Workpackage 4
Inverse modelling from satellite observations for improving sources of aerosol and gas precursors, as well as their evolution along transport

Leaders laboratories: LOA
Participants: ICARE

 

Contact

Oleg Dubovik
LOA

Scientific objectives

Introduction

Materials and methods

Ongoing studies and results

Key publications


 

Scientific objectives

Inverse modelling of global chemistry/transport models (CHIMERE, GOES_CHEM) combined with spaceborne observations to tune trace gas and aerosol emissions. The polarimetric (e.g. POLDER/PARASOL) satellite observations provide sensitivity to aerosol parameters such as composition (via sensitivity to refractive index), shape (spherical particles or not like dust), size, which allows to identify the aerosol type. Hyper spectral (e.g. IASI/METOP or Tanso-fts/Gosat) satellite observations provide all at once trace gas and aerosol concentrations as well and aerosol size. Using the results of satellite inversions improves inverse modelling algorithm and provides improved emissions fields to the modelling community.

Illustration: Cartoons that depict schematic views of models and satellite combinations (relevant for aerosols).


Introduction

New inverse modeling algorithm intends to tune chemical transport models by the global satellite aerosol data and to derive/correct the global sources of atmospheric aerosols. This approach compares the model outcome with the real observations at large space and time scales and corrects the least known model parameters (e.g. aerosol or gas sources and sinks) relying on better-known model components (e.g. meteorological fields, advection and diffusion processes, deposition). However, while the inverse modeling of atmospheric trace gases has been established and rather commonly used, the atmospheric aerosol has generally higher temporal and spatial variability than gases and its properties are characterized by a larger number of parameters (including particle size, shape, composition, etc.), which explains the fact that there are only few studies discussing application of an inverse modeling approach for atmospheric aerosol. 



Materials and methods

Satellite Data :

  • POLDER/PARASOL : global information about aerosol loading, particle composition, shape, etc.
  • CALIPSO : information aerosol vertical distribution
  • FCI/METEOSAT : information on diurnal aerosol variability
  • IASI/METOP and Tanso-fts/Gosat: information on aerosol and gas aerosol precursors

Chemical/Transport Model :

  • Regional chemistry-transport model (CHIMERE)
  • Global aerosol transport model (GEOS_CHEM)

Inverse algorithm :

  • GRASP (Generalized Retrieval of Aerosol and Surface Properties)
  • Information Content (IC) analysis

Ongoing studies and results

  • Development of inverse modelling algorithm GRASP (Generalized Retrieval of Aerosol and Surface Properties) : unified algorithm for characterizing atmospheric and surface properties gathered from a variety of remote sensing observations
  • Application of GRASP on PARASOL data archive (full resolution 5x6 km2) : it provides retrievals everywhere including both bright and dark land surfaces and aerosol parameters such as single scattering albedo
  • Reconstructing the flux and altitude of the volcanic SO2 emissions with an hourly resolution from the Infrared Atmospheric Sounding Interferometer (IASI) ; regional chemistry-transport model CHIMERE is used to describe the dispersion of SO2 released in the atmosphere
  • Evaluation of the consistency of the detection and characterization of volcanic ash plumes (optical thickness, effective particle size, mass concentration) from different thermal infrared instruments (Case study: Eyjafjallajökull eruption in May 2010)
  • Retrieval of CO2 and CH4 column concentrations and aerosols characterization using spectral synergy from thermal IR to visible spectrum : spectral synergy allows obtaining up to almost seven different aerosol parameters and retrieving CO2 and CH4 total columns simultaneously in the presence of one aerosol layer.


Illustration : Upper panel : GEOS-CHEM emissions (Dust, BC and OC) denoted as “Prior” (source by default) and retrieved using satellite retrieval denoted as “Posterior”. Lower panel : average spectral AODs retrieved by GRASP and predicted by GEOS-CHEM (default values). It is clear that default GEOS-CHEM emissions strongly underestimated BC and OC emissions and overestimated dust emissions. Results are averaged over 2 weeks in Sep 2008.


Key publications

  • Boichu, M., Clarisse, L., Péré, J.-C., Herbin, H., Goloub, P., Thieuleux, F., Ducos, F., Clerbaux, C., Tanré, D. 2015, Temporal variations of flux and altitude of sulfur dioxide emissions during volcanic eruptions: implications for long-range dispersal of volcanic clouds, Atmospheric Chemistry and Physics, 15, 8381-8400, doi:10.5194/acp-15-8381-2015
  • Dubovik, O., Lapyonok, T., Litvinov, P., Herman, M., Fuertes, D., Ducos, F., Lopatin, A., Chaikovsky, A., Torres, B., Derimian, Y., Huang, X., Aspetsberger, M., and Federspiel C.: GRASP: a versatile algorithm fo characterizing the atmosphere, SPIE: Newsroom, Doi: 10.1117/2.1201408.005558, Published Online: 19 /09/2014. http://spie.org/x109993.xml.