Sie können Operatoren mit Ihrer Suchanfrage kombinieren, um diese noch präziser einzugrenzen. Klicken Sie auf den Suchoperator, um eine Erklärung seiner Funktionsweise anzuzeigen.
Findet Dokumente, in denen beide Begriffe in beliebiger Reihenfolge innerhalb von maximal n Worten zueinander stehen. Empfehlung: Wählen Sie zwischen 15 und 30 als maximale Wortanzahl (z.B. NEAR(hybrid, antrieb, 20)).
Findet Dokumente, in denen der Begriff in Wortvarianten vorkommt, wobei diese VOR, HINTER oder VOR und HINTER dem Suchbegriff anschließen können (z.B., leichtbau*, *leichtbau, *leichtbau*).
This study aims to integrate methodologies for carbon footprint (CF) assessment in higher education institutions (HEIs) using the Greenhouse Gas Protocol and ISO 14064:2018 standards with a hybrid process-economic approach. The research addresses current gaps in HEI CF assessment, where scope 3 emissions are often unreported, and establishes a framework for systematic quantification across the educational value chain.
Methods
A hybrid assessment model was developed, combining process-based and economic-based approaches through environmentally extended input–output (EEIO) analysis. The methodology integrates complementary aspects of the GHG Protocol and ISO 14064:2018, enabling the categorisation of emissions across organisational boundaries. The framework was applied to the University of Coimbra (Portugal) as a case study, with data collected from 2019 to 2023 using sustainability reports, financial statements, and mobility surveys. Thirty-six activities were compiled across four ISO 14064 categories and nine GHG Protocol sub-categories. Data quality was assessed using a tiered approach with sensitivity analysis performed through the one-at-a-time method to identify emission hotspots.
Results and discussion
Analysis revealed that the University of Coimbra emitted 29.53 kt CO₂-eq in 2023, representing a 59.69% increase from 2019, with a carbon intensity of 0.87 t CO₂-eq/person. Scope 3 emissions constituted 78.38% of total emissions, primarily from construction activities (40.35%), commuting (32.18%), and purchased goods (8.05%). Without the hybrid approach, significant indirect emissions would have remained unquantified. The comprehensive assessment revealed the University’s CF to be substantially higher than previously reported figures, which excluded scope 3 emissions. Benchmarking against other Portuguese institutions positioned the University of Coimbra’s carbon intensity (0.87 t CO₂-eq/person) within expected national ranges.
Conclusions
The hybrid CF assessment methodology effectively captures comprehensive emission profiles in HEIs, particularly revealing previously unreported scope 3 impacts. This pioneering application to a Portuguese university establishes a benchmark for subsequent assessments while highlighting significant institutional emission sources. The framework enables meaningful comparative analyses between institutions and can be readily adapted for national and international HEIs seeking to enhance their sustainability assessments and develop targeted carbon reduction strategies.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
Climate emergency requires sustainability assessments that focus on the impact of climate change. Regarding higher education institutions (HEI), these assessments have evolved from simple operational controls (such as water and energy consumption) to environmental impact assessments, mainly greenhouse gas (GHG) accounting (Paredes-Canencio et al. 2024). Currently, there is no internationally accepted method for consistently and comparably measuring and reporting GHG emissions from HEI (Valls-Val and Bovea 2021). While several internationally recognised methodologies are available for carbon footprint (CF) accounting, such as the GHG Protocol and ISO 14064:2018, their sector-agnostic nature provides challenges and opportunities for HEI applications (International Organization for Standardization 2018; World Business Council for Sustainable Development & World Resources Institute 2004).
HEI activities involve many decision-makers, stakeholders, and activities, making the CF assessment particularly complex. The GHG Protocol, developed by the World Resources Institute (WRI) and the World Business Council for Sustainable Development (WBCSD), is the most commonly used standard for GHG accounting worldwide (World Business Council for Sustainable Development & World Resources Institute 2004). This methodology assists in determining the CF in HEI regarding operational scopes, identifying improvement areas, and establishing climate targets (Cano et al. 2023; Herth and Blok 2022). The protocol’s widespread adoption stems from its definition of emission across three scopes: (1) direct emissions from sources that are controlled or owned by an organisation; (2) indirect emissions from the purchase of electricity, steam, heat, or cooling; (3) indirect upstream and downstream value chain emissions. According to this standard, scopes 1 and 2 are compulsory, and scope 3 is optional. The lack of a mandatory scope 3 often fails to capture unique aspects of HEI operation, such as goods consumption and mobility patterns.
