I - Quasi-3-D Seismic Reflection Imaging and Wide-Angle Velocity Structure of Nearly Amagmatic Oceanic Lithosphere at the Ultraslow-Spreading Southwest Indian Ridge (Momoh et al 2017)

The results from 3-D processing of 2-D seismic data shot along 100 m spaced profiles in a 1.8 km wide by 24 km long box during the SISMOSMOOTH 2014 cruise allow us to understand the oceanic crust formation at an end-member mid-ocean ridge environment of nearly zero melt supply.

1 .A series of bright and weak amplitude reflectors, which dip southward beneath the elevated northernwall in our study area, have been imaged and interpreted as the damage zone of the active axial detachment fault. Although the velocity model does not constrain the basement in the footwall compartment of this detachment fault due to insufficient data coverage, we observed that the velocity in the hanging wall next to the fault is 0.5¿1 km/s lower than the other parts of the model at any given depth, suggesting more extensive fracturing. The series of subparallel reflectors outlining the inferred damage zone occurs in a domain that is 2¿2.8 km thick, and dip 50 to 60° down to 5 km below the seafloor. This inferred damage zone could have developed as a finite system due to strain localization on a succession of individual frac- tures or due to distributed instantaneous deformation over a thick domain. A combination of both scenar- ios is also possible. Whichever is likely, we proposed that this pervasive fracturing may provide a mechanism to explain the serpentinization process on a crustal scale.

Figure 1 : (a) Location of the study area at the Southwest Indian Ridge (SWIR, 64°E longitude). (b) Bathymetric map of the study area at the easternmost Southwest Indian Ridge showing the boundary between a corridor of nearly amagmatic smooth seafloor (dark shading on map) and volcanic seafloor (dashed blue line; Cannat et al., 2006). The breakaway of the inferred active axial detachment fault is mapped at the top of the northern axial valley wall, and its emergence is shown in purple (Sauter et al., 2013). Proposed inactive detachment fault breakaways are indicated in the southern Antarctic plate. These breakaways are drawn following the apex of successive ridge-parallel elongated massifs. The red thick line indicates the ridge axis (Sauter et al., 2013). The white rectangle indicates an area where the seafloor reflec- tivity was surveyed by a towed ocean-bottom instrument (Sauter et al., 2013), which is shown in Figure 1c. The thin white lines in the narrow box within the larger white rectangle show eighteen 2-D multichannel seismic lines, spaced by 100 m that we processed in a 3-D manner in this study. These lines are numbered 13 to 31 (from east to west) and image a 1.8 km wide and 24 km long box. The white thinner lines on the bathymetric map are other seismic lines acquired during the SISMOSMOOTH cruise. The large white circles indicate the location of ocean-bottom seismometers, and the intersecting white line across axis is the corresponding shot line. The smaller white circles are other ocean- bottom seismometers deployed in a 3-D network during the SISMOSMOOTH cruise. (c) Acoustic backscatter image (located as a white rectangle on the bathymetric map in Figure 1b). Basalts in volcanic seafloor have strong reflectivity, as do scarps in the surrounding ultramafic seafloor. The intersecting ridge axis, emergence, and breakaway of the active detachment fault and the breakaway of one of the inactive faults are indicated as in Figure 1b. A zoom of the 3-D box is shown as a white line, with three black lines indicating the bordering and central lines constituting the box.

2. We have also imaged several north dipping reflective events in the hangingwall (Antarctic) plate of the axial detachment fault. For these events, we propose two alternative hypotheses: one in which the reflec- tors represent damage zones formed in the footwall of a previously active north dipping axial detachment fault and the other hypothesis is that the reflectors formed in response to more recent tectonic extension and incipient magmatism in the hanging wall of the presently active, south dipping axial detachment fault. We propose that flat-lying or weakly dipping discontinuous but coherent reflective events, also located in the hanging wall domain below the axial valley, may correspond to crystallized melt sills in the velocity-defined crust and upper mantle. An alternative interpretation for the deepest events would be that they reflect contrast between locally enhanced serpentinization zones and fresher peridotite. Melt sills, if present, are volumetrically small, and there are no indications in the P wave tomography model for kilometer-scale bodies of partially melted material at crustal or upper mantle depth in the axial domain.

