Thesis (MScAgric)--Stellenbosch University, 2003. / ENGLISH ABSTRACT: In the past it was very difficult to visualise the extent and distribution of variability in growth
vigour within vineyards. The advancement of remote sensing technology has changed this
however, establishing new methods to assess and manage variability. Even though the
causes and effects of within-vineyard variability in growth vigour are complex, new
technologies offer better ways of monitoring, researching and managing these factors. In
addition to the possible benefits of aerial or satellite remote sensing, new methods of
mapping soil spatial variability as well as advances in georeferencing technologies supply
precision tools to both researchers and producers. The scientific advancement of the
technology, however, is currently the most important aspect needed. This is crucial to
answer and explore fundamental questions regarding the use of the technology and the
interpretation of results within the framework of the plant's observed reaction. Only then
can the possible applications in vineyard management be optimised to address the
management problems of extensive within-vineyard variation in growth vigour. The choice
will always be to use the technology to manage the observed variability in order to limit the
negative impacts of a heterogeneous harvest, or to identify the variability and its causes
for the implementation of management practices aimed at a more homogenous vineyard
and harvest. Whatever the case, extensive research is needed to provide tight correlations
between information gathered with new technologies to assess variability and plant status,
such as multi- and hyperspectral analysis, and ground-truthed results in the vineyard. Only
then will it become evident which methods and analyses would be useful in the drive
towards in-depth analysis and management of vineyards within the concept of precision
viticulture and its derived advantages.
With this in mind, the aim of this study was to establish an experimental model to use
remote sensing technologies to identify and classify within vineyard variability with a
subsequent analysis of the causes of variability and the effects on the plants. The targeted
experimental model was a vineyard with highly heterogeneous above-ground growth. An
aerial photograph of the vineyard was studied and manipulated to yield image pixel values
used to quantify the degree of variability for different plots, which were chosen according
to different plot layouts. Soil conditions were assessed on both a global and plot level, with
extremely high pH and low resistance values in the soil in combination with erratic soil
preparation practices found to be the main cause of variability. Soil physicochemical
condition was also assessed during a soil profile pit study. Significant differences were
found between several soil-related parameters measured for the higher and lower vigour
levels and a strong correlation was also found between the resistance of a saturated soil
paste and the image pixel values.
Vegetative measurements also yielded highly significant differences between the
vigour levels and confirmed the suitability of the vineyard to study within-vineyard variability. Some of these measurements were also strongly correlated with soil conditions
as well as image pixel values. Trunk circumference proved to be an excellent measure for
the level of variability, being linked strongly to canopy characteristics, soil conditions as
well as the image pixel values. Leaf water potential measurements also yielded significant
differences between the vigour levels.
Harvest data and wine analyses showed the effect that vigour differences can have on
grape composition and wine quality, even though the differences found here were much
less than expected. Even though no clear preference was shown between the wines made
from the different vigour levels, the lower vigour wine was considered fruitier. The overall
quality of both experimental wines was however very high, considering that experimental
winemaking techniques has been used.
Hyperspectral measurements also confirmed differences between the vigour levels
through a narrow-band NOVI (normalised difference vegetation index). It was also possible
to show differences in certain biochemical compounds between the vigour levels on both a
leaf and canopy level. Wavelength regions corresponding to carotenoid, chlorophyll a and
chlorophyll b showed different spectral reactions in the leaves of more stressed (lower
vigour) canopies, indicating possibilities for further studies.
This study and its results is the first of its kind in the South African wine industry and
paves the way for more focussed and in-depth analyses of the use of specifically multiand
hyperspectral data to accurately assess within-vineyard vigour variability and the
management thereof to yield optimum quality grapes for a specific wine target. Moreover,
the approach adopted in this study is also echoed in other international research programs
in prominent wine countries. The availability of scientific research regarding the optimal
use and limitations of these technologies has the potential to revolutionise production
management practices in the next few years in the viticultural industry. / AFRIKAANSE OPSOMMING: In die verlede was dit baie moeilik om die omvang en verspreiding van groeikragvariasie
binne 'n wingerd te visualiseer. Die vordering gemaak op die gebied van
afstandswaarneming-tegnologie het egter nuwe metodes beskikbaar gestel waardeur
hierdie variasie in wingerde gemonitor en bestuur kan word. Selfs al is die oorsake en
invloede van binne-wingerd-groeikragvariasie kompleks, verskaf nuwe tegnologieë
verbeterde metodes om hierdie variasie te monitor, te bestuur en na te vors. Saam met die
moontlike voordele wat lugfoto's en satelliet-afstandswaarneming teweegbring, verskaf
nuwe metodes om ruimtelike variasie in grondfaktore te karteer, asook vordering in geoverwysingstegnologie,
presisiehulpmiddels aan produsente én navorsers. Die
wetenskaplike vordering van dié tegnologie is tans van groot belang. Die belang daarvan
is om fundamentele vrae te ondersoek en te beantwoord rakende die gebruik van die
tegnologie en die interpretasie van resultate binne die raamwerk van die waargeneemde
reaksie in die plant. Dit sal die weg baan vir optimale toepassing van die tegnologie in
wingerdbestuur om sodoende die bestuursprobleme wat deur binne-wingerdgroeikragvariasie
teweeggebring word, aan te spreek. Die voorkeurkeuse is om dié
tegnologie aan te wend om hierdie variasie te bestuur sodat die negatiewe impak van In
heterogene oes teengewerk kan word, of om die variasie te identifiseer vir die
implementering van bestuurspraktyke gemik op die skep van 'n meer homogene wingerd
en oes. Dit is noodsaaklik dat uitgebreide navorsing gedoen word om noue verwantskappe
vas te stel tussen inligting wat ingewin is met behulp van nuwe tegnologieë wat die
variasie in plantstatus monitor, soos multi- en hiperspektrale analise, en inligting wat op
grondvlak ingewin is. Hieruit sal dit duidelik wees watter metodes en analises die
doeltreffendste is vir in-diepte analises en die bestuur van wingerde binne die konsep van
presisie-wingerdkunde.
