Effects of spatial and temporal variations in aerosol properties on mean cloud albedo

Abstract

The mean cloud albedo over a spatial or temporal domain depends not only on the mean cloud condensation nucleus (CCN) spectrum but also on the CCN spectrum variation. When the variation of CCN spectrum is neglected, cloud albedo calculated using the mean CCN spectrum is positively biased. The CCN spectrum variation due to either variation in size distribution or chemical composition, as well as its corresponding effect on mean cloud albedo, was investigated using data collected at Pt. Reyes, California, during the July 2005 Marine Stratus Experiment (MASE). Even when the variation in chemical composition during the entire 28‐day project is neglected, the error in mean cloud albedo is small, and the corresponding error in mean upwelling irradiance is less than 0.5 W/m2. This small and nonsystematic error over such an extended period suggests that for study of mean cloud albedo and upwelling irradiance, the CCN spectrum can be parameterized using the average particle chemical composition or particle activation diameter based on location or air mass type. In contrast, neglecting the variation in aerosol size distribution or CCN spectrum results in positive bias in mean cloud albedo. The bias increases superlinearly with the relative standard deviation of CCN concentration over the domain of interest. On the basis of the MASE data, the average bias in mean upwelling irradiance within grid cells of typical global models ranges 0.4–0.5 W/m2 when only the variation of aerosol size distribution is neglected and increases to 0.5–0.7 W/m2 when the variation of CCN spectrum is neglected. The bias in mean upwelling irradiance can potentially, albeit infrequently, reach 12 W/m2 within grid cells of typical global models if the mean cloud albedo is derived using the mean size distribution or mean CCN spectrum. This suggests accurate evaluation of mean cloud albedo requires the variation of CCN spectrum or aerosol size distribution be taken into consideration, at least for high variability of CCN concentration.