• Acarreta, J. R., , J. F. De Haan, , and P. Stammes, 2004: Cloud pressure retrieval using the O2–O2 absorption band at 477 nm. J. Geophys. Res., 109, D05204, doi:10.1029/2003JD003915.

    • Search Google Scholar
    • Export Citation
  • Ahmad, Z., , P. K. Bhartia, , and N. Krotkov, 2004: Spectral properties of backscattered UV radiation in cloudy atmospheres. J. Geophys. Res., 109, D01201, doi:10.1029/2003JD003395.

    • Search Google Scholar
    • Export Citation
  • Ahn, C., , O. Torres, , and P. K. Bhartia, 2008: Comparison of Ozone Monitoring Instrument UV aerosol products with Aqua/Moderate Resolution Imaging Spectroradiometer and Multiangle Imaging Spectroradiometer observations in 2006. J. Geophys. Res., 113, D16S27, doi:10.1029/2007JD008832.

    • Search Google Scholar
    • Export Citation
  • Bond, T. C., 2001: Spectral dependence of visible light absorption by carbonaceous particles emitted from coal combustion. Geophys. Res. Lett., 28, 40754078.

    • Search Google Scholar
    • Export Citation
  • Chand, D., , T. L. Anderson, , R. Wood, , R. J. Charlson, , Y. Hu, , Z. Liu, , and M. Vaughan, 2008: Quantifying above-cloud aerosol using spaceborne lidar for improved understanding of cloudy-sky direct climate forcing. J. Geophys. Res., 113, D13206, doi:10.1029/2007JD009433.

    • Search Google Scholar
    • Export Citation
  • Chand, D., , R. Wood, , T. L. Anderson, , S. K. Satheesh, , and R. J. Charlson, 2009: Satellite-derived direct radiative effect of aerosols dependent on cloud cover. Nat. Geosci., 2, 181184, doi:10.1038/ngeo437.

    • Search Google Scholar
    • Export Citation
  • Colarco, P., , M. Schoeberl, , B. Doddridge, , L. Marufu, , O. Torres, , and E. Welton, 2004: Transport of smoke from Canadian forest fires to the surface near Washington, D.C.: Injection height, entrainment, and optical properties. J. Geophys. Res., 109, D06203, doi:10.1029/2003JD004248.

    • Search Google Scholar
    • Export Citation
  • Dave, J. V., , and C. L. Mateer, 1967: A preliminary study on the possibility of estimating total atmospheric ozone from satellite measurements. J. Atmos. Sci., 24, 414427.

    • Search Google Scholar
    • Export Citation
  • de Graaf, M., , P. Stammes, , O. Torres, , and R. B. A. Koelemeijer, 2005: Absorbing Aerosol Index: Sensitivity analysis, application to GOME and comparison with TOMS. J. Geophys. Res., 110, D01201, doi:10.1029/2004JD005178.

    • Search Google Scholar
    • Export Citation
  • de Graaf, M., , P. Stammes, , and E. A. A. Aben, 2007: Analysis of reflectance spectra of UV-absorbing aerosol scenes measured by SCIAMACHY. J. Geophys. Res., 112, D02206, doi:10.1029/2006JD007249.

    • Search Google Scholar
    • Export Citation
  • Deirmendjian, D., 1964: Scattering and polarization properties of water clouds and hazes in the visible and infrared. Appl. Opt., 3, 187196.

    • Search Google Scholar
    • Export Citation
  • Dirksen, R. J., , K. Folkert Boersma, , J. de Laat, , P. Stammes, , G. R. van der Werf, , M. Val Martin, , and H. M. Kelder, 2009: An aerosol boomerang: Rapid around-the-world transport of smoke from the December 2006 Australian forest fires observed from space. J. Geophys. Res., 114, D21202, doi:10.1029/2009JD012360.

    • Search Google Scholar
    • Export Citation
  • Dubovik, O., , and M. D. King, 2000: A flexible inversion algorithm for retrieval of aerosol optical properties from sun and sky radiance measurements. J. Geophys. Res., 105, 20 67320 696.

    • Search Google Scholar
    • Export Citation
  • Dubovik, O., , B. Holben, , T. F. Eck, , A. Smirnov, , Y. J. Kaufman, , M. D. King, , D. Tanré, , and I. Slutsker, 2002: Variability of absorption and optical properties of key aerosol types observed in worldwide locations. J. Atmos. Sci., 59, 590608.

    • Search Google Scholar
    • Export Citation
  • Duncan, B. N., , R. V. Martin, , A. C. Staudt, , R. Yevich, , and J. A. Logan, 2003: Interannual and seasonal variability of biomass burning emissions constrained by satellite observations. J. Geophys. Res., 108, 4100, doi:10.1029/2002JD002378.

    • Search Google Scholar
    • Export Citation
  • Eck, T. F., and Coauthors, 2003: Variability of biomass burning aerosol optical characteristics in southern Africa during the SAFARI 2000 dry season campaign and a comparison of single scattering albedo estimates from radiometric measurements. J. Geophys. Res., 108, 8477, doi:10.1029/2002JD002321.

    • Search Google Scholar
    • Export Citation
  • Fromm, M., , O. Torres, , D. Diner, , D. Lindsey, , B. Vant Hull, , R. Servranckx, , E. P. Shettle, , and Z. Li, 2008: Stratospheric impact of the Chisholm pyrocumulonimbus eruption: 1. Earth-viewing satellite perspective. J. Geophys. Res., 113, D08202, doi:10.1029/2007JD009153.

    • Search Google Scholar
    • Export Citation
  • Hansen, J. E., 1969: Exact and approximate solutions for multiple scattering by cloudy and hazy planetary atmospheres. J. Atmos. Sci., 26, 478487.

    • Search Google Scholar
    • Export Citation
  • Hao, W. M., , and M.-H. Liu, 1994: Spatial and temporal distribution of tropical biomass burning. Global Biogeochem. Cycles, 8, 495503.

    • Search Google Scholar
    • Export Citation
  • Haywood, J. M., , and K. P. Shine, 1997: Multi-spectral calculations of the direct radiative forcing of tropospheric sulphate and soot aerosols using a column model. Quart. J. Roy. Meteor. Soc., 123, 19071930.

    • Search Google Scholar
    • Export Citation
  • Haywood, J. M., , S. Osborne, , P. N. Francis, , A. Kiel, , P. Formenti, , T. W. Andreae, , and P. Kaye, 2003: The mean physical and optical properties of regional haze dominated by biomass burning aerosol measured from the C-130 aircraft during SAFARI 2000. J. Geophys. Res., 108, 8473, doi:10.1029/2002JD002226.

    • Search Google Scholar
    • Export Citation
  • Herman, J. R., , P. K. Bhartia, , O. Torres, , C. Hsu, , C. Seftor, , and E. Celarier, 1997: Global distribution of UV-absorbing aerosols from Nimbus 7/TOMS data. J. Geophys. Res., 102, 16 91116 922.

    • Search Google Scholar
    • Export Citation
  • Hoffer, A., , A. Gelencsér, , P. Guyon, , G. Kiss, , O. Schmid, , G. P. Frank, , P. Artaxo, , and M. O. Andreae, 2006: Optical properties of humic-like substances (HULIS) in biomass-burning aerosols. Atmos. Chem. Phys., 6, 35633570.

    • Search Google Scholar
    • Export Citation
  • Hsu, N. C., and Coauthors, 1996: Detection of biomass burning smoke from TOMS measurements. Geophys. Res. Lett., 23, 745748.

  • Hsu, N. C., , J. R. Herman, , and S.-C. Tsay, 2003: Radiative impacts from biomass burning in the presence of clouds during boreal spring in Southeast Asia. Geophys. Res. Lett., 30, 1224, doi:10.1029/2002GL016485.

    • Search Google Scholar
    • Export Citation
  • Hu, Y., , M. Vaughan, , Z. Liu, , K. Powell, , and S. Rodier, 2007: Retrieving optical depths and lidar ratios for transparent layers above opaque water clouds from CALIPSO lidar measurements. IEEE Geosci. Remote Sens. Lett., 4, 523526.

    • Search Google Scholar
    • Export Citation
  • Jaffe, D., , I. Bertschi, , L. Jaeglé, , P. Novelli, , J. S. Reid, , H. Tanimoto, , R. Vingarzan, , and D. L. Westphal, 2004: Long-range transport of Siberian biomass burning emissions and impact on surface ozone in western North America. Geophys. Res. Lett., 31, L16106, doi:10.1029/2004GL020093.

    • Search Google Scholar
    • Export Citation
  • Jethva, H., , and O. Torres, 2011: Satellite-based evidence of wavelength-dependent aerosol absorption in biomass burning smoke inferred from ozone monitoring instrument. Atmos. Chem. Phys. Discuss., 11, 72917319, doi:10.5194/acpd-11-7291-2011.

    • Search Google Scholar
    • Export Citation
  • Johnson, M. S., and Coauthors, 2010: Modeling dust and soluble iron deposition to the South Atlantic Ocean. J. Geophys. Res., 115, D15202, doi:10.1029/2009JD013311.

    • Search Google Scholar
    • Export Citation
  • Kallos, G., , A. Papadopoulos, , P. Katsafados, , and S. Nickovic, 2006: Transatlantic Saharan dust transport: Model simulation and results. J. Geophys. Res., 111, D09204, doi:10.1029/2005JD006207.

    • Search Google Scholar
    • Export Citation
  • Kaufman, Y. J., , I. Koren, , L. A. Remer, , D. Tanré, , P. Ginoux, , and S. Fan, 2005: Dust transport and deposition observed from the Terra-Moderate Resolution Imaging Spectroradiometer (MODIS) spacecraft over the Atlantic Ocean. J. Geophys. Res., 110, D10S12, doi:10.1029/2003JD004436.

    • Search Google Scholar
    • Export Citation
  • Keil, A., , and J. M. Haywood, 2003: Solar radiative forcing by biomass burning aerosol particles during SAFARI 2000: A case study based on measured aerosol and cloud properties. J. Geophys. Res., 108, 8467, doi:10.1029/2002JD002315.

    • Search Google Scholar
    • Export Citation
  • King, M. D., , S. Platnick, , C. C. Moeller, , H. E. Revercomb, , and D. A. Chu, 2003: Remote sensing of smoke, land, and clouds from the NASA ER-2 during SAFARI 2000. J. Geophys. Res., 108, 8502, doi:10.1029/2002JD003207.

