The Morphotectonics and Its Evolutionary Dynamics of the Central

Home , Morphotectonics, Southwest Indian Ridge

Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95 DOI: 10.1007/s13131-013-0394-1 http://www.hyxb.org.cn E-mail: [email protected]

The morphotectonics and its evolutionary dynamics of the central Southwest Indian Ridge (49° to 51°E) LIANG Yuyang1,2,3, LI Jiabiao3*, LI Shoujun3, RUAN Aiguo3, NI Jianyu3, YU Zhiteng3, ZHU Lei4 1 Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China 2 University of Chinese Academy of Sciences, Beijing 100049, China 3 Key Laboratory of Submarine Geosciences, the Second Institute of Oceanography, State Oceanic Administration, Hangzhou 310012, China 4 China Ocean Mineral Resources Research and Development Association, Beijing 100860, China

Received 12 May 2013; accepted 18 August 2013

©The Chinese Society of Oceanography and Springer-Verlag Berlin Heidelberg 2013

Abstract The morphotectonic features and their evolution of the central Southwest Indian Ridge (SWIR) are discussed on the base of the high-resolution full-coverage bathymetric data on the ridge between 49°–51°E. A comparative analysis of the topographic features of the axial and flank area indicates that the axial topography is alternated by the ridge and trough with en echelon pattern and evolved under a spatial-temporal migration especially in 49°–50.17°E. It is probably due to the undulation at the top of the mantle asthenosphere, which is propagating with the mantle flow. From 50.17° to 50.7°E, is a topographical high terrain with a crust much thicker than the global average of the oceanic crust thickness. Its origin should be independent of the spreading mechanism of ultra-slow spreading ridges. The large numbers of volcanoes in this area indicate robust magmatic activity and may be related to the Crozet hot spot according to RMBA (residual mantle Bouguer anomaly). The different geomorphological feature between the north and south flanks of the ridge indicates an asymmetric spreading, and leading to the development of the OCC (oceanic core complex). The tectonic activity of the south frank is stronger than the north and is favorable to develop the OCC. The first found active hydrothermal vent in the SWIR at 37°47′S, 49°39′E is thought to be associated with the detachment fault related to the OCC. Key words: ultra-slow spreading, multibeam bathymetry, morphotectonics, oceanic core complex, Southwest Indian Ridge Citation: Liang Yuyang, Li Jiabiao, Li Shoujun, Ruan Aiguo, Ni Jianyu, Yu Zhiteng, Zhu Lei. 2013. The morphotectonics and its evolutionary dynamics of the central Southwest Indian Ridge (49° to 51°E). Acta Oceanologica Sinica, 32(12): 87–95, doi: 10.1007/ s13131-013-0394-1

1 Introduction of hydrothermal vents are not reduced in ultra-slow spreading Ultra-slow spreading ridges have been a global hot research ridges. In 2007, the ﬁrst active high-temperature hydrothermal field because of their scientific and resource significance. A typ- ﬁeld on ultra-slow mid-ocean ridge was discovered at the loca- ical ultra-slow spreading rate usually is lower than 12 mm/a, but tion 37°47′S, 49°39′E of the Southwest Indian Ridge (SWIR). It sometimes it is defined as being no more than 20 mm/a (Dick was also found that the distribution density of hydrothermal et al., 2003). Previous seafloor spreading models of mid-ocean vents between 49°–52°E was as high as 2.5/100 km, close to that ridges were mainly based on the studies about slow-fast spread- of 36°–38°N of the MAR where there is a sufficient magma sup- ing ridges such as the EPR (East Pacific Rise) and MAR (Mid-At- ply (Tao et al., 2011). Since the polymetallic sulfide deposit as- lantic ridge), so it is necessary to consider the ultra-slow spread- sociated with hydrothermal activity is one of important poten- ing ridges in order to understand some unique phenomena tial mineral resources, ultra-slow spreading ridges also have an such as the extremely thinning oceanic crust, oblique spread- important resource significance. ing, non-transform discontinuities and distinct segmentation The SWIR, whose spreading rate is only 14 mm/a with its pattern. This is also important for improving our knowledge western part being slightly faster than the eastern, has a num- about the dynamic mechanism of global seafloor spreading and ber of unique characteristics of ultra-slow spreading. Following the interaction between different spheres of the earth. Studies a series of scientific cruises (Fig. 1), progress has been made on slow-fast spreading ridges showed that hydrothermal vents in study on the morphotectonics of this area. Dick et al. (2003) are mainly controlled by the heat sources of melt bodies and, depicted the axial topography of 9°–25°E, and combining with therefore, correlated with spreading rate. It might be inferred gravity and magnetic data and geological sampling illustrated that the hydrothermal activity would decrease on an ultra-slow the structural features of the oblique-orthogonal spreading, spreading ridge because of its relatively limited magmatic ac- the magmatic accretion and amagmatic accretion. Cannat et tivity. However, recent investigations revealed that the activities al. (2006) analyzed the topography of both flanks of 61°–66°E

Foundation item: The National Natural Science Foundation of China under contract No. 91028006; the Dayang 115 under contract No. DYXM- 115-02-3-01. * Corresponding author, E-mail: [email protected] 88 LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95

