Assessment of atmospheric aerosols from two reanalysis products over Australia

S.K. Mukkavilli, A.A. Prasad, R.A. Taylor, J. Huang, R.M. Mitchell, A. Troccoli, M.J. Kay

Research output: Contribution to journalArticlepeer-review

45 Citations (Scopus)
26 Downloads (Pure)

Abstract

Assessments of atmospheric aerosols from reanalysis are important for understanding uncertainty in model simulations, and ultimately predictions, such as for solar power or air quality forecasts and assessments. This study intercompares total aerosol optical depth (AOD) and dust AOD (DAOD) from two global reanalyses datasets, the European Centre for Medium-Range Weather Forecasts (ECMWF) Monitoring Atmospheric Composition and Climate (MACC) and the NASA Modern-Era Retrospective Analysis for Research-2 (MERRA-2). These are evaluated against AeroSpan (Aerosol characterisation via Sun photometry: Australian Network) ground observations which forms part of the Aerosol Robotic Network (AERONET) over the Australian continent for the 2002–2012 period. During dust storms, AeroSpan/AERONET AOD measurements were missing due to cloud screening. To overcome validation limitations in sun photometry for dust events, a nephelometer's scattering coefficient is qualitatively compared against reanalysis of DAOD at a key dust storm activation site, Tinga Tingana in South Australia (~200 km east of Lake Eyre). A specific extreme event that occurred in 2009 originating from the Lake Eyre basin, a major dust source covering one-sixth of Australia, was studied. The results show that MERRA-2 reanalysis overestimates monthly total AOD twice as much compared to AeroSpan/AERONET ground observations but seems better correlated against AeroSpan/AERONET than ECMWF/MACC. Mean data of MERRA-2 time series over 10 years provide lower DAOD values and lower dust aerosol estimates than ECMWF/MACC reanalysis (over the Lake Eyre basin with spatial averaging). Specifically at Tinga Tingana, the correlation from MERRA-2 (0.45 correlation) and ECMWF/MACC (0.43 correlation) against AeroSpan/AERONET's AOD were similar. Between MERRA-2 and ECMWF/MACC decade long daily gridded DAOD, the correlation coefficient was high at 0.73, again indicating similarity between the datasets. MERRA-2 total AOD correlation is significantly higher (by 0.26) against AeroSpan/AERONET than ECMWF/MACC. MERRA-2 also provides higher AOD values in extreme cases which may correspond to dust storms. During dust storms, a hybrid strategy using nephelometers and hourly reanalysis from MERRA-2 is able to identify dust storms better than AeroSpan/AERONET. Overall, this work can enable and inform better aerosol data assimilation into forecast models such as for solar energy, agriculture or air quality over Australia.
Original languageEnglish
Pages (from-to)149-164
Number of pages16
JournalAtmospheric Research
Volume215
Early online date5 Sep 2018
DOIs
Publication statusPublished - 1 Jan 2019

Keywords

  • Reanalysis
  • Aerosol optical depth
  • AERONET
  • Australian dust AOD
  • ECMWF/MACC
  • MERRA-2

Cite this