Introduction

Academic institutions play a crucial and significant role in helping society to meet the climate and environmental challenges proposed by international frameworks, such as Green Deal (COM 640 2019) and the framework for achieving climate neutrality (COM 80 2020) focussed on achieving climate neutrality in the short/medium term. Higher Education Institutions (HEI), as organisations committed to education and research, play a significant role in preparing responsible graduates involved in maintaining sustainable development, and they themselves have to be an example for their students and staff as well as for society as a whole. For this reason, calculating, tracking and reporting their own carbon footprint (CF) is a starting point from which to become sustainable organisations.

The term “carbon footprint” is defined by the IPCC Guidelines (2006) as “a representation of the effect on climate in terms of the total amount of greenhouse gases (GHG) that are produced, measured in units of CO2e as a result of the activities of an organization”. GHG emissions can be calculated for each source using the following formula:

$$E_{S} = \, AD_{S} \times \, EF_{S}$$

where the GHG emissions from a specific source (ES) are obtained from the product between the activity data from that specific source (ADS), which represents a quantitative measure of the source expressed in units (for example litres of petrol or kWh of electricity), and its respective GHG emission factor (EFS), which is a coefficient that allows activity data to be converted into GHG emission. Once the total GHG emissions from all sources have been calculated, they are added up to quantify the total CF in units of carbon dioxide equivalent (CO2e). This is a common unit for describing GHG emissions, for any quantity and type of GHG it signifies the amount of CO2, which would have the equivalent global warming impact.

Although organisations contribute significantly to GHG emissions, methodological guidance for them is less developed and less prescriptive than for products (Robinson et al. 2015). There are different international standards for calculating the CF of organisations. Amongst others, the most notable regulatory frameworks are the GHG Protocol (2004), ISO 14064–1 (2006) and ISO/TR 14069 (2013), PAS 2050 (2011) and PAS 2060 (2014). Although initially they were applied to verify the requirements for quantifying GHG emissions within organisations under the Kyoto Protocol (2008), their use is currently becoming widespread in other types of organisations that are voluntarily interested in calculating and communicating their CF.

Higher Education Institutions, also known as universities, are establishments devoted to post-secondary education and research that award academic degrees in different disciplines. Therefore, as organisations engaged in education, research and community services, they play an important role in sustainable development and the fight against climate change (Cordero et al. 2020). CF is a very useful tool for exercising a greater degree of control over activities that impact on the environment (Robinson et al. 2018) and also provides a baseline on which to evaluate the effect of future mitigation efforts on-campus (Letete et al. 2011).

Moreover, the role of HEI in sustainability is already recognised by different international declarations, such as the Talloires Declaration (TD 1990) or the Cre-Copernicus University Charta (Copernicus 1993), associations/networks, such as the CRUE’s Sectoral Sustainability Commission (CRUE 2002), the Association for the Advancement of Sustainability in Higher Education (AASHE 2005), the American College and University Presidents’ Climate Commitment (ACUPCC 2007), which was rebranded as the Carbon Commitment (CC 2015), the International Sustainable Campus Network (ISCN 2007) or the Global Universities Partnership on Environment for Sustainability (GUPES 2012), as well as rankings, such as the Times Higher Education-World University Ranking (THE 2004), the Sustainability Monitoring, Assessment and Rating System (STARS 2013) or the UI GreenMetric World University Ranking on Sustainability (GreenMetric 2010).

For these reasons, universities, as an example of sustainable organisations, should take a leading role in the fight against climate change and thus in the calculation, monitoring, reporting, reduction or even offsetting of their CF. However, as a preliminary step for calculating the CF of HEI, it is necessary to understand their activities that contribute to climate change by creating a greenhouse gas emissions inventory (Bailey and LaPoint 2016). HEI typically consists of a mixture of buildings used for classrooms, laboratories, offices, canteens, residences, etc. Some of them have their own power plants, transport circuits, water systems or health services, mainly depending on the number of students they host. Any of these activities have emission sources contributing to the CF, which need to be identified and quantified. This task can become complicated depending on the type and size of the institution. In this study, the use of university buildings and the material needed to carry out academic activities are considered. Santovito and Abiko (2018) offered recommendations on how to prepare the GHG inventory, identified some relevant emission sources and allowed a better visualisation of the opportunities for GHG mitigation. Yet, there is no specific standardised methodology for conducting the inventory and calculating the GHG emissions generated for the case of educational institutions.

