Abstract The upper Awash runs across a volcano-sedimentary succession dated from the Early to Middle Pleistocene and located on the western margin of the northern portion of the Main Ethiopian Rift. The succession lies above Late Miocene to Pliocene lava flows, domes and large volume ignimbrites. The succession formed within a fluvial system that developed within transversal rift faults. The stratigraphy consists of primary volcanic deposits interbedded with reworked sediments emplaced in a low-energy floodplain developed in a subsiding area. The lower part of the volcaniclastic sequence is dominated by a 1.2 Ma old, low aspect ratio, pyroclastic density current deposit (Kella ignimbrite). This eruption was followed by an eruptive stasis and reorganization of the drainage system. Tephra deposition in the floodplain climaxed again between 0.9 and 0.7 Ma and was associated with extensive tephra reworking. Sedimentation rates significantly decreased after 0.6 myr ago, probably owing to declining volcanic activity. Dynamic interaction between tectonics and volcanic activity created a complex sedimentary environment preserving numerous artefacts (lithic tools), animal and hominin remains. Stratigraphic correlation is based on the interpretation of the basin evolution and has relevance for the reconstruction of the palaeoenvironment and the interpretation of the palaeontological and archaeological data.

Abstract Outer marginal collapse (OMC), a recently proposed process by which top-rift and base-salt unconformities formed near sea level may subside rapidly to 2.5–3 km at continental margins as mantle exhumation or seafloor spreading begins, needs further examination. We examine salt deposition at three margins and find that the differing positions and volumes of salt can be related to different durations of salt deposition as OMC and subsequent mantle exhumation proceed. Along NW Florida, salt is thin but deep and is interpreted as having formed at the start of OMC, before drowning further to abyssal depths. In the Campos Basin, salt is thick and extends across tens of kilometres of interpreted exhumed mantle, interpreted as having formed during the entire period of OMC before spreading onto mantle during exhumation. In the Santos Basin, salt is thick and extends across c. 100 km of interpreted exhumed mantle and/or oceanic crust, arguably requiring ‘lateral tectonic accommodation’, whereby salt deposition persists near global sea level across the conjugated salt basin during mantle exhumation beneath mobile salt. The supposition that OMC can account for salt deposition in three different basins without invoking problematic 1.5–2 km-deep subaerial depressions provides further support for the process.