Introduction: Biodiversity plays a fundamental role in maintaining forest ecosystem stability and functioning and can enhance forest productivity through mechanisms such as species complementarity and reduced intraspecific competition. Despite extensive research on biodiversity–productivity relationships, the simultaneous effects of tree species diversity, stand structure, and environmental factors on volumetric growth in the Hyrcanian forests have received limited attention. Therefore, this study aimed to model annual volumetric growth in the Hyrcanian forests and assess the importance of biotic, structural, and environmental variables using parametric and non-parametric approaches.
Materials and Methods: Data were collected from 258 permanent sample plots over a nine-year period. Annual volumetric growth (m³ ha⁻¹ yr⁻¹) was considered as the dependent variable. Explanatory variables included tree species diversity indices, stand basal area (m² ha⁻¹), basal area of the largest trees within each sample plot (m² ha⁻¹), wind speed, air temperature, solar radiation, and topographic wetness index (TWI) were included as explanatory variables. Multiple linear regression, multilayer perceptron (MLP), and radial basis function (RBF) neural networks were developed, and model performance was evaluated using R², RMSE, and BIAS.
Results: The results showed that the MLP model outperformed multiple regression (R² = 0.57) and the RBF neural network in predicting annual volumetric growth, achieving a coefficient of determination of 0.85 (R² = 0.85)Sensitivity analysis of the MLP model revealed that tree species richness had the highest individual contribution (30.9%) to volumetric growth prediction, Stand structural variables collectively accounted for approximately 29%, while environmental factors each contributed around 10% to the explained variation.
Conclusion: The findings suggest that annual volumetric growth in the Hyrcanian forests Is associated with complex and nonlinear interactions among biodiversity, stand structure, and environmental conditions. Artificial neural networks provide an effective framework for capturing such ecological complexity. These results highlight the importance of conserving tree species diversity as a key component of strategies to maintain long-term forest productivity and sustainability in the Hyrcanian forests.