- Top of page
- Materials and methods
Like many agricultural products, a particular wine may be defined and protected by a legal framework. This is of great interest to producers because it marks the identity of the product in an increasingly globalised market and provides decisive criteria of acceptability for consumers in terms of guaranteed quality (Galgano et al. 2008, Tonietto 2008, Gonzálvez et al. 2009). Thus, Protected Geographical Indications (PGIs) have appeared as place-based names that convey the geographical, cultural and/or historical identity of the product (Bowen 2010). As indicated by Council Regulation (EC) no. 479/2008 for European wines (European Commission 2008), PGIs are based on the essential or exclusive attribution of their quality and particular characteristics to the geoclimatic environment involved in vineyard development and wine production.
In recent years, some studies have attempted to determine the degree of interaction between grapes, musts or wines and the natural environment, according to one of the three main generators of typicity in the vineyard – climate, soil and landscape – generally by analysing phenolic and flavour composition to assess the effects.
The influence of climate on agricultural crops is mainly mediated by sunlight exposure, rainfall and temperature (Hidalgo 2002). Given that climate has been considered the most relevant factor, many studies are focused on its effects on different crops, such as for quinoa (Bois et al. 2006), wheat (Li et al. 2010) or rice (Mohammed and Tarpley 2010).
For instance, among reports of viticulture, Bergqvist et al. (2001) examined the influence of sunlight exposure on colour intensity, anthocyanin accumulation and total phenolic concentration of Cabernet Sauvignon and Grenache grapes. Anthocyanins and total phenolics were reported to increase linearly only with diffuse but not direct sunlight exposure, suggesting that the effect of sunlight on fruit composition was dependent on the grape berry temperature resulting from the exposure. Similar conclusions were reported by Spayd et al. (2002) for Merlot grapes. Nevertheless, the flavonol, anthocyanin and quercetin level in Pinot Noir grapes with a moderate to high sunlight exposure was higher than for those located in more shaded positions on the vine (Price et al. 1995).
Sadras and Soar (2009) examined the ability of the yield components of Shiraz vines to buffer a 2–4°C increase in ambient temperature at different phenological stages and showed that only budburst transiently accelerated development because of this slight increase in temperature, in comparison with that of the controls. Yasui et al. (2002) studied the effect of weather conditions on grape resveratrol concentration, reporting that high temperature during grape development had an adverse effect. Therefore, grape yield and composition appear to be dependent on climatic attributes.
The effect of soil has also been evaluated because it provides the agricultural system with nutrients, water, oxygen and mechanical support (Wang et al. 2003). Some of these studies focused on how vine water status may influence the phenolic concentration of grapes or the profile of volatile compounds in wine (Kennedy et al. 2002, Ojeda et al. 2002, Koundouras et al. 2006). Two responses to water deficit in the vineyard were observed in these studies: an indirect and always positive influence on phenolic concentration due to berry size reduction, and a direct effect on biosynthesis that could be positive or negative, depending on the type of phenolic compound, the period of grape development affected and the severity of the water deficit.
Texture and fertility features of vineyard soil have also been examined. For example, Choné et al. (2001) investigated the influence of soil, in terms of texture and organic matter, on grapes, must composition and wine quality for Cabernet Sauvignon in the Médoc area (France). The authors suggested that nitrogen deficit in soils had a positive effect on the anthocyanin and tannin concentration of wine in spite of a reduction in vine vigour and yield production. Andrés de Prado et al. (2007) also reported an approach to assessing wine properties and relating them to the influence of soil. In line with previous reports, the most fertile soils with the greatest water-holding capacity produced wines of poor quality in terms of phenolic concentration, flavour and colour intensity.
Gómez-Míguez et al. (2007) examined the colour and aroma profiles of white wines from Designation of Origin Condado de Huelva (Spain) in order to evaluate the effect of clay–sand composition of vineyard soils, concluding again that the type of soil could directly affect both colour and aroma characteristics of wines.
