Atmospheric Chemistry and Physics (Aug 2019)

Observationally constrained analysis of sea salt aerosol in the marine atmosphere

  • H. Bian,
  • H. Bian,
  • K. Froyd,
  • K. Froyd,
  • D. M. Murphy,
  • J. Dibb,
  • A. Darmenov,
  • M. Chin,
  • P. R. Colarco,
  • A. da Silva,
  • T. L. Kucsera,
  • G. Schill,
  • G. Schill,
  • H. Yu,
  • P. Bui,
  • M. Dollner,
  • B. Weinzierl,
  • A. Smirnov

DOI
https://doi.org/10.5194/acp-19-10773-2019
Journal volume & issue
Vol. 19
pp. 10773 – 10785

Abstract

Read online

Atmospheric sea salt plays important roles in marine cloud formation and atmospheric chemistry. We performed an integrated analysis of NASA GEOS model simulations run with the GOCART aerosol module, in situ measurements from the PALMS and SAGA instruments obtained during the NASA ATom campaign, and aerosol optical depth (AOD) measurements from the AERONET Marine Aerosol Network (MAN) and from MODIS satellite observations to better constrain sea salt in the marine atmosphere. ATom measurements and GEOS model simulations both show that sea salt concentrations over the Pacific and Atlantic oceans have a strong vertical gradient, varying up to 4 orders of magnitude from the marine boundary layer to free troposphere. The modeled residence times suggest that the lifetime of sea salt particles with a dry diameter of less than 3 µm is largely controlled by wet removal, followed by turbulent process. During both boreal summer and winter, the GEOS-simulated sea salt mass mixing ratios agree with SAGA measurements in the marine boundary layer (MBL) and with PALMS measurements above the MBL. However, comparison of AOD from GEOS with AERONET/MAN and MODIS aerosol retrievals indicated that the model underestimated AOD over the oceans where sea salt dominates. The apparent discrepancy of slightly overpredicted concentration and large underpredicted AOD could not be explained by biases in the model RH affecting the particle hygroscopic growth, as modeled RH was found to be comparable to or larger than the in situ measurements. This conundrum could at least partially be explained by the difference in sea salt size distribution; the GEOS simulation has much less sea salt percentage-wise in the smaller particle size range and thus less efficient light extinction than what was observed by PALMS.