    • Search Google Scholar
    • Export Citation
  • Kirchstetter, T. W., , T. Novakov, , and P. V. Hobbs, 2004: Evidence that the spectral dependence of light absorption by aerosols is affected by organic carbon. J. Geophys. Res., 109, D21208, doi:10.1029/2004JD004999.

    • Search Google Scholar
    • Export Citation
  • Kittaka, C., , D. M. Winker, , M. A. Vaughan, , A. Omar, , and L. A. Remer, 2011: Intercomparison of column aerosol optical depths from CALIPSO and MODIS-Aqua. Atmos. Meas. Tech., 4, 131141.

    • Search Google Scholar
    • Export Citation
  • Liao, H., , and J. H. Seinfeld, 1998a: Effect of clouds on direct aerosol radiative forcing of climate. J. Geophys. Res., 103, 37813788.

    • Search Google Scholar
    • Export Citation
  • Liao, H., , and J. H. Seinfeld, 1998b: Radiative forcing by mineral dust aerosol: Sensitivity to key variables. J. Geophys. Res., 103, 31 63731 646.

    • Search Google Scholar
    • Export Citation
  • Penning de Vries, M. J. M., , S. Beirle, , and T. Wagner, 2009: UV aerosol indices from SCIAMACHY: Introducing the Scattering Index (SCI). Atmos. Chem. Phys., 9, 95559567.

    • Search Google Scholar
    • Export Citation
  • Peters, K., , J. Quaas, , and N. Bellouin, 2009: Effects of absorbing aerosols in cloudy skies: A satellite study over the Atlantic Ocean. Atmos. Chem. Phys., 11, 13931404, doi:10.5194/acp-11-1393-2011.

    • Search Google Scholar
    • Export Citation
  • Platnick, S., , M. D. King, , A. Ackerman, , W. P. Menzel, , B. A. Baum, , J. C. Riedi, , and R. A. Frey, 2003: The MODIS cloud products: Algorithms and examples from Terra. IEEE Trans. Geosci. Remote Sens., 41, 459473.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., , and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 22612287.

  • Russell, P. B., and Coauthors, 2005: Aerosol optical depth measurements by airborne sun photometer in SOLVE II: Comparison to SAGE III, POAM III and airborne spectrometer measurements. Atmos. Chem. Phys., 5, 13111339.

    • Search Google Scholar
    • Export Citation
  • Russell, P. B., and Coauthors, 2010: Absorption angstrom exponent in AERONET and related data as an indicator of aerosol composition. Atmos. Chem. Phys., 10, 11551169, doi:10.5194/acp-10-1155-2010.

    • Search Google Scholar
    • Export Citation
  • Spurr, R. J. D., 2006: VLIDORT: A linearized pseudo-spherical vector discrete ordinate radiative transfer code for forward model amd retrieval studies in multilayer multiple scattering media. J. Quant. Spectrosc. Radiat. Transfer, 102, 316342.

    • Search Google Scholar
    • Export Citation
  • Stammes, P., , M. Sneep, , J. F. de Haan, , J. P. Veefkind, , P. Wang, , and P. Levelt, 2008: Effective cloud fractions from the Ozone Monitoring Instrument: Theoretical framework and validation. J. Geophys. Res., 113, D16S38, doi:10.1029/2007JD008820.

    • Search Google Scholar
    • Export Citation
  • Swap, R. J., , H. J. Annegarn, , J. T. Suttles, , M. D. King, , S. Platnick, , J. L. Privette, , and R. J. Scholes, 2003: Africa burning: A thematic analysis of the Southern African Regional Science Initiative (SAFARI 2000). J. Geophys. Res., 108, 8465, doi:10.1029/2003JD003747.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , P. K. Bhartia, , J. R. Herman, , Z. Ahmad, , and J. Gleason, 1998: Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: Theoretical basis. J. Geophys. Res., 103, 17 09917 110.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , P. K. Bhartia, , J. R. Herman, , A. Sinyuk, , and B. Holben, 2002: A long-term record of aerosol optical thickness from TOMS observations and comparison to AERONET measurements. J. Atmos. Sci., 59, 398413.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , A. Tanskanen, , B. Veihelman, , C. Ahn, , R. Braak, , P. K. Bhartia, , P. Veefkind, , and P. Levelt, 2007: Aerosols and surface UV products from OMI observations: An overview. J. Geophys. Res., 112, D24S47, doi:10.1029/2007JD008809.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , Z. Chen, , H. Jethva, , C. Ahn, , R. Freitas, , and P. K. Bhartia, 2010: OMI and MODIS observations of the anomalous 2008–2009 Southern Hemisphere biomass burning seasons. Atmos. Chem. Phys., 10, 35053513.

    • Search Google Scholar
    • Export Citation
  • Waquet, F., , J. Riedi, , L. C. Labonnote, , P. Goloub, , B. Cairns, , J. L. Deuzé, , and D. Tanré, 2009: Aerosol remote sensing over clouds using A-Train observations. J. Atmos. Sci., 66, 24682480.

    • Search Google Scholar
    • Export Citation
  • Wen, G., , A. Marshak, , and R. F. Cahalan, 2008: Importance of molecular Rayleigh scattering in the enhancement of clear sky reflectance in the vicinity of boundary layer cumulus clouds. J. Geophys. Res., 113, D24207, doi:10.1029/2008JD010592.

    • Search Google Scholar
    • Export Citation
  • Wilcox, E. M., 2010: Stratocumulus cloud thickening beneath layers of absorbing smoke aerosol. Atmos. Chem. Phys., 10, 11 76911 777, doi:10.5194/acp-10-11769-2010.

    • Search Google Scholar
    • Export Citation
  • Wilcox, E. M., , Harshvardhan, , and S. Platnick, 2009: Estimate of the impact of absorbing aerosol over cloud on the MODIS retrievals of cloud optical thickness and effective radius using two independent retrievals of liquid water path. J. Geophys. Res., 114, D05210, doi:10.1029/2008JD010589.

    • Search Google Scholar
    • Export Citation
  • Zarzycki, C. M., , and T. M. Bond, 2010: How much can the vertical distribution of black carbon affect its global direct radiative forcing? Geophys. Res. Lett., 37, L20807, doi:10.1029/2010GL044555.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (top) OMI UV-AI, (middle) CALIOP backscatter, and (bottom) MODIS true color image composite for an aerosol event observed on 4 Aug 2007 off the coast of Central Africa and over the southeastern Atlantic Ocean. The CALIOP track (thick black line) is also shown on UV-AI and RGB images.

  • View in gallery

    TOA reflectance sensitivity as a function of cloud optical depth for no-aerosol and for a BC carbonaceous aerosol-laden atmosphere. The cloud layer is located between 1.0 and 1.5 km and the aerosol layer is at 3 km above the surface. Calculations were carried out for solar and satellite zenith angles of 20° and 32°, respectively, and the relative azimuth angle 120°.

  • View in gallery

    The UV–aerosol index sensitivity to varying cloud optical depth for no-aerosol, gray aerosol, and colored aerosol models. Aerosol and cloud location and viewing geometry are as in Fig. 2.

  • View in gallery

    Relative change in the UV–aerosol index sensitivity for varying aerosol mean height under different values of cloud optical depth. Shown in inset is the derivative of the aerosol index with regard to height as a function of cloud optical depth.

  • View in gallery

    Retrieval domain of AAI vs reflectance at 388 nm simulated using a radiative transfer model for several values of AOD at 388 nm (0.0, 0.156, 0.780, 1.56, 3.9, 6.24, and 9.36) and COD (0.0, 2, 5, 10, 20, 30, 40, and 50).

  • View in gallery

    Comparison of MODIS and OMI cloud optical depth retrieved on 13 Oct 2006. No aerosols were present above clouds during this event. The dotted line indicates the one-to-one agreement, the solid line is the linear fit, and the dashed lines represent the ±20% uncertainty of the MODIS retrieval for CODs larger than 5. Significantly larger uncertainty exists for COD smaller than 5 (Platnick et al. 2003).

  • View in gallery

    Spatial distribution of above-cloud (left) AOD (388 nm) and (right) COD for the 4 Aug 2007 aerosol event.

  • View in gallery

    As in Fig. 6, but for the aerosol event over the southeastern Atlantic Ocean on 4 Aug 2007.

  • View in gallery

    Relation between coincident aerosol optical depth and UV aerosol index measured by OMI on 4 Aug 2007 over the study region.

  • View in gallery

    Composite of above-cloud and clear region AOD derived using cloud–aerosol and aerosol-only OMAERUV algorithms on 4 Aug 2007.

  • View in gallery

    Retrieved (left) AOD and (right) COD fields on 31 Aug 2005.

  • View in gallery

    As in Fig. 6, but for the aerosol event over the southeastern Atlantic Ocean on 31 Aug 2005.

  • View in gallery

    Scatterplot of retrieved 865-nm AOD from POLDER observation (x axis) and 388-nm AOD from OMI measurements on 4 Aug 2008. The difference in magnitude is related to the small aerosol particle size and difference in reporting wavelengths (see text for details).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 169 169 29
PDF Downloads 129 129 22

Retrieval of Aerosol Optical Depth above Clouds from OMI Observations: Sensitivity Analysis and Case Studies

View More View Less
  • 1 NASA Goddard Space Flight Center, Greenbelt, Maryland
  • | 2 Department of Atmospheric and Planetary Sciences, Hampton University, Hampton, Virginia
  • | 3 NASA Goddard Space Flight Center, Greenbelt, Maryland
© Get Permissions
Full access

Abstract

A large fraction of the atmospheric aerosol load reaching the free troposphere is frequently located above low clouds. Most commonly observed aerosols above clouds are carbonaceous particles generally associated with biomass burning and boreal forest fires, and mineral aerosols originating in arid and semiarid regions and transported across large distances, often above clouds. Because these aerosols absorb solar radiation, their role in the radiative transfer balance of the earth–atmosphere system is especially important. The generally negative (cooling) top-of-the-atmosphere direct effect of absorbing aerosols may turn into warming when the light-absorbing particles are located above clouds. The actual effect depends on the aerosol load and the single scattering albedo, and on the geometric cloud fraction. In spite of its potential significance, the role of aerosols above clouds is not adequately accounted for in the assessment of aerosol radiative forcing effects due to the lack of measurements. This paper discusses the basis of a simple technique that uses near-UV observations to simultaneously derive the optical depth of both the aerosol layer and the underlying cloud for overcast conditions. The two-parameter retrieval method described here makes use of the UV aerosol index and reflectance measurements at 388 nm. A detailed sensitivity analysis indicates that the measured radiances depend mainly on the aerosol absorption exponent and aerosol–cloud separation. The technique was applied to above-cloud aerosol events over the southern Atlantic Ocean, yielding realistic results as indicated by indirect evaluation methods. An error analysis indicates that for typical overcast cloudy conditions and aerosol loads, the aerosol optical depth can be retrieved with an accuracy of approximately 54% whereas the cloud optical depth can be derived within 17% of the true value.

