1 Introduction

This work assesses the environmental impact associated with a wine produced in Portugal (i.e. the white vinho verde) and exported worldwide. This wine is exclusively produced in the northern part of Portugal, in the Demarcated Region of Vinho Verde. The wine sector assumes an important position in Portugal, as the country is among the 12 leading countries for vine plantation (OIV 2010). The white vinho verde is the white wine with the largest production share in Portugal. In 2010–2011, about 50 million litres of white vinho verde were produced in Portugal, of which around 30 % were exported (IVV 2010; CVRVV 2011).

The main stages associated with the wine include activitites during viticulture, wine production (from vinification to storage) and distribution (wine commercialisation). However, several other activities also occur. This includes the production and transport of materials used in viticulture (for example phytosanitary products) and during wine production (including wine additives).

Several studies have evaluated the environmental impacts of wine using life cycle assessment (LCA). Environmental product declarations (EPDs) and Product Category Rules (PCRs) have also been published for wine (IEPDS 2011; EPD 2008a, b). These LCA and EPD studies vary on the wine production country, type of wine, production practices (conventional and biological) and, among other aspects, on the life cycle stages considered. Several LCA studies present a cradle to grave approach including wine distribution (Aranda et al. 2005; Gonzalez et al. 2006; Gazulla et al. 2010; Bosco et al. 2011; Point et al. 2012). Other studies perform a cradle to gate approach considering wine distribution (Ardente et al. 2006; Petti et al. 2006). The remainder consider cradle to grave or cradle to gate approaches, however, distribution is not analysed (Notarnicola et al. 2003; Pizzigallo et al. 2008). In summary, the studies demonstrate that the production of glass bottles and viticulture are important stages regarding the impacts of the wine life cycle.

However, a study that analyses a complete set of the relevant inputs and outputs associated with viticulture, wine production and distribution and concerning the life cycle of wine that is sold worldwide is not available in the reviewed literature. We have found that the available studies may not be complete either in the inventory list of materials and energy considered from viticulture (by excluding for instance, the production and the emission resulting from the application of the phytosanitary products) or in the complete assessment for the impacts categories associated with the activities from the main stages of the wine life cycle. Studies that cover a more complete set of environmental impacts lack in the assessment of at least one, of the following impact categories: land competition, ozone depletion, human toxicity, freshwater and marine ecotoxicities (Gazulla et al. 2010; Point et al. 2012; Vázquez-Rowe et al. 2012). Moreover, the studies reviewed lack analysis of the influence of wine distribution (for wines commercialised to several countries for which a long-range transport occurs) in the impact results (Gazulla et al. 2010). This paper aims to cover some of the identified lacks.

The purpose of this study is to perform a LCA of a wine exclusively produced in Portugal, i.e. the white wine designated locally as white vinho verde. This is to identify the environmental impacts occurring along the wine life cycle stages of viticulture, wine production, distribution and glass bottle production. For that purpose, we have made use of data provided by the largest producer of vinho verde in Portugal, Aveleda S.A. In 2008–2009, this company sold about 25 % of the national production of the white vinho verde and exported worldwide around 50 % of its wine. The results obtained for several impact categories are presented for each life cycle stage. A sensitivity analysis is done to identify the parameters that most influence the results of the impact assessment.

2 Methods

2.1 Scope definition

The stages considered include activities taking place during (1) viticulture (grapes growing), (2) wine production (from vinification to wine storage), (3) distribution and (4) bottles production. The data used pertains to the wine produced in the campaign 2008–2009 and refers to 0.75 l of white vinho verde as the functional unit.

The system boundary includes the materials, water and energy used in the activities taking place during viticulture, wine production and the white glass bottle production. Worldwide wine distribution is also included, as well as the transportation of grapes, wine and must and other products used during wine production (Santos 2010; Machado 2011) (Fig. 1).

