Estimation of sunflower yield using multi-spectral satellite data (optical or radar) in a simplified agro-meteorological model.

Summary This paper aims to compare the crop yield retrieval performances, obtained by assimilating the leaf area index derived from multi-temporal satellite signatures (i.e. reflectances and backscattering coefficients) into an agro-meteorological model. The study is based on the Multispectral Crop Monitoring experimental campaign, conducted in 2010 by the CESBIO laboratory. During the agricultural season of sunflower, regular satellite images were quasi-synchronously acquired by 6 sensors (Formosat-2, Spot-4/5, TerraSAR-X, Radarsat-2 and Alos), over a region located in the south west of France. Calibration and validation steps take advantage of the dense network of monitored fields. Among the wide range of the tested image configurations (multi-frequency and multi-polarization), promising results are offered by optical and co-polarized C-band (i.e. HH and VV) data for yield estimate, with correlation superior to 0.74.