Answers are grounded in this paper and cite their sources.
Leaf area index (LAI) is a key parameter for characterizing vegetation canopies and an essential input for land surface models. Statistical regression models of vegetation indices (VIs) derived from spectral reflectance are widely used for LAI retrieval. However, traditional VIs like the normalized difference vegetation index (NDVI) experience saturation issues in dense vegetation. To address the limitations of conventional VIs, Red-Edge (RE) VIs have been developed. While satellites like Sentinel-2 include RE bands, their short time series limit long-term monitoring. In contrast, Landsat lacks RE bands but offers decades of data, prompting research into machine learning-based RE band reconstruction. This study proposes a robust methodology for RE VIs reconstruction based on random forest model, which establishes relationships between Landsat 8 reflectance and Sentinel-2 VIs. The method was implemented at Ground-Based Observations for Validation (GBOV) sites, and the results show that reconstructed RE VIs, with their distinct sensitivity to dense vegetation structure, offer advantages in estimating high-value LAI (Rel-Bias = -0.07 when LAI>3), mitigating the saturation issue associated with NDVI (Rel-Bias = -0.10 when LAI>3). In addition, in comparison to NDVI, the reconstructed RE VIs exhibit better temporal patterns, particularly for land cover types with high biomass. The results confirmed the viability of the RE VIs reconstruction from Landsat 8 reflectance, which achieves satisfactory accuracy and captures the temporal vegetation patterns. These findings underscore the potential to leverage the extensive historical Landsat data, thereby enhancing long-term vegetation monitoring capabilities.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: