A machine learning model for estimating tree age based on trunk radius and radial growth rate

Due to substantial differences in radial growth among tree species and environments, accurate tree age estimation remains challenging. This study developed a generalized multi-species tree age estimation framework by incorporating approximate relative radial growth rate ( RGR ) into machine learning and ensemble models. A total of 186 stem discs from four tree species were collected at a sampling height of 1.3 m above ground. Annual rings were sequentially grouped from the outermost ring toward the pith to derive stem radius ( R ), age at the sampling height, average annual-ring width ( AAW ), and RGR . Four machine learning models, including Backpropagation Neural Network, Random Forest, Extreme Gradient Boosting, and Support Vector Regression, were trained using R and RGR as input variables, and their predictions were integrated using a weighted ensemble strategy. The ensemble model achieved an R 2 of 0.865, a prediction accuracy of 85.5%, an RMSE of 4.318 years, and an MAE of 3.058 years. Compared with the model using R and TreeSpecies , incorporating RGR increased R 2 from 0.589 to 0.865 and reduced RMSE from 7.345 to 4.318 years. These results indicate that RGR partially reduces the model’s dependence on explicit species information. The proposed framework shows potential for multi-species tree age estimation and provides a basis for developing generalized large-scale models for tree age estimation.