Retrieval of aerosol optical depth in vicinity of broken clouds from reflectance ratios: Sensitivity study

We conducted a sensitivity study to better understand the potential of a new method for retrieving aerosol optical depth (AOD) under partly cloudy conditions. This method exploits ratios of reflectances in the visible spectral range and provides an effective way to avoid three-dimensional (3D) cloud effects. The sensitivity study is performed for different observational conditions and random errors in input data. The results of the sensitivity study suggest that this ratio method has the ability to detect clear pixels even in close proximity to clouds. Such detection does not require a statistical analysis of the two-dimensional (2D) horizontal distribution of reflected solar radiation, which makes it suitable for operational retrievals. In comparison with previously suggested approaches, the ratio method has the capability to increase the 'harvest' of clear pixels. Similar to the traditional independent pixel approximation (IPA), the ratio method has a low computational cost for retrieving AOD. In contrast to the IPA method, the ratio method provides much more accurate estimations of the AOD values under broken cloud conditions: pixel-based and domain-averaged estimations of errors in AOD are about 25% and 10%, respectively. Finally, both the ratio-based cloud screening and the accuracy of domain-averaged ratio-based AOD values do not suffer greatly when 5% random errors are introduced in the reflectances.