Geoffrey King

We examine the links between geomorphological processes, specific landscape features, surface water drainage, and the creation of suitable habitats for hominins. The existence of mosaic (i.e., heterogeneous) habitats within hominin site landscape reconstructions is typically explained using models of the riverine and gallery forest settings, or the pan or lake setting.We propose a different model: the Tectonic Landscape Model (TLM), where tectonic faulting and volcanism disrupts existing pan or river settings at small-scales (∼10e25 km). Our model encompasses the interpretation of the landscape features, the role of tectonics in creating these landscapes, and the implications for hominins. In particular, the model explains the underlying mechanism for the creation and maintenance of heterogeneous habitats in regions of active tectonics. We illustrate how areas with faulting and disturbed drainage patterns would have been attractive habitats for hominins, such as ustralopithecus, and other fauna. Wetland areas are an important characteristic of surface water disturbance by fault activity; herefore we examine the tectonically-controlled Okavango Delta (Botswana) and the Nylsvley wetland (South Africa) as modern
examples of how tectonics in a riverine setting significantly enhance the faunal and floral biodiversity. While tectonic landscapes may not have been the only type of attractive habitats to hominins, we propose a suite of landscape, faunal, and floral indicators, which when recovered together suggest that site environments may have been influenced by tectonic and/or volcanic activity while hominins were present. For the fossil sites, we interpret the faulting and landscapes around australopithecine-bearing sites of the Middle Awash (Ethiopia) and Makapansgat, Taung, and Sterkfontein (South Africa) to illustrate these relationships between landscape features and surface water bodies. Exploitation of tectonically active landscapes may explain why the paleoenvironmental signals, anatomy, diets, as well as the fauna associated with Australopithecus appear largely heterogeneous through time and space. This hypothesis is discussed in light of potential preservation and time-averaging effects which may affect patterns visible in the fossil record. The model, however, offers insight into the landscape processes of how such habitats are formed. The landscape features and range of habitat conditions, specifically the wetter, down-dropped plains and drier, uplifted flanks persist in close proximity for as long as the fault motion continues. The Tectonic Landscape Model provides an alternative explanation of why mixed habitats may be represented at certain sites over longer timescales.

Climate shifts at decadal scales can have environmental consequences, and therefore, identifying areas that act as environmental refugia is valuable in understanding future climate variability. Here we illustrate how, given appropriate geohydrology, a rift basin and its catchment can buffer vegetation response to climate signals on decadal timescales , therefore exerting strong local environmental control. We use time-series data derived from Normalised Difference Vegetation Index (NDVI) residuals that record vegetation vigour, extracted from a decadal span of MODIS images, to demonstrate hydrogeological buffering. While this has been described previously it has never been demonstrated via remote sensing and results in relative stability in vegetation vigour inside the delta, compared to that outside. As such the Delta acts as a regional hydro-refugium. This provides insight, not only to the potential impact of future climate in the region, but also demonstrates why similar basins are attractive to fauna, including our ancestors, in regions like eastern Africa. Although vertebrate evolution operates on time scales longer than decades, the sensitivity of rift wetlands to climate change has been stressed by some authors, and this work demonstrates another example of the unique properties that such basins can afford, given the right hydrological conditions. Rift basins and associated wetlands may offer attractive habitats and their role in human evolution has been stressed by several authors 1–4. The Okavango Delta is the largest wetland in southern Africa and renowned for its high floral and faunal biodiversity 5. It covers an area of over 40,000 square km and consists of a smooth, (relief ≤2 m; slope 1:3400) conically-shaped, alluvial fan located on the Cubango-Okavango River which feeds a network of distributary channels and flanking swamps that form the Delta itself 6,7. While more accurately an alluvial megafan (or fan delta) rather than a delta we preserve the current terminology to avoid confusion. Located in the Kalahari Basin, the region has been subject to sedimentation throughout the Cenozoic, and was once at the centre of palaeo-lake Makgadikgadi 8–12 , prior to its drainage as a result of continued tectonic activity associated with the westward propagation of the East African Rift and river capture by the Zambezi River 10,13,14. The Delta is now located in an asymmetric graben, demarcated by a series of sub-parallel faults, including the Thamalakane Fault and is in-filled with at least 300 m of sediment deposited sometime in the last 1 Ma 10,15–17. Podgorski and colleagues 12 report a combination of electromagnetic, borehole and seismic data and suggest that this sediment fill consists of palaeo-megafans in the Okavango Basin overlain, by palaeo-lake Makgadikgadi deposits and the current delta/megafan. Recent renewed movements, post the palaeo-lake, along the Thamalakane and Kunyere faults now separate the Okavango and Makgadikgadi basins 10. A series of topographic cross-sections illustrate the geomorphological features of the Delta and wider basin. The Okavango catchment is dominated by aeolian sands and outcrops of weathered bedrock (Fig. 1). The annual sediment discharge within the Cubango-Okavango River consists mainly of bed and solute loads (bed: 170,000 tonnes; solute: 360,000 tonnes) while suspended sediment forms only a minor component (8 mg/l or 39,000 tonnes) 18,19. As a consequence of the high sediment porosity (very well-sorted with rounded grains) over 80 to 90% 20 of seasonal flood water infiltrates into the Delta from the Cubango-Okavango River with peak flows in the Panhandle in April and early May (Fig. 1). Much of this shallow water is then transpired by Delta vegetation as this slow moving flood wave takes over four months to reach the south eastern extremity of the Delta (i.e., in