40° 45° 50° 55° 60° 65° 70° 75° E

a 25° S

30°

35° 36° 40° 44° 48° 52° E 34° S b 38° 40° 42°

46° 45°

50° RMBA −120−80 −40 04080 mGal

−4 000 −3 000 −2 000 −1 000 0

Depth/m

Fig.1. Bathymetry map (ETOPO2) of SWIR and the study area, the shaded relief image illuminating from N45°E at an angle of 45° (a) and regional RMBA of SWIR (b) (Zhang et al., 2013). and identified three distinct types of seafloor: (1) volcanic tonics and discuss their structural characteristics and dynamic seafloor, with bounding horsts and grabens in a similar man- mechanism. ner to the abyssal hills described at faster spreading seafloor; (2) smooth seafloor, probably associated with amagmatic ac- 2 Data and methods cretion; and (3) corrugated seafloor, probably associated with The bathymetric data mainly come from the full-coverage detachment fault-caused OCCs. Sloan et al. (2012) focused on multibeam survey during the leg for 3D OBS seismic explora- the flank topography of 54°–67°E and discussed the effect of tion at the SWIR (49°–51°E) of cruise DY115-21 in 2010, with the spreading rate, sub-axis mantle temperature, and volcanic or integrated multibeam data collected from other cruises of this non-volcanic crust on topographic parameters such as root area. The data was collected by Simrad EM120 multibeam echo mean square height, lineament azimuth and characteristic sounder system on board the R/V Dayang I. It was mounted width of abyssal hills. Based on analysis of the on-axis deep tow under the hull with a 12 kHz transmitting/receiving transducer side scan sonar data and off-axis multibeam bathymetric data and contains 191 beams arrayed over an arc of 150°and each of both 58°30′–60°12′E and 63°23′–65°45′E, Mendel et al. (2003) beam width up to 1°. The maximum depth of the survey was thought that there was a magmato-tectonic cycle of the axial 11 000 m. Multibeam processing systems were set up on the evolution. vessel to carry out real-time processing for the sound veloc- Ultra-slow spreading ridge is an important frontier in the ity, heave, roll, pitch and heading offered by the velocity pro- deep sea research, but so far publications are rare. The features files and motion sensor unit (U-Phins & Octans). The drafts at of magmatic and amagmatic segments at the central rift valley start and end of the survey were recorded, and the linear draft of ultra-slow spreading ridges, the tectonic role the detachment correction for depth was made in post-processing. A full-cov- fault played in the spreading process, and the distinct tectonic erage multibeam sounding bathymetric survey was carried out evolution and dynamic mechanism of ultra-slow spreading over the region of 49°–51°E of the SWIR, with a total measuring ridges are some of the un-answered questions. length of 5 930 km. According to the principles of topographi- With little sediment cover at the spreading center of the cal continuous variation and adjacent swath comparison (Li, SWIR, the seafloor topography is basically a direct reflection of 1999), all the data were examined and edited to remove errone- regional tectonics and deep response. Using the high-resolu- ous points with the CARISHIPS 6.0 software. The depth data was tion multibeam bathymetric data of the SWIR (49°–51°E) col- exported as an ASCII file with the HIPS, and was subsequently lected during voyages over several years, in combination with gridded at 50 m spacing using universal Kriging, shaded with relevant research results from regional gravity-magnetic data artificial illumination, and displayed on a workstation using and 3D OBS detection, we analyze the segmental morphotec- Global Mapper V 7.01 (Fig. 2). LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95 89

49.0° 49.2° 49.4° 49.6° 49.8° 50.0° 50.2° 50.4° 50.6° 50.8° E

37.4° S

37.6°

37.8°

38.0° −4 000 −3 000 −2 000 −1 000 Depth/m

Fig.2. Multibeam shaded relief image of the study area. The shaded relief image illuminating from the N45°E at an angle of 60° and multibeam survey lines being mainly N-S and E-W trending. For location see Fig. 1.

