A User-Friendly Google Earth Engine Web Application for Vegetation Trait Mapping and Time Series Analysis From Sentinel-2 and Sentinel-3 Imagery

A Google Earth Engine (GEE) web application is introduced for automated mapping of a suite of quantitative vegetation traits (e.g., LAI, fAPAR, GPP, chlorophyll/nitrogen/water content) from Sentinel-2 and Sentinel-3 imagery, covering both surface reflectance and top-of-atmosphere radiance. The system operationalizes validated Gaussian Process Regression (GPR) models based on radiative transfer simulations within a fully cloud-based workflow, enabling on-demand generation of vegetation trait maps at 20–300 m resolution without requiring coding expertise. Key innovations include: (1) dynamic model loading for 30 GPR models enabling estimation of 11 vegetation traits, (2) automated masking/compositing for trait mapping, and (3) Savitzky-Golay smoothing for noise-reduced temporal analysis. The application demonstrates the potential of GEE to bridge advanced machine-learning-based retrieval methods and practical, operational Earth observation use.