Keywords

1 Introduction

Melt generation and the resulting volcanic activity associated with a seafloor spreading axis generally depend on the spreading rate (e.g., Small 1998; Macdonald 1998). Along back-arc ridges, the correlation between spreading rate and volcanism is more complex owing to the influence of plate subduction processes (Martinez and Taylor 2002; Taylor and Martinez 2003). Taylor and Martinez (2003) systematically analyzed global back-arc basin basalt and proposed that melt generation along back-arc ridges can vary markedly with distance from the volcanic front, mainly due to subduction-induced compositional changes in the mantle rather than the seafloor spreading rate itself. The spreading center of the Southern Mariana Trough shows an axial high morphology similar to fast-spreading ridges, despite its slow to intermediate spreading rate. The back-arc spreading center is very close to the volcanic arc in this region, although the volcanic front is not clear south of 12°30′N. Previous studies have shown that the magmatic budget increases along the spreading axes in the Southern Mariana Trough because the spreading axis possibly captures heat and melt supply from the volcanic front (e.g., Martinez et al. 2000; Fryer 1996; Taylor and Martinez 2003).

Although previous studies have shown a high magma budget in the Southern Mariana Trough, the style of volcanism (e.g., the distribution of the volcanic product) and lava morphology and texture have not yet been studied. During the TAIGA project, we conducted fine-scale acoustic observations using the AUV Urashima and visual observations using the submersible Shinkai 6500 in the hydrothermal areas of the Southern Mariana Trough (Seama et al., Chap. 17). One of the target areas is the axial zone of back-arc spreading, where it is possible to observe the most recent volcanic activity and active hydrothermal vents. In this chapter, we show the micro-bathymetry and side-scan sonar imagery collected by the AUV together with photographs taken during the submersible dives. We describe the fine-scale volcanic and tectonic features in the area. This survey gives us the first sub-meter scale observations with ground references along the Southern Mariana Trough, and enables us to obtain a better understanding of volcanism at back-arc spreading centers under the considerable influence of arc volcanism.

2 Geological Background

The Mariana Trough is a back-arc basin located behind the Mariana Trench, where the Pacific Plate subducts under the Philippine Sea Plate. The current rate of spreading is approximately 40 mm/year near Guam (13°24′N) (Kato et al. 2003; Martinez et al. 2000) and, based on this spreading rate, the ridge has been categorized as a slow to intermediate spreading ridge.

The spreading axis lies in the eastern part of the basin, indicating asymmetric seafloor accretion (Yamazaki et al. 2003; Deschamps and Fujiwara 2003; Deschamps et al. 2005; Asada et al. 2007). Abyssal hills have not been clearly observed in the eastern off-axis area owing to thick sedimentary coverage and/or overprinting of later arc volcanism (Fryer 1996; Martinez et al. 2000). North of 14°N, the spreading center of the Mariana Trough is morphologically similar to slow-spreading mid-ocean ridges, having a deep crustal graben flanked by a zone of abyssal hills (Seama et al. 2002; Yamazaki et al. 2003), as expected from its slow spreading rate. However, the southern part of the Mariana Trough, where the back-arc spreading axis approaches the volcanic arc (within 10 km at around 13°20′N) (Martinez et al. 2000), shows a broad and smooth morphological cross section and lacks a deep crustal graben (Fig. 36.1). The morphology of this spreading axis is thus similar to that of fast-spreading ridges (Martinez et al. 2000; Martinez and Taylor 2002).

Fig. 36.1
figure 1

Regional map of the Southern Mariana Trough. The bathymetry data obtained using a SeaBeam 2112 system on cruise YK09-08 aboard the R/V Yokosuka are superimposed on the ETOPO1 dataset (Amante and Eakins 2009). The box on transect shows the area shown in Fig. 36.2. The inset shows the cross section along the black line 1 − 1′

Full size image

The similarity in morphology to fast-spreading ridges suggests that the spreading ridge of the Southern Mariana Trough receives a considerably higher magma supply than elsewhere along the trough (e.g., Fryer 1996; Martinez et al. 2000; Becker et al. 2010). The highly inflated region along the Southern Mariana Trough is centered at 12°57′N, where the axis forms a broad shallow plateau almost a kilometer wide. The high magma supply in the area is also supported by the existence of a thick crust determined from gravity analysis (Kitada et al. 2006) and by the existence of a melt lens detected by a multi-channel reflection survey (Becker et al. 2010).