The morphotectonics is an effective tool to study the mor- combining with gravity, magnetic and seismic data. Lonrenço phology of the earth’s surface and its tectonism. It was first et al. (1998) introduced the morphotectonic method into the formed, developed and matured for the land environment be- geodynamics study of Azores Triple Junction among the Ameri- cause there it is relatively easy to obtain the morphology infor- can, European and African plates. Multibeam bathymetric data mation. On the other hand, the oceanic bathymetry technol- of Cadiz Bay (boundary of the Eurasian and African plates), ogy went through a long way to be developed from the early together with seismic data, were used to deduce its dynami- rope-hammer measurement, classic echo sounding method, cal evolution since Late Cretaceous, and it is believed that its side-scan sonar and single/double frequency single-beam echo morphotectonic feature was mainly built by the reactivation of sounding to multi-beam bathymetric echo sounding. The mor- deep faults (Rosas et al., 2009; Terrinha et al., 2009). Carbottle photectonics was also gradually introduced into the field of ma- and Macdonald (1992) proposed the formative mechanism of rine geology due to this bathymetric technology development. OSC (Overlapping spreading center) based on their study at the Because of the development of multi-beam bathymetry tech- axial ridge area of the EPR. According to the topographic fea- nology, a large number of high-resolution seafloor topography tures of the Atlantic ridge, discussions about its segmentation, can be easily retrieved. By now, the morphotectonics has played structural characteristics and spreading mode were carried out an important role in the tectonic study of hydrothermal areas, (Purdy et al.,1990; Grindlay et al.,1992; Sempéré et al.,1993). subduction zones, marginal basins and mid-ocean ridges. Using The morphotectonic method is effective to study tectonic near-bottom high-resolution bathymetry data and digital pho- features of young seafloor. To discuss the seafloor spreading tographic imagery, Humphris and Kleinrock (1996) presented pattern and tectonic dynamic mode of SWIR (49°–51°E), we the fine structure of the hydrothermal vents of TAG in Atlantic, made the tectonic interpretation of this full coverage multi- and proposed a geological model of hydrothermal vents in this beam bathymetric data set in the present contribution. Based area. Li et al. (2004) analyzed the multibeam data of the accre- on a comparative analysis of the axial and off-axial morphology tionary wedge and diapiring structure of the subduction zone in the study area, we discussed the topographical spatial-tem- at the Manila trench, and discussed its structural characteristics poral evolution due to seafloor spreading process, compared and dynamic mechanism by combining with other geological the geomorphic differences of both eastern and western areas and geophysical data. Kukowski et al. (2001) and Grando and along the ridge, and attempted to identify the different factors McClay (2007) discussed the morphotectonics and structural affecting the topography. styles of the Makran accretionary wedge offshore Pakistan and Iranre respectively based on multi-beam data. Papanikolaou et 3 Morphotectonic analysis al. (2002) used high-resolution multibeam bathymetric data of The SWIR, bordering the African and Antarctic plates, starts the North Aegean Sea to calculate tectonic deformation accu- from the northeast at the Rodriguez Triple Junction, and ends to mulation during the last 4–5 Ma. On the basis of high precision the southwest at the Bouvet Triple Junction, with a total length and full coverage multibeam bathymetric data, Wu et al. (2003) of about 8 000 km. It is generally oblique spreading with an carried out quantitative analysis of the typical morphotectonics angle approximately 60° to the spreading direction and the full of Okinawa Trough basin, and interpreted its geological process spreading rate 14 mm/a (Dick et al., 2003). The study area is cor- and trough formation mechanism (Wu et al., 2003). Based on responding to Segments 27 and 28 (Cannat et al.,1999), located the multibeam bathymetric data of South China Sea, Li et al. at the central SWIR, between the Indomed (46.0°E) and Gal- (2002, 2011, 2012) analyzed the morphotectonics of the east- lieni fault zones (52.2°E) (Fig. 1). It is a product of Rodriguez’s ern and southwestern sub-basin, and established the Cenozoic westward propagation (Patriat and Segoufin,1988). Except for evolution model of the seafloor spreading in South China Sea by the area of 50.17°–50.7°E, the axial region between the Indomed 90 LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95

and Gallieni fault zones is occupied by the rift valley with an 500 m from the bottom, extending E-W for about 15 km with an NEE trending. The rift valley bents slightly, but no transform average width of about 3 km. Sub-segments B and D have well faults were observed. The terrain between 50.17°E and 50.7°E developed E-W trending linear structures with stepped normal has the highest topography of the entire segment, and the rift faults at both walls. Volcanoes can be found on the AVR of the valley is absent at its axial region. Inversion of the OBS seismic sub-segment D (Fig. 5a), indicating robust magmatism. data, which was collected by the same cruise, suggests that this The axial part of the eastern area is nearly E-W trending, oceanic crust thickness is up to 8–10 km and can be tracked off- without rift valley, and slightly higher than both its flanks. The axis at least for 4 Ma. Such a crust thickness is much thicker than volcanic center lying between 50.4°–50.5°E is the highest terrain the global average of oceanic crust, about 6 km (Chen, 1992). At with a depth of less than 1 600 m. The water depth increases 37°47′S, 49°39′E, an active high-temperature hydrothermal ﬁeld from this center towards both sides and has a gradual transition on the ultraslow mid-ocean ridge was found during the cruise towards the eastern Trough E of the western area. DY115-17 in 2007 (Tao et al., 2011). 3.2 Flank topography 3.1 Axial topography The fractures and structural trends in the flanks present dif- The study area can be divided by 50.17°E into two quite dis- ferent characteristics and can be divided into two areas, similar tinct parts in terms of the axial topography. The western part is to the situation of the axial parts. The eastern area (50.17°–50.7°E) occupied by an NEE trending rift valley where there are some generally extends E-W, with fractures that usually have small axial volcanic ridges (AVRs). In contrast, the eastern part has the fault planes and a single style. The western area (49°–50.17°E) highest terrain without rift valleys, occupied by dense volcanic is relatively complex: fractures (including large main faults and activities especially at about 50.4°–50.5°E. secondary faults) are present, with different trends (Figs 2 and 3). The western part, essentially a rift valley, can be divided into In the western area, two groups of fractures, nearly E-W and five sub-segments: two ridges with AVRs at the center of the NEE trending, both facing to the axis, are present. The large valley, and three troughs without AVRs. The ridges and troughs fractures with NEE and E-W strikes are mostly high-angle nor- appear alternatively along the axis and are arranged in en ech- mal faults. Those at the southern flank dip slightly gentler than elon pattern (Fig. 3, hereinafter referred to as A-E). The adjacent the northern ones; an individual fault may run along its trend ridges and troughs are slightly overlapped at their ends, and the for about 10 km, with fault offsets between 300 m and 500 m. axes are offset left-stepped at that position without a transform These NEE and E-W trending large fractures are inclined, head- fault. The deepest point of the troughs reached 4 000 m while tail connected to form several longer fractures of 20–40 km in the shallowest point of the ridges was 2 800 m deep. Sub-seg- length. These fractures have a polyline shape, similar to the rift ments A, C and E are troughs with ellipse long axis deviated valley boundary morphology, which probably represent earlier from E to NEE slightly; the valley width here is about 15 km, and rift boundaries. Thus, this area can be subdivided into several stepped normal faults can be observed on both walls. The N-S tectonic phases along these fractures. Likewise, the small frac- topographical profiles are roughly “U” shaped (Fig. 4). The bot- tures, E-W trending, with fault offsets of less than 300 m and toms of Troughs A and C are smooth, and linear structures were fault plane width no more than 10 km, are well developed at the not observed. The Trough E is shallower than both A and C, and northern flank, distributed in clusters. Generally, in the western its bottom relatively rough; nearly E-W trending linear struc- area, faults are better developed at the southern flank, and the ture is present and volcanoes can be observed in its eastern escarps here are steeper and larger, than those of the northern part (Fig. 5b). Sub-segment B and D have much shallower water flank. depth with AVRs at the valley center; the AVRs uplift more than In contrast with the western area, large escarpments are

