Very early prediction of wine yield based on satellite data from VEGETATION

• Type: Article in International Scientific Journal
• Date: 2010

Authors

Mario Cunha (Author) FCUP; Andre R S Marcal (Author) FCUP; Lisa Silva (Author) Other

Journal

Title: International Journal of Remote Sensing

Vol. 31 No. 12

Pages: 3125-3142

ISSN: 0143-1161

Publisher: Taylor & Francis

Indexing

ISI Web of Knowledge - 53 Citations

ISI Web of Science

Scopus - 64 Citations

Current Contents

Scientific classification

FOS: Engineering and technology > Environmental engineering

Other information

Authenticus ID: P-003-BEA

DOI: 10.1080/01431160903154382

Abstract (EN): A forecast model for estimating the annual variation in regional wine yield based on remote sensing was developed for the main wine regions of Portugal. Normalized Difference Vegetation Index (NDVI) time-series obtained by the VEGETATION sensor, on board the most recent Satellite Pour l'Observation de la Terre (SPOT) satellite, over the period 1998-2008 were used for four test sites located in the main wine regions of Portugal: Douro (two sites), Vinhos Verdes and Alentejo. The CORINE (Coordination of Information on the Environment) Land Cover maps from 2000 were initially used to select the suitable regional test sites. The NDVI values of the second decade of April of the previous season to harvest were significantly correlated to the wine yield for all studied regions. The relation between the NDVI and grapevine induction and differentiation of the inflorescence primordial or bud fruitfulness during the previous season is discussed. This NDVI measurement can be made about 17 months before harvest and allows us to obtain very early forecasts of potential regional wine yield. Appropriate statistical tests indicated that the wine yield forecast model explains 77-88% of the inter-annual variability in wine yield. The comparison of official wine yield and the adjusted prediction models, based on 36 annual data records for all regions, shows an average spread deviation between 2.9% and 7.1% for the different regions. The dataset provided by the VEGETATION sensor proved to be a valuable tool for vineyard monitoring, mainly for inter-annual comparisons on a regional scale due to their high data acquisition rates and wide availability. The accuracy, very early indication and low-cost of the developed forecast model justify its use by the winery and viticulture industry.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 18