Fig. 1
figure 1

Life cycle of the white vinho verde. The dashed line identifies the stages not included in the analysis

Full size image

Some activities were not considered here, as for instance, the transportation of man labour during the grapes harvesting. This is due to the fact that harvesting occurs automatically, i.e. only 4 % of the harvested area uses manual labour. This amount was considered to be negligible. Other activities as the potential impacts regarding the consumption of refrigerated wine were not taken into account. The vinho verde is always refrigerated before consumption and this may occur in the wine shop or at the consumer home. Despite the fact that the potential impact due to refrigeration may be significant, data regarding energy consumption for refrigeration is not readily available and thus this stage was excluded from this study. The same occurs for the consumer transport to purchase wine. This activity was also not considered. Other activity not included is the transportation of wastes as it was considered to be negligible. The transport of phytosanitary products used in the viticulture stage was also excluded due to the lack of data about distances travelled and type of transport. Our study does not consider the capital goods production, maintenance and end of life. Others activities concerning the cork stoppers production as well as the labelling materials used in the bottles were also not considered in the study.

2.2 Inventory

The analysis was performed by using SimaPro software (SimaPro 2011). The Ecoinvent database (Ecoinvent 2011) was employed as the principal source of background data. Table 1 shows the activities and its duration associated with the viticulture, wine production and distribution stages for Aveleda S.A. wine campaign of 2008–2009.

Table 1 Time frame associated to each life cycle stage for the wine campaign 2008–2009
Full size table

The inventory data from viticulture were supplied by the case company or estimated using activity data from the company and emission factors from literature. Aveleda S.A. is the largest producer of white vinho verde and despite owning vineyards it purchases around 90 % of grapes and about 40 % of must and wine used to produce wine. Grapes and must are produced by other local producers within the Demarcated Region of Vinho Verde. For the purpose of the analysis, we have considered that the practices of other producers are similar to Aveleda’s. The grape vines are not irrigated.

Table 2 lists the inventory for viticulture calculated based on data supplied by Aveleda S.A. about the vine area, grapes and wine production and productivity, materials, water and energy used as well as the wastes produced. Background data associated with the production of materials used in viticulture is estimated using Ecoinvent (2011).

Table 2 Direct inputs and outputs from viticulture
Full size table

The emissions resulting from the application of phytosanitary products, fertilisers and from diesel consumption in viticulture were estimated. The emissions associated with the use of phytosanitary products are, according to many authors, mainly released to soil (Audsley 1997; Hauschild 2000; Margni et al. 2002; Milà i Canals 2003; EMEP/EEA 2009). This study considers that 75 % of the emissions of the active chemical are released to soil. The remaining is emitted to air. This approach was considered to occur for all phytosanitary products used. The value 75 % is not explicitly recommend to use in the literature, thus we have performed a sensitivity analysis, regarding the amount of the active chemicals released to air and soil, as result of the application of the phytosanitary products.

The emissions of nitrogen compounds to air such as nitrous oxide (N2O), nitrogen oxides (NOx) and ammonia (NH3) and the emission of nitrate (NO3 −) to water, were estimated based on the methodology described in Intergovernmental Panel on Climate Change (IPCC 2006a). Nitrogen compounds are emitted to air and water as a result of the use of fertilisers (synthetic and solid manure). The emissions of N2O to air (i.e. the N2O direct emissions from managed soils and N2O indirect emissions from volatilisation and from runoff and leaching) were estimated based on emission factors taken from IPCC (2006a). Besides, the calculated N2O indirect emissions from volatilisation considered also emission factors specific for the chemical compounds occurring in the synthetic fertilisers, i.e. the ammonium nitrate, ammonium sulphate and the urea ammonium nitrate. These emission factors, that were also used to estimate NH3 emissions, were collected from Nemecek and Kägi (2007). We have assumed that the N2O indirect emissions are all volatilised as NH3. Emission factors for NO3 − are from IPCC (2006a). Although not often included in the inventory of wine-related studies, the emissions of NOx were estimated as being 0.21 times the total emissions of N2O (Nemecek and Kägi 2007). This equation, valid for non-irrigated systems (as stated in Nemecek and Kägi 2007), was used in order to cover a wide number of pollutants associated with viticulture.

