Non-stationary temperature extremes in South Korea: An extreme value analysis of global warming impacts

Abstract

Understanding the evolution of temperature extremes under global warming is crucial for effective regional climate adaptation. In South Korea—a region undergoing rapid urbanization and facing rising heat-related risks—there is a growing need to better understand these changes, particularly in relation to global climate drivers. To meet this need, this study employs a non-stationary Generalized Extreme Value (GEV) model that incorporates global mean surface temperature (GMST) as a covariate. Daily maximum and minimum temperatures from 60 nationwide weather stations were analyzed, focusing on summer and winter extremes over durations ranging from 1 to 15 days during the period 1974–2023.

The results reveal distinct regional and seasonal patterns of non-stationarity. Winter extremes have become milder, with broader temperature distributions and fewer cold events, indicating a weakening of cold spells. However, in terms of adverse societal and environmental impacts, summer extremes are more vulnerable. Tropical night events exhibit the greatest sensitivity to global warming, particularly along the west and south coasts, where their frequency and persistence have increased sharply. Heatwaves have also intensified, though less dramatically. These changes appear linked to large-scale climate dynamics, including intensification of the North Pacific Subtropical High and warming of surrounding seas, which enhance atmospheric moisture transport toward the Korean Peninsula. By integrating observational data with a dynamic statistical framework, this study provides new insights into the changing nature of temperature extremes and their drivers—offering a valuable basis for targeted adaptation strategies in South Korea.