UNLIKELY BACKBONES SUPPORTING FLAT SUBDUCTION

Distribution of earthquakes in a nearly subhorizontal plane has been attributed to flat subduction of the oceanic lithosphere. Numerical models with a high transition temperature for the eclogitisation in the subducted crust, and assuming the contribution of the overthrusting continent, favour flat subduction in comparison with the sole subduction of an oceanic plateau (van Hunen et al., 2004). However, an uncertain “correlation” has been made between narrow subducting plateaus and the location of flat subduction, and several of the flat slab regions in the Earth are not correlated to overthrusting continents. In the Andes, the subduction of seamounts and ridges has currently been considered a main mechanism for flat subduction through enhanced buoyancy of the oceanic plate due to hydration of these oceanic backbones. In the ‘flat slab' region of western Argentina, Vp, Vs and Vp/Vs distributions surrounding a projection of the Juan Fernández ridge (Wagner et al., 2005), are seen in this communication as problematic. Also, the eruption of Late Miocene Pocho lavas at a distance of about 750 km from the trench, is in conflict with thermal models indicating that magma generation at distances larger than 600 km during flat-slab subduction is unlikely (English et al., 2003). In a magnetotelluric profile running east of the eastern edge of the ‘flat slab' region of western Argentina, a structure with its top at about 100 km depth, and at 800-900 km inboard of the trench, shows enhanced electrical conductivity, and may imply that partial melting occurs to at least 250 km, or that melt is supplied from the 410 km discontinuity (Booker et al., 2004). Instead of flat subduction, it is proposed under several rheological parametrizations that earthquakes in a subhorizontal plane at 70-110 km depth beneath large regions of the Andes are occurring in the continental lithosphere due to low radiogenic heat generation of these geologically aged or allochthonous terranes. This may be confirmed in the case of Ecuador where the real Wadati-Benioff zone with normal dip was found beneath earthquakes observed at 70-80 km depth (Guillier et al., 2001). In Perú and Argentina, the subhorizontal plane is observed at 100-110 km depth beneath the old terranes, and studies of local seismicity do not show until now a normally dipping Wadati-Benioff zone.