49.2° 49.4° 49.6° 49.8° 50.0° 50.2° 50.4° 50.6° 50.8° E

37.4° S

37.6°

E D C 37.8° B A

38.0° 8 4 9 1 5 10 2 6 11 3 7 12

Fig.3. Structural interpretation of the study area. 1 represents uplift zones, 2 OCC, 3 rift valley, 4 AVRs, 5 supposed AVRs, 6 troughs, 7 supposed troughs, 8 NTDs, 9 escarpments, 10 conical volcanoes, 11 crater volcanoes, and 12 flat-topped volcanoes. LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95 91

SNSN SN 1

1 1 3

3 5 5

5 7

9 9 7

13 11

0 510 15 km 0 51015 20 km 0 510 15 km Trough A Trough C Trough E

Fig.4. Topographical profiles of Troughs A, C and E. For location see Fig. 2.

a d

−2 500

−3 000

−3 500

Depth/m

−500

b −1 000 −2 000 e

−1 500 −2 500

−3 000 −2 000

−3 500 −2 500

Depth/m −3 000

Depth/m f c

−2 000

−2 500

−3 000

Depth/m

Fig. 5. Volcanoes in the study area. a. Ridge D, b. Trough E, c. Zone III, d. the uplift off-axis north about 15 km at the eastern area, e. the axial part at the eastern area, and f. the uplift off-axis south about 15 km at the eastern area. For location see Fig. 2. 92 LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95

49.3° 49.4° 49.5° 49.6° 49.7° 49.8° 49.9° E

37.5° S

37.6°

37.7°

−4 000 −3 000 −2 000 −1 000

Depth/m

Fig.6. Topographical partitioning of the north flank at the western area. scarce in the eastern area, only present at the uplifted zones width is 3–7 km with decreasing trend towards the north, cover- about 15 km off-axis. Here, it is nearly E-W trending high-angle ing an area of about 40 km2. small faults that are well developed, with offsets of only about This OCC possesses general morphological features of OCCs 100 m and fault plane widths of about 5 km. E-W trending struc- with a complete shape. Its top is about 200 m higher than the tures are dominant in this area. base, and the surface is smooth compared with surrounding ter- As shown in Fig. 6, the northern flank, 49.3°–50°E, can be di- rain, dipping to the north with a low angle. Several high-angle vided into E-W arrayed zones according to topographical char- fractures dipping to the north may have been developed within acteristics: the Zones I and III have a shallow water depth with the OCC (Figs 7 and 8). The linear ridge on its southern side has well-developed E-W linear structures, while the Zone II has a collapsed somewhat, making the initial ruptured surface less relatively large water depth and smooth topography without continuous, and the blocks on its western side may have been the presence of linear structures except for several large frac- subjected to the same detachment fault (Figs 7 and 8). tures. Several volcanoes exist in the Zone III (Fig. 5c). Further- more, the three zones are all NNE-SSW trending. In contrast, such a geomorphological zonation pattern is not present at the 4 Discussion south flank. 4.1 Ridge-trough topography and its evolutionary mechanism 3.3 Volcanic eruption area at 50.5°E The axial topography, which alternated with ridge and The terrain centered by 50.5°E has higher topography than trough and arranged in en echelon pattern, is associated with both its east and west sides, and axial valley is absent. Addition- ultra-slow oblique spreading ridges exclusively. Such topogra- ally, a large number of volcanoes can be identified in this terrain. In a range of approximately 3 000 km2, around 80 volcanoes are 49.85° 49.8° 49.75° 49.7° 49.65° 49.6° E present, with a distribution density of up to 2.7/100 km2. According to the shape, these volcanoes can be divided into three categories: conical, crater and flat-topped volcanoes. In 37.95° addition, there are a few poorly defined volcanoes, which may S result from the collapse of crater volcanoes. Among the 79 volcanoes identified, there are 43 conical, 21 crater and 15 flat- 37.9° topped ones (Fig. 3). The size of these volcanoes is almost the same, with a diameter about 1 km and height 100–200 m, and some conical volcanoes could be higher than 200 m. The dis- 37.85° tribution of these three categories of volcanoes does not have a distinguishable pattern, except that the flat-topped ones tend to appear in the uplift zones. However, collectively they tend to 37.8° be distributed in the E-W trending zones, concentrating in the axial part and on the uplift zones 15 km off-axis (Figs 3, 5d, 5e and 5f). 37.75°

