| Abstract |
Plant communities are unevenly distributed in space, shaped by both abiotic and biotic factors. Several methods have been developed for delineating their extent, including the spatial analysis of vegetation patterns using tools such as vegetation maps, climate-based simulations, and the use of characteristic species distribution. However, limited knowledge exists about which species are most suitable for this purpose. In this study, we aimed to delimit the seasonally dry tropical forest (SDTF) in Mexico based on the distribution area of vascular plant species endemic to Mexico and registered to this biome. Endemic species serve as key indicators for delineating biomes, highlighting regions with stable conditions and unique evolutionary and biological characteristics. The occurrence records of species were obtained from the Global Biodiversity Information Facility (GBIF) database, and ecological niche models were generated using the ENMTML package in R. The boundaries of the SDTF were delineated by stacking species distribution models, grouping endemic species according to the proportion of their occurrence records located within the SDTF: i) ≥50% of records (SDTF 50%), ii) ≥75% (SDTF 75%), and iii) 100% (SDTF 100%). Model performance was evaluated using Kappa, sensitivity, and specificity metrics. We validated our results using Asteraceae points distributed across Mexico's major biomes and analyzed confusion matrices. A total of 3,673 endemic species were registered, and 228 met the criteria for species distribution modeling. Of these, 96% yielded models with high predictive accuracy. Among the three approaches, the model based on high-affinity species (SDTF 75%) performed best in terms of all evaluation metrics, delineating approximately 14% of Mexico's surface as SDTF. In conclusion, high-affinity species serve as reliable indicators for delineating plant communities with well-defined environmental characteristics, facilitating both the precise delineation of biomes and their application in conservation ecology and in other biomes.
|