Anzeige
ISO 14064:2018 provides a methodology for inventory management, calculation, and reporting of GHG emissions (International Standard Organization 2018). It defines the approach to quantify and qualify processes for CF, grouping GHG emissions into six categories: one direct (emissions and uptakes generated within the organisation’s boundaries) and five indirect emissions (purchased energy, transportation, products used by the organisation, use of the organisation’s products, and other sources). Notably, this standard accounts for carbon uptake activities, making it particularly valuable for HEIs with significant green areas. The alignment of the standard with the GHG Protocol creates opportunities for a more comprehensive accounting of emissions in educational contexts, studying in depth the most relevant indirect emissions in each context (Cano et al. 2023).
However, GHG Protocol and ISO 14064:2018 have primarily been applied at the industrial and corporate level, where emissions are predominantly concentrated in scopes 1 and 2 (Osorio et al. 2022; World Business Council for Sustainable Development & World Resources Institute 2004). While previous studies, such as Cano et al. (2023) and Kiehle et al. (2023), have applied these standards separately to HEIs, their integration offers enhanced capabilities, mainly in the categorisation of emissions (Cano et al. 2023; Kiehle et al. 2023). This presents unique challenges for HEIs, whose GHG emissions profile differs from the abovementioned organisations. Recent studies indicate that scope 3 emissions typically correspond to the largest share of HEI CF (World Business Council for Sustainable Development & World Resources Institute 2004). The incomplete consideration of scope 3 emissions might result in underestimating the CF, particularly regarding indirect emissions from the educational value chain. For example, in HEIs, an exhaustive evaluation requires considering the mobility of students and staff, the acquisition of teaching materials, and research-related issues, all of which fall within scope 3. Collecting data for scope 3 emissions presents unique challenges in the HEI context. These include (i) complex and often opaque supply chains for educational and research materials, (ii) diverse and fluctuating student and staff commuting patterns, (iii) difficulty in tracking business travel across decentralised academic departments, (iv) incomplete procurement records with varying levels of detail, and (v) limited institutional control over upstream and downstream activities (Helmers et al. 2021; Li et al. 2021). Additionally, HEIs frequently lack standardised protocols for collecting and classifying scope 3 data, creating inconsistencies in inventory development and hindering meaningful comparisons between institutions (Larsen et al. 2013). These challenges help explain why many HEIs limit their carbon accounting to scopes 1 and 2, potentially overlooking the majority of their actual emissions.
Valls-Val and Bovea (2021) show that the CF in HEIS ranges from 0.06 to 10.94 t CO2-eq/student.year, with a mean of 2.67 t CO2-eq/student globally (Valls-Val and Bovea 2021). This wide variation stems from three key factors: (i) methodological differences in boundaries (scope 1, 2, 3), (ii) institutional characteristics (size, location, and main activities), and (iii) diverse reporting metrics (student, employee, area). Moreover, Herth and Blok (2022) and Li et al. (2021) highlighted data gaps in HEI inventories (Herth and Blok 2022; Li et al. 2021). Common gaps include insufficient data granularity and highly aggregated information, making it challenging to identify GHG hotspots. For instance, data is rarely disaggregated by time and location (department or building), limiting GHG tracking. Moreover, additional gaps identified were as follows: (i) a lack of knowledge about the categorisation of emissions, their scope, and their sources; (ii) lack of relevant data (mainly in the scope 3) in the inventory of emissions, leading to an underestimation of the CF; and (iii) a lack of transparency in the inventory, which makes benchmarking difficult.
Calculating indirect emissions in HEIs is often resource-intensive, requiring significant time, data, and expertise. Some of the main difficulties are related to the lack of specific data for activities and processes within the management structures of HEIs. This difficulty is due to its inherent complexity and variable size (often involving a cluster of institutions), such as faculties, departments, and institutes. For example, tracking movements between campuses, canteens, and residences or monitoring the value chain of goods procurements present significant challenges (Helmers et al. 2021, 2022). These activities are complex to quantify and frequently require estimation approaches. Environmentally extended input–output (EEIO) is a commonly used method for calculating CF based on macroeconomic data (Larsen et al. 2013). This method allows consideration of GHG emissions embodied in the inputs and outputs of economic data, leveraging readily available monetary flows for different activities (Schoenaker and Delahaye 2017). Several hybrid models, containing both process and economic data, have been developed to address data gaps and estimate scope 3 emissions (Kiehle et al. 2023; Larsen et al. 2013; Robinson 2017; Yue et al. 2016).
Anzeige
Herth and Blok (2022) applied a hybrid model to the Delft University of Technology, showing that the emissions quantification based on the EEIO were around 70% of the total footprint, contributing to scope 3 (80% of total CF), which highlights the importance of the hybrid model for the comprehensive quantification of emissions on the value chain (Herth and Blok 2022). A hybrid process allows more comprehensive inventory management, filling gaps in the data inventory, and avoiding underreporting. Moreover, financial data is usually easier to access as HEI must report yearly economic activities. Portuguese HEIs are showing an increasing interest in CF assessment and are motivated by sustainable rankings and pioneering opportunities. Available data from HEIs indicates that the carbon intensity varies between 0.20 and 1.66 t CO2-eq/person across Portuguese HEIs (Deda et al. 2023). However, only a few HEIs reported the CF, and when they did, only aggregate data were provided without a detailed inventory description. In addition, the methodological variations, such as emission factors and data quality, make meaningful comparisons challenging.
