Relative Importance of Environmental and Spatial Predictors for Slipper Orchid (Paphiopedilum spp.) Distribution in Thailand Using Random Forest
Issued Date
2026-04-01
Resource Type
ISSN
16866576
eISSN
26730014
Scopus ID
2-s2.0-105038218451
Journal Title
International Journal of Geoinformatics
Volume
22
Issue
4
Start Page
160
End Page
176
Rights Holder(s)
SCOPUS
Bibliographic Citation
International Journal of Geoinformatics Vol.22 No.4 (2026) , 160-176
Suggested Citation
Pholgerddee P., Pattanakiat S., Nakmuenwai P., Sattraburut T., Phutthai T. Relative Importance of Environmental and Spatial Predictors for Slipper Orchid (Paphiopedilum spp.) Distribution in Thailand Using Random Forest. International Journal of Geoinformatics Vol.22 No.4 (2026) , 160-176. 176. doi:10.52939/ijg.v22i4.4945 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/116747
Title
Relative Importance of Environmental and Spatial Predictors for Slipper Orchid (Paphiopedilum spp.) Distribution in Thailand Using Random Forest
Author's Affiliation
Corresponding Author(s)
Other Contributor(s)
Abstract
Slipper orchids (Paphiopedilum spp.) represent some of Southeast Asia's most ecologically specialized and conservation-sensitive orchid taxa. Their distribution in Thailand is shaped by intricate environmental gradients, encompassing climate, topography, vegetation structure, and geographic location. Nonetheless, comprehensive national assessments of habitat appropriateness for Paphiopedilum species are scarce. This research utilized a species distribution modelling methodology to ascertain critical environmental factors and forecast appropriate habitats for Paphiopedilum species throughout Thailand. Seventy occurrence reports were obtained from biodiversity databases, field surveys, and herbarium collections. A Random Forest modelling framework was employed utilizing twelve environmental predictors that encompass climatic, non-climatic, and spatial factors. A balanced presence-absence dataset was created, and model performance was assessed using various statistical criteria, including the Brier score. The model attained a Brier score of 0.05, signifying elevated predictive accuracy. Climatic factors accounted for 38.30% of overall variable importance, followed by non-climatic variables at 37.40% and spatial predictors at 24.30%. The model exhibited robust predictive efficacy, with an AUC of 0.94, sensitivity of 0.80, specificity of 0.76, and an overall classification accuracy of 0.78. Principal predictors encompassed latitude, longitude, elevation, vegetation structure (NDVI), and temperature-associated factors. Optimal habitats were predominantly situated in the mountainous areas of northern and western Thailand, with supplementary fragmented patches observed in certain locations of southern Thailand. The results underscore the significant influence of topographic gradients, temperature, and vegetation structure on the ecological niche of Paphiopedilum species. These findings establish a scientific foundation for pinpointing priority conservation zones and facilitating long-term management strategies for slipper orchid ecosystems in Thailand.