2.1 Data Collection

We conducted a search in the Scopus electronic database. We chose the Scopus database because it is considered the largest citation database of peer-reviewed literature (^{Aldieri et al., 2019}; 2018; ^{Randhawa et al., 2016}). The following combination of keywords was used: “environmental sustainability” AND (“universit*” OR “higher education” OR “tertiary education” OR “post-secondary”). The environmental sustainability search was conducted in the titles, abstracts and keywords of articles while the terms related to higher education were searched only in titles, in order to avoid bias caused by authors’ institutional affiliation. As supported by ^{Raban and Gordon (2020)}, we believe that searching in titles can provide a reasonable and representative sample of the field of study.

Initially, we retrieved 191 publications, from which we selected only those documents published between 1989 and 2019. Then we further restricted the sample to include only scientific articles - on the basis that they have clearer peer-reviewed procedures - that were published in English. The final sample comprised 105 articles (Table 2). The reference information of the publications was imported to Mendeley and downloaded in CVS format, as required by VOSviewer, the tool used to analyze and visualize bibliometric networks.

2.2 Data Analysis

To analyze the publications retrieved, we followed the same procedure performed by ^{Dias et al. (2019)} and ^{Terán-Yépez et al. (2020)} to analyze scientific publications. Firstly, we present a descriptive analysis of the results, analyzing the number of articles by year of publication and determining the leading journals and most productive countries.

Secondly, we performed a keyword co-occurrence analysis in order to identify the key research themes from 1989 to 2019. This type of analysis is also known as keyword network analysis or co-word network analysis. To analyze keyword network structures, items and ties (links) must be identified. Items are the individual components - in our case, the keywords provided by authors - while ties are the interrelationships among the components (keyword co-occurrences) (^{Yoo et al., 2019}). Items and ties together constitute a network.

Furthermore, between items there can be no more than one link. Each link has a positive numerical value called strength. The higher this value, the stronger the link (^{Van Eck & Waltman, 2017}). The strength indicates, for instance, the number of publications in which two terms or keywords occur together. In short, a network is a set of items together with the links between the items. A set of items can form clusters. In VOSviewer an item may belong to only one cluster, with clusters labeled using cluster numbers.

To visualize a cluster in a map, we have to consider the weight, which indicates the importance of the item. Items with a higher weight are considered more important than an item with a lower weight. In a keyword co-occurrence map, items with a higher weight are shown more prominently and the color of an item is determined by the cluster to which it belongs. The weight of the item also determines the size of the label and the circle of an item (^{Van Eck & Waltman, 2017}).

Additionally, we performed a bibliometric analysis called bibliographic coupling to select the articles that were content-analyzed. This method is an effective way to group literature trends graphically based on the co-occurrence of references between two documents. A bibliographic coupling analysis is a citation method that allows researchers to evaluate the degree of similarity between two documents, based on the co-occurrence of their references. The main advantage of this technique, when compared with co-citation or direct citation methods, is that it allows recently published articles to be included (^{Lerena et al., 2019}).

Each article was then read in full and categorized into a cluster according to the topic addressed and the subject of discussion. Thus, cluster 1 includes 12 documents, cluster 2 includes 11 articles, and cluster 3 includes 9 studies. The full-text analysis of the articles in Atlas.ti enabled a broader discussion of the contributions of each study and the similarity between studies.

3.1 Evolution of Scientific Production

A review of scientific production provides insight into the evolution of the literature of any field of research (^{Iturralde et al., 2017}; ^{Navarro et al., 2017}; ^{Terán-Yépez et al., 2020}). In that sense, in this first subsection we present descriptive results on the number of articles published by year and determine the leading journals and most productive countries.

Figure 1 shows the number of articles on university environmental sustainability published between 1989 and 2019. Two main periods can be identified: the first from 1994 to 2011, when on average 1.3 articles were published per year, and a second period, between 2012 and 2019, when this figure jumps to an average of 10 articles. We note that the number of articles about environmental sustainability in the context of higher education has increased sharply in the last five years (Figure 1). This is due to increased attention given to serious environmental concerns and the pressure that universities face to implement institutional policies to meet these challenges.

