila subduction system. The accretionary prism is and taper angle Tsunami earthquake Splay faults composed of the outer wedge and the inner wedge separated by the slope break. The backstop structure and a 0.5–1 km thick subduction channel are interpreted in the Introduction seismic Line 973 located in the northeastern South China Sea. The clear dećollement horizon reveals the oceanic Convergent margins are the site of subduction processes sediment has been subducted beneath the accretionary including accretion, tectonic underplating and subduction prism. A number of splay faults occur in the active outer erosion. Subduction accretion and subduction erosion wedge. Taper angles vary from 8.0° ± 1° in the North (tectonic erosion) have been observed and elucidated in the Luzon segment, 9.9° ± 1° in the seamount segment to past decades. The observations of the subduction system 11° ± 1° in the West Luzon segment. Based on variations can be explained by the above two end-member models. between the taper angle and orthogonal convergence rates Subduction accretion preferentially occurs in regions of in the world continental margins and comparison between slow convergence (\7.6 cm year-1) and/or trench sedi- our results and the global compilation, different segments ment thickness [1 km (Clift and Vannucchi 2004). The of the Manila subduction system fit well the global pattern. material within the wedge is shortened horizontally by folding and thrusting similar to analog models (Lallemand et al. 1994). A weaker base of the sediments enhances the & J. Zhu ( ) Z. Sun X. Qiu H. Xu W. Zhan growth of the wedge (Ellis et al. 1999; Beaumont et al. Key Laboratory of Marginal Sea Geology, South China Sea Institute of Oceanology, Chinese Academy of Sciences (CAS), 1999). Accretionary wedge (prism) has been analyzed in 164 West Xingang Rd., Guangzhou 510301, People’s Republic terms of a static force balance using critical wedge theory of China (Davis et al. 1983; Dahlen 1990) and dynamic Coulomb e-mail: [email protected] wedge theory (Wang and Hu 2006). Accretionary prisms H. Kopp are formed by the external tectonic addition of sedimentary Helmeholtz Center for Ocean Research Kiel (GEOMAR), rocks through lithospheric convergence at a subduction Wischhofstrasse 1-3, 24148 Kiel, Germany zone (Moore and Silver 1987). Commonly between the accretionary prism and the underthrusting oceanic crust, a S. Li College of Marine Geosciences, Ocean University of China, dećollement zone is observed (Bangs and Westbrook 1991; Qingdao 266100, China Bangs et al. 1999; Park et al. 2002; Bangs et al. 2003) 123 Mar Geophys Res along the plate boundary. Megasplay faults are very long reflector (BSR) was imaged in the accretionary prism and thrust-type reverse-faults that rise from the subduction seismic attributes indicate that gas hydrate and free gas plate boundary megathrust and intersect the sea floor at the layers are present at the place of BSR (Deng et al. 2006). landward edge of the accretionary prism (Park et al. 2002; The integrated seismic interpretations were related to the Moore et al. 2007). Slip on the megasplay faults in the plate regional tectonic evolution of the northeastern South China boundary contributed to generating devastating historic Sea (Li et al. 2007) and tectonic features, such as frontal tsunami (Park et al. 2002; Moore et al. 2007). dećollement and out-of-sequence of thrusts in the accre- In the Manila subduction system, oceanic lithosphere of tionary wedge (Lin et al. 2009). The purpose of this paper the South China Sea subducts eastward beneath the Phil- is to present a new compilation and analysis of multi- ippine Sea plate along the Manila trench (Bowin et al. channel seismic and bathymetric data which have increased 1978; Taylor and Hayes 1980) (Fig. 1). Most recent our understanding the subduction accretion and erosion in researches focus on the region between Taiwan and Luzon. the convergent continental slope along the Manila trench. The Manila subduction zone includes a very well devel- The main aim is to establish the segmentation of the sub- oped fore arc basin system, the West Luzon Trough and the duction system and to determine which end-member model North Luzon Trough (Ludwig 1970; Pautot and Rangin or mixture model controls the individual segments. 1989). The disappearance of the Manila trench to the west of Henchun ridge/peninsula seems to correlate with a sudden drop in the number of outer rise earthquakes there Tectonic setting of the Manila subduction system (Kao et al. 2000). The accurate northward extension of the Manila trench to southern Taiwan or as a buried feature is The onset of rifting of the proto-China margin occurred still debated (Bowin et al. 1978; Lewis and Hayes 1984; during the latest Cretaceous or Paleocene (65 ± 10 Ma) Liu et al. 1997; Hsiung and Yu 2011). In the east of the (Taylor and Hayes 1983) and is manifested as thereafter South China Sea, a historical M7.5 tsunami event on seafloor spreading evidenced by magnetic anomalies in the February 14, 1934 was reported (Heck 1947; Engdahl and South China Sea marginal basin (Taylor and Hayes 1980, Villasenor 2002; Duong et al. 2009). The generated loca- 1983; Briais et al. 1993; Hsu et al. 2004; Hsu and Sibuet tion of this event could be offshore in the South China Sea 2004; Barckhausen and Roeser 2004) in Middle Oligocene/ at 18°N, 118°E (Berninghausen 1969). The exact location early Middle Miocene time (32–17 Ma). The extinct of this event is the latitude 17°240N, the longitude spreading center has a complex configuration trending 119°11.40E and the focal depth of 35 km (see earthquake N60°E and is dissected by N50°W trending transform catalogues shown in Engdahl and Villasenor 2002) located faults (Pautot and Rangin 1989). The lithosphere of the between the Line 10 and Line 11 (Fig. 1) within the obli- South China Sea becomes older both to the north and to the que seamount subduction domain. Tsunami height may south away from the east–west trending relict spreading reach 14 m near the Philippines and southwest of Taiwan center now located between 15°N and 16°N (Fig. 1) (Bo- at the Manila trench (Ha et al. 2009). The Stewart Bank has win et al. 1978; Taylor and Hayes 1980, 1983). Based on been proposed as a seamount protruded through the forearc the reconstruction of the South China Sea region, the in the surface (Pautot and Rangin 1989), and it is not clear Manila trench was located at the eastern side of the South whether the seamount subduction is related to this event China Sea basin in the early Miocene when the seafloor although it is very close to this tsunami event (Fig. 1). The spreading has proceeded (Taylor and Hayes 1983) or at the subduction convergence rate decreases southwards from southern side in the middle Miocene when southward 98 mm year-1 to 52 mm year-1 as established by GPS subduction of proto-South China Sea initiated at 45 Ma measurements (Rangin et al. 1999) and no large difference (Hall 2002). The Manila subduction system progressively with other investigations (Seno et al. 1993; Kuo et al. evolves from normal subduction to initial arc-continent 1999). Subduction accretion clearly dominates the north collision of the Taiwan orogen (Suppe 1984; Kao et al. Manila subduction system, however, in the central and the 2000). Seismicity patterns indicate a clear subduction zone, southern portion it remains unclear which mechanism although the deepest seismic events occur only to about dominates which segment. 200 km (Kao et al. 2000; Zhu et al. 2005). Between about In 2001, a 240 multichannel seismic (MCS) section 116°E and 118°Eat15°N, many of the seamounts of the (Line 973, Fig. 2) was acquired by Guangzhou Marine Scarborough chain have formed along the axis of the Geological Survey (GMGS) of the Chinese Ministry of central segment (Pautot et al. 1986) (Fig. 1). The Scar- Land and Resources (Li 2005). Seismic interpretations of borough Seamount Chain is present near the axis of the this line were published and addressed questions in the seafloor spreading magnetic lineation (Taylor and Hayes different aspects (Deng et al. 2006; Ding et al. 2006;Li 1980; 1983; Briais et al. 1993; Hsu et al. 2004; Barck- et al. 2007; Lin et al. 2009). A clear bottom-simulating hausen and Roeser 2004). The Philippine Sea plate is 123 Mar Geophys Res

