Top Banner
1 Applying Precision Agriculture Methodology to the Viticulture of Douro Demarcated Region Fernando A. Santos (*) (*) Universidade de Trás-os-Montes e Alto Douro (UTAD). Rural Engineering Department. [email protected] Abstract The INTERREG - COREA project, entitle “The News Information and Communication Technologies (NTIC) used in Atlantic Arc Region” (*) had the purpose of applying Precision Agriculture techniques to the viticulture of Douro Region, in order to analyse data variability between and within parcels and between the installations forms (terraces and slope vineyards). In this study several environmental data were determined at various geo-referenced sites, namely atmospheric, plant and soil (at two levels) data, and yield, must and wine parameters, being each one “divided” in three levels and related to the wine sensorial characteristics in order to get three different quality wines. Using ANOVA various possible scenarios were envisage. In some cases parameters were significantly different within and between the parcels, in other cases no significant differences were detected under any circumstance. In the vast majority of cases however, the parameters were different when comparing the parcels but not inside them. Considering the high number of parameters determined for each group of data a Principal Component Analysis (PCA) was performed in order to select the most important, according the interpretation of the first factor. (*) The whole report can be consulted in: www.utad.pt/~fsantos/pub-fas/COREA-RelFinalIng.pdf
19

Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

Apr 21, 2019

Download

Documents

dangkhanh
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

1

Applying Precision Agriculture Methodology to the Viticulture of Douro

Demarcated Region

Fernando A. Santos (*)

(*) Universidade de Trás-os-Montes e Alto Douro (UTAD). Rural Engineering

Department. [email protected]

Abstract

The INTERREG - COREA project, entitle “The News Information and Communication

Technologies (NTIC) used in Atlantic Arc Region” (*) had the purpose of applying

Precision Agriculture techniques to the viticulture of Douro Region, in order to analyse

data variability between and within parcels and between the installations forms (terraces

and slope vineyards).

In this study several environmental data were determined at various geo-referenced

sites, namely atmospheric, plant and soil (at two levels) data, and yield, must and wine

parameters, being each one “divided” in three levels and related to the wine sensorial

characteristics in order to get three different quality wines.

Using ANOVA various possible scenarios were envisage. In some cases parameters

were significantly different within and between the parcels, in other cases no significant

differences were detected under any circumstance. In the vast majority of cases

however, the parameters were different when comparing the parcels but not inside them.

Considering the high number of parameters determined for each group of data a

Principal Component Analysis (PCA) was performed in order to select the most

important, according the interpretation of the first factor.

(*) The whole report can be consulted in:

www.utad.pt/~fsantos/pub-fas/COREA-RelFinalIng.pdf

Page 2: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

2

Keywords: precision viticulture, vineyard, wine, enology

1. Introduction

The project “The News Information and Communication Technologies (NTIC) used in

Atlantic Arc Region Agricultural Services” is a part of an INTERREG-COREA study,

and is composed of Portuguese, Spanish and French research teams. The project had

two parts, one named “Axe I” pertaining to the identification and spreading of national

sites with relevant information for NTIC and, the second one, named “Axe II”, related

to experimental work on precision agriculture. The Portuguese team was focused on

precision viticulture, implemented at the S. Bárbara Experimental Station, which is

located nearby the Pinhão town (Douro Demarcated Region), property of Official

Agricultural Services.

This experimental work included the selection and identification (Figure 1) of several

geo-referenced sites within each parcel where environmental, plant and soil parameters

were determined.

Figure 1

For the yield, must and wine data characterisation each parcel was “divided” in three

blocks being the parameters determined for each one; in the terraced vineyards the three

blocks were discriminated in the sideslope according the topographic contours and, in

the slope vineyards, according the same level; these blocks are the “units” to

performance the final vineyard variability management.

The experimental data analysis and its comparison with the wines sensorial data,

attributed by a wine tasting panel (Alvelos et al., 2000) allow the drivers identification

more relevant to the wine quality. The knowledge on parameters variation within and

Page 3: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

3

between parcels can be used to do a differentiated vintage, in order to get different

quality wines (Krstic & Smart, 2002).

2. Material and methods

2.1. Parcels selection and its characterisation

The identification and characterisation of selected parcels (Table 1) were conceived to

guarantee the same lineage (Aragonês - Tinta Roriz), no matter the parcels were to be

installed in terraces, with one and two rows, or slope vineyards, at different altitudes

and exposition. Variability in the parcels’ settings had as main purpose to emphasize the

differences within and between parcels (Krstic & Smart, 2002).

Table 1

2.2. Environmental data characterisation

The environmental data were determined several times during the plants’ vegetative

cycle and included the air temperature and humidity, as well as, the plants and soil

temperatures.