Figure 1 Number of articles published by year 

On the other hand, results show that most articles concern the areas of social sciences; environmental science; energy; and business, management and accounting. These four areas make up 79% of publications. Moreover, the leading journals in university environmental sustainability research are the Journal of Cleaner Production, the International Journal of Sustainability in Higher Education, and Sustainability (Table 3).

The first of these journals covers topics such as sustainable consumption, environmental and sustainability assessment, and education for sustainable development. The second has a clear focus on new advances in sustainability in a higher education context, including topics such as curricular innovation, energy, water, recycling, and waste management. The journal Sustainability, meanwhile, focuses specifically on research on challenges relating to sustainability and socioeconomic, scientific and integrated approaches to sustainable development.

The United States accounts for the highest share of publications on environmental sustainability in higher education (30.1%), followed by the United Kingdom and Australia. Only one Latin American country, Brazil, appears in the list of leading countries (Table 4).

3.2 Influential Keywords in University Environmental Sustainability Research

A keyword co-occurrence analysis was performed in order to identify the main topics in environmental sustainability research in a higher education context. We selected a threshold of 6 occurrences, which left 24 keywords. They keyword “article”, which has no clear meaning, was then excluded from the network, and thus ultimately the 23 most relevant terms were extracted (Table 5). This resulted in a map of associations between terms and their relationships in thematic clusters (^{Dias et al., 2019}). Three thematic areas were identified: environmental education, environmental sustainability integration, and environmental protection.

The whole keyword co-occurrence network is represented in the following map (Figure 2):

Figure 2 Keyword co-occurrence map 

4.1 Environmental Education

Sustainability education is becoming one of the main missions of universities. Higher education institutions are making increasing efforts in terms of learning, research and operations with the aim of educating for sustainability (^{Cortese & Hattan, 2010}). However, in practice, these efforts do not always produce the expected results because students and graduates know little about how to align their personal and professional goals with environmental sustainability (^{Cortese & Hattan, 2010}).

The first cluster of this study focuses on curriculum, environmental education, and education for sustainable development. In a study that set out to determine how to promote environmental education for sustainable development by linking higher education and non-governmental organizations (NGOs), ^{Haigh (2006)} draws our attention to the fact that environmental sustainability education is not only a curricular issue. According to ^{Christie et al. (2015)}, most efforts by universities in environmental sustainability education have focused on management and operations rather than teaching and learning. This may be due to the fact that in many cases, environmental sustainability on campus has been motivated by financial incentives or regulatory pressure (^{Christie et al., 2015}).

Environmental education can be beneficial to sustainable development when it is addressed as a process that goes beyond classrooms and is not restricted to formal education, and when it becomes a lifelong process that is not simply limited to the understanding of environmental sustainability issues. Therefore, the main challenge for environmental sustainability education is pushing the boundaries of formal academic education and achieving an impact on the environmental awareness of not just all students, but all people (^{Haigh, 2006}). Effective practices, for example, include enabling students to conduct research on global problems, case studies, and current sustainability decisions, so that they stay up-to-date in the face of an ever-changing world (^{Dmochowski et al., 2016}). However, when designing strategies, it is important to consider that students may have different socioeconomic, cultural and ethical backgrounds and different academic performance, which may condition their expectations in terms of learning environmental sustainability (^{Foster, 2010}). In the case of teachers, it is important to constantly motivate them to think about their courses and their relationship with sustainability. It is also important to promote collaboration with other teachers and the exchange of academic texts to improve teaching practice (^{Dmochowski et al., 2016}).

