TL;DR
This study leverages 5 years of Orbiting Carbon Observatory-2 (OCO-2) satellite data to independently quantify carbon dioxide (CO2) emissions from Chinese cities and industrial areas. By developing a novel method to detect and attribute CO2 plumes, the researchers provide a valuable, independent check on national emission inventories, demonstrating the potential of satellite observations for global climate target verification.
What Problem Does This Paper Address?
The Paris Agreement requires countries to report their greenhouse gas emissions, but self-reporting mechanisms can have biases. Independent verification is crucial for assessing collective progress towards climate goals. While satellite observations offer a promising avenue for this, existing studies have been limited by instrument capabilities and focus on only the largest emission sources. This paper addresses the challenge of using OCO-2, an instrument not specifically designed for detailed emission monitoring, to observe and quantify CO2 emissions from a broader range of urban and industrial areas in China, a major global emitter. The goal is to demonstrate the feasibility of using satellite data to provide a more comprehensive and independent evaluation of emission inventories.
Methods
The core of the methodology involves a multi-step filtering process applied to OCO-2 data to identify and quantify CO2 emissions:
- XCO2 Local Enhancement Identification: The researchers first identify localized increases in column-averaged dry-air mole fraction of CO2 (XCO2) along OCO-2 tracks. This involves fitting a curve to the data within a moving window to separate the plume signal from the background, using a Gaussian plume model to represent the enhancement. Strict criteria are applied to ensure the identified enhancement is statistically significant and robust.
- Attribution to Emission Sources: The identified CO2 enhancements are then linked to nearby human emission sources using the spatially explicit Multi-resolution Emission Inventory for China (MEIC) dataset. A Gaussian plume model is used to simulate plumes from known emission sources and match them to the observed satellite data. This step aims to attribute the observed CO2 plumes to specific cities or industrial clusters.
- Cross-sectional CO2 Flux Estimation: For plumes that can be unambiguously attributed to a single, isolated emission source or cluster, the researchers estimate the cross-sectional CO2 flux. This involves calculating the area under the fitted CO2 enhancement curve (CO2 line density) and multiplying it by the wind speed perpendicular to the OCO-2 track. Wind direction is optimized within a certain range to maximize the correlation between modeled and observed plumes.
Datasets
- OCO-2 (Orbiting Carbon Observatory-2) retrievals: Version 9r of the bias-corrected XCO2 data, spanning September 2014 to August 2019. This dataset provides column-averaged dry-air mole fractions of CO2.
- MEIC (Multi-resolution Emission Inventory for China): A detailed inventory providing locations of individual industrial point sources and 0.1° × 0.1° area source emissions for China. This dataset is crucial for attributing observed CO2 plumes to specific emitters.
- ERA5 reanalysis data: Used to obtain local wind fields (wind speed and direction) at various pressure levels, which are essential for plume transport modeling.
Key Insights and Ideas
- OCO-2's Potential Beyond Design: The study demonstrates that OCO-2, primarily designed for monitoring natural carbon cycle fluxes, can be effectively adapted to detect and quantify anthropogenic CO2 emissions from urban and industrial areas, even with its sparse sampling.
- Bridging the Gap in Emission Monitoring: By observing plumes from 46 cities and industrial regions, the research significantly expands the number of sources that can be independently verified compared to previous studies, offering a more comprehensive view of emissions.
- Importance of Localized Data: The detailed spatial information from MEIC, combined with OCO-2's localized observations, is critical for attributing CO2 plumes to specific emission clusters, highlighting the value of high-resolution emission inventories.
- Seasonal Differences in Emissions: The study notes that satellite-based estimates are higher in the cold season, potentially due to human respiration not accounted for in inventories, and lower in the warm season, possibly due to plant carbon uptake influencing the signal.
Key Results
- The study successfully identified and quantified CO2 emissions from 46 cities and industrial regions in China using 5 years of OCO-2 data.
- The aggregated emissions from these regions amounted to approximately 1.3 Gt CO2 yr⁻¹, representing about 13% of mainland China's annual emissions.
- The number of cities whose emissions were constrained by OCO-2 was 3 to 10 times larger than in previous studies.
- Satellite-based emission estimates showed broad consistency with the MEIC inventory but more significant differences with global gridded datasets like EDGAR and ODIAC, particularly at the city level.
- The total observed emissions from the studied regions (1.25 Gt CO2 yr⁻¹) were comparable to bottom-up estimates from ODIAC (1.13 Gt CO2 yr⁻¹), EDGAR (1.38 Gt CO2 yr⁻¹), and MEIC (1.17 Gt CO2 yr⁻¹).
Limitations and Future Work
- Sparse Sampling: OCO-2's sparse sampling, due to its narrow swath and reliance on clear-sky conditions, limits the number of observations and the ability to capture diurnal emission cycles.
- Instrument Design: OCO-2 was not specifically designed for detailed emission monitoring, which introduces challenges in data processing and interpretation.
- Wind Uncertainty: Uncertainties in wind speed and direction, even with optimization, can affect the accuracy of emission flux estimations.
- Future Work: The authors suggest that future satellite missions with denser spatial sampling (e.g., CO2 Monitoring mission, GeoCarb) will significantly improve the ability to monitor emissions. Refining data selection criteria and improving atmospheric transport models are also areas for future development. The approach's simplicity makes it applicable globally, and further work could focus on applying it to other regions.
Summary updated