Predicting species-specific diameter growth rate for Caribbean trees using mixed-effects extreme gradient boosting
| Authors: | Sheng-I Yang, Thomas Brandeis, Eileen Helmer, Humfredo Marcano-Vega |
| Year: | 2025 |
| Type: | Scientific Journal |
| Station: | Southern Research Station |
| DOI: | https://doi.org/10.1016/j.foreco.2025.122520 |
| Source: | Forest Ecology and Management |
Abstract
Caribbean islands encompass diverse forest ecosystems which provide valuable ecosystem services to their inhabitants. Currently, researchers rely on global generic or broad regional models or default values to predict tree species growth rates, which, in some cases, have been developed in temperate forests. The ability to understand the factors influencing the growth rates of Caribbean forest ecosystems, including both native and non-native species, has implications for projecting changes in and managing these forests under climate change scenarios. The objective of this study was to predict species-specific diameter growth rates for Caribbean trees with mixed-effects extreme gradient boosting (XGBoost), which is a hybrid approach combining a mixed model and a machine learning algorithm. The predictability of the models with and without the inclusion of climatic and environmental variables as predictors was examined. Long-term data collected by the US Department of Agriculture (USDA), Forest Service, Forest Inventory and Analysis (FIA) program in Puerto Rico and the U.S. Virgin Islands were used in analyses.
Results show that mixed-effects XGBoost is advantageous for providing species-specific predictions from both fixed and random effects, as well as capturing the remaining variability from XGBoost. Models produced more accurate predictions of growth rate for forests in the U.S. Virgin Islands than Puerto Rican forests. Among 30 variables examined, average diameter at breast height, average total tree height, average height-diameter ratio and average competition index play the most important roles for both islands. Adding topography- and climaterelated variables can improve the prediction accuracy of annual diameter increment. This work will serve as a working example to demonstrate the application of the methodology to continuously monitor forest resources for the vulnerable ecosystems in Caribbean as well as other mixed-species forests.