Corresponding author address: Omar Torres, NASA Goddard Space Flight Center, Code 6140, Greenbelt, MD 20771. E-mail: omar.o.torres@nasa.gov

Abstract

A large fraction of the atmospheric aerosol load reaching the free troposphere is frequently located above low clouds. Most commonly observed aerosols above clouds are carbonaceous particles generally associated with biomass burning and boreal forest fires, and mineral aerosols originating in arid and semiarid regions and transported across large distances, often above clouds. Because these aerosols absorb solar radiation, their role in the radiative transfer balance of the earth–atmosphere system is especially important. The generally negative (cooling) top-of-the-atmosphere direct effect of absorbing aerosols may turn into warming when the light-absorbing particles are located above clouds. The actual effect depends on the aerosol load and the single scattering albedo, and on the geometric cloud fraction. In spite of its potential significance, the role of aerosols above clouds is not adequately accounted for in the assessment of aerosol radiative forcing effects due to the lack of measurements. This paper discusses the basis of a simple technique that uses near-UV observations to simultaneously derive the optical depth of both the aerosol layer and the underlying cloud for overcast conditions. The two-parameter retrieval method described here makes use of the UV aerosol index and reflectance measurements at 388 nm. A detailed sensitivity analysis indicates that the measured radiances depend mainly on the aerosol absorption exponent and aerosol–cloud separation. The technique was applied to above-cloud aerosol events over the southern Atlantic Ocean, yielding realistic results as indicated by indirect evaluation methods. An error analysis indicates that for typical overcast cloudy conditions and aerosol loads, the aerosol optical depth can be retrieved with an accuracy of approximately 54% whereas the cloud optical depth can be derived within 17% of the true value.

Corresponding author address: Omar Torres, NASA Goddard Space Flight Center, Code 6140, Greenbelt, MD 20771. E-mail: omar.o.torres@nasa.gov

1. Introduction

Because of the nature of the physical processes driving the emission and atmospheric injection of desert dust and carbonaceous aerosols (i.e., the wind’s lifting power of soil particles over arid and semiarid areas and the strong convective activity associated with anthropogenic biomass burning and wild fire events), large amounts of these light-absorbing particles reach the free troposphere more often than other aerosol types that generally reside in the boundary layer such as sulfate and sea salt aerosols. Elevated layers of desert dust and carbonaceous particulate are therefore frequently observed above clouds where they are mobilized by the prevailing winds and transported thousands of kilometers away from their original sources. The spring and summer transoceanic transport of desert dust across the Atlantic Ocean and the spring flow of aerosol from Asian sources across the Pacific Ocean to North America and sometimes reaching Northern Europe have been well documented using both satellite observations (Kaufman et al. 2005) and model calculations (Kallos et al. 2006; Johnson et al. 2010). Smoke layers originate from well-known regional sources of agricultural biomass burning mainly in Africa and South America (Duncan et al. 2003; Torres et al. 2010). Smoke layers can also be the result of wild fires in both hemispheres (Colarco et al. 2004; Dirksen et al. 2009) that are transported across the oceans or from northern regions to the midlatitudes from their original distant remote sources to populated centers where they often contribute to the observed high levels of local pollution (Jaffe et al. 2004). Elevated aerosol layers, often lofted above clouds, are also observed in connection with pyroconvective events (Fromm et al. 2008). Much of the long-range transport of desert dust and carbonaceous aerosols, however, takes place above low-level layers of altostratus and stratocumulus clouds.

Quantifying the fraction of the atmospheric aerosol load above clouds is important for air quality considerations as some of these aerosols eventually descend into the boundary layer, contributing to the enhancement of pollution levels. Knowledge of the amount and type of aerosol above clouds is also of interest for radiative forcing considerations. The general cooling effect of these aerosol types under clear-sky conditions may turn into a warming effect when located above clouds. Depending on their light absorption capacity, above-cloud aerosols may have a positive radiative forcing effect at the top of the atmosphere (TOA) (Keil and Haywood 2003; Haywood and Shine 1997; Chand et al. 2008), whose magnitude depends on the aerosol single scattering albedo (SSA) and cloud fraction in addition to the total aerosol load. Liao and Seinfeld (1998a,b) examined the radiative forcing of desert dust and found that the highest radiative forcing occurs for dust layers located above clouds. This effect is particularly important for carbonaceous aerosols in which the combined absorption effect of its organic and black carbon components acts on the incoming solar radiation from the UV to the near-IR, producing potentially large effects on the local climate. The radiative effect of carbonaceous aerosols above stratiform clouds off the coasts of Angola and Namibia was estimated during the Southern African Regional Science Initiative (SAFARI) 2000 field experiment by making use of radiative transfer calculations (Keil and Haywood 2003) and airborne-measured profiles of aerosol particle number concentration, scattering coefficients, and meteorological parameters (Haywood et al. 2003). The study indicated that the normal cooling effect of the smoke aerosol layer of about −13 W m−2 over a cloud-free area turned into a warming effect of 11.5 W m−2 for the same aerosol layer located above a cloud deck, which is similar to the expected radiative forcing effect associated with an increase in surface albedo. Changes in cloud morphology may also result as a consequence of atmospheric heating (by aerosol absorption) of the air above the cloud deck. As suggested by Wilcox (2010), this aerosol-induced heating of the atmosphere may lead to the optical thickening of the cloud resulting from enhanced liquid water path and lower cloud top associated with the warming above the cloud.

Agriculture-related biomass burning in the tropical regions accounts for nearly 80% of the global amount including both anthropogenic and naturally produced burning (Hao and Liu 1994). It is estimated that although BC above clouds makes up 20% of the total burden, it accounts for about 50% of the BC radiative forcing effect (Zarzycki and Bond 2010). To properly estimate the regional and global forcing effect of carbonaceous and desert dust aerosols it is therefore necessary to quantify the atmospheric aerosol load above clouds.

In this paper we present a method of deriving aerosol optical depth (AOD) over clouds making use of near-UV observations by the Ozone Monitoring Instrument (OMI) on the Aura satellite. The high near-UV sensitivity to aerosol absorption (Torres et al. 1998, 2007) is further enhanced in the presence of underlying clouds because of the high reflectance from the cloud-top layer. A general description of the near-UV capability of aerosol detection is presented in section 2 followed by a detailed discussion of a sensitivity analysis of near-UV observations to aerosol presence above clouds in section 3. A description of the retrieval procedure and error analysis is presented in section 4, and results of application of the inversion procedure to two case studies are discussed in section 5. Discussion and conclusions are presented in section 6.

2. Satellite detection of aerosols above clouds

The detection and characterization of aerosols in cloud-free scenes using near-UV satellite observations is an established aerosol remote sensing technique developed based on Total Ozone Mapping Spectrometer (TOMS) observations (Torres et al. 1998). The near-UV technique is particularly sensitive to absorbing aerosols as explained below. A long-term record (1979–92 and 1996–2001) of aerosol optical depth (Torres et al. 2002) was developed from TOMS observations. Currently, near-UV observations are used in the OMI UV aerosol algorithm (OMAERUV) to derive aerosol optical depth and single scattering albedo under cloud-free conditions (Ahn et al. 2008; Torres et al. 2007). In this paper we describe initial results of a recently developed approach using OMI observations to derive the optical depth of aerosol layers above clouds.

Aerosols above clouds, or for that matter above any bright background, are not easily detectable by conventional satellite remote sensing approaches that measure aerosols based on the increased reflectance associated with particle scattering. The difficulty lies on the fact that aerosol scattering contribution to the total upwelling reflectance at the top of the atmosphere is very small compared to the reflectance of the bright cloud background. Thus, to be able to separate the aerosol scattering signal from the measured total, one should be able to characterize the reflective properties of the bright background (i.e., clouds) to a level of accuracy probably unachievable. Aerosol absorption effects, however, can be advantageously used for the detection and characterization of absorbing particles even over highly reflective backgrounds.

It has been demonstrated that aerosols that absorb UV radiation can be detected and characterized using the unambiguous particle absorption signal in the measured levels of UV reflectance (Torres et al. 1998). The unique advantage of the near-UV detection capability of absorbing aerosols is the large Rayleigh scattering component that makes available for absorption by aerosol particles a significantly larger number of photons than at longer wavelengths (Torres et al. 1998). The most commonly known near-UV aerosol product is the absorbing aerosol index (AAI), a parameter used to detect UV-absorbing aerosols over ocean and all land surfaces.

The AAI was developed as a by-product of improvements to the column ozone amount retrieval algorithm applied to UV observations by the TOMS instrument. It is basically a residual quantity resulting from the comparison between measured and calculated radiances in the range 330–390 nm where trace gas absorption effects are negligible. The calculated radiance is obtained using a simple model of the earth–atmosphere system consisting of a molecular atmosphere bounded at the bottom by a Lambert equivalent reflector (LER) (Dave and Mateer 1967). The LER reflectivity ρ is derived from measurements at wavelength λ0. A key assumption in this model representation of the earth–atmosphere system is that the reflectivity of the column atmosphere’s lower boundary is wavelength independent in the near UV. This hypothetical surface is intended to account for the wavelength-dependent effects of surface, clouds, and aerosols that are not explicitly included in the radiative transfer calculations.

The residue rλ is defined as the log of the ratio of the actually measured radiance to the calculated value assuming an effective surface reflectivity :
e1
This residual quantity provides a measure of the error in predicting the actually observed spectral contrast using the LER model. Examination of global maps of the residue clearly indicate that UV-absorbing aerosols are by far the most important residue source (Hsu et al. 1996; Herman et al. 1997); this has become a powerful tool of absorbing aerosols detection that has been used in a variety of applications. Hence, the term “absorbing aerosol index” was coined to refer to this residual quantity. The AAI is the same as rλ when λ0 > λ, and it is defined as −rλ when λ0 < λ so that it is always positive in the presence of absorbing aerosols. In the OMAERUV algorithm AAI is calculated for λ0 = 388 nm and λ = 354 nm.