Our study area is the neo-volcanic zone of the axial high between 12°56′30″N and 12°57′30″N (Box in Fig. 36.1). The area is the most inflated part along the spreading axis (Martinez et al. 2000; Baker et al. 2005). In the study area, there are two known sites of hydrothermal activity: the Snail site and the Yamanaka site (Fig. 36.2). The Snail site (12°57′12″N, 143°37′12″E) was discovered by an American group using the remotely operated vehicle Jason (Wheat et al. 2003). The site is characterized by several high- and low-temperature hydrothermal vents with clear fluid coming up through cracks in outcrops. The Yamanaka site (12°56′42″N, 143°36′48″E) is located approximately a kilometer southwest of the Snail site and was discovered by a Japanese group using the Shinkai 6500 (Kakegawa et al. 2008). Inactive chimneys and low-temperature simmering have been observed at the Yamanaka site (Kakegawa et al. 2008).

Fig. 36.2
figure 2

Bathymetry map obtained by the 400 kHz multibeam system mounted on the AUV Urashima. Contour interval is 1 m. The enlarged views in boxes a–f are shown in Fig. 36.5. Triangles indicate the locations of two hydrothermal sites

Full size image

3 Data Acquisition

Meter-scale, high-resolution, bathymetric and side-scan sonar data were acquired by the AUV Urashima (JAMSTEC), which was built in 1998 (Tsukioka et al. 2005; Kasaya et al. 2011). The AUV Urashima is fitted with a 120 kHz side-scan sonar (SSS) for obtaining backscattering intensity data, a 1–6 kHz charp sub-bottom profiler (SBP) for obtaining sub-seafloor sediment information (EdgeTech2200). It is also fitted with a 400 kHz multi-beam echo sounder (MBES) to obtain meter-scale bathymetry data and backscattering intensity data simultaneously (Seabat7125), a depth meter, an altimeter, and a conductivity, temperature, depth, and dissolved oxygen level (CTDO) sensor.

The Urashima Dive-91 was conducted at the back-arc spreading center between 12°56′30″N and 12°57′30″N (Fig. 36.2) during the YK09-08 cruise in 2009 (Okino and Shipboard scientific party 2009). The survey was done along seven ~2 km long survey lines parallel to the spreading axis at intervals of roughly 100 m. The obtained data cover an area approximately 2 km long and 1 km wide that includes the neo-volcanic zone. The average survey altitude and speed of the AUV during the YK09-08 cruise were about 100 m and 2 knots, respectively. The expected across-track resolutions for the AUV’s acoustic imagery are several meters for the 400 kHz MBES and approximately 7.5 cm for the 120 kHz SSS (when acoustic velocity in seawater is 1,500 m/s). The acoustic beam along-track footprint is 2–5 m (beam width is 0.5° for MBES and 0.9° for SSS). The data processing method is described in Asada et al. (Chap. 37).

In this area, the Shinkai 6500 submersible performed a total of six dives in 2003 and 2005, and then three more dives after the AUV observation during the YK10-11 cruise in 2010 (Kojima and Shipboard scientific party 2011). We utilized video images and photographs obtained during the dives as ground references for the acoustic imagery data.

4 Results and Discussion

4.1 Overview of the Survey Area

The micro-bathymetry collected by the AUV-attached 400 kHz multibeam sonar is shown in Fig. 36.2. The northwestern part of the survey area is characterized by well-developed faults; on the contrary, the southeastern part is shallower and dominated by volcanic structures. Hereafter, we refer to these parts as the western and eastern areas following the nomenclature of Yoshikawa et al. (2012) (Fig. 36.2). The mosaic image of the 120 kHz side-scan sonar data and its geological interpretation are shown in Fig. 36.3a and b, respectively.

Fig. 36.3
figure 3

(a) Mosaic imagery of 120 kHz side-scan sonar data. Darker color indicates lower backscattering intensity. Bathymetry is shown by contours. Tracks of nine submersible dives used for ground reference are indicated. Boxes a and b indicate the locations of Fig. 36.4a and b, respectively. (b) Map showing the interpretation of the side-scan sonar imagery. High-backscattering lumpy terrain and low-backscattering terrain are indicated by purple and black colors, respectively. Orange areas indicate relatively large hills and gray areas indicate smooth terrain. Boxes indicate the same as in (a)

Full size image

The western area occupies approximately 30 % of the surveyed area. The side-scan sonar image shows the existence of extensive linear features (Fig. 36.3a, b). These linear features, sometimes associated with acoustic shadows, are generally interpreted as faults, fissures, lava flow channels, and levees. A 30 m high, rectangular hill lies in the area (Fig. 36.2) and is cut by the linear features. The orientation of the linear features is typically NNE–SSW to NE–SW and corresponds well with the orientation of the ridge axis. Both side-scan sonar intensity and SBP data suggest that the sediment cover is very thin or absent in this area.

