A Network Analysis for Environmental Assessment in Wine Supply Chain

: In the agri-food sector, the Life Cycle Assessment method (LCA) is used to evaluate the environmental impact of a product. Within agri-food products, wine is among the most analysed products, not only for its economic importance but also for the environmental impact of its activity. The paper aims to identify the main trends in the wine sector revolving around environmental evaluation using the LCA method in the academic literature. The aim is to investigate the literature on life cycle assessment analysis of grape and wine production through the systematic grouping of papers into clusters of research. So, the purpose is to discuss the gaps and insights identiﬁed by the study in order to aid in the development of a comprehensive state of the art on the topic. Scopus and Web of Science were used to search all articles following a clear and replicable protocol. The results (keywords) were subjected to co-occurrence analysis using VOSviewer, after which the articles were further analysed. Through a bibliographic coupling analysis, the research results were grouped through a network analysis that allowed identifying the research trends on the topic. Three clusters were identiﬁed containing the main lines of research on the subject. The results show that nowadays the literature is focusing on concerns related to climate change and consumer awareness on sustainability issues and certiﬁcations as well as environmental impacts generated mainly in the production phase in the vineyard. The research results are of interest for future research on LCA analysis in the wine sector in order to contribute to the discussion on the current model in the global wine sector.


Introduction
Environmental sustainability in agricultural production is one of the main and debated areas of discussion in the academic literature [1,2]. Among agri-food products, wine is among the most analysed products, not only for its economic importance [3] but also for the environmental impact of its activity. The wine supply chain can be described through different phases, from the cultivation of the vine and grape production, to the transformation of the product during winemaking, transport and distribution, up to the production of by-products and waste, relevant in a circular economy perspective [4].
Given the importance of viticulture in the characterization of the territory in which it is practiced, the question is of great importance and needs further investigation. From the point of view of the multifunctionality of agriculture, it is necessary to evaluate the ecosystem function of the vineyard and not only the merely productive and therefore economic function. In this context, an economic assessment cannot be distinguished from an environmental sustainability assessment [1,2]. The main gaps in the research on wine analysis of the environmental impacts of grapes and wine production. We performed a bibliographic correspondence analysis by identifying three groups of search lines. Then, we analysed each research cluster and opportunities for new emerging research trends still under development. Our research question focuses on the contents of the scientific literature, with the aim of investigating emerging issues in the field of environmental analysis in the wine sector. The overall goal is to provide evidence-based insights for researchers, policy makers and stakeholders in the wine supply chain interested in the topic of environmental assessment in the context of wine.
The article is organized as follows: The first section deals with the introduction, Section 2 describes the method; Section 3 discusses the findings of the literature review; the discussion of the results is described in Section 4; finally, the concluding observations and limitations of this study are presented in Section 5.

Literature Searching Procedure
The review was conducted to analyse academic literature from scientific databases and to summarize the main trends of environmental assessment in the wine and grape sector. A systematic literature review was carried out based on strict keyword search criteria. The research was carried out in October 2021.
The review was carried out following the Preferred Reporting Items for Systematic Review and meta-Analysis (PRISMA) protocol [2,43,44]. The PRISMA procedure provides a detailed, replicable, scientific and transparent protocol [45]. The research was conducted through the online core collection of Elsevier Scopus and Web of Science databases, which are the main online scientific research databases widely recognized by the scientific community for the collection of reliable and multidisciplinary research. In Figure 1, a flowchart with the selection procedure and exclusion criteria [2,44,46,47] is presented. The thematic research area was identified in the LCA analysis in the wine sector with reference to both the agricultural phase of cultivation of the raw material (grape) and the vinification and transformation process of grapes (wine) (Identification phase). The literature searching procedure was conducted by a combination of keywords in the databases. The keywords "wine" OR "grape" AND "lca" OR "life cycle assessment" OR "life-cycle-assessment" were used. The same search query was used on both databases, and the same criteria to ensure the completeness of the data were chosen. The following string was used: ("wine" OR "grape") AND ("lca" OR "life cycle assessment" OR "life-cycle assessment").
