Aerosol optical depths over oceans: A view from MISR retrievals and collocated MAN and AERONET in situ observations. Issue 22 (19th November 2013)

• Record Type: Journal Article
• Authors: Witek, Marcin L.; Garay, Michael J.; Diner, David J.; Smirnov, Alexander
• Is Part Of: Journal of geophysical research. Volume 118:Issue 22(2013)
• Page Start: 12, 620
• Page End: 12, 633
• Publication Date: 2013-11-19
• Subjects: Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2013JD020393
• Languages: English
• ISSNs: 2169-897X

Abstract

<abstract abstract-type="main"> <title>Abstract</title> <p>[1] In this study, aerosol optical depths over oceans are analyzed from satellite and surface perspectives. Multiangle Imaging SpectroRadiometer (MISR) aerosol retrievals are investigated and validated primarily against Maritime Aerosol Network (MAN) observations. Furthermore, AErosol RObotic NETwork (AERONET) data from 19 island and coastal sites is incorporated in this study. The 270 MISR/MAN comparison points scattered across all oceans were identified. MISR on average overestimates aerosol optical depths (AODs) by 0.04 as compared to MAN; the correlation coefficient and root‐mean‐square error are 0.95 and 0.06, respectively. A new screening procedure based on retrieval region characterization is proposed, which is capable of substantially reducing MISR retrieval biases. Over 1000 additional MISR/AERONET comparison points are added to the analysis to confirm the validity of the method. The bias reduction is effective within all AOD ranges. Setting a clear flag fraction threshold to 0.6 reduces the bias to below 0.02, which is close to a typical ground‐based measurement uncertainty. Twelve years of MISR data are analyzed with the new screening procedure. The average over ocean AOD is reduced by 0.03, from 0.15 to 0.12. The largest AOD decrease is observed in high latitudes of both hemispheres, regions with climatologically high cloud cover. It is postulated that the screening procedure eliminates spurious retrieval errors associated with cloud contamination and cloud adjacency effects. The proposed filtering method can be used for validating aerosol and chemical transport models.</p> </abstract>