Fig. 1 The bathymetric map along the Manila trench. Bathymetric of the Ms 7.5 tsunami earthquake on 14 February 1934 (location from data is from the GEBCO 1-min grid (http://www.gebco.net/data_ Engdahl and Villasenor 2002). A red square the location of ODP Site and_products/). Number 1 to 16 shows the self-deﬁned bathymetric 772 (Harding et al. 1990). A double dotted line the location of the lines from the Line 1 to Line 16. White line shows the shot locations extinct spreading center. Yellow arrow lines three segments of the of the seismic Line 973 and red line AB indicates in Fig. 2. Seismic Manila trench. The convergence rates and direction along the Manila Line M-2005 is from McIntosh et al. (2005). C-2003 seismic line is trench are from Rangin et al. (1999) and Seno et al. (1993). The from Chi et al. (2003). L-1970-1 and L-1970-2 are from Ludwig seaﬂoor topography is contoured and marked at 200, 1,000, 2,000, (1970). Black circles Ms [6.0 earthquakes (data from IRIS event 3,000 and 4,000 m. NLS North Luzon segment, WLS West Luzon catalogs). Red triangles the locations of volcanoes (data from the segment, SCS seamount chain segment, NLT North Luzon Trough, Smithsonian Institution Global Volcanism Program ‘‘Volcanoes of WLT West Luzon Trough, ELT East Luzon Trough, SB Stewart Bank, the World’’ http://www.volcano.si.edu/gvp/). A red star the location LG Lingayen Gulf moving northwestward at 310° with a rate of Benham Rise is underlain by a series of basement ridges 71 mm year-1 (Seno et al. 1993). The Benham Rise is a trending N–S or NNE-SSW, roughly parallel to the trend of broad topographic high centered east of Luzon at about the East Luzon Trough (Fig. 1). In the north off the 17°N (Fig. 1). The crest of the Benham Rise is about southern tip of Taiwan, the accretionary wedge in the 2,000–3,000 m deep and reaches to within 38 m of the sea incipient arc-continent zone was divided into lower slope, surface at Benham Bank (Lewis and Hayes 1983). The upper slope and backthrust domains (Lin et al. 2009). In the

