A travers l'analyse de matériel macrobotanique (charbon et graines / fruits) issus de trois sites majeurs du Later Stone Age (LSA) d'Afrique australe, cette étude s'intéresse à reconstituer les conditions environnementales durant l'Holocène dans la région, en particulier au cours des deux à trois derniers millénaires. L'étude de ce matériel, accompagné de la constitution d'une collection moderne de références de la végétation du bassin Kgalagadi, vise à mieux cerner l'impact quel'environnement a pu avoir sur les sociétés humaines (subsistance, peuplements, etc.) à une période-clé du (LSA), qui marque l'émergence des pratiques pastorales dans la région. L'étude des environnements est ici primordiale pour interpréter les conditions qui ont pu favoriser l'arrivée de ces premiers éleveurs et leurs lieux d'installation. En complément, une approche ethnographique a été menée, en travaillant avec les communautés locales dans la région de l'Erongo en Namibie, qui peuvent aider à comprendre et reconstituer les pratiques passées d'utilisation de la végétation. / Through macrobotanical (wood charcoal and seeds / fruit) analysis from three major Later Stone Age (LSA) sites located in the Kgalagadi Basin in southern Africa, this study is interested in reconstructing the environmental conditions during the Holocene in the region, in particular in the last two to three millennia. Initially this work couples archaeological macrobotanical analysis with the construction of modern reference material in an effort to trace the environmental / vegetal evolution as well as to comprehend socio-ecological and socio-environmental dynamics in Southern Africa during the late Holocene. The scope of this study covers the period spanning the last 3000-2000 years, with the main objective of understanding what relationships humans had with their environment at a time linked with the arrival or the appearance of the first herding practices in that part of Africa. These herding practices are believed to be accompanied by significant human movement from eastern or central Africa southwards. Favourable environmental conditions may have influenced their routes as well as settlement choices, and these are aspects that this archaeobotanical study aims to address. This study also employed an ethnographic approach, working with local communities in the Erongo region of Namibia, so as to make inferences to past vegetation utilisation practices while at the same time discerning and reconstituting past human activities.

Scholars of the Late Pleistocene in Southern Africa have recently sought to develop models explaining long-term variation between Middle Stone Age and Late Stone Age assemblages in terms of variability between “macrolithic” vs. “microlithic” toolmaking systems associated with shifts in hunter-gatherer ecology and land use patterns. While it has often proven extremely difficult to actually test many models, recently developed methods allow us to do so in novel ways. In this dissertation, I use new archaeological data from excavations of two sites in the Namib Desert, as well as new approaches to sourcing lithic artifacts to examine the hypothesis that contrasts between terminal Pleistocene (ca. 15-20 ka BP) and early Holocene (ca. 6-12 ka BP) occupation phases at the two sites represent adaptive responses primarily driven by changes in fluvial regimes and the resource productivity of riparian corridors. Analyzing the lithic assemblage compositions and locating probable source areas for raw materials suggests that terminal Pleistocene groups likely centered land use strategies more toward upland areas east of the study sites and periodically followed broad riparian corridors into the desert itself. Early Holocene groups expanded their ranges and more intensively targeted resources on the open desert plains, dunes, and beaches of the coastal lowlands. My results suggest environmental change may be partially responsible for driving this shift, but new data and methodological tools are needed to address factors like fluctuations in regional population size that may have been driving shifts in the late Pleistocene record of this unique region of Southern Africa.