1. A Sharper View of Fuzzy Objects: Warm Clouds and their Role in the Climate System as seen by Satellite Ralf Bennartz University of Wisconsin – Madison

2. Outline What do we want to know? Cloud microphysical parameters Cloud fraction and albedo Conclusions and outlook

3. Clouds are the single-most important factor modulating climate sensitivity….. … and we know very little about forcings and feedbacks IPCC 2007

4. 21 climate models- one scenario

5. How can observations help narrow down these uncertainties?

6. What do we need to know? Cloud droplet number concentration (N) Liquid water path (LWP) Drizzle/rain? Rain water path (RWP) Cloud fraction Albedo

7. (Sub) adiabatic stratiform clouds LWC

8. LWP from VIS/NIR: Factor f depends on stratification of cloud (e.g. f=5/9 for adiabatic, f=2/3 for vertically uniform) N for an adiabatic cloud related to optical depth and LWP. C depends weakly on temperature and width of droplet spectrum Basic physical relations

9. LWP from VIS/NIR: Factor f depends on stratification of cloud (e.g. f=5/9 for adiabatic, f=2/3 for vertically uniform) N for an adiabatic cloud related to optical depth and LWP. C depends weakly on temperature and width of droplet spectrum MW: Very direct measurement. Mass absorption coefficient depends slightly on temperature, σ r depends on rain water content too. Basic physical relations

10. Bennartz et al. (2010)

11. Basic physical relations Bennartz et al. (2010)

12. Three input variables (τ VIS r eff τ MW )  Three output variables (N,LWP,RWP) Basic physical relations

13. Three input variables (τ VIS r eff τ MW )  Three output variables (N,LWP,RWP) Auxiliary data, assumptions: Cloud top height Cloud top temperature Drizzle/rain particle size distribution Width of cloud droplet spectrum Basic physical relations

14. Direct physical relationship between MW/VNIR optical properties and cloud physical properties. Errors and uncertainties due to input and auxiliary parameters can be specified and dependencies can be explicitly spelled out. Validity of assumptions can be assessed from observations. No unknown unknowns (though a lot of known unknowns). Basic physical relations

15. Bennartz et al. (2010), Bennartz (2007), Rausch et al. (2010) Products

16. Cloud droplet number concentration

18. Bennartz et al. (2011)

19. Cloud droplet number concentration Bennartz et al. (2011)

20. Application example: Constraining cloud-aerosol interaction processes in GCMs Storelvmo, Lohmann, Bennartz (2009) Differences in cloud droplet activation account for about 65 % (1.3 W/m 2 ) of the total spread in shortwave forcing (2 W/m 2 ) in IPCC AR4 models

21. Deviations from adiabaticity  valid ‘targets’ can be identified by checking multi-spectral agreement of r eff with adiabatic assumption. Open issues & proposed solutions

22. Effective Radius

23. Deviations from adiabaticity  valid ‘targets’ can be identified by checking multi-spectral agreement of r eff with adiabatic assumption. Open issues & proposed solutions

24. Deviations from adiabaticity  valid ‘targets’ can be identified by checking multi-spectral agreement of r eff with adiabatic assumption. Physical explanation for differences in r eff still outstanding: partly cloudy pixels, subscale variability, rain, 3D…. Probably all of them to some extend Open issues & proposed solutions

25. Validating models: LWP observations against iRAM Lauer et al. (2009)

26. Special Sensor Microwave/Imager

27. Liquid Water Path [kg/m 2 ] Local Time [hours]

28. Liquid Water Path [kg/m 2 ] Local Time [hours]

29. Liquid Water Path [kg/m 2 ] Local Time [hours]

30. Liquid Water Path [kg/m 2 ] Local Time [hours]

31. Liquid Water Path [kg/m 2 ] Local Time [hours]

32. Conclusions Satellite observations provide a wealth of information about cloud microphysical and radiative properties. Simple physically-based cloud models provide an excellent framework to interpret satellite observations and brigde gap between observations and GCMs Caveats, uncertainties, and the limits of our current understanding need to be conveyed. Working closely with GCM/RCM community.

33. Acknowledgements Collaborations: A. Lauer (U Hawaii), T. Storelvmo (Yale), U. Lohmann (ETH), J. Rausch (UW-AOS), L. Borg, (UW-AOS), A. Heidinger, (NOAA), M. Foster (CIMSS), R. Leung (PNNL), J. Fan (PNNL), C. O’Dell (CSU), R. Wood (UW) Funding: NASA MODIS Science Team, NOAA/Joint Center for Satellite Data Assimilation, DOE Regional Climate Modeling