Simona Meiler, Chahan M. Kropf , Jamie W. McCaughey, Chia-Ying Lee, Suzana J. Camargo, Adam H. Sobel, Nadia Bloemendaal, Kerry Emanuel, David N. Bresch

Future tropical cyclone risks will evolve due to climate change and socio-economic development, entailing substantial uncertainties. An uncertainty and sensitivity analysis of these risks is thus vital, yet outcomes are influenced by the chosen model setup. This study investigates how much future tropical cyclone risks are driven by climate and socio-economic changes; quantifies the uncertainty from propagating alternate representations of these systems through the entire risk modeling chain; and evaluates how strongly each model input contributes to output uncertainty. By comparing these three elements - drivers, uncertainty, and sensitivity - across four distinct TC models, we can derive findings that are generalizable beyond individual model setups. We find that average TC risk will increase 1-5% by 2050 globally, with maximum risk increases ranging from 10-400% by 2100, depending on hazard model choice and region, while the dominant source of uncertainty shifts with modeling choices. Finally, we differentiate between aleatory, epistemic, and normative uncertainties, offering guidance to reduce them and inform better decision-making.