The Broadband Cloud Albedo Susceptibility of Liquid Clouds as Inferred from MODIS Observations

Satellite observations have been used to study the modification of cloud microphysical and optical properties by anthropogenic aerosols (i.e., the "first indirect" or "Twomey" effect). However, the challenges in determining actual cloud changes are many, and originate mainly from the difficulties in trying to quantify a partial derivative (i.e., a temporal change in cloud properties due only to changes in aerosol amount and type, while all relevant dynamic/thermodynamic quantities remain fixed). Rather than confronting the formidable task of assessing the partial derivative of cloud properties in a particular place and time, we have adopted an alternative approach where satellite retrievals are used to estimate the radiative response or sensitivity to some specified change in cloud droplet number concentration (CDNC). Global spatial and temporal distributions of the albedo sensitivity can be used to estimate the range of future radiative forcings, as well as provide an additional constraint in indirect effect modeling studies. Our method is thus an attempt to quantify through observations, the radiative impact of well-posed hypothetical scenarios of CDNC changes. In this regard, we revisit the previously introduced concept of "cloud albedo susceptibility", a sensitivity parameter given by the cloud albedo change for a differential change in CDNC under constant liquid water content conditions. Our work expands on previous susceptibility studies to include relative CDNC changes, perturbations in cloud droplet asymmetry parameter and single scattering albedo in addition to extinction, and broadband top-of-atmosphere calculations derived from global cloud retrievals, observational broadband surface albedo maps, and atmospheric state profiles. Cloud properties are obtained from recent MODIS Collection 5 gridded (1 degree) liquid cloud optical property joint histograms; atmospheric and surface albedo data used for the susceptibility calculations are the same data sets used in the MODIS retrievals. The dependencies of "relative susceptibility" on the joint optical thickness and effective radius histogram are very different from absolute susceptibility. Relative susceptibility also requires fewer assumptions (e.g., constant liquid water content) and is well suited for determining the effect of hypothetical/modeled CDNC perturbations. Monthly susceptibility climatologies for differential as well as a range of relative CDNC changes will be shown.