Abstract
Recent studies identify large uncertainties in the projections of tropical cyclone (TC) activity due to discrepancies in tropical Pacific sea surface temperature (SST) warming patterns. While observational datasets consistently reveal a La Niña-like warming pattern [0.15°C–0.25°C (10 yr) −1 relative cooling in the eastern equatorial Pacific], over 80% of CMIP6 models project an erroneous El Niño-like trend. These discrepancies arise from biases in cloud feedbacks, Walker circulation strength, and oceanic upwelling processes. This review examines the key mechanisms shaping observed versus modeled warming patterns, evaluates the complex link between tropical SST patterns and TC activity, and explores the feasibility of storm-resolving models for improving TC projections. We propose that pattern-conditioned TC projections using convection-permitting models, alongside physics-informed interpretations, offer a path forward in reducing uncertainties in future climate predictions.