Several reviews can be found in the literature. Some of them are focussed on analysing methodological aspects of the CF calculation, Fenner et al. (2018) for the building sector or Durojaye et al. (2020) in general, highlighting both the lack of standardisation in spite of the different frameworks developed for that purpose. Others are specific reviews dedicated to specific sectors, such as construction/buildings (Onat and Kucukvar (2020) for the construction industry, Che Muhammad Fatihi Hafifi Wahid et al. (2019) for highway developments and Schwartz et al. (2018) for refurbished and new buildings), population (Purwanto et al. (2019) for settlement activities and Heinonen et al. (2020) for consumptions in settlement, region, city or country), food and drink (Navarro et al. (2017) for wine and wineries, Brade and Brade (2014) for milk and milk products production, Rugani et al. (2013) for wine, Nijdam et al. (2012) for animal food and their substitutes and Pirlo (2012) for milk production), healthcare (Rizan et al. (2020) for surgical operations and Alshqaqeeq et al. (2020) for hospital services), metal (Nilsson et al. (2017) for Cu and Zn production from primary and secondary sources), tourism (Sun et al. (2020) for transport, accommodation, catering, shopping, entertainment, telecommunications, etc.), water (Cornejo et al. (2014) for water reuse and desalination) and Information and Communication Technology (ICT) (Grimm et al. (2014) for workplace hardware, server, networks and IT-services). However, no specific reviews have been done for educational activities, as Fig. 1 shows. So, this study fills this gap, encompassing research activity in CF in this field, from the date of publication of the first CF framework (2004), to the present (March 2021).

Fig. 1
figure 1

Other reviews in the literature classified by sector and years covered (n = number of articles)

Full size image

Although there are no general reviews related to this subject, the literature contains a few studies focussed on comparing the CFs of different HEIs belonging to specific associations in a specific geographical area. So that, Sinha et al. (2010) compared the CF from institutions that were signatories of the American College and University Presidents’ Climate Commitment (ACUPCC) and Bailey and LaPoint (2016) compared the CF from nine universities located in Texas (USA). Both studies applied the Clean Air Cool Planet Calculator (CA-CP 2020) to compile and model the emission data from the institutions compared in each study. However, when trying to compare the CF from different studies, the scope, boundaries, emission sources, emission factors, etc. are specifically defined for each case study, making it difficult to carry out comparisons amongst different HEIs. This fact highlights the lack of a common framework.

Taking into account this context, the aim of this study is to carry out a review of studies calculating the CF of HEI worldwide in order to identify the most common practises related to the methodological aspects of the calculation and to compare results. This will make it possible to establish a common framework that facilitates comparability of the studies. The paper is structured as follows: Sect. 2 presents a four-stage methodology used to select the studies under review and as the basis for comparing those studies in general terms, as well as their methodology and results; Sect. 3 presents the results obtained after applying the methodology; Sect. 4 discusses the results; and, lastly, Sect. 5 draws some final conclusions.

Research methodology

This systematic review follows a structure designed to achieve consistency, robustness and transparency in research. The methodology guides the selection of case studies focussed on calculating the CF in HEI and the evaluation rules to identify the information to be extracted. The research methodology has four stages, as shown in Fig. 2 and described below:

  • Stage 1 aims to identify the literature focussed on quantifying the CF of HEI and whose content included fully detailed and defined case studies with a consistent methodology and results.

  • Stage 2 includes the general mapping of the literature selected in stage 1, considering temporal aspects and the main descriptive characteristics of the institution under analysis (location, size, etc.).