3.4 OCC terrain At the southern flank of the western area, an S-N trend- −4 000 −3 000 −2 000 −1 000 ing terrain with a typical OCC morphology was discovered on just the southern side of the sub-segment D (Figs 3 and 7). The Depth/m northern edge of its corrugated surface is 10 km away from the axial ridge, extending towards the south for 10 km, and its E-W Fig.7. Structural interpretation of the OCC area. LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95 93

S N OCC detachment surface spreading center

3 000 m

01020 30 40 50 km a. Profile of L1

S detachment surface N vent spreading center

3 000 m

010203040 50 km b. Profile of L2

Fig.8. Profile across the OCC (a) and profile across the hydrothermal vent at 37°47′S, 49°39′E (b). For location see Fig. 2. phy was also observed at Mohns Ridge (obliquity 30°–40°) and topography. The seafloor spreading toward N20°E with a half- Reykjanes Ridge (obliquity 30°–40°) which both belong to Arc- spreading rate of 7 mm/a, the lift valley trending N75°E, and the tic ultra-slow spreading ridge system (Van Wijk and Blackman, Zones I, II and III trending N30°E, results in an estimate of an 2007). However, no similar topography has been reported at along-axis topography propagating rate of 1.72 mm/a (Fig. 6). faster spreading ridges so far. The along-axis propagation of the alternating ridge-trough The ridge-trough topography is a manifestation of locally topography will lead to topography changing over time for a focused magmatism which tends to occur at ultra-slow spread- given axial part, and this is somewhat similar to the theory of ing ridges. Ridges are correlated with focused magmatism seg- Mendel et al. (2003) that there was a magmato-tectonic evolu- ments with shallow depth, while troughs are corresponding to tionary cyclicity at the axial part. To explain the focused mag- poor magmatism segments with large depth (Sauter et al., 2001; matism at ultra-slow spreading ridges, some geological models Cannat et al., 2003; Michael et al., 2003; Sauter et al., 2004). have beeen established (Sauter et al., 2001; Cannat et al., 2003; Ridges and troughs firstly formed at the axial part, and subse- Sauter et al., 2004). However, these models are unable to explain quently migrated to both flanks as seafloor spreading. As a re- the ridge-trough alternation and migration at the axial part over sult, Cannat et al. (2006) thought that the smooth topography time, or the magmato-tectonic cyclicity. To explain the spatial- of both flanks at the east section of the SWIR (approximately temporal variation of magmatism at the SWIR, we proposed a 61°–66°E) may represent poor magma seafloor; Mendel et al. model as shown in Fig. 9. There is a kind of small-scale mantle (2003) believed that the depressions, observed at the off-axis flow waves existing at the surface of asthenosphere mantle, and area of two sections (58°30′–60°12′E and 63°23′–65°45′E) of the these waves lead to the focused magmatism. The wave crests SWIR, corresponded to the axial NTDs. In addition, fractures will correspond to the position of axial ridges, while the wave associated with the ridge and trough topography were also troughs correspond to axial troughs. The movement of mantle developed in different styles. For the ridge topography, the rift boundary faults generally strike perpendicular to the spreading direction, and intra-rift faults are well developed with strikes oceanic crust also perpendicular to the spreading direction; however, for the trough topography, the rift boundary faults' strikes are usually lithospheric mantle slightly oblique to the spreading direction, with few intra-rift faults (Dick et al., 2003). melt mantle As outlined above, at the northern flank, the Zones I and III mantle flow are different from the Zone II in terms of water depth, linear structure development and volcano occurrence. In combi- Fig.9. Model for the ridge-trough topography. nation with the topographical features of the axial part of the western area, it may be the case that the Zones I and III were flow leads to the wave propagation, causing the axial topogra- originated from the accretion of the ridge sub-segments at the phy change. axial part, whilst the Zone II was originated from that of the ridge sub-segment. 4.2 High topography at 50.5°E Since the Zones I, II and III are all NNE-SSW trending (Figs The terrain of 50.17°–50.7°E, where rift valleys are absent, 3 and 6), we infer that the three zones were originated from the has the almost highest topography over the entire segment. It accretion of correspondent axial part. Such NNE-SSW trending has been proposed that magma supply at the SWIR is limited, is inconsistent with the spreading direction and probably sug- leading to a very thin crust, and such high topography is a re- gests that the ridge-trough topography alternating at the axial sult of mantle uplifting due to regional isostatic compensation part is propagating westward. Consequently, the NNE-SSW (Zhou and Dick, 2013). However, morphotectonic evidence trending of these zones is a vector sum of both seafloor spread- does not appear to support this argument. First, the large num- ing and along-axis propagation of the alternated ridge-trough bers of volcanoes identified suggest abundant magma supply 94 LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95

