Modeling habitat suitability of Quercus castaneifolia in the Hyrcanian forest: a comprehensive integration of environmental factors for conservation insights

An accurate understanding of how tree species distribution changes along the environmental gradient is essential in applying conservation strategies in old-growth temperate deciduous forests. This study focuses on Quercus castaneifolia within the Hyrcanian forest, covering 1.9 million hectares. We used species distribution modeling (SDM) with various machine learning algorithms, including random forest (RF), support vector machine (SVM), k nearest neighbor (kNN), maximum entropy (MaxEnt), and generalized linear model (GLM). The models were applied alongside sixteen environmental variables formatted as raster with a spatial resolution of 1 × 1 km to predict suitable habitats for Q. castaneifolia in this forest. The models were performed with a tenfold cross-validation and evaluated using the area under the curve (AUC). The results showed that while the importance of variables influencing the distribution of Q. castaneifolia varied among different models, soil bulk density consistently emerged as the most important factor in predicting the species presence. All models showed good performance, with RF exceeding the others with a mean AUC of 0.77, followed by MaxEnt (AUC = 0.75), SVM (AUC = 0.72), kNN (AUC = 0.71), and GLM (AUC = 0.70). The main outcome of our study includes spatial maps depicting suitable habitats for Q. castaneifolia based on their distribution pattern across the Hyrcanian forest. The suitable habitats are concentrated in the eastern parts of the Hyrcanian forest, which are characterized by lower precipitation and higher temperature than other parts. This study provides insight into sustainable forest restoration and future plantations using Q. castaneifolia in the Hyrcanian forest. In a broader context, the study helps understand species distribution patterns and guide conservation strategies.