Détection de situations à risques pour la culture du soja à partir de données satellitaires TRMM et MODIS

Abstract : The soybean cultivation has proven for 10 years to be the first economical activity in the Brazilian state of Mato Grosso. The strong spatio-temporal variability of rainfalls acts upon the agricultural calendar and then can cause important yield losses. For instance, the planting dates that occur before the wet season onset can be linked to potential yield losses. The aim of this study is to detect those situations using satellite data. On one hand, multi-temporal data of vegetation index proposed by the MODIS-TERRA satellite (16-days of temporal resolution and 250 metres of spatial resolution) allow analysing phenological cycles. Then, planting dates are estimated through a threshold-based method. On the other hand, rainfall estimates from TRMM at a 0, 25° spatial resolution allow detecting the wet season onset. The TRMM and MODIS data are then crossed in order to detect situations where the planting dates were anticipated as compared to the wet season onset. It appears than 20% of the soybean planted during the year 2005-2006 corresponded to such a situation. It represented 1 400 000 ha over the entire state of Mato Grosso.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00321444
Contributor : Vincent Dubreuil <>
Submitted on : Sunday, September 14, 2008 - 9:31:00 PM
Last modification on : Friday, July 5, 2019 - 12:42:03 PM

Identifiers

  • HAL Id : hal-00321444, version 1

Citation

Damien Arvor, Vincent Dubreuil, Margareth Simões Penello Meirelles. Détection de situations à risques pour la culture du soja à partir de données satellitaires TRMM et MODIS. XXIe Colloque de l'Association Internationale de Climatologie, 2008, Montpellier, France. pp.99-104. ⟨hal-00321444⟩

Share

Metrics

Record views

311