In the first phase of Identification (Figure 1), the search for keywords initially produced a total result of 397 records, of which 181 were identified through Scopus and 216 through the Web of Science database. The next phase for the selection of the relevant literature took place through the Screening and Eligibility [2,44,46,47]. In the Screening phase, applying the primary exclusion criteria-in this bibliographic search phase, only academic articles published in indexed journals were included-the selected articles identified by Web of Science and Scopus were reduced from 397 to 359 records. Only articles and reviews were considered for research purposes; 38 records were eliminated in this phase, including Books, Chapters, Proceedings, Editorials and Reports. Subsequently, in the Eligibility phase, the duplicates of the articles that came from both databases were eliminated, and therefore the number of articles for this study was reduced from 359 to 230, thus excluding 129 duplicate records. In order to consider the literature as highly visible within the scientific community, 4 non-English articles were excluded during the subsequent Eligibility phase [2,4,44]. In the last phase of the Included stage, a sample of 226 documents was selected to answer our research question.

Bibliographic Analysis of Co-Occurrence Keywords
The results obtained from both scientific research databases were analysed through a co-occurrence analysis method using the VOSviewer software. This JAVA-based software was established in 2010 by Van Eck and Waltman in the Centre for Science and Technology Studies (CSTS) of the University of Leiden as an aid in designing maps created on network data [48]. This tool develops network analyses by elaborating bibliometric maps [49] through a quantitative method, which combines several factors, provide the visual graphical visualization of various forms of data network of scientific publications.
In this phase, keywords, article titles and abstracts (TITLE_ABS_KEY) of the articles previously collected within the scientific databases, and then selected through the PRISMA protocol, were initially loaded into the VOSviewer software. Using the VOSviewer software, maps of the co-occurrence network of the keywords of all the articles selected from the databases under examination were drawn up in the research period available on the databases (1996-2021). VOSviewer is software used to identify and explore bibliometric maps and matches of co-citations and co-occurrence keywords [50,51] and to create distance maps indicating the strength of the relationships between the elements [52]. The software uses two standard weights, such as the number and total strength of the links, to graphically visualize the nodal network [53].
The statistical analysis of keywords using the VOSviewer tool allows to identify the most used and recurring terms in the academic literature and their relationships, from which it is then possible to derive the main research topics relating to the field analysed but also the research topics that are more recent and still under explored, thus allowing assessments to be made not only of current research but also possible predictions on the themes of possible future trends [54]. The file containing the information of the articles selected for the study was imported into VOSviewer to develop a single network analysis of the three research data above (TITLE_ABS_KEY), based on their relevance and co-occurrence [55]. The "map based on bibliographic data" option was preferred; "co-occurrence" was chosen as the type of analysis; the unit of analysis identified was that of the "keywords"; the counting method preferred was the "full count"; the minimum number of occurrences of a term was 4; and the Number of terms designated was the total number of items, in line with other authors [55,56]. This selection was selected to guarantee higher accuracy in the examination of the results. Several units of analysis can be used in this type of analysis, including journals, publications and authors [48]. The present paper used "publication" as a unit of analysis [57].

Overview of Selected Papers
Information regarding the title, author(s) and year of publication of the selected papers are summarized in Appendix A (Table A1).
The Figure 2 shows the development of academic literature in the period 1996-2021, taking into consideration the number of articles per year. Generally, before 2011, there were fewer than five papers per year. Since 2012, the trend has always been constantly growing. During the first years, from 1996 to 2009, there was a low interest in these issues, with an average of one publication per year. Despite a general upward trend in the number of publications, there was a decline in the number from 19 to 16 in 2017. The peak has been reached in recent years, especially in 2020, with 38 published papers. This shows that interest in this research has increased.
