A method of cloud field detection over Antarctica during the polar night using AVHRR data

An improved algorithm for cloud field detection over the Antarctic region is presented. The algorithm is a combination of a conventional split window method and a spatial coherence method. The conventional split window method utilizes the brightness temperature and the brightness temperature difference between split window channels. The spatial coherence method makes use of the standard deviation of the brightness temperature difference between split window channels as an index of cloud field homogeneity. The algorithm, in principle, enables us to detect an inhomogeneous cloud field which has not been well detected with only the conventional split window method. This algorithm was applied to Advanced Very High Resolution Radiometer (AVHRR) data during the polar night over the Antarctic region. It is found that the improved algorithm seems to work well for an inhomogeneous cloud field as expected, although there still remains uncertainty related to the vertical structure of the cloud field. In order to investigate the vertical structure in detail, further analyses are needed such as comparison of the result to ground-based meteorological observation, radiative transfer simulation, and analysis of TIROS-N Operational Vertical Sounder (TOVS) data.