123 Mar Geophys Res accretionary wedge (south of 21.5°N), the oceanic crust of Fig. 2 Part of the time-migrated seismic image of Line 973 located at c the South China Sea is thrust beneath the Luzon arc with a the Manila subduction system (location shown in Fig. 1). The subduction system is composed of South China Sea basin, the 55° east dipping Wadati–Benioff zone (Kao et al. 2000). accretionary prism and North Luzon Though (NLZ). The clear The accretionary wedge widens from 80 km in the south dećollement horizon indicates the high amplitude. Buried subducted (20.5°N) to 180 km near the southern tip of Taiwan (Lin seamounts are identified in the seismic section. A megathrust fault et al. 2009). In southern Taiwan, the full scale arc-conti- along the plate boundary is imaged and a number of splay faults intersect the seafloor in the outer wedge domain. The backstop nent collision is manifested by the accretion Luzon arc onto structure is located in the inner wedge domain. The lower slope, the Eurasian margin, which forms the Coast Range in transition zone and the upper slope interpreted by Lin et al. (2009) are eastern Taiwan (Suppe 1984; Teng 1990). A 3-D archi- marked in the figure. A clear bottom-simulating-reflector was tecture of the intersection of the Philippine Sea plate with observed (Deng et al. 2006). The compressed North Luzon Trough is clearly imaged the Eurasian plate in northern Taiwan was presented and the wedge-shaped Eurasian plate on top of the Ryukyu subducting plate is connected to the Eurasian plate on the (Ludwig 1970; Chi et al. 2003; McIntosh et al. 2005) Ryukyu side and coupled to the NW moving Philippine Sea (Figs. 1, 3). plate by friction at the plate interface (Wu et al. 2009a; Malavieille et al. 2002). Subduction polarity reversal appears at the arc-continent collision zone in Taiwan and it Results has been interpreted as a mechanism of slab breakoff (Teng et al. 2000; Clift et al. 2003). The geometry of rift basins in Structure of the Manila subduction system the southwest of Tainwan and transform faults of the passive margin of Eurasia has been shown to play a critical The seafloor morphological characteristic of the Manila role in the evolution of the Taiwan collision, such as the trench curves between Taiwan in the north and Mindoro in clastic sediment of the orogen (Teng 1990; Lin et al. 2003) the south (Fig. 1). The oblique subduction of the South and the kinematic history (Mirakian et al. 2012) and seis- China Sea crust caused the material input and output in the motectonics (Kao et al. 2000; Gourley et al. 2007).The subduction system. A well developed forearc basin system, tectonic features of the Taiwan orogeny and subduction the West Luzon Trough and the North Luzon Trough lies zones in the southern Taiwan are investigated by the trenchward of the volcanic arc and forms the Manila sub- TAIGER experiment using geological and geophysical duction system forearc (Ludwig 1970; Pautot and Rangin methods (Liu et al. 2007; Lee et al. 2009; Wu et al. 2007, 1989) (Fig. 1). The seismic Line 973 was reinterpreted and 2009b; McIntosh et al. 2009; Deng et al. 2013). detailed descriptions by compared to the previous literature (Deng et al. 2006; Ding et al. 2006; Li et al. 2007) and showed the clear structure of subduction zone and indicates Data and methods some tectonic features, such as the huge accretionary wedge, underthrust sediment and megathrust (Lin et al. Multichannel seismic and bathymetric data were collected 2009). Deng et al. (2006) addressed the formation of BSR and applied in this study. A 240 channels multichannel in the upper slope by analysis of the seismic line 973. We seismic (MCS) data (Fig. 2) was acquired in 2001 with a apply this seismic image to address the volume of accre- 12.5 m trace interval and 50 m shot interval. The total tionary wedge, subducted seamount and megasplay faults volumes of air guns are 3,000 cubic inch and record times in the plate boundary, and most of seismic interpretations are 10 s (Li 2005). Seismic interpretations of this line were are similar to Lin et al. (2009) (Fig. 2). published and addressed questions in the different way According to the morphological features and calculated (Deng et al. 2006; Ding et al. 2006; Li et al. 2007; Lin et al. wedge taper along the Manila trench, the Manila trench is 2009). In this study we attempt to quantify the forearc composed of three segments from Taiwan to the south- slope angle and taper angle by defining bathymetric lines ernmost of Luzon: the North Luzon segment, the seamount perpendicular to the trench (Fig. 1) and this method is chain segment and the West Luzon segment (yellow dotted similar to that used by Kopp et al. (2009). The bathymetric arrows in Fig. 1). The detail structure characteristics of data was extracted using 1 km grid interval with the linear three segments are indicated here, respectively. interpolation method along the strike lines. These bathymetric lines straddle over the Manila trench with the length North Luzon segment of 220 km. The mean forearc slope angle is calculated close to trench with the distance of 40 km within the The North Luzon segment is located between Taiwan and the accretionary prism. The dip angle of the subducted plate is southern tip of the North Luzon Trough. The bathymetric calculated from the published velocity-depth models data indicate the connection of the northernmost Manila 123 Mar Geophys Res

123 Mar Geophys Res

Fig. 3 Velocity–depth models (a) stride over the Manila subduction system. a The seismic line M-2005 is from McIntosh et al. (2005). b The seismic line C-2003 is from Chi et al. (2003). c Seismic lines L-1970-1 and d L-1970-2 are after Ludwig (1970) (location shown in Fig. 1). e A schematic diagram shows the relationship among the slope angle, dip angle and taper angle for the accretionary prism. FC forearc crust, HC Hengchun core, NLA North Luzon arc, NLT North (b) Luzon Trough, WLT West (e) Luzon Trough

(c)

(d)

trench and the South China continental margin southwest of to the Wadati–Benioff zone (Lin et al. 2009). The morpho- Taiwan (Fig. 1). The arc-continent collision zone consists of logical characteristics of the Manila trench gradually dis- three structural styles, i.e., lower slope domain, upper slope appear northward, accompanied by the decreasing width of domain and backthrust domain (Lin et al. 2009; Ku and Hsu the forearc basin and increasing width of the accretionary 2009) (Fig. 2) as interpreted from multichannel seismic prism. Part of the seismic Line 973 intersects the trench and surveying. In the incipient arc-continent collision zone slope domains (Fig. 2, location AB in Fig. 1). A slope break (between 21.5° and 23°N), crustal seismicity characterizes separates the 108 km wide accretionary prism into the outer the forearc and arc regions in addition to earthquakes related wedge and the inner wedge. A frontal de´collement horizon

