Geological setting of the Maldives

The Maldives archipelago in the central equatorial Indian Ocean is an isolated tropical carbonate platform constituting the central and largest part of the Chagos-Laccadives Ridge, which is located southwest of India (Figure F1). A north-south–oriented double row of atolls encloses the Inner Sea of the Maldives (Figure F2). The atolls are separated from each other by interatoll channels, which deepen toward the Indian Ocean (Purdy and Bertram, 1993). The Inner Sea is a bank-internal basin with water as deep as 550 m. The Maldives carbonate sedimentary succession is almost 3 km thick; it accumulated since the Eocene, away from any terrigenous input (Aubert and Droxler, 1992; Purdy and Bertram, 1993).

The archipelago comprises about 1200 smaller atolls, lying near or slightly above the sea surface. Discontinuous marginal rims formed by such small atolls surround lagoons with water depths up to 50–60 m. These rims are interrupted by deep passages, allowing strong currents within the atoll lagoons that rework and redeposit sediment and influence the growth of patch reefs (Ciarapica and Passeri, 1993). Modern marginal reefs are composed of robust-branching corals and coralline algae, whereas the lagoonal reefs show domal corals and detrital sand and rubble facies (Gischler et al., 2008). Muddy sediments are present only in the smaller atolls’ lagoons protected by a continuous marginal reef rim (Ciarapica and Passeri, 1993; Gischler, 2006).

The oceanward margins of the Maldives archipelago are generally steeply inclined, with dips of 20–30 to 2000 m water depth. On the Inner Sea side, stepped atoll slopes have the same dip angles but reach to water depths of only 150 m, where the gradient rapidly declines (Fürstenau et al., 2010). The Inner Sea is characterized by periplatform ooze deposition (Droxler et al., 1990; Malone et al., 1990), locally accumulated into sediment drift bodies (Betzler et al., 2009).

The climate and oceanographic setting of the Maldives is dictated by the seasonally reversing Indian monsoon system (Tomczak and Godfrey, 2003). Southwestern winds prevail during Northern Hemisphere summer (April–November), and northeastern winds prevail during winter (December–March). Winds generate ocean currents, which are directed westward in the winter and eastward in the summer. Interseasonally, a band of Indian Ocean equatorial westerlies establish, enforcing strong, eastward-flowing surface currents with velocities up to 1.3 m/s. Currents reach water depths of 200 m and more with only slightly reduced velocities (Tomczak and Godfrey, 2003). Within the modern atolls’ passages, currents can exhibit velocities up to 2 m/s at the surface (Preu and Engelbrecht, 1991), accounting for winnowing in the passages and lagoons, where hard bottoms form (Ciarapica and Passeri, 1993; Gischler, 2006).

The Maldives formed on lower Paleogene (60–50 Ma) volcanic basement (Duncan and Hargraves, 1990). The long-term subsidence rate of the Maldives is roughly 0.03–0.04 mm/y based on deep core data from Well ARI-1 (Figure F3) (Aubert and Droxler, 1996; Belopolsky and Droxler, 2004a). In contrast, sedimentological data from the Rasdhoo atoll indicate a maximum subsidence rate of 0.15 mm/y during the past 135,000 y (Gischler et al., 2008). Faulting of the Maldives archipelago is reported to be restricted to pre-Miocene times (Purdy and Bertram, 1993). Stratigraphic data for the chronostratigraphic framework for seismic interpretation are derived from data from Wells NMA-1 and ARI-1 and Site 716 and are described by Purdy and Bertram (1993), Aubert and Droxler (1996), Belopolsky and Droxler (2004a), and Rio et al. (1990). Carbonate lithofacies, paleobathymetric evaluations, and biostratigraphic age determinations are based on cuttings and sidewall core analyses for Shell exploration Well ARI-1 first published by Aubert and Droxler (1996). A vertical seismic profile is used for time-depth conversion and to tie well data to seismic data (Figure F3, lower panel). For Site 716, which is covered by two high-resolution seismic lines (Figure F4), a simple time-depth estimation is made based on existing whole-core P-wave velocity measurements.

