An estimate of aerosol indirect effect from satellite measurements with concurrent meteorological analysis

Many studies have used satellite retrievals to investigate the effect of aerosols on cloud properties, but these retrievals are subject to artifacts that can confound interpretation. Additionally, large‐scale meteorological differences over a study region dominate cloud dynamics and must be accounted for when studying aerosol and cloud interactions. We have developed an analysis method which minimizes the effect of retrieval artifacts and large‐scale meteorology on the assessment of the aerosol indirect effect. The method divides an oceanic study region into 1° × 1° grid boxes and separates the grid boxes into two populations according to back trajectory analysis: one population contains aerosols of oceanic origin, and the other population contains aerosols of continental origin. We account for variability in the large‐scale dynamical and thermodynamical conditions by stratifying these two populations according to vertical velocity (at 700 hPa) and estimated inversion strength and analyze differences in the aerosol optical depths, cloud properties, and top of atmosphere (TOA) albedos. We also stratify the differences by cloud liquid water path (LWP) in order to quantify the first aerosol indirect effect. We apply our method to a study region off the west coast of Africa and only consider single‐layer low‐level clouds. We find that grid boxes associated with aerosols of continental origin have higher cloud fraction than those associated with oceanic origin. Additionally, we limit our analysis to those grid boxes with cloud fractions larger than 80% to ensure that the two populations have similar retrieval biases. This is important for eliminating the retrieval biases in our difference analysis. We find a significant reduction in cloud droplet effective radius associated with continental aerosols relative to that associated with oceanic aerosols under all LWP ranges; the overall reduction is about 1.0 mm, when cloud fraction is not constrained, and is about 0.5 mm, when cloud fraction is constrained to be larger than 80%. We also find significant increases in cloud optical depth and TOA albedo associated with continental aerosols relative to those associated with oceanic aerosols under all LWP ranges. The overall increase in cloud optical depth is about 0.6, and the overall increase in TOA albedo is about 0.021, when we do not constrained cloud fraction. The overall increases in cloud optical depth and TOA albedo are 0.4 and 0.008, when we only use grid boxes with cloud fraction larger than 80%.