XCO 2 in an emission hot-spot region: the COCCON Paris campaign 2015

Felix Vogel, Matthias Frey, Johannes Staufer, Frank Hase, Grégoire Broquet, Irène Xueref-Rémy, Frédéric Chevallier, Philippe Ciais, Mahesh Kumar Sha, P. Chélin, Pascal Jeseck, Christof Janssen, Yao Té, Jochen Groß, Thomas Blumenstock, Qiansi Tu, J. Orphal

TL;DR

This study uses a network of portable FTIR spectrometers deployed around Paris to measure column-averaged CO2 (XCO2) concentrations during a 2015 spring campaign. The data reveal distinct XCO2 patterns upwind and downwind of the city, which are well-reproduced by the CHIMERE atmospheric transport model, highlighting the significant impact of urban fossil fuel emissions on local CO2 levels.

What Problem Does This Paper Address?

Accurately quantifying greenhouse gas (GHG) emissions, particularly from urban areas, is crucial for effective climate change mitigation. Cities are major contributors to global fossil fuel CO2 (FFCO2) emissions, but their exact contribution and the spatial distribution of these emissions are often uncertain. Traditional methods of measuring atmospheric CO2 provide background levels, making it difficult to isolate and quantify emissions from specific urban hot spots. This paper addresses the challenge of using ground-based measurements to understand and quantify urban CO2 emissions by deploying a network of instruments to capture the influence of a major city (Paris) on atmospheric CO2 concentrations.

Methods

The study employed a network of five portable Fourier Transform Infrared (FTIR) spectrometers (EM27/SUN) strategically placed in and around the Paris metropolitan area. These instruments measure column-averaged dry air mole fractions of CO2 (XCO2) by analyzing the absorption of sunlight. The campaign ran from April 28 to May 13, 2015. To interpret the observed XCO2 variations, the researchers utilized the CHIMERE atmospheric transport model, configured with a 2 km resolution for the Paris region. This model simulates the transport and mixing of CO2, allowing for the separation of contributions from local fossil fuel emissions, biogenic fluxes, and boundary conditions. The study also analyzed spatial gradients of XCO2 between different measurement sites to better isolate the impact of local urban emissions.

Datasets

The primary dataset used in this study consists of XCO2 measurements collected by the five EM27/SUN FTIR spectrometers during the Paris campaign (April 28 - May 13, 2015). For modeling, the study used:

  • IER emission inventory: A 1 km resolution emission model for France, used for anthropogenic emissions within the Paris region.
  • EDGAR V4.2 database: Used for anthropogenic emissions outside the Paris region.
  • CTESSEL: The land-surface component of the ECMWF forecasting system, used for natural fluxes (Net Ecosystem Exchange - NEE).
  • Copernicus Atmosphere Monitoring Service (CAMS) CO2 forecasts: Used for boundary conditions.
  • ECMWF operational analyses: Used for meteorological forcing in the CHIMERE model.
  • Whole Atmosphere Community Climate Model (WACCM) version 6: Provided the a priori CO2 profile for the modeling.

Key Insights and Ideas

  • Urban CO2 plumes are detectable: Even in a densely populated region like Paris, the network of instruments successfully captured distinct differences in XCO2 between upwind and downwind locations, clearly indicating the influence of urban emissions.
  • Biospheric uptake matters, even in cities: The study found that despite significant urban emissions, the natural uptake of CO2 by vegetation (biospheric uptake) plays a major role in the daily cycle of XCO2, causing a noticeable decrease during the afternoon.
  • Fossil fuel emissions dominate local gradients: When analyzing the difference in XCO2 between upwind and downwind sites, the study's modeling framework indicated that these local gradients are primarily driven by fossil fuel CO2 emissions from the Paris metropolitan area.
  • Model-data agreement improves with gradient analysis: Focusing on the difference in XCO2 between sites (gradients) rather than absolute values helped to reduce the impact of uncertainties in boundary conditions and biogenic fluxes, leading to better agreement between model simulations and observations.