The emissions of fossil carbon dioxide (CO2) resulting from urea application were calculated according to IPCC (2006a). Air emissions from diesel combustion during the use of agricultural machinery were estimated based on the emission factors from the European Monitoring and Evaluation Programme/European Environment Agency (EMEP/EEA 2009) by using the tier 2 approach.

Must and wine purchased by Aveleda S.A. come from local producers. About 85 % of the must and wine travel between 9 and 46 km and the remaining 15 % travel between 35 and 60 km. During the campaign 2008–2009, grapes were purchased from about 40 producers. About 83 % of the grapes travelled to Aveleda less than 60 km, around 15 % travelled between 61 and 110 km and the remaining grapes travelled between 110 and 374 km. Table 2 lists the t.km associated with grapes, wine and must transportation from local producers to Aveleda. Grapes are transported in a tractor and trailer and wine and must in lorries >32 t.

The inventory data from wine production was mainly supplied by the case company (Table 3). This specifically refers to materials, water and energy used as well as water emissions and solid wastes produced.

Table 3 Direct inputs and outputs from wine production
Full size table

Air emissions associated with fermentation and liquefied petroleum gas (LPG) use were estimated based on activity data from Aveleda S.A. and emission factors from literature. During fermentation, biogenic CO2 and nonmethane volatile organic compounds (NMVOC) are released. The emission factor for CO2 was taken from Peynaud (1981) and Soares (1998) and the emissions of NMVOC were calculated from EMEP/EEA (2009). The emissions released during LPG consumption in wine production (LPG is used to generate vapour used to sterilise the empty wine bottles) were estimated using emission factors from IPCC (2006b) and EMEP/EEA (2009).

The materials used during wine production, the enologic products, the filtration and cleaning agents and the other wine additives, are transported by truck to the company. Most of the products are purchased from a national retailer and transported at a distance less than 50 km. Exceptions are the sulphur dioxide (SO2) and a few cleaning agents, for which the supplier is located at a distance of about 350 km. Table 3 lists the t.km associated with the transportation of wine products as the additives, disinfectants, enologic products and filtering materials to Aveleda. These are transported in vans <3.5 t.

White glass bottles were assumed to be manufactured out of about 61 % recycled glass. This value was retrieved from the ecoinvent database (version 2.2) for white packaging glass (Ecoinvent 2011).

In respect to the distribution, in 2009, about 50 % of the total wine produced by Aveleda S.A. was sold in Portugal. The remaining wine was exported to several countries. Nearly 40 % of the total exportation was delivered by truck, boat and plane, to six countries, namely USA, Germany, France, Angola, Canada and Brazil.

Information regarding the sales destination countries, amount transported and type of transport (truck, boat and plane) was also made available by Aveleda S.A. The company sells in 38 countries, from which 20 are located in Europe, seven in America, six in Africa, four in Asia and one in Oceania. The road transport (about two thirds of the transportations) is mainly done by 16–32 t lorries. However, transoceanic tankers and flight freights are also used. Around one third of the transportations occur by boat and the flights represent a small amount. Emissions were estimated for lorries, transoceanic tankers and flight freights by using the Ecoinvent (2011) database. Table 4 expresses in t.km the amount of wine transported in Distribution.

Table 4 Inventory data for the distribution of vinho verde
Full size table

2.3 Impact assessment

SimaPro (version 7.3.2) was used to model the life cycle of wine using midpoint indicators of environmental impact (CML 2001 impact assessment method). Results are obtained for the characterisation step. This widely accepted method was chosen also to allow the comparison of our results with a large number of related studies on wine impact assessment.

The amount of biogenic CO2 sequestered from the atmosphere during vine and grape growth was assumed to be equal to the amount of CO2 that is released back to the atmosphere due to the oxidation of the carbon contained in the pruning wastes, as well as due to oxidation of the carbon contained in the grapes along the downstream life cycle stages. Data taken from Ecoinvent also assumes neutrality for biogenic CO2. Therefore, biogenic CO2 was not considered in the global warming impact category.