  • Stage 3 goes deeper into the analysis of specific aspects related to the calculation of CF, that is, methodologies applied, goal definition, scopes, source emissions, etc.

  • Stage 4 focuses on the comparison amongst the CF of the HEI analysed and the identification of the causes underlying the differences found.

Fig. 2
figure 2

Methodology

Full size image

Results

Stage 1: identification of articles

The Scopus database was used as the main search engine for selecting the literature, using “carbon footprint, university”, “greenhouse, university”, “carbon footprint, higher education” and “greenhouse, higher education” as strings in the article title and keywords. To avoid limitations due to the capacity of the database search, other sources, such as GoogleScholar and ScienceDirect were used with these same strings to complete the list of articles to be analysed. In addition, the list obtained was completed with specific articles found in the list of references of those articles focussed on literature reviews. By applying this procedure, 84 articles were found. A first screening of the source title and article title/keywords resulted in the rejection of those that did not correspond to indexed research articles or conference papers, and those that did not focus on the calculation of the carbon footprint of HEI, respectively. In addition, duplicates were also rejected, thus reducing the sample to 55 articles. A second screening of the content of the abstracts was conducted and articles that were not aimed at calculating the carbon footprint of one or more HEI were excluded, thereby further reducing the sample to 44 articles. After a third screening of the full text, only articles that included fully detailed and defined case studies with a consistent methodology and results were included. A final sample of 35 articles was selected, as reported in Table 1. A descriptive content analysis was carried out, considering the aspects detailed in Fig. 2.

Table 1 Universities’ carbon footprint
Full size table

Other studies were not included in this review although they did calculate the CF, because this calculation was linked to specific university activities. For example, Chung et al. (2014) and Sippel et al. (2018) calculated the CF due to some students’ campus activities in Tajen University (Taiwan) and in the University of Applied Science in Konstanz (Germany), respectively, while Kulsuwan et al. (2019) only took into account the students' electricity consumption in Mahidol University Amnat Charoen Campus (Thailand). Barros et al. (2018), Pérez-Neira et al. (2020) and Rao et al. (2017) calculated the CF of transportation habits in the Federal University of Technology (Brazil), the University of León (Spain) and the Symbiosis International University (India), respectively, while Beardsley and Morton (2009) did the same for university sponsored air travels. In addition, Schwarz and Bonhotal (2018) determined the CF of a compost facility at Cornell University (USA), Stephan et al. (2020) examined the embodied CF in materials on the Parkville campus of the University of Melbourne (Australia) and Song et al. (2016) calculated that of scientific publications at Dalian University of Technology (China). On the other hand, Filimonau et al. (2021) compared the carbon intensity of on-campus and off-campus higher education, taking advantage of the unique opportunity of the COVID-19 pandemic.

Stage 2: general mapping

As a starting point of the review, a general mapping of the 35 research papers reported in Table 1 was carried out by analysing the following general aspects related to the time period analysed and the main characteristics of the institution:

  • Location of university It can also be observed in Fig. 3 that studies are distributed all around the world: 20% in Europe, 34% in Asia, 6% in Africa, 6% in Oceania, and 23% and 11% in North and South America, respectively. The CF can be highly conditioned by the climatic conditions of the location of the higher education institution, so it is necessary to subsequently analyse the influence of its value with the location of the HEI.

  • Type of institution 86% of the studies correspond to public higher education institutions, 11% to private ones and the remaining 3% to national ones. Therefore, a greater involvement of public higher education institutions is observed.

  • Year analysed 34% of the studies were conducted between 2000 and 2010, while the remaining 66% were carried out in the decade 2011–2020. As Fig. 3 shows, there is a significant increase in the number of articles published over the years, mainly in 2020. This may be due to the fact that HEI is more aware of environmental issues and also their commitment to contribute to their decarbonisation as a sustainability strategy.

  • Time period analysed Some studies calculated the carbon footprint for one academic year (29%), while others did so for a fiscal year (68%) and only one of them considered a single semester (Kandananond 2017). The results are lower when only one semester is calculated, so this study is not comparable with the rest.

  • Boundary 43% of the studies analyse the whole institution (I, in Table 1), 48% analyse only one campus of the institution (C, in Table 1), while the remaining 9% analyse only one building/school (B, in Table 1). It is thus observed that universities prefer to carry out the analysis of a single campus, as the different campuses often act independently.

  • Size of the institution As Table 1 reports, two main parameters are considered for measuring the size of the institution: people and area. The parameter people is considered in 86% of the studies, although it is measured according to three different variables: students in 77% of the studies, employees in 57% of the studies and per capita (as the sum of students and employees) in 66% of the studies. The parameter area is considered in 43% of the studies. This aspect is important as it allows a comparison of CF to be carried out amongst different institutions.

Fig. 3
figure 3

Temporal evolution of the literature related to calculation of CF of higher educational institutions, by country

Full size image

Stage 3: methodological aspects

A content analysis of the selected articles was performed to identify the main methodological aspect followed in each one. This was a feedback process defining the criteria used to classify each article in order to support the results and discussion of findings. The methodological elements evaluated for each article selected are reported in Table 1 and analysed below:

  • Standard applied for the CF calculation As described in the introduction section, there are different international standards for calculating the CF. Only 17% of the studies reviewed fail to indicate the standard used as a basis. Of the rest, 54% use GHG Protocol (2004), 20% use (IPCC Guidelines 2006), 11% use ISO 14,064–1 (2006), and PAS 2050 (2011) is used by Budihardjo et al. (2020) and Thurston and Eckelman (2011).

  • Aim of the study Although the main aim of all the studies is to evaluate the CF of the institution/campus or building/school, three different secondary objectives were also identified. Some studies analyse the evolution of emissions over time (40%) while the rest only calculate emissions during a specific year (60%). In addition, 46% of the studies compare their emissions with the CF of other universities and 80% of the studies analysed the effect of different measures aimed at reducing the CF.

Table 2 reports the specific measures applied in the literature to reduce the CF of higher education institutions, classified by the scope they have an influence over. In addition to the measures reported in Table 2, there are other strategies that affect specific aspects of the institutions. At the buildings level, strategies, such as designs for low-energy buildings are proposed by Baboulet and Lenzen (2010) and Ozawa-Meida et al. (2013), infrastructure interventions are carried out using techniques and materials with a low carbon footprint by Varón-hoyos et al. (2021) or minimising the construction of new buildings by Jung et al. (2016) and Riedy and Daly (2010). Moreover, establishing a structured sustainability office/group responsible for monitoring, tracking and advocating for sustainability initiatives is proposed by Riedy and Daly (2010) and Bailey and LaPoint (2016), and integrating an environmental management system, is proposed by Rodríguez-Andara et al. (2020), since it will facilitate the calculation of emissions and their mitigation. In specific HEI, which use animals for teaching or research purposes, Butt (2012) proposed decreasing the number of grazing animals and increasing the per animal productivity or decreasing the amount of dung and urine added to the pasture through restricted grazing.

Table 2 Specific actions to reduce the CF
Full size table

23 % of the studies analysed the improvement that would be obtained by applying these actions by accurately calculating the tCO2e saved or obtaining the specific percentage of reduction in emissions, while the rest of the studies reviewed only comment on the recommendations (see “Action CF improvement analysis (%)” row in Table 2).

  • Scope/Emission sources Traditionally, organisations have accounted for just Scope 1 (all direct emissions) and Scope 2 (indirect emissions from electricity purchased and used by the HEI) in their CF reports because they are the easiest and cheapest to assign and calculate. Quantifying Scope 3 (other indirect emissions) can be quite challenging for HEI, particularly for the largest and highly decentralised ones. Table 3 reports the emission sources for each scope considered in the studies reported in Table 1 and in Santovito and Abiko (2018), which offers recommendations on preparing the campus inventory.