Key Results

  • Observed XCO2 in the Paris region ranged from 397.27 to 404.66 ppm, with a mean of 401.26 ppm.
  • All sites showed similar diurnal patterns: XCO2 increased at night and decreased significantly (typically 2-3 ppm) during the afternoon.
  • Downwind sites consistently showed higher XCO2 concentrations than upwind sites, demonstrating the impact of urban emissions.
  • The CHIMERE model, with a 2 km resolution over Paris, successfully reproduced the observed diurnal cycles and the relative differences between sites.
  • Analysis of modeled XCO2 gradients showed that fossil fuel emissions were the dominant contributor to the observed differences between upwind and downwind sites.
  • The model-data agreement for spatial gradients (AXCO2) was good, with a slope of 1.07±0.09 for the PIS-RES pair, though a potential underestimation of urban source impact was noted for the MIT-RES gradient.

Limitations and Future Work

  • Short campaign duration: The 2-week campaign duration limited the ability to perform in-depth analyses of seasonal variations or the impact of specific ambient conditions (e.g., traffic, temperature) on emissions.
  • Model bias: A general offset was observed between modeled and observed XCO2, with the model often showing higher background concentrations. This is attributed to uncertainties in the CO2 forecast used for boundary conditions.
  • Emission inventory refinement: The study suggests that the emission model for the eastern outskirts of Paris might need improvement, as the modeled gradient for MIT-RES indicated a potential underestimation of local FFCO2 impact.
  • Future work: The authors advocate for longer-term observations across different seasons to better understand biogenic and anthropogenic CO2 fluxes. They also highlight the development of remotely controllable shelters for FTIR instruments, which will facilitate the establishment of permanent spectrometer arrays around cities for continuous monitoring and improved urban emission estimates. The study also points to the need for satellite observations with higher precision and lower biases to complement ground-based networks.

Summary updated

Abstract

Providing timely information on urban greenhouse gas (GHG) emissions and their trends to stakeholders relies on reliable measurements of atmospheric concentrations and the understanding of how local emissions and atmospheric transport influence these observations. Portable Fourier transform infrared (FTIR) spectrometers were deployed at five stations in the Paris metropolitan area to provide column-averaged concentrations of CO2 (XCO2) during a field campaign in spring of 2015, as part of the Collaborative Carbon Column Observing Network (COCCON). Here, we describe and analyze the variations of XCO2 observed at different sites and how they changed over time. We find that observations upwind and downwind of the city centre differ significantly in their XCO2 concentrations, while the overall variability of the daily cycle is similar, i.e. increasing during night-time with a strong decrease (typically 2–3 ppm) during the afternoon. An atmospheric transport model framework (CHIMERE-CAMS) was used to simulate XCO2 and predict the same behaviour seen in the observations, which supports key findings, e.g. that even in a densely populated region like Paris (over 12 million people), biospheric uptake of CO2 can be of major influence on daily XCO2 variations. Despite a general offset between modelled and observed XCO2, the model correctly predicts the impact of the meteorological parameters (e.g. wind direction and speed) on the concentration gradients between different stations. When analyzing local gradients of XCO2 for upwind and downwind station pairs, those local gradients are found to be less sensitive to changes in XCO2 boundary conditions and biogenic fluxes within the domain and we find the model–data agreement further improves. Our modelling framework indicates that the local XCO2 gradient between the stations is dominated by the fossil fuel CO2 signal of the Paris metropolitan area. This further highlights the potential usefulness of XCO2 observations to help optimize future urban GHG emission estimates.
BibTeX
@inproceedings{vogel2019xco,
  author = {Felix Vogel and Matthias Frey and Johannes Staufer and Frank Hase and Grégoire Broquet and Irène Xueref-Rémy and Frédéric Chevallier and Philippe Ciais and Mahesh Kumar Sha and P. Chélin and Pascal Jeseck and Christof Janssen and Yao Té and Jochen Groß and Thomas Blumenstock and Qiansi Tu and J. Orphal},
  title = {{XCO <sub>2</sub> in an emission hot-spot region: the COCCON Paris campaign 2015}},
  booktitle = {Atmospheric},
  year = {2019},
  doi = {10.5194/acp-19-3271-2019},
}
Figure