Cloud Geometric Thickness and Improved Cloud Boundary Detection with the GOES ABI

Quantification of the boundaries of cloud layers is difficult from the perspective of passive sensors, including the Advanced Baseline Imager (ABI) onboard GOES-16 and GOES-17. Most information about clouds viewed by the ABI is concentrated near cloud top. Cloud base, which is particularly relevant for aviation applications, is oftentimes completely ambiguous using ABI-only observations. The presence of multilayer clouds further complicates cloud boundary retrievals, leading to errors in cloud top assignment and an inability to detect lower clouds when obscured by thick, upper clouds.

This work will present a number of methods we are using to improve operational cloud boundary retrievals from the ABI. Chiefly, observations from the CloudSat radar and CALIPSO lidar are being used to train the ABI retrievals and glean information about cloud extent that can then be applied globally. Several examples of cloud systems viewed from the ABI, and transected by CloudSat and CALIPSO, will be presented to demonstrate how the picture of cloudiness from ABI can be augmented by the active sensors to gather information that can in turn be utilized for ABI-only retrievals. Active sensor "truth" of cloud layer boundaries will be compared to ABI-only based classifications of clouds, revealing areas of potential improvement.

Progress on a number of methods designed to improve the ABI cloud layers product will be discussed, including the use of numerical weather prediction layer humidity information, and initial results from a multispectral methodology we are developing. In particular, we are utilizing the ratio of reflectances between the ABI 1.37 micrometer cirrus channel and 0.65 micrometer red channel to aid in determining the cloud top of low, optically thick cloud layers that are partially obscured under a higher cirrus layer. An associated nighttime retrieval using the 3.9 and 11.2 micrometer bands will also be discussed.