The Maldives comprise an approximately 3 km thick shallow-water carbonate succession (Belopolsky and Droxler, 2004a). Carbonate production established during the early Eocene when flat-topped carbonate banks began to form on topographic highs created by the volcanic basement during the Eocene to early Oligocene. During the late Oligocene, bank margins typically had elevated rims, which separated bank-interior areas from the open ocean. During the early Miocene, these banks partially drowned and carbonate production became restricted to narrow bands at the respective most oceanward areas. During the Miocene, bank margins prograded toward the Inner Sea, as recognized in different versions of reflection seismic data, irrespective of seismic resolution; however, details of the interpretation differ (Purdy and Bertram, 1993; Aubert and Droxler, 1996; Belopolsky and Droxler, 2004a). Aubert and Droxler (1996) differentiated the prograding margins into 4 Neogene units (N2–N5), with Unit N2 comprising the main phase of bank-margin progradation (Figure F4). Units N3–N5 were seen as its waning stage, with Unit N3 sediments interpreted as tending to be preferentially accumulated in front of the prograding bank margins. Unit N3 isochrons reveal that this unit is locally linked to areas of bank-margin disintegration, which in turn are associated with partial drowning and channel erosion (Aubert and Droxler, 1996). New higher resolution data prompted reinterpretation of the Units N3–N5 as large drift sequences. During the upper Miocene and Pliocene, the Inner Sea basin was filled while bank margins dominantly aggraded (Belopolsky and Droxler, 2004a) but also showed further partial drowning after Unit N5 deposition (Aubert and Droxler, 1996).

In the new high-resolution multichannel seismic data, 10 sequences are recognized in the lower and middle Miocene strata of the Maldives. They are interpreted to have formed in response to sea level–driven accommodation space variations (Betzler et al., 2013a) (Figure F4C). Platform Sequences (PS) 1–6 show development from a shallow ramp to a steep-flanked, reef-rimmed carbonate platform. Bank-edge reefs protect the lagoon, where back-reef aprons occur. At the PS6–PS7 transition, a switch from dominantly aggrading to dominantly prograding bank margins occurs. PS7–PS10 are composed of deposits formed in response to forced regression overlain by deposits formed during reflooding of the bank margins. These sequences lack bank-edge reefs, similar to other carbonate platforms of the Indo-Pacific realm, where community replacement in a neritic environment occurred (Halfar and Mutti, 2005; Betzler et al., 2013a).

Neogene sequences of the Maldives: middle Miocene to Pleistocene

The upper middle Miocene is characterized by the appearance of large-scale lobate clinoform bodies, attesting the onset of current amplification (Figures F4, F5, F6) in the Inner Sea (Betzler et al., 2013a; Lüdmann et al., 2013). These bodies are attached to passages where parts of barrier reefs drowned while relict banks and atolls grew elsewhere (Betzler et al., 2009, 2013a). Lobes are interpreted as “mega spillovers” fed by easterly currents and reworked by a current system flowing obliquely or normally to this main stream (Figures F6, F7). This current pattern filled the Inner Sea from west to east (Lüdmann et al., 2013). Starting with Drift Sequence (DS) 6, the opening of a southern gateway introduced northward flow of bottom waters in the Inner Sea, leading to deposition of giant elongated drifts at the eastern flank of the basin, filling it from east to west. Because the current swept away most of the material around the atolls, the system was not able to prograde and the steady subsidence was compensated by aggradation (Betzler et al., 2013a; Lüdmann et al., 2013).

DS1–DS9 are mostly conformable in the central part of the Inner Sea, but they display downlaps and onlaps at the basin margins. Two main types of bottom current–controlled deposits occur: (1) a prograding wedge to clinoform type near the passages between the atolls and (2) a mounded shape type located along the atoll flanks with a broad moat of 2.2–2.5 km in updip direction (Lüdmann et al., 2013). The latter is interpreted as a plastered drift body of a giant elongated drift, migrating along the basin flank under a weak current regime (Faugères et al., 1999). Narrower moats (500–800 m) at the western flank of the Inner Sea are interpreted as reflecting a high-velocity current focused in the west and a wider, slower current in the east. This partitioning is in line with monsoonal-triggered current reversals.

Two additional drowning steps affected the Maldives in the upper Miocene and lower Pliocene (Betzler et al., 2009, 2013a) (Figure F8). Flat-topped and atoll-shaped banks are interpreted to have drowned quickly, whereas mound-shaped banks are interpreted to have undergone sequential drowning under elevated nutrient fluxes (Betzler et al., 2009), similar to banks described by Zampetti et al. (2004) offshore Malaysia. The drowned banks are elongated parallel to the extension of the Kardiva Channel, which is characterized by throughflow of monsoon-driven surface currents and globally driven bottom currents (Lüdmann et al., 2013). This shape is interpreted as reflecting current control, similar to the recent atolls, which are elongated because of the action of waves and currents (Purdy and Bertram, 1993).

