Predicting climate change impacts on Suaeda japonica distribution in East Asian salt marshes using ensemble modeling

Halophytes in salt marshes provide important ecosystem services, such as carbon sequestration, coastal protection, nutrient cycling, biodiversity support, and nitrogen and phosphorus removal. However, climate change threatens their survival by reducing habitat availability. This study employed ensemble species distribution modeling integrating Random Forest, Generalized Boosted Models, Generalized Additive Models, and Maximum Entropy to predict the current and future habitat of Suaeda japonica-a major component of South Korean and Japanese salt marshes-under the SSP2-4.5 and SSP5-8.5 climate scenarios across East Asia for 2050 and 2100. Eight key environmental variables were systematically selected (from a total of 48). The ensemble model demonstrated excellent predictive performance (TSS = 0.987 and ROC = 0.999), superior to that of individual models. Temperature range emerged as the most important variable (importance = 0.629). Current distribution modeling identified optimal habitats along the western coast of South Korea and Kyushu, Japan. Future projections revealed minimal differences in the emission scenarios by 2050, but substantial changes by 2100 under SSP5-8.5. Notably, habitat suitability was predicted to increase along the eastern coast of China while decreasing in the traditional South Korean and Japanese strongholds, representing a distributional inversion phenomenon. These findings inform the development of regionally appropriate conservation strategies and emphasize the importance of protecting existing core populations while preparing for habitat restoration in emerging suitable areas. The study provides strong evidence on the urgency of blue carbon management and halophytic species conservation under climate change.