Accurate prediction of individual tree mortality is essential for informed decision making in forestry. In this study, we proposed machine learning models to forecast individual tree mortality within the temperate Larix gmelinii var. principis-rupprechtii forests in Northern China. Eight distinct machine learning techniques including random forest, logistic regression, artificial neural network, generalized additive model, support vector machine, gradient boosting machine, k-nearest neighbors, and naive Bayes models were employed, to construct an ensemble learning model based on comprehensive dataset from this specific ecosystem. The random forest model emerged as the most accurate, demonstrating 92.9% accuracy and 92.8% sensitivity, making it the best model among those tested. We identified key variables impacting tree mortality, and the results showed that a basal area larger than the target trees (BAL), a diameter at 130 cm (DBH), a basal area (BA), an elevation, a slope, NH4-N, soil moisture, crown density, and the soil’s available phosphorus are important variables in the Larix Principis-rupprechtii individual mortality model. The variable importance calculation results showed that BAL is the most important variable with an importance value of 1.0 in a random forest individual tree mortality model. By analyzing the complex relationships of individual tree factors, stand factors, environmental, and soil factors, our model aids in decision making for temperate Larix gmelinii var. principis-rupprechtii forest conservation.

Crown width (CW) is a critical metric for characterizing tree-canopy dimensions; however, its direct measurement remains labor-intensive and is often impractical in inaccessible crowns. Consequently, CW is frequently derived from projections, which are susceptible to multiple sources of imprecision, including canopy density, crown irregularity, terrain heterogeneity, and the observer’s vantage point, especially in structurally complex natural forests. While deep neural network (DNN) models show substantial potential for CW prediction, their performance in heterogeneous forests remains uncertain. We developed DNN models integrated with a Height Threshold Method (HTM) to predict individual-tree CW in the natural mixed forests of Northern China, dominated by Larix principis-rupprechtii and Picea asperata. Our study further compared the relative importance of feature engineering versus model architectural complexity in predictive accuracy and identified the key ecological variables governing CW. The model performance was evaluated through the coefficient of determination (R²), mean square error (MSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). Field surveys of 34 representative sample plots produced 1884 individual-tree records. The main results were as follows: (1) all DNNs avoided overfitting, and were statistical stable under ten-fold cross-validation; (2) the optimized DNN3-2 model (tuned hidden layer count, neurons/hidden layer, L₂ regularization, and dropout) achieved peak performance, explaining 69% of CW variance with residuals with stable variance and excellent coverage properties; (3) tree size, neighborhood competition, species identity, and site quality were the most important predictors; and (4) stand parameters calculated from competitive neighborhoods defined by the HTM, particularly mean stand crowding, Simpson’s index (1-D), and Shannon’s index (H′), significantly improved prediction accuracy. By integrating DNN with the HTM, our approach allows for accurate prediction of individual-tree CW in natural mixed forests of Northern China, dominated by Larix principis-rupprechtii and Picea asperata.

Accurate estimation of individual tree Above-Ground Biomass (AGB) is essential for assessing forest carbon sequestration. This study integrates multi-source Unmanned Aerial Vehicle (UAV) remote sensing (LiDAR and RGB) with machine learning to estimate the AGB of Larix principis-rupprechtii in a natural secondary forest. We applied an instance segmentation approach to identify individual trees and extract structural and spectral features, which were subsequently optimized before model training. Our results demonstrate that models utilizing combined multi-source features significantly outperformed those relying on a single data source. The Extreme Gradient Boosting (XGBoost) algorithm achieved the best performance, with an R² of 0.770 using the combined feature set. SHapley Additive exPlanations (SHAP) interpretation revealed that structural attributes—particularly tree height and crown volume—were the most influential predictors, underscoring their greater importance over spectral information. This study presents an effective and interpretable framework for accurate tree-level AGB estimation, supporting scalable monitoring of regional forest carbon dynamics.