Millets and sorghum are amongst the most common drougth-tolerant crops, central to the livelihhod of c. 600 milions people living in drylands. To date, our understanding of the role of these cereals in past economies has been hindered by preservation issues and the limited number of systematic archaeobotanical research programs. By building predictive models that combine published ethnographic literature and environmental datasets on a global level, we were able to generate hypotheses about agricultural systems in drylands. Our ecological modelling of traditional agricultural systems has revealed that the relationship between annual precipitation and agricultural viability is not as strong as previously considered. Other factors such as growing cycles duration, soil nutrient availability and water holding capacity appear to be much more determinant in shaping traditional agroecosystems.
The ability of the models to predict local agricultural practices was tested against ethnoarchaeological observations collected in various dry areas (Ethiopia, Sudan, Pakistan). Archaeobotanical data from archaeological sites, was then used to assess the model’s predictions when applied to the archaeological record. According to our results, rainfed agriculture of drought- resistant crops was already in place earlier than previously thought and supported the economy of complex societies. Our work forwards the understanding of how human communities developed long-term sustainable, resilient agricultural strategies and challenges mianstream views of dryland adaptation. This is especially significant in the current context of climate instability and increasing population, which calls for immediate action.
