Abstract
Abstract. A model of the sea surface bidirectional reflectance distribution function (BRDF) is presented for the visible and near-IR channels (over the spectral range 550 nm to 1.6 μm) of the dual-viewing Along-Track Scanning Radiometers (ATSRs). The intended application is as part of the Oxford-RAL Aerosols and Clouds (ORAC) retrieval scheme. The model accounts for contributions to the observed reflectance from whitecaps, sun-glint and underlight. Uncertainties in the parametrisations used in the BRDF model are propagated through into the forward model and retrieved state. The new BRDF model offers improved coverage over previous methods, as retrievals are possible into the sun-glint region, through the ATSR dual-viewing system. The new model has been applied in the ORAC aerosol retrieval algorithm to process Advanced ATSR (AATSR) data from September 2004 over the south-eastern Pacific. The assumed error budget is shown to be generally appropriate, meaning the retrieved states are consistent with the measurements and a priori assumptions. The resulting field of aerosol optical depth (AOD) is compared with colocated MODIS-Terra observations, AERONET observations at Tahiti, and cruises over the oceanic region. MODIS and AATSR show similar spatial distributions of AOD, although MODIS reports values which are larger and more variable. It is suggested that assumptions in the MODIS aerosol retrieval algorithm may lead to a positive bias in MODIS AOD of order 0.01 at 550 nm over ocean regions where the wind speed is high.
Highlights
The Intergovernmental Panel for Climate Change (IPCC) has identified aerosols as among the most uncertain contributions to radiative forcing (Penner et al, 2001, Forster et al, 2007)
Two-channel aerosol retrieval algorithms are presented for Advanced Very High Resolution Radiometer (AVHRR) by Higurashi and Nakajima (1999) and Mishchenko et al (1999): these use fixed glint-based surface reflectances calculated as described in Nakajima and Tanaka (1983) and Mishchenko and Travis (1997), respectively
The reader is referred to Zeisse (1995) for more details; such extreme viewing geometries do not occur for the (A)Along-Track Scanning Radiometers (ATSRs) views, calculation of the reflectance at such geometries is required for the integration to obtain Rbd and Rdd for the retrieval forward model
Summary
The Intergovernmental Panel for Climate Change (IPCC) has identified aerosols as among the most uncertain contributions to radiative forcing (Penner et al, 2001, Forster et al, 2007). A previous version of the algorithm described here and applied to ATSR-2 data, involving a nadir-view aerosol retrieval algorithm, is detailed by Thomas et al (2009b) and Thomas et al (2010) This took a similar approach for sea surface reflectance allowed the absolute magnitude to vary (while fixing the spectral shape of the surface). Two-channel aerosol retrieval algorithms are presented for AVHRR by Higurashi and Nakajima (1999) and Mishchenko et al (1999): these use fixed glint-based surface reflectances calculated as described in Nakajima and Tanaka (1983) and Mishchenko and Travis (1997), respectively They consider the impacts of the simple reflectance model on the retrieved AOD and Angstrom exponent, noting that it can be significant for cases of high wind or pigment concentrations, or low aerosol loadings. This, together with vicarious calibration against stable bright ground targets, means that the visible channel reflectances are known to an accuracy of 2–3% (Smith et al, 2002, 2008)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
