The reflectance spectra of vegetation carry potentially useful information that can be used to determine chemical composition and discriminate between vegetation classes. If compared with analytical methods such as conventional chemical analysis, reflectance measurement provides non-destructive, economic, near real-time data. Presently, results from reflectance measurements performed by spectroradiometry on leaves and branches of Vitis vinifera L. cv. Merlot and Cabernet Sauvignon from two vineyards in south Brazil are reported. The vineyards had different geological histories but were subjected to the same management. The objectives were to detect spectral differences between the vineyards, and to correlate these differences to variations in foliar traits like the chemical composition of vine leaves. To that end, seven vine parcels were selected for reflectance measurements and chemical analyses (of eleven elements) of vine leaves, and correlations between reflectance and chemical composition were looked for. An initial investigation by discriminant analysis applied to reflectance data of leaves and branches and to grape varieties as well allowed for good separation between vineyards and varieties (> 90% accuracy). By further investigating the correlations between leaf chemical composition and reflectance along the wavelength domain covered by the measurements, we found several well-determined wavelengths with Pearson correlation coefficients r > 0.7. Abundances of elements could be modelled up to 94% accuracy. These preliminary results, which have to be validated, suggest that variations in soil properties induce chemical differences in vine leaves that can be detected by reflectance measurements. Applications of this observation include the assessment of the chemical content of vine leaves by spectroradiometry as a fast, low-cost alternative to chemical analytical methods.
Presenting author: Jorge Ducati – Remote Sensing Center, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil
Additionnal authors: Adriane Thum | Universidade do Vale do Rio dos Sinos, São Leopoldo, Brazil
Email: ducati@if.ufrgs.br
Keywords: vineyard geology – chemical abundances – spectroradiometry – multivariate analysis