Search

Under the Copernicus programme, an operational CO2 Monitoring Verification and Support system (CO2MVS) is being developed and will exploit data from future satellites monitoring the distribution of CO2 within the atmosphere. Methods for estimating CO2 emissions from significant local emitters (hotspots; i.e. cities or power plants) can greatly benefit from the availability of such satellite images that display the atmospheric plumes of CO2. Indeed, local emissions are strongly correlated to the size, shape, and concentration distribution of the corresponding plume, which is a visible consequence of the emission. The estimation of emissions from a given source can therefore directly benefit from the detection of its associated plumes in the satellite image. In this study, we address the problem of plume segmentation (i.e. the problem of finding all pixels in an image that constitute a city or power plant plume). This represents a significant challenge, as the signal from CO2 plumes induced by emissions from cities or power plants is inherently difficult to detect, since it rarely exceeds values of a few parts per million (ppm) and is perturbed by variable regional CO2 background signals and observation errors. To address this key issue, we investigate the potential of deep learning methods and in particular convolutional neural networks to learn to distinguish plume-specific spatial features from background or instrument features. Specifically, a U-Net algorithm, an image-to-image convolutional neural network with a state-of-the-art encoder, is used to transform an XCO2 field into an image representing the positions of the targeted plume. Our models are trained on hourly 1 km simulated XCO2 fields in the regions of Paris, Berlin, and several power plants in Germany. Each field represents the plume of the hotspot, with the background consisting of the signal of anthropogenic and biogenic CO2 surface fluxes near to or far from the targeted source and the simulated satellite observation errors. The performance of the deep learning method is thereafter evaluated and compared with a plume segmentation technique based on thresholding in two contexts, namely (1) where the model is trained and tested on data from the same region and (2) where the model is trained and tested in two different regions. In both contexts, our method outperforms the usual segmentation technique based on thresholding and demonstrates its ability to generalise in various cases, with respect to city plumes, power plant plumes, and areas with multiple plumes. Although less accurate than in the first context, the ability of the algorithm to extrapolate on new geographical data is conclusive, paving the way to a promising universal segmentation model trained on a well-chosen sample of power plants and cities and able to detect the majority of the plumes from all of them. Finally, the highly accurate results for segmentation suggest the significant potential of convolutional neural networks for estimating local emissions from spaceborne imagery.

Dataset and weights release for preprint "Segmentation of XCO2 images with deep learning: application to synthetic plumes from cities and power plants" Joffrey Dumont Le Brazidec et al.

Urban areas, where more than 55% of the global population gathers, contribute more than 70% of anthropogenic fossil fuel carbon dioxide (CO_(2ff)) emissions. Accurate quantification of CO_(2ff) emissions from urban areas is of great importance for formulating global warming mitigation policies to achieve carbon neutrality by 2050. Satellite-based inversion techniques are unique among “top-down” approaches, potentially allowing us to track CO_(2ff) emission changes over cities globally. However, their accuracy is still limited by incomplete background information, cloud blockages, aerosol contamination, and uncertainties in models and priori emission inventories. To evaluate the current potential of space-based quantification techniques, we present the first attempt to monitor long-term changes in CO_(2ff) emissions based on the OCO-2 satellite measurements of column-averaged dry-air mole fractions of CO₂ (X_(CO₂)) over a fast-growing Asian metropolitan area: Lahore, Pakistan. We first examined the OCO-2 data availability at global scale. About 17% of OCO-2 soundings over the global 70 most populated cities from 2014 to 2019 are marked as high-quality. Cloud blockage and aerosol contamination are the two main causes of data loss. As an attempt to recover additional soundings, we evaluated the effectiveness of OCO-2 quality flags at the city level by comparing three flux quantification methods (WRF-Chem, X-STILT, and the flux cross-sectional integration method). The satellite/bottom-up emissions (OCO-2/ODIAC) ratios of the high-quality tracks with reduced uncertainties in emissions are better agreed across the three methods compared to the all-data tracks. This demonstrates that OCO-2 quality flags are useful filters of low-quality OCO-2 retrievals at local scales. All three methods consistently suggested that the ratio medians are greater than 1, implying that the ODIAC slightly underestimated CO_(2ff) emissions over Lahore. Additionally, our estimation of the a posteriori CO2ff emission trend was about 734 kt C/year (i.e., an annual 6.7% increase). 10,000 Monte Carlo simulations of the Mann-Kendall upward trend test showed that less than 10% prior uncertainty for 8 tracks (or less than 20% prior uncertainty for 25 tracks) is required to achieve a greater-than-50% trend significant possibility at a 95% confidence level. It implies that the trend is driven by the prior and not due to the assimilation of OCO-2 retrievals. The key to improving the role of satellite data in CO₂ emission trend detection lies in collecting more frequent high-quality tracks near metropolitan areas to achieve significant constraints from X_(CO₂) retrievals.