The eastern area occupies approximately 70 % of the surveyed area. The bathymetry map (Fig. 36.2) shows that this area consists of several mounds, ring-shaped craters, and minor ridges. These volcanic structures are mostly undeformed by faults and are aligned in the NNE–SSW direction, forming the neo-volcanic zone. The relative elevation of these features is 5–10 m. On the side-scan sonar image, we recognize a few linear features that are mostly not associated with acoustic shadows (Fig. 36.3). Generally, such linear features are interpreted as faults or fissures with small vertical throw. Two hydrothermal sites lie in the eastern area (Fig. 36.2). We were unable to recognize any chimney-like structures at and around the Snail site (Fig. 36.4a). A small chimney-like feature was observed on the sonar image at the Yamanaka site (triangle in Fig. 36.4b). The Yamanaka site is situated on an approximately 25 m-high, flat-topped mound (Fig. 36.2), that adjoins another flat-topped mound to the southwest. Other several chimney-like structures were also recognized in between two mounds (circle in Fig. 36.4b). The surface of these mounds corresponds to high backscattering and lumpy terrain (described in the next paragraph) on the sonar imagery. Both side-scan sonar intensity and the SBP data suggest that the sediment cover is very thin or absent in this area, too.

Fig. 36.4
figure 4

The 120 kHz side-scan sonar images (left) and 400 kHz multibeam bathymetry (right) of two hydrothermal sites: (top) Snail site, (bottom) Yamanaka site. The locations of maps are shown by boxes in Fig. 36.3a. Orange triangles indicate the active hydrothermal areas recognized by visual observation. Circle in the bottom figure (Yamanaka site) indicate the location at which we recognized a few chimney-like structures on the sonar imagery

Full size image

4.2 High-Backscattering Lumpy and Smooth Terrains

We categorize the high-backscattering terrain into two groups: lumpy terrain and smooth terrain. Figure 36.5 shows the typical facies of these terrains on the sonar image, and the distribution of these terrains is shown in Fig. 36.3b.

Fig. 36.5
figure 5

Lumpy, smooth, and low-backscattering terrains on the side-scan sonar image. (a) 400 kHz side-scan sonar imagery, (b) 120 kHz side-scan sonar imagery, (c) interpretation image of the 120 kHz side-scan sonar imagery, (d) lumpy terrain (400 kHz), (e) smooth terrain (400 kHz), (f) low-backscattering terrain (400 kHz), (g) lumpy terrain (120 kHz), (h) smooth terrain (120 kHz), and (i) low-backscattering terrain (120 kHz). White dotted lines on the 400 kHz side-scan sonar image indicate the edge of the swath

Full size image

The lumpy terrain is characterized by densely spaced small-scale bumps (lumps) (Figs. 36.5d, g). The typical size of each bump is 20–30 m and the relief is less than a few meters. No dominant direction is recognized for each bump or the distribution pattern. The lumpy terrain occupies most part of the eastern area (Fig. 36.3b).

The smooth terrain is the area exhibiting a smooth surface with finer dots (Figs. 36.5e, h). The terrain shows relatively high backscattering intensity and has no prominent pattern on the sonar image. The smooth terrain covers a large part of the western area (Fig. 36.3b)

4.3 Low-Backscattering Terrains

We recognized at least 49 sites of smooth surfaces with low-backscattering signatures on the side-scan image (Fig. 36.3b). These sites, which we refer to as low-backscattering terrain, can be distinguished from the lumpy terrain (Fig. 36.5f, i). The low-backscattering terrain appears as a very fine and homogeneous pattern with few acoustic shadows. The boundaries between the low-backscattering terrain and the lumpy terrain are distinct in some places but ambiguous in others. This variation may be attributed to the differences in age and/or morphological relationships between them. The relative proportion of low-backscattering terrain to lumpy terrain within the observed area was approximately 10 %. The low-backscattering terrains were observed in both the eastern and western areas (Fig. 36.3b). Figure 36.6 shows close up views of the low-backscattering terrains with a sonar intensity profile across these terrains. The low-backscattering terrains are observed in areas with various morphologies: on the top and slope of several mounds, minor ridges, and the slope and bottom of ring-shaped craters.