2.4 Sensitivity analysis

A set of parameters, considered to be the most uncertain, were changed in order to assess its influence in the environmental impact results (Table 5). In viticulture, the parameters include those associated with the emission to air and water of nitrogen compounds (NO3 −, NH3, N2O (direct and indirect)) caused by application of fertilisers. We also changed the NOx emission between a minimum and a maximum as a function of the total emission of N2O. We changed the amount of manure used in fertilisation by ±20 % and the CO2 emission as result of the urea application by lowering it by 50 % relative to the reference case. Finally, for viticulture, we also changed the emissions to soil and air associated with the use of the active chemical of the 12 phytosanitary compounds (listed in Table 2).

Table 5 List of parameters and respective changes considered in the sensitivity analysis
Full size table

For wine production, factors changed were the emission factors of CO2, methane (CH4) and N2O associated to the use of LPG in the sterilisation of bottles occurring in the production line. These were changed between a minimum and a maximum within a range defined in EMEP/EEA (2009). The results are obtained by considering simultaneously, in one hand all minimum values of emission factors and, on the other hand, all maximum values.

The amount of recycled glass used in bottle production is also a source of uncertainty. The Portuguese situation, in 2009, is that 55 % of the glass packaging has been recycled (APA 2011).

However, the amount of white packaging glass being recycled for wine bottle production purposes is not available. Moreover, no inventory data is readily available concerning the production of white glass in Portugal. Also, no data is available in respect to the materials and energy needs as a function of the amount of cullets used to produce white glass for packaging. For these reasons, the execution of a sensitivity analysis on the percentage of recycled glass incorporated in the production of white packaging glass, although important, is left outside the scope of this paper. More research is needed on this topic and this must be done with the cooperation of the glass production industry.

In the distribution, the sensitivity analysis focused on changing the distance travelled. The distances changed by ±30 % for all the 38 countries and for each transportation mode (truck, boat and plane).

3 Results

Results show that the viticulture is the life cycle stage that mostly contributes to all impact categories. The contribution to each category is larger than 50 % (Table 6).

Table 6 Results (expressed in absolute values and in percentage of contribution) from the characterisation step presented for each impact category and for all the life cycle stages considered
Full size table

Bottles production is the second contributor. The contribution varies from about 4 % (eutrophication) to 26 % (acidification). Wine production and distribution are the subsequent contributors. The contribution of wine production varies between 0.6 % (to land competition) and about 13 % (from marine aquatic and sediment ecotoxicity 100a). The contribution of distribution ranges from 0 % (to land competition) to 14 % (to photochemical oxidation).

Viticulture has a large contribution to the majority of the environmental problems. We identify below the pollutants responsible for about two thirds of the overall contribution of viticulture to each impact category.

Eutrophication is caused by the emissions of NO3 − (to water) due to the use of fertilisers. Land competition is caused by the soil area needed for grapes growing. The ozone depletion problem is related to the emission of CFC and halon occurring, respectively, during the production of a phytosanitary compound (the acetamide aniline compound, phenexamid) and the production of diesel. Terrestrial ecotoxicity is due to the emission of glyphosate (to soil) and vanadium (to air). Glyphosate emission to soil is due to the phytosanitary application and the emission of vanadium occurs during the production of fertilisers.

Global warming problems are essentially due to the emission of CO2 and N2O. CO2 emissions are mainly associated with the combustion of diesel, used in the agricultural machinery, and the CO2 emitted during application of urea. The emission of N2O is essentially occurring due to the application and production of fertilisers.

The problems of fresh water and sediment ecotoxicity are due to emission of glyphosate (to air) and barium, formaldehyde, polyaromatic hydrocarbons (PAHs) and zinc (to water). The emission of glyphosate occurs during the application of this phytosanitary product. Barium and PAH are emitted during the production of diesel. Formaldehyde results from the production of glyphosate and zinc occurs during the production of a phytosanitary compound (i.e. the dithiocarbamate compound, metiram).