2.3. Plants data characterisation

The plant characterization, done after the sprout load correction (Escudero & Zaballa,

2000), in order to have the same number of grape flowers in all plants consisted in the

determination of the photosynthetic activity (SPAD), leaves dry weight and area,

macronutrients concentration (nitrogen, phosphorous and potassium) and micronutrients

(calcium, magnesium, boron, iron, copper, zinc and manganese) in leaves and pruning

weight (Bramley, 2001). .

The photosynthetic activity determination, done several times during the plants’ life

cycle, was carried out in leaves gathered at the first hours of the day. For this

Page 4: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

4

determination, leaves were picked up from the plants’ grape region located close to the

geo-referenced sites. The SPAD determination along the plants’ life cycle, had the

purpose of knowing the nutritional state of plants, namely the nitrogen content.

Leaves used in SPAD determinations were used to determine leaf area and dry weight.

The leaves picked up in 2005 during the grape veraison were grinded and analysed in

the UTAD Soil Laboratory to determine their chemical composition, namely nitrogen,

phosphorous and potassium concentrations. In 2006, two determinations were done, one

during the flowering and the other in veraison (Bramley, 2001); in this year

micronutrients such as (calcium, magnesium, boron, iron, copper, zinc and manganese)

were also determined.

To get the pruning wood average weight per plant (Figure 2), material was collected

from plants located close to the geo-referenced site; the pruning wood is a good

indicator of yield (Krstic & Smart, 2002).

Figure 2

2.4. Soil data characterisation

For soil surveying, samples were taken next to the geo-referenced sites at of 0 - 20 cm

and 20 - 40 cm depths intervals (Figure 3). Samples collected at the same depth, from

three geo-referenced close sites (stations), were joined getting so nine samples per

parcel, accomplishing so 72 samples in all parcels (4 parcels x 9 stations x 2 depths).

Figure 3

The data determined at the UTAD Soil Laboratory refers to soil texture, pH in H2O and

KCl, organic matter (OM), assimilable phosphorous (P2O5), assimilable potassium

(K2O), calcium (Ca), magnesium (Mg), potassium (K), nitrogen (Na), extracted boron

in boiling water (B), exchangeable acidity (AT), sum of exchangeable bases (SBT),

Page 5: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

5

cation exchangeable capacity (CTCe) and base saturation (GSBe). These analysis were

done by the Egner-Riehm method of extraction, using an extraction agent solution of

lactic acid, ammonium acetate and acetic acid, tampon at pH of ± 3.5.

2.5. Grapes, must and wine data

These data were measured in samples resulting from the joining of the three close

stations (blocks), of each parcel from which resulted 12 samples (4 parcels x 3 blocks).

The vintage date was decided based on the evolution of grapes ripening.

2.5.1. Grapes weight and composition

To determine the grapes weight, about 20 berries were collected along the fruit ripening

(from middle July until the vintage, in the beginning of September), in each geo-

referenced site, joining them in the three blocks of each parcel.

The data determinations of grapes composition during the vegetative cycle were the

probable alcohol, sugar content, total acidity and pH (Rizzon & Miele, 2003). Some

grapes were frozen to be used later for sugar analysis content, pH, total acidity, total

phenols and total anthocyanins.

2.5.2. Must characterization

The characterisation of must used the last picked up grapes. In each microvinification,

at least 30 kg were used, which means that ± 100 kg of grapes were necessary to

characterize the must of each parcel. Must characterization was based on probable

alcohol, total acidity and pH determinations (Rizzon & Miele, 2003).

2.5.3. Wine analytical parameters

The musts microvinifications took place according to usual techniques of red wines

vinifications with long maceration, to the extraction of aroma and phenols. Once the

alcoholic fermentation ended, controlled malolactic fermentation were started with the

Page 6: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

6

addition of lactic bacteria and, when finished, the wines were microbiologically

stabilized by drawing off and sulphurous addiction, being transferred to a refrigerator

chamber at about 4ºC, for a period of nearly 20 days, until the tartaric stabilization had

occurred, after what the wines were bottled (Silva, 1994).

Analytical determinations made during the microvinification were the alcohol content,

volume mass, non reducer dry extract, reducer sugar, total dry extract (non reducer dry

extract + reducer sugar), pH, volatile acidity, fix acidity, total acidity (volatile acidity +

fix acidity), total phenols, colour intensity, character, ashes, ashes alkalinity, ashes

inorganic phosphates and anthocyanins (Rizzon & Miele, 2003).

2.6. Wine sensorial tests

The wines resulting from the twelve microvinifications were subject to visual and

tasting tests, concerned with colour intensity, aroma, including red fruits and floral

aroma, body, astringency and total acidity (Andrade et al., 2006).

A classification was attributed to the several characteristics, varying the sensorial scale

evaluation from 0 to 5 and the final marks from 0 to 20 (Alvelos et al., 2000).

