Protecting civilians in inter-communal conflict through remote sensing of climate impacts

Inter-communal conflicts are known for their high toll on civilians. To enable a targeted deployment of peacekeepers and humanitarian aid, United Nations agencies and non-governmental organizations require reliable early warning systems that allow them to anticipate the time and place of future violent outbreaks. However, inter-communal conflict events occur sporadically and are difficult to anticipate. We propose the use of highly disaggregated remote sensing data on vegetation condition for understanding and modeling the data generating process of inter-communal conflict, and discuss how it contributes to extant research on the conflict-climate nexus, conflict prediction, and to operational early warning systems in the field. Drawing on the case of Mali, we empirically validate the predictive performance of vegetation condition, illustrating how it can improve forecasts of inter-communal conflict and inform anticipatory action.