Other authors like ^{Clark et al. (2011)} analyze the purpose and design of environmental programs. They point out that often these programs fail in their aim to develop training activities with students to enhance their abilities as problem solvers due to three weaknesses: 1) a lack of clarity about the core goals of environmental programs, leading the participants to replace them with goals that may not be relevant; 2) the poor alignment of multiple academic disciplines into environmental programs, due to ineffective frameworks and communication; 3) a diverse array of perspectives, pedagogies, methodologies, and courses, which makes it even more difficult for students to acquire, integrate and apply knowledge about environmental issues.

^{Pesonen (2003)} emphasizes the importance of including environmental aspects not only at the undergraduate level but also higher levels of training like master's degrees. For example, for a corporate environmental management course, ^{Pesonen (2003)} argues that it is important to consider at least three components: 1) identifying the environmental impacts of organizations and products; 2) environmental impact management; and 3) creation of eco-competitiveness. Thus, as can be seen, in this cluster, curricular issues are one of the main topics addressed. In the same vein, ^{Hurlimann (2009)}, referring specifically to disciplines such as urban planning, maintains that students wish to be trained in at least the following aspects: open space planning, sustainable transportation planning, energy efficiency, and environmental systems. Paradoxically, students consider that the climate change adaptation and mitigation teaching they receive is inadequate, which points to a need for increased awareness-raising given that mitigation and adaptation are two fundamental pillars for tackling climate change.

Finally, other articles selected for analysis in this cluster focus on the conceptualization or operationalization of sustainability in the context of higher education. ^{Aleixo et al. (2018)} stress that sustainability in universities can be developed from different areas: campus operations, community engagement, institutional frameworks, on-campus experiences, research, education, and assessment and reporting. For ^{Marta et al. (2018)}, the integration of leaders, faculty, administrative staff, students and external stakeholders should be a fundamental part of any university environmental sustainability strategy. On the other hand, ^{Wright and Wilton (2012)} found, in a sample of facilities management directors, that the term environmental sustainability is generally associated with minimal impact generation on the environment and reduced use of natural resources. Understanding how sustainability is conceptualized by members of the university community is an early stage in all institutional policy-making processes, because this term can take on different meanings depending on the context and the person (^{Wright & Wilton, 2012}).

4.2 Environmental Sustainability Integration

Cluster 2 involves the main concept of this study (environmental sustainability) as well as terms related to recycling and waste management in the higher education context. The articles in this cluster have been critically analyzed in order to identify the most discussed aspects surrounding the integration of environmental sustainability into higher education institutions. We wish to start by noting that the integration of environmental sustainability into higher education responds to institutional pressures (^{Wright, 2002}).

It could be said that the interest in integrating environmental concerns into universities stems from principles, initiatives and statements such as the Stockholm Declaration on the Human Environment (1977), the Ninth International Association of Universities Round Table: The Kyoto Declaration (1993) or the International Conference on Environment and Society-Education and Public Awareness for Sustainability: Declaration of Thessaloniki (1997) (^{Wright, 2002}), which have led universities to adopt strategies to contribute to major global issues. However, universities sometimes adhere to these initiatives only for public relations purposes and not out of a genuine interest in environmental sustainability (^{Wright, 2002}).

Supporting the arguments of ^{Wright (2002)}, ^{Ralph and Stubbs (2014)} maintain that while it is important that universities adhere to international initiatives, in most cases these initiatives fail to become effective actions, resulting in few systemic transformations. For instance, according to ^{Ralph and Stubbs (2014)}, universities have so far focused on recognizing the impact of energy use or waste generated, thus increasing the number of universities concerned with reducing the ecological footprint of their campuses, but which experience serious difficulties in the development of “green” curricula. This has some impact on campus operations but achieves little progress in environmental sustainability in intellectual and conceptual terms (^{Zen et al., 2016}). At the same time, the number of universities reporting their actions on initiatives like Global Compact grows at a slower pace than the number of commercial businesses does.

