• Temperate conifers show complex canopies, tropics simplest across Yunnan’s three climatic zones • Crown width critical in tropics: missing data drops R² to 0.53, inclusion boosts to 0.97 • Random forest models outperform others, excel in complex forests (R²=0.98) • DBH drives coniferous biomass; random forest validates strong correlation (R²=0.98) Forest canopy structure and biomass are crucial to the estimation of carbon sink function, while the intensification of global climate change is significantly affecting the structure and function of forest ecosystems, especially in different climatic zones. In this study, we assessed the effects of climatic variations on canopy volume, stem biomass estimation, and model accuracy in three contrasting climatic zones of Yunnan Province, China: Pudacuo (characterized by a cold and humid plateau climate), the Ailao Mountains (featuring a subtropical monsoon mountain climate), and Xishuangbanna (exhibiting a tropical rainforest climate). Based on crown widthand employing multiple modeling approaches (nonlinear mixed-effects, random forest, multiple regression, deep learning, and decision tree), we examined the relationships between forest structural complexity and biomass estimation within each region. The results showed the random forest model demonstrated a higher predictive accuracy across all the climatic zones, particularly in forests with complex structures. Crown width are critical for accurate biomass estimation in tropical and subtropical regions. Specifically, in Xishuangbanna, omitting crown width resulted in a substantial underestimation of biomass (R² = 0.53), whereas its inclusion significantly improved model performance (R² = 0.97). In contrast, the relatively simple canopy structure in Pudacuo showed less sensitivity to the absence of crown width, while (Diameter at breast height) DBH was identified as the primary factor influencing biomass, with the random forest model achieving the highest prediction accuracy (R² = 0.98). Furthermore, canopy structural complexity of FSCI (Forest canopy structure complexity index) was highest in temperate coniferous forests and lowest in tropical rainforests, indicating marked structural differences among forest types under distinct climatic and disturbance regimes. These findings offer valuable insights for enhancing the accuracy of biomass estimation and advancing the understanding of forest structural complexity.
zhou et al. (Sun,) studied this question.