In this paper, a hybrid CF model is proposed for the GHG assessment of HEIs, aggregating process-based and EEIO-based approaches, capitalising on the complementary strengths of the GHG Protocol and ISO 14064:2018. Unlike previous hybrid approaches, the methodology proposed in this paper addresses HEI operational complexities while maintaining cross-standardisation for institutional comparability. This methodology enables comprehensive CF assessment by leveraging readily available institutional data, providing a framework that balances accuracy with practicality for the educational sector. Hence, the aims of this study are:
To align ISO 14064:2018 categories with the GHG Protocol for comprehensive assessment in HEIs
To apply a hybrid method combining process-based and EEIO-based approaches to fill data gaps
To validate the methods effectiveness through a case study application
To establish a reference framework for calculating the CF in Portuguese HEIs and other international HEIs, allowing for benchmarking
The proposed approach directly addresses the identified literature gaps while guiding other HEIs to overcome their current challenges. The methodology is described in Sect. 2 and shows how these standards can be integrated, maximising synergies while addressing HEI-specific difficulties, such as economic-based emission factors. In Sect. 3, the proposed approach is applied to the University of Coimbra (UC) due to its significant role in the national context, data, and current lack of scope 3 calculation.
2 Methods
The proposed methodology integrates the complementary strengths of the GHG Protocol and ISO 14064:2018 to address the data challenges of HEI CF assessment. GHG studies have typically applied these standards separately; this study offers a comprehensive scope coverage through unified categorisation between the referenced standards, exemplifying inventory management tailored to educational contexts for HEI decision-makers. This integration leverages the GHG Protocol’s organisational approaches and scope-based categorisation while incorporating ISO 14064:2018 detailed inventory management and specific guidance for carbon sinks.
2.1 Model description
GHG emissions are quantifiable per unit of activity (ACi), corresponding to the flows of resources that impact the boundaries of a system under study. All levels of activities, ACi, are needed to assess the CF of HEI, which can be obtained in a bottom-up or top-down strategy, depending on data availability and the institutional structure. The bottom-up approach gathers disaggregated data (e.g. building-level energy consumption). The top-down approach is employed when only aggregated data is available (e.g. total HEI energy consumption in the HEI). The PB is primarily used for scope 1 and 2 emissions and is available through utility bills and facility records. EEIO data are helpful for scope 3 emissions, where physical data is unavailable and usually available from financial records (Gamarra et al. 2019). When the PB data source is insufficient or incomplete, as is common in top-down approaches, physical and monetary data may be combined in the same inventory to avoid underreporting. This combination transforms the inventory into a hybrid one.
Emission factors (EFi) represent the quantity of a pollutant emitted per unit of activity. An EFi is obtained by calculating the global warming potential (GWP) for each ACi in terms of carbon equivalent. The GWP approach was established to allow comparisons of the impacts of different atmospheric gases on global warming. According to the Intergovernmental Panel on Climate Change (IPCC), GWP is ‘an index measuring the radiative forcing following an emission of a unit mass of a given substance, accumulated over a chosen time horizon, relative to that of the reference substance, carbon dioxide’. Thus, GWP corresponds to the combined effect of the different times (e.g. 100 years is used in this study) in which these substances remain in the atmosphere and their effectiveness in causing radiative forcing. According to the GHG Protocol, the GHGs included in the CF are CO2, N2O, CH4, SF6, HFC, and PFC (World Business Council for Sustainable Development & World Resources Institute 2004).
Moreover, EFi are available for PB and EEIO data; the former is related to physical and operational indicators (e.g. MWh, tons), while the latter has monetary indicators (e.g. €). The EFi also considers different life cycle stages depending on the emission source. These stages represent the number of processes accounted for within each EFi, such as raw material production, utilisation, and disposal. For example, cradle-to-gate factors encompass emissions from raw material extraction through production, which is particularly relevant for purchased goods and equipment. Process-based EFi typically considers only gate-to-gate emissions, whilst economic EFi derived from EEIO analysis provides cradle-to-gate coverage quality. The CF is calculated by the sum of the multiplication of ACi and the EFi for each emission source (i), following Eq. (1),
$$CF= {\sum }_{i=1}^{S}A{C}_{i} x E{F}_{i}$$
(1)
where S is the number of emission sources, on the other hand, the hybrid methodology combines process-based (PB) and economic-based (EEIO) approaches to fill data gaps and improve the quantification of CF in HEIs. While PB approaches are accurate for direct emissions (scope 1) and energy consumption (scope 2), they have difficulties capturing indirect emissions due to the unavailability of detailed physical data (Herth and Blok 2022). The practical implementation follows the steps and requirements of both standards, which are demonstrated in detail in the case study. From the GHG Protocol, the incorporation considered its reporting principles (relevance, completeness, consistency, transparency, and accuracy), organisational and operational boundaries, and emissions reporting structure. ISO 14064:2018 was considered in the objectives definition, scope alignment, quantification methodology, and results presentation. Usually, the functional unit is defined as one year of normalised operation (e.g. per student or faculty) to enable comparisons between institutions of different sizes.
