Journal of Environmental Biology

Modeling and mapping potential distribution of Crimean juniper (Juniperus excelsa Bieb.) using correlative approaches

K?rşad ?ZKAN¹*, ?zdemir ŞENT?RK², Ahmet MERT³ and Mehmet G?ven? NEGİZ³

¹Suleyman Demirel University, Faulty of Forestry, Department of Forest Engineering, 32260, Isparta, Turkey

²Mehmet Akif Ersoy University, G?lhisar Vocational School, 15400, G?lhisar, Burdur, Turkey

³Suleyman Demirel University, Sutculer Vocational School, 32950, Sutculer, Isparta, Turkey

*Corresponding Authors Email : kursadozkan@gmail.com

 Publication Data

Paper received:

17 July 2013

Revised received:

15 October 2013

Re-revised received:

18 June 2014

Accepted:

22 September 2014

Abstract

Modeling and mapping potential distribution of living organisms has become an important component of conservation planning and ecosystem management in recent years. Various correlative and mechanistic methods can be applied to build predictive distributions of living organisms in terrestrial and marine ecosystems. Correlative methods used to predict species' potential distribution have been described as either group discrimination techniques or profile techniques. We attempted to determine whether group discrimination techniques could perform as well as profile techniques for predicting species potential distributions, using elevation (ELVN), parent material (ROCK), slope (SLOP), radiation index (RI) and topographic position index (TPI))? as explanatory variables. We compared potential distribution predictions made for Crimean juniper (Juniperus excelsa Bieb.) in the Yukarı g?kdere forest district of the Mediterranean region, Turkey, applying four group discrimination techniques (discriminate analysis (DA), logistic regression analysis (LR), generalized addictive model (GAM) and classification tree technique (CT)) and two profile techniques (a maximum entropy approach to species distribution modeling (MAXENT), the genetic algorithm for rule-set prediction (GARP)).? Visual assessments of the potential distribution probability of the applied models for Crimean juniper were performed by using geographical information systems (GIS). Receiver-operating characteristic (ROC) curves were used to objectively assess model performance. The results suggested that group discrimination techniques are better than profile techniques and, among the group discrimination techniques, GAM indicated the best performance.  

 Key words

Crimean juniper, Habitat selection models, Model selection, Potential distribution, Predictive accuracy,? Predictive mapping

Copyright ? 2015 Triveni Enterprises. All rights reserved. No part of the Journal can be reproduced in any form without prior permission. Responsibility regarding the authenticity of the data, and the acceptability of the conclusions enforced or derived, rest completely with the author(s).