top of page
Search

Can Methane Emissions & Mining Actions be Correlated?

  • john06025
  • Apr 3
  • 5 min read

GeoSynergy has been actively involved in the development, and evaluation, of wind-rotated TROPOMI CH4 [1] for the assessment of methane emissions over coal mines, amongst other facilities. One such example involved the assessment of a coal mine cluster in NSW (Fig.1). In this case, background CH4 has been subtracted, using a bounding box to the south of the wind-rotated product. Even at the monthly level of granularity, we can see significant CH4 emissions above background.

Fig.1 Wind-rotated TROPOMI CH4 over a cluster of coal mines in NSW.
Fig.1 Wind-rotated TROPOMI CH4 over a cluster of coal mines in NSW.

We were interested in whether the observed emissions could be correlated with mine actions also visible from space. We have had some success in using wind-rotated TROPOMI CH4 to identify approximate point sources. As described by Maasakkers et al. TROPOMI products are aggregated over a fine grid of rotation centers, and the point source is then inferred from the optimal enhancement.


However, in the last few years, new methane sensing products have emerged, which can detect small instantaneous emissions from coal mines (and other sites) at high resolution. This is clearly advantageous, over time-aggregates, in the current context.


Tanager-1, launched on August 16, 2024, is a hyperspectral imaging satellite with a spatial resolution of 30m, a revisit time of ~3.5 days, and a methane detection threshold as low as ~50 kg per hour [2] (Fig.2).

Fig.2 Examples of high resolution instantaneous methane emissions from coal mines in Australia, using the new Tanager-1 product [3]. Top Left: 500 kg/h. Top Right: 1.1k kg/h. Bottom Left: 1.2k kg/h. Bottom Right: 919 kg/h. Quantification is preliminary and provided by CarbonMapper [2, 3].
Fig.2 Examples of high resolution instantaneous methane emissions from coal mines in Australia, using the new Tanager-1 product [3]. Top Left: 500 kg/h. Top Right: 1.1k kg/h. Bottom Left: 1.2k kg/h. Bottom Right: 919 kg/h. Quantification is preliminary and provided by CarbonMapper [2, 3].

Another relevant satellite product is ISS EMIT. Mounted on the International Space Station, EMIT’s hyperspectral sensor detects emissions as low as ~50–100 kg CH₄ per hour, at 60 m resolution [4]. While its spectral precision improves methane isolation, its dust-focused mission design and irregular revisit times (3–16 days) limit its utility.  


We screened opencast coal mines, within a NSW cluster, for instantaneous methane emissions, and found the following example of an EMIT product (Fig.3).

Fig.3 ISS EMIT methane emission from a coal mine in NSW, 2023.01.25 [3]. Not quantified.
Fig.3 ISS EMIT methane emission from a coal mine in NSW, 2023.01.25 [3]. Not quantified.

Focusing on the period around June 2023, in the first instance, we reviewed Sentinel-2 products. Sentinel-2 products have a spatial resolution of 10m in the RGB channels, and a revisit time of ~5 days (Fig.4).

Fig.4 Sentinel-2 time series over the AOI. Top Left: ISS EMIT methane emission over base map. Top Right to Bottom Right: Sentinel-2 RGB time series. Note the increasingly visible coal bed, as the time series progresses. The 24th Jan product contains a small cloud occlusion.
Fig.4 Sentinel-2 time series over the AOI. Top Left: ISS EMIT methane emission over base map. Top Right to Bottom Right: Sentinel-2 RGB time series. Note the increasingly visible coal bed, as the time series progresses. The 24th Jan product contains a small cloud occlusion.

Based on this time series, we can reasonably suggest that the EMIT methane emission corresponds with mining activity between 24th and 29th Jan 2023, during which we can see a coal seam being exposed. A Red channel time series, of closer crops and 10m resolution, is shown in Fig.5.

Fig.5 Sentinel-2 Red channel time series over the AOI. The yellow marker indicates the suggested mining activity, which may correlate with the observed EMIT emission.
Fig.5 Sentinel-2 Red channel time series over the AOI. The yellow marker indicates the suggested mining activity, which may correlate with the observed EMIT emission.

This provides useful intelligence. However, the available cloud-free Sentinel-2 products limit our ability to correlate the emission to mining actions within a 5 day window. In this case, we can do better using Landsat 8/9 products.


Launched on 11 February 2013, and 27 September 2021, respectively, Landsat 8 and 9 together have a combined revisit time of ~8 days, and a spatial resolution of 15m in the panchromatic band. For our AOI, we identified 5 Landsat products as follows, note the WRS-2 path and row:


  1. Landsat 9, path 89, row 83, 2023-01-09

  2. Landsat 8, path 89, row 83, 2023-01-17

  3. Landsat 8, path 90, row 82, 2023-01-24

  4. Landsat 9, path 89, row 83, 2023-01-25

  5. Landsat 8, path 89, row 83, 2023-02-02


Notably, because of the adjacent and overlapping pathing of Landsat 8 and 9, we were able to obtain products for both 24 and 25 January 2023. i.e. directly before, and on the day of, the observed methane emission (Fig.6).