Partial drowning steps triggered by currents

Marine records suggest onset of monsoonal-triggered marine upwelling at ~8.5 Ma (Kroon et al., 1991). An increase in sediment flux into the Indian Ocean occurred around 11 Ma (Rea, 1992; Zheng et al., 2004), and a peak in the sedimentation rates of the Indus Fan occurred between 16 and 10 Ma (Clift et al., 2008). Onset and monsoon-intensification steps correlate with the Maldives partial drowning steps (Betzler et al., 2003a). It is therefore proposed that upwelling from monsoonal currents shaped the atolls in the past, controlling sediment production and reef growth. Thus, timing of the partial platform drowning and monsoon evolution is linked. Under monsoon conditions, upwelling injected nutrients into surface waters, affecting the carbonate banks (Betzler et al., 2009). Even short-term seasonal upwelling forces adaptation of the carbonate factory (Reijmer et al., 2012), controlling a biotic association with low growth potential more vulnerable to sea level change effects. This process caused the barrier reef demise, which was replaced by a string of current-shaped relict reefs, separated by passages accommodating the throughflow of currents (Betzler et al., 2009, Lüdmann et al., 2013).

Seismic studies/site survey data

Table T1 provides characteristics of horizons tied into the available seismic data of the Maldives. The stratigraphic framework was established for the succession above the Oligocene/Miocene boundary, as the imaging quality of older strata is regarded to be too poor for interpretation. It is based on published horizons O/M, EM1, E/MM, MM3, MM5 (Belopolsky and Droxler, 2004a), and PB2 (Purdy and Bertram, 1993), which set a reliable stratigraphic framework throughout the Shell and Elf seismic grid and are tied into the newer high-resolution seismic data (Figure F5). In addition to the established horizons, the new high-resolution seismic lines allow us to trace a better approximation of the base of the middle Miocene (bMMio), as defined by Belopolsky and Droxler (2004a) at Well ARI-1, and the base of the early Pliocene (bEPlio), defined at Site 716 (Rio et al., 1990), throughout the seismic grid.

Supporting site survey data for Expedition 359 are archived at the IODP Site Survey Data Bank. Site survey data consist of multichannel reflection seismic lines and parasound lines acquired during R/V Meteor cruise M74/4 in December 2007 (Betzler et al., 2009, 2013a; Lüdmann et al., 2013) and the integration of the published low- to medium-resolution industrial reflection seismic Lines 1973/74 shot for Elf (Purdy and Bertram, 1993; Aubert and Droxler, 1996) and Lines 1989/90 for Shell (Belopolsky and Droxler, 2004), respectively. The Shell seismic data set covers the Inner Sea and the interatoll passages. The Elf seismic grid also transects most of the Maldivian atolls and offers good penetration depth across the atolls and their drowned parts. Stratigraphic interpretation of seismic data is made via correlation to published data of exploration Wells NMA-1 and ARI-1 and Site 716. This study is complemented by multibeam and parasound data, which were continuously recorded during the Meteor cruise.

The high-resolution seismic data set consists of approximately 1400 km of reflection seismic profiles. Seismic signals were generated by 2 clustered generator-injector (GI) guns, each with a volume of 45 in³ for a 105 in³ generated injector volume. A digital 144-channel streamer array with an active length of 600 m and an asymmetric group interval was used. The data were digitized with 7 SeaMUX 24-channel 24-bit digitizing modules, configured in 6 multiple arrays totaling 144 channels. The shotpoint distance during the entire cruise was 12.5 m. The dominant frequencies center around 100–120 Hz. Processing of reflection seismic data was done using the software package ProMAX 2-D (Halliburton-Landmark). The data are processed to zero phase, filtered in time and f-k domain, and corrected for dip moveout. In basinal areas, suppression of multiple reflections was achieved by predictive deconvolution of prestacked data. Amplitude losses were compensated by a power function. Interpretation and visualization was done using the software package Petrel (Schlumberger). Depending on depth, the vertical resolution of the newly acquired data is approximately 4–6 m compared to only 10–25 m of the Shell seismic data (Belopolsky and Droxler, 2004). The vertical resolution of the Elf seismic data is lower. Seismic interpretation was performed on time-migrated data in time domain. As the continuity of the reflections in part is weak, the instantaneous phase was also used for tracing.

An overview of the acquisition parameters and processing steps is as follows:

  • Survey speed: average 5 nm;
  • 44-channel digital streamer;
  • Active length: 600 m;
  • Group interval: asymmetric (average = 6.25);
  • Shot interval: 12.5 (25 m p65 only);
  • Sample rate: 1 ms;
  • Offset source, first receiver: 55 m;
  • Fold: 72 (p65 only) to 144;
  • Common depth point (CDP) interval: 12.5 m;
  • Source: 2 GI guns, total volume 300 in3;
  • Source frequency: 80–120 Hz;
  • Software: Halliburton-Landmark ProMAX 2-D;
  • Trace editing;
  • Bandpass: Ormsby, frequency domain, 20-25-200-220;
  • Minimum phase predictive deconvolution (multiple suppression);
  • Designature (remove of source signature; transformation to zero-phase);
  • Spherical divergence correction;
  • Surface consistent amplitude recovery;
  • Normal moveout (NMO) correction;
  • Dip moveout (DMO) correction;
  • Stack (mean);
  • Fast explicit frequency domain (FD) time migration (12.5 m interval, max. freq. = 200 Hz);
  • Automatic gain control (AGC) (1024 ms);
  • SEG-Y export (ibm reel); and
  • CDP locations: Projection UTM_WGS84_43N.

Subbottom profiles were recorded using the Parasound system on the Meteor in 2007. The system operates with 2 frequencies (18 and 22 kHz) emitted in a 4 cone from 2 hull-mounted transducers. The Parasound system uses the parametric effect caused by the interference of 2 frequencies in the water column. The effective signal of the Parasound is a 4 kHz wavelet, which results a seismic footprint with a diameter of 7% of the water depth. The seafloor penetration depth can reach up to about 200 m but varies strongly with the lithology, grain size, and gas load of the sediment.

Recorded data were stored in the Parasound native format ps3 and the raw-data format ASD for later processing. For onboard visualization, ps3 data were converted with the tool ps3segy (Hanno Keil, University of Bremen, Germany) into the standard seismic data format SEG-Y. Onboard data processing was performed with the software ReflexW (Sandmeier Software) and comprises AGC and amplitude normalization along the profile. Long profiles were subject to trace stacking in order to reduce the data volume. Processed data were stored in SEG-Y format for further visualization and interpretation in the software packages Kingdom Suite (IHS) and Petrel (Schlumberger).

Geological setting of the Kerala-Konkan Basin

Proposed Site KK-03B is positioned ~220 km offshore India at the edge of the Kerala-Konkan Basin (Figures F1, F9) that formed during the Mesozoic rifting of Gondwanaland (Kalaswad et al., 1993). Climate and the oceanographic setting of the site is similar to the Maldives and controlled by the seasonally reversing Indian monsoon system (Tomczak and Godfrey, 2003). A basement ridge, the Pratap Ridge, runs parallel to India’s coast in the upper continental slope region, delineating a series of offshore basins under the shelf and upper slope (Naini and Talwani, 1982; Subrahmanyam et al., 1991). These basins contain 2–4 km of sediment (Rao, 2001; Campanile et al., 2008). Farther offshore, the Chagos-Laccadive Ridge parallels the southwestern margin of India south of ~15N and continues across the Indian Ocean as the trace of the Réunion hotspot (Duncan, 1990). The drill site is located on line with the Chagos-Laccadive Ridge just north of it where it fades away as a bathymetric feature (Collet et al., 2008) (Figure F10). A basement high near the site comes to within 450 m of the seafloor. Isopach maps show slightly less than 1000 m of sediment in this region (Campanile et al., 2008) compared to more than 3 km north of the site in the Bombay Basin and south in the Kerala-Konkan Basin proper (Figure F9). Western India has a well-defined Cretaceous escarpment, the Western Ghats, that runs parallel to the coast at an average height of 1200 m. Monsoonal rains fall preferentially at the coast under this orographic influence (Xie et al., 2006). Surface waters at Site KK-03B are exposed to low-salinity summer monsoon conditions from direct rain, from fluvial discharge of Ghats rivers, as well as from the Bay of Bengal low-salinity sea-surface currents coming around the southern tip of India in winter (Jensen, 2001).

Drilling by the National Gas Hydrate Program of India at Site NGHP 01-1A located nearby yielded a continuous 300 m thick upper Eocene-recent sequence of hydrate- and turbidite-free, pelagic foraminifer-rich sediments with secondary siliciclastics (Collet et al., 2008) (Figures F11, F12). Seismic data show parallel stratification at the site without structural complications. The age model for Site NGHP 01-1A by Flores et al. (in press) shows that sedimentation was continuous for the last 35 My with bulk average sedimentation rates increasing significantly after ~9–8 Ma, with a gradual increase since ~18 Ma. Site KK-03B is expected to recover a similar but expanded section. These changes in sedimentation rate are accompanied by changes in (1) terrigenous vs. pelagic components; (2) clay mineral assemblage toward more illite dominated (Phillips et al., in press), indicative of physical weathering; and (3) possible increased inputs of C4-type plant-derived organic carbon indicative of arid conditions (Johnson et al., in press).