Journal metrics

Journal metrics

  • IF value: 3.248 IF 3.248
  • IF 5-year value: 3.650 IF 5-year 3.650
  • CiteScore value: 3.37 CiteScore 3.37
  • SNIP value: 1.253 SNIP 1.253
  • SJR value: 1.869 SJR 1.869
  • IPP value: 3.29 IPP 3.29
  • h5-index value: 47 h5-index 47
  • Scimago H index value: 60 Scimago H index 60
Atmos. Meas. Tech., 7, 1185-1199, 2014
https://doi.org/10.5194/amt-7-1185-2014
© Author(s) 2014. This work is distributed under
the Creative Commons Attribution 3.0 License.
Research article
08 May 2014
Quantification of uncertainty in aerosol optical thickness retrieval arising from aerosol microphysical model and other sources, applied to Ozone Monitoring Instrument (OMI) measurements
A. Määttä1,2, M. Laine1, J. Tamminen1, and J. P. Veefkind3,4 1Finnish Meteorological Institute, Helsinki, Finland
2University of Helsinki, Department of Mathematics and Statistics, Helsinki, Finland
3Royal Netherlands Meteorological Institute, De Bilt, the Netherlands
4Delft University of Technology, Department of Civil Engineering and Geosciences, Delft, the Netherlands
Abstract. Satellite instruments are nowadays successfully utilised for measuring atmospheric aerosol in many applications as well as in research. Therefore, there is a growing need for rigorous error characterisation of the measurements. Here, we introduce a methodology for quantifying the uncertainty in the retrieval of aerosol optical thickness (AOT). In particular, we concentrate on two aspects: uncertainty due to aerosol microphysical model selection and uncertainty due to imperfect forward modelling. We apply the introduced methodology for aerosol optical thickness retrieval of the Ozone Monitoring Instrument (OMI) on board NASA's Earth Observing System (EOS) Aura satellite, launched in 2004. We apply statistical methodologies that improve the uncertainty estimates of the aerosol optical thickness retrieval by propagating aerosol microphysical model selection and forward model error more realistically. For the microphysical model selection problem, we utilise Bayesian model selection and model averaging methods. Gaussian processes are utilised to characterise the smooth systematic discrepancies between the measured and modelled reflectances (i.e. residuals). The spectral correlation is composed empirically by exploring a set of residuals. The operational OMI multi-wavelength aerosol retrieval algorithm OMAERO is used for cloud-free, over-land pixels of the OMI instrument with the additional Bayesian model selection and model discrepancy techniques introduced here. The method and improved uncertainty characterisation is demonstrated by several examples with different aerosol properties: weakly absorbing aerosols, forest fires over Greece and Russia, and Sahara desert dust. The statistical methodology presented is general; it is not restricted to this particular satellite retrieval application.

Citation: Määttä, A., Laine, M., Tamminen, J., and Veefkind, J. P.: Quantification of uncertainty in aerosol optical thickness retrieval arising from aerosol microphysical model and other sources, applied to Ozone Monitoring Instrument (OMI) measurements, Atmos. Meas. Tech., 7, 1185-1199, https://doi.org/10.5194/amt-7-1185-2014, 2014.
Publications Copernicus
Download
Share