J5B.3 Detection and Characterization of Contrails Using Artificial Intelligence: Insights from a 2023 Field Campaign

Tuesday, 30 January 2024: 9:00 AM
338 (The Baltimore Convention Center)
Jay P. Hoffman, CIMSS, Madison, WI; and T. Rahmes, A. J. Wimmers, W. Feltz, M. J. Foster, J. Feltz, and C. Phillips

Contrails, also known as aviation-induced cirrus (AIC), are high-altitude ice particle clouds resulting from the condensation and freezing of water vapor within the exhaust plume of aircraft. While the radiative impacts of these clouds are substantial, they have yet to be well quantified. Recent efforts have been published employing diverse artificial intelligence (AI) methodologies and imagery acquired from the Advanced Baseline Imager (ABI) aboard the Geostationary Operational Environmental Satellite (GOES), aimed at contrail detection. This study particularly centers on adapting a Convolutional Neural Network (CNN) and a U-Net architecture for the identification of contrails within thermal infrared (IR) satellite images. A field campaign in October 2023 involving Boeing and NASA test flights utilized high-temporal-resolution GOES ABI imagery to study the entire lifecycle of contrails. Alongside contrail identifications, estimations of cloud heights are furnished. An experimental iteration of the operational ABI Cloud Top Height Algorithm (ACHA) is implemented, optimized to yield cloud heights tailored for ice cloud retrievals. The AI-centric approach is tested to enhance the resolution of ABI imagery, yielding contrail height retrievals at an impressive resolution of 500 meters.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner