Uso de medidas espectrales para la caracterizaci?n del vi?edo

Trabajo Fin de M?ster. M?ster en Ingenier?a Agron?mica. Curso 2013-2014.

[EN]Field radiometry based on high spectral capacity of spectroradiometers is an alternative to the high-cost direct or destructive measurements for monitoring agricultural vegetation. This technique has not scale restrictions (i.e., remote sensing) and, unlike the laboratory experiments, it reflects the actual conditions of the plants and their environment. Parameters related with phenology, status and physiognomy either at leaf, plant or canopy level can potentially be estimated and monitored using measurements of the reflected energy in the visible and near infrared spectrum range. In this work, hyperspectral indices derived from combination of reflectance measurements were proposed for estimating biophysical parameters of vines plants. A vineyard of 100 has of Vitis vinifera L. cv. Tempranillo was studied in the Castilla y Le?n region of Spain (41.18? N, 5.21? W, 717 m a.s.l.). The vines are trellised to a vertical shoot position and equispaced 1.5 m. Seventeen sampling points were selected within the vineyard. Two vines at each side of these sampling points were monitored during the complete growing season of 2013. The biophysical parameters that have been controlled in these 34 vines were Leaf Area Index (LAI), chlorophyll relative content, and vegetation water content (VWC), all of them estimated at leaf and plant level. LAI and VWC were directly estimated trough destructive measurements, and chlorophyll was measured by means of a Minolta SPAD-502. The spectral measurements were taken with a spectroradiometer Ocean Optics USB4000, with a spectral range of 500-1100 nm and spectral resolution of 0.21 nm. Both direct and indirect measurements were acquired approximately every 20-30 days at noon (5 measurements in total) along the vineyard growing season. Several hyperspectral indices have been used in this work, related with canopy structure and vigour (Normalized Difference Vegetation Index, NDVI; Soil-Adjusted Vegetation Index, SAVI), pigments (Photochemical Reflectance Index, PRI; Transformed Chlorophyll Absorption in Reflectance Index, TCARI; Chlorophyll Normalized Difference Index, CNDI; Greennes index; and two Carotenoids indices, CAR), and water status (Water Index, WI). Correlation among LAI, VWC and chlorophyll measurements with hyperspectral indices was established and evaluated. Soil moisture observations were also included in the comparisons. Regarding the coefficient of correlation (R) between vine parameters and hyperspectral indices, non significant correlations were obtained with the NDVI as well as poor results in terms of R from the carotenoids and the Greeness indices. Even though, good results were found for SAVI, TCARI, PRI and CNDI indices for LAI, VWC and chlorophyll characterization, with a high number of significant correlations (R>0.60), specially for the SAVI. Detailed results will be presented at the conference. The results suggested that field radiometry might provide valuable information to predict growing vine attributes. Hyperspectral indices SAVI, TCARI, PRI and CNDI