Figure 2 : (a) Perspective view of the 3-­D processed seismic volume with inline 13 on the forefront. The displayed box is 19 km long and 1.8 km wide (Figure 1). It is viewed from the SSE. (b) 3-D perspective view of the line drawings of the interpreted events (reflectors) picked in each of the 19 inlines of the 3-D time-­migrated seismic volume. (c) Same as Figure 14b, but viewed from the east. Four distinct categories of events are observed: south dipping events (in purple), north dipping events (in green), and two categories of subhorizontal reflectors occurring in the shallow TWT (cyan) and deep TWT (orange) regions of the processed volume. Seafloor topography is derived from picking the seafloor reflectivity from poststack time-migrated volume, and interpolating data between the picks.

3. Crustal thickness derived from the P wave velocity model ranges between 4 and 5km, where the thickest portion is observed in the hanging wall close to the active south facing axial detachment fault. The crust, on the basis of existing geological evidence and the geophysical data analyzed in this paper, is proposed to be composed mainly of variably serpentinized peridotite.

4. Vertical velocity profiles extracted from our Pwave velocity model display nearly linear velocity to depth gradients, similar to those observed over ultramafic seafloor at the 23°N MAR (Canales et al., 2000). Velocity values deeper than 1 km into the basement, however, are higher at 23°N MAR, which may be due to less internal deformation in the exhumed ultramafics and to more abundant gabbroic intrusions. Our velocity profiles also show similarities with the average P wave velocity structure of OCT basement from the Iberia-Newfoundland magma-poor rifted margin (Minshull, 2009).

Figure 3: Sketch integrating the results from the 3-D seismic reflection volume and wide-angle P wave velocity model. The dashed thick purple markers trace the bright amplitude reflectors in the footwall of the axial detachment fault. The thinner dashed purple lines indicate areas where weak amplitude reflectors are persistent with similar southern dips. The principalfault plane (i.e., the limit between the hanging wall and footwall plate) isinferred and shown as a continuous purple line. The green dashed lines arethe north dipping reflectors imaged in the axial valley domain (hangingwall of the active detachment fault). The base of the crust, which we proposeas corresponding to the transition from serpentinized peridotite to freshmantle, is sketched on the 7.5 km/s velocity contour. This contour is representedonly in the domain where it is well resolved by the OBS data.

II - Internal Structure of the Oceanic Lithosphere at a Melt-starved Ultra-slow Spreading Mid-Ocean Ridge: Insights from 2-D Seismic Reflection Data (Momoh et al subm.)

Extensive outcrops of serpentinized peridotite in melt-starved spreading corridors of the ultraslow easternmost Southwest Indian Ridge are hypothesized to be due to slip on successivelong-offset normal faults that alternate polarity (flip-flop detachment faults). We investigate the nature of the oceanic crust which forms under these conditions, using seismic reflectiondata acquired during the SISMOSMOOTH 2014 cruise. Using 3-D binning, the seismic profiles were binned elastically, while three of the profiles shot closely were merged into oneto take advantage of the larger air gun source volume. Using a poststack imaging sequence,we observe several types of reflectors at crustal and infracrustal depths, in the axial valley and off-axis. Correlating our seismic observations with Residual Mantle Bouguer gravity anomalies and seafloor observations, we find that our results are explicable in the frame-work of the flip-flop hypothesis of detachment faulting. Reflectors imaged down to 5 km into the basement and interpreted as due to damaged zones outlining [UdMO1] the detachment faults dip 50° atthe early stages, while at late stages after developing offsets >10 km, they dip 25°.Other reflectors observed in the crust are interpreted as moderate offset (< 200 m) normal faults accommodating deformation and alteration in the hanging wall and channelling the sparsemelt to the seafloor. We interpret these and other observed seismic reflectors in the frame of a two-phase evolutionary sequence over the lifetime of two successiveflip-flop detachmentfaults: exhumation, footwall flexure, damage, serpentinization and incipient magmatism in the footwall of one detachment fault; followed by further tectonic damage, alteration and incipient magmatism in the hanging wall of the next detachment fault.