Human toxicity is associated with the emission of heavy metals to air (as nickel and arsenic), benzene (to water) and PAH to both water and air. Nickel and arsenic are associated with the production of fertilisers, electricity and diesel. The emission of benzene to water occurs mainly during the production of phytosanitary compounds (as the acetamide aniline compound, phenexamid and other unspecified). The emission of benzene to air is also due to the production of phytosanitary compounds (as the acetamide aniline compound, phenexamid) but also to the production of diesel. The emission of PAH to air occurs during the production of fertilisers and electricity. The emission of PAH to water occurs during diesel production.

The problems of marine aquatic and sediment ecotoxicity are caused by the emission of vanadium, and nickel (to air) and tributyltin (to water). Tributyltin is a biocide used to prevent the growth of marine organisms on the hulls of large ships to water during barge tanks transport. Vanadium and nickel are associated with the production of fertilisers and electricity and tributyltin is mainly due to the production of diesel. Abiotic depletion is due to the consumption of fossil-based energy resources as oil crude, natural gas and coal. Acidification is mostly associated with the emission of SO2 and NOx to air. The emission of SO2 is mainly caused during the production of electricity and diesel. NOx is essentially associated with the diesel combustion is the agricultural machinery. Finally, the photochemical oxidation results from emission of SO2 and CO. SO2 emission is mainly due to the production of electricity and diesel and CO emission is mainly associated with the combustion of diesel used during the operations of agricultural machinery.

The main environmental problems associated with glass bottle production are acidification (26 %), photochemical oxidation (24 %), freshwater sediment ecotoxicity (23 %), freshwater aquatic ecotoxicity (22 %), terrestrial ecotoxicity (20 %) and abiotic depletion (19 %). Acidification and photochemical oxidation are mainly caused by the emission of SO2 to air. Freshwater aquatic and sediment ecotoxicity are due to the emission of heavy metals (mainly selenium (emitted to air), nickel and barium emitted to water). Finally, abiotic depletion results mostly from the use of natural gas and crude oil.

Wine production problems relate mainly to marine aquatic and sediment ecotoxicity (both contributing to 13 %), acidification and photochemical oxidation (contribution of about 12 %) and human toxicity (contribution of 10 %). Ecotoxicity is due to the release of heavy metals to air (mainly vanadium). Photochemical oxidation is mainly associated with the emission of SO2 and the human toxicity are due to the release of heavy metals (as nickel and arsenic) and PAH to air.

The environmental problems caused during distribution include photochemical oxidation (contribution of 14 %), and marine sediment ecotoxicity (contribution of 10 %). Photochemical oxidation is mainly caused by the emission of CO and SO2 associated with diesel combustion during truck transport. The problem of marine sediment ecotoxicity 100a relate to the use of biocides (i.e. tributyltin) used to prevent the growth of marine organisms on the hulls of large ships to water during barge tanks transport.

The results of the sensitivity analysis show that changing some parameters does not influence the impact assessment results for the variation range analysed. This is valid for a number of parameters analysed including the amount of manure used, the emissions of CO2 associated with the urea use and the emissions to air associated with the use of the 12 phytosanitary products. Other parameters, not influencing the results, include the distances travelled during wine distribution for a large number of countries.

However, some parameters influence significantly the calculated impact categories. Table 7 shows the parameters for which the changes in the characterisation results are equal or larger than 1 % for at least one impact category, when compared to the results obtained for the reference case. Significant changes are calculated for parameters SA1 (NO3 − leaching to groundwater), SA2 (N2O direct emission from managed soil) and SA4 (N2O indirect emission from runoff and leaching) (see Table 7).

