Understanding the contribution of structural diversity to stand biomass for carbon management of mixed forests using machine learning algorithms

The structural properties of mixed stands and its potential impact on forest carbon sink function have attracted the attention of forest managers. Comprehensively understanding how stand factors and structural diversity influence forest biomass is critical for enhancing carbon management. However, data and information on biomass variability and its relationships to stand structural features are still insufficient. The purpose of this study was to develop models that delineated the relationships between stand biomass (above and below ground) and stand factors, with a particular emphasis on structural diversity in natural mixed forests. Four machine learning (ML) algorithms named support vector machine (SVM), artificial neural network (ANN), random forest (RF) and boosted regression trees (BRT) were trained. The results indicated that SVM and ANN provided more accurate stand biomass estimations than RF and BRT algorithms. The ANN, incorporating tree size diversity, achieved the highest accuracy (R² = 0.9255 ± 0.0421), followed by SVM (R² = 0.9252 ± 0.0385). Structural diversity proved to be a reliable predictor of biomass estimation in mixed stand surpassing stand average height traditionally used. The positive correlation between stand biomass and structural diversity with the relative importance value ranged from 10.21 to 15.32% suggested that complex stand structural diversity facilitated larger biomass accumulation. Thus, our study offered a ML protocol for predicting stand biomass of natural coniferous-broadleaved mixed forests, and suggested that using comprehensive management measures such as properly promoting tree differentiation can help forest managers enhance ecosystem carbon.