The Aboveground Biomass Estimation of the Grain for Green Program Stands Using UAV-LiDAR and Sentinel-2 Data

Area of Science:

• Forestry and Ecological Remote Sensing

Background:

• Aboveground biomass (AGB) is vital for assessing the Grain for Green Program (GGP) success, forest health, and carbon sequestration.
• Limited research exists on AGB estimation in GGP stands, particularly those dominated by sparse forests.

Purpose of the Study:

• To develop optimal tree height-diameter at breast height (DBH) growth models for key tree species.
• To construct a high-quality AGB sample dataset using airborne LiDAR and tree species data.
• To estimate GGP stand AGB using Sentinel-2 data and a gradient boosting decision tree (GBDT) algorithm.

Main Methods:

• Development of species-specific tree height-DBH models.
• Integration of airborne LiDAR and tree species data for AGB sample dataset creation.
• Application of Sentinel-2 data and GBDT algorithm for AGB estimation.

Main Results:

• The AGB sample dataset showed high consistency with field data (R² = 0.89).
• The GBDT model achieved high accuracy in AGB estimation (R² = 0.96, RMSE = 560 g/m²).
• Key variables for AGB estimation included tasseled cap greenness (TCG), RENDVI, and VDVI, highlighting the importance of vegetation indices.

Conclusions:

• The study presents a practical and accurate method for estimating AGB in GGP stands.
• The findings support sustainable forest management and ecological conservation by providing reliable biomass data.
• Vegetation indices derived from remote sensing data are crucial for accurate AGB estimation in sparse forest stands.