By this we do not mean that sustainable campus strategies are inadequate. On the contrary, sustainable campuses are a policy widely adopted by universities, but they are not, per se, a guarantee of environmental sustainability. Environmental sustainability is about much more: students, teachers and administrative staff are required to engage in real problems and actively engage in generating solutions (^{Filho et al., 2017}). As ^{Filho et al. (2017)} suggest, there is a need to go beyond the mere dissemination of information; if you really hope to change attitudes and behaviors, it is important to engage the community, reform curricula, and determine governance mechanisms.

Also in reference to barriers to the integration of environmental sustainability, but adopting a micro-perspective, ^{Lozano (2006)} analyzes the reasons why sustainability has not really been incorporated and implemented throughout most universities across the world. ^{Lozano (2006)} argues that the poor adoption of environmental sustainability initiatives by university institutions is largely due to individual barriers associated with resistance to change, and posits at least three barriers that make it difficult to integrate environmental sustainability into universities. The first relates to resistance to the idea itself, that is, resistance generated by the lack of information or the lack of a clear statement on the subject or because we disagree with the idea. The second barrier relates to a sense of loss of control, power, or status, which can lead to feelings of incapacity. The third type of barrier is associated with values: for example, the person may be aware that change is necessary, but disagrees as a result of cultural differences or religious, sexual or other beliefs.

Faced with the problem of a lack of real commitment from universities in adopting environmental sustainability strategies, ^{Filho et al. (2015)} suggest that the main concern for universities discussing environmental sustainability should be how to translate sustainability into real practices that transform society. According to ^{Filho et al. (2015)}, universities could make a real impact if they are able to develop holistic approaches that include strategies related to curriculum, the campus, the community, and research. However, although this is an internal issue for universities, it should be well understood that environmental sustainability issues require interdisciplinary solutions developed with multiple external stakeholders.

Finally, this cluster includes five case studies that describe the experiences of university institutions in environmental sustainability. Case studies are a research design that offers a deeper insight into the research phenomenon. One of the case studies (^{Zhang et al., 2011}) highlights how sustainable waste management has positioned the University of Southampton as one of the leading universities in environmental sustainability. In a similar way to ^{Lozano (2006)}, ^{Zhang et al. (2011)} found that resistance to change is one of the most recurrent barriers and has been the biggest obstacle to developing sustainability projects. Despite this, the university's success in implementing its sustainability strategy was largely due to the integration of students through both traditional and creative methods, thereby making them think for themselves about how and why they should care about recycling.

Another case study focuses on Michigan State University students and shows that the student community is aware of the benefits of recycling, but lacks sufficient knowledge about which materials can be recycled and where (^{Kaplowitz et al., 2009}). However, overall the results of the study are not negative; on the contrary, they suggest that students may be very interested in learning about the recycling process and that sustainability policymakers should conduct communication campaigns that do much more than just explain the reasons for recycling.

The third case study describes a project to examine the environmental and economic impacts of composters at Kean University (^{Mu et al., 2016}). This space has helped students to gain familiarity with and relate to sustainable technologies and real environmental problems. The fourth case study focuses on building a culture of environmental sustainability at the University of Michigan (UM) (^{Levy & Marans, 2012}). The practical recommendations put forward by the authors to promote the development of pro-environmental behaviors revolve around three main dimensions: education (through online tutorials, certified courses for managers and staff, training courses for teachers on how to integrate sustainability into curricula); engagement (open calls for ideas, competitions on recycling or energy use, provision of funds); and evaluation and monitoring (development of indicators, promotion of a culture of measurement and reporting of environmental performance).

The fifth and final case study focuses on Vytautas Magnus University (VMU), Lithuania, and highlights the role of this institution in the development of environmental sustainability policies, internally (as an organization) and externally (as an agent of the region) (^{Dagiliūtė & Liobikienė, 2015}). The authors conclude that universities, through sustainable policy, research, curriculum, and their role in the region, can have a significant impact on society, stressing, as we mentioned at the beginning of this discussion, that each university is under ever greater national and international pressure to engage in more profound environmental sustainability actions.