123 Mar Geophys Res beneath the accretionary prism indicates the accreted sedi- megathrust fault is clear marked in the seismic Line 973 ments above and underthrust oceanic sediments below beneath the outer wedge (Fig. 2). The slip on the splay (Fig. 2). Pelagic and hemipelagic sediments deposited on faults is thought to play an important role in tsunami oceanic crust are transported into the trench by plate motions genesis (Park et al. 2002; Moore et al. 2007; Kopp et al. (Lewis and Hayes 1984). Above the frontal dećollement, a 2009). series of imbricate splay faults cut through the outer wedge and intersect the seafloor (Fig. 2). The inner wedge, which Seamount chain segment serve as large backstop structure (Fig. 2) mainly consists of sedimentary rocks with P-wave velocities of 1.6–4.8 km s-1 The Scarborough seamount chain obliquely subducts (Deng et al. 2013), which reveal the low sound impedence beneath the Manila trench between 15°N and 17°N contrast within the inner wedge. At the Ocean Drilling Pro- (Fig. 1). Based on a morphostructural study (Pautot et al. gram (ODP) Site 772 (119°420E, 16°390N) (location see 1986), seamount chain indicates a NW–SE direction Fig. 1), only one lithostratigraphic unit consisted of greenish spreading direction in the 150–200 km wide axial region of gray siliciclastic silty clay and claystone was recognized the South China Sea and most of N50°E striking normal although it reaches only a depth of 361 mbsf. and was not fault scarps formed during a spreading episode throughout penetrated to basement (Harding et al. 1990). the axial region. The volume and size of the seamounts Beneath the trench a subducted seamount or buried ridge decreasing from west to east along this axial region was may facilitate the thin oceanic sediment thickening to fill into observed by the detailed Seabeam survey (Pautot and a subduction channel and to uplift this channel (Fig. 2) like Rangin 1989). Stewart Bank (indicated in Fig. 1) was Ecuador margin (Sage et al. 2006; Collot et al. 2008). The proposed as a piece of the subducted Scarborough se- two-way time of the subduction channel shows 0.5–0.7 s in amount, which has protruded through the forearc in recent the seismic record section (Fig. 2). Assuming the P-wave time (Pautot and Rangin 1989). A reflection profile indi- velocities of 2.2–3.0 km s-1 similar to the velocities of cates the Stewart Bank and the West Luzon Trough are sediment lens in Ecuador margin (Sage et al. 2006), its separated by a fault zone (Hayes and Lewis 1984). The thickness will be 0.5–1 km. Seamount subduction may cause collision of seamounts with the Manila trench has caused the geometry variation of the subduction channel (Fig. 2). uplift of the forearc (Pautot and Rangin 1989). Under the accretionary prism, a transition crust was proposed Our study reveals that the mean forearc slope angle in by McIntosh et al. (2005), however the boundary of the the north of the seamount chain segment varies from 2.0° transition crust to oceanic crust from the arc-continent col- to 3.5° (Fig. 4, from Line 10 to Line 12) and varies from lision to the south is unclear. The typical characteristics of 5.5° to 5.7° (Fig. 4, from Line 13 to Line 14) in the south the North Luzon segment are a large accretionary prism (the of the segment. Some of the seamounts are coming close to lower slope) (Fig. 3), with decreasing width from about the trench (Fig. 1 at Line 11 and Line 12). The variation of 200 km to about 50 km along the strike of trench from north the forearc slope is correlated to the different tectonic to south (Fig. 4, from Line 1 to Line 8), and variations of the mechanism. The higher forearc slope angle may suggest North Luzon Trough structure within the upper slope. The the oblique subducted seamount has protruded through the size of the accretionary prism varies significantly along trench and uplifted the accretionary prism or oversteepen- strike of the trench and is correlated with the local thickness ing of the slope plays a role. Subduction of a buried se- of turbidite sediments within the trench (Fig. 4). The amount was observed in the north Ecuador margin, which bathymetric slope angle as the forearc slope angle is calcu- drives dećollement thrust to rise and steepens the lower lated over a distance of 40 km perpendicular to the trench slope close to trench (Collot et al. 2008). The uplift of the axis. The mean forearc slope angles are about 1.2°–3.6° in accretionary prism due to the shape of subducting high was the northern segment (Fig. 4). modeled by sandbox experiments (Dominguez et al. 2000). The large accretionary prism is composed of the outer wedge and the inner wedge with the width of about 105 km West Luzon segment along the seismic Line 973 (Fig. 2). The outer wedge represents a compressive zone with discrete localization of West Luzon segment is mainly composed of the Manila deformation along the thrust faults (Kopp et al. 2009). A trench, the frontal prism and West Luzon Trough (Figs. 1, compressive thrust zone is characterized by the imbricate 4). The West Luzon Trough is separated from the North thrusts identified beneath the outer wedge at the seismic Luzon Trough by a large westward-trending submarine section distance of 65–90 km (Fig. 2, left top). In the ridge complex from the coast near Lingayen Gulf (Fig. 1). seismic section the inner wedge shows the uplift charac- The northern border of the ridge (Stewart Bank) marks an teristic compared to outer wedge separated by slope break important structure change between north and central Lu- (Fig. 2). Its composition and ages are not clear. A zon (Ludwig 1970). In general the sediment fill of the West 123 Mar Geophys Res

Fig. 4 Forearc slope angle variation from the Line 1 to Line 16 along the accretionary prism or frontal prism between trench axis and north/ the Manila subduction system. The bathymetric data is extracted from west Luzon Trough. NLT North Luzon Trough, SSC Subduction the GEBCO 1-min grid data. The dashed line marks the boundary of Seamount Chain, WLT West Luzon Trough

Luzon Trough is not intensely deformed and sediment 1984). The seismic reflection, gravity and bathymetric thickness in the Trough reaches approximately 4.5 km profiles demonstrate that the Manila trench continues based on velocities of 1.7–3.5 km s-1 (Hayes and Lewis southeastward to 12°400N, where it continues onshore as a 1984; Lewis and Hayes 1984). The basement is composed belt of active faults and of gravity lows along southwest of deformed sediments and perhaps igneous rocks with Mindoro (Karig 1983). velocities of 4.1–5.7 km s-1 (Fig. 3). The landward flank Based on the taper analysis and calculation from the of the Trough is floored by Zambales ophiolite, which bathymetric data, the mean forearc slope angle ranges exposed onshore east of the Trough (Hayes and Lewis from 5.3° to 5.7° (Fig. 4). The internal structure and

