Distribution Prediction Model of a Rare Orchid Species (<i>Vanda bicolor</i> Griff.) Using Small Sample Size

Abstract

Advancement in field of GIS and Information Technology has taken conservation works and strategies a step further as most conservation works are now dependent on these technologies. The present study explores the prediction ability of MAXENT using a very low sample size by applying jackknife analysis over a well defined smaller region and using only climate data. <i> Vanda bicolor </i> is a horticulture important orchid grown in certain patches of North Eastern region of India and the species considered to be “ Vulnerable ” . Present study reports a distribution prediction model using different geo-climatic parameters for a small area. Model validation by ground truth ing gives a significant success ful result which clearly defines the ability of MAXENT prediction model to give high success rate (71%) with low training samples. Use of the low sample size over a larger area results in unstable models however application of these samples in smaller radius around the occurrence points could provide good working models.