Met inagneming van hierdie aspekte, was die doel van hierdie studie om 'n
eksperimentele model daar te stel waardeur afstandswaarneming-tegnologie gebruik kan
word om variasie binne wingerde te identifiseer en te klassifiseer deur analises van die
oorsake van hierdie variasie en invloede op die plant. Die geteikende eksperimentele
model was 'n wingerd met hoogs heterogene bogrondse groei. 'n Lugfoto van die wingerd
is bestudeer en gemanipuleer om pixelwaardes te verskaf wat die graad van variasie vir
verskillende eksperimentele plotte, wat aan die hand van verskillende plotuitlegte gekies
is, te kwantifiseer. Grondtoestande is bestudeer op 'n globale én plotvlak, met uiters hoë
pH en lae weerstande in kombinasie met verkeerde grondvoorbereidingspraktyke, wat
geïdentifiseer is as die hoofoorsake vir die hoë vlakke van variasie. Grondfisiese en
-chemiese toestand is ook tydens profielgatstudies bestudeer. Betekenisvolle verskille is
gevind tussen verskeie grondverwante parameters gemeet vir plotte met onderskeidelik laer en hoër groeikrag, en 'n sterk verwantskap is gevind tussen grondweerstand en
pixelwaardes, soos vanaf die lugfoto bepaal.
Vegetatiewe metings het ook betekenisvolle verskille tussen die hoër en laer
groeikragvlakke opgelewer, wat die geskiktheid van die wingerd vir die studie van binnewingerdvariasie
in groeikrag bevestig het. Van hierdie metings was ook nou verwant aan
grondtoestande, asook beeldpixelwaardes. Stamomtrek was 'n uitstekende maatstaf vir
die vlakke van variasie, aangesien dit nou verwant was aan lowertoestande,
grondtoestande, asook beeldpixelwaardes. Blaarwaterpotensiaal-metings het ook
betekenisvolle verskille tussen die hoër en laer groeikragvlakke opgelewer.
Oesdata en wynanalise het die uitwerking van groekragverskille op druifsamestelling
en wynkwaliteit uitgewys, selfs al was die verskille wat gevind is minder as wat verwag is.
Hoewel geen duidelike voorkeur tussen die wyne afkomstig van verskillende
groeikragvlakke uitgewys kon word nie, was die wyn wat van die laer-groeikrag stokke
gemaak was, meer vrugtig. Die algemene kwaliteit van beide wyne was egter baie hoog as
in ag geneem word dat eksperimentele wynmaakprosedures gevolg is.
Hiperspektrale metings het ook die verskille tussen groeikragvlakke bevestig deur 'n
nou-bandwydte NDVI ("normalised difference vegetation index"). Dit was ook moontlik om
verskille in sekere biochemiese komponente tussen die groeikragvlakke op 'n blaar- én
lowervlak uit te wys. Golflengte-areas ooreenstemmend met karotenoïed, chlorofil a en
chlorofil b het verskillende spektrale reaksies in die blare met hoër stresvlak (laer
groeikrag) lowers ten toon gestel. Dit het moontlikhede vir verdere navorsing uitgewys.
Hierdie studie en die resultate wat verkry is, is die eerste van sy soort in die Suid-
Afrikaanse wynbedryf. Dit baan die weg vir meer gefokusde en in-diepte analise van die
gebruik van spesifiek multispektrale en hiperspektrale data om binne-wingerdgroekragvariasie
akkuraat te monitor en te bestuur met die oog op optimum wynkwaliteit
vir 'n spesifieke produkdoelwit. Die aanslag van hierdie navorsing is ook sigbaar in ander
prominente wynproduserende lande. Beskikbaarheid van wetenskaplike navorsing
rakende die optimale gebruik en tekortkominge van hierdie tegnologieë het die potensiaal
om produksiebestuurspraktyke in die wingerdbedryf in die komende jare te
revolusionaliseer.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:sun/oai:scholar.sun.ac.za:10019.1/53304 |
| Date | 12 1900 |
| Creators | Strever, Albert E. |
| Contributors | Vivier, M. A., Archer, E., Stellenbosch University. Faculty of AgriSciences. Dept. of Viticulture and Oenology. Institute for Wine Biotechnology. |
| Publisher | Stellenbosch : Stellenbosch University |
| Source Sets | South African National ETD Portal |
| Language | en_ZA |
| Detected Language | English |
| Type | Thesis |
| Format | 147 p. : ill. |
| Rights | Stellenbosch University |
Page generated in 0.0036 seconds