In Figure 3, journals containing at least two papers are represented. Most of the papers are published in the Journal of Cleaner Production with 64 total articles, followed by the International Journal of Life Cycle Assessment with 23 articles and Science of the Total Environment with 19 articles, Sustainability with 15 papers, the Journal of Environmental Management (7) and the Journal of Industrial Ecology (4). All other journals achieved lower values, as mentioned in Figure 3. During the first years, from 1996 to 2009, there was a low interest in these issues, with an average of one publication per year. Despite a general upward trend in the number of publications, there was a decline in the number from 19 to 16 in 2017. The peak has been reached in recent years, especially in 2020, with 38 published papers. This shows that interest in this research has increased.
In Figure 3, journals containing at least two papers are represented. Most of the papers are published in the Journal of Cleaner Production with 64 total articles, followed by the International Journal of Life Cycle Assessment with 23 articles and Science of the Total Environment with 19 articles, Sustainability with 15 papers, the Journal of Environmental Management (7) and the Journal of Industrial Ecology (4). All other journals achieved lower values, as mentioned in Figure 3.
In reference to the nationality of the authors of the studies, most of these are located in Europe, in particular in Italy and Spain, but widespread are authors from North and South America, Asia, South Africa and Australia. Specifically, Europe is the most productive continent. The country with the largest number of papers is Italy (68), followed by Spain (55), France (24), the United Kingdom (20), the Netherlands (13), Portugal (10), Germany (9), Sweden (8), Peru (8), Greece (6) and Luxembourg (6). These countries compete with Brazil (12), the USA (10), Canada (10), Australia (8) and China (7). The other countries achieved a record of less than five papers ( Figure 4). During the first years, from 1996 to 2009, there was a low interest in these issues, with an average of one publication per year. Despite a general upward trend in the number of publications, there was a decline in the number from 19 to 16 in 2017. The peak has been reached in recent years, especially in 2020, with 38 published papers. This shows that interest in this research has increased.
In Figure 3, journals containing at least two papers are represented. Most of the papers are published in the Journal of Cleaner Production with 64 total articles, followed by the International Journal of Life Cycle Assessment with 23 articles and Science of the Total Environment with 19 articles, Sustainability with 15 papers, the Journal of Environmental Management (7) and the Journal of Industrial Ecology (4). All other journals achieved lower values, as mentioned in Figure 3. In reference to the nationality of the authors of the studies, most of these are located in Europe, in particular in Italy and Spain, but widespread are authors from North and

Bibliometric Analysis of the Themes
After the selected papers have been collected, VOSviewer software has been adopted to visualize the networks among the data by creating graphical bibliometric maps [58].
The VOSviewer tool provides the essential functionalities to visualize bibliometric

Bibliometric Analysis of the Themes
After the selected papers have been collected, VOSviewer software has been adopted to visualize the networks among the data by creating graphical bibliometric maps [58].
The VOSviewer tool provides the essential functionalities to visualize bibliometric networks and co-occurrence links between keywords [50]. VOSviewer is a computer program developed to generate and investigate bibliometric maps [50]. Keywords contained in the article titles and abstracts are explored based on their occurrence to create a map of the occurrences of all terms used in the 226 selected papers. We built a map of the most frequent keywords. We have selected "co-occurrence" as the analysis type, "all keywords" as the unit of analysis, and "four shared keywords" as the minimum level. Then, VOSviewer converted data into a graphic form and categorized frequent keywords into three main clusters in the network view visualization [59]. Larger circles and map labels explain greater importance and meaning. Similarly, colour keywords belong to the same cluster [50] ( Figure 5). In the graphic maps, larger circles imply higher relevance for a theme [50]. In Figure 5, red, blue and green colours differentiate the clusters. All three clusters corroborate the research flows obtained from the bibliographic coupling. Nodal outputs based on bibliometric analysis aid to understand the proximity and relevance of keywords and discover possible gaps and insights. The size of the circles represents how often keywords are displayed. The distance between the two circles indicates their correlation. In this study, the keywords in the title and abstract of selected papers were included in the final analysis. Three main clusters have been created, which means groups of themes. The first cluster was designated by the red colour and covered 13 elements. The second cluster was expressed by the green colour and involved 11 elements. The third cluster includes 8 elements in blue, as shown in Figure 5.