ABSTRACT Recent activity of normal faults in Greece has produced steep limestone fault scarps at the base of the mountain fronts. For example, on the Sparta fault located in the Peloponnese, and responsible for the 464 B.C. M˜7 earthquake, a continuous fresh scarp cuts limestone bedrock and indurated conglomerates. The scarp is nearly continuous dipping at 65-68o with well-preserved slickensides. The maximum height of the scarp is 10-12 metres, progressively decreasing towards the ends. The few local variations are associated with active streams where hangingwall erosion causes the scarp to be locally higher. The regularity of the scarp is powerful evidence that the footwall and hangingwall surfaces were originally continuous and the scarp surface represents fault slip alone. It also suggests that there was no significant erosion or deposition on the hanging-wall (except near active gullies) since the scarp began to form. Such observations have led to the suggestion that numerous well-preserved limestone escarpments around the eastern Mediterranean, similar to those in Sparta, are post-glacial in age. Using 36Cl cosmogenic dating we tested the foregoing ideas. Limestones (largely calcite) contain an abundance of calcium, which is a major target element for cosmogenic 36Cl production. Samples were collected from the limestone scarp surface to recover the continuous exposure history of the scarp and also from the footwall and hanging wall surfaces as well as a depth profile in the hanging wall wedge. The concentration of 36Cl and of stable chlorine has been measured by accelerator mass spectrometry (AMS) at the LLNL-CAMS for each of the samples (about 150). The result confirm: 1- that the fault scarp was formed between about 2000 B.P. and 13 ka B.P. as the result of 5 earthquakes (including the known event in 464 B.C.) with similar slip amplitudes of about 2m and with time intervals ranging from 500yr to 4500yr (Benedetti et al., GRL, 2002), 2- that both hanging wall and footwall surfaces have exposure ages of 40-50 ka, and 3- that the colluvial wedge was formed during three episodes of deposition that occurred between 50 to 70 ka B.P. It is not clear whether those episodes, that are also stratigraphically recognizable, are seismically or climatically related. The climatic hypothesis is that during the cold and dry climate of the last glaciation the slopes were unstable so that the free-faces formed during repeated earthquakes were immediately buried. At the end of the glaciation, as a consequence of the wet and warm climate, slopes stabilized and earthquake slip began to accumulate and to form the present day cumulative escarpements.