This review analyses the environmental evolution of the Ethiopian highlands in the late Quaternary. The late Pleistocene (20,000-12,000 C-14 years BP) was cold and dry, with (1) low lake levels in the Rift Valley, (2) large debris fans on the flanks of Lake Abhe basin, and (3) the Blue Nile transporting coarse bedload. Then, a period with abundant and less seasonal rains existed between 11,500 and 4800 C-14 years BP, as suggested by increased arboreal pollen, high river and lake levels, low river turbidities and soil formation. Around 5000-4800 C-14 years BP, there was a shift to more and conditions and more soil erosion. Many phenomena that were previously interpreted as climate-driven might, however, well be of anthropic origin. Thick sediment deposits on pediments as well as an increase of secondary forest, scrub and ruderal species in pollen diagrams are witnesses of this human impact. One important aspect of the late Quaternary palaeoenvironment is unclear: changes in Nile flow discharges and Rift Valley lake levels have been linked to changes in precipitation depth. Most authors do not take into account changes in land use in the highlands, nor changes in the seasonality of rain, both of which can lead to a change in runoff coefficients. Tufa and speleothem deposition around 14,000 years ago tend to indicate that at the end of the Last Glacial Maximum (LGM), conditions might have been wetter than generally accepted. The most important present-day geomorphic processes are sheet and rill erosion throughout the country, gullying in the highlands, and wind erosion in the Rift Valley and the peripheral lowlands. Based on existing sediment yield data for catchments draining the central and northern Ethiopian highlands, an equation was developed allowing to assess area-specific sediment yields: SY = 2595A(-0.29) (n = 20; r(2) = 0.59) where SY = area-specific sediment yield (t km(-2) year(-1)), and A = drainage area (km(2)). With respect to recent environmental changes, temporal rain patterns, apart from the catastrophic impact of dry years on the degraded environment, cannot explain the current desertification in the driest parts of the country and the accompanying land degradation elsewhere. Causes are changing land use and land cover, which are expressions of human impact on the environment. Deforestation over the last 2000-3000 years was probably not a linear process in Ethiopia. Studies on land use and land cover change show, however, a tendency over the last decades of increasing removal of remnant vegetation, which is slowed down or reversed in northern Ethiopia by a set-aside policy. Ongoing land degradation requires urgent action at different levels of society. Soil and water conservation (SWC) structures are now widely implemented. Local knowledge and farmer's initiatives are integrated with introduced SWC techniques at various degrees. Impact assessments show clear benefits of the soil conservation measures in controlling runoff and soil erosion. In high rain areas, runoff management requires greater emphasis during the design of soil conservation structures. In such areas, investment in SWC might not be profitable at farm level, although benefits for society are positive. This pleads in favour of public support. The present land degradation in the Ethiopian highlands has a particular origin, which includes poverty and lack of agricultural intensification. Causes of these are to be found in the nature of past and present regional social relations as well as in international unequal development. This review strengthens our belief that, under improved socio-economic conditions, land husbandry can be made sustainable, leading to a reversal of the present desertification and land degradation of the Ethiopian highlands. (C) 2003 Elsevier B.V. All rights reserved.