3. Results and discussion

All data were introduced in a worksheet where columns represent the several variables

and the rows the cases. In order to analyse and interpret the data, it was used statistical,

geographical and cartographic interpretation software. The first one allows for variance,

regression, principal component and cluster analysis data, the second to figure out its

spatial distribution and the third to draw variable distribution maps.

This paper pertains solely to the results of wine sensorial data and final marks attributed

by a jury in the first year, which include the colour intensity (Cor), aroma (Aroma), red

Page 7: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

7

fruit aroma (FrVer), floral aroma (Floral), body (Corpo), astringency (Adst), total

acidity (AcTt) and final marks (NFinal).

The results for the twelve wines, obtained from the twelve analysed regions (blocks),

are depicted in Table 2.

Table 2

The final grades presents values that are significantly different between parcels, but not

when comparing the vineyards installation forms (F=4.14, P=0.048 and F=1.70,

P=0.222, respectively); it is impossible to analyse the variation within the parcels

because no data is available on the variations inside each block. For the parcels the

highest grades were obtained in Bateiras (BaG2) and Bico dos Casais (BCG2) (> 13)

and the lowest in Cardanhas (CaG2 and CaG3) (< 12). According the installation forms

mean values were 12.9 in case of terraces and 12.3 for slope vineyards.

Comparing the data from two years it is noticed that in the last one the wine quality was

lower, keeping Bateiras as the best parcel and Cardanhas the worst. Comparing the

different installations forms one realises that, in the last year, the wines from terraces

are significantly better (F=8.72, P=0.014) than those from slope vineyards, a result that

did not occurred in the first year. This probably means that in years of good wines, there

is no significant difference between vineyards installations forms but when the quality

is lower the differences are noticeable.

The scale followed by the panellist (Table 3) indicates that the wines with a grade ≥ 15

are very good, with a final grade ≥ to 13 and < 15 are good, with a grade ≥ 10 and < 13

are regular, and with a grade <10 are mediocre. Within the classification of regular,

three levels were considered.

Table 3

Page 8: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

8

To define three wine quality levels among the twelve blocks, a cluster analysis was

executed, with the following results (Figure 4):

- cluster 1 - 8 wines (AmG1, AmG2, AmG3, BaG1, BaG3, BCG1, BCG3 and CaG1),

considered wines of average quality;

- cluster 2 - 2 wines (BaG2 and BCG2), considered the best quality wines;

- cluster 3 - 2 wines (CaG2 and CaG3), considered the worst quality wines.

Figure 4

The level 1 includes the wines classified by the panellist with grades > 13 (good), the

level 2 the wines classified with a grade between 12 - 13 (regular +) and the level 3 the

wines classified with a grade < 12 (regular and regular -).

Relatively to the PCA (Andrade et al., 2006), and according the interpretation of factor

one, the more relevant sensorial parameters (loadings > 0.70) are the aroma,

adstringency, colour and structure.

4. Conclusions

The differentiation between and within parcels characteristics, must be taken in

consideration to a differentiated vintage, with the purpose of getting grapes lots

qualitatively different, that should be converted separately into wine. The identification

and characterisation of quality major potential zones allows the improvement of the

other ones (Bramley, 2001), although there will be always some parcels, or part of it,

that will have better grapes (Krstic & Smart, 2002).

Considering the determined data, and for this situation, we notice:

- for the environmental data there are some, like the atmospheric temperature, after the

firsts hours of the day, are generally significantly different between and within parcels;

Page 9: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

9

- for the plants data, like leaves composition, the variance is not, generally, significant

between and within parcels;

- for the data soil we find different situations but, generally, we can say that between

parcels the variance is significant, and inside the parcels it can be or not;

- for the yield, must and wine data, where is not possible to analyse the variance inside

parcels, there are some parameters that are significant different between the parcels but

others ones, no.

Considering the variance analysis of all parameters, it is believed that it is necessary to

go on these trials in order to get broadly consistent data from year to year and increase

the relationship between site-soil-water-vine-wine; before reach these goals the

principal component analysis give important information on the variance of all

parameter, why must be taken on a predictive value (Bramley, 2001). Considering that

the spatial pattern is relatively consistent between years (Profffit & Hamilton, 2001) it is

more important to know the within parcel variability.

The understanding of the relationship between the inputs of grape production system

and their output is so required to reach the purpose of the precision viticulture.

To finish we can say that the vinification techniques can overcome some lack of grapes

quality, but the wines are better when the grapes are of good quality, why it is necessary

to select the better ones to get the betters wines (Silva, 1994).

Page 10: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

10

Figures list

Figure 1- Georeferenciation sites and its identification

Figure 2- Determination of the average plant pruning wood weight.

Figure 3- Soil samples collecting

Figure 4- Cluster analyse for the several blocks according the final grades.