While the developed framework provides the theoretical foundation, the case study will demonstrate how this guidance is fulfilled in practice, mainly focusing on scope 3 classification and quantification. The source of emissions and the classification of activities are described in the following subsection. To quantify the CF of an HEI, emissions from upstream activities, the operation phase, and downstream activities must be considered. From the combination of the GHG Protocol and ISO 14064:2018, twenty-two sub-categories emerged, categorised into six main categories, as indicated in Table 1, which includes a brief description and examples of application in the HEI context and the association with the scopes of the GHG Protocol.
Table 1
ISO 14064:2018 categories with GHG Protocol scopes and sub-categories regarding HEI operations
Category
GHG scope
Sub-category
Main activities
C1—Direct
1
A. Stationary Combustion
Burning of fuels for heating, cooling, and electricity generation in campus facilities
1
B. Mobile Combustion
Burning of fuels in vehicles for transportation and operations
1
C. Fugitive Gases
Accidental releases or leaks from equipment and infrastructure
D. Sinks
Activities that capture carbon dioxide
C2—Indirect from purchased energy
2
E. Purchased Electricity
Electricity is sourced from external suppliers to power buildings and facilities
3
F. Purchased Steam, Heating & Cooling
Steam, heating, and cooling services from external providers
C3—Indirect from transportation
3
G. Upstream Distribution & Transportation
Transportation of goods, materials, and supplies to and from campus
3
H. Business Travel
Travel for academic, research, and administrative purposes
3
I. Employee Commuting
Staff and students are commuting daily to and from campus
C4—Indirect from products and processes
3
J. Purchased Goods
Procurement of goods for operational purposes
3
K. Capital Goods
Acquisition of long-term assets and infrastructure
3
L. Fuel & Energy
Upstream emissions of purchased fuels and energy, losses
3
M. Waste
Management, treatment, and disposal
C5—Indirect from products sold
3
N. Upstream Leased Assets
Emissions from assets leased by the HEI for various functions
3
O. Downstream Distribution & Transportation
Transportation and distribution of sold products from the campus to final destinations
3
P. Processing of Products
Emissions from processing and manufacturing activities
3
Q. Use of Sold Products
Emissions from the usage phase of sold products
3
R. End-of-Life of Sold Products
Disposal, recycling, or treatment of products
3
S. Downstream Leased Assets
Emissions from assets leased to the institution by external entities
3
T. Franchises
Operations of franchised businesses
C6—Indirect from other sources
3
U. Investments
From the institution's financial investments and endowments
3
V. Other
Others not previously categorised
Sub-category Sinks (D), in category C1 from ISO 14064:2018, is not yet considered in the GHG Protocol, although it is expected to be included soon. The following section outlines data collection from the different emission sources in Table 1.
The collection of GHG emissions is usually done yearly, and since the main goal is to track emissions over time, a reference year is considered a baseline. Moreover, total emissions may be provided as absolute values or normalised by area or per capita. For example, organisational boundaries encompass all university-controlled facilities and operations, including academic units, support facilities, and research centres. The operational boundaries specify the carbon sources considered in the organisational boundaries, such as building energy usage, transportation emissions from commuting and business travel, and others. The distribution of carbon emissions among the various activities and processes within the HEI follows an impact allocation strategy. The allocation is made by studying the nature of the inventory, looking at the activities, and seeing how they interfere with the boundaries of the system to define what is evaluated. For example, impacts can be allocated to sub-divisions of the system, such as the impact of food only being related to canteen services, while the impacts of nitrogen purchases are only related to laboratory consumption.