A keyword cloud was also created to show the frequency and interconnections of keywords that occurred more than four times in the papers selected for this research. "Life cycle assessment", "environmental impact", "wine", "carbon footprint", "sustainability" and "consumption" and "greenhouse gas emission" seem to be much-studied topics. The strengths of low keywords indicate that more research is needed, as this is still an emerging stage.
It has been indicated that "life-cycle assessment" is also the most frequent keyword (with 190 occurrences), followed by "wine" (with 88 occurrences) and "environmental impact" (with 66 occurrences) and "carbon footprint" (with 51 occurrences). All other keywords scored lower. In this study, the keywords in the title and abstract of selected papers were included in the final analysis. Three main clusters have been created, which means groups of themes. The first cluster was designated by the red colour and covered 13 elements. The second cluster was expressed by the green colour and involved 11 elements. The third cluster includes 8 elements in blue, as shown in Figure 5.
A keyword cloud was also created to show the frequency and interconnections of keywords that occurred more than four times in the papers selected for this research. "Life cycle assessment", "environmental impact", "wine", "carbon footprint", "sustainability" and "consumption" and "greenhouse gas emission" seem to be much-studied topics. The strengths of low keywords indicate that more research is needed, as this is still an emerging stage.
It has been indicated that "life-cycle assessment" is also the most frequent keyword (with 190 occurrences), followed by "wine" (with 88 occurrences) and "environmental impact" (with 66 occurrences) and "carbon footprint" (with 51 occurrences). All other keywords scored lower.
The keywords that have appeared the most and with stronger interconnected links are "life-cycle assessment" (total connection strength 633), "environmental impact" (total connection strength 302) and "vineyard" (total connection strength 302), which had a strong link with "carbon footprint" and "wine".
The literature published on the measurement of the environmental impacts of the wine production chain using LCA methodologies includes the following three aspects: wine industry (red), agricultural phase (blue) and winemaking phase (green). Through the analysis of keywords and the analysis of their recurrence, it was possible to identify three clusters of main research topics ( Figure 5).
The (1) first cluster, in red colour, involved documents that study the environmental impacts in the wine industry, referring to sustainability indicators in the sustainable management of the wine industry, as well as the measurement of impacts along the industry and consumer demand for sustainability certifications.
The (2) second cluster, in blue colour, contains articles that investigate the environmental impacts in the agricultural phase, with a particular focus on the vineyard management and, therefore, the analysis of the life cycle delimited to the agricultural phase for the grape production.
The (3) third cluster, in green colour, concerned environmental assessment documents concerning the life cycle analysis in the vinification phase for the wine production, with particular attention to the materials used for packaging but also to the generation of products and waste.
In Figure 6, the temporal evolution over the years of the topics investigated in the literature and the keywords used are reported. The most used keywords starting from the first half of 2016 were "assessment lca", "life cycle analysis", "food", "wine industry" and "emission". These words indicate that the lines of research in 2016 were mainly dedicated to the literature on the LCA methodology in the wine industry and the consequent measurement of the impacts of this with particular interest in the emissions generated. The (2) second cluster, in blue colour, contains articles that investigate the environmental impacts in the agricultural phase, with a particular focus on the vineyard management and, therefore, the analysis of the life cycle delimited to the agricultural phase for the grape production.
The (3) third cluster, in green colour, concerned environmental assessment documents concerning the life cycle analysis in the vinification phase for the wine production, with particular attention to the materials used for packaging but also to the generation of products and waste.
