Spectral Discrimination of Grape Varieties and a Search for Terroir Effects Using Remote Sensing

Abstract

Satellite images are used to determine the reflectance dependency on wavelength in different grape varieties (cabernet sauvignon, merlot, pinot noir, and chardonnay). The terroir influence is investigated through a study of vineyards in France, Brazil and Chile. Statistical techniques (ANOVA, cluster and discriminant analysis) are applied. Results indicate that there are consistent spectral features, mainly in the near infrared, which can lead to variety identification. Discriminant functions were derived; these separate grape varieties for the regions studied. Spectral features are affected by terroir effects, since the reflectance spectra showed similarities between regions, specially for cabernet sauvignon; phenological factors, expressed by the NDVI, further contribute to variety differentiation. It is concluded that remote sensing data are effective for terroir and grape variety studies.

Acknowledgements

The authors are grateful to the managers of Chateau Giscours and Chateau Duhart Milon, at Bordeaux region; to Champagne Louis Roederer; to Errazuriz Estate, at Aconcagua Valley and Viña Viu Manent at Colchagua; and to Vinícola Lidio Carraro, at Encruzilhada do Sul, who kindly provided maps and opened their properties to field studies.