Assessment of Tropical Cyclone Activity From a Statistical‐Dynamical Downscaling Model in the Northern Pacific and Atlantic Oceans

Abstract This study investigates the interannual variability of tropical cyclones (TCs) from 1951 to 2019 using observational data sets and synthetic TCs generated by the Columbia HAZard (CHAZ) model, a hybrid TC risk model. Analysis of TC genesis and landfall indicates that CHAZ reasonably captures the spatial distribution of TC activity during the Northern Hemisphere TC season in the tropical western Pacific, eastern Pacific, and Atlantic Ocean basins. The CHAZ simulation of TC genesis, track distribution, and landfall latitude is most accurate in the North Atlantic, where interannual variability is also best represented. Observed interannual correlation coefficients between TC genesis and landfall generally exceed 0.5 in the western Pacific and North Atlantic, but are lower in the eastern Pacific; CHAZ simulates higher correlations across all three basins than observations. Further analysis examines the influence of seven climate modes: El Niño‐Southern Oscillation, Pacific Meridional Mode (PMM), Atlantic Meridional Mode (AMM), Tropical Upper‐Tropospheric Trough in the Pacific (TUTT_Pac) and Atlantic (TUTT_Atl), Western Pacific Subtropical High, and North Atlantic Subtropical High. In the western and eastern North Pacific Ocean basins, heightened TC activity is observed during positive PMM and negative TUTT_Pac phases. In the Atlantic, La Niña, positive AMM, and negative TUTT_Atl phases enhance TC activity. Interactions between modes can amplify or dampen their individual effects; notably, the combination of negative TUTT_Atl and positive AMM phases increases TC landfall occurrence in the Atlantic by over 50% relative to climatological norms in both observations and CHAZ.