The AAI detection capability of absorbing aerosols is not limited to cloud-free, dark scenes. It can also identify the presence of absorbing particles above bright backgrounds such as ice–snow-covered surfaces and cloud decks. Figure 1 shows a typical aerosol-above-cloud event on 4 August 2007 over the southern Atlantic Ocean as seen by three different A-Train sensors. The top panel depicts the OMI UV absorbing aerosol index indicating the unmistakable presence of an absorbing aerosol layer above the horizontally extended cloud feature shown in the Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) true color image composite (bottom panel). The cloud layer covers a large region from the coast of Namibia and Angola extending into the ocean in a northwest direction. The center panel illustrates the vertical distribution of clouds and aerosols in terms of the Cloud–Aerosol Lidar with Orthogonal Polarization (CALIOP) 532-nm attenuated backscatter measured along the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) spacecraft orbital track shown on both the OMI and MODIS images. The CALIOP’s transect shows the presence of a geometrically thin low-level cloud with cloud top at 1 km above sea level. An aerosol layer at about 2 km above the cloud is observed over the first half of the image. Toward the middle of the shown orbital section the aerosol layer is observed over a cloud-free area.

Fig. 1.
Fig. 1.

(top) OMI UV-AI, (middle) CALIOP backscatter, and (bottom) MODIS true color image composite for an aerosol event observed on 4 Aug 2007 off the coast of Central Africa and over the southeastern Atlantic Ocean. The CALIOP track (thick black line) is also shown on UV-AI and RGB images.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

TOMS and OMI AAI observations over clouds have been used in conjunction with other satellite measurements to estimate the radiative effects of aerosol layers above clouds (Hsu et al. 2003; Peters et al. 2009). Other applications include the use of Scanning Imaging Absorption Spectrometer for Atmospheric Cartography (SCIAMACHY)-derived AAI for identification of aerosol layers above cloud and the analysis of aerosol absorption effect on visible and near-IR TOA radiance measurements by the same sensor (de Graaf et al. 2007).

Aerosols above clouds have also been detected and their optical depth estimated by polarization measurements by the Polarization and Directionality of Earth Reflectances (POLDER) sensor (Waquet et al. 2009). This approach uses multisensor measurements aboard the A-Train constellation to derive aerosol optical depth above cloud taking advantage of the unique angular dependence associated with the presence of the polarized radiation of above-cloud aerosols. The modification in the polarized light reflected by cloud top by the aerosol layer permits retrieving aerosol optical depth using a single scattering approximation for aerosol signal.

Active spaceborne measurements can also be used to derive the optical depth of aerosols above clouds. The standard CALIOP AOD product is derived as the vertically integrated aerosol extinction coefficient over the vertical extent of the identified aerosol feature. This approach requires an assumption of the extinction-to-backscatter or lidar ratio (Kittaka et al. 2011). Alternate transmission methods for directly deriving AOD above clouds from CALIOP observations of depolarization (Hu et al. 2007) and color (Chand et al. 2008) ratios have been investigated. Unlike the standard approach, the transmission-based lidar techniques do not require assumption on the lidar ratio.

3. Sensitivity analysis

a. Aerosol and cloud models

In this section we discuss the results of a sensitivity analysis of the near-UV reflectance at the top of the atmosphere to aerosol and cloud properties. Radiative transfer calculations at 354 and 388 nm were carried out for an atmosphere containing both aerosols and clouds using the Vector Linearized Discrete Ordinate Radiative Transfer code (VLIDORT) developed by Spurr (2006). VLIDORT calculates Jacobians and radiances for reflectances at the top of a multilayered atmosphere. It takes into account the atmosphere’s curvature making use of a pseudospherical approximation for solar beam attenuation and includes a sphericity correction for off-nadir viewing at large angles. The VLIDORT code accounts for multiple scattering effects by molecules, aerosols, and clouds.

In the following discussion, the cloud droplet size distribution is represented by the C1 modified gamma distribution suggested by Deirmendjian (1964), which is commonly used in cloud radiative transfer modeling (Rossow and Schiffer 1999; Ahmad et al. 2004). The assumed distribution with a maximum radius of 15 μm yields a 6.0-μm radiatively effective radius. Two aerosol models representative of carbonaceous aerosols are used in the calculations. The assumed bimodal particle size distributions and complex refractive index of the selected aerosol models are based on multiyear Aerosol Robotic Network (AERONET) statistics of measurements at locations where carbonaceous aerosols are typically observed (Dubovik et al. 2002). AERONET retrievals of aerosol particle size and optical properties use both direct sun and sky radiance measurements as input to an inversion algorithm (Dubovik and King 2000). The reported measurements are representative of the atmospheric column aerosol load and are therefore suitable for the analysis of spaceborne remotely sensed data. The first model, representative of black carbon (BC), has a wavelength independent imaginary refractive index, while the second one assumes a wavelength dependent absorption refractive index associated with organic carbon (OC) consistent with observations (Kirchstetter et al. 2004). Table 1 lists the aerosol parameters used in the calculations.

Table 1.

Aerosol parameters: mode radius r, standard deviation σ, and particle concentration N for fine and coarse modes, and refractive index. The corresponding SSA is 0.85 for BC (wavelength independent) and 0.85 and 0.84 at 388 and 354 nm, respectively, for the OC model.

Table 1.

A 0.5-km-thick cloud located between 1 and 1.5 km above sea level (900 and 850 hPa, respectively) is assumed. The aerosol vertical distribution is represented by a single layer whose concentration follows a Gaussian distribution characterized by the height of maximum concentration (peak) and full-width-half-maximum. Note that this yields an aerosol profile in which the entire aerosol load is located above the cloud. The above described vertical distribution of clouds and aerosols resembles actual conditions observed by the CALIOP lidar over the southern Atlantic Ocean as shown in Fig. 1.

b. Reflectance sensitivity to aerosol type and cloud parameters

The sensitivity of the satellite measured reflectance at 354 and 388 nm to cloud-top height was first examined in the no-aerosol case. For a cloud of optical depth 10, systematically varying cloud-top level between 950 and 100 mb produced maximum differences of −3% reflectance at 354 nm and −1% at 388 nm resulting from the shielding by clouds of molecular scattering. Negligible reflectance change was observed associated with variations in cloud drop effective radius.

Figure 2 shows the calculated top of the atmosphere reflectance at 388 nm as a function of cloud optical depth (COD) for a column atmosphere containing cloud layer between 1.0 and 1.5 km and an aerosol layer with peak particle concentration at 3 km above the surface and a full-width at half maximum of 0.5 km. Results for a no-aerosol case and two different absorbing aerosol models of optical depth 1 (at 388 nm) are presented.

Fig. 2.
Fig. 2.

TOA reflectance sensitivity as a function of cloud optical depth for no-aerosol and for a BC carbonaceous aerosol-laden atmosphere. The cloud layer is located between 1.0 and 1.5 km and the aerosol layer is at 3 km above the surface. Calculations were carried out for solar and satellite zenith angles of 20° and 32°, respectively, and the relative azimuth angle 120°.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

The solid line in Fig. 2 depicts the relationship between the calculated TOA 388-nm reflectance and COD between 0 and 50. The observed reflectance increases monotonically with COD as expected. The cloud diffuse radiation is by far the most important source of the observed TOA reflectance enhancement. Although molecular and aerosol scattering are also present (Wen et al. 2008), their contribution to the observed TOA reflectance for overcast conditions, which may be important for optically thin clouds, is negligible at large values of cloud optical depth.

The results associated with an aerosol layer of single scattering albedo of 0.95 (at 388 nm) representative of BC are indicated by the dotted line in Fig. 2. For weakly absorbing aerosols, scattering dominates over absorption effects and the net result is an increase of TOA reflectance values above those of the cloud-only case for optically thin clouds. Aerosol scattering and absorption effects cancel each other at a COD value of about 5. At larger COD values the increased cloud reflectance triggers additional aerosol absorption, resulting in TOA reflectance values lower than the ones for the no-aerosol case. For the largest modeled COD (50) the weakly absorbing aerosol produces a 16% TOA decrease in relation to the no-aerosol case. For a stronger absorbing aerosol case (SSA = 0.85) the absorption effect dominates over scattering even at low COD values as indicated by the dashed line in Fig. 2. For a COD value of 50, the observed TOA reflectance is reduced by 33%.

c. Aerosol index sensitivity to spectral aerosol absorption

The resulting AAI values for clouds only and above-clouds absorbing aerosol layers are shown in Fig. 3. For clouds only the AAI is slightly negative, reaching a value of −0.6 at COD 5. It then increases with increasing COD, becoming zero at COD 20 and reaching about 0.25 at COD 50. Although the AAI of clouds depends on observing geometry (solar and viewing zenith angles, and relative azimuth angle), for solar zenith angles lower than 60 it oscillates around zero between −1.0 and 1.0 (not shown) with the slightly negative values associated with COD smaller than about 20 and small positive values for larger CODs. These results shown in Fig. 3 are consistent with previously reported sensitivity analysis using Mie theory (Torres et al. 1998; Herman et al. 1997) that show the weak spectral dependence associated with the scattering effects of nonabsorbing large-size particles. For typical AOD values, large aerosol particles yield AAIs close to zero whereas for nonabsorbing aerosol particles smaller than about 0.3 μm and typical AOD values the resulting AAI is negative, reaching values as low as about −1 for large aerosol loads. The slightly larger cloud-only AAI values in Fig. 3 than reported by Torres et al. (1998) are due to the much larger optical depth of clouds.

Fig. 3.
Fig. 3.

The UV–aerosol index sensitivity to varying cloud optical depth for no-aerosol, gray aerosol, and colored aerosol models. Aerosol and cloud location and viewing geometry are as in Fig. 2.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

The small AAI values calculated for cloud droplets in the current work differ from recently published calculations (Penning de Vries et al. 2009) in which clouds are reported to produce negative AAI values (as large as −1.5) for all values of COD. A likely explanation for the noted discrepancy is the Penning de Vries et al. approximate representation of the angular dependence of scattering by clouds and aerosols using a Henyey–Greenstain (H–G) phase function instead of actual Mie calculations. Using the H–G function to represent clouds does not allow us to account for the angular effects associated with varying size of the scattering particle (Hansen 1969).

For absorbing aerosols located above clouds, the radiative coupling among cloud diffuse radiation, molecular scattering, and particle absorption enhances the observed near-UV spectral contrast, and therefore the resulting AAI is larger than the one associated with the same aerosol layer under cloud-free conditions. The AAI magnitude will also depend on the spectral dependence of aerosol absorption as determined by the wavelength dependence of the imaginary component of the index of refraction. The effect of the above-cloud absorbing aerosols on the calculated AAI as a function of COD for several values of the angstrom absorption exponent (AAE) (Bond 2001) and SSA 0.85 (388 nm) is also shown in Fig. 3. In all cases the calculated positive AAI is a strong function of the COD of the underlying cloud layer. The slope of the observed COD-AAI relationship decreases for COD values larger than 20. The resulting AAI values increase rapidly as the AAE varies from a value of 1.0 (dotted line), an indicator of nonwavelength dependent imaginary refractive index (representative of BC), to values of 2.0 (dot-dashed line) and 2.8 (dashed line) representative of wavelength-dependent aerosol absorption. A detailed analysis of the AAI sensitivity to spectral aerosol absorption is presented in Jethva and Torres (2011).

The presence of organic carbon in the carbonaceous aerosols produced by biomass burning has been demonstrated by a variety of laboratory and field studies (Kirchstetter et al. 2004; Hoffer et al. 2006; Russell et al. 2010). A recent analysis combining OMI near-UV observations, radiative transfer calculations, and ground-based measurements of aerosol optical depth (Jethva and Torres 2011) has further confirmed the presence of OC as an important aerosol component of biomass-burning generated smoke. Jethva and Torres (2011) showed that allowing for the wavelength dependence of the imaginary refractive index in the near UV (an indication of OC presence) produced significantly better agreement of OMI-retrieved AOD with sun photometer observations than when using an aerosol model with the wavelength-independent imaginary refractive index characteristic of black carbon. Thus, the OC aerosol model has been selected as more representative of the actual carbonaceous aerosol load in the OMAERUV algorithm. In the remainder of this work we will use the OC aerosol model.

d. Aerosol index sensitivity to aerosol–cloud separation

The AAI dependence on aerosol layer height over cloudless scenes is a well-documented AAI property (Torres et al. 1998; De Graaf et al. 2005). We will now examine the resulting height dependence when the absorbing aerosol layer lies above a cloud deck of varying reflectance as determined by the COD. As in the preceding analysis, the cloud layer is located between 1.0 and 1.5 km. The height of the absorbing aerosol layer is assumed to be 2.5, 3.0, 4.0, and 5.0 km above the surface. The resulting relative change in AAI value as a function of aerosol layer height and COD is depicted in Fig. 4. The largest percent change is observed for the cloud-free case (COD = 0) and quickly goes down with increasing COD. For a COD value of 30, the AAI associated with a height change of 2.5 km is about 25%, compared to more than 75% for the cloud-free case. Thus, unlike in the cloud-free case where the aerosol layer height is a large source of uncertainty for the quantitative interpretation of the AAI (Torres et al. 1998), in the presence of underlying clouds the AAI sensitivity to aerosol layer height is reduced. The inset in Fig. 4b illustrates the rate of change of AAI per 1-km change in aerosol layer height as a function of cloud optical depth. The observed rapid drop in AAI sensitivity to cloud–aerosol separation with increasing COD is an advantage from a retrieval point of view as it introduces less uncertainty in the derivation of AOT due to wrong assumptions about the aerosol height above the cloud.

Fig. 4.
Fig. 4.

Relative change in the UV–aerosol index sensitivity for varying aerosol mean height under different values of cloud optical depth. Shown in inset is the derivative of the aerosol index with regard to height as a function of cloud optical depth.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

4. Retrieval approach

a. Simultaneous aerosol–cloud retrieval

In the preceding sections it has been shown that the magnitude of the observed AAI associated with aerosol layers above clouds depends on the reflectance of the underlying cloud (COD), the aerosol load (AOD) and single scattering albedo, the AAE parameter, and the aerosol–cloud separation. For testing of the current research retrieval algorithm the values of some of these parameters can be prescribed based on independent observations. Aerosol–cloud separation, for instance, is obtainable from spaceborne CALIOP lidar profiles. A justifiable assumption of the 388-nm single-scattering albedo can be made from retrievals by the cloud-free OMAERUV algorithm (Torres et al. 2007) in the vicinity of the cloud–aerosol feature under investigation if one assumes this value is representative of the SSA of the above-cloud aerosol load. Assumption of the value of the AAE parameter is required. In principle, COD information from other sensors in the A-Train (i.e., MODIS) could be used as input to the inversion algorithm. This option, however, implies the use of erroneous COD data as the MODIS cloud optical product will be most probably also affected by the absorption effects of the overlying aerosol layer (Wilcox et al. 2009). In this work we demonstrate a retrieval approach that accounts for the spatial variability of clouds and aerosols by simultaneously deriving AOD and COD over overcast scenes covered by opaque clouds.

The retrieval scheme uses a set of precalculated values of AAI and 388-nm reflectances for several values of COD (0, 2, 5, 10, 20, 30, 40, 50) and AOD (0, 0.1, 0.5, 1.0, 2.5, 4.0) for a fixed value of SSA and assumed AAE. Cloud and aerosol layer heights are prescribed based on CALIPSO observations. With the above stated assumptions the interpolating domain reduces to a two-dimensional array on COD and AOD as illustrated in Fig. 5. As shown in the web-like diagram, an observation set of AAI and the 388-nm reflectance is associated with a set of values of COD and AOD.

Fig. 5.
Fig. 5.

Retrieval domain of AAI vs reflectance at 388 nm simulated using a radiative transfer model for several values of AOD at 388 nm (0.0, 0.156, 0.780, 1.56, 3.9, 6.24, and 9.36) and COD (0.0, 2, 5, 10, 20, 30, 40, and 50).

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

b. Error analysis

The accuracy of the retrieved parameters depends on the uncertainty associated with the assumed values of aerosol layer height above the cloud and the wavelength-dependent aerosol single scattering albedo. In this section we discuss the results of an error analysis assuming a set of “true” conditions consisting of an aerosol layer of known optical depth and single scattering albedo located at a height Z above the surface, and a 0.5-km-thick cloud layer with cloud top at 1.5 km and known optical depth. The assumed cloud–aerosol vertical structure is representative of the typically observed cloud–aerosol layers over the southern Atlantic Ocean off the coasts of Angola and Namibia during that region’s biomass burning season (King et al. 2003; Chand et al. 2009). The uncertainty of the spectrally dependent single scattering albedo is evaluated by assuming a fixed SSA value at 388 nm and an associated AAE value that describes the wavelength dependence of the aerosol absorption optical depth. The true value of Z is assumed to be 4 km. An SSA value of 0.85 typical of biomass burning smoke aerosols for this region (Eck et al. 2003) and an AAE value of 2.2 as reported by Kirchstetter et al. (2004) during the SAFARI 2000 field campaign are assumed. Radiative transfer calculations were carried out for a nominal AOD value of 0.5 and COD values of 5 and 10. The assumed values of Z and SSA are perturbed by ±2 km and ±0.03, respectively, which represent typical variability associated with these parameters (Chand et al. 2009; Eck et al. 2003) over the southern Atlantic region. The assumed AAE parameter is varied between 1.8 and 2.6 or ±0.4 to account for possible variations in aerosol composition (i.e., OC content). Although the actual natural variability of AAE is largely unknown, the tested range allows us to accounting for some uncertainty in its actual value. The AAE perturbation is equivalent to a change from the nominal spectral dependence of 15% in the values of the imaginary refractive index (i.e., k354 = 1.15k388) to 10% and 20%.

Table 2 shows the obtained AOD and COD retrieval errors associated with the underestimation (subscript “und”) or overestimation (subscript “ovr”) of Z, SSA, and AAE by the amounts indicated in the left column. For a nominal COD value of 5, maximum AOD retrieval errors take place when Z is underestimated (40%), SSA is overestimated (48%), and AAE is underestimated (23%), yielding a combined largest AOD error estimate of 67%. Minimum AOD errors, on the other hand, result when Z and AAE are overprescribed and SSA is underestimated, producing a combined AOD underestimation error of about −34%. The overall AOD error goes down with increasing cloud optical depth. For a more typical COD value of 10 the overall AOD error range decreases to −27% to 54%.

Table 2.

Percentage error in retrieved AOD and COD associated with the uncertainty of the prescribed values of Z, SSA, and AAE.

Table 2.

Expected errors in retrieved COD are smaller than those associated with the AOD retrieval. For a nominal COD value of 5 the overall error range associated with the above described uncertainties is −4% to 7% and gets larger as the cloud becomes optically ticker. An error range between −6% and 17% is expected for COD 10.

5. Case studies

a. Cloud-only case: 13 October 2006

We will first examine the near-UV capability of retrieving COD from single-channel measurements. This test is necessary to assess the results of the combined AOD–COD retrieval algorithm discussed in this section. In the absence of intervening aerosols cloud optical depth can be derived from solar reflectance measurements at wavelengths where gas absorption interference is negligible. Cloud optical depth is routinely derived from MODIS observations on the Terra and Aqua satellites at 860 nm over the oceans and 645 nm over land (Platnick et al. 2003). In the cloud-only retrieval application the observations are inverted using the forward calculations for AOD zero shown in Fig. 5. This is equivalent to a single channel retrieval using OMI’s 388-nm reflectance measurements. This approach was used to derive the optical depth of the cloud field over the southern Atlantic Ocean on 13 October 2006. To ensure that broken clouds were excluded, the retrieval was only applied to those OMI scenes characterized by overcast conditions or unity MODIS geometric cloud fraction (CF). The absence of absorbing aerosols above the clouds was verified by making sure the AAI was close to zero. A comparative analysis of MODIS geometric CF and OMI 388-nm reflectivity (introduced in section 2) indicated that overcast conditions can be present at OMI reflectivity ρ values as low as 0.20. Thus, a 0.20 minimum ρ threshold value was used. OMI retrieved COD was compared with collocated MODIS COD observations. Given the large difference in spatial resolution between the two sensors (MODIS 0.5 × 0.5 km2 vs OMI 13 × 24 km2), linear averaging of the MODIS COD product over the OMI pixel was necessary. Data points in Fig. 6 are color-coded as a function of the associated OMI-derived reflectivity. The comparison of the OMI research COD (388 nm) and MODIS operational COD (860 nm) products in Fig. 6 indicates a close agreement between the retrievals with correlation coefficient 0.92 and slope 0.96. The OMI COD values appear slightly lower than those from MODIS for R values lower than 0.30, probably associated with the presence of broken and/or transparent clouds. The close OMI–MODIS agreement in retrieved COD illustrates the sensitivity of OMI observations to cloud presence and is consistent with the reported agreement in effective CF calculated by Stammes et al. (2008) from MODIS-retrieved geometric CF (Platnick et al. 2003) and OMI-retrieved effective CF from observations of O2–O2 absorption at 477 nm (Acarreta et al. 2004).

