1. Introduction Invisible clouds in this study mean super-thin clouds which cannot be detected by MODIS but are classified as clouds by CALIPSO. These sub-visual clouds may exist globally and may have effects on Earth-atmosphere radiation budget and remote sensing of aerosols. In this study, 12-month (Jan 1 – Dec 31, 2007) CERES, MODIS, CALIPSO, and AIRS measurements are analyzed for these clouds. Wenbo Sun Science Systems and Applications, Inc. Mail Stop 420, NASA Langley Research Center, Hampton, VA23693, USA wenbo.sun-1@nasa.gov Study Invisible Clouds for Glory Aerosol Product Glory Science Team Meeting, August 10-12, 2011, New York

2. Total attenuated backscatter at 532nm from CALIPSO lidar

3. 20km CERES FOV 1km MODIS Pixel 333m CALIPSO Lidar Shot. Cloud coverage percentage is calculated using along-CALIPSO-track CALIPSO and MODIS data.. Radiation energy budget effect of invisible clouds is estimated on CERES FOVs of MODIS clear and CALIPSO cloudy. Method and Data CCCM data – CERES, CALIPSO, MODIS, and MOA AIRS data – L3 daily 1°x1° gridded standard retrieval product V5 CCCM data

4. MODIS-derived 12-month clear percentage of CERES FOVs CALIPSO-derived cloudy percentage in MODIS-clear CERES FOVs

5. Daytime Purely Clear 12-month CERES FOVs Sampling Distribution Daytime Invisibly Cloudy Nighttime Purely Clear Nighttime Invisibly Cloudy

6. Zonal and altitude distribution of invisible cloud occurrence frequency (in the unit of CERES FOV number) for daytime (left panel) and nighttime (right panel) ocean Zonal and altitude distribution of invisible cloud

7. One-year zonal mean of invisible cloud heights for nighttime oceans The extent of Hadley cell is a metric of climate change. Invisible clouds provide a novel way for satellite remote sensing of Hadley cell. Invisible clouds correlate with general circulations Regular ice clouds do not well correlate with general circulation cells Sun and Lin (2011) ACPD

8. Instantaneous CERES SW flux is converted to diurnal 24-hour mean value by using previously made lookup tables from CERES TRMM processing-orbit data (Loeb & Manalo-Smith 2005). Invisibly thin clouds have ~2.5 Wm -2 diurnal mean SW cooling effect. Daytime Invisible Clouds’ Radiation Effect

9. Comparison of CERES outgoing LW flux for clear (filled circle) and invisibly cloudy (open circle) cases Nighttime Clear Sky and Invisible Cloud Radiation

10. The CERES LW flux difference between clear and invisibly cloudy FOVs could be a result of water vapor absorption. This makes the quantification of the invisible clouds’ effect on LW radiation difficult. Humidity and temperature difference between clear and invisibly cloudy environment Daytime zonal mean instantaneous column water vapor amount from AMSR-E (filled circle) and AIRS (open circle) for clear (black and red) and invisibly cloudy (blue and green) ocean Daytime zonal mean instantaneous temperature profiles from AIRS for clear (thin curve) and invisibly cloudy (thick curve) ocean

11. Comparison of modeled outgoing LW flux for clear (filled circle) and invisibly cloudy (open circle) cases using atmospheric profiles of clear CERES FOVs. Modeled Invisible Clouds’ Radiation Effect

12. Comparison of CERES and modeled LW flux for clear CERES FOVs Modeling LW flux for daytime and nighttime ocean using atmospheric profiles from MOA and AIRS dataset Sun et al. (2011) JGR daytime nighttime Invisible cloud effect

13. Zonal mean MOD04 aerosol optical depth at 0.55 µm for daytime ocean Statistics of 1km x 1km areas with matched and unmatched cloud masks from CALIPSO and MOD04 Effect of invisible clouds on MODIS aerosol product ~25% MOD04 aerosol product is contaminated by invisible clouds

14. Updated Work Outline 1.Use CALIPSO and MODIS data to study the distribution of invisible clouds; 2.Develop a polarized radiative transfer model and use RSP data to validate the model; 3.Use RSP data plus model to study the BPDFs of different surfaces; 4.Use model to study the sensitivity of the polarized reflectance at 1.37 µm to invisible clouds, aerosol, low clouds, and surface optical properties; 5.Develop the algorithm to use Glory 1.37-µm polarized radiance to retrieve the physical properties of invisible clouds; 6.Develop the algorithm to remove the invisible clouds’ effect from Glory radiances, so that the corrected radiances are suitable for retrieval of aerosol and cloud phase.

15. 1. Adding-Doubling radiative transfer model: This can calculate full Stokes vector (I, Q, U, V). 2. Atmospheric profiles: Standard Atmosphere now. 3. Spectral gas absorption: Line-by-Line and k-distribution plus ozone cross-section table. 4. Molecular scattering: Rayleigh. 5. Particulate absorption and scattering: Mie for water clouds (Gamma size distribution); FDTD for aerosols (lognormal size distribution with fine and coarse mode); FDTD plus GOM for ice clouds (lognormal or measured size distributions). 6. Surface reflection model: Lambert surface for land now. Cox & Munk + foam for wind-roughened ocean. The polarized radiative transfer model

16. Sea-Salt AOT = 0.075 Sensitivity of clear ocean total reflectance and DOP to wind speed

17. Loeb et al (2005) CERES SW anisotropic factors in the principal plane Water Clouds Ice Clouds Model results have excellent agreement with CERES data in total radiance angular anisotropy, except the ice cloud specular reflection, since we assume pure randomly oriented ice crystals in the model.

18. Conclusion 1.CALIPSO shows that ~50% of the clear sky are actually covered by invisible clouds. 2.These clouds have big impact on global radiation energy budget, but are not observed and studied sufficiently. 3.These clouds may introduce significant uncertainties into Glory aerosol data if not identified and properly removed from Glory product. 4.For these super-thin clouds, Glory APS is the most precise instrument to measure their physical properties. 5.We will use Glory and CALIPSO data to study the physical properties of invisible clouds. These properties will then be used to reduce the uncertainty in the Glory aerosol and cloud phase products.