Long-Range Transport of Coal Mine Methane Emissions Causes Source Mis-Attribution: A Satellite-Based Estimation Approach Incorporating Data Uncertainty

Accurately estimating and attributing methane (CH4) emissions is critical for climate change analysis and mitigation policy, but complex meteorology and terrain and incomplete knowledge of source geospatial distrubiton challenges current satellite-based and transport model-based methods. One of the largest coal mining regions by density is found in the high-elevation and mountainous regions of Shanxi China, and represents an ideal laboratory to quantify how emitted CH4 from such regions transports into the free troposphere and is subsequently mis-attributed downwind. We employ Empirical Orthogonal Function (EOF) analysis on five years of daily TROPOMI satellite observations, and WRF-STILT simulation to reveal the most important spatial-temporal causes of changes in methane concentration in the areas overlapping with the high coal mine methane in Shanxi and subsequent urban and agricultural downwind regions in Henan. Then we use ground observation data as physical constraints to construct a physics-based lightweight methane emission method to estimate methane emissions, while explicitly considering the impact of satellite data uncertainty on the emission results.

This work found that the contribution of the global background methane concentration to the methane in this region is approximately 40% and matches well with the upward trend modulated by seasonal cycles. Next this work identified a local mode on 5.3% of days that reveals a slow build-up and rapid release of CH4 from mining areas to middle-tropospheric loadings over downwind agricultural areas. During the 32 days of most substantial atmospheric transport, 0.064Mt of coal mine methane emissions slowly built up over basins in Shanxi and were transported over agricultural areas of Henan, accounting for over two thirds of the net 0.10Mt downwind increase. Applying our light-weight physically constrained emissions framework properly identifies the sources in these regions and effectively filters these observed long-range transported events, enabling more reliable emission estimates from existing satellite data. Failure to filter these events will lead to a substantial underestimation of fossil-fuel methane sources, since current isotopic constraint approaches do not sample middle tropospheric air.