Interest in the processes by which human populations expanded out of Africa after about 2 million years ago to colonise the rest of the world has never been higher in the scientific and public consciousness, thanks to new fossil and archaeological finds, palaeoclimatic data and DNA-based information on phylogenetic histories. What has been largely left out of account is the far-reaching effects on human adaptive strategies and patterns of dispersal of physical landscapes subject to geological instability associated with rifting, plate motions, seismic activity, volcanism and sea level change. These processes dominate the geological history of the regions of Africa and southern Eurasia where early finds are concentrated, but have received little attention in comparison with climate change. We aim to place these processes centre-stage, not simply as occasional disruptions that interfere with the smooth flow of biological and sociocultural evolution, and hamper the archaeological task of cultural and palaeoenvironmental reconstruction, but as ‘agents without intent’ that have interacted intimately with, and selected for, new strategies of human adaptation and social action at every scale, from the local to the continental and from the short-term event of a single earthquake to wholesale reconfiguration of the physical landscape. We will focus our attention on the major episodes of dispersal associated with archaic and anatomically modern humans (H. ergaster/erectus and H. sapiens sapiens), and on the western Arabian Peninsula and adjacent Red Sea as a key but little investigated intermediate pathway between Africa and Eurasia. We will apply new concepts and techniques of tectonic and volcanic geomorphology to the reconstruction of ancient landscapes and habitats in the region, including comparative studies from sample regions in Africa and southern Eurasia, in order to characterise the landscape conditions that have attracted or discouraged human settlement. We will use these studies to focus the search for new sites in the Arabian sector, and develop systematic strategies for exploring submerged coastlines and coastal landscapes, which offered extensive and qualitatively different opportunities for human settlement when sea levels were lower throughout most of the past 2 million years, but which remain a huge area of ignorance. We will use high resolution studies of coastal shell mounds and shoreline geomorphology in recent millennia to guide the search for and interpretation of traces of coastal settlement in earlier periods and underwater. The result will be a new framework within which to synthesise the archaeological, palaeoclimatic and biological data relating to human dispersal, and a decisive breach of two longstanding beliefs that have obstructed advances in understanding, that ancient landscape settings cannot be reconstructed in regions of rapid geological change, and that exploration of submerged landscapes is too costly and unrewarding to be worth taking seriously.