Correlation between model-calculated anthropogenic aerosols and satellite-derived cloud optical depths: Indication of indirect effect?

[1] We consider two independently derived data sets. The first represents the annually averaged distribution of anthropogenic aerosols over East Asia as derived by a coupled regional climate/chemical transport model. The other is the annually averaged distributions of cloud optical depths and cloud amount over East Asia derived by the International Satellite Cloud Climatology Project (ISCCP) for 1990, 1991, 1992, and 1993. We find a remarkable similarity in the distributions of model-calculated anthropogenic aerosols and ISCCP-reported cloud optical depths, with both exhibiting a region of enhanced values extending over the east central portion of China, between the Sichuan Basin and the Yangtze Delta, and then in an easterly direction over the East China Sea, Japan, and South Korea, and the western edge of the Pacific Ocean. Linear regression between the estimated aerosol column burdens and the cloud optical depths yields an r(2) > 0.6, indicating that the correlations are statistically significant at a confidence level that is >99.9% and that more than 60% of the variation in the cloud optical depths is related to variations in the anthropogenic aerosol abundances. Multivariate analysis involving the distributions of boundary layer relative humidity and precipitation over East Asia, as well as that of the model-calculated anthropogenic aerosols and the ISCCP-reported cloud properties, indicates that the relationship between anthropogenic aerosols and cloud optical depth is unique to these two variables and not symptomatic of a more general mechanism involving the hydrologic cycle. Trend analysis of the ISCCP data suggests that there was an upward trend in cloud optical depths over areas in East Asia impacted by air pollution during the early 1990s that would have corresponded to the likely increase in anthropogenic aerosol concentrations over this period in East Asia in response to growing anthropogenic emissions. A likely explanation for these findings is that there is a mechanistic coupling between anthropogenic aerosol concentrations and cloud optical properties; one such mechanism is the so-called first and second indirect effect by which aerosols enhance the optical depths and albedos of clouds by increasing the number of droplets within clouds and suppressing precipitation from clouds, respectively. The regressions further suggest that the cloud optical depths increase on average by 0.16 for each 1 mg m(-2) increase in the column-integrated anthropogenic aerosol burden. Simple box-model calculations suggest that this is equivalent to a cooling over the model domain from anthropogenic aerosols via the indirect effect that is a factor of about 1.5 times that from the direct effect. Accounting for a possible underestimate in model-simulated aerosol concentrations over the model domain of as much as a factor of 0.6 would lower the estimated cooling from the indirect effect to about 1 times that from the direct effect. In contrast to the results using ISSCP-derived cloud optical depths, the correlation between the model-calculated anthropogenic aerosols and average cloud amount over the model domain was relatively weak and inconsistent. This result arose perhaps because of a cancelling of the competing influences on cloud lifetime and frequency by the second indirect effect and the so-called semi direct effect (i.e., the suppression of clouds by absorbing aerosols).