Page 11: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

11

Figure 1- Georeferenciation sites and its identification

Page 12: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

12

Figure 2- Determination of pruning wood weight.

Page 13: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

13

Figure 3- Soil samples collecting

Page 14: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

14

CASE CLUSTER DISTANCE

AmG1 1 0.24

AmG2 1 0.20

AmG3 1 0.16

BaG1 1 0.25

BaG2 2 0.24

BaG3 1 0.15

BCG1 1 0.23

BCG2 2 0.24

BCG3 1 0.16

CaG1 1 0.26

CaG2 3 0.28

CaG3 3 0.28

Tree Diagram f or 12 Cases

Ward`s methodEuclidean distances

0 1 2 3 4 5 6

Linkage Distance

CaG3

CaG2

BCG2

BaG2

BCG3

BCG1

CaG1

BaG1

AmG2

BaG3

AmG3

AmG1

Figure 4- Cluster analyse for the several blocks according the final grades.

Page 15: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

15

Tables list

Table 1- General characterisation of selected parcels

Table 2- Results of wine tasting done in 2005

Table 3- Wines classification according the final grade

Page 16: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

16

Table 1- Identification and characterisation of selected parcels

Parcel name Parcel number Area (ha) Installation type

Amendoal (Am) 42 0.397 1 row terraces

Bateiras (Ba) 1 1.130 2 row terraces

Bico dos Casais (BC) 25 0.353 Slope vineyards

Cardanhas (Ca) 23 0.353 Slope vineyards

Page 17: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

17

Table 2- Results of wine sensorial data and final grade in 2005 (*)

AmG1 AmG2 AmG3 BaG1 BaG2 BaG3 BCG1 BCG2 BCG3 CaG1 CaG2 CaG3

Cor 3.83 3.83 3.67 4.00 4.17 3.50 3.17 4.17 3.17 3.50 3.00 2.83

Aroma 3.67 3.50 3.33 3.83 3.50 3.33 3.33 3.50 3.50 3.67 2.83 2.67

FrVer 2.16 1.83 2.00 2.50 2.00 2.00 2.33 2.00 2.33 2.50 1.50 1.67

Floral 0.00 0.33 0.33 0.67 0.33 0.50 0.83 1.33 0.50 0.67 0.83 1.00

Corpo 3.50 3.33 3.33 3.17 3.67 3.17 3.50 3.50 3.33 2.83 2.83 2.67

Adst 3.00 3.50 3.00 3.50 3.83 3.00 3.33 3.00 3.17 3.17 3.00 2.00

AcTt 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.00

NFinal 12.83 12.50 12.83 12.67 13.50 12.83 12.67 13.83 12.50 12.17 11.83 10.67

(*) The scope of the sensorial data is between 0 and 5 and, the final grade, between 0 and 20.

Page 18: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

18

Table 3- Wines classification according the final sensorial data

Classification Mediocre Regular - Regular Regular + Good Very good

Grades < 10 ≥ 10 e < 11 ≥ 11 e < 12 ≥ 12 e < 13 ≥ 13 e < 15 ≥ 15

Index 0 1 2 3 4 5

Page 19: Applying Precision Agriculture Methodology to the ... · ... (AT), sum of exchangeable bases (SBT), 5 ... volatile acidity, fix acidity, total acidity (volatile acidity + fix acidity),

19

References

Alvelos H; Cabral J; Amaral B; Barros P (2000). Statistical methods for controlling

wine tasting panels. Ciência Técnica Vitivinícola 15 (1):15-26.

Andrade M; Paim A; Silva J (2006). Aplicação de métodos analíticos e

quimiométricos para caracterização de vinhos comercializados no Recife. 29a Reunião

Anual da Sociedade Brasileira de Química

Bramley R (2001). Progress in the development of precision viticulture - variation in

yield, quality and soil properties in contrasting Australian vineyards. CSIRO Land and

Water and Cooperative Research Centre for Viticulture. Private Bag nº 2, Glen

Osmond, SA 5064, Australia.

Escuder E; Zaballa O (2000). Ensayos de aclareo manual de racimos en la D.O. La

Rioja. Agro Técnica 811: 61-68.

Proffit T; Hamilton R (2001). Precision viticulture - Technology to Optimize

Vineyard Performance. Australian Wine Industry. Technical Conference. Adelaide -

Australia.

Rizzon L; Miele A (2003). Avaliação da cv. Merlot para elaboração de vinho tinto.

Ciências Tecnológicas Alimentares, Campinas, 23 (Supl); 156-161, Dez.2003

Silva H (1994). A enologia preventiva: a estabilidade dos vinhos. Vida Rural 1596: 32-

34.

Krstic M; Smart R (2002). Possible applications of new world viticulture in Portugal.

Revista de Vinhos. II Conferência Revista de Vinhos.