Analyzing Local Carbon Dioxide and Nitrogen Oxide Emissions From Space Using the Divergence Method: An Application to the Synthetic SMARTCARB Dataset

Abstract

Since the Paris Agreement was adopted in 2015, the role of space-based observations for monitoring anthropogenic greenhouse gas (GHG) emissions has increased. To meet the requirements for monitoring carbon dioxide (CO2) emissions, the European Copernicus programme is preparing a dedicated CO2 Monitoring (CO2M) satellite constellation that will provide CO2 and nitrogen dioxide (NO2) observations at 4 km2 resolution along a 250 km wide swath. In this paper, we adapt the recently developed divergence method to derive both CO2 and nitrogen oxide (NOx) emissions of cities and power plants from a CO2M satellite constellation by using synthetic observations from the COSMO-GHG model. Due to its long lifetime, the large CO2 atmospheric background needs to be removed to highlight the anthropogenic enhancements before calculating the divergence. Since the CO2 noise levels are large compared to the anthropogenic enhancements, we apply different denoising methods and compare the effect on the CO2 emission estimates. The annual NOx and CO2 emissions estimated from the divergence maps using the peak fitting approach are in agreement with the expected values, although with larger uncertainties for CO2. We also consider the possibility to use co-emitted NOx emission estimates for quantifying the CO2 emissions, by using source-specific NOx-to-CO2 emission ratios derived directly from satellite observations. In general, we find that the divergence method provides a promising tool for estimating CO2 emissions, alternative to typical methods based on inverse modeling or on the analysis of individual CO2 plumes.

Building a bridge: characterizing major anthropogenic point sources in the South African Highveld region using OCO-3 carbon dioxide snapshot area maps and Sentinel-5P/TROPOMI nitrogen dioxide columns

“…one year). Recently, Hakkarainen et al (2022) proposed an extension of the divergence method to CO 2 based on synthetic CO 2 observations that could also be applied to future satellite observations over areas with multiple point sources close to each other.…”

“…Also, no plumes can be observed during persistently cloudy conditions or during night-time. This clear-sky bias will remain an issue for any emission estimation method based on passive remote sensing observations, including those methods based on temporal averaging (Hakkarainen et al 2022).…”

“…The improved spatial resolution has also fostered new methodological developments for the emission estimation, such as the divergence method (Beirle et al 2019(Beirle et al , 2021, which enabled the compilation of a global catalogue of NO 2 emissions from point sources. The divergence method can be potentially applied to CO 2 observations from the upcoming CO2M mission (Hakkarainen et al 2022). Recent advances in space-based CO 2 observations have been provided by the NASA's OCO-3 mission (Eldering et al 2019) on the International Space Station (ISS).…”

On the theory of the divergence method for quantifying source emissions from satellite observations

“…Recently, the method has also been applied to other satellites such as GEMS NO 2 data (Xu et al, 2023). A study of its application to synthetic CO 2 data for the upcoming CO2M mission can be found in Hakkarainen et al (2022), which notes that a background and noise removal step must be applied to make the method robust for CO 2 data. Since its inception by Beirle et al (2019), notable improvements to the method include a different order of operations (differentiatethen-average instead of average-then-differentiate) as proposed in de Foy and Schauer (2022), computing the results on the satellite swaths themselves before remapping to a uniform grid as proposed in de Foy and Schauer (2022), skipping computations proportional to the wind divergence and adding a correction term proportional to the gradient of the topography as proposed in Sun (2022), and approximating the divergence operation using finite-difference stencils that can account for non-pixel-aligned transport as proposed in Sun (2022) and Veefkind et al (2023).…”

Segmentation of XCO2 images with deep learning: application to synthetic plumes from cities and power plants

“…The SMARTCARB simulations were run with the COSMO-GHG model for a domain centred on Berlin and covering several neighbouring power plants (Jänschwalde, Lippendorf, Boxberg and others). The simulations were used to generate synthetic CO2M observations (Kuhlmann et al, 2020b) and to assess different plume detection and inversion methods (Kuhlmann et al, 2019a(Kuhlmann et al, , 2020a(Kuhlmann et al, , 2021Hakkarainen et al, 2022). The model fields consist of hourly data over one year with a spatial resolution of 0.01 • and sixty vertical layers from 0 to 24 km.…”