Tecomella undulata under threat: The impact of climate change on the distribution of a valuable tree species using a machine learning model

PLOS One | https://doi.org/10.1371/journal.pone.0326609 July 9, 2025 8 / 18 Predicting the geographical distribution of T. undulata The plant species T. undulata exhibits a marked affinity for the southern regions of Iran, with its natural geographical distribution primarily concentrated in this area. The provinces demonstrating the highest habitat suitability include Fars and Bushehr, which provide optimal environments for the species to thrive. Adjacent southern provinces, which comprise Kerman, Hormozgan, and Sistan-Baluchistan, also offer considerable levels of habitat suitability for T. undulata (Fig 4). No suitable habitats were found for this species in most of the central and northern regions of the country. Only in some areas in the north and northeast of Iran, suitable habitats with medium to low suitability were observed. One such area can be seen in Golestan province, where this species exhibits a moderate level of habitat suitability. Therefore, the habitat suitability map reveals a clear geographic distribution pattern for T. undulata, indicating a strong presence in the south of the country and a lack of suitable habitats in other regions of the country. Based on Table 2, we analyzed changes in the distribution of T. undulata under current and future climate conditions. Unsuitable (< 0.1) areas. These areas are predicted to increase under both scenarios (SSP245 and SSP585) and across all future timeframes (2030, 2050, 2070, 2090). This means that in both scenarios and all future years, regions

deemed unsuitable for T. undulata are predicted to expand. Under scenario SSP245, the unsuitable areas would increase by 6.5% by 2030 and up to 9.4% by 2090. Under scenario SSP585, the increase is slightly smaller initially (5.6% by 2030) but still reaches a 13% increase by 2090. Low (0.1–0.4) suitable areas. Under both scenarios, these areas are generally predicted to decrease across all future timeframes. Under scenario SSP245, the low suitable areas would decrease by 18.6% by 2030, and the decrease would continue at a varying rate, with a 17.4% reduction by 2090. Under scenario SSP585, a similar trend is observed, with a 17.2% decrease by 2030 and a 14.9% decrease by 2090. Fig 3. Relative Variable Importance. https://doi.org/10.1371/journal.pone.0326609.g003