[Monitoring aboveground biomass in Larix olgensis plantations using bi-temporal unmanned aerial vehicle laser...

This study aimed to accurately monitor plantation biomass change, support the assessment of forest carbon dynamics, and improve forest resource management. We focused on Larix olgensis plantations from three age groups at the Mengjiagang Forest Farm, Heilongjiang Province. By using dual-temporal UAV LiDAR point cloud data, we constructed canopy height models to extract canopy height statistics, canopy cover, and vertical volume occupied by the canopy for both periods. The Weibull function was used to fit the height distribution of the canopy surface, obtaining threshold, scale, and shape parameters. A comparative analysis was conducted on the accuracy and applicability of two biomass change estimation methods: the direct method (modeling biomass change directly using the difference between two periods' features) and the indirect method (predicting biomass for each period separately and calculating the difference). The results showed that the indirect method achieved high fitting accuracy for biomass [R2 =0.94, relative root mean square error (rRMSE)=9.2%], and the indirectly derived biomass change also exhibited high estimation accuracy [Bias=-0.16 Mg·hm^-2, mean absolute error (MAE)=4.78 Mg·hm^-2]. The canopy volume and the Weibull threshold parameter were identified as key predictors. In contrast, the direct method showed lower fitting accuracy for biomass change (R2 =0.60, rRMSE=26.2%) and correspondingly lower prediction accuracy (Bias=0.37 Mg·hm^-2, MAE=4.95 Mg·hm^-2). Predictions based on the indirect method indicated that young forest of the Mengjiagang Forest Farm had the fastest biomass growth (mean stand biomass change rate was 6.48 Mg·hm^-2·a^-1), followed by middle-aged forests (5.29 Mg·hm^-2·a^-1), while near-mature forests showed the slowest growth (3.75 Mg·hm^-2·a^-1). This study validated the effectiveness of bi-temporal UAV LiDAR data in plantation biomass monitoring and offered a technical and methodological reference for forest biomass monitoring and carbon stock assessment.