Retrieval and Attribution of Tropical Cyclone Vertical Tilt From SAR and Infrared Satellite Imagery

Abstract Tropical cyclones (TCs) are destructive natural hazards, and their vertical tilt, defined as the displacement between the sea surface center and cloud top center, plays a key role in storm structure and intensity changes. This study constructed a global satellite‐based tilt data set from 1,024 paired synthetic aperture radar (SAR) and infrared (IR) images, with SAR locating the sea surface center and IR identifying the cloud top center, spanning 264 TCs during 2016–2024. Statistical analyses reveal systematic links between TC tilt and TC intensity, motion, and latitude, with large tilts frequently occurring during weakening. Attribution analyses using interpretable regression and deep learning models show vertical wind shear, tropospheric temperature differences, TC intensity, and cloud‐related indices as dominant controls, with shear orientation strongly constraining tilt direction. These findings highlight the value of satellite‐derived tilt mapping, while deep learning attribution quantifies nonlinear interactions among key factors, advancing our understanding of TC vertical structure.