Robert Wood with Duli Chand , Tad Anderson, Bob Charlson

1. Direct Radiative Effect of aerosols over clouds and clear skies determined using CALIPSO and the A-Train Robert Wood with DuliChand, Tad Anderson, Bob Charlson VOCALS RF04, 21 November 2008

2. Effect of aerosol layer on TOA SW radiation 0.0 0.9 0.99 0.999 DRE > 0 (warming) Coakley and Chylek (1974) Single scattering albedo (approx) DRE < 0 (cooling) 00 0.2 0.4 0.6 0.8 1.0 Surface albedo

3. biomass burning aerosol above cloud stratocumulus clouds MODIS Aqua RGB (enhanced) 13 Aug 2006 SE Atlantic 500 km

4. Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations

5. CALIPSO lidar (CALIOP) • Backscatter profiles at 532 and 1064 nm • Parallel and perpendicular polarization for 532 nm channel

6. Aerosol layers over clouds seenwith CALIPSO over SE Atlantic Ocean (13 Aug 2006)

7. Biomass burning fires (2006, monthly)

8. Direct radiativeforcing (DRF) AEROCOMModels(Schulz et al. 2006)

9. Aerosol optical thickness retrieval methodologies

10. Measurements of Lidar ratio (S) From Anderson et al. (2000), J. Geophys. Res.

11. Retrieval: color ratio method (CR)(Chand et al. 2006, J. Geophys. Res.) • CALIPSO data, integrated attenuated backscatter at 532 and 1064 nm (g532and g1064) • Determine color ratiocwater = g1064/g532 from layers classified as cloud (z < 3 km) • Unobstructed liquid clouds should have c = 1, and so deviations from this represent aerosols above clouds • Use Beer-Lambert law to obtain AOD of aerosol layer: =1 ideally, but use unobstructed cloud to calibrate Angstrom exponent

12. Depolarization ratio method (DR) (Hu et al. 2007, Chand et al. 2007) • Use depolarization d of cloud layer, combined with its integrated attenuated backscatter g, to derive AOD of overlying layer Extinction to backscatter ratio for water clouds (19 sr) Self-calibration coefficient

13. Comparison of DR and CR aerosol optical depths assuming å = 2 for CR method • Daytime resultsaresimilar Increasing Angstrom exponent assuming å = 2

14. Cloud layer top heights

15. Angstrom exponent for layers above cloud

16. Aerosol optical depth for layers above cloud (by month 2006) June July August Sep Oct Nov

17. Biomass burning fires (2006, monthly)

18. AOD and winds at 600 hPa -25 -20 -15 -10 -5 0 5 10 15

19. Determining the direct radiative effect of elevated aerosol layers above the partly cloudy boundary layer (Chand et al. 2009, Nature Geosciences)

20. Radiative transfer model • DISORT radiative transfer model • Aerosol properties needed are AOD (from CALIPSO), single scattering albedo (w=0.85,Leahy et al. 2006), Angstrom exponent (CALIPSO), asymmetry factor (g = 0.62) • Cloud properties are cloud optical depth and cloud effective radius (MODIS), and cloud fraction • Ocean surface albedo = 0.06 • Determine aerosol radiative effect for clear sky, cloudy sky, and all-sky (Jul-Oct 2006/2007)

21. Absorption of solar radiation by aerosols Single scattering albedo Aitken mode particle conc. Scattering coefficient Absorption coefficient  Single scattering albedo a (10-7 m-1) s (10-6 m-1) From Clarke and Charlson, 1985, Science

22. Absorption: Single scattering albedo Data from SAFARI-2000 field campaign Frequency of occurrence 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1.0 single scattering albedo at 550 nm From Leahy et al. (2007), Geophys. Res. Lett., 34, L12814, doi:10.1029/2007GL029697

23. Wavelength dependence of  Aerosol is relatively more absorptive at longer wavelengths Leahy et al. (2007) From Bergstrom et al. (2007), Atmos. Chem. Phys.

24. Effect of aerosol upon radiative fluxes Absorption AOD • DRE (TOA) • Radiative forcing efficiency

25. RFE is determined primarily by cloud cover

26. Dependence on single scattering albedo • Critical cloud fraction increases strongly with SSA (brighter albedo required for positive DRE)

27. Inter-model standard deviation of aerosol direct radiative forcing (AEROCOM, Schulz et al. 2006)

28. Aerosol all-sky direct radiative forcing 5-20oS, 10oE-10oW • DFall=DFclr + C(DFcld-DFclr) • Inter-model cloud cover variations explain 70% of the variance in the aerosol directradiative forcing (DRF) • Model prediction of cloud cover important for accurate quantification of aerosol DRF DFall

29. Summary • Novel method using color ratio of cloud targets used to derive aerosol optical depth of elevated biomass burning layers over clouds over SE Atlantic • Radiative transfer calculations and MODIS cloud optical properties data used to determine direct radiative effect of elevated aerosol layers • Remarkably linear dependence of RFE upon cloud fractional coverage of low clouds (critical cloud fraction) • Inter-model differences in DRF are not only related to aerosol radiative properties, but are strongly dependent upon model cloud cover

30. Questions • What is the climate response to aerosol forcing by biomass burning aerosols over the South Atlantic? • Simulations using CAM (with Naoko Sakaeda, Phil Rasch) • What are the global mean effects of aerosols above clouds? • Global analysis CALIPSO/DISORT. Use this to constrain models (DuliChand) • Passive remote sensing of aerosols above clouds using MODIS?

31. Community Atmospheric Model (CAM) Simulations (Naoko Sakaeda, Phil Rasch) • Preliminary analysis of 20-year CAM simulations • Present-day AOD tuned to match CALIPSO measurements • Figure shows change in low cloud cover (biomass burning aerosols – no biomass burning aerosols) 0.20 0.15 0.10 0.05 0 -0.05

33. AEROCOM Models (Schulz et al. 2006) Direct radiative forcing for cloudy skies