Investigating the influence of different environmental variables in modeling the distribution of yew (Taxus baccata L.) using the MAXENT model in Hyrcanian forests

Message:
Article Type:
Research/Original Article (دارای رتبه معتبر)
Abstract:
Background and Objective
Mapping species distributions is vital for evaluating regional conservation priorities, informing planning efforts, and guiding management actions. Species Distribution Models (SDMs) are analytical-statistical tools that use field occurrence data and environmental variables to estimate the geographic range of species. The yew tree (Taxus baccata), a relict and long-lived species, is among the most ecologically valuable trees in the Hyrcanian forests. Mapping its current distribution and identifying potential suitable habitats—as well as understanding the environmental drivers of its presence—are essential steps toward effective conservation of this endangered species. Among SDMs, the Maximum Entropy (MaxEnt) model is widely regarded as one of the most effective, particularly for presence-only data. It performs well even with a limited number of occurrence records (as few as five points) and is capable of modeling complex, non-linear relationships between predictors and species presence. Its simplicity and ease of use have made MaxEnt the most commonly applied SDM technique in ecological studies. The main goal of this research was to identify which environmental variables most influence the distribution of the yew tree.
Material and Methods
To achieve this, the primary habitats of yew in the Hyrcanian forests across Golestan, Mazandaran, and Gilan provinces were identified, and 1,614 presence points were recorded. Environmental variables included bioclimatic layers from the WorldClim database, soil parameters from SoilGrids, and topographic variables derived from a 1-km resolution Digital Elevation Model (DEM). The MaxEnt model was run under four scenarios: (M1) using only climatic variables; (M2) combining climatic and topographic variables; (M3) combining climatic and soil variables; and (M4) integrating all environmental variables while accounting for multicollinearity. Model tuning involved testing five regularization multipliers (0.5, 1, 2, 3, and 4) in combination with various feature classes (L, LQ, H, LQH, LQHP). Threshold-dependent evaluation metrics—particularly the omission rate—were used to identify the optimal parameter settings that maximized model performance. After selecting the best configuration, modeling was performed using the block method with 5,000 background points. Model accuracy was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC), a widely used measure in SDM evaluation.
Results
Model 3 (climatic + soil variables) produced the lowest AUC, while Model 2 (climatic + topographic variables) improved performance slightly, raising the AUC from 0.93 to 0.94. The highest predictive accuracy (AUC = 0.96) was achieved by Model 4, which incorporated all environmental variables after removing multicollinearity—indicating that accounting for variable interdependence enhances model reliability. Among all variables, the most influential predictors of yew distribution were the bioclimatic variables bio2, bio3, bio7, and bio18, along with elevation and slope. Collectively, these six variables explained 90% of the variation in yew presence, while soil variables contributed just 1.02% to the model's predictive power.
Conclusion
This study enhances our understanding of how abiotic factors shape the distribution of the endangered yew tree and underscores the significance of specific environmental predictors. These insights can inform the delineation of potential habitats and support targeted conservation planning in the Hyrcanian forests of northern Iran. Ultimately, the findings provide a scientific foundation for developing prioritized and coordinated conservation strategies to safeguard this valuable species within its native range.
Language:
Persian
Published:
Journal of Forest Research and Development, Volume:11 Issue: 1, Spring 2025
Pages:
25 to 39
https://www.magiran.com/p2866382  
سامانه نویسندگان
  • Corresponding Author (2)
    Seyed Jalil Alavi
    Associate Professor Department of Forestry, Tarbiat Modares University, Tehran, Iran
    Alavi، Seyed Jalil
اطلاعات نویسنده(گان) توسط ایشان ثبت و تکمیل شده‌است. برای مشاهده مشخصات و فهرست همه مطالب، صفحه رزومه را ببینید.
مقالات دیگری از این نویسنده (گان)