123 Mar Geophys Res stratigraphic development of the Trough (forearc basin) is location of ODP Site 772 is situated in the seamount seg- closely linked to the growth and uplift of subduction ment and it is far away from the West Luzon segment. complex (accretionary prism) (Lewis and Hayes 1984). Comparing to accretionary and erosive margins in the However, the width of the accretionary prism is about world (Lallemand et al. 1994; Clift and Vannucchi 2004), 20–38 km in the West Luzon segment and is smaller than the north Luzon and seamount segment all fall into the widths of prism (70–140 km) in the North Luzon segment. typical accretionary margins and the West Luzon segment So the volume of the accretionary prism is smaller than the falls into the erosive margins due to large orthogonal North Luzon segment and it termed as the frontal prism convergence rates and slope angles (Fig. 6). The width of here (Fig. 4). The multichannel seismic profile (BB0 see the accretionary prism is about 29–35 km (Fig. 4 shown in Fig. 4 in Lewis and Hayes 1984) indicates a smaller frontal line 13-line 16, Fig. 5 shown between latitude 14° and prism corresponding to the lower slope variation in the 16°300). The accretionary mechanism dominates the North bathymetric profiles (Fig. 4 line 11–16). Luzon segment indicated with a less than 3° slope angles (Fig. 5). The width of the accretionary prism quickly grows from 44.4 km at latitude 19°60 to 130 km at latitude 21°480 Discussion in the accretionary setting (Fig. 5). The subduction accretion and erosion may be coeval Subduction accretion and erosion along the Manila trench, especially in the local area with the seamount subduction segment (forearc slope angles Subduction accretion has been observed by the previous may be less or larger than 3°, Fig. 5). The hypocenter geophysical surveying in the North Luzon segment (Bowin projection through the Manila trench indicated the slab dip et al. 1978; Karig 1983; Hayes and Lewis 1984; Liu et al. is about from 10° to 45° at shallow depths and it steepens 1997; Sibuet et al. 2002; Chi et al. 2003; McIntosh et al. gradually southward (Bautista et al. 2001). The motion of 2005). It is important to quantify the degree of this tectonic the upper plate will also cause the large variations of the process. The dip angle of the subducted plate beneath the orthogonal convergence rate and in this paper we assume prism is 6.7° at the distance of 100–140 km along the that the upper plate an inert thing and segments of the seismic line M-2005 and 4.7° at the distance of 40–80 km subduction zone shows only show the modern segmenta- along the seismic line C-2003 (Fig. 3). The mean dip angle tion. The variation of the orthogonal convergence rate of the subducted plate is 5.7° ± 1° in the North Luzon when Luzon volcanic arc moved in the past will cause a segment. The orthogonal convergence rate is 30 km m.y.-1 large uncertainty to define accretionary margins or erosive in the North Luzon segment and 90 km m.y.-1 in the West margins. Based on the paleomagnetic study, the Philippine Luzon segment (Taylor and Hayes 1980; Karig 1983; Seno sea plate indicated a discontinuous clockwise rotation at et al. 1993; Rangin et al. 1999; Clift and Vannucchi 2004). The mean slope angle in the North Luzon segment is 2.3° ± 1° (Fig. 4, Line1–Line10), 4.2° ± 1° in the seamount segment and 5.3° in the West Luzon segment. Taper angles of the different segments along the Manila trench are 8.0° ± 1° in the North Luzon segment, 9.9° ± 1° in the seamount segment and 11° ± 1° in the West Luzon segment. The mean trench sediment thickness is about 2 km in the North Luzon and seamount chain segment and about 1 km in the West Luzon segment (Ludwig et al. 1979; Taylor and Hayes 1980). Erosive convergent margins typically show a larger than 3° bathymetric slope angle (Clift and Vannucchi 2004) and commonly indicate a small frontal prism at the deformation front. This is supported by the multichannel seismic section BB0 at the latitude of 15°N (Lewis and Hayes 1984) and it indicates a prism width of 15 km. Our results show that the steep slope indicates in the West Luzon segment and implies the erosive mechanism could control the West Luzon segment (slope angles larger than 3° between lati- Fig. 5 Diagrams showing the relationship between the forearc slope angle and prism width variation. Red dots curve shows the width 0 tude 14° and 16°30 , Fig. 5). However, this implication variation of the accretionary/frontal prism. The black square curve requires high resolution seismic data to prove this. The shows the variation of the forearc slope angle 123 Mar Geophys Res

Fig. 6 Global compilation of convergent margins modiﬁed from Clift and Vannucchi (2004). The NLS and SCS clearly map in the accretionary regime, whereas the WLS indicates the erosive regime

the different time (Hall et al. 1995a, b) and it is worth could be recorded by subsurface parameters of subduction considering that the collision point between Luzon and zone (slope angle, taper angle and convergence rate) or by South China is migrating towards the southwest with time the rock record. In the seamount chain segment, when the (Clift et al. 2008). Subduction erosion removes crustal large convergence rate variation with the same taper angle material from a forearc wedge above subducted plate and will cause this segment to fall into the erosive margins was explained by abrasion and hydrofracturing due to ﬂuid (Fig. 6). The accurate results require the detailed velocity- overpressure (von Huene et al. 2004) or by basal erosion depth models in the seamount segments and the West along the plate boundary (Ranero and von Huene 2000; Luzon segment. We suggest that subducted seamount Zhu et al. 2009). The transition from accretion to erosion or facilitate to the activity of megathrust faults along the plate coeval exist along the same margin may be caused by the boundary, therefore caused the slip of faults and most rebuilding and adjusting of plate tectonic. This process likely contributed to generating historic tsunamis.

