Mapping Canopy-Level Crop Traits Using Top-of-Atmosphere Sentinel-2 Data in Google Earth Engine

Abstract

To take advantage of the vast amount of remote sensing data, cloud computing platforms such as Google Earth Engine (GEE) open new possibilities to develop crop trait retrieval models applicable to any corner of the world. In the present study, we implemented hybrid models directly in GEE for processing Sentinel-2 (S2) top-of-atmosphere (TOA) reflectance data into crop traits. To achieve this, a training dataset was generated using the leaf-canopy RTM PROSAIL in combination with the atmospheric model 6SV. Gaussian processes regression (GPR) retrieval models were then established for 4 canopy-level crop traits namely: leaf area index, canopy chlorophyll content, canopy water content and canopy dry matter content. Successful reduction of the training dataset by 78% was achieved using the active learning technique Euclidean distance-based diversity (EBD). The EBD-GPR model showed moderate to good performance against in situ data over an independent study site (Grosseto, Italy). Obtained maps compared against ESA Sentinels' Application Platform (SNAP) vegetation estimates showed high consistency of both retrievals. Finally, local and national scale maps were successfully generated in GEE, with additionally providing uncertainties. In summary, the proposed retrieval workflow demonstrates the possibility of routine processing S2 TOA data into crop trait maps at any place on Earth as required for operational agricultural applications.