Accounting for the effects of surface BRDF on satellite cloud and trace-gas retrievals: a new approach based on geometry-dependent Lambertian equivalent reflectivity applied to OMI algorithms

Abstract

Abstract. Most satellite nadir ultraviolet and visible cloud, aerosol, and trace-gas algorithms make use of climatological surface reflectivity databases. For example, cloud and NO 2 retrievals for the Ozone Monitoring Instrument (OMI) use monthly gridded surface reflectivity climatologies that do not depend upon the observation geometry. In reality, reflection of incoming direct and diffuse solar light from land or ocean surfaces is sensitive to the sun-sensor geometry. This dependence is described by the bidirectional reflectance distribution function (BRDF). To account for the BRDF, we propose to use a new concept of geometry-dependent Lambertian equivalent reflectivity (LER). Implementation within the existing OMI cloud and NO 2 retrieval infrastructure requires changes only to the input surface reflectivity database. The geometry-dependent LER is calculated using a vector radiative transfer model with high spatial resolution BRDF information from the Moderate Resolution Imaging Spectroradiometer (MODIS) over land and the Cox-Munk slope distribution over ocean with a contribution from water-leaving radiance. We compare the geometry-dependent and climatological LERs for two wavelengths, 354 and 466 nm, that are used in OMI cloud algorithms to derive cloud fractions. A detailed comparison of the cloud fractions and pressures derived with climatological and geometry-dependent LERs is carried out. Geometry-dependent LER and corresponding retrieved cloud products are then used as inputs to our OMI NO 2 algorithm. We find that replacing the climatological OMI-based LERs with geometry-dependent LERs can increase NO 2 vertical columns by up to 50 % in highly polluted areas; the differences include both BRDF effects and biases between the MODIS and OMI-based surface reflectance data sets. Only minor changes to NO 2 columns (within 5 %) are found over unpolluted and overcast areas.

The version 3 OMI NO<sub>2</sub> standard product

“…With the currently adopted AMF estimates we anticipate an overall reduction in the OMI-derived top-down anthropogenic NO x emissions and surface concentrations. However, applying a new geometry-dependent Lambertian equivalent reflectivity in AMF calculation would result in increasing tropospheric VCDs (Vasilkov et al, 2017) and derived top-down NO x emissions and surface concentrations.…”

“…In the free troposphere, W i,trop increases with altitude and strongly depends on the cloud radiance fraction and optical centroid pressure Vasilkov et al, 2009). In the boundary layer and under cloud-free conditions, W depends most strongly on altitude and surface pressure and reflectance (Vasilkov et al, 2017). The AMF for a stratospheric or tropospheric column is computed as the vertical integral of the NO 2 profile shape weighted average of W (Eq.…”

The importance of surface reflectance anisotropy for cloud and NO<sub>2</sub> retrievals from GOME-2 and OMI

“…Vasilkov et al (2017) also found increased differences in the tropospheric AMFs due to surface BRDF effects on cloud parametes, but as reported by Lin et al (2014), there can also be compensating effects.…”

“…The indirect effect of biased cloud parameters on NO 2 retrievals combined with effects on clear-sky AMFs received less attention. Vasilkov et al (2017) and Lin et al (2015) addressed the indirect effects and showed that the effects on cloud parameters could enhance or compensate the direct effect on clear-sky AMFs.…”

“…A number of recent studies have attempted to account for the effects of anisotropic reflectance on both cloud and NO 2 retrievals (Lin et al (2014);Vasilkov et al (2017)), but a global assessment in-5 cluding the full range of possible retrieval conditions is still missing. An additional incentive to account for surface reflectance anisotropy is that the currently available LER climatologies (Kleipool et al (2008); Tilstra et al (2017)) describe the spatial variation in albedo at a scale (0.5°×0.5°-1°×1°) coarser than the OMI or GOME-2 pixel itself (13 × 24 km 2 / 80 × 40 km 2 ).…”