The influence of landscape, as a combination of geology and topography, on grape and/or wine characteristics has not been so widely investigated. Mateus et al. (2001), however, studied the effect of vineyard altitude on the proanthocyanidin composition of two grape varieties and showed that a low altitude favoured the biosynthesis of total catechin, low molecular weight oligomers and total extractable proanthocyanidins. Conversely, it is interesting to note that cultivation altitude for the same two grape cultivars had the opposite effect on anthocyanin biosynthesis because a higher concentration of anthocyanins was reported at a higher altitude (Mateus et al. 2002).
As observed, many authors have assessed and demonstrated the influence of a single parameter of the natural environment on grape, must or wine composition. Research is scarce, however, into the combined environmental effects leading to an improved understanding of how geographical origin influences vineyard behaviour. Tesic et al. (2001a) evaluated some climatic and soil factors that determine the phenological stages of vineyard sites, relating the precocity of veraison and flowering to fruit composition and wine attributes. The same authors developed the ‘site index’ concept as a combination of those climatic and soil attributes of the vineyard and correlated it to several viticultural variables (Tesic et al. 2001b). Van Leeuwen et al. (2004) examined the overall effect of some climatic and soil factors on the growth and development of vines, reporting the proportion of variance attributable to soil and vintage for each vine experimental parameter. They did not, however, study either the relationships between soil and climatic variables or the combined effects of both factors on vine experimental parameters.
In order to study a combination of effects, several statistical techniques of multivariate analysis may be applied. Among them, analysis of variance and principal component analysis (PCA) are the most frequently used (Frau et al. 1997). The first evaluates the existence of significant differences in the parameters analysed between the samples, whereas PCA removes the redundant information by generating a new set of independent parameters as a result of the best linear combination of the original ones. In this way, PCA can, first, explain the existing relationships among the different variables considered and, second, provide an overview of the capacity of the variables to explain the characteristics of samples from different regions (Douglas et al. 2001, Kallithraka et al. 2001, Morlat and Bodin 2006, Goldner and Zamora 2007, Vilanova et al. 2007).
All these studies, by applying PCA, were able to differentiate wines from different viticultural regions according to their physicochemical, mineral and/or sensory characteristics. They did not, however, relate these differences among the wines to the agroclimatic factors involved in their production.
In addition, each viticultural region has its own oenological tradition, heritage culture and history. Thus, in wine production, the human factor may be significant in differentiating the final product from different viticultural regions. The question remains as to whether this differentiation is caused by winemakers or whether they simply amplify the differences already present in grapes and those shaped by the agroclimatic fingerprint. There are no studies, however, that explore the influence of both the natural environment (considered as a combination of soil, climate and landscape factors) and grape characteristics on the PGI appellations in terms of a whole terroir concept and without considering the human factor.
For this reason, the present research is aimed at establishing a methodology that considers and analyses a combination of the grape, soil, climate and landscape attributes of the viticultural region. In order to validate this methodology, the wine region of the Balearic Islands, an archipelago located in the Mediterranean Sea, off the east coast of Spain, has been chosen. Four PGIs, called ‘Vi de la Terra’ (VT), coexist in this region: VT Mallorca, VT Illa de Menorca, VT Ibiza and VT Formentera, each one located on a separate island.
The legal recognition of these four PGIs dates from the last decade (in 2007, 2002, 2003 and 2004, respectively); however, their differentiation has always been socially and historically recognised because the Balearic Islands vine-growing subregions are discrete islands. Nevertheless, the validity of these PGIs still has not been demonstrated according to their geographical origin as required by the Council Regulation (EC) no. 479/2008 (European Commission 2008).
The principal aim of this study was to establish a methodology to evaluate the appropriateness of the recognition of the PGIs based on the grape and agroclimatic attributes of their vineyards. To validate the proposed methodology, Merlot and Cabernet Sauvignon cultivars from the vine-growing region of the Balearic Islands were chosen. For each cultivar, multivariate analysis techniques were applied to identify, in a first approach, the different behaviour patterns related to the PGIs by analysing a combination of the grape characteristics and the soil, climate and landscape conditions involved in its production.