In Figure 6, the temporal evolution over the years of the topics investigated in the literature and the keywords used are reported. The most used keywords starting from the first half of 2016 were "assessment lca", "life cycle analysis", "food", "wine industry" and "emission". These words indicate that the lines of research in 2016 were mainly dedicated to the literature on the LCA methodology in the wine industry and the consequent measurement of the impacts of this with particular interest in the emissions generated. Trends have changed since 2017, and the most frequent keywords in the literature have been "life-cycle assessment", "wine", "vineyard", "climate change", "carbon footprint", "environmental impact", "energy", "packaging" and "product". This indicates that literature trends have shown greater attention to issues relating to climate change and  Trends have changed since 2017, and the most frequent keywords in the literature have been "life-cycle assessment", "wine", "vineyard", "climate change", "carbon footprint", "environmental impact", "energy", "packaging" and "product". This indicates that literature trends have shown greater attention to issues relating to climate change and environmental impacts mainly linked to the agricultural phase of production and product sustainability certifications. Finally, the most recent keywords in the literature starting from 2018 concerned "sustainability", "industry", "management", "water footprint", "consumption" and "footprint". This indicates that the issues starting from 2018 were mainly addressed to the sustainable management of the wine industry and consumer awareness towards sustainability issues.
The colours suggest the density of the terms, varying from green (lowest density) to yellow (highest density). As can be perceived in Figure 7, prominent search terms include "life-cycle assessment", "wine" and "environmental impact". Among the slightly less prominent terms, we also observe various terms such as "carbon footprint", "impact", "consumption" and "sustainability". The main cluster (marked in red) has 13 recurring keywords, as shown in Figure 8. The second cluster (green) has 11 recurring keywords. The third cluster (blue) includes 8 recurring keywords. We also found that the first cluster is the largest in terms of the number of items it contains and the most recent in terms of the topics it contains. The second cluster is relatively small, and the most prevalent and most frequent keywords are placed in this cluster. The third cluster was found to be smaller than the others, as it is probably in the incubation phase, but it has its important influence on sustainable grape and wine production.  The main cluster (marked in red) has 13 recurring keywords, as shown in Figure 8. The second cluster (green) has 11 recurring keywords. The third cluster (blue) includes 8 recurring keywords. We also found that the first cluster is the largest in terms of the number of items it contains and the most recent in terms of the topics it contains. The second cluster is relatively small, and the most prevalent and most frequent keywords are placed in this cluster. The third cluster was found to be smaller than the others, as it is probably in the incubation phase, but it has its important influence on sustainable grape and wine production.
As for the occurrences of keywords per cluster, the highest occurrences are placed in the second cluster. The keyword with the highest value of 190 is "life-cycle assessment", followed by "wine" with a value of 88. Both keywords are placed within the second cluster (green). Other significant keywords are placed in the first cluster: "environmental impact" with 66 values, "carbon footprint" with 51 values, "consumption" with 36 values and "sustainability" with 35 values. In the second cluster, the keywords "impact" with 36 values and "greenhouse gas emission" with 33 values also stand out. Lower values were obtained for all other keywords (Figure 9).
The main cluster (marked in red) has 13 recurring keywords, as shown in Figure 8. The second cluster (green) has 11 recurring keywords. The third cluster (blue) includes 8 recurring keywords. We also found that the first cluster is the largest in terms of the number of items it contains and the most recent in terms of the topics it contains. The second cluster is relatively small, and the most prevalent and most frequent keywords are placed in this cluster. The third cluster was found to be smaller than the others, as it is probably in the incubation phase, but it has its important influence on sustainable grape and wine production. As for the occurrences of keywords per cluster, the highest occurrences are placed in the second cluster. The keyword with the highest value of 190 is "life-cycle assessment", followed by "wine" with a value of 88. Both keywords are placed within the second cluster (green). Other significant keywords are placed in the first cluster: "environmental impact"      Table 1 summarizes the 3 most relevant keywords in this study. The keyword "Lifecycle assessment" is confirmed as the most important word. This belongs to the second cluster (green), with 31 links, 633 total link strength and 190 occurrences. The second most important keyword is "environmental impact" belonging to the first cluster (red), with 31 links, 302 total link strength and 66 occurrences. Finally, the third most important word is "vineyard" belonging to the third cluster (blue) with 28 links, 122 total link strength and 23 occurrences.