Fig. 6.
Fig. 6.

Comparison of MODIS and OMI cloud optical depth retrieved on 13 Oct 2006. No aerosols were present above clouds during this event. The dotted line indicates the one-to-one agreement, the solid line is the linear fit, and the dashed lines represent the ±20% uncertainty of the MODIS retrieval for CODs larger than 5. Significantly larger uncertainty exists for COD smaller than 5 (Platnick et al. 2003).

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

b. Aerosols above clouds: 4 August 2007

The presence of smoke layers above stratocumulus clouds off the coasts of Namibia and Angola is a common occurrence during the biomass burning season from August through October. Low-level stratus clouds form a few hundred meters above sea level. Since the continental surface is about 1 km above sea level, westward smoke advection results in the accumulation of carbonaceous aerosols above low-level clouds. The retrieval method described in section 4 has been applied to the aerosol event depicted in Fig. 1. Note that during this event a section of the aerosol strip detected by CALIOP is located above a thin, spatially continuous cloud deck at 1 km above sea level (between about 21° and 11°S) whereas cloudy and cloud-free conditions are intermittently present below the aerosol layer between 11° and 5°S. Cloud presence above the aerosol layer is observed north of 5°S. For retrieval purposes, CALIOP observations shown in Fig. 1 have been used to prescribe cloud location (top 1.5 km, bottom 1 km) and aerosol layer height (3 km). A 388-nm aerosol SSA value of 0.88 has been assumed based on OMAERUV algorithm retrieval (Torres et al. 2007) over a cloud-free region at 12°S, 10°E. The spectral dependence of near-UV aerosol absorption has been characterized based on the work by Kirchstetter et al. (2004), who found that an AAE value of about −2.2 adequately described the spectral dependence of aerosol absorption optical depth of carbonaceous particles resulting from biomass burning during the SAFARI 2000 campaign (Swap et al. 2003). An AAE value of −2.2 is equivalent to a 14% increase in imaginary refractive index from 388 to 354 nm as assumed in this analysis. Retrievals were carried out for OMI scenes with R larger than 0.20 and AAI values larger than unity. Note that in the presence of absorbing aerosols above clouds, the actual cloud reflectivity is likely larger than the OMI observed one.

Figure 7 depicts the retrieved AOD (left panel) and COD (right panel) fields. The number of possible retrievals is limited by the upper COD limit of 50 currently used in the research retrieval algorithm. Therefore for CODs larger than 50, neither COD nor AOD values are reported. A clear south–north gradient in the obtained AOD is observed across the aerosol layer with minimum values in the 0.3–0.6 range along the southern edge of the layer and maxima of 1.8 or larger along the northern boundary of the aerosol layer. The COD field shows three localized areas of COD larger than about 12 whereas COD background values of about 5–10 predominate everywhere else.

Fig. 7.
Fig. 7.

Spatial distribution of above-cloud (left) AOD (388 nm) and (right) COD for the 4 Aug 2007 aerosol event.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

A qualitative assessment of the derived aerosol optical depth can be done by examining certain characteristics of the retrieved AOD and COD fields to obtain an idea of the reliability of the results or the lack thereof. A quick visual inspection of the spatial variability of the two retrieved quantities indicates that the derived AOD and COD fields are not generally correlated. For instance, low AOD and high COD values are observed at the region in the vicinity of 14°S, 2°E, whereas the opposite scenario (i.e., high AOD and low COD values) occurs in the area surrounding the clear area at 12°S, 7°E.

Since Aqua MODIS and Aura OMI observations are only a few minutes apart, one can evaluate the OMI COD retrieval by directly comparing it to MODIS results. As was previously demonstrated in this work, when no aerosols are found above clouds OMI-retrieved COD is in excellent agreement with the MODIS operational COD product (Fig. 6). In spite of the difference in spatial resolution between the two sensors the different cloud features in the MODIS product are easily recognizable in the OMI COD image (not shown). A scatterplot of retrieved COD by the two sensors is shown in Fig. 8. Because of the aerosol absorption effect on the MODIS 860-nm reflectance (Wilcox et al. 2009), OMI-retrieved COD values are expected to be higher than those of MODIS. Overall a good level of agreement between the two retrievals is observed as indicated by the high correlation coefficient (0.85), a slope of near 1 (1.07), and a low y intercept (−0.19). A careful analysis in terms of cloud reflectivity shows that for ρ values less than 0.25 (blue filled circles) OMI consistently retrieves CODs lower than MODIS values. This result probably indicates that OMI clouds of reflectivity less than 0.25 are either nonopaque or nonuniform (i.e., broken clouds), or both, which would allow radiation from below the cloud or from cloud-free parts of the pixel to reach the sensor, producing an underestimate of the cloud optical depth. For clouds with reflectivity between 0.25 and 0.30 (green triangles) OMI and MODIS COD tend to be in agreement for COD values up to about 7. At larger values the OMI COD is again underestimated in relation to MODIS, probably due to not meeting the required overcast conditions. Thus, it seems that for ρ values less than 0.30, the overcast and opaque cloud condition is not met and therefore the OMI COD error is comparable to or larger than the aerosol absorption–induced error in the MODIS product. For ρ values larger than 0.30 (red crosses), on the other hand, OMI consistently retrieves higher than MODIS values for COD values up to 15. The lower MODIS values are most probably associated with the aerosol absorption effects at 860 nm that would result in the underestimation of the optical depth as discussed by Wilcox et al. (2009). Figure 9 shows a scatterplot between AOD and AAI, indicating the expected high correlation between these two quantities.

Fig. 8.
Fig. 8.

As in Fig. 6, but for the aerosol event over the southeastern Atlantic Ocean on 4 Aug 2007.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

Fig. 9.
Fig. 9.

Relation between coincident aerosol optical depth and UV aerosol index measured by OMI on 4 Aug 2007 over the study region.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

Another way of qualitatively assessing the retrieved AOD fields is to analyze the spatial continuity between cloudy and cloud-free areas. An inherent assumption made here is that the AOD field does not change significantly from overcast scene to the adjacent clear region. Here, we employ standard OMAERUV algorithm to derive AOD fields over cloud-free region. Figure 10 shows a composite of AOD retrievals over cloudy and cloud-free scenes. No obvious spatial discontinuities in AOD along the northern edge of the cloudy region and adjacent clear region are observed. Missing data, however, are observed along the northernmost edge of the cloud and especially over the southernmost boundary in the transition from overcast to totally clear conditions. The occurrence of these pockets of missing data is expected as none of the OMI algorithms deals with partial cloudiness conditions.

Fig. 10.
Fig. 10.

Composite of above-cloud and clear region AOD derived using cloud–aerosol and aerosol-only OMAERUV algorithms on 4 Aug 2007.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

c. Aerosols above clouds: 31 August 2005

A second sample retrieval was carried out for an aerosol event on 31 August 2005 over the same region. Similar assumptions concerning aerosol–cloud vertical distribution to those in the 4 August 2007 case were used for this event. A 388-nm SSA value of 0.90 retrieved over a clear area at 16°S, 14°E was assumed together with an AAE exponent of −2.2 as in the previously discussed example.

The MODIS true color image composite and the OMI AAI data on 31 August 2005 (not shown) illustrate the presence of a smoke layer located above a large horizontally extended cloud deck over the southern Atlantic Ocean. Retrieved AOD and COD fields for this event are depicted in Fig. 11. The shown AOD field includes retrievals under both overcast and cloud-free conditions. Peak AOD values larger than 3.0 are observed over the southeastern region of the cloud off the coast of Namibia where cloud optical depth values in the vicinity of 5 were retrieved. Lower AOD values of about 1.0 are observed over the optically densest part of the cloud where the COD reaches values close to 20. As in the previous case no correlation between the retrieved AOD and COD fields is observed. The OMI-retrieved COD again correlates very well with the MODIS COD product as shown in Fig. 12. As in the previous case, the OMI-derived COD is generally higher than the MODIS retrieval for reflectivity larger than 0.30 and somewhat lower than the MODIS results for lower reflectivity values. In this case, larger OMI–MODIS differences are observed than on the 4 August 2007 case, which is consistent with the retrieved larger aerosol load.

Fig. 11.
Fig. 11.

Retrieved (left) AOD and (right) COD fields on 31 Aug 2005.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

Fig. 12.
Fig. 12.

As in Fig. 6, but for the aerosol event over the southeastern Atlantic Ocean on 31 Aug 2005.

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

d. Assessment of AOD results

The validation of satellite retrievals of aerosol optical depth is only possible by means of sun photometer measurements as it is routinely done to validate spaceborne quantification of the atmospheric aerosol load for cloud-free conditions. For a direct quantitative assessment of above-cloud aerosol optical depth retrievals, as reported in this work, airborne sun photometer measurements (Russell et al. 2005) would be required. In the absence of those observations, however, a semiquantitative evaluation can be attempted by comparing the OMI-obtained AOD results to those obtained by other satellite-based techniques capable of detecting, and possibly quantifying, aerosol amounts above clouds. Such retrievals have been attempted by application of active (Kittaka et al. 2011; Chand et al. 2008) CALIOP observations as well as passive POLDER measurement of polarized radiation (Waquet et al. 2009) described in section 2.

A comparison of the CALIOP version 3.01 standard AOD (532 nm) to OMI AOD at 388 nm for the 4 August 2007 case discussed in section 5b was carried out. Although a correlation between the two retrievals is apparent, the standard CALIOP AOD is significantly smaller than the OMI results. The wavelength difference does not explain the large observed differences. Given the current provisional nature of the CALIOP AOD product as indicated by the official CALIPSO Quality Statement, we defer further evaluation of OMI above-cloud aerosol retrievals using CALIOP data as a future work activity when a more definitive CALIOP AOD product may be available. Results of above-cloud aerosol optical depth by the method described here were compared to an equivalent POLDER AOD product as described by Waquet et al. (2009). Figure 13 shows a scatterplot of POLDER AOD at 865 nm compared to OMI’s 388 nm for the event of 4 August 2008 documented in Waquet et al. (2009). A well-defined correlation between the two measurements is observed. Although OMI-retrieved AODs are about 3 times as large as POLDER’s, the difference can be explained by the small aerosol particle size typical of carbonaceous aerosols and the difference in reporting wavelength. A reduction of results to a common wavelength (not shown) indicates a close one-to-one agreement.