123 Mar Geophys Res

Conclusions Berninghausen WH (1969) Tsunami and seismic seiches of Southeast Asia. Bull Seismol Soc Am 59(1):289–297 Bowin C, Lu RS, Lee CS, Schouten H (1978) Plate convergence and Time-migrated seismic Line 973 images the structure of accretion in Taiwan-Luzon region. AAPG Bull 62:1645–1672 the accretionary prism which consists of outer wedge and Briais A, Patriat P, Tapponnier P (1993) Updated Interpretation of inner wedge separated by the slope break. The dećollement Magnetic Anomalies and Seafloor Spreading Stages in the South horizon in the outer wedge indicates the oceanic sediment China Sea: implications for the Tertiary Tectonics of Southeast Asia. J Geophys Res 98(B4):6299–6328. doi:10.1029/92JB02280 has been subducted beneath the accretionary prism and Chi WC, Reed DL, Moore G, Nguyen T, Liu CS, Lundberg N (2003) generates a 0.5–1 km thick subduction channel above the Tectonic wedging along the rear of the offshore Taiwan subducted oceanic crust. According to the variation of accretionary prism. Tectonophysics 374:199–217. doi:10.1016/j. forearc slope angle, taper angle and the volume variations tecto.2003.08.004 Clift PD, Vannucchi P (2004) Controls on tectonic accretion versus of the accretionary prism, the Manila subduction system erosion in subduction zones: implications for the origin and consists of three segments, i.e., the North Luzon segment, recycling of the continental crust. Rev Geophys 42:RG2001. doi: the seamount chain segment and the West Luzon segment. 10.1029/2003RG000127 Taper angle variation from 8.0° ± 1° in the North Luzon Clift PD, Schouten H, Draut AE (2003) A general model of arc- continent collision and subduction polarity reversal from Taiwan segment, 9.9° ± 1° in the seamount segment to 11° ± 1° and the Irish Caledonides. In: Larter RD, Leat PT (eds) Intra- in the West Luzon segment suggest that subduction oceanic subduction systems: tectonic and magmatic processes. accretion dominates the north Luzon and seamount chain Special Publication 219, Geological Society, London, pp 81–98. segment, but the steep slope shows in the West Luzon doi:10.1144/GSL.SP.2003.219.01.04 Clift PD, Lin ATS, Carter A, Wu F, Draut AE, Lai TH, Fei Y, segment and implies the erosive mechanism controls the Schouten H, Teng L (2008) Post-collisional collapse in the wake West Luzon segment. of migrating arc-continent collision in the Ilan Basin, Taiwan. In: Draut AE, Clift PD, Scholl DW (eds) Formation and applications Acknowledgments We thank the crew members of the R/V Tanbao of the sedimentary record in arc collision zones. Geological for their efforts for collecting the Line 973 seismic data. We used the Society of America Special Paper 436, pp 257–278. doi:10.1130/ GMT software (Wessel and Smith 1991) and Seismic Unix software 2008.2436(12) package (Stockwell 1999) to plot several figures. Reviews by guest Collot JY, Agudelo W, Ribodetti A, Marcaillou B (2008) Origin of a Editor Peter Clift and two anonymous reviewers greatly improved the crustal splay fault and its relation to the seismogenic zone and manuscript. This study has been supported by the National Nature underplating at the erosional north Ecuador–south Colombia Science Foundation of China (Grant No.: 41176054 and Grant No.: oceanic margin. J Geophys Res 113:B12102. doi:10.1029/ 41006030) and the Knowledge Innovation Program of the Chinese 2008JB005691 Academy of Sciences (SQ200910), and the Joint Funds of the Dahlen FA (1990) Critical taper model of fold-and-thrust belts and National Natural Science Foundation of China and Guangdong accretionary wedges. Annu Rev Earth Planet Sci 18:55–99. doi: province (Grand No.: U0933006). 10.1146/annurev.ea.18.050190.000415 Davis DM, Suppe J, Dahlen FA (1983) Mechanics of fold-and-thrust belts and accretionary wedges. J Geophys Res 88:1153–1172. doi:10.1029/JB088iB02p01153 References Deng H, Yan P, Liu H, Luo W (2006) Seismic data processing and the characterization of a gas hydrate bearing zone offshore of the Bangs N, Westbrook G (1991) Seismic modeling of the decollement southwestern Taiwan. TAO 17(4):781–797 zone at the base of the Barbados ridge accretionary complex. Deng J, Wang T, Yang B, Lee C, Liu C, Chen S (2013) Crustal J Geophys Res 96:3853–3866. doi:10.1029/90JB02138 velocity structure off SW Taiwan in the northernmost South Bangs N, Shipley T, Moore J, Moore G (1999) Fluid accumulation China Sea imaged from TAIGER OBS and MCS data. Mar and channeling along the north Barbados ridge decollement Geophy Res. doi:10.1007/s11001-013-9166-8 thrust. J Geophys Res 104:20399–20414. doi:10.1029/1999JB Ding W, Yang S, Chen H, Cheng X, Wu N (2006) Arc-continent 900133 collision orogeny in offshore Taiwan during Neogene. Chin J Bangs N, Christeson G, Shipley T (2003) Structure of the Lesser Geol 41:195–201 (in Chinese) Antilles subduction zone backstop and its role in a large Dominguez S, Malavieille J, Lallemand S (2000) Deformation of accretionary system. J Geophys Res 108(B7):2358. doi: margins in response to seamount subduction: insights from 10.1029/2002JB002040 sandbox experiments. Tectonics 19(1):182–196. doi:10.1029/ Barckhausen U, Roeser, HA (2004) Seafloor spreading anomalies in 1999TC900055 South China Sea revisited. In: Clift P, Wang P, Kuhnt W, Hayes Duong NA, Thuy NN, Kimata F, Meilano I, Yen LTH (2009) D (eds) Continent–ocean interactions within east Asian marginal Assessment of bathymetry effects on tsunami propagation in the seas. Geophysical Monograph, 149, American Geophysical east Vietnam Sea. Adv Nat Sci 10(4):479–489 Union, Washington, DC, pp 121–125. doi:10.1029/149GM07 Ellis S, Beaumont C, Pfiffner OA (1999) Geodynamic models of Bautista BC, Bautista MLP, Oike K, Wu FT, Punongbayan RS (2001) crustal-scale episodic tectonic accretion and underplating in A new insight on the geometry of subducting slabs in northern subduction zones. J Geophys Res 104(B7):15169–15190. doi: Luzon, Philippines. Tectonophysics 339:279–310. doi:10.1016/ 10.1029/1999JB900071 S0040-1951(01)00120-2 Engdahl ER, Villasenor A (2002) Global seismicity: 1900–1999. In: Beaumont C, Ellis S, Pfiffner A (1999) Dynamics of sediment Lee WHK, Kanamori H, Jennings PC, Kisslinger C (eds) subduction-accretion at convergent margins: short-term modes, International handbook of earthquake and engineering seismol- long-term deformation, and tectonic implications. J Geophys Res ogy, Part A, Chapter 41. Academic Press, London, pp 665–690. 104(B8):17573–17601. doi:10.1029/1999JB900136 doi:10.1016/S0074-6142(02)80244-3

