Atmospheric Environment: X (Jan 2023)
Analysis of atmospheric CO2 variability in the Marseille city area and the north-west Mediterranean basin at different time scales
Abstract
This study presents the first analysis of the variability of atmospheric CO2 in the area of the Marseille city (France). It addresses the role of anthropogenic emissions, natural fluxes and atmospheric boundary layer height (ABLH) dynamics on CO2 variability at the diurnal, synoptic, seasonal and multi-annual scales. A regional network based on 4 in-situ observation sites of CO2, CO and NOx was deployed between 2013 and 2018. One urban site (CAV) located in Marseille center was set up in collaboration with the regional air quality monitoring agency ATMOSUD. A second site (SME) was installed at the coastal edge of Marseille at the border of the Mediterranean Sea. The two other sites belonging to the ICOS (integrated Carbon Observing System) national atmospheric greenhouse observation network, are located in natural areas at the Observatoire de Haute Provence (OHP, 80 km north of Marseille) and at Cape Corsica (ERSA, 330 km east of Marseille) and are defined as regional background sites. The comparison between the sites was performed on the period common to all sites (1 July 2016–13 February 2018). The datasets are calibrated on the reference World Meteorological Organization scales for CO2 and CO with high precision and accuracy levels. At all sites, the mean annual CO2 growth rate is found to be quite similar to the Mauna Loa (Hawaii) reference site one, but mean annual CO2 concentrations are higher of several ppm at both urban sites than at both background sites. The diurnal cycle shows a higher amplitude at the urban sites (14.5 ppm at SME; 18.8 ppm at CAV) than at the background sites (5.3 ppm at OHP; 0.5 ppm at ERSA), as in other urban studies. While the urban stations are influenced by large urban anthropogenic emissions (mostly from traffic and heating, especially in winter), both background sites are mainly influenced by natural fluxes. At ERSA, the CO2 diurnal cycle is found to be primarily controlled by the small air-sea CO2 fluxes. At OHP, the diurnal variability of CO2 is mainly driven by the activity of vegetation (photosynthesis and respiration) and ABLH dynamics. For similar reasons, atmospheric CO2 concentrations are also characterized by larger seasonal variations in the city (29.2 ppm at CAV and 20.3 ppm at SME, respectively) than at OHP (13.1 ppm) and at SME (13.9 ppm). The influence of local, regional and remote anthropogenic emissions is assessed through a classification of the datasets by wind conditions. Similarly to other urban studies, a dome of several tens of ppm of CO2 gets formed over the city at low wind speed (less than 4 m s−1). For higher wind speeds (4–10 m s−1), the influence of regional and remote emissions on atmospheric CO2 is function of wind direction, varying from a few ppm at the background sites to a plume of more than 10 ppm at the urban ones. For very strong winds, the CO2 plume gets diluted. Finally local breezes, although not much frequent and more occurrent in summer, partly control atmospheric CO2 concentrations in Marseille. Additional local meteorological measurement sites would help to better characterize breezes in Marseille. Also, our study shows that additional background sites closer to the city on the path of the dominant winds would help to better constrain Marseille CO2 urban dome and plume. The NW and W sectors show a higher CO2 concentration variability even for strong winds, with likely an impact of the industrial area of Fos-Berre north-west of Marseille. Furthermore, as CO and NOx are used to assess the role of anthropogenic emissions vs natural fluxes on CO2, future dedicated campaigns using carbon isotopes will help to decipher the role of fossil fuel combustion sources vs modern ones on CO2 in Marseille. Finally, remote sensing measurements would be useful to better assess the impact of ABLH on atmospheric CO2 in the coastal area of Marseille where atmospheric dynamics are quite complex.