Fig. 13.
Fig. 13.

Scatterplot of retrieved 865-nm AOD from POLDER observation (x axis) and 388-nm AOD from OMI measurements on 4 Aug 2008. The difference in magnitude is related to the small aerosol particle size and difference in reporting wavelengths (see text for details).

Citation: Journal of the Atmospheric Sciences 69, 3; 10.1175/JAS-D-11-0130.1

6. Summary and conclusions

We have discussed the basis, and presented examples, of an inversion procedure to derive the optical depth of the atmospheric aerosol load above clouds. The retrieval technique uses measurements of backscattered near-UV measurements by the OMI sensor on the Aura satellite. The OMI-derived absorbing aerosol index and the observed 388-nm reflectance are fed to an inversion procedure that simultaneously retrieves both aerosol and cloud optical depth. Results of a sensitivity analysis indicate that the magnitude of the aerosol index associated with absorbing aerosols above clouds depends on cloud and aerosol optical depth, wavelength-dependent aerosol single scattering albedo (or AAE), and, to a lesser extent, on aerosol layer–cloud separation. Thus, to simultaneously retrieve AOD and COD, assumptions as to the values of the other parameters or external information about them is needed. An error analysis was carried out to estimate the uncertainty in the retrieved values of aerosol and cloud optical depth associated with the uncertainty in the assumed values of single scattering albedo and aerosol–cloud separation. Results indicate that the combined uncertainty of ±0.03 in single scattering albedo and ±2 km in aerosol–cloud separation yields an AOD error between −26% and 54% for typical cloud and aerosol layer optical depths of 10 and 0.5, respectively. Retrieval errors decrease with increasing cloud optical depth. Errors in retrieved COD are smaller than 20% in most cases, which is comparable to the reported uncertainty of the MODIS product for optical depths larger than about 10 (Platnick et al. 2003).

The retrieval approach was applied to two aerosol events over the southern Atlantic Ocean off the west coast of southern Africa on 4 August 2007 and 31 August 2005 where a persistent widespread layer of carbonaceous aerosols transported from the adjacent land region accumulates above a uniform low-level stratocumulus cold deck during August and September. In the absence of airborne sun photometer measurements for a direct assessment of the validity of the retrieved AOD results, a qualitative analysis was carried out. For both events, the OMI-retrieved AOD and COD fields are uncorrelated with each other, which indicates the absence of retrieval biases associated with the bright reflecting background. An examination of the spatial homogeneity of the AOD retrieval over overcast and cloud-free scenes on 4 August 2007 shows no obvious discontinuities in the clear-to-cloudy transition regions except for the expected lack of retrievals over partly cloudy scenes.

Direct comparisons of MODIS- and OMI-retrieved cloud optical depths during both events show a very good correlation. When the satellite-derived scene reflectivity is lower than about 0.25, OMI-retrieved CODs are smaller than those of MODIS, indicating the presence in the OMI scene of broken and/or transparent clouds. For scene reflectivities larger than 0.30, overcast opaque cloud conditions seem to exist and OMI COD retrievals are higher than those of MODIS. The larger OMI COD values are consistent with the expected error in MODIS COD associated with the effect of aerosol absorption on the near-IR radiances (860 nm) used in the retrieval. The good correlation between the two COD measurements and the explainable difference in OMI–MODIS retrieval results lends confidence to the results of the OMI two-parameter retrieval given that the OMI and MODIS COD retrieval approaches are fundamentally different from each other.

A quantitative evaluation of the derived AOD was carried out by comparing OMI AOD results to equivalent CALIOP and POLDER retrievals. The comparison to the lidar-based product was not conclusive given the current provisional nature of the CALIOP AOD product. The comparison to POLDER results, on the other hand, yielded a very high correlation. The observed difference in magnitude of the two measurements can be explained in terms of the predominantly small aerosol particles and the difference in reporting wavelengths.

Acknowledgments

The authors thank F. Waquet from the Universitè des Sciences et Technologies de Lille for making available retrieval results from POLDER measurements for comparison to our results. We are also thankful to the valuable feedback of three anonymous reviewers.

REFERENCES

  • Acarreta, J. R., , J. F. De Haan, , and P. Stammes, 2004: Cloud pressure retrieval using the O2–O2 absorption band at 477 nm. J. Geophys. Res., 109, D05204, doi:10.1029/2003JD003915.

    • Search Google Scholar
    • Export Citation
  • Ahmad, Z., , P. K. Bhartia, , and N. Krotkov, 2004: Spectral properties of backscattered UV radiation in cloudy atmospheres. J. Geophys. Res., 109, D01201, doi:10.1029/2003JD003395.

    • Search Google Scholar
    • Export Citation
  • Ahn, C., , O. Torres, , and P. K. Bhartia, 2008: Comparison of Ozone Monitoring Instrument UV aerosol products with Aqua/Moderate Resolution Imaging Spectroradiometer and Multiangle Imaging Spectroradiometer observations in 2006. J. Geophys. Res., 113, D16S27, doi:10.1029/2007JD008832.

    • Search Google Scholar
    • Export Citation
  • Bond, T. C., 2001: Spectral dependence of visible light absorption by carbonaceous particles emitted from coal combustion. Geophys. Res. Lett., 28, 40754078.

    • Search Google Scholar
    • Export Citation
  • Chand, D., , T. L. Anderson, , R. Wood, , R. J. Charlson, , Y. Hu, , Z. Liu, , and M. Vaughan, 2008: Quantifying above-cloud aerosol using spaceborne lidar for improved understanding of cloudy-sky direct climate forcing. J. Geophys. Res., 113, D13206, doi:10.1029/2007JD009433.

    • Search Google Scholar
    • Export Citation
  • Chand, D., , R. Wood, , T. L. Anderson, , S. K. Satheesh, , and R. J. Charlson, 2009: Satellite-derived direct radiative effect of aerosols dependent on cloud cover. Nat. Geosci., 2, 181184, doi:10.1038/ngeo437.

    • Search Google Scholar
    • Export Citation
  • Colarco, P., , M. Schoeberl, , B. Doddridge, , L. Marufu, , O. Torres, , and E. Welton, 2004: Transport of smoke from Canadian forest fires to the surface near Washington, D.C.: Injection height, entrainment, and optical properties. J. Geophys. Res., 109, D06203, doi:10.1029/2003JD004248.

    • Search Google Scholar
    • Export Citation
  • Dave, J. V., , and C. L. Mateer, 1967: A preliminary study on the possibility of estimating total atmospheric ozone from satellite measurements. J. Atmos. Sci., 24, 414427.

    • Search Google Scholar
    • Export Citation
  • de Graaf, M., , P. Stammes, , O. Torres, , and R. B. A. Koelemeijer, 2005: Absorbing Aerosol Index: Sensitivity analysis, application to GOME and comparison with TOMS. J. Geophys. Res., 110, D01201, doi:10.1029/2004JD005178.

    • Search Google Scholar
    • Export Citation
  • de Graaf, M., , P. Stammes, , and E. A. A. Aben, 2007: Analysis of reflectance spectra of UV-absorbing aerosol scenes measured by SCIAMACHY. J. Geophys. Res., 112, D02206, doi:10.1029/2006JD007249.

    • Search Google Scholar
    • Export Citation
  • Deirmendjian, D., 1964: Scattering and polarization properties of water clouds and hazes in the visible and infrared. Appl. Opt., 3, 187196.

    • Search Google Scholar
    • Export Citation
  • Dirksen, R. J., , K. Folkert Boersma, , J. de Laat, , P. Stammes, , G. R. van der Werf, , M. Val Martin, , and H. M. Kelder, 2009: An aerosol boomerang: Rapid around-the-world transport of smoke from the December 2006 Australian forest fires observed from space. J. Geophys. Res., 114, D21202, doi:10.1029/2009JD012360.

    • Search Google Scholar
    • Export Citation
  • Dubovik, O., , and M. D. King, 2000: A flexible inversion algorithm for retrieval of aerosol optical properties from sun and sky radiance measurements. J. Geophys. Res., 105, 20 67320 696.

    • Search Google Scholar
    • Export Citation
  • Dubovik, O., , B. Holben, , T. F. Eck, , A. Smirnov, , Y. J. Kaufman, , M. D. King, , D. Tanré, , and I. Slutsker, 2002: Variability of absorption and optical properties of key aerosol types observed in worldwide locations. J. Atmos. Sci., 59, 590608.

    • Search Google Scholar
    • Export Citation
  • Duncan, B. N., , R. V. Martin, , A. C. Staudt, , R. Yevich, , and J. A. Logan, 2003: Interannual and seasonal variability of biomass burning emissions constrained by satellite observations. J. Geophys. Res., 108, 4100, doi:10.1029/2002JD002378.

    • Search Google Scholar
    • Export Citation
  • Eck, T. F., and Coauthors, 2003: Variability of biomass burning aerosol optical characteristics in southern Africa during the SAFARI 2000 dry season campaign and a comparison of single scattering albedo estimates from radiometric measurements. J. Geophys. Res., 108, 8477, doi:10.1029/2002JD002321.

    • Search Google Scholar
    • Export Citation
  • Fromm, M., , O. Torres, , D. Diner, , D. Lindsey, , B. Vant Hull, , R. Servranckx, , E. P. Shettle, , and Z. Li, 2008: Stratospheric impact of the Chisholm pyrocumulonimbus eruption: 1. Earth-viewing satellite perspective. J. Geophys. Res., 113, D08202, doi:10.1029/2007JD009153.

    • Search Google Scholar
    • Export Citation
  • Hansen, J. E., 1969: Exact and approximate solutions for multiple scattering by cloudy and hazy planetary atmospheres. J. Atmos. Sci., 26, 478487.

    • Search Google Scholar
    • Export Citation
  • Hao, W. M., , and M.-H. Liu, 1994: Spatial and temporal distribution of tropical biomass burning. Global Biogeochem. Cycles, 8, 495503.

    • Search Google Scholar
    • Export Citation
  • Haywood, J. M., , and K. P. Shine, 1997: Multi-spectral calculations of the direct radiative forcing of tropospheric sulphate and soot aerosols using a column model. Quart. J. Roy. Meteor. Soc., 123, 19071930.

    • Search Google Scholar
    • Export Citation
  • Haywood, J. M., , S. Osborne, , P. N. Francis, , A. Kiel, , P. Formenti, , T. W. Andreae, , and P. Kaye, 2003: The mean physical and optical properties of regional haze dominated by biomass burning aerosol measured from the C-130 aircraft during SAFARI 2000. J. Geophys. Res., 108, 8473, doi:10.1029/2002JD002226.

