Assimilation of a biophysical parameter estimated by remote sensing using SPOT 4 and 5 data into a sugarcane yield forecasting model

Bappel Eric, Bégué Agnès, Martiné Jean-François, Pellegrino Anne, Siegmund Bertrand. 2005. Assimilation of a biophysical parameter estimated by remote sensing using SPOT 4 and 5 data into a sugarcane yield forecasting model. In : Proceedings of the XXV ISSCT Congress, 30 January - 4 February 2005, Guatemala. Ed. Hogarth. Guatemala : ATAGUA, 4 p. ISBN 99922-2-211-5 ISSCT Congress. 25, Guatemala, Guatemala, 30 Janvier 2005/4 Février 2005.

Résumé : In the context of the SUCRETTE project, (Système de sUivi de Ia Canne à sucRE par TélédéTEction) CIRAD (Centre International de Recherche Agronomique pour le Développement) has initiated a project to monitor growth of sugarcane in Reunion Island (55°60' E longitude, 20°90' S latitude). Ground measurements (Leaf Area Index) using a PCA Licor 2000 were taken from June 2002 to December 2002, coupled with SPOT 4&5 (Satellite Pour l'Observation de la Terre) images acquired during the same period. Yield estimates could be improved by combining remotely sensed data with growth models. The objective of this study was to determine a relationship between leaf area index (LAI) and the Normalized Difference Vegetation Index (NDVI) generated from SPOT 4&5 images, and to establish whether there is a benefit in using remotely estimated LAI instead of simulated LAI to estimate yield. A strong relationship between LAI and NDVI was obtained using an exponential function (R2=0.86). A time series of 12 Spot images was used to produce NDVI profiles for each field. Gaps between dates were filled by fitting an empirical curve. LAI was estimated at a daily time step from the NDVI curve. The growth model Mosicas was forced using the estimated LAI. Actual yields were compared with yields simulated with and without LAI forcing. The coefficient of determination results indicate that R2 between observed and simulated yields is equal to 0.42 and 0.66 for simulation without forcing and for simulation with LAI forcing, respectively. The root mean square error of 19.3 t/ha without LAI forcing decreased to 12.8 t/ha using LAI forcing.

Mots-clés Agrovoc : canne à sucre

Classification Agris : U30 - Méthodes de recherche
F01 - Culture des plantes

Auteurs et affiliations

• Bappel Eric
• Bégué Agnès, CIRAD-AMIS-UPR Spatialisation (FRA)
• Martiné Jean-François, CIRAD-CA-UPR Systèmes canniers (REU)
• Pellegrino Anne
• Siegmund Bertrand, CIRAD-CA-UPR Systèmes canniers (REU)

Source : Cirad - Agritrop (https://agritrop.cirad.fr/526424/)

Métriques

[ Page générée et mise en cache le 2024-04-02 ]