Modeling of bonitet classes according to the site characteristics in red pine (Pinus brutia Ten.) stands

This thesis quantified the ecological thresholds that determine site productivity (bonitet) of natural Turkish red pine (Pinus brutia Ten.) stands within the Antalya Forest Regional Directorate, produced spatial distribution maps, and projected future patterns under climate-change scenarios. Stand-inventory site-index data were modelled with multiple linear regression and regression trees, whereas presence-only layers of good, medium, poor site classes and gap-dominated (degraded) red pine areas extracted from forest management plans were analysed with Maximum Entropy (MaxEnt). MaxEnt outputs were further forced with four Shared Socio-economic Pathway (SSP) trajectories (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) to 2100. The MLR model (R² = 0.41) showed that higher temperature, solar-radiation, topographic illumination and topographic position indices depress site index. CART (R² = 0.52) captured non-linear thresholds, identifying tai ≤ 26.5 and radiation index ≤ 0.185 as the best rule for high-quality sites. MaxEnt predicted the good site class within bio5 = 29.5–35 °C, bio12 = 650–1000 mm, bio17 = 22–36 mm and clastic/metamorphic lithologies (AUC = 0.76). Moderate and poor classes were separated by increasingly limiting climate terrain thresholds, while steep slopes plus very low dry season rainfall delineated potential degraded areas. SSP projections indicate a contraction of good-quality habitats at low elevations and a shift towards 800–1200 m by 2100. Consequently, in-situ genetic conservation, climate-smart revision of management plans, site-adapted silviculture, strengthened fire prevention, and long-term monitoring are recommended to sustain the ecological and economic functions of P. brutia forests and to build climate-resilient stands in the region.