Table 7 Sensitivity analysis results, calculated for the characterisation step, expressing the changes for each impact category related to the reference case
Full size table

We observed that these changes are significant for only a few impact categories. SA1 leads to changes in eutrophication due to changes in NO3 − emissions. SA2, SA3 and SA4 lead to changes in global warming 100a due to the variation on the emission of N2O which has a large global warming potential.

Others parameters influence, by a comparatively small amount, multiple categories (see Table 7). This is the case for SA5 (NOx emissions from fertilisers use), SA6 (NH3 emissions), SA7 (emissions to soil from the use of phytosanitary compounds), SA8 (emission factors for CO2, CH4 and N2O from LPG used for bottles sterilisation during wine production), SA9 (distance travelled by truck in Portugal) and SA10 (distance travelled by boat to USA).

SA5 leads to small changes in the categories acidification (−1, +2 %) and eutrophication (0, +1 %). SA6 caused changes in acidification (0, +2 %) and eutrophication (0, +1 %). SA7 affects freshwater aquatic ecotoxicity 100a (+2 %), terrestrial ecotoxicity 100a (+4 %) and freshwater sediment ecotoxicity 100a (+1 %). The parameter SA8 leads to a change in global warming 100a (±1 %) and photochemical oxidation (+1 %). The parameter SA9 leads to changes in a large number of environmental problems, varying from ±1 % (acidification, human toxicity, marine aquatic and sediment ecotoxicity 100a and terrestrial ecotoxicity) to ±4 % (photochemical oxidation). SA10 leads to small changes for the categories of marine aquatic and sediment ecotoxicity 100a (± 1 %).

4 Discussion and conclusions

Viticulture is the life cycle stage with the largest relative contribution to the overall environmental impact. This stage contributes by more than 50 % to all environmental problems. Bottle production appears as the second stage due essentially to the fossil fuels used during glass bottles production. We have found that these two stages have a larger contribution to the impact when compared to wine production and distribution. This is also valid for a number of reviewed wine studies (e.g. Gazulla et al. 2010; Point et al. 2012). Furthermore, the environmental problems for wine production are, alike for the bottle production stage, associated with the fossil fuels used during wine production.

We observed that despite the fact that 50 % of the wine is being distributed to 37 countries located worldwide, distribution has a relatively lower impact when compared to viticulture and bottle production. Environmental problems caused by distribution are mainly associated with the emission of gases from diesel combustion and use of biocides associated with the maritime transportation. Based on these results, we may then conclude that wine distribution, for this specific case in which a long-range transportation occurs, does not influence significantly the overall environmental problems associated with the wine life cycle.

The comparison with other recent published wine-related LCA studies shows that the impact assessment results may differ. A significant difference occurs for global warming. We summarise here the differences found for the viticulture and distribution stages, which are the stages for which major variations occur. The values found for viticulture expressed in kilograms CO2 eq. per FU (0.75 l of wine) are the following: 0.33 (Bosco et al. 2011), 0.50 (Gazulla et al. 2010), 0.80 (Point et al. 2012) and 2.5 (Vázquez-Rowe et al. 2012). The value we calculated for white vinho verde is equal to 2.0 kg CO2 eq./FU. These differences may occur due both to different degrees of mechanisation in the agricultural procedures and to different amounts and composition of the fertilisers used (manure and synthetic). The methods used to estimate the emissions may also be the cause. Thus, further investigation is needed to understand these differences.

For the distribution stage, the study from Bosco et al. (2011) calculates 0.44 kg CO2 eq./FU referring to a wine distributed from the winery gate in Tuscany to a wine platform located 1,600 km away. The study from Gazulla et al. (2010) calculates 0.076 kg CO2 eq./FU for domestic distribution within Spain and 0.085 kg CO2 eq./FU for distribution to UK. Our study calculates 0.18 kg CO2 eq./FU for the distribution in Portugal and 0.052 kg CO2 eq./FU for the worldwide distribution. That is to say that about 78 % of the GWP value is due to the domestic distribution. We conclude that indicators for global warming potential vary largely even among wines produced and commercialised within Europe. Therefore, more research is needed to better understand the differences in the magnitude among these values.

