In southern Africa, Mapungubwe (AD 1200–1300) is regarded as the earliest precolonial state system that emerged by exploiting floodplain ecologies for extensive agriculture and trade for immense surplus wealth, significant pillars that helped transforming kinships into bureaucratic, class-based societal organisations. By the end of 13th century AD, significant climate related changes associated with unpredictable rainfall trends and regional drying would have triggered Mapungubwe’s decline (Huffman 1996, 2000; Meyer 1998). This model is highly critiqued because its conception solely relies on archival research and insufficient palaeo-environmental records. Several studies show that, within major climatic events worldwide, landscapes often experience different conditions at the local and regional scale. By approaching Mapungubwe in its regional and local context, the Shashe-Limpopo basin, this thesis develops the first, multi- scalar analysis of past hydrological changes in relation to the evolution of complex societies in southern Africa by using an integrated geoarchaeological approach that combines statistical modelling of past rainfall records, geochemical analysis, soil micromorphology, and computational modelling. Analysis of 19th century rainfall records and hydrological interfaces (Geographic Information System, Remote Sensing and Hydrologic Engineering Center River Analysis System) suggests that flooding linked to climate variability would have been an imminent threat. Floodplain soil records and Optically-Stimulated Luminescence-dated deposits (AD 1237–1697) further suggest that changing environmental conditions did not necessary lead to the abandonment of different parts of Mapungubwe landscape. Rather, agricultural groups responded to climatic and environmental changes by reworking their landscape as an alternative management strategy. Accordingly, this thesis argues that prolonged occupation across the middle Limpopo valley would have been associated with comprehensive human resilience and adaptation strategies to changing floodplain ecologies. Moreover, it urges that we consider alternative pillars of state formation (e.g. competition and viable catchment dynamics) which may have offered institutional supremacy to reinforce socio-political ideologies along the Shashe- Limpopo basin. As such, this study contributes a new multidisciplinary theoretical and methodological framework to decode the long-term occupation history, particularly concerning resource use and sustainability in southern Africa with resonance for other drylands around the world.