Combined use of agro-meteorological model and multi-temporal satellite data (optical and radar) for monitoring corn biophysical parameters

Ameline Maël, Fieuzal Remy, Betbeder Julie, Berthoumieu Jean-François, Baup Frederic. 2017. Combined use of agro-meteorological model and multi-temporal satellite data (optical and radar) for monitoring corn biophysical parameters. In : Fifth recent advances in quantitative remote sensing. Sobrino, José A. (ed.). Valencia : Universitat de Valencia, 287-291. ISBN 978-84-9133-201-5 International Symposium on Recent Advances in Quantitative Remote. 5, Valencia, Espagne, 18 Septembre 2017/22 Septembre 2017.

Résumé : This work evaluates the assimilation of Green Area Index (GAI) derived from radar and optical imagery into an agrometeorological model named SAFY-WB (Simple Algorithm For Yield model combined with a Water Balance model) to retrieve corn biophysical parameters. Radar satellite information is provided by the Sentinel-I A (SI-A) mission through two angular normalized orbits allowing a repetitiveness from 12 to 6 days. Optical images are provided by Spot-5-Take-5 and Landsat-8 missions all along the crop cycle. A nonlinear relationship between GAIopt (GAI derived from optical data) and the ratio radar signal (ơVH VV), allows to createa GAlsar (GAI derived from radar data) (R2 = 0. 75). Then, these GAI are assimilated into the model to optimize the simulations by using three configurations: GAIsar, GAIopt or a combination of radar and optical (GAlsar-opt). Results show that in the calibration and validation steps, the GAIsar is mainly suitable for initializing the model during the first crop stages. The GAIopt is suited all along the crop cycle but limitations remains during the first phenological stages, when cloud cover is important over the studied region. The last configuration (GAIsar-opt) uses the benefit of both sensors: GAIsar to substitute GAIopt at the beginning of the crop cycle and GAIopt to retrieve the vegetation dynamic from flowering. Finally, the model, controlled by SAR and optical data, is able to accurately simulate the dynamic of GAI (R2<0.99 and rRMSE< 11%), Dry Stem Mass (DSM), Total Dry Mass (TDM) and yield through Dry Grain Mass (DGM) (R2=0.82).

Auteurs et affiliations

• Ameline Maël, CESBIO (FRA)
• Fieuzal Remy, CESBIO (FRA)
• Betbeder Julie, CIRAD-ES-UPR BSef (FRA)
• Berthoumieu Jean-François
• Baup Frederic, CESBIO (FRA)

Source : Cirad-Agritrop (https://agritrop.cirad.fr/590364/)

Métriques

[ Page générée et mise en cache le 2019-10-03 ]