Science & Research

This work demonstrates that MethaneAIR observations can be used to detect and quantify methane plumes from Concentrated animal feeding operations (CAFOs), showing that agricultural methane emissions are highly variable and can exceed inventory estimates under detectable conditions. By combining targeted scene selection, wavelet denoising, and divergence-integral flux estimation, the study highlights the value of airborne and satellite measurements for improving agricultural methane monitoring and emissions verification.

Methane point source emissions can be difficult to detect in satellite imagery when the signals are weak and hidden by background noise. This work developed an automated method for methane plumes detection that improves the visibility of plumes while reducing false detections, decreasing the need for time-consuming manual inspection. The method identifies more low-volume emissions across MethaneSAT and MethaneAIR, helping build a more complete understanding of methane emission distributions.

A new study used satellite data from MethaneSAT collected across multiple days and regions in the United States, Mexico, Turkmenistan, Uzbekistan, Iran, and Iraq to quantify and map oil and gas methane emissions at high-resolution. By scanning vast areas in fine detail, researchers found that methane emissions vary not only between regions, but also between counties and districts within the same oil and gas basin. These findings give policymakers a more precise, localized view of emissions hotspots, helping identify where the industry should prioritize repairs to more effectively slow climate warming.

Note: This is a preprint, and we expect the final published version to become available soon.

This study advances MethaneSAT’s ability to automatically detect and segment methane plumes by combining deep learning, cross-sensor transfer learning, and physics-informed postprocessing. By leveraging MethaneAIR observations to overcome limited labeled satellite data, the team developed an instance-segmentation framework that reliably identifies individual methane plumes and supports both high-sensitivity emissions screening and high-confidence source attribution. These improvements strengthen MethaneSAT’s capacity to map methane emissions and provide actionable insights for climate mitigation efforts.

Note: This is a preprint, and we expect the final published version to become available soon.

New analysis of MethaneSAT data reveals that methane leaks from global oil and gas operations are roughly 60% higher than what companies and governments have previously reported. Results from this study show that certain regions in North America, the Middle East, and Central Asia are losing far more natural gas than others, making them prime targets for urgent repairs. Ultimately, current emissions levels are about ten times higher than international climate targets, highlighting a massive gap between current industry practices and global methane mitigation goals.