VEGETOS: An International Journal of Plant ResearchOnline ISSN: 2229-4473
Print ISSN: 0970-4078

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Predicting Geographical Distributions of Endemic species in the Arid desert Landscapes using three Modeling methods

Predicting Geographical Distributions of Endemic species in the Arid desert Landscapes using three Modeling methods

Species distribution models are very useful to estimate a species’ geographical distribution potential especially in vast and arid desert areas, where a plenty of endemic and/or rare species are sparsely populated. Eastern Central Asian desert (ECAD) is one of the taxonomically important geographical units. Modeling the species geographical distribution in the area is challenging because the known distribution sites of species necessary for model formation are limited. To identify methods most suitable for modeling in these conditions, three leading models namely the Maximum Entropy (Maxent), Genetic Algorithm for Rule-set Prediction (GARP) and Domain at predicting distributions of 13 endemic genera in ECAD were taken into account. The chosen species in this area vary in distribution characteristics and with available distribution points ranging from 8 to 109. The receiver operating characteristic (ROC) method and the jackknife test were employed to evaluate model predictivity. For species with occurrence localities ranging from 27 to 109, we found that Maxent and Domain performed similarly according to the area under the ROC curve (AUC), however; the visualized predictions of Domain were found greatly affected by the spatial structure of the know distribution sites of species. GARP presented the lowest AUC values among the three methods tested here and overprediction. For species with limited occurrence localities ranging from 8 to 16, jackknife tests indicated that both Maxent and GARP yielded statistically significant predictions, while Domain failed..

Special Features

Full Text

View

Track Your Manuscript

Share This Page

Media Partners

Associations