    • Search Google Scholar
    • Export Citation
  • Herman, J. R., , P. K. Bhartia, , O. Torres, , C. Hsu, , C. Seftor, , and E. Celarier, 1997: Global distribution of UV-absorbing aerosols from Nimbus 7/TOMS data. J. Geophys. Res., 102, 16 91116 922.

    • Search Google Scholar
    • Export Citation
  • Hoffer, A., , A. Gelencsér, , P. Guyon, , G. Kiss, , O. Schmid, , G. P. Frank, , P. Artaxo, , and M. O. Andreae, 2006: Optical properties of humic-like substances (HULIS) in biomass-burning aerosols. Atmos. Chem. Phys., 6, 35633570.

    • Search Google Scholar
    • Export Citation
  • Hsu, N. C., and Coauthors, 1996: Detection of biomass burning smoke from TOMS measurements. Geophys. Res. Lett., 23, 745748.

  • Hsu, N. C., , J. R. Herman, , and S.-C. Tsay, 2003: Radiative impacts from biomass burning in the presence of clouds during boreal spring in Southeast Asia. Geophys. Res. Lett., 30, 1224, doi:10.1029/2002GL016485.

    • Search Google Scholar
    • Export Citation
  • Hu, Y., , M. Vaughan, , Z. Liu, , K. Powell, , and S. Rodier, 2007: Retrieving optical depths and lidar ratios for transparent layers above opaque water clouds from CALIPSO lidar measurements. IEEE Geosci. Remote Sens. Lett., 4, 523526.

    • Search Google Scholar
    • Export Citation
  • Jaffe, D., , I. Bertschi, , L. Jaeglé, , P. Novelli, , J. S. Reid, , H. Tanimoto, , R. Vingarzan, , and D. L. Westphal, 2004: Long-range transport of Siberian biomass burning emissions and impact on surface ozone in western North America. Geophys. Res. Lett., 31, L16106, doi:10.1029/2004GL020093.

    • Search Google Scholar
    • Export Citation
  • Jethva, H., , and O. Torres, 2011: Satellite-based evidence of wavelength-dependent aerosol absorption in biomass burning smoke inferred from ozone monitoring instrument. Atmos. Chem. Phys. Discuss., 11, 72917319, doi:10.5194/acpd-11-7291-2011.

    • Search Google Scholar
    • Export Citation
  • Johnson, M. S., and Coauthors, 2010: Modeling dust and soluble iron deposition to the South Atlantic Ocean. J. Geophys. Res., 115, D15202, doi:10.1029/2009JD013311.

    • Search Google Scholar
    • Export Citation
  • Kallos, G., , A. Papadopoulos, , P. Katsafados, , and S. Nickovic, 2006: Transatlantic Saharan dust transport: Model simulation and results. J. Geophys. Res., 111, D09204, doi:10.1029/2005JD006207.

    • Search Google Scholar
    • Export Citation
  • Kaufman, Y. J., , I. Koren, , L. A. Remer, , D. Tanré, , P. Ginoux, , and S. Fan, 2005: Dust transport and deposition observed from the Terra-Moderate Resolution Imaging Spectroradiometer (MODIS) spacecraft over the Atlantic Ocean. J. Geophys. Res., 110, D10S12, doi:10.1029/2003JD004436.

    • Search Google Scholar
    • Export Citation
  • Keil, A., , and J. M. Haywood, 2003: Solar radiative forcing by biomass burning aerosol particles during SAFARI 2000: A case study based on measured aerosol and cloud properties. J. Geophys. Res., 108, 8467, doi:10.1029/2002JD002315.

    • Search Google Scholar
    • Export Citation
  • King, M. D., , S. Platnick, , C. C. Moeller, , H. E. Revercomb, , and D. A. Chu, 2003: Remote sensing of smoke, land, and clouds from the NASA ER-2 during SAFARI 2000. J. Geophys. Res., 108, 8502, doi:10.1029/2002JD003207.

    • Search Google Scholar
    • Export Citation
  • Kirchstetter, T. W., , T. Novakov, , and P. V. Hobbs, 2004: Evidence that the spectral dependence of light absorption by aerosols is affected by organic carbon. J. Geophys. Res., 109, D21208, doi:10.1029/2004JD004999.

    • Search Google Scholar
    • Export Citation
  • Kittaka, C., , D. M. Winker, , M. A. Vaughan, , A. Omar, , and L. A. Remer, 2011: Intercomparison of column aerosol optical depths from CALIPSO and MODIS-Aqua. Atmos. Meas. Tech., 4, 131141.

    • Search Google Scholar
    • Export Citation
  • Liao, H., , and J. H. Seinfeld, 1998a: Effect of clouds on direct aerosol radiative forcing of climate. J. Geophys. Res., 103, 37813788.

    • Search Google Scholar
    • Export Citation
  • Liao, H., , and J. H. Seinfeld, 1998b: Radiative forcing by mineral dust aerosol: Sensitivity to key variables. J. Geophys. Res., 103, 31 63731 646.

    • Search Google Scholar
    • Export Citation
  • Penning de Vries, M. J. M., , S. Beirle, , and T. Wagner, 2009: UV aerosol indices from SCIAMACHY: Introducing the Scattering Index (SCI). Atmos. Chem. Phys., 9, 95559567.

    • Search Google Scholar
    • Export Citation
  • Peters, K., , J. Quaas, , and N. Bellouin, 2009: Effects of absorbing aerosols in cloudy skies: A satellite study over the Atlantic Ocean. Atmos. Chem. Phys., 11, 13931404, doi:10.5194/acp-11-1393-2011.

    • Search Google Scholar
    • Export Citation
  • Platnick, S., , M. D. King, , A. Ackerman, , W. P. Menzel, , B. A. Baum, , J. C. Riedi, , and R. A. Frey, 2003: The MODIS cloud products: Algorithms and examples from Terra. IEEE Trans. Geosci. Remote Sens., 41, 459473.

    • Search Google Scholar
    • Export Citation
  • Rossow, W. B., , and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 22612287.

  • Russell, P. B., and Coauthors, 2005: Aerosol optical depth measurements by airborne sun photometer in SOLVE II: Comparison to SAGE III, POAM III and airborne spectrometer measurements. Atmos. Chem. Phys., 5, 13111339.

    • Search Google Scholar
    • Export Citation
  • Russell, P. B., and Coauthors, 2010: Absorption angstrom exponent in AERONET and related data as an indicator of aerosol composition. Atmos. Chem. Phys., 10, 11551169, doi:10.5194/acp-10-1155-2010.

    • Search Google Scholar
    • Export Citation
  • Spurr, R. J. D., 2006: VLIDORT: A linearized pseudo-spherical vector discrete ordinate radiative transfer code for forward model amd retrieval studies in multilayer multiple scattering media. J. Quant. Spectrosc. Radiat. Transfer, 102, 316342.

    • Search Google Scholar
    • Export Citation
  • Stammes, P., , M. Sneep, , J. F. de Haan, , J. P. Veefkind, , P. Wang, , and P. Levelt, 2008: Effective cloud fractions from the Ozone Monitoring Instrument: Theoretical framework and validation. J. Geophys. Res., 113, D16S38, doi:10.1029/2007JD008820.

    • Search Google Scholar
    • Export Citation
  • Swap, R. J., , H. J. Annegarn, , J. T. Suttles, , M. D. King, , S. Platnick, , J. L. Privette, , and R. J. Scholes, 2003: Africa burning: A thematic analysis of the Southern African Regional Science Initiative (SAFARI 2000). J. Geophys. Res., 108, 8465, doi:10.1029/2003JD003747.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , P. K. Bhartia, , J. R. Herman, , Z. Ahmad, , and J. Gleason, 1998: Derivation of aerosol properties from satellite measurements of backscattered ultraviolet radiation: Theoretical basis. J. Geophys. Res., 103, 17 09917 110.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , P. K. Bhartia, , J. R. Herman, , A. Sinyuk, , and B. Holben, 2002: A long-term record of aerosol optical thickness from TOMS observations and comparison to AERONET measurements. J. Atmos. Sci., 59, 398413.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , A. Tanskanen, , B. Veihelman, , C. Ahn, , R. Braak, , P. K. Bhartia, , P. Veefkind, , and P. Levelt, 2007: Aerosols and surface UV products from OMI observations: An overview. J. Geophys. Res., 112, D24S47, doi:10.1029/2007JD008809.

    • Search Google Scholar
    • Export Citation
  • Torres, O., , Z. Chen, , H. Jethva, , C. Ahn, , R. Freitas, , and P. K. Bhartia, 2010: OMI and MODIS observations of the anomalous 2008–2009 Southern Hemisphere biomass burning seasons. Atmos. Chem. Phys., 10, 35053513.

    • Search Google Scholar
    • Export Citation
  • Waquet, F., , J. Riedi, , L. C. Labonnote, , P. Goloub, , B. Cairns, , J. L. Deuzé, , and D. Tanré, 2009: Aerosol remote sensing over clouds using A-Train observations. J. Atmos. Sci., 66, 24682480.

    • Search Google Scholar
    • Export Citation
  • Wen, G., , A. Marshak, , and R. F. Cahalan, 2008: Importance of molecular Rayleigh scattering in the enhancement of clear sky reflectance in the vicinity of boundary layer cumulus clouds. J. Geophys. Res., 113, D24207, doi:10.1029/2008JD010592.

    • Search Google Scholar
    • Export Citation
  • Wilcox, E. M., 2010: Stratocumulus cloud thickening beneath layers of absorbing smoke aerosol. Atmos. Chem. Phys., 10, 11 76911 777, doi:10.5194/acp-10-11769-2010.

    • Search Google Scholar
    • Export Citation
  • Wilcox, E. M., , Harshvardhan, , and S. Platnick, 2009: Estimate of the impact of absorbing aerosol over cloud on the MODIS retrievals of cloud optical thickness and effective radius using two independent retrievals of liquid water path. J. Geophys. Res., 114, D05210, doi:10.1029/2008JD010589.

    • Search Google Scholar
    • Export Citation
  • Zarzycki, C. M., , and T. M. Bond, 2010: How much can the vertical distribution of black carbon affect its global direct radiative forcing? Geophys. Res. Lett., 37, L20807, doi:10.1029/2010GL044555.

    • Search Google Scholar
    • Export Citation
Save