Su et al. (2026) Annual canopy height mapping in Italy from Landsat and multi-source LiDAR for disturbance and recovery monitoring

Identification

• Journal: Ecological Informatics
• Year: 2026
• Date: 2026-09-09
• Authors: Yang Su, Nikola Besic, Saverio Francini, Xianglin Zhang, Yidi Xu, Giovanni D’Amico, Gherardo Chirici, Martin Schwartz, Ibrahim Fayad, Sarah Brood, Agnès Pellissier‐Tanon, Ke Yu, Haotian Chen, Haoruo Li, Songchao Chen, Alexandre d’Aspremont, Philippe Ciais
• DOI: 10.1016/j.ecoinf.2026.104036

Research Groups

• CNRS & D´epartement d'Informatique, ´Ecole Normale Sup´erieure – PSL
• Laboratoire des Sciences du Climat et de l'Environnement, CEA CNRS UVSQ Orme des Merisiers
• Universit´e Paris-Saclay, AgroParisTech, INRAE, UMR ECOSYS
• Universit´e de Lorraine, G´eodata Paris, IGN, Laboratoire d'Inventaire Forestier (LIF)
• Department of Science and Technology of Agriculture and Environment (DISTAL), University of Bologna
• Department of Agriculture, Food, Environment and Forestry, Universit`a degli Studi di Firenze
• Institute of Crop Sciences, Chinese Academy of Agricultural Sciences / Key Laboratory of Crop Physiology and Ecology
• State Key Laboratory of Efficient Utilization of Arable Land in China，the Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences
• School of Ecological and Environmental Sciences, East China Normal University
• International Institute for Applied Systems Analysis (IIASA)

Short Summary

This study presents an annual 30 m canopy height dataset for Italy spanning 2004–2024, developed by integrating Landsat imagery with multi-source LiDAR observations in a time-series deep learning framework. The dataset provides high predictive accuracy and outperforms existing products.

Objective

• To develop a framework for annual canopy height mapping using Landsat time series, ALS, GEDI, deep learning, and Bayesian fusion.
• To create a harmonized, spatially consistent, and temporally continuous canopy height dataset for Italy from 2004 to 2024.

Study Configuration

• Spatial Scale: National scale (Italy)
• Temporal Scale: Annual, spanning 2004–2024

Methodology and Data

• Models used: U-Net models with Bayesian Model Averaging (BMA) for fusion
• Data sources: Landsat imagery, ALS data, GEDI data, Italian National Forest Inventory (NFI) data, reference data for forest disturbance

Main Results

• High predictive accuracy of the annual canopy height dataset, with a mean absolute error of 3.98 m.
• The dataset outperforms existing products, including a Europe-scale PlanetScope-based dataset and a global Sentinel-based canopy height product.
• The study demonstrates the utility of the dataset for forest dynamics monitoring by deriving a canopy height change-based disturbance product and tracking post-disturbance recovery.

Contributions

• This study provides an effective framework for annual canopy height mapping that supports disturbance tracking, recovery monitoring, and broader assessments of forest structural dynamics.
• The dataset offers a harmonized, spatially consistent, and temporally continuous record of canopy height across Italy from 2004 to 2024.

Funding

• Not specified in the provided text.

Citation

@article{Su2026Annual,
  author = {Su, Yang and Besic, Nikola and Francini, Saverio and Zhang, Xianglin and Xu, Yidi and D’Amico, Giovanni and Chirici, Gherardo and Schwartz, Martin and Fayad, Ibrahim and Brood, Sarah and Pellissier‐Tanon, Agnès and Yu, Ke and Chen, Haotian and Li, Haoruo and Chen, Songchao and d’Aspremont, Alexandre and Ciais, Philippe},
  title = {Annual canopy height mapping in Italy from Landsat and multi-source LiDAR for disturbance and recovery monitoring},
  journal = {Ecological Informatics},
  year = {2026},
  doi = {10.1016/j.ecoinf.2026.104036},
  url = {https://doi.org/10.1016/j.ecoinf.2026.104036}
}