Table 3 Emission sources considered in the literature
Full size table

Apart from the emission sources reported in Table 3, which are commonly considered in the general literature focussed on the calculation of CF, other specific emission sources are considered in the studies reviewed, such as procurement of glassware, plasticware and capital woods (e.g. equipment and setups for scientific laboratories) (Sangwan et al. 2018), farm machinery (Ologun and Wara 2014) and trips for parents/friends visiting the campus (Ozawa-Meida et al. 2013).

Moreover, not all the source emissions are always considered in the same scope. Leakage of refrigerants is usually considered in Scope 1, although Sangwan et al. (2018) included them in Scope 3. On the contrary, other emission sources which are usually considered in Scope 3 are included in Scope 1 in some studies. For example, the purchase of fertilisers is included in Scope 1 by Bailey and LaPoint (2016) and Clabeaux et al. (2020), water supply by Budihardjo et al. (2020), Syafrudin et al. (2020) and Ullah et al. (2020) and wastewater treatment by Clabeaux et al. (2020) and Criollo et al. (2019). Finally, fuel used in power generators, which is usually considered in Scope 1, is included in Scope 2 by Güereca et al. (2013).

Figure 4 shows a graphic representation of the percentage distribution of the emission sources considered in the studies reviewed. As can be seen, not all emission sources have the same level of consideration. It can be observed that 100% of the studies consider scope 2, 86% consider scope 1 and 94% consider some source from scope 3. In scope 1, 72% consider stationary consumption, 83% consider the vehicle fleet, while only 33% consider the leakage of refrigerants. In scope 2, all the studies consider emissions from purchasing electricity and only Gu et al. (2019) considered the generation of electricity (renewable energy). Regarding scope 3, the five sources of emission that are considered the most are commuting (75%), generation of waste (75%), business trips (56%) and the consumption/procurement of paper (47%) and water (36%).

Fig. 4
figure 4

Emission sources considered in the articles analyse

Full size image

Regarding the emission factors applied for each source of emissions, it can be observed that they are quite variable and depend mainly on the country, in which the institution under study is located. Current emission factors are available from many handbooks, government publications and the literature searches of appropriate research papers and journals. Some studies do not report the reference source of the emission factors (7 %) and none of them report the specific list of emission factors that are applied. Some studies used emission factors provided by official government sources, for example, Butt (2012) from New Zealand, Criollo et al. (2019) from UK, Mendoza-Flores et al. (2019) from Mexico, Ridhosari and Rahman (2020) from Indonesia or Rodríguez-Andara et al. (2020) from Spain. Most of them are based on relevant sources, such as IPCC, DEFRA or EPA.

  • Data source Data related to the consumption of energy, fuel, water, etc. are usually obtained directly from primary sources (field data obtained directly from the institution under analysis) in the studies reviewed, that is, through bills, counters, etc. and for the same base year, which allows reliable and accurate results to be obtained. For commuting and paper consumption, however, data are commonly obtained through a survey carried out with students and/or employers, which involves a certain degree of uncertainty since it depends on the veracity of the answers and the respondents’ level of involvement. Nevertheless, due to limited data availability, some studies need to make assumptions, such as Letete et al. (2011) that considers the consumption of LPG, acetylene and transport in November 2007 as the average consumption from January to October and the same electricity consumption for the previous year; other studies extrapolate daily or monthly data to annual data (Quintero-Núñez et al. (2015) calculates an average daily electricity consumption for winter and summer and extrapolates it to the total number of working days; Almufadi and Irfan (2016) extrapolate daily transport data to annual data and Iskandar et al. (2020b) calculate monthly carbon footprint), Güereca et al. (2013) extrapolates data from one building to the rest of the institution analysed and Ozawa-Meida et al. (2013) extrapolates commuting information extracted from surveys to the whole university. Other studies also used surveys to estimate data for the consumption of fuel in generators (Ologun and Wara, 2014), for electricity consumption by lighting/equipment (Güereca et al., 2013) and for LPG consumed in residential colony (Ullah et al. 2020).

