Qi et al. (2025) Retrieving forest LAI from Landsat via 3D look-up table generated by realistic LiDAR scenes

Identification

• Journal: Remote Sensing of Environment
• Year: 2025
• Date: 2025-12-18
• Authors: Jianbo Qi, Siying He, Xun Zhao, Su Ye, Tianjia Chu, Zhexiu Yu, Simei Lin, Huaguo Huang
• DOI: 10.1016/j.rse.2025.115204

Research Groups

• Advanced Interdisciplinary Institute of Satellite Applications, Faculty of Geographical Science, Beijing Normal University, Beijing 100875, China
• State Key Laboratory of Soil Pollution Control and Safety, Zhejiang University, Hangzhou 310058, China
• Zhejiang Key Laboratory of Agricultural Remote Sensing and Information Technology, College of Environmental and Resource Sciences, Zhejiang University, Hangzhou 310058, China
• College of Forestry, Southwest Forestry University, Kunming 650224, China
• School of Soil and Water Conservation, Southwest Forestry University, Kunming 650224, China
• College of Forestry, Fujian Agriculture and Forestry University, Fuzhou 350002, China
• University Key Lab for Geomatics Technology & Optimize Resource Utilization in Fujian Province, Fuzhou 350002, China
• State Key Laboratory of Remote Sensing and Digital Earth, Faculty of Geographical Science, Beijing Normal University, Beijing 100875, China
• State Forestry and Grassland Administration Key Laboratory of Forest Resources and Environmental Management, Beijing Forestry University, Beijing 100083, China

Short Summary

This study developed a novel 3D Look-Up Table (3D-LUT) approach for retrieving forest Leaf Area Index (LAI) from Landsat imagery by reconstructing realistic 3D forest scenes using LiDAR data to parameterize a 3D Radiative Transfer Model, achieving high accuracy across various forest types.

Objective

• To develop and validate a novel 3D Look-Up Table (3D-LUT) approach for retrieving forest Leaf Area Index (LAI) from Landsat imagery, accounting for forest heterogeneity through LiDAR-based scene reconstructions to parameterize a 3D Radiative Transfer Model (RTM).

Study Configuration

• Spatial Scale: Regional-scale forest LAI retrieval, validated across 16 National Ecological Observatory Network (NEON) sites and 8 Integrated Carbon Observation System (ICOS) sites, representing different forest types.
• Temporal Scale: Not explicitly defined, but implied multi-year observations from Landsat imagery and continuous monitoring sites (NEON, ICOS).

Methodology and Data

• Models used:
  • Large-scalE remote Sensing data and image Simulation (LESS) (3D RTM)
  • PATH_RT (analytical RTM based on 3D path-length distribution and spectral invariant theory)
  • PROSAIL model (for comparison, e.g., Simplified Level-2 Prototype Processor (SL2P))
• Data sources:
  • Landsat imagery (satellite)
  • Airborne LiDAR data (for 3D scene reconstruction)
  • Field observations from 16 NEON sites and 8 ICOS sites (in-situ validation)
  • High-resolution Global LAnd Surface Satellite (Hi-GLASS) LAI product (for intercomparison)
  • MODIS LAI product (for intercomparison)

Main Results

• The proposed 3D-LUT algorithm achieved high-accuracy LAI retrieval across four forest types (Deciduous Broadleaf Forest, Deciduous Needleleaf Forest, Evergreen Broadleaf Forest, and Evergreen Needleleaf Forest).
• Validation against in situ data showed Root Mean Square Error (RMSE) ranging from 0.93 to 1.20 m²/m² and Mean Absolute Error (MAE) from 0.73 to 1.00 m²/m².
• Intercomparison revealed that PROSAIL-based algorithms (e.g., SL2P) tend to underestimate forest LAI.
• The proposed algorithm showed strong overall agreement with the Hi-GLASS LAI product and MODIS LAI product, supporting its reliability for regional-scale forest LAI retrieval.

Contributions

• Development of a novel 3D Look-Up Table (3D-LUT) approach for forest LAI retrieval from Landsat, which explicitly accounts for forest heterogeneity.
• Integration of realistic 3D forest scene reconstructions from airborne LiDAR data to parameterize a 3D RTM, moving beyond idealized homogeneous layers or simple geometric objects.
• Coupling of the LESS 3D RTM with the efficient PATH_RT analytical model for building type-specific LAI LUTs.
• Advancement in the application of LiDAR in quantitative remote sensing retrieval by generating simulated datasets from realistically reconstructed 3D forest structures.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Qi2025Retrieving,
  author = {Qi, Jianbo and He, Siying and Zhao, Xun and Ye, Su and Chu, Tianjia and Yu, Zhexiu and Lin, Simei and Huang, Huaguo},
  title = {Retrieving forest LAI from Landsat via 3D look-up table generated by realistic LiDAR scenes},
  journal = {Remote Sensing of Environment},
  year = {2025},
  doi = {10.1016/j.rse.2025.115204},
  url = {https://doi.org/10.1016/j.rse.2025.115204}
}