Anomalously shallow bottom-simulating reflections on the upper Amazon deep-sea fan record gas hydrate response to upward fluid/heat flux
Résumé
The Amazon River supplies one of the world's largest deep sea fans, a sedimentary depocentre over 10 km thick that occupies an area of nearly 330,000 km 2 from the shelf-break to water depths greater than 4000 m. Gas hydrates have been inferred to occur within the Amazon fan, based on seismic profiles showing discontinuous bottom simulating reflectors (BSRs) on the upper fan, and the presence of gassy sediments and pore water freshening in sediment cores of ODP Leg 155 on the lower fan. Gas hydrate dissociation during glacial-interglacial changes in sea level has been hypothesized to act as a trigger for slope instabilities within the fan. The Amazon fan is also characterized by remarkably high rates of sedimentation, especially during the late Quaternary when average sedimentation rates reached 50 mm/year during glacial stages. Rapid deposition favours overpressure build-up, which can drive deformation at different scales. On the Amazon fan this includes gravitational collapse above multiple deep detachment surfaces, expressed near the seafloor in a large-scale extensional-and-compressional system characterized by normal faults on the shelf and upper slope (above ~500 mbsl) and thrust-folds at greater depths (from ~1000 mbsl to 2000 mbsl). Some of these structures can reach the seafloor generating scarps up to 500 m in relief attesting to ongoing deformation. Our interpretation of a regional grid of 2D and 3D multichannel seismic reflection data, combined with modeling of the gas hydrate stability zone, provides new information on the distribution and character of discontinuous BSRs on the Amazon fan and their relation to fluid flow through structures within the compressive domain. BSRs are only observed in water depths of 1200-2000 m, mainly on the northwest part of the main depocentre, forming a series of elongate BSR 'patches' up to 140 km long and 10-50 km wide that coincide with the crests of thrust-folds. The BSRs mainly lie at 200-300 mbsf, as strong reflections of negative polarity that cut across deformed strata within the thrust-folds and fade away within intervening basins. In places the BSRs rise to shallower depths, within 150 m of the seafloor, in some cases beneath fluid escape features observed on 3D seismic, including pock-marks and probable mud volcanoes linked to migration pathways within the thrust-folds. The regional methane hydrate stability zone (RMHSZ) was modeled using the phase boundary for pure methane in equilibrium with water of 3.5% salinity and inputs for bathymetry (Gebco08), bottom-water temperatures (World Ocean Database) and geothermal gradients (International Heat Flow Commission, ODP and published sources). The results show the upper limit of the RMHSZ to lie between 500-600 mbsl, its exact position being controlled by bottom-water temperatures that vary over seasonal and longer timescales. The RMHSZ thickens rapidly downslope in response to low geothermal gradients within the upper-central fan (17-20˚C/km), to maximum thicknesses of over 1 km (in 2500-3000 m water depth) but decreases to 305 m on the lower fan (> 4,000 m water depth). On the upper fan, the modeled depth of the RMHSZ (700-900 mbsf) is several times greater than the maximum depths of the BSR observed on seismic sections. Anomalously shallow BSRs could be explained either by highly saline pore fluids, or by higher sub-seafloor temperatures. Modeling of the RMHSZ using a range of constant geothermal gradients shows the BSR patches to correspond to values that vary across their widths (of 10 km or more) from 40˚C/km at their edges to as high as 90˚C/km where shallowest. This variation in gradients corresponds to vertical variations in temperature of up to 7.5˚C across the observed sub-seafloor depth range of the BSRs (150 m). The coincidence of BSRs with thrust-folds versus their absence elsewhere (over water depths of 500-4500 m), suggests that compressive structures are key elements for the migration of fluids from depth. The fact that BSRs are only observed above thrust-folds could be explained in terms of higher fluid flux, in particular by a supply of free