Search for: All records

Abstract Cloud radiative feedback impacts early tropical cyclone (TC) intensification, but limitations in existing diagnostic frameworks make them unsuitable for studying asymmetric or transient radiative heating. We propose a linear Variational Encoder‐Decoder (VED) framework to learn the hidden relationship between radiative anomalies and the surface intensification of realistic simulated TCs. The uncertainty of the VED model identifies periods when radiation has more importance for intensification. A close examination of the radiative pattern extracted by the VED model from a 20‐member ensemble simulation on Typhoon Haiyan shows that longwave forcing from inner core deep convection and shallow clouds downshear contribute to intensification, with deep convection in the downshear‐left quadrant having the most impact overall on the intensification of that TC. Our work demonstrates that machine learning can aid the discovery of thermodynamic‐kinematic relationships without relying on axisymmetric or deterministic assumptions, paving the way for the objective discovery of processes leading to TC intensification in realistic conditions.