Quality should be assessed based on the quality and granularity of the ACs and EFs data. In the case of ACs, it is vital to check the evaluation period—many may have missing information in the sample. The ACs details are important to reduce the number of estimates and better divide the results. On the other hand, the EFs must also be evaluated in terms of time because some impact sources emit differently over time, as with contracted electricity. Tiers were adopted to show the models’ limitations transparently and foster continuous improvement of the models. For this assessment, three classification levels were adopted: tier 1 EF uses default indicators from international databases, tier 2 EF country-specific are used, and tier 3 is site-specific. The higher the tier of the EFs, the better, as they more closely reflect the reality of the system being studied. Data underwent cross-validation against multiple sources where available. The geographical context was a primary consideration in tiering emission factors, with Portuguese data using the same technology classified as tier 3 (highest quality), Portuguese data using different technologies as tier 2, and European or international data as tier 1. Emission factors were assigned quality scores for each criterion, providing transparency regarding potential areas of uncertainty. The tier classification reveals the completeness and limitations of the compiled inventory. This assessment identified priority areas for future data improvement and provides stakeholders with information on the robustness of different aspects of the analysis. Bottom-up approaches typically enable higher-tier EFs through detailed data collection, while top-down approaches can achieve up to tier 2 with regional economic data.
Sensitivity analyses are performed to identify the GHG emission hotspots using the one-at-a-time (OAT) method, which is appropriate for single-year assessment (Heijungs 2010). The OAT systematically varies one parameter (± 10%) while keeping others constant, aiming to understand the parameter’s influence on the total CF. This variation level strikes a balance between capturing meaningful parameter sensitivities without exceeding realistic data uncertainty ranges for institutional data. The top 10 impactful activities were selected as parameters based on their relative contribution to total emissions to ensure the analysis focused on potentially significant factors for decision-making. Consequently, the OAT gives an understanding of critical inventory variables, which must be qualitatively improved with priority in the subsequent reporting cycles. The CF results are reported through multiple analytical perspectives to ensure understanding and validate the methodology's effectiveness. A breakdown by activity type identifies specific emissions sources and their relevance in a particular year to understand the GHG profile and major contributors. Performance indicators are needed as operational efficiency GHG metrics and to facilitate comparisons. The performance, called carbon intensity, is normalised per capita, considering both students and staff. The analysis results identified hotspots and improvement opportunities.
2.2 Case study
2.2.1 Context
The University of Coimbra (UC) in Portugal was selected as a case study to validate the proposed hybrid methodology. First, as one of the world’s oldest HEIs and a UNESCO World Heritage Site, UC presents a complex structure necessary to test the comprehensive CF assessment. Its community has over 32,000 people dispersed across 4 campuses, encompassing over 100 buildings and infrastructures, including 16 food units, 13 university residences, 2 stadiums, 1 theatre, 2 museums, 16 libraries, and extensive research infrastructure. The third reason is the limitation of current reporting of GHG emissions, which is familiar to other Portuguese HEIs. While UC has published the CF in the annual sustainability reports, they only accounted for scope 1 (stationary natural gas consumption and mobile fossil fuel consumption) and scope 2 (electrical energy consumption). Scope 3 has not yet been reported because the existing data collection does not allow for a reliable analysis (University of Coimbra 2023). Over the last 5 years, the average CF published by the UC was 6.04 kt CO2-eq, with a standard deviation of 0.55 kt CO2-Eq. (9.15%). In 2023, there was a carbon intensity of 0.19 t CO2-eq/person. The carbon footprint assessment covers calendar years (January–December) to align with institutional financial and sustainability reporting cycles. Applying the created methodology thus addresses both the scope limitation and transparency gaps in current reporting practices.
2.2.2 Framework application
The organisational boundaries encompass UC’s primary operations, which align with the official documents assessed, including the academic unities (departments, R&D centres), support facilities (social services, cultural extension units), and associated infrastructure. The functional unit is defined as UC’s annual service delivery—including teaching, research, administrative functions, and technology transfer—measured per individual. Normalising emissions data by the size of the institutional community (students and staff) facilitates comparisons with other HEIs from diverse contexts. However, it is important to recognise that variations in organisational structure, research intensity, and disciplinary focus can lead to significantly different emissions profiles, which are not solely determined by community size. While per-capita normalisation offers a useful baseline for benchmarking, meaningful comparisons require careful consideration of each institution’s unique characteristics, research activities, and operational conditions.
Data collection followed a top-down approach since, for most activities, only aggregated data were available in the official documents with no facility-level disaggregation. There were no cut-off criteria; an attempt was made to add as many activities as possible to the inventory. Figure 1 indicates the flows accounted for within each value chain stage, with the number of activities counted in brackets.
Despite best efforts to collect comprehensive institutional data, several specific limitations warrant consideration: such as the (i) temporal coverage was inconsistent across few activities, with some data sources providing continuous annual records while others offered only periodic snapshots; (ii) the top-down approach, dictated by the availability of institutional data, prevented facility-level disaggregation for most activities; (iii) procurement data often lacked the granularity needed to precisely match economic sectors in the EEIO model, requiring allocation assumptions. Supplementary Material A (Table A.3) shows a data quality overview for the UC inventory.