Fig.6 a: Landsat products, which overlap over our AOI, Landsat 8 24th Jan (top), and Landsat 9 25th Jan (bottom). b and c: closer crops, showing the relative position of our AOI to the point of overlap.
Fig.6 a: Landsat products, which overlap over our AOI, Landsat 8 24th Jan (top), and Landsat 9 25th Jan (bottom). b and c: closer crops, showing the relative position of our AOI to the point of overlap.

Now, plotting the time series of Landsat 8/9 products, we can narrow the suggested source of the emission (Fig.7). i.e. a mining action that exposed the new, small region of coal visible in the 25th Jan product, but not visible in the 24th Jan product.

Fig.7 Time series of Landsat 8/9 products (panchromatic band, 15m resolution). We can see a small region of coal exposed, next to the pre-existing coal bed (yellow arrow).
Fig.7 Time series of Landsat 8/9 products (panchromatic band, 15m resolution). We can see a small region of coal exposed, next to the pre-existing coal bed (yellow arrow).

The above discussion raises the question of whether mining actions, and their corresponding emissions, could be modelled. Whilst intuitive that coal mines produce methane, there is also some existing supporting research that links various mine properties to emission profiles. For example, Global Energy Monitor's Global Coal Mine Tracker uses three factors to estimate CH4 emissions from coal mines: production, gas content at mining depth, and emission factor coefficient. Their work is based on a study by Kholod et al. [6].


An important research question is whether sufficiently fine-grained mining actions, and methane emissions, can be remote sensed. We have discussed the emergence of new methane-sensing satellite products above. In terms of remote sensing on-ground mining actions, there are a variety of products visible spectra products, from Sentinel-2 (10m spatial resolution), to the very high resolution WorldView-3 (0.3m) [7].


We might also consider InSAR products, which have long been used for change detection applications. Importantly, these products have the advantage of being cloud-invariant. InSAR products can also be used to produce DEMs, via interferometry, which can inform mining depth (a proposed predictor of emissions). Wang et al. have proposed using Sentinel-1 for coal mine activity monitoring [8], and commercial InSAR products, such as ICEYE, can of course produce superior DEMs, enabling more granular change detection [9].


A DEM of an opencast coal mine, derived from Sentinel-1 interferometry, is shown below (Fig.8).

Fig.8 DEM from Sentinel-1 interferometry of an opencast coal mine. Left: full region, AOI boxed in white. Right: crop to the AOI, note the elevation dip corresponding with the opencast coal mine (white arrow).
Fig.8 DEM from Sentinel-1 interferometry of an opencast coal mine. Left: full region, AOI boxed in white. Right: crop to the AOI, note the elevation dip corresponding with the opencast coal mine (white arrow).

In conclusion, the emergence of quantified, high-resolution methane satellite products may open the door to mine action/emission modeling. In the context of the National Greenhouse and Energy Reporting (NGER) framework, the ability to associate emission profiles, with specific mining activities, could significantly enhance strategic planning for stakeholders. Although not discussed here, correlating emission behavior with site characteristics naturally extended to other sectors.


References


[1] Maasakkers, J.D., Varon, D.J., Elfarsdóttir, A., McKeever, J., Jervis, D., Mahapatra, G., Pandey, S., Lorente, A., Borsdorff, T., Foorthuis, L.R. and Schuit, B.J., 2022. Using satellites to uncover large methane emissions from landfills. Science Advances, 8(31), p.eabn9683. https://www.science.org/doi/full/10.1126/sciadv.abn9683



[3] Carbon Mapper open data portal. https://carbonmapper.org/about 


[4]. Thorpe, A.K. et al. [2023] ‘EMIT: New capabilities for methane detection from the International Space Station’, Geophysical Research Letters, 50[4], p. e2023GL103713. 



[6] Kholod, N., Evans, M., Pilcher, R.C., Roshchanka, V., Ruiz, F., Coté, M. and Collings, R., 2020. Global methane emissions from coal mining to continue growing even with declining coal production. Journal of Cleaner Production, 256, p.120489.


[7] WorldView-3 satellite product specification. https://earth.esa.int/eogateway/missions/worldview-3


[8] Wang, L., Yang, L., Wang, W., Chen, B, 2021. Monitoring mining activities using Sentinel-1A InSAR coherence in open-pit coal mines. Remote Sensing, 13(21), p.4485.


[9] ICEYE product specification. https://sar.iceye.com/latest/

 
 
 

Comments

bottom of page