We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
NICHE BASED DISTRIBUTION MODELLING AND MAPPING OF BRUTIAN PINE IN THE GÖLHISAR DISTRICT.
- Authors
ŞENTÜRK, Özdemir
- Abstract
The purpose of this work was to elucidate the fundamental characteristics and application areas of various modelling techniques that are widely employed in current ecological modelling research. Using five distinct distribution modelling techniques, possible species distribution modelling and mapping of the Brutian pine species in the Gölhisar district were conducted. The data was collected from Brutian pine species in 400 sampling plots in the area. The variables used in the models were elevation, slope, aspect, radiation index, heat index, topographic position index and bedrock types. Logistic regression, classification tree, random forest, generalized additive model and maximum entropy were used as the species distribution modelling methods. Receiver Operating Characteristics (ROC) curves were created and the performance of the species distribution models was evaluated with the Area Under the ROC curve (AUC). The statistical analyses revealed that the best models were generalized additive model, random forest, classification tree, maximum entropy and logistic regression, respectively. Elevation and bedrock types had the highest contribution to the Brutian pine distribution models. The outputs of the generalized additive model technique that had the highest AUC value were mapped. Some ecological and statistical differences were found between the models and their reasons were presented. Compared to the methods commonly used in species distribution modelling studies, generalized additive model technique has a specific smoothing function which ensures both fittings between the environmental changes and explanatory curves and more accurate ecological interpretation of the models obtained.
- Subjects
RECEIVER operating characteristic curves; SPECIES distribution; RANDOM forest algorithms; ECOLOGICAL models; LOGISTIC regression analysis; REGRESSION trees
- Publication
Journal of Forestry Society of Croatia / Sumarski List Hrvatskoga Sumarskoga Drustva, 2024, Vol 148, Issue 11/12, p541
- ISSN
0373-1332
- Publication type
Academic Journal
- DOI
10.31298/sl.148.11-12.2