The relevance of aerosol optical depth to cumulus fraction changes: a five-year climatology at the ACRF SGP site

Abstract. The objective of this study is to investigate, by observational means, the magnitude and sign of the actively discussed relationship between cloud fraction N and aerosol optical depth τ_a. Collocated and coincident ground-based measurements and Terra/Aqua satellite observations at the Atmospheric Radiation Measurement (ARM) Climate Research Facility (ACRF) Southern Great Plains (SGP) site form the basis of this study. The N–τ_a relationship occurred in a specific 5-year dataset of fair-weather cumulus (FWC) clouds and mostly non-absorbing aerosols. To reduce possible contamination of the aerosols on the cloud properties estimation (and vice versa), we use independent datasets of τ_a and N obtained from the Multi-filter Rotating Shadowband Radiometer (MFRSR) measurements and from the ARM Active Remotely Sensed Clouds Locations (ARSCL) value-added product, respectively. Optical depth of the FWC clouds τ_cld and effective radius of cloud droplets r_e are obtained from the MODerate resolution Imaging Spectroradiometer (MODIS) data.

We found that relationships between cloud properties (N,τ_cld, r_e) and aerosol optical depth are time-dependent (morning versus afternoon). Observed time-dependent changes of cloud properties, associated with aerosol loading, control the variability of surface radiative fluxes. In comparison with pristine clouds, the polluted clouds are more transparent in the afternoon due to smaller cloud fraction, smaller optical depth and larger droplets. As a result, the corresponding correlation between the surface radiative flux and τ_a is positive (warming effect of aerosol). Also we found that relationship between cloud fraction and aerosol optical depth is cloud size dependent. The cloud fraction of large clouds (larger than 1 km) is relatively insensitive to the aerosol amount. In contrast, cloud fraction of small clouds (smaller than 1 km) is strongly positively correlated with τ_a. This suggests that an ensemble of polluted clouds tends to be composed of smaller clouds than a similar one in a pristine environment. One should be aware of these time- and size-dependent features when qualitatively comparing N–τ_a relationships obtained from the satellite observations, surface measurements, and model simulations.