  • Tool for CF calculation Most of the studies reviewed (83%) do not use any commercial/free CF calculation tools; instead, the calculations are performed with their own means. The remaining studies applied different tools: Clean Air Cool Planet Carbon Calculator (CA-CP 2020) was used by Bailey and LaPoint (2016), Klein-Banai et al. (2010) and Moerschbaecher and Day (2010), Economic Input–Output Life Cycle Assessment on-line tool (EIO-LCA 2020) was employed by Thurston and Eckelman (2011), IELab (2020) was implemented by Stephan et al. (2020), Umberto Software (Umberto 2020) was the tool used by Sangwan et al. (2018) and SimaPro (Pre Consultants 2019) was employed by (Ullah et al. 2020).

Comparison of results

The annual CF is reported in all the studies reviewed, although they are not comparable due to the different sizes of the institutions/campuses/schools/faculties analysed. In order to create a meaningful study despite these differences and to obtain a good basis for comparing the CF, a normalisation based on different criteria (student, employee, capita and area) is applied, as reported in Table 4. In addition, comparing this rate with the location of the HEI also reveals differences: 5.25/2.30/2.25/1.77/0.67 t CO2e/student on average for North America, Africa, Europe, Asia and South America, respectively. As a general result, it can be observed that the range of variation is very wide, regardless of the standardisation criteria applied. Moreover, other studies, especially those focussed on assessing the CF of university purchases (t CO2e/€ purchased), obtain less representative results due to the variety of potential purchases: 0.38 kgCO2e/€ (Larsen et al. 2013), 0.34 kgCO2e/€ (Ozawa-Meida et al. 2013) or 2.81 kgCO2e/€ (Alvarez et al. 2014). This fact clearly denotes a lack of homogeneous criteria when conducting the study.

  • Disaggregation of results All the studies calculate the CF for the defined boundary. However, 29% of the studies disaggregate the results per building of the institution/campus/school and 88% disaggregate them by the source causing the GHG emissions.

  • Offset emissions Only 14% of the studies have calculated their compensation potential, obtaining an emission offset ranging from 0.09% to 18%. However, many studies (26%) recognise the importance of incorporating university reforestation. For this reason, they include offset emissions as one of their recommendations for reducing the Carbon Footprint (see “Offset emissions” row in Table 3).

Table 4 Normalised CF according to different units
Full size table

Discussion

Clear differences have been identified when the normalised CF from different HEI are compared. The main reason for this is the lack of a single international standardisation method for calculating the CF for organisations and specifically for educational institutions, which present certain peculiarities when compared to organisations in other areas. This study has highlighted the main aspects that require international consensus in order to obtain comparable results.

The first aspect is related to the need to establish a common time metric and functional unit. As activities in HEI do not remain constant over time—including face-to-face teaching, exams, holidays, etc. (Bailey and LaPoint 2016)—it is desirable not to extrapolate data from a specific period. It seems to be more advisable to consider the fiscal year (instead of the academic year) as a temporal basis for calculating the CF in educational institutions, since emission factors are revised and published annually. Hence, it is recommendable to implement mechanisms for keeping historical records.

Furthermore, consensus is needed as regards the reference unit applied to normalise the CF, in order to be able to make comparisons amongst HEI. It would be recommendable to use kg CO2e/student, since it is directly related to the function of the HEI. It is also in agreement with the definition of the functional unit in the LCA methodology (ISO 14040 2006).

To analyse the values of the CF in greater depth, the evolution of the CF/student over time is represented in Fig. 5. It can be seen that, in spite of the differences amongst the studies reviewed, a trend is sensed towards the reduction in the CF over time, probably due to the increasing environmental awareness of HEI.

