Tree-level biomass estimation using unmanned aerial vehicle (UAV) imagery and allometric equation

Abstract

Estimating forest biomass is imperative for comprehensively understanding the function of forest in regulating climate, providing theoretical support for vegetation management. Constructing allometric equations rationally is essential for accurate tree-level biomass estimation without destructive sampling, and optimizing the sample size for fitting allometric equations ensures a desirable balance between accuracy and cost. In this study, the optimal sample size and the best allometric equation for biomass estimation were discussed using unmanned aerial vehicles (UAVs) imagery and field measurements of tree height (H), diameter at breast height (DBH) and crown radius (R_c) in an artificial Pinus tabuliformis forest. Results demonstrated that the optimal sample size for accurately estimating tree-level biomass with minimal manpower and time costs was 130. Besides, the estimating precision of allometric equations can be enhanced by increasing the number of suitable variables, altering the variables combination, and modifying functional forms. The proposed allometric equation based on H and R_c in this study outperformed common equations in estimating Pinus tabuliformis forest biomass. This equation achieved a coefficient of determination (R²) of 0.72 and a root-mean-square error (RMSE) of 8.56 kg for biomass estimation, owing to its utilization of multivariate analysis and exclusive application of logarithmic transformation to the dependent variable. Moreover, the study revealed that the total biomass of 1490 planted Pinus tabuliformis trees in this study area was 67.3 t. This research offers valuable insights into accurately estimating tree-level forest biomass, which is essential for addressing challenging ecological issues and formulating rational forest management policies.