Abstract
Abstract
Low tropical cyclone (TC) occurrence in the vicinity of Hawai‘i introduces high uncertainty in assessing regional TC activity, motivating the use of modeling approaches. This work assesses TC activity using synthetic TCs generated by the Columbia Hazard model (CHAZ), forced with environmental fields from coupled climate models. Simulations span the historical period (1980–2014), mid-century (2040–2070), and the late 21st century (2070–2100) under high greenhouse gas emissions that represent a “worst case” scenario and allow for a larger signal for TC activity changes. We analyze projected changes in track density, translation speed, maximum sustained wind speed, and TC trajectories, and examine sensitivity to the choice of moisture variable in the CHAZ genesis index. By the late 21st century, translation speeds and TC duration show no significant change, while small but significant increases are found in occurrences of maximum sustained wind speeds over 100 kts. Depending on the moisture variable, projections show either no change or an increase in the frequency of nearby and landfalling TCs. By the late century, modulation of TC intensity by ENSO is reduced. However, we find a projected dipole change in ENSO-dependent TC track density, with El Niño shifting TCs southwest of the island chain compared to La Niña, though landfalling TC frequency shows no significant difference between ENSO phases. Finally, return periods for both nearby and landfalling TCs shorten, suggesting elevated hazard despite low event counts. These projections are subject to substantial uncertainty arising from CMIP6 biases, CHAZ structural limitations, and the limited observational record.