Tovaranonte, Jantrararuk4; Barfod, Anders S.5; Balslev, Henrik5; Overgaard, Anne Blach5; Svenning, J.-C.5
1 Department of Bioscience, Science and Technology, Aarhus University2 Department of Bioscience - Ecoinformatics and Biodiversity, Department of Bioscience, Science and Technology, Aarhus University3 Department of Mathematics, Science and Technology, Aarhus University4 Department of Mathematics, Science and Technology, Aarhus University5 Department of Bioscience - Ecoinformatics and Biodiversity, Department of Bioscience, Science and Technology, Aarhus University
As a consequence of the decimation of the forest cover in Thailand from 50% to ca. 20 % since the 1950ies, it is difficult to gain insight in the drivers behind past, present and future distribution ranges of plant species. Species distribution modeling allows visualization of potential species distribution under specific sets of assumptions. In this study we used maximum entropy to map potential distributions of 103 species of palms for which more than 5 herbarium records exist. Palms constitute key-stone plant group from both an ecological, economical and conservation perspective. The models were built on information extracted from more that 1,900 geo-referenced herbarium vouchers and a number of carefully selected predictor variables (four climatic, five environmental and two spatial variables). The performances of different models were compared using the Receiver Operating Characteristics (ROC) and the Area Under the Curve (AUC). All models performed well with AUC scores above 0.95. The predicted distribution ranges showed high suitability for palms in the southern region of Thailand. It also shows that spatial predictor variables are important in cases where historical processes may explain extant distribution patterns.