Iranian Society of Forestry Iranian Journal of Forest 2008-6113 17 3 2025 11 22 Snow as a natural driver and its consequences on the structure of mixed broadleaf forest stands in the Hyrcanian region (Case study: Liresara forest, Nowshahr) Snow as a natural driver and its consequences on the structure of mixed broadleaf forest stands in the Hyrcanian region (Case study: Liresara forest, Nowshahr) 485 509 238052 10.22034/ijf.2025.531403.2054 FA M Amiri Associate Prof., Dept. of Environmental Engineering, Faculty of Natural Resources, Semnan University, Semnan, Iran. 0000-0003-4920-3602 H Ravanbakhsh Assistant Prof., Forest Research Division, Research Institute of Forests and Rangelands, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran. 0000-0003-0990-0112 M Mostafa 3Assistant Prof., Dept. of Soil, Plant and Food Sciences (DISSPA), University of Bari, 70199 Bari, Italy M Mohammady Associate Prof., Dept. of Environmental Engineering, Faculty of Natural Resources, Semnan University, Semnan, Iran. Journal Article 2025 07 09 Introduction: Snow plays a crucial role in shaping the ecology of mountain forests, significantly impacting the distribution, structure, and dynamics of plant communities. In the Hyrcanian forests, especially at higher elevations, snowfall influences not only soil moisture and water storage but also the structural integrity of forest stands. This can lead to notable physical disturbances, such as broken trunks and branches, tree uprooting, and change in species composition. This study aims to explore how various tree species in the Hyrcanian forests respond to disturbances caused by snow, an area that has not been extensively researched until now. We will assess changes in tree density, diameter distribution, basal area, volume, and the vulnerability of key species to heavy snowfall. The outcomes of this research will enhance our understanding of how Hyrcanian forests adapt to climate change, providing valuable insights that can inform the development of adaptive management strategies for future environmental conditions. Material and Methods: To evaluate the damage caused by snow, three forest compartments that had experienced the most severe impacts were selected within Series 6 of the Liresara Forest, Mazandaran Province, Iran. Within each compartment, 25 sample plots measuring 40 × 40 meters were established using a systematic-random sampling method. All trees with a diameter at breast height (DBH) greater than 7.5 cm were quantitatively and qualitatively measured. The resulting data were analyzed using chi-square tests, Student’s t-tests, and Duncan’s multiple range test. Subsequently, a Random Forest algorithm was employed to model the severity of tree damage, with model performance evaluated based on overall accuracy, Kappa coefficient, balanced accuracy, and the area under the receiver operating characteristic curve (AUC). Results: The results indicated that alder species showed the highest level of damage (22%), followed by gleditsia and beech. In terms of damage types, stem breakage (53%) and uprooting (33.6%) were the most frequent, while bending (13.4%) was the least common. Data analysis revealed that trees with diameters of 25–35 cm and heights of 20–25 m were the most vulnerable. This pattern was primarily attributed to the young age and high stand density, resulting from clear-cutting operations conducted in the 1980s. The random forest model effectively ranked the relative importance of variables, identifying slenderness coefficient and total tree volume as the most influential factors associated with snow damage. Furthermore, high AUC values demonstrated the model's strong capability in accurately predicting various damage severity classes. Conclusion: The findings of this study demonstrated that morphometric characteristics of trees, such as tree slenderness ratio (TSC) and total volume, play a crucial role in the vulnerability of forest species in the Hyrcanian region to heavy snowfall. The results can inform forest stand management and contribute to mitigating. Modeling results using the random forest algorithm confirm with high accuracy that young trees in lower diameter and average height classes are the most vulnerable due to their unbalanced structure and the high density of forest stands resulting from past clear-cutting operations. Additionally, Caucasian alder was identified as the most sensitive, with (22%) of its individuals showing damage, likely due to the morphological characteristics of its brittle branches and shallow root system. From a management perspective, this study emphasizes the need to adjust the density of young stands through silvicultural interventions (such as thinning) to reduce slenderness ratio and increase trees' mechanical stability. Furthermore, the selection of less vulnerable species in reforestation programs for degraded forests could be a key strategy for adapting to climate change and reducing future damages. Introduction: Snow plays a crucial role in shaping the ecology of mountain forests, significantly impacting the distribution, structure, and dynamics of plant communities. In the Hyrcanian forests, especially at higher elevations, snowfall influences not only soil moisture and water storage but also the structural integrity of forest stands. This can lead to notable physical disturbances, such as broken trunks and branches, tree uprooting, and change in species composition. This study aims to explore how various tree species in the Hyrcanian forests respond to disturbances caused by snow, an area that has not been extensively researched until now. We will assess changes in tree density, diameter distribution, basal area, volume, and the vulnerability of key species to heavy snowfall. The outcomes of this research will enhance our understanding of how Hyrcanian forests adapt to climate change, providing valuable insights that can inform the development of adaptive management strategies for future environmental conditions. Material and Methods: To evaluate the damage caused by snow, three forest compartments that had experienced the most severe impacts were selected within Series 6 of the Liresara Forest, Mazandaran Province, Iran. Within each compartment, 25 sample plots measuring 40 × 40 meters were established using a systematic-random sampling method. All trees with a diameter at breast height (DBH) greater than 7.5 cm were quantitatively and qualitatively measured. The resulting data were analyzed using chi-square tests, Student’s t-tests, and Duncan’s multiple range test. Subsequently, a Random Forest algorithm was employed to model the severity of tree damage, with model performance evaluated based on overall accuracy, Kappa coefficient, balanced accuracy, and the area under the receiver operating characteristic curve (AUC). Results: The results indicated that alder species showed the highest level of damage (22%), followed by gleditsia and beech. In terms of damage types, stem breakage (53%) and uprooting (33.6%) were the most frequent, while bending (13.4%) was the least common. Data analysis revealed that trees with diameters of 25–35 cm and heights of 20–25 m were the most vulnerable. This pattern was primarily attributed to the young age and high stand density, resulting from clear-cutting operations conducted in the 1980s. The random forest model effectively ranked the relative importance of variables, identifying slenderness coefficient and total tree volume as the most influential factors associated with snow damage. Furthermore, high AUC values demonstrated the model's strong capability in accurately predicting various damage severity classes. Conclusion: The findings of this study demonstrated that morphometric characteristics of trees, such as tree slenderness ratio (TSC) and total volume, play a crucial role in the vulnerability of forest species in the Hyrcanian region to heavy snowfall. The results can inform forest stand management and contribute to mitigating. Modeling results using the random forest algorithm confirm with high accuracy that young trees in lower diameter and average height classes are the most vulnerable due to their unbalanced structure and the high density of forest stands resulting from past clear-cutting operations. Additionally, Caucasian alder was identified as the most sensitive, with (22%) of its individuals showing damage, likely due to the morphological characteristics of its brittle branches and shallow root system. From a management perspective, this study emphasizes the need to adjust the density of young stands through silvicultural interventions (such as thinning) to reduce slenderness ratio and increase trees' mechanical stability. Furthermore, the selection of less vulnerable species in reforestation programs for degraded forests could be a key strategy for adapting to climate change and reducing future damages.

https://www.ijf-isaforestry.ir/article_238052_7bc8336e8d0eb8c755d94f9ab7024ac2.pdf