« Back

Abstract:

Both the European continent and Romania are occupied by significate forested areas, often represented by rare species or even virgin forests. The natural distribution of the black pine (Pinus nigra ssp. Banatica) in the Banat region of Romania is still a pending spatial problem. Therefore, the main aim of this study is to assess the sustainability of morphometric parameters in modelling of Banat black pine in Domogled and Cernei Mountains, using Geographic Information Systems instruments. A set of morphometric parameters (altitude, depth, density, slope, and curvature), econometric climatic indicators (Martonne and FAI) and Vegetation indexes (NDVI and SAVI) have been derived from a 10m DEM, Sentinel 2 satellite images and WorldClim gridded data. For the modeled data, a contour of the area where the Banat black pine is found, which was validated by in situ mapping, was used. For the purpose of this study the dependent data set was split up into presence – absence of the Banat black pine and also the data was split into training and validation for both categories. For spatial analysis of the data ArcGIS 10.8.1 was used and for the statistical analysis R software was used. We obtained four models that used between two and six independent variables. All models were validated using AUC and R² McFadden and obtained very good scores (between 0,886 and 0,904 and between 0,37 and 0,42 respectively). The most important predictors were found to be Martonne Aridity Index, slope, and SAVI. Based on our finding we can conclude that a geomorphometric modeling approach can be used, successfully, in explaining the spatial distribution of a tree species in a marginal environment.