Emission factors were selected for their geographical, technological, and temporal proximity regarding the inventory system, giving preference to national data when available. When local data was not found, EF from similar technological contexts were gathered. In terms of quantified gases, all the EF accounted for at least the three primary GHGs: CO2, CH4, and N2O. The GWPs were sought based on the latest IPCC report, AR6 (2023), but AR5 (2013) and AR4 (2007) were also used. Life cycle stage coverage varied according to the activity data. Production-stage factors (cradle-to-gate) dominated purchased goods and energy sources, while complete coverage (cradle-to-grave) was achieved for transport and capital goods. Disposal-stage factors (gate-to-grave) were used for the waste category, considering open-loop disposal. Complete details of the data inventory, emission factors, and calculation methods are provided in the Supplementary Material B and C. Table B.1. presents the comprehensive activity data for the activities across the 5-year period (2019–2023), including measurement units and data sources. Table C.1 documents all emission factors used, including their life cycle coverage, geographical representativeness, tier classification, and reference sources.
3 Results
3.1 Carbon footprint
Table 2 summarises the estimated CF macro indicators per calendar year. The sample mean was 19.53 kt of CO2-eq/year with a standard deviation of 8.03 kt of CO2-eq/year, and this variation is mainly caused by two factors: the fact that the sample has caught the COVID-19 pandemic and that the academic population has grown by 16.49% since 2019. The maximum CF was 29.53 kt CO2-eq in 2023, 59.69% more than the baseline year (2019). The sixfold increase in construction expenditure is one of the main reasons the CF is so far above the baseline values. As expected, the CF observed in 2020 decreased significantly because of regular activities, corresponding to the lowest amount (9.35 kt CO2-eq). Regarding the intensity, for the sample studied, the carbon intensity is almost 0.63 t CO2-eq/person.year on average. In 2023, it reached 0.87 t CO2-eq/person.year, 37.08% higher than the baseline year. A detailed breakdown of activity-specific emissions is provided in the Supplementary Material (Table D.1), showing the contribution of each activity to the total carbon footprint.
Table 2
Five-year analysis of the CF for the UC
Year
Total
Intensity
Value chain (%)
Activity (%)
kt CO2-eq/year
t CO2-eq/person.year
Upstream
Own operation
Downstream
PB
EEIO
2023
29.53
0.87
96.74
2.91
0.35
55.51
44.49
2022
25.29
0.80
96.26
3.29
0.44
60.17
39.83
2021
15.01
0.52
95.01
4.37
0.62
68.80
31.20
2020
9.35
0.32
90.36
8.10
1.54
75.08
24.92
2019
18.49
0.64
92.67
6.17
1.16
84.51
15.49
As HEI is not a manufacturing industry, most CF comes from upstream (and indirect) activities, representing, on average, 94.21% of the total CF. On the other hand, the EEIO contribution increases while the PB amounts decline regarding the total CF. This shift reflects fundamental changes in the university operations, remarkably increasing capital investment in infrastructure and equipment. While EEIO analysis provides a valuable scope 3 assessment, several limitations should be highlighted. Sector aggregation in EEIO models may misrepresent specific activities in HEIs, such as potentially overestimating construction impacts while underestimating specialised research equipment emissions (Lenzen 2011). Additionally, temporal discrepancies and price variations across economic sectors can affect accuracy. Hybrid approaches mitigate these limitations by prioritising process-based data for major emission sources, while using EEIO data to fill gaps where physical data is unavailable (Crawford et al. 2018). Comparing the results with similar studies reveals important patterns. While the carbon intensity of UC (0.87 t CO₂-eq/person) falls within the lower range of global values (0.06–10.94 t CO₂-eq/person) reported by Valls-Val and Bovea (2021), it exceeds the average for Portuguese institutions (0.51 t CO₂-eq/person). This difference primarily stems from our comprehensive scope 3 accounting, which represented 78.38% of total emissions, similar to findings from Herth and Blok (2022) and Cano et al. (2023), who reported scope 3 contributions around 80%. The relative contribution of emission sources also shows interesting variations. While transportation consistently emerges as a major source across studies, our finding of construction activities as the dominant source (40.35%) differs from Larsen et al. (2013), who identified electricity consumption as the primary contributor. This variation likely reflects differing construction activity levels, electricity grid carbon intensities, and inventory completeness across institutions.
