Determination of model maps for the potential distribution of Anatolian black pine (Pinus nigra Arnold.) in natural forest areas in the Central Black Sea region

Abstract

Özdemir ŞENTÜRK 1*, Serkan GÜLSOY 2, Mehmet Güvenç NEGİZ 3, Fatih KARAKAYA 2 ORCID: https://orcid.org/0000-0003-2513-6070; https://orcid.org/0000-0003-2011-8324; https://orcid.org/0000-0002-6338-3307; https://orcid.org/0000-0002-6534-6382 1 Gölhisar Vocational School, Departmant of Forestry, Burdur Mehmet Akif Ersoy University, Gölhisar, Burdur, Turkey 2 Faculty of Forestry, Department of Forest Engineering, Isparta University of Applied Sciences, Isparta, Turkey 3 Sütçüler Prof. Dr. Hasan GÜRBÜZ Vocational School, Department of Forest Engineering, Isparta University of Applied Sciences, Sütçüler, Isparta, Turkey *Corresponding author: e-mail: osenturk@mehmetakif.edu.tr Abstract The purpose of this study was to determine potential distribution modeling and mapping of Anatolian black pine in the Vezirköprü district. Presence / absence data of the species was collected from 586 sample areas. Environmental variables (elevation, slope, bedrock, radiation index etc.) were obtained from the digital maps. In addition, climatic variables were downloaded from WorldClim database. Generalized Additive Model and Classification and Regression Tree were used for potential distribution modeling of the species. The validation value of Generalized Additive Model was 0.84 while the cross-validation test value was found to be 0.82. Also, the ROC values of the tree model were found to be 0.804 for the training data set and 0.750 for the testing data set. According to the Classification and Regression Tree method, the locations above 650 m and without meta-sandstone are suitable for the potential distribution of the species. Under the condition that there is meta-sandstone, the areas where the temperature is between 7.6-11.0 °C and the slope degree is more than 23% coincide with the potential distribution of the species. Also, Generalized Additive Model showed that places where gabbro, ophiolitic melange, serpentine, and mixed material were seen as main bedrock type, sloping sites where average elevation was from 600 to 1150 m and temperature is between approximately 8.5-11.3 °C were the most suitable conditions for the potential distribution of the species in the district. In both models, it was determined that especially rock formations, climatic variable, and altitude are effective in the potential distribution of the species. On the other hand, although there are some differences in the variables in the models, the potential distribution maps formed by these variables overlapped quite. As a result, the information obtained in the study is important for the forestry practices such as afforestation, regeneration and the monitoring the species under future climate change conditions. Keywords: generalized additive model; classification and regression tree; species distribution model; Vezirköprü