4.3 Environmental Protection

How to lead people to adopt pro-environmental behaviors? This question serves as a basis for several articles in this cluster. Environmental awareness, in the words of ^{Talay et al. (2004)}, is defined as care, concern, and sensitivity to environmental problems. However, just being conscious is not enough; saving the planet requires that people adopt more rational consumption habits and a lifestyle based on environmental protection. The problem in the case of developing countries, as explained by ^{Talay et al. (2004)}, is that it becomes difficult to change behaviors and attitudes towards the environment in contexts where people are struggling just to survive and therefore find it harder to give any thought to environmentally sustainable habits.

^{Yuan et al. (2013)} report that students are more aware of sustainability issues than their parents and faculty members. However, all three groups seem to be more aware of global or local environmental issues than university environmental concerns. ^{Yuan et al. (2013)} analyzed environmental sustainability, sustainable curricula, staff development and rewards, research and development, student opportunities, awareness of management systems, and social responsibility.

While promoting environmental behaviors is an arduous task due to an often muddled understanding of the concept of "sustainability" in the curriculum, research and campus operations context (^{Mader, 2013}), the difficulty of addressing the problem of environmental protection should not deter us from seeking to do so. For this reason, universities, as educational and life-training centers, play a fundamental role in safeguarding the planet. In that sense, ^{Grady-Benson and Sarathy (2016)} point to a shift from individual environmental responsibility to collective mobilization, whereas in the second cluster of this study, most authors described the development of individual actions, from a university as an organization or institution, to reduce the carbon footprint. What Grady-Benson and Sarathy (2016) suggest is for students to act collectively on campus and join national movements to tackle climate policies that compromise the well-being of the planet.

To achieve collective mobilization, for ^{Grady-Benson and Sarathy (2016)}, first it is necessary to understand how students perceive environmental problems. For this reason, one of the main contributions by ^{Campbell-Arvai (2015)} in relation to the promotion of environmentally responsible habits to prevent food waste in universities is a three-strategy approach: 1) conducting campus-wide campaigns that call on students to assess their food choices and environmental impacts; 2) university executives should make efforts to purchase food with little environmental impact in terms of plastic or non-organic waste, for example; and 3) developing environmentally friendly interventions to promote healthy habits and pro-environmental and pro-social behaviors. Also, in reference to interventions to improve students' awareness in terms of meal choices, ^{Whitehair and Shanklin (2012)} suggest using simple messages with students to introduce behavioral changes to address food waste.

In addition, there are two studies in this cluster that are closely related because they mainly address reducing the environmental impact within universities. ^{Iriarte et al. (2015)}, for example, analyze greenhouse gas emissions at the University of Talca (Chile) using the Greenhouse Gas Protocol Corporate Standard (the GHG Protocol) by the World Resources Institute (WRI). In particular, we consider this to be a robust methodology that could be replicated by Latin American universities, because it measures direct emissions from fuel consumption and student commuting by an institutional bus and fugitive and indirect emissions from the organization's electricity consumption. ^{Lukman et al. (2018)} also set out the case of a university's environmental impact assessment in Slovenia, but unlike ^{Iriarte et al. (2015)}, they report the use of a life cycle assessment (LCA) methodology. The main dimensions proposed by this methodology are the construction and demolition of buildings, cleaning, heating, lighting, painting, materials, and waste management.

Finally, ^{Sáez-Martínez et al. (2014)} set out the fundamental role of universities in achieving sustainable innovations in partnership with companies, researchers and citizens, among other agents, as part of innovation ecosystems. This is another route that universities have taken in recent years in order to contribute to sustainable development while strengthening their social and relational capital. Collaboration between universities and other agents to achieve sustainable innovations (i.e. circular economy, waste management, sustainable packaging) is much more complex than collaboration for traditional innovations, because environmental issues demand the involvement of a much larger number of agents, who may have different interests.