Direct emissions (scope 1) stayed relatively stable, ranging from 0.04 to 0.03 t CO2-eq/person.year, representing, on average, 4.97%. Purchased electricity (scope 2) CF declined from 20.99% in 2019 to 8.87% in 2023. Even though the consumption increased, the national EF decreased, reflecting the progressive decarbonisation of Portugal’s electricity grid. The heavy contributions came from indirect emissions from product use and transportation (scope 3), collectively accounting for 42.46% and 35.92% on average, respectively. While transportation absolute CF remained about 0.3 t CO2-eq/person.year (without 2020 and 2021), the product used grew from 0.16 to 0.47 t CO2-eq/person.year in 5 years. Figure 2 provides a detailed view using GHG Protocol sub-categories to identify operational patterns and shifts. The most noticeable change was in capital goods, which rose from the third position (15.49%) in 2019 to the dominant contributor (44.49%) in 2023; in absolute values, it presented about a threefold increase from 2019. Transportation relative share decreased from 47.47 to 33.62%, remaining the second-largest contributor. Purchased electricity experienced the most substantial relative decline, from 20.99 to 8.85%, dropping one position. Purchased goods showed consistent growth, climbing from 5.64 to 8.05%, surpassing stationary combustion (which declined from 6.57 to 3.26%) to become the fourth largest contributor. The purchased goods also increased in absolute terms by 127.89%, reaching 2.37 kt CO2-eq/year. The remaining sub-categories—fuel and energy, mobile combustion, and waste—maintained relatively stable but minor contributions. The effect of the sinks was residual, capturing, on average, 0.99% of the total CF per year.
Fig. 2
Evolution of CF by GHG Protocol sub-categories from 2019 to 2023
A detailed examination of individual activities for the most recent year (2023) was conducted to achieve the highest level of granularity in the analysis (Fig. 3). Construction was the dominant CF source (11.91 kt CO2-eq/year), highlighting how infrastructure significantly impacts. Commuting activities from gasoline cars (6.40 kt CO2-eq/year) represented the second-largest contributor, followed by public transportation (2.64 kt CO2-eq/year), electricity consumption (2.61 kt CO2-eq/year), and retail trade (1.23 kt CO2-eq/year), closing the top five. Food-related activities, while individually modest, collectively accounted for 2.34 kt CO2-eq/year, with animal protein sources as the most significant CF sources—swine (0.70 kt CO2-eq/year), fish (0.58 kt CO2-eq/year), and chicken (0.36 kt CO2-eq/year). Infrastructure support activities like water consumption (0.02 kt CO2-eq/year) and waste management (0.10 kt CO2-eq/year) show relatively low CF. The CF from photovoltaic generation (0.02 kt CO2-eq/year) enabled long-term emission reductions, even though the production is residual compared to the UC demand. Some activities, such as electronic waste, medical waste, and upstream gasoline, were excluded due to their relative nullity.
Fig. 3
CF of UC in 2023 (kt CO2-eq) classified by emissions from the value chain to ISO 14064:2018 and GHG protocolThe OAT sensitivity analysis varied each activity individually by ± 10% to identify hotspots. Construction emerged as the most sensitive contributor (± 4.04%), followed by transportation modes, with gasoline cars (± 2.17%) and buses (± 0.90%) showing significant influence. Electricity consumption (± 0.89%) and retail trade (± 0.42%) demonstrated moderate sensitivity, while natural gas (0.38%), swine products (0.24%), and fish consumption (0.20%) showed lower sensitivity levels. Electric cars (0.16%) and scooters (0.14%) exhibited minimal impact on overall carbon footprint fluctuations despite their inclusion in the model. As predicted, hotspot activities showed the highest CF, with upstream emissions accounting for the top 5
The findings from this study have several important implications for HEI sustainability policies. First, the significant contribution of scope 3 emissions (78.38% of total) suggests that carbon reduction strategies focused exclusively on campus operations will have a limited impact on overall institutional carbon footprints. Effective policies must address the entire value chain, including procurement practices, commuting infrastructure, and construction specifications. Construction activities, as the dominant emission source (40.35%), represent a critical area for policy intervention. Institutions should implement sustainable construction requirements that consider embodied carbon alongside operational energy efficiency. For existing infrastructure, extending building lifespans through renovation rather than replacement can significantly reduce construction-related emissions. Commuting emissions (32.18%) highlight the importance of sustainable transportation policies. Institutions should prioritise accessible public transportation, cycling infrastructure, carpooling programmes, and strategic scheduling to reduce commuting frequency. The substantial contribution of gasoline vehicles suggests that policies promoting alternative transportation modes could yield significant emission reductions. Food procurement (8.05%) offers another policy opportunity, with plant-based menu options potentially reducing food-related emissions without requiring significant infrastructure investments.
