Grassland Aboveground Biomass Estimation using Vegetation Indices with Machine Learning Algorithms

As indispensable ecological assets, grasslands play a crucial role in preserving biodiversity, sequestering carbon, and maintaining ecological balance, thereby underscoring their essential contribution to environmental well-being. In grassland ecosystems, aboveground biomass (AGB) serves as a pivotal indicator for ecosystem management and a fundamental component in the terrestrial carbon cycle. Building on this, the integration of remote sensing technology and cloud computing platforms, such as Google Earth Engine (GEE), represents a significant advancement in the extensive and rapid assessment of grassland AGB. This technological evolution provides innovative methods for enhanced environmental monitoring and more sophisticated data analysis.In this study, multiple inversion models were developed based on the relationship between AGB and vegetation indices. Each model was subjected to hyperparameter tuning through Leave-One-Out Cross-Validation (LOOCV) and adjustments in feature selection to enhance their specificity. Subsequently, a thorough evaluation of each model's performance was conducted, establishing a benchmark for rapid AGB inversion in grassland productivity analysis. This methodology not only improves the precision of the models but also offers critical insights for advancing the field.