Spatio-temporal leaf area index mapping in open-canopy forests from Sentinel-2 imagery: evaluating hybrid and empirical approaches

This repository contains the complete dataset, field measurements, and source codes supporting our research on Leaf Area Index (LAI) retrieval in Open-Canopy Forests (OCFs). This data is shared to ensure full transparency, repeatability, and open-access compliance for our manuscript submitted to the journal Ecological Informatics. The repository is organized into three main directories:01_Field_Data: Contains in-situ LAI measurements collected from two independent study sites (SS1 in 2023 and SS2 in 2016), used for model training and spatio-temporal validation.02_Empirical_Approach: Includes the extracted Sentinel-2 spectral/structural data and the source codes (scripts) used to train and evaluate standard Machine Learning Regression Algorithms (MLRAs).03_Hybrid_Approach: Contains the physical modeling framework codes (PROSPECT-D + 4SAIL2) for generating simulation Look-Up Tables (LUTs) and executing the hybrid LAI inversion based on MLRAs (SVR). These materials allow researchers to fully reproduce our comparative analysis between standard empirical models and structurally-aware hybrid inversion frameworks.