Remote sensing monitoring of point source methane column concentration in typical coal mining areas of Anhui Province

Methane is a radioactive and chemically reactive gas in the atmosphere, affecting both the Earth's radiation balance and atmospheric chemical processes. As the second largest greenhouse gas, its annual average concentration is more than 2.5 times that before the Industrial Revolution. Methane's global warming potential is approximately 84 times that of carbon dioxide on a 20-year timescale and about 30 times on a 100-year timescale. Methane's atmospheric lifetime is only about 12 years, far shorter than carbon dioxide's (which lasts for hundreds of years). Controlling methane emissions has a faster effect on mitigating global warming and is particularly important for achieving carbon neutrality.

The working principle of methane monitoring satellites is to estimate the emission of methane by detecting the spectral absorption of methane at different wavelengths using greenhouse gas monitoring payloads. Inversion is a top-down method that combines observational data and atmospheric transport models to optimize emissions (Cai Xiaoli, 2025). CH4 inversion methods mainly include atmospheric chemical transport model algorithms, physical algorithms, multi-band multi-channel methods, and deep learning algorithms.

This paper takes 19 typical coal mining areas in Anhui Province as the research object. Based on AHSI hyperspectral remote sensing data from the Gaofen-5 02 satellite (GF5B), the matched filtering algorithm is used to carry out remote sensing monitoring of methane column concentration in typical coal mining areas, aiming to provide data support and scientific basis for the dynamic supervision of regional methane emission sources.

During the Carboniferous-Permian period, the North China Plate underwent tectonic evolution, developing multiple coal-bearing strata and forming large coal bases such as the Huaihe River Basin and western Shandong. The Huaihe River Basin, spanning five cities in northern Anhui, holds coal reserves of 33.817 billion tons and is an important energy base in East China. Geologically, it is located in the Xuhuai Block on the southeastern edge of the North China Plate, divided into the Huaihe River Basin and the Huaihe River South Coalfield by the Bengbu Uplift.

Figure 1. Tectonic outline of the North China Plate, tectonic location of the Lianghuai Coalfield, and distribution of abandoned coal mines (Jiang Jianming, 2024).

Figure 2. Remote sensing image of the study area and distribution map of mineral deposits.

Matched filtering is a data-driven method where the spectral information can be characterized as a superimposed spectral perturbation on the average spectrum, caused by variations in methane column concentration. This algorithm maximizes the signal-to-noise ratio and is commonly used for methane point source remote sensing inversion. Based on the Beer-Lambert theorem, it is linearized using a first-order Taylor expansion:

Where k is the unit methane absorption coefficient, characterizing the ability of atmospheric methane to absorb radiation at a specific wavelength. That is, the sum of the absorption cross-sections of methane molecules in the vertical direction, which is determined by both the absorption path length and the methane absorption cross-section._theThe average radiation intensity μ received by the satellite can be used as an approximation, i.e., the reference spectrum. The target spectrum t(I)_theThe value represents the perturbation of background radiation intensity caused by the increased methane column concentration. The coefficient k can be obtained through a radiative transfer model.

Assuming that the spectral change is caused only by the change in methane concentration and that the absorption characteristics of methane do not change, the methane column concentration enhancement can be obtained by performing Gaussian log-likelihood on equation (3-1) and optimizing the target formula, as shown in equation (3-2):

In the formula, C is the background covariance matrix, while μ approximately expresses I._theThe formula for calculating t(μ) is t(μ) = μ·k

Remote sensing monitoring of methane column concentration in 19 typical coal mining areas of Anhui Province was conducted using AHSI data from the Gaofen-5 02 satellite and matched filtering (MF). The spatial distribution map of methane column concentration enhancement and the spatial distribution map of plume in some coal mining areas are shown below:

Figure 3. Spatial distribution of methane column concentration enhancement in a typical coal mining area.

Figure 4. Location of Dingji Coal Mine and Spatial Distribution of Methane Plume

Figure 5. Location of Pansan Coal Mine and Spatial Distribution of Methane Plume

Figure 6. Location of Zhujidong Coal Mine and Spatial Distribution of Methane Plume

This study used AHSI hyperspectral remote sensing data from the Gaofen-5 02 satellite (GF5B) and employed a matched filtering algorithm to conduct remote sensing monitoring of methane column concentrations in 19 typical coal mining areas in Anhui Province. However, due to imperfections in the accuracy verification system and interference from complex meteorological conditions, the remote sensing inversion of methane point source column concentrations exhibits significant uncertainties, necessitating further in-depth research.

References

[1] Cai Xiaoli, Bao Yunfei, Huang Qiaolin. A review of atmospheric methane detection and inversion technology based on satellite remote sensing [J]. Space Return and Remote Sensing, 2025, 46(01):160-173.

[2] Jiang Jianming, Han Feng, Ding Hai, et al. Comprehensive survey, development and utilization status and prospect of remaining resources in abandoned coal mines in Lianghuai Coalfield, Anhui Province [J]. China Mining, 2024, 33(08):46-58.

[3] Li Fei, Sun Shiwei, Zhang Yongguang, et al. Remote sensing inversion and analysis of typical methane super emission sources in China and the United States by the hyperspectral imager of Gaofen-5 02 satellite [J]. Journal of Remote Sensing, 2024, 28(08):1986-2001.

[4] Qin Kai, He Qin, Kang Hanshu, et al. Research progress and prospect of satellite remote sensing of methane emissions in coal industry [J]. Acta Optica Sinica, 2023, 43(18):118-130.

[5] Jiang Yuhan. Performance study of satellite remote sensing detection of methane point sources [D]. Chinese Academy of Meteorological Sciences, 2025.