123 Mar Geophys Res

Gourley JR, Byrne T, Chan YC, Wu F, Rau RJ (2007) Fault geometries American Geophysical Union Geophysical Monograph Series illuminated from seismicity in central Taiwan: implications for 27, pp 57–78. doi:10.1029/GM027p0057 crustal scale structural boundaries in the northern Central Range. Lewis SD, Hayes DE (1984) A geophysical study of the Manila Tectonophysics 445:168–185. doi:10.1016/j.tecto.2007.08.013 trench, Luzon, Philippines, 2, Forearc basin structural and Ha DM, Tkalich P, Soon CE, Megawatif K (2009) Tsunami stratigraphic evolution. J Geophys Res 89:9196–9214. doi: propagation scenarios in the South China Sea. J Asian Earth 10.1029/JB089iB11p09196 Sci 36:67–73. doi:10.1016/j.bbr.2011.03.031 Li J (2005) Evolution of China’s marginal seas and its effect of Hall R (2002) Cenozoic geological and plate tectonic evolution of SE natural resources. Ocean press, Beijing Asia and the SW Pacific: computer-based reconstructions and Li CF, Zhou Z, Li J, Hao H, Geng J (2007) Structures of the animations. J Asian Earth Sci 20:353–434. doi:10.1016/S1367- northeasternmost South China Sea continental margin and ocean 9120(01)00069-4 basin: geophysical constraints and tectonic implications. Mar Hall R, Ali JR, Anderson CD, Baker SJ (1995a) Origin and motion Geophys Res 28:59–79. doi:10.1007/s11001-007-9014-9 history of the Philippine Sea plate. Tectonophysics 251:229–250. Lin A, Watts A, Hesselbo S (2003) Cenozoic stratigraphy and doi:10.1016/0040-1951(95)00038-0 subsidence history of the South China Sea margin in the Taiwan Hall R, Fuller M, Ali JR, Anderson CD (1995b) The Philippine Sea region. Basin Res 15:453–478. doi:10.1046/j.1365-2117.2003. Plate: magnetism and reconstructions. In: Taylor B, Natland J 00215.x (eds) Active margins and marginal basins of the Western Pacific. Lin A, Yao B, Hsu S, Liu C, Huang C (2009) Tectonic features of the AGU, Washington, DC, pp 371–403. doi:10.1029/GM088p0371 incipient arc-continental collision zone of Taiwan: implications Harding BW, Storms MA, et al (1990) Proc. ODP, Init Repts, 124E: for seismicity. Tectonophysics 479:28–42. doi:10.1016/j.tecto. College Station, TX (Ocean Drilling Program). Site 772, 75–89. 2008.11.004 doi:10.2973/odp.proc.ir.124E.1990 Liu C, Huang I, Teng L (1997) Structure features off southwestern Hayes DE, Lewis SD (1984) A geophysical study of the Manila Taiwan. Mar Geol 137:305–319. doi:10.1016/S0025-3227(96) trench, Luzon, Philippines, 1, Crustal structure, gravity, and 00093-X regional tectonic evolution. J Geophys Res 89:9171–9195. doi: Liu C, Feng L, Chuang H, Chen L (2007) Structural variations of an 10.1029/JB089iB11p09171 cccretionary wedge in the transition zone from subduction to Heck NH (1947) List of seismic sea waves. Bull Seismol Soc Am collision offshore southern Taiwan. Eos Transactions American 4:269–286 Geophysical Union 88(52), Fall Meet Suppl Abstract T41D–07 Hsiung K, Yu H (2011) Morpho-sedimentary evidence for a canyon- Ludwig WJ (1970) The Manila trench and West Luzon Trough, III, channel-trench interconnection along the Taiwan-Luzon plate Seismic-refraction measurements. Deep Sea Res 17:553–571. margins, South China Sea. Geo Mar Lett 31:215–226. doi: doi:10.1016/j.bbr.2011.03.031 10.1007/s00367-010-0226-7 Ludwig WJ, Kumar N, Houtz RE (1979) Profiler-sonobuoy measure- Hsu S, Sibuet J (2004) Continent-ocean transition of the northern ments in the South China Sea basin. J Geophys Res South China Sea and off southwestern Taiwan. Mar Geophys 84:3505–3518. doi:10.1029/JB084iB07p03505 Res 25(1–2):1–4. doi:10.1007/s11001-005-0729-1 Malavieille J, Lallemand SE, Dominguez S, Deschamps A, Lu CY, Hsu S, Yeh Y, Doo W, Tsai C (2004) New bathymetry and magnetic Liu CS, Schnurle P, Crew AS (2002) Arc-continent collision in lineations in the northernmost South China Sea and their tectonic Taiwan: new marine observations and tectonic evolution. GSA implications. Mar Geophys Res 25:29–44. doi:10.1007/s11001- Special Papers 358:189–213. doi:10.1130/0-8137-2358-2.187 005-0731-7 McIntosh K, Nakamura Y, Wang TK, Shih RC, Chen A, Liu CS Kao H, Huang GC, Liu CS (2000) Transition from oblique subduction (2005) Crustal scale seismic profiles across Taiwan and the to collision in the northern Luzon arc–Taiwan region: constraints western Philippine Sea. Tectonophysics 401:23–54. doi:10.1016/ from bathymetry and seismic observations. J Geophys Res j.bbr.2011.03.031 105:3059–3079. doi:10.1029/1999JB900357 McIntosh K, Van Avendonk HJ, Liu C, Lee C, Wang TK, Hsu S, Wu Karig DE (1983) Accreted terranes in the northern part of the F (2009) Overview and early highlights of the TAIGER project Philippine Archipelago. Tectonics 2(2):211–236. doi:10.1029/ marine, active-source seismic program. Eos Transactions, Amer- TC002i002p00211 ican Geophysical Union, 90(52), Fall Meet Suppl Abstract Kopp H, Hindle D, Klaeschen D, Oncken O, Scholl D (2009) T54B–01 Anatomy of the western Java plate interface from depth- Mirakian D, Crespi J, Byrne T, Huang C, Ouimet W, Lewis J (2012) migrated seismic images. Earth Planet Sci Lett. doi:10.1016/j. Tectonic implications of nonparallel topographic and structural epsl.2009.09.043 curvature in the higher elevations of an active collision zone, Ku C, Hsu S (2009) Crustal structure and deformation at the northern Taiwan. Lithosphere 5(1):49–66. doi:10.1130/L232.1 Manila trench between Taiwan and Luzon islands. Tectono- Moore JC, Silver EA (1987) Continental margin tectonics: submarine physics 466:229–240. doi:10.1016/j.tecto.2007.11.012 accretionary prisms. Rev Geophys 25:1305–1312. doi:10.1029/ Kuo LC, Punongbayan RS, Ramos EG (1999) GPS observation of RG025i006p01305 crustal deformation in the Taiwan-Luzon Region. Geophys Res Moore GF, Bangs NL, Taira A, Kuramoto S, Pangborn E, Tobin HJ Lett 26:923–926. doi:10.1029/1999GL900148 (2007) Three dimensional splay fault geometry and implications Lallemand SE, Schnu¨rle P, Malavieille J (1994) Coulomb theory for tsunami generation. Science 318. doi:10.1126/science.1147195 applied to accretionary and nonaccretionary wedges: possible Park J, Tsuru T, Kodaira S, Cummins P, Kaneda Y (2002) Splay fault causes for tectonic erosion and/or frontal accretion. J Geophys branching along the Nankai subduction zone. Science 297: Res 99(B6):12033–12055. doi:10.1029/94JB00124 1157–1160. doi:10.1126/science.1074111 Lee C, Wang T, Huang Y, Van Avendonk HJ, Lin J, Lallemand S, Pautot G, Rangin C (1989) Subduction of the South China Sea axial Klingelhoeher F (2009) Marine TAIGER OBS experiment and ridge below Luzon (Philippines). Earth Planet Sci Lett 92:57–69. its future prospects. Eos Transactions, American Geophysical doi:10.1016/j.bbr.2011.03.031 Union, 90(52), Fall Meet Suppl Abstract T54B–03 Pautot G, Rangin C, Briais A, Tapponnier P, Beuzart P, Lericolais G, Lewis S, Hayes DE (1983) Northward-propagating subduction along Mathieu X, Wu J, Han S, Li H, Lu Y, Zhao J (1986) Spreading eastern Luzon, Philippine Islands. In: Hayes D (ed) The Tectonic direction in the central South China Sea. Nature 321:150–154. and geologic evolution of southeast Asian Seas and Islands, 2, doi:10.1038/321150a0

