Initial assessment of AIRS cloud phase determination

An assumption of cloud thermodynamics phase is typically a prerequisite in the retrievals of cloud properties, such as cloud particle size, optical thickness, and water content. An existing bispectral infrared cloud phase algorithm has been used by Moderate Resolution Imaging Spectroradiometer (MODIS) for years. The primary advantage of a pure IR approach is that it can be applied to both day and night time scenes. Radiative transfer simulations by Nasiri and Kahn (2008) indicate that the bispectral IR phase algorithm has limitations of misclassifying the water and ice clouds in the mid-temperature range between 250 and 265 K. An improvement has been proposed that would make use of the greater sensitivity to phase sensitivity provided by hyperspectral observations, such as those from Atmospheric Infrared Sounder (AIRS). AIRS provides copious information in the atmospheric window that can be used in the remote sensing of clouds.

A synthesized dataset including geolocation fields, radiances and brightness temperatures for multiple infrared channels, and global surface emissivity using satellite observations from the AIRS, CALIPSO/CALIOP and MODIS instruments has been generated. This dataset is being used to develop a hyperspectral IR cloud phase retrieval algorithm for AIRS. In this presentation, we will show the results of applying the AIRS cloud phase algorithm to one month of data along the CALIPSO track. These results will be evaluated against the CALIPSO/CALIOP 1-km cloud layer product retrievals of several quantities, such as depolarization ratio, attenuated backscatter, and mid-layer cloud temperature. The sensitivity of IR cloud phase classification algorithm will be tested with respect to cloud height, cloud opacity, and mid-layer cloud temperature.