in this area. Second, at the western area, the Trough E is appar- 2010). Smith (2013) proposed two seafloor spreading modes for ently shallower than both Troughs A and C, with volcanoes even upper mantle exhumation in the context of slow and ultra-slow arising at its eastern bottom, indicating that it has been signifi- spreading ridges, respectively: (1) mode of asymmetric spread- cantly affected by magmatic activity of this area. Finally, gravity ing at slow spreading ridges, in which magmatic accretion model and preliminary inversion of OBS data both show that only happens on one side of the axis while detachment faults this area has an 8–10 km thick oceanic crust, which can be at are developed on the other side, then OCCs are formed due to least traced off-axis 4 Ma. Therefore, magma supply should be continued extension; and (2) mode of amagmatic spreading at sufficient here to create the very thick oceanic crust and high ultra-slow spreading ridges, in which detachment faults are de- relief. Furthermore, because of the robust magmatic eruptions, veloped at the axial part one after another almost without mag- the alternating ridge-trough axial topography, existing in the matism. western area where magma supply is limited, abruptly stopped The spreading rate of the study area is only about 14 mm/a, here. but based on structural analysis of the OCC situated at the Combined with the global bathymetric data, it can be found south of sub-segment D it is found that this area has been prob- that the axial topography, with alternating ridges and troughs, ably developed under a similar mechanism to the asymmetric arranged in en echelon pattern, not only exists within the study spreading mode (Fig. 8a). The structural style is asymmetric at area, but is also widely developed at the entire segment between two flanks. The northern side has only high-angle normal faults the Indomed and Gallieni transform faults. Assuming that the while detachment fault was developed on the southern side. volcanic eruption area originally had been a trough morphol- Further, the OCC was generated under continued extension ogy (Fig. 3), such alternating ridge-trough axial morphotecton- with several normal faults. ics would be formed throughout the entire segment. According The active hydrothermal vent at 37°47′S, 49°39′E is prob- to the analysis outlined above, the main factor of large-scale ably mainly ascribed to the detachment fault shown in Fig. 8b. volcanic eruption and overall shallow water depth at 50.5°E As a low-angle and deep fault, the detachment fault is able to should be independent of the spreading mechanism of ultra- provide heat and water circulation channels needed by the for- slow spreading ridges where magma supply is limited, and may mation of hydrothermal vent. As shown in Figs 7 and 8b, at the be associated with hot spots. Based on large-scale RMBA, Zhang detachment fault’s hanging wall, normal faults tend to be devel- et al. (2013) found that it existed Marion-Del Cano-Crozet man- oped due to tectonic extension; because the oceanic crust here tle plume hot spots in the southwest of the study area, and in is thin, these normal faults can easily connect to the detach- particular it formed a hot spots—mid-ocean ridge interaction ment fault beneath the seafloor, becoming a water circulation between the Crozet hotspot and the study area, leading to large- channel. Under such an asymmetric spreading mode, the area scale volcanic eruptions at 50.5°E (Fig. 1b). between the axial ridge and OCC’s termination is a favorable Within the magmatic activity area at 50.5°E, topography is position for the development of hydrothermal vents. It can be high and volcanoes are densely distributed at the central axial inferred from the vent locations that the area affected by this part and the uplifted zones of both flanks (Figs 2, 5d, 5e and detachment fault is far beyond the OCC. 5f), while in the terrain between the central axial part and the As outlined above, in the western area, the regular geomor- uplift zones the topography is relatively low and volcanoes are phological pattern is not so clear at the southern flank, as com- sparsely distributed. This observation indicats that volcanic ac- pared with the northern flank. Faults are better developed at tivity of this area varies in intensity over time. The uplifted zone the southern flank, and are gentler and wider than those at the about 15 km off-axis and its volcanoes should be the product northern flank. These phenomena also suggest that the spread- of the last volcanic actively phase at the axial part, and it was ing of the western area is asymmetric, and tectonism played a subsequently separated and migrated to both sides with sea- more important role at the southern flank than at the north. The floor spreading. At about 50.5°E, the central axial topography is mode shown in Fig. 8b may be common at the southern flank. slightly higher than both its southern and northern sides, and a volcanic center is more than 3 km in diameter (although much 5 Conclusions smaller than the uplifts off-axis), which probably represent the The axial topography, with alternating ridges and troughs, start of a new volcanic actively phase. arranged in en echelon pattern, probably exists at ultra-slow Compared with the western area, because of the sufficient spreading ridges exclusively. The ridges and troughs represent magma supply, fractures are less well-developed in this area, magmatism dominated sections and tectonism dominated sec- and large fractures are absent. The sufficient magma supply tions, respectively, and correspond to magmatic and tectonic made the magmatism predominate in the process of seafloor spreading, respectively. Such a ridge-trough topography is sub- spreading, while tectonic extension became relatively weak. jected to spatial-temporal change, probably due to the undulation at the top of the mantle asthenosphere, whose wave shape 4.3 OCC and asymmetric spreading propagates with the mantle flow. The along-axis propagation Continued extension on a single fault will lead to significant of the ridge-trough alternation topography, together with the flexural rotation of its footwall. If extension continues for more off-axis migration of the axial topography, generates a zona- than about 5 km, continued rotation means that the exhumed tion pattern, which is more clearly defined at the northern flank fault surface becomes nearly horizontal from its initial dip of than the south. The zonation has a trending direction deviated about 60°, and the seafloor domes as a result of regional iso- from the spreading direction, suggesting that the ridge-trough static compensation (Smith et al., 2012). OCCs are developed topography is propagating westward. by such a mechanism, and it exhumes the lower crust or upper There is a very thick oceanic crust existing at the eastern mantle. Studies have shown that OCC development is related area, and its origin should be independent of the spreading to the axial magma supply (Buck et al., 2005; Schouten et al., mechanism of ultra-slow spreading ridge system. According to LIANG Yuyang et al. Acta Oceanol. Sin., 2013, Vol. 32, No. 12, P. 87–95 95