123 Mar Geophys Res

Ranero CR, von Huene R (2000) Subduction erosion along the Teng LS, Lee CT, Tsai YB, Hsiao LY (2000) Slab breakoff as a Middle America convergent margin. Nature 404:748–752. doi: mechanism for flipping of subduction polarity in Taiwan. 10.1038/35008046 Geology 28:155–158. doi:10.1130/0091-7613(2000 Rangin C, Le Pichon X, Mazzotti S, Pubellier M, Chamot-Rooke N, von Huene R, Ranero C, Vannucchi P (2004) Generic model of Aurelio M, Walpersdorf A, Quebral R (1999) Plate convergence subduction erosion. Geology 32:913–916. doi:10.1130/G20563.1 measured by GPS across the Sundaland/Phillippine Sea Plate Wang K, Hu Y (2006) Accretionary prisms in subduction earthquake deformed boundary: the Phillippines and eastern Indonesia. Geophys cycles: the theory of dynamic Coulomb wedge. J Geophys Res J Int 139:296–316. doi:10.1046/j.1365-246x.1999.00969.x 111:B06410. doi:10.1029/2005JB004094 Sage F, Collot JY, Ranero CR (2006) Interplate patchiness and Wessel P, Smith WHF (1991) Free software helps map and display subduction-erosion mechanisms: evidence from depth-migrated data. EOS Trans Am Geophys Union 72:445–446. doi:10.1029/ seismic images at the central Ecuador convergent margin. 90EO00319 Geology 34:997–1000. doi:10.1130/G22790A.1 Wu FT, Lavier LL, Taiger Teams (2007) Collision tectonics of Seno T, Stein S, Gripp AE (1993) A model for the motion of the Taiwan and TAIGER Experiments. Eos Transactions, American Philippine Sea plate consistent with NUVEL-I and geologic data. Geophysical Union, 88(52), Fall Meet Suppl Abstract T51A– J Geophys Res 98:17941–17948. doi:10.1029/93JB00782 0321 Sibuet JC, Hsu SK, Le Pichon X, Le Formal J, Reed D, Moore G, Liu Wu FT, Liang WT, Lee JC, Benz H, Villasenor A (2009a) A model C (2002) East Asia plate tectonic since 15 Ma: constraints from for the termination of the Ryukyu subduction zone against Taiwan region. Tectonophysics 344:103–134. doi:10.1016/j. Taiwan: a junction of collision, subduction/separation, and bbr.2011.03.031 subduction boundaries. J Geophys Res 114:B07404. doi: Stockwell JW (1999) The CWP/SU: Seismic Un*x package. Comput 10.1029/2008JB005950 Geosci 25(4):415–419. doi:10.1016/S0098-3004(98)00145-9 Wu FT, Kuo-Chen H, Lavier LL, Unsworth MJ, Bertrand EA (2009b) Suppe J (1984) Kinematics of arc-continent collision, flipping of Toward a tectonic synthesis of Taiwan with TAIGER data. Eos subduction, and back-arc spreading near Taiwan. Mem Geol Soc Transactions, American Geophysical Union, 90(52), Fall Meet China 6:21–34 Suppl Abstract T52C–05 Taylor B, Hayes DE (1980) The tectonic evolution of the South China Zhu J, Qiu X, Zhan W, Xu H, Sun L (2005) Focal mechanism Sea Basin. In: Hayes, DE (ed) Tectonic and geologic evolution solutions and its tectonic significance in the trench of the eastern of Southeast Asian Seas and Islands, Geophys Monogr Ser 23, South China Sea. Acta Seismol Sin 18:280–289. doi:10.1007/s AGU, Washington, DC, pp 89–104. doi:10.1029/GM027 11589-005-0076-y Taylor B, Hayes DE (1983) Origin and history of the South China Zhu J, Kopp H, Flueh ER, Klaeschen D, Papenberg C, Planert L Basin. In: Hayes, DE (ed) Tectonic and geologic evolution of (2009) Crustal structure of the central Costa Rica subduction Southeast Asian Seas and Islands, Geophys Monogr Ser 27. zone: implications for basal erosion from seismic wide-angle AGU, Washington, DC, pp 23–56. doi:10.1029/GM027p0023 data. Geophys J Int 178:1112–1131. doi:10.1111/j.1365-246X. Teng LS (1990) Geotectonic evolution of late Cenozoic arc-continent 2009.04208.x collision in Taiwan. Tectonophysics 183:57–76. doi:10.1016/ j.bbr.2011.03.031

123