Automated detection and monitoring of methane super-emitters using satellite data
Abstract. A reduction in anthropogenic methane emissions is vital to limit near-term global warming. A small number of so-called super-emitters is responsible for a disproportionally large fraction of total methane emissions. Since late 2017, the TROPOspheric Monitoring Instrument (TROPOMI) has been in orbit providing daily global coverage of methane mixing ratios at a resolution of up to 7 × 5.5 km2, enabling the detection of these super-emitters. However, TROPOMI produces millions of observations each day, which together with the complexity of the methane data, makes manual inspection infeasible. We have therefore designed a two-step machine learning approach using a Convolutional Neural Network to detect plume-like structures in the methane data and subsequently apply a Support Vector Classifier to distinguish emission plumes from retrieval artefacts. The models are trained on pre-2021 data, and subsequently applied to all 2021 observations. We detect 2974 plumes in 2021 with a mean estimated source rate of 44 t h−1 and 5–95th percentile range of 8–122 t h−1. These emissions originate from 94 persistent emission clusters and hundreds of transient sources. Based on bottom-up emission inventories, we find that most detected plumes are related to urban areas/landfills (35 %), followed by plumes from gas infrastructure (24 %), oil infrastructure (21 %) and coal mines (20 %). For twelve (clusters of) TROPOMI detections, we "tip-and-cue" targeted observations and analysis of high-resolution satellite instruments to identify the exact sources responsible for these plumes. Using high-resolution observations from GHGSat, PRISMA and Sentinel-2, we detect and analyze both persistent and transient facility-level emissions underlying the TROPOMI detections. We find emissions from landfills and fossil fuel exploitation facilities, for the latter we find up to ten facilities contributing to one TROPOMI detection. Our automated TROPOMI-based monitoring system in combination with high-resolution satellite data allows for the detection, precise identification and monitoring of these methane super-emitters, which is essential for mitigating their emissions.
Berend J. Schuit et al.
Status: final response (author comments only)
- RC1: 'Comment on acp-2022-862', Anonymous Referee #1, 21 Feb 2023
- RC2: 'Comment on acp-2022-862', Anonymous Referee #2, 06 Mar 2023
- AC1: 'Comment on acp-2022-862', Berend Schuit, 21 Apr 2023
Berend J. Schuit et al.
Berend J. Schuit et al.
Viewed (geographical distribution)
This paper describes a two-step machine learning approach that uses a Convolutional Neural Network (CNN) to detect plume-like structures in TROPOMI methane data and then applies a Support Vector Classifier (SVC) to distinguish emission plumes from retrieval artefacts. The CNN is trained using hand-selected scenes from 2018-2020 and then applied to 2021 observations. This is an important topic because TROPOMI collects millions of measurements over the globe each day, and future missions will collect even more. Automated approaches are therefore needed to process these data and reliably identify emission plumes. In general, this manuscript does represent a substantial contribution. However, the methods section (section 2.2, 2.3) needs a substantial revision to make it more understandable to the average reader of Atmospheric Chemistry and Physics, who may not be familiar with these machine learning techniques.
The description of the choice and configuration of the CNN is very short. Four reasons are cited to justify the choice of this particular machine-learning method (L131-136), but readers not familiar with these methods might not know that CNNs are commonly chosen for image recognition and pattern recognition. We learn that “the same convolutional kernel scans the entire image”, but never told what this kernel is or where it comes from or what a “pooling layer” (figure 2) is or does.
The description of the CNN training process is even more obscure and confusing. We are told that “For the training process, the class weight parameter is set to the ratio between the number of plumes (828 positives) and negatives (2242), …”, but we are never told what the class weight parameter is or how sensitive the solution might be to this setting. The paragraph that follows (L148-163) makes the training process look more like black magic, where the user utters a few magic words (Keras, ReLU, ADAM, softmax) and wondrous things happen. All of these terms are used without reference to refereed scientific papers. Instead, the reader is sent to a web page (Chollet et al., 2015) with a sales pitch and code and then a github site (O’Malley et la.. 2019). What part of these code distributions are used here? All of them? The only real reference in this paragraph is Li et al. (2018), which describes one of two approaches used for optimizing hyperparameters. Neural network training is a major of this paper. Additional insight into these methods is essential to gain the acceptance and understanding of this Earth Science audience. At a minimum, we need to understand the specific inputs and outputs of these methods and how the results are validated against standards. A few additional figures illustrating these topics would be great.
The discussion of Feature Engineering (Section 2.3, L193-199) and list of features in Table C1 is more helpful, but still unnecessarily confusing. I was surprised that the feature vector included the CNN score (0, 1) as feature, that is apparently no less or more important than any other. We are told that the algorithm operates on 32x32 pixel scenes using a 41 x 1 feature vector. However, then we learn (L205) that “In our binary classification problem, the CAM visualizes which regions of the deepest feature maps(the 8x8, deepest max-pooling layer in Figure 2) lead to an activation of the plume class.” Figure 2 shows only two pooling layers and does not mention where the 8x8, deepest max-pooling layer. We are then told (L207-208) that “This spatial activation is calculated using the gradients between all internal 64 feature maps and the fully-connected layer.” Where did we learn about internal 64 feature maps? At this point, I was totally lost.
Smaller issues, concerns and editorial suggestions:
L66: “As such the hyperspectral …” à “As such, the hyperspectral …”
L78: “atmospheric conditions and identify plume signatures” à “atmospheric conditions to identify plume signatures …”?
L85: “we target three high-resolution satellite instruments …” The word “target” is ambiguous here, because it sometime means that you point a satellite instrument (i.e., GHGSat) at a target. From the context, I believe you mean “we use data from three high-resolution satellite instruments …” Is that correct?
L90: “We use two 90 machine learning models in sequence to detect plumes in the TROPOMI methane data. First we apply a Convolutional Neural Network to detect plume-like structures in TROPOMI methane atmospheric mixing ratio data, then we use additional atmospheric parameters and supporting data to further distinguish between genuine methane plumes and retrieval artefacts. We then use (targeted) high-resolution methane observations to pinpoint the responsible sources.”
Sections 2.2 – 2.4 – see comments above.
L314: “allows multiple close by plumes” à “allows multiple nearby plumes”
L345, L358, L371. It appears that the same, 10m wind field is used in the analysis of GHGSat, PRISMA, and sentinel-2, but two different notations are used. For GHGSat (L345), it is called with U10 winds from GEOS-FP, while for the other two satellites, it is called “GEOS-FP 10m wind data”. It would be good to use consistent nomenclature.
L351: “hyperspectral 30x30 km2 images at a spatial resolution of 30x30 m …” If the pixels and images are nearly square, it would be better to describe their dimensions as 30 km x 30 km and 30 m x 30 m, respectively.
L352: “The minimum revisit time can be up to 7 days with ±20% across-track pointing.”
L356: “location of interest on a future moment in time.” à “location of interest in the future.”
L361: “capable of the detection of methane” à “capable of detecting methane”
L362: “with a pixel resolution of 20 m” à is this “with a pixel resolution of 20 m x 20 m”?
L369: “that similarly to Varon et al. (2021)” à “", that, like Varon (2021) uses ..."
L375: “identifies 26,444 scenes (3.3 %) as containing” à “identifies 26,444 scenes (3.3 %) that contain”