RMBA analysis, it may be associated with hot spots. Volcanic Lourenço N, Miranda J M, Luis J F, et al. 1998. Morpho-tectonic analysis activity of this area varies in intensity, and so as the hot spots— of the Azores Volcanic Plateau from a new bathymetric compila- mid-ocean ridge interaction. The sufficient magma supply tion of the area. Marine Geophysical Researches, 20(3): 141–156 Mendel V, Sauter D, Rommevaux Jestin C, Patriat P, et al. 2003. Mag- made magmatism virtually predominate in process of seafloor mato-tectonic cyclicity at the ultra-slow spreading Southwest spreading in this area. Indian Ridge: Evidence from variations of axial volcanic ridge Asymmetric spreading exists in the study area, and com- morphology and abyssal hills pattern. Geochemistry Geophysics paratively the south has more intensive tectonism, leading to Geosystems, 4: doi: 10.1029/2002gc000417 the difference in morphological characteristics between the Michael P J, Langmuir C H, Dick H J B, et al. 2003. Magmatic and amag- southern and northern flanks. The OCC in the study area was matic seafloor generation at the ultraslow-spreading Gakkel developed under an asymmetric spreading mode. The active ridge, Arctic Ocean. Nature, 423: 956–962 hydrothermal vent at 37°47′S, 49°39′E should be associated with Papanikolaou D, Alexandri M, Nomikou P, et al. 2002. Morphotectonic structure of the western part of the North Aegean Basin based on the detachment fault, and the terrain between the axial ridge swath bathymetry. Marine Geology, (190): 465–492 and OCC’s termination is a favorable location for the develop- Patriat P, Segoufin J. 1988. Reconstruction of the Central Indian Ocean. ment of hydrothermal vents. Tectonophysics, 155(1–4): 211–234 Purdy G M, Sempere J C, Schouten H, et al. 1990. Bathymetry of the References mid-atlantic ridge, 24°–31°N: A map series. Marine Geophysical Researches, 12(4): 247–252 Buck W Roger, Lavier Luc L, Poliakov Alexei NB. 2005. Modes of faulting at mid-ocean ridges. Nature, 434(7034): 719–723 Rosas F M, Duarte J C, Terrinha P, et al. 2009. Morphotectonic charac- Cannat M, Rommevaux Jestin C, Sauter D, et al. 1999. Formation of the terization of major bathymetric lineaments in Gulf of Cadiz (Af- axial relief at the very slow spreading Southwest Indian Ridge rica–Iberia plate boundary): Insights from analogue modelling (49°to 69°E). Journal of Geophysical Research, 104(B10): 22825– experiments. Marine Geology, 261(1–4): 33–47 22843 Sauter D, Mendel V, Rommevaux Jestin C, et al. 2004. Focused magma- Cannat M, Rommevaux Jestin C, Fujimoto H. 2003. Melt supply varia- tism versus amagmatic spreading along the ultra-slow spread- tions to a magma-poor ultra-slow spreading ridge (Southwest ing Southwest Indian Ridge: Evidence from TOBI side scan Indian Ridge 61°to 69°E). Geochemistry Geophysics Geosystems, sonar imagery. Geochemistry Geophysics Geosystems, 5: doi: 4: doi: 10.1029/2002gc000480 10.1029/2004gc000738 Cannat M, Sauter D, Mendel V, et al. 2006. Modes of seafloor genera- Sauter D, Patriat P, Rommevaux Jestin C, et al. 2001. the Southwest tion at a melt-poor ultraslow-spreading ridge. Geology, 34(7): Indian Ridge between 49°15′E and 57°E: foused accretion and 605–608 magma redistribution. Earth and Planetary Science Letters, 192: Carbotte S, Macdonald K. 1992. East Pacific Rise 8°–10°30′N: Evolu- 303–317 tion of ridge segments and discontinuities from SeaMARC II and Schouten H, Smith D K, Cann J R, et al. 2010. Tectonic versus magmatic three-dimensional magnetic studies. Journal of Geophysical Re- extension in the presence of core complexes at