Fig. 5
figure 5

Normalised carbon footprint: temporal evolution of CF per student

Full size image

As set out in the previous section, one of the main reasons for the differences observed amongst the normalised CF in the studies reviewed is the scope and the emission sources considered in each of them. Figure 6 shows the normalised CF for those studies that allow its calculation (78% of the studies reviewed). The bar that represents the value of the CF/student for each study is shown in colour, which depends on the contribution of each scope to the CF: green for scope 1, blue for scope 2 and yellow/red/purple for scope 3.

Fig. 6
figure 6

Emission sources

Full size image

It should also be noted that the CF from each study are not directly comparable, since they do not include the same emission sources, as Figs. 4 and 6 show. Some of them only consider a single emission source, while other studies use almost all those reported in Table 3. However, the contribution of each emission source is quite variable depending on the value of its corresponding emission factor. Figure 7 represents the average contribution that each source emission has on the CF of studies reviewed.

Fig. 7
figure 7

Contribution of emission sources

Full size image

As can be seen in Fig. 7, the major contribution to the CF in HEI comes from electricity consumption in buildings (scope 2), followed by daily commuting to and from the campus (scope 3), as also remarked by Bailey and LaPoint (2016) for universities located in U.S., Butt (2012) in New Zealand, Güereca et al. (2013) in México and Jung et al. (2016) in Korea. Scope 3 emissions represent the main source of CF in most of the HEI where they were counted. Scope 3 accounted for around 79% of the total CF of Montfort University (Ozawa-Meida et al. 2013), around 68% for the Curicó Campus of Talca University (Vásquez et al. 2015) and up to 80% in Castilla-La Mancha University (Gómez et al. 2016).

This discussion stresses the need for a standardised framework for calculating the CF of organisations and specifically for HEI. In addition, it is also important to remark the need to include scope 3 in the near future (it is usually optional), since it represents a large percentage of emissions and is a good source of action to reduce emissions.

Conclusion

Footprint assessment increases the level of environmental awareness of the population and provides a baseline to measure the impact of future policies and technical measures to reduce consumption and its associated GHG emissions. The results of HEI carbon footprinting will increase environmental awareness in the student population, which could then spill over to the larger population. With this goal in mind, the number of HEI that calculate their CF is gradually increasing and this sector has been able to reduce its environmental impact while increasing its efficiency.

This study conducted a global literature review focussed on carbon footprint analyses in the HEI sector and also pinpointed the main gaps in the literature. Although several standards have been developed in response to the need for transparency in reporting GHG emissions, there is no internationally accepted method for measuring, reporting and verifying offsets of GHG emissions from HEIs in a consistent and comparable way. Consequently, carbon footprint studies often yield widely divergent results and comparison of the CF of HEI is difficult. Indeed, even after normalising differences in the student population, the metric tons CO2-e per student varied between the HEI that were compared. This may be due to the time period considered, taking different methodologies or tools into account, the fact that each of them have incorporated different GHG emission sources in their scopes or have used different methods to obtain activity data, and even due to the use of specific/generic emission factors for each source. In addition, the application of action plans and offsetting projects to compensate the CF can also contribute to the differences in the CF from one university to another.

Hence, this study illustrates the importance of developing comprehensive and consistent GHG inventories for HEI, listing the GHG emission sources included and the factor emissions considered so that HEI is not compared unfairly.

Moreover, the findings demonstrated the importance of taking the Scope 3 emissions into account when analysing the carbon footprints of HEIs, because carbon emissions related to Scopes 2 and 3 correspond to the major portion of total emissions of the sector. For example, emissions associated with commuting by employees and students are often considered a significant carbon emitter, but are not always included as they are optional.

Therefore, after this review, it can be concluded that there are no standardised criteria for reporting the GHG emissions of universities or higher education centres, mainly with regard to aspects related to organisational boundaries for scope 3 emissions and emission factors. The absence of common criteria results in inventories that, apart from problems of comparability, do not indicate clearly the possible opportunities for mitigation actions.

For this reason, and as a futures research proposal, it is necessary to develop methodologies and a simplified tool for calculating the carbon footprint of HEI. The goal is to obtain results that can be compared with other universities and that all opportunities for reducing emissions can be identified and considered.