Overall Comment
Interest in the issues of environmental sustainability is increasingly widespread in the agri-food sector and in particular in the wine sector. In this study, through an analysis of the literature and a bibliometric analysis of the themes, their interconnections and new research topics, it has been possible to deepen the main problems studied and those recently emerging in the academic literature on the theme of the sustainable management of production in viticulture. Among the agri-food products, the wine industry arises as one of the main sectors in which environmental assessment is most used, given its importance from an economic point of view but also for its generation of outputs such as emissions, by-products and waste. The analysis of the bibliometric network provides an overview of the main aspects that characterise the problem of environmental assessment in the wine sector, allowing the investigation of the relationships between keywords. The integration of the analysis of social networks and bibliometric science has led to a useful approach capable of capturing the multidimensional nature of environmental assessment by analysing a large amount of literature data. The maps of the network show that environmental assessment is a complex issue that affects multiple research fields. The keyword map of the co-occurrence network has shown that the main focus of research in recent years has shifted towards greater attention to issues related to climate change and consumer awareness on the issues of sustainability, certifications and related environmental impacts to the agricultural production phase. Through the analysis of the keywords and the analysis of their cooccurrence, it was possible to identify three main clusters of research topics: environmental impacts in the wine industry, the agricultural phase and the winemaking phase.

Environmental Impact in the Wine Industry
Viticulture, distribution and packaging of wine are the main environmental impacts of the wine industry [60].
The interest of producers is increasingly aimed at reducing the environmental impact of their production through environmental strategies, for example, through ecolabeling [61,62] or eco-design for the supply of novel environmentally friendly products [63].
Today, in the wine sector, the LCA tool is of great interest to producers as it is used for the assessment of environmental loads along all wine production processes, to consent the identification and reduction of hotspots in the life cycle of the wine production and to transmit this information for communication purposes to the consumer [64,65].
The use of an environmental assessment of emissions is indirectly able to favour a more equitable and transparent distribution of the responsibilities of the environmental loads produced in a process amongst the various stakeholders in the wine chain [66,67]. Most of the studies in the literature aim to identify the critical points of the wine life cycle; some authors calculate an extensive detailed choice of LCA impact categories, others are oriented towards the design of new indicators for the wine sector [68] and others turn to specific single indicators as in the case of CF assessment [65,69].
In the literature, there are numerous studies that use the LCA methodology, which, thanks to the specific and detailed impact categories, are able to evaluate the environmental loads linked with different phases of the wine life cycle [63,65]. However, in order to make the results of LCA studies available to supply chain stakeholders and policy makers [70], the need to develop specific indicators, such as carbon footprint (CF) and water footprint (WF), has emerged in life cycle studies [71,72].
Indicators such as the CF and WF are increasingly widespread, which aim to assess the environmental impact related to the production and consumption of agri-food products [73,74]. The CF and WF analyses of the products are developed with an LCA analysis, which permits the assessment of the impacts "from the cradle to the grave", according to the requirements of the respective international reference standards [12,13]. However, the function of a single indicator method, as in the case of the WF or the CF, in the dissemination of results has limits and risks associated with the representation of a single environmental impact [75,76]. This limit is much debated in the literature, which has overcome it as a single indicator, such as the CF, can still represent other underlying environmental impacts as it is strongly linked to the use of energy [76]. In the wine sector, the CF indicator is closely related to several environmental concerns and management processes [77].
Numerous articles are available in the literature that focus on conceptual and methodological aspects [78] or that use these indicators specifically in the wine sector [79,80]. As regards the WF, several case studies exist in the literature with a focus on both grape-wine production [81] and on the bottle of wine [82].