- Top of page
- Materials and methods
The attribution of quality and particular characteristics of wine to its geographical origin has become important for both producers and consumers. In the last decade, several authors have studied the characterisation and differentiation of wines according to their geographical origin by applying PCA on a data matrix composed only of wine characteristics, such as their aroma profile, sensory attributes, phenolic compounds and/or mineral composition, independent of the grape and agroclimatic variables. For instance, Kallithraka et al. (2001) achieved a classification of Greek red wines from three areas of production, in terms of anthocyanins and sensory analysis. Miele et al. (2010) discriminated red wines from five out of seven Brazilian viticultural regions by means of applying PCA to their physicochemical composition. Douglas et al. (2001) showed the nature and magnitude of the differences among Riesling wines from two Canadian terroirs on the Niagara Peninsula; based on the sensory attributes, their research documented the appropriateness of two unique appellations within that vine-growing region.
As observed, none of these studies explored the influence of the natural environment and grape features on the different viticultural subregions, considering the human factor in all cases. Therefore, to check the existence of differences among vine-growing subregions or recognised appellations of quality prior to wine production, the present study has proposed a methodology to evaluate simultaneously the grape and agroclimatic characteristics of the vineyard.
According to the results obtained in PCA, a specific role may be assigned to the first three principal components for both Merlot and Cabernet Sauvignon cultivars. As observed in Figure 1 for the Merlot cultivar, while PC1 presented temperature conditions and texture properties of soil as dominant descriptors, PC2 was mainly represented by grape characteristics and PC3, by rainfall and landscape conditions and fertility properties of the vineyard. In the case of the Cabernet Sauvignon cultivar (Figure 2), PC1 was mainly described by maximum temperature conditions and texture properties of the vineyard, whereas PC2 presented elevation, rainfall, minimum temperature and fertility properties of soil as dominant features, and PC3, colour characteristics of macerated grape juice. Thus, for both grape varieties, there were two dimensions in PCA focused on natural environmental or agroclimatic descriptors (PC1 and PC3 in the case of Merlot cultivar, and PC1 and PC2 for Cabernet Sauvignon cultivar) and another one representative of grape attributes (PC2 and PC3, respectively, for the Merlot and Cabernet Sauvignon cultivars).
As observed, an almost similar distribution of the variables into the first three principal components was detected for Merlot and Cabernet Sauvignon cultivars, which showed the existence of relationships between these parameters, completely independent of the grape cultivar. For instance, colour tonality and colour intensity showed an inverse relationship in both cases, whereas silts content and maximum temperature parameters presented a direct one. Soto et al. (2011) also noted the highest colour tonalities for the analysed wines which had the lowest colour intensities and vice versa, while Koundouras et al. (2006) had already observed a direct relationship between the proportion of silts in vineyard soils and mean maximum temperature, maybe because of the fact that their study was also carried out in the Mediterranean basin.
In the present study, the first three principal components allowed the identification of different behaviour patterns of the samples according to their geographical origin, because the four clusters obtained at the distribution of the sample dataset in the PCA plot matched with the four PGIs recognised in the wine region of the Balearic Islands. The maximum temperature conditions and silts content of the vineyard soil were the most important descriptors of the pattern distribution of the samples in the PCA plane because of their presence in the first principal component of both Merlot and Cabernet Sauvignon multivariate results. In fact, samples from the four Balearic Islands PGIs were mainly separated along the PC1 (Figures 1, 2). According to the literature, both descriptors have a major influence on vine development and consequently on the characteristics of the product (Andrés de Prado et al. 2007, Soar et al. 2008).
From the multivariate results for Merlot, it could be concluded that PC1 and PC3, both related to agroclimatic features, explained the different behaviour patterns better and differentiated the four Balearic Islands PGIs; meanwhile, PC2 (related to grape characteristics) displayed differences only between VT Illa de Menorca and VT Ibiza (Figure 1).
The PCA results for Cabernet Sauvignon revealed that the principal component related to grape attributes (PC3) could identify pattern differences and separate sites in terms of their appellation of quality, to a greater extent than in the case of Merlot. In any case, for both Merlot and Cabernet Sauvignon cultivars, it was essential to consider together the first three principal components, based on the overall grape, soil, climate and landscape attributes of the vineyard, to divide the samples into different clusters that matched with their corresponding PGIs.
It is important to point out that the proposed methodology has been developed only from a single season and that for a complete validation and definition of the existence of these PGIs or those from another vine-growing region on a geographical basis, it would be necessary to apply the methodology over several seasons.