Forest Research and Development

Forest Research and Development

مدل‌سازی مطلوبیت رویشگاهی توس (Betula litwinowii Doluch.) در جنگل‌های هیرکانی با استفاده از مدل آشیان بوم‌شناختی Maxent

Document Type : Scientific article

Author
Department of Forest Science and Engineering, Faculty of Natural Resources and Marine Sciences, Tarbiat Modares University
10.30466/jfrd.2026.57136.1811
Abstract
abstract
Background and objectives: The Hyrcanian forests constitute a continuous belt of temperate deciduous broadleaf forests extending along the northern slopes of the Talesh and Alborz Mountains and are listed as a UNESCO World Heritage site. Owing to their long evolutionary history and high species richness, these forests provide habitat for dozens of tree and shrub species and form an important component of the country’s biodiversity. However, increasing anthropogenic pressures resulting from the unsustainable exploitation of this ecosystem—including overgrazing, expansion of agricultural lands and mining activities, and infrastructure development—together with climate change, have posed serious threats to its structure and dynamics. This challenge is even more critical for rare species such as Betula litwinowii Doluch., whose distribution is limited to a small number of habitats with restricted area located at elevations above 2200 m and in the semi‑alpine zones of the Hyrcanian region. The light‑demanding nature of this species, its low tolerance to high air temperatures, and its relatively weak competitiveness against dominant species of the Hyrcanian forests have further limited its natural distribution in these forests. Furthermore, its patchy distribution with relatively large distances between populations and its high sensitivity to climatic changes further emphasize the importance of accurately identifying suitable habitats and analyzing the environmental factors influencing its niche. Considering the extent of environmental pressures and the limited presence data available for this species, the use of the MaxEnt ecological niche model can provide an efficient tool for analyzing its potential distribution, determining limiting factors, and identifying priority areas for conservation. Therefore, the aim of this study was to model the habitat suitability of Betula litwinowii in the Hyrcanian forests, evaluate model performance, determine the relative contribution of environmental variables, and produce classified maps of habitat suitability in order to provide a scientific basis for the conservation management of this rare species.
Methodology: This study was conducted across the Hyrcanian forests within an elevational range of approximately 0 to 2800 m. Occurrence data of Betula litwinowii were collected through field surveys and recorded using a GPS device. To reduce spatial sampling bias, only one presence record was retained within each 1‑km grid cell. After filtering the records, the final dataset was used as input for the model. Among the initial 25 environmental variables (19 bioclimatic variables and 6 topographic variables), those showing high collinearity were removed using Pearson correlation analysis and cluster analysis. The remaining variables included Bio1, Bio2, Bio7, Bio8, Bio12, and Bio15, as well as the topographic variables Northness, Eastness, TPI, and TRI. Modeling was performed in the R environment using the packages tidymodels, sf, terra, tidysdm, and tidyterra. Model parameters were tuned using a grid search over a range of 0.5 to 5 for the Regularization Multiplier and different combinations of Feature Classes. Model performance was evaluated using the AUC and TSS indices with repeated five‑fold cross‑validation. To enhance the applicability of the results for conservation planning, the species occurrence probability map was classified into five suitability classes ranging from unsuitable to very high suitability (less than 0.1 to greater than 0.7).
Results: The analysis of the relative importance of variables based on the permutation method showed that Bio8 (mean temperature of the wettest quarter) was the most influential factor controlling the distribution of this species, contributing 43.47%. This was followed by the Terrain Ruggedness Index (TRI) with 20.81% and Bio15 (precipitation seasonality) with 17.50%. In total, these three variables accounted for approximately 81% of the model’s explanatory power. Other variables, such as annual mean temperature (Bio1), Eastness, and Northness, had lower contributions to the model. Examination of the response curves indicated that Betula litwinowii exhibited a unimodal, bell‑shaped response to Bio8, reaching its maximum suitability at temperatures close to 0 °C. For Bio15, the highest habitat suitability was observed at low values of precipitation seasonality, while increasing seasonal variability in precipitation resulted in a sharp decline in the probability of species occurrence. In contrast, habitat suitability showed an increasing trend with greater terrain ruggedness and reached its peak at the highest TRI values. The final habitat suitability map indicated that the most suitable habitats were mainly concentrated as relatively continuous patches in Mazandaran Province. In Golestan and Gilan provinces, the spatial extent of suitable habitats was more limited compared with Mazandaran. Area estimation of the suitability classes showed that 11.72% of the forest areas in Mazandaran, 1.8% in Gilan, and 0.49% in Golestan fall within the moderate to very high suitability classes.
Conclusion: The results of this study confirm that the MaxEnt model is a highly efficient and reliable tool for explaining the ecological niche of Betula litwinowii in the Hyrcanian forests. The determining role of three key factors—mean temperature of the wettest quarter, terrain ruggedness, and precipitation stability (low precipitation seasonality)—indicates that the persistence of this rare species is strongly associated with the presence of rocky microrefugia located in upper elevations and near the upper forest limit, where winters are cold and moisture availability remains relatively stable throughout the year. The spatial pattern derived from the modeling also reveals the strong habitat limitation of this species in the western and eastern parts of the Hyrcanian region and its main concentration in the central areas (Mazandaran Province). From a management perspective, these findings can provide a scientific basis for identifying priority conservation areas, designing restoration programs, implementing landscape‑based management strategies, allocating conservation resources, monitoring climate‑driven changes, and preventing local extinction of birch populations. Overall, the application of ecological niche models such as MaxEnt plays a crucial role in identifying suitable habitat domains and supporting conservation‑oriented decision‑making in the Hyrcanian forests, and can provide effective guidance for the sustainable management of this valuable species.
Keywords
Subjects


Articles in Press, Accepted Manuscript
Available Online from 10 October 2026