MABL-AI: A Machine Learning Approach to Data Fusion for High Resolution Satellite Profiles of the Marine Atmospheric Boundary Layer

MABL-AI is an artificial intelligence (AI) data fusion framework for deriving high resolution thermodynamic profiles of the Marine Atmospheric Boundary Layer (MABL). MABL-AI uses available single-source satellite retrievals and other complementary data sources to upscale satellite retrieved temperature and moisture profiles of the MABL (0-5 km above sea level) to a vertical resolution of 50 m capable of capturing important MABL features, including inversions and MABL boundaries. This programÆs Phase II objectives include, amoung others, the expansion of MABL-AI to global maritime environments, optimizing the deep learning framework, performing robust validation and verification against in situ measurements, developing the prototype MABL-AI software package for integration into Navy operational systems, and testing MABL-AI performance in a variety of use cases including in a near-real-time capacity as part of a Navy field campaign.