This research article presents a comprehensive analysis of the distribution
patterns of Rhododendron arboreum in Nepal, encompassing the current
distribution and future projections. We employed an ensemble approach by
combining eight selected models with AUC > 0.8. Resulting ensemble model
demonstrated remarkable predictive accuracy, achieving an AUC value of 0.97
and a TSS value of 0.77. Our analysis unveiled that within total land area of
Nepal, approximately 115,317.1 km2 were deemed unsuitable for Rhododendron
arboreum, while 32,198.89 km2 were identified as suitable habitats. Future
projections for 2050 and 2070 under the SSP2-4.5 and SSP5-8.5 scenarios showed
significant changes in habitat suitability. By 2050, under SSP2-4.5, 10,710.39 km2
were projected to be lost, while 4,792.78 km2 were gained. Under SSP5-8.5, more
significant loss of 14,237.64 km2 and a smaller gain of 5,696.41 km2 were
anticipated. In 2070, further losses in suitable habitat were projected under both
scenarios. A consistent upward movement in the suitable habitat elevation range
was observed from the current period to the projected years of 2050 and 2070.
These findings underscore the urgency of implementing proactive conservation
measures to safeguard Rhododendron arboreum in the face of climate change. The
research provides valuable insights for informing conservation strategies and
emphasizes the ecological significance of this species in Nepal. Considering
distribution patterns and elevation shifts, our study contributes to a
comprehensive understanding of the species dynamics and supports effective
conservation planning.
Keywords: Biomod2; Climate change; Ensemble modeling; Rhododendron
arboreum; Species distribution modeling
