Quantum Computing for Monitoring Methane Emissions Author

Methane is one of the most important anthropogenic greenhouse gases, and timely detection of methane emissions is essential for climate monitoring, mitigation policy, and operational leak detection. In recent years, satellite-based Earth observation has become an increasingly relevant source of information for this purpose, as it enables large-scale monitoring of emission sources. At the same time, extracting actionable information from such data remains computationally challenging. For example, atmospheric inverse modeling and plume detection algorithms require methods that can handle high-dimensional inputs, temporal dependencies, and limited availability of labeled real-world examples. These characteristics make methane monitoring a useful application domain in which to assess whether emerging computational paradigms, including quantum computing, can offer practical added value.Quantum computing is often considered as a promising route for accelerating selected computational primitives, such as linear algebra and optimization. However, in applied machine learning, the relevant question is not whether a theoretical speedup or improved learning capability exists, but whether it survives under realistic data-access, training, and output requirements. For that reason, this project evaluated where quantum approaches might plausibly contribute within the methane-monitoring pipeline. Three candidate challenges were considered in this pipeline. The first is the updating of methane emissions inventories through inverse modeling. The second is synthetic data generation. The third is plume detection from multitemporal Sentinel-2 imagery. Following this initial assessment, a challenge was selected for quantum solution development and benchmarking. This report summarizes experimental results, insights, and recommendations for future work.