Description:
Environmental research related to military activities and warfare is sparse and fragmented by discipline. Although achieving military objectives will likely continue to trump any concerns related to the environment during active conflict, military training during peacetime has environmental consequences. This research aims to quantify how much pollution is emitted during regular military exercises which has implications for climate change. Focusing on major military training exercises conducted in Australia, we assess the impact of four international exercises held within a dedicated military training area on pollution levels. Leveraging high-frequency data, we employ a machine learning algorithm in conjunction with program evaluation techniques to estimate the effects of military training activities. Our main approach involves generating counterfactual predictions and utilizing a "prediction-error" framework to estimate treatment effects by comparing a treatment area to a control area. Our findings reveal that these exercises led to a notable increase in air pollution levels, potentially reaching up to 25% relative to mean levels during peak training hours.