earticle

영어

The purpose of this study is to predict regions suitable for the growth of Lilium cernuum and analyze changes in the distribution of these areas under climate change. This study employed two statistical models from species distribution modeling, namely Generalized Additive Models (GAM) and Multivariate Adaptive Regression Splines (MARS), as well as two machine learning models, Random Forest (RF) and Maximum Entropy Algorithm (MaxEnt). To minimize uncertainties inherent in statistical models, an ensemble model was applied to integrate and address these variabilities. The prediction of potential habitats for Lilium cernuum demonstrated high predictive accuracy, with AUC values exceeding 0.9 across all four models. MaxEnt achieved the highest AUC value (0.983, sd = 0.007), while GAM recorded the lowest (0.903, sd = 0.048). The results revealed that areas with a very high probability of Lilium cernuum distribution are currently concentrated along the Baekdudaegan mountain range, extending from Jirisan National Park in South Korea to the Hamgyeongnam-do region in North Korea, primarily in high-altitude forested areas. By 2070, under both RCP 4.5 and RCP 8.5 scenarios, potential habitats are projected to shift to the Primorskiy region in Russia. These findings suggest that with ongoing climate change, the current potential habitat of Lilium cernuum , primarily located along the Baekdudaegan range on the Korean Peninsula, is expected to contract significantly by 2070, with a concomitant northward migration toward the Central Sikhote-Alin mountain range in the Primorskiy region of Russia, favoring high-altitude forested zones.