Ensemble modeling of Begonia octopetala in the coastal hills of Peru to 2065

Abstract

This research synthesizes the assessment of the climate impact on the endemic species Begonia octopetala in the coastal hills ecosystem of Peru up to 2065 using ensemble modeling, which combines machine learning and statistical algorithms (Maxent, Random Forest, BioClim0, and GLM). The model was fed with CMIP6 bioclimatic variables to project habitat suitability under Shared Socioeconomic Trajectories (SSP1-2.6 and SSP5-8.5). The results reveal a critical loss of suitability under the pessimistic scenario (SSP5), projecting the total disappearance of suitable habitat by 2065. Only the SSP1 scenario shows the persistence of suitable microclimates, albeit with severe fragmentation. These findings quantify the species' vulnerability to radiative forcing and underscore the need to implement adaptive conservation strategies based on identified climate refugia