slow-spreading search: Solid Earth, 97(B5): 6959–6982 ridges from a visualization of faulted seafloor topography. Geol- Chen Yongshun John. 1992. Oceanic crustal thickness versus spreading ogy, 38(7): 615–618 rate. Geophysical Research Letters, 19(8): 753–756 Sempéré J C, Lin Jian, Brown H, et al. 1993. Segmentation and morpho- Dick H J B, Lin Jian, Schouten H. 2003. An ultraslow-spreading class of tectonic variations along a slow-spreading center: The Mid-At- ocean ridge. Nature, 426: 405–412 lantic Ridge (24°00′N–30°40′N). Marine Geophysical Researches, Grando G, McClay K. 2007. Morphotectonics domains and structural 15(3): 153–200 styles in the Makran accretionary prism, offshore Iran. Sedimen- Sloan H, Sauter D, Goff J A, et al. 2012. Abyssal hill characterization at tary Geology, 196(1–4): 157–179 the ultraslow spreading Southwest Indian Ridge. Geochemistry Grindlay N R, Fox P J, Vogt P R. 1992. Morphology and tectonics of the Geophysics Geosystems, 13: doi: 10.1029/2011gc003850 Mid-Atlantic Ridge (25°–27°30′S) from Sea Beam and magnetic Smith D. 2013. Mantle spread across the sea floor. Nature Geoscience, data. Journal of Geophysical Research: Solid Earth, 97(B5): 6983– 6(4): 247–248 7010 Smith D, Escartín J, Schouten H, et al. 2012. Active long-lived Humphris S E, Kleinrock M C. 1996. Detailed morphology of the TAG faults emerging along slow-spreading Mid-Ocean Ridges. Active Hydrothermal Mound: Insights into its formation and Oceanography, 25(1): 94–99 growth. Geophysical Research Letters, 23(23): 3443–3446 Tao Chunhui, Lin Jian, Guo Shiqin, et al. 2011. First active hydrothermal Kukowski N, Schillhorn T, Huhn K, et al. 2001. Morphotectonics and vents on an ultraslow-spreading center: Southwest Indian Ridge. mechanics of the central Makran accretionary wedge off Paki- Geology, 40(1): 47–50 stan. Marine Geology, (173): 1–19 Terrinha P, Matias L, Vicente J, et al. 2009. Morphotectonics and strain Li Jiabiao. 1999. Multibeam Sounding Princaples Survey Technologies partitioning at the Iberia-Africa plate boundary from multibeam and Data Processing Methods. Beijing: China Ocean Press and seismic reflection data. Marine Geology, 267(3–4): 156–174 Li Jiabiao, Ding Weiwei, Gao Jinyao. 2011. Cenozoic evolution model Van Wijk J W, Blackman D K. 2007. Development of en echelon of the sea-floor spreading in South China Sea: new constraints magmatic segments along oblique spreading ridges. Geology, from high resolution geophysical data. Chinese Journal Geo- 35(7): 599–602 physics, 54(12): 3004–3015 Wu Ziyin, Jin Xianglong, Li Jiabiao, et al. 2003. The typical tectonic Li Jiabiao, Ding Weiwei, Wu Ziyin. 2012. The propagation of seafloor morphologies of Okiniwa Trough and their geology meaning (in spreading in the southwestern subbasin, South China Sea. Chi- Chinese). Submarine Geosciences Symposium. Beijing: China nese Science Bulletin, 57(24): 3182–3191 Ocean Press: 113–123 Li Jiabiao, Jin Xianglong, Gao Jinyao. 2002. Morpho-tectonic study on Zhang Tao, Lin Jian, Gao Jinyao. 2013. Mamatism and tectonic process- late-stage spreading of the Eastern Subbasin of South China Sea. es in the area of hydrothermal vent on Southwest Indian Ridge. Science in China Series D: Earth Sciences, 45(11): 978–989 Science in China Series D: Earth Sciences, doi:10.1007/s11430- Li JiaBiao, Jin Xianglong, Ruan Aiguo. 2004. Indentation tectonics in 013-4630-5 the accretionary wedge of middle Manila Trench. Chinese Sci- Zhou Huaiyang, Dick H J. 2013. Thin crust as evidence for depleted ence Bulletin, 49(12): 1279–1288 mantle supporting the Marion Rise. Nature, 494(7436): 195–200