In the literature, the analysis of the CF is acquiring a role of great interest in the issues of sustainability in the wine sector [77]. This indicator offers both producers and consumers the opportunity to reduce uncertainties and information asymmetries within an increasingly globalized wine market [77]. The CF indicator, in fact, quantifies the greenhouse gas emissions that are the basis of the wine life cycle, taking into consideration all the phases of the product, from the agricultural phase of viticulture, to the transformation during winemaking, to the bottling, distribution and production of waste and by-products [77].
The literature analysing the CF is quite large and diverse [79,80,83], including both studies that perform a complete LCA analysis of a bottle of wine [74] and studies that instead focus on single stages of the production process [64,65,69,[84][85][86][87][88], as well as others studying supply chain analysis [60,89] and others who instead carry out comparative analyses between the management of conventional and unconventional viticulture [90]. Furthermore, single problem approaches are commonly used in the literature, but a more comprehensive analysis should be employed instead, as a single indicator does not adequately track the pressure on the environment [74].

Environmental Impacts in the Agricultural Phase
The literature highlights the need for particular attention to vineyard management and therefore to the analysis of the life cycle delimited to the agricultural phase for grape production [15][16][17][18][19][20]. The research on the adoption of LCA in the wine sector has shown an unequivocal environmental hotspot in the agricultural phase [2] due to the wide use of different agronomic techniques [15][16][17][18][19][20], which require an extensive use of inputs [21].
The viticulture phase presents the widest range of variation in results among the impact categories considered. However, four categories of environmental impact are considered in all LCA studies in the wine sector: Carbon Footprint (CF), Abiotic Depletion (AD), Acidification Potential (AP) and Eutrophication Potential (EP). EP is the only impact category for which viticulture is the most impacting phase from an environmental point of view [22]. According to the literature on the analysis of the life cycle assessment in the wine sector, the planting phase of the vineyard is the one that represents the greatest impact on the CF of the wine and, on the contrary, the pre-production phase does not generate a significant impact in this sense [73]. The use of nitrogen fertilizers and N 2 O emissions are the factors that most determine the carbon footprint in the agricultural phase [73].
The review of the studies in the literature indicates the main wine hotspots in the agricultural phase of the vineyard due to the high inputs used in the vineyard (fuel, fertilizers and pesticides) [77]. Some authors have compared the management of wine production by comparing organic and conventional treatment [64,65,[91][92][93][94][95], finding that the CF indicator value for organic wine "from the cradle to the door" is about 25% lower than for conventional wine [77].
Organic or biodynamic cultivation systems can significantly decrease the environmental impacts of viticulture [90] because they avoid the use of synthetic fertilizers and pesticides; moreover, the application of organic fertilizers improves the soil structure and closes the cycle of biological matter [91]. However, the results in the literature are not unanimous, both because viticulture is the phase with the widest variability of results and because the results are influenced by the definition of the functional unit. For example, the study carried out by Falcone et al. [92] shows that the combination of conventional practices with the espalier training system allows the best environmental performance thanks to the higher yield per hectare.

Environmental Impacts in the Vinification Phase
The environmental assessment in the vinification phase takes into consideration all the steps necessary for the wine production: transport of the grapes to the cellar; destemming and crushing; fermentation and filtration. Although the major environmental hotspot is in the agricultural phase, however, the transformation phase of wine production also generates considerable environmental impact, mainly due to the packaging materials used but also to the generation of products and waste [22]. Packaging production is the most impactful phase of the wine life cycle. The studies examined specified that the highest impact value (for all impact categories considered) was due in particular to the production of primary packaging production of glass bottles. Furthermore, the differences in the winemaking processes of the different types of wine in the studies examined do not determine high variations in environmental impacts to compete with the impacts produced by the other phases [22].
