Spatially continuous estimation of urban forest aboveground biomass with UAV-LiDAR and multispectral scanning: An allometric model of forest structural diversity

Abstract

Aboveground biomass (AGB) is a key parameter for assessing the carbon sequestration potential of urban ecosystems. However, traditional empirical models for AGB estimation often have poor transferability in urban environments, leading to overestimation or underestimation and limiting the ability to create continuous spatial maps of AGB. Recently, the relatively stable allometric relationships between forest structure and AGB have been further validated. With the increasing use of UAV remote sensing to monitor forest structural diversity (FSD) in urban areas, there is an urgent need to develop a method for quickly and accurately estimating AGB using FSD. This study focuses on an urban forestry demonstration base as the research area, aiming to establish an allometric growth model based on FSD to estimate AGB, grounded in the power-law relationship between forest structure and AGB. By systematically defining FSD, integrating UAV-LiDAR and multispectral data, and performing regression analysis, allometric modeling, model comparison, and accuracy assessment of extracted indicators, we thoroughly explored the optimal parameter combinations and estimation accuracy for estimating urban forest AGB using the FSD allometric model. The results show that combining FSD indicators through allometric relationships can improve AGB estimation accuracy to 80 % (R²_b=0.80, RMSE_b=2.79 kg/m², MAE_b=2.19 kg/m²), surpassing models that use only simplified FSD indicators (R²_b=0.63). Additionally, the proposed method captures nonlinear relationships and complex interactions better than traditional MLR, avoiding the overfitting that can occur with RF and XGBoost. This study confirms that allometric relationships with FSD indicators can be used for AGB prediction, highlighting the biological and physiological significance of FSD. It provides an alternative solution for rapid and large-scale AGB assessment in Urban forest.