Enhancing Methane Plume Quantification via Anisotropic Diffusion

Methane is a potent greenhouse gas with a global warming potential significantly higher than carbon dioxide, making its detection and quantification crucial for climate change mitigation. Accurate estimation of methane emissions relies on the ability to detect and delineate plumes from satellite observations, which is challenging due to sensor noise, background variability, and the diffuse nature of many emission sources. Conventional plume retrieval techniques, such as thresholding and median filtering, often fail to preserve plume structure and underestimate emission rates. To address these limitations, this study proposes an anisotropic diffusion (AD)-based energy minimization framework that enhances methane plume images by suppressing background noise while preserving fine plume details. The proposed method is tested on two controlled release sites using Sentinel-2 shortwave infrared (SWIR) bands B11 and B12 and compared against the standard median filtering approach. The results demonstrate that the AD-enhanced plume images lead to higher Integrated Mass Enhancement (IME) values and more accurate emission rate estimates, showing closer agreement with in situ measurements than conventional methods.