Using satellite data and machine learning to study conflict-induced environmental and socioeconomic destruction in data-poor conflict areas: The case of the Rakhine conflict

Abstract This paper studies socioeconomic and environmental changes in the neighboring areas Bangladesh-Myanmar border from 2012 to 2019, thus covering period before after 2017 Rakhine conflict Myanmar outflux of refugees across Bangladesh. Given scarcity costliness traditional data collection methods such areas, uses a novel methodological model based on very-high-resolution satellite imagery, nighttime machine-learning algorithms generate reliable reusable for comparative assessment impacts conflict. Assessments welfare risks using this approach can be accurate scalable different regions times when other are unavailable. Key findings are: general livelihood situation has worsened income sources shrunk Rakhine; forced migration damaged ecologically fragile two countries; destruction aquaculture wetland ecosystems is observed deforestation rate reached 20% 13% Bangladeshi side border. The results provide guidance policymakers international actors as they work repatriate victims minimize conflict’s security consequences. methodology applied data-poor refugee world.