MODELING THE IMPACTS OF CLIMATE CHANGE ON TROPICAL CYCLONE FORMATIONS IN THE WESTERN NORTH PACIFIC

Abstract

A dynamical-statistical modeling approach was used to investigate climate change impacts on tropical cyclone (TC) formations in the western North Pacific (WNP). Reanalysis data and climate model projections were analyzed for 1981–2050 to identify both past and future impacts. The projections came from the Community Climate System Model version 4 (CCSM4). The primary analysis tool was a statistical model that uses information about large-scale environmental factors (LSEFs) to predict the probability of TC formation. This model was built using data on the LSEFs and TC formations during the last several decades. We forced the statistical model with the reanalysis LSEFs and with the climate model LSEFs outputs from the climate model to produce formation probability analyses for 1981–2018. The two methods for calculating the probabilities both showed increasing trends during this period for the central WNP region in which most WNP formations occur. These trends are consistent with the observed increase in TC formations in this region during this period. The climate model probabilities based on long-term mean, El Niño, and La Niña LSEFs for 1981–2018 were also consistent with observed TC formations. For future decades, the climate model projections showed multi-decadal increases in both TC formation probabilities and the WNP area in which TC formation is most probable. We recommend continued use of dynamical-statistical methods as observational data and climate models improve.