Fig. 36.6
figure 6

Detailed view of bathymetric features of the low-backscattering terrains. From left to right, panels show multibeam bathymetry maps, side-scan sonar imagery, interpretation images of side-scan sonar imagery, and intensity along lines shown on the side-scan sonar images. Locations of maps are shown in Fig. 36.2. Red- and gray-colored areas in graphs indicate the positions of acoustic shadow and low-backscattering terrain, respectively. The horizontal and vertical axes are pixel number and gray-scale intensity (darker shades indicate lower number)

Full size image

4.4 Visual Observation Using the Manned Submersible, Shinkai 6500

The submarine lava shows various surface forms depending on the physical and chemical property of magma, effusion rate, and local surface morphology (Griffiths and Fink 1992; Gregg et al. 1996; Fink and Griffiths 1998; Umino et al. 2000; Tominaga and Umino 2010). Pillow lavas display bulbous, spherical, or elongate tubular patterns and sheet flows exhibit smooth, lobate, rippled, wrinkled, ropy, whorly, hackly, or jumbled configurations (Fox et al. 1998; Gregg and Fink 1995; Kennish and Lutz 1998; Fundis et al. 2010). Along our observed nine transects of the Shinkai 6500 dives, the seafloor was mostly covered by bulbous pillow lava with sporadic jumbled, wrinkled, and/or partially fractured (or hackly) types of sheet lava. No axial summit troughs, pillars, or collapse features were observed, even though these features are frequently observed along the fast-spreading East Pacific Rise (Gregg and Chadwick 1996; Gregg and Fornari 1998; Kennish and Lutz 1998; Gregg et al. 2000; White et al. 2002; Haymon and White 2004; Tanaka et al. 2007).

The lumpy terrain on the sonar image corresponds to the flatly distributed bulbous pillow lava (Fig. 36.7a). We found jumbled or wrinkled (and partly hackly) sheet lava in areas consisting of low-backscattering terrains (Fig. 36.7b). We can recognize fragmented pillow lavas in the smooth terrain. We carefully investigated the boundaries between the low-backscattering terrain and the lumpy terrain in the visual records but could not recognize the age difference between theses terrains (Fig. 36.7d, e). Because of the little sedimentary cover over both terrains, differences in backscattering strength likely correspond to differences in the surface morphologies of lava.

Fig. 36.7
figure 7

Photos taken by submersible showing a large variety in lava morphology and the corresponding side-scan sonar imagery. (a) Photo of pillow lava in lumpy terrain. (b) Jumbled sheet lava in low-backscattering terrain. (c) Geological interpretation of side-scan sonar imagery. Low-backscattering terrains are shown in black. Red line, green line, and white line indicate fault, fissure, and nadir of AUV track and edge of swath, respectively. (d) Photo of the boundary of pillow and sheet lava. (e) Schematic interpretation of photo (d)

Full size image

Linear features on a hill in the western area (Fig. 36.3b) are faults, which displace the surrounding pillow lavas. Thin sedimentary cover, which was not detected by SBP observations, was observed on the seafloor in the western area. Although turbid water was observed at the foot of the hill, we did not discover any hydrothermal signature in the western area.

Very little sediment was observed in the eastern area. At the Snail site, a hydrothermal plume appeared to be coming up through the cracks of large rocky outcrops, but we did not find any chimney-like structures. The Snail site is located in a valley surrounded by three small mounds. The flat-topped hill on which the Yamanaka site is developed is covered by bulbous pillows at the top and elongated pillows on the slope. These geologic features are consistent with our interpretation of acoustic observations.

5 Conclusions

  1. 1.

    We observed the axial zone of the back-arc spreading center in the Southern Mariana Trough using the AUV Urashima and the Shinkai 6500 submersible. The neo-volcanic zone is mostly occupied by lumpy terrain with high backscattering intensity. The lumpy terrain consists of flatly distributed pillow lava.

  2. 2.

    Our high-resolution AUV observations using side-scan sonar imagery revealed the existence of low-backscattering signatures. These low-backscattering terrains are associated with a variety of bathymetric features. The submersible observations revealed that the terrain mostly consists of wrinkled and/or jumbled sheet lava, and the edges of the low-backscattering terrains are coincident with the boundaries of pillow and sheet lava types.

  3. 3.

    The proportion of low-backscattering terrain within our observed area is estimated to be approximately 10 %. Since the low-backscattering is interpreted as sheet lava and lumpy/smooth terrains are as bulbous or fractured pillow lavas, the sheet lava occupied ~10 % of seafloor when it is assumed that all of the low-backscattering areas indicate the occurrence of sheet lava.

  4. 4.

    We did not observe axial summit troughs, pillars, or collapse features, all of which are common along the fast spreading East Pacific Rise, in our study area. We also observed a few pillow mounds, which are commonly observed along the intermediate spreading Mid Atlantic Ridge