Assessment of LAI derived from HLS Landsat Sentinel-2, S2 THEIA, and SNAP for crops in Bekaa (Lebanon)

Abstract

Background: Leaf area index (LAI) is an essential indicator of crop development and growth. Proper satellite-based LAI estimates at the farm-level often require near-daily imagery at medium to high spatial resolution. The combination of data from different ongoing satellite missions, Sentinel 2 (ESA) and Landsat 8 (NASA), provides this opportunity. Objectives: In this study, we evaluated the leaf area index generated from three products, namely: the harmonized surface reflectance produced by NASA, SNAP biophysical model, and L2A THEIA’s product from Sentinel-2 for the agricultural scheme in Bekaa (Lebanon). Methods: For this purpose, we used a broad set of in-field LAI measurements collected in a wide variety of canopy structures during the 2018 and 2019 growing seasons. The dynamics of LAI and crop height were monitored during the 2019 growing season. We further assessed the validity of existing LAI models and evaluated the relationship between the studied vegetation indices and the ground measured LAI. Also, crop-specific height – LAI and above-ground biomass– LAI equations were generated. Results: Results show when comparing the measured LAI to the LAI Models existing in Literature that LAI models, which were derived from EVI2 statistically performed better than other models for the combined crops. LAI derived from the artificial neural network through ESA’s SNAP biophysical processor is underestimated. Additionally, the red-edge bands used in the S2 SeLI LAI algorithm offers an improved LAI crop biophysical parameter retrieval with low errors following the HLS LAI Models. Also, our findings show that the LAI-VIs relationship is crop-specific. Among the examined indices, EVI2 outperformed other indices for the crops combined (R2:0.6, RMSE: 1.00, and p-value <0.0001), thus, EVI2 derived from the HLS product can be identified as a best suited for a unified algorithm.