As regards the environmental performance of the types of white and red wine in terms of their carbon footprint (CF), the results of the studies are affected by a wide variability, and therefore it is difficult to define which type of wine is more sustainable from an environmental point of view. The results of the literature in this sense are not unanimous. On the one hand, some authors found that white wines perform slightly better in terms of CF than red wines thanks to the use of wooden barrels for aging during the storage period in the vinification phase [91]. The white winemaking phase, on the other hand, usually requires a process at a lower temperature, and this involves greater energy consumption. In light of this, this aspect requires further scientific investigation.
Logistics is also energy-intensive in the wine industry [89]. As noted in the literature, the emissions related to the delivery of wine are a significant part, but only a part of the total carbon emissions produced by a company [89]. With the growing attention on reducing greenhouse gas emissions, wineries are facing increasing pressure to demonstrate their commitment to minimizing their CF, also encouraging consumers to consider their own contributions [89]. Concerning the economic implications of introducing a CF label, it is necessary to consider the consumer's awareness of this information and his purchasing behaviour for products that present such an indication of sustainability, which can lead to a change in the perceived qualitative characteristics of the product. On the one hand, the presence of these environmental indicators constitutes an element of differentiation that can lead the consumer to gain more awareness towards this quality attributes [96]. In this context, future studies should assess whether there exists a premium price for wines with a sustainable label. One of the conditions for eco-labels to be effective in the market is that consumers are willing to pay a premium price to support the higher costs of more sustainable management those producers are called to sustain [97].

Conclusions
Environmental assessment in the wine sector offers benefits to both producers and consumers in an increasing globalization of the market. Specific indicators, such as the quantification of greenhouse gas emissions generated throughout the life cycle of wine, are increasingly required by the industry, which must respond to new and increasingly specific needs of consumers, who are increasingly attentive to sustainability issues in their decisions of purchases.
The present study offers a critical analysis of the literature on the use of the LCA method to evaluate the sustainability of the wine sector in order to highlight the main research areas, the trends of recent years and the new research trajectories on the subject.
The results of this study show that today the focus of the literature is on issues related to climate change and consumer awareness on sustainability issues and certifications, as well as environmental impacts generated mainly in the phase of agricultural production.
This bibliometric analysis provides key information for further developments in academic literature in the LCA topic in the wine supply chain. Furthermore, it has been found that this field is still underdeveloped, and scarce research has been conducted yet, so the topic needs further research. Moreover, another avenue for the academic world could be to compare LCA in viticulture and wine management with other marketing hypotheses, such as consumer involvement in LCA issues and the impacts generated in the wine supply chain. These concepts have been well considered in marketing [98,99], but research on this topic is still scarce.
The main limitation in this paper is due to the mapping citations and analysing clusters technique because one of the criteria was to include in the analysis only documents with at least four citations; therefore, current studies could be included in this research, regardless of their relative importance. For this reason, we suggest repeating the bibliometric analysis in the future to include also recent research to contribute to the discussion on the topic. Bibliometric analysis also has the limit of being too sensitive to the construction of the search query and to the selection of the data source, which limits it to a use mainly of an exploratory nature. Furthermore, the literature searching criteria, applied in the PRISMA procedure has inevitably excluded some papers on this topic.
Among the limitations of the study, it is important to highlight that the selected documents show a widespread variability in the definition of the system boundaries, and a lack of accessibility of original and site-specific inventory data in the LCA methodology. Furthermore, environmental impact values in the vineyard stage present a high variability, mainly due to the different types of production methods. The results could be influenced by other technical factors such as the grape variety, the pedoclimatic conditions, the chemicalphysical characteristics of the soil and the climate. These variables can have a great influence on the results and may constitute a limitation of the present study and require further scientific input through future studies.
Future studies could investigate the environmental assessment linked to an intensification in the recycling rate of glass used for bottling but also the development of innovations such as lighter glass bottles or even the replacement of glass with different packaging materials that are at the same time suitable for preserving the quality of the wine. Ultimately, the attention to the sustainability of the wine sector for environmental assessment could therefore shift to the packaging phase that needs further investigation in the future. These indications suggest that future studies on the subject are needed.