A complete assessment of the environmental problems allows a more supported decision on how to reduce the overall impact associated with the wine life cycle. This study shows the importance of using a more complete set of inventory data concerning specifically the life cycle of phytosanitary products and their consequent impacts in air, soil and water. In fact, by considering the production of phytosanitary products it was possible to draw important conclusions associated with the toxicity (human, terrestrial, freshwater and sediment ecotoxicities) related to the production and use of these products. In the wine campaign of 2008–2009, the number of phytosanitary interventions was comparatively large by comparison to other years. This was to avoid the reduction of grapes productivity caused by the mildew disease that could affect in this specific campaign largely the vineyard. For this specific situation, the most significant results indicate that the acetamide–aniline compounds (more specifically the phenexamid) are responsible for the emission of substances contributing to human toxicity. The use of glyphosate is contributing to terrestrial, freshwater and sediment ecotoxicities. Also, the production of the dithiocarbamate compound (more specifically the metirame) is causing freshwater and sediment ecotoxicities. The comparison among the toxicity categories (human, terrestrial, freshwater and marine aquatic and sediment ecotoxicities) also allows to conclude that the magnitude of the indicators for human toxicity (3.1E-01 kg 1.4-dB eq./UF), marine aquatic (3.7E-01 kg 1.4-dB eq./UF) and marine sediment (4.0E-01 kg 1.4-dB eq./UF) are relatively large when compared to the other toxicity categories. Therefore, human toxicity, marine aquatic and sediment seem to be the larger contributors to toxicity problems. This is true for all the life cycle stages considered.

We conclude that viticulture is also the stage where large sensitivities are found. Changes in the sensitivity parameters lead to comparatively large effect for few impact categories as eutrophication and global warming. Results show that the most influential parameters relate to the emission of nitrogen compounds resulting from the use of manure and synthetic fertilisers. Changes in these parameters (SA1, SA2 and SA4) affect eutrophication and global warming.

The other parameters do not influence significantly the impact categories analysed. Changes in the parameter SA3 leads to changes in the global warming potential. Changes in SA5 and SA6 change the acidification and eutrophication categories. Changes in the parameters associated with the emission of phytosanitary compounds to soil (parameter SA7) affect some of the toxicity related categories and changes in the emission factors specific for LPG (parameter SA8) result in small changes for global warming and photochemical oxidation. The changes in the distance travelled in Portugal by truck (parameter SA9) influence by small amount (from ±1 to ±4 %) several categories. Finally, the distance travelled to USA by barge tank (parameter SA10) show to affect only by a small amount the maritime toxicity problems.

A limitation of this study relates with the exclusion of some activities, namely waste management activities; the transportation of phytosanitary products, production and transport of cork, labels and caps; the consumer transportation to wine shops, wine refrigeration before consumption and the bottle end of life. This is due to two reasons: (1) data were not readily available at Aveleda, S.A. and estimation may result very uncertain and (2) these activities were considered by others as relatively less important when compared to the main stages here considered (as in Gazulla et al. 2010 and Point et al. 2012). However, we refer, in line with the study from Point et al. (2012) that among the abovementioned activities, the consumer transportation to purchase wine might be a relatively important stage and thus should not be neglected. In summary, the different results obtained indicate that a more complete study, covering more wine life cycle stages, is needed in order to reveal the relative importance of them.

Finally and from the review made, we conclude that for the few available wine-related LCA studies, the impact categories results are not easily comparable. This may be due to different methodological options as, for instance, the inventory parameters considered, the methods used to estimate emissions and the impact assessment methods. This also highlights the need of developing harmonised PCRs to allow the comparison among the results of LCA studies for wine.

Another possible reason relates to distinct procedures used to produce wine. Concerning future studies, it is important for the decision making on reducing wine impacts that the studies performed made available a more transparent list of inventory data associated with all stages from the wine life cycle covering also a large number of problems such as the toxicity related impacts.