Policy development should focus on high-impact, low-implementation-barrier activities. For construction emissions, this includes establishing green building standards for new projects, prioritising renovation over new construction, and specifying low-carbon materials. For commuting emissions, early priorities include developing comprehensive mobility surveys to establish detailed baselines, installing electric vehicle charging infrastructure, and enhancing bicycle storage facilities. Medium-term implementation should focus on infrastructure development and procurement policy revision. This includes transitioning campus fleets to electric vehicles, implementing sustainable procurement guidelines with carbon intensity criteria, and developing IT infrastructure to support remote work and learning. Long-term strategies should address structural systems, including campus energy systems, spatial planning to reduce travel needs, and integration of carbon accounting into routine financial reporting. Throughout implementation, regular monitoring and verification are essential to track progress and refine strategies based on measured outcomes.
3.3 National benchmarking
Comparing carbon footprints across higher education institutions (HEIs) poses inherent methodological challenges that demand cautious interpretation. Although per-capita normalisation is widely adopted in HEI carbon assessments (Helmers et al. 2021; Larsen et al. 2013; Valls-Val and Bovea 2021), it has well-documented limitations when applied to institutions with fundamentally different operational characteristics. A review of Portuguese HEIs shows that carbon intensity varies significantly across institutions (0.20 to 1.66 t CO₂-eq/person) due to factors beyond community size (Deda et al. 2023). Research-intensive universities with extensive laboratory facilities, engineering schools with energy-intensive equipment, and medical institutions with specialised infrastructure typically exhibit higher per-capita emissions than liberal arts colleges or business schools, regardless of student and staff numbers. Additionally, factors such as campus age, geographic location, energy grid carbon intensity, and institutional policies significantly influence emission profiles independently of community size. Despite these constraints, per-capita comparison remains the most established approach in the field due to the absence of more sophisticated normalisation methods that adequately account for institutional diversity (Deda et al. 2023).
In this study, benchmarking regarding the CF of some Portuguese HEIs was performed based on online official information. The benchmarking analysis was conducted with careful consideration of methodological differences between institutions. To ensure meaningful comparisons, harmonisation steps were implemented. First, carbon footprints were normalised per person (students and staff) to account for institutional differences. Second, emissions were categorised according to the GHG Protocol scopes and sub-categories to enable structured comparison despite varying inventory approaches. Table 3 shows the CF of those 6 HEIs, with the systems, sub-categories, and community (students and workers) displayed. In this case, the UC CF compared was the one calculated in this study.
Table 3
CF intensities of Iberian HEIs in specific years regarding GHG scopes
Institution
Year
People
System
Sub-categories considered
Ref
University of Lisbon (UL)
2019
6,161
One campi
Mobile combustion, fugitive gases, purchased electricity, business travel, commuting, purchased goods, waste
(Faculdade de Ciências da Universidade de Lisboa 2019
)
University of Coimbra (UC)
2023
33,878
HEI
Stationary and mobile combustion, carbon sinks, purchased electricity, commuting, purchased goods, capital goods, fuel & energy, waste
University of Minho (UM)
2022
23,485
HEI
Mobile combustion, purchased electricity, commuting, purchased goods, waste
When comparing institutions with different system boundaries, direct comparisons were limited to matching scopes (e.g. comparing only scope 1 and 2 emissions when scope 3 data were unavailable for certain institutions). Methodological differences were explicitly documented in Table 3, which includes information on system boundaries, institution size, and inventory data. This transparency allows understanding the limitations of cross-institutional comparisons. For institutions with partial data, no extrapolation was attempted, and comparisons were qualified with appropriate caveats about data limitations. While some assessments cover entire institutions (UC, UM, NOVA), others focus on specific campuses (UL) or faculties (UP). Although UP was the first to carry out these calculations, the CF was last published in 2012. The scope of included activities also differs substantially, with UL and the studied case study from UC presenting the most comprehensive assessments. Most institutions limit their analysis to basic operational emissions, mainly focusing on electricity consumption and direct fuel use. In addition, little information was found on the activities and emission factors considered, except for the IPC, which carried out a public study on CF. Figure 4 presents the carbon intensities disaggregated by scope, revealing significant variations in reported emissions.
Fig. 4
Carbon intensity comparison across Portuguese HEIs
Based on the existing information for each scope, and considering the last year reported, the national average carbon intensity is 0.77 t CO2-eq/person.year, with scope contributions averaging 0.15, 0.38, and 0.47 t CO2-eq/person.year for scopes 1, 2, and 3, respectively. The values are within the global lower limit in the literature, as global intensity ranges from 0.06 to 10.94 t CO2-eq/person.year (Valls-Val and Bovea 2021). NOVA reports the highest intensity (1.58 t CO2-eq/person.year), closely followed by UL (1.55 t CO₂-eq/person.year), while UC assessment (0.87 t CO2-eq/person.year) occupies the third position. These higher values likely reflect more comprehensive scope 3 accounting rather than operational differences. UM demonstrates intermediate intensity (0.66 t CO2-eq/person.year), while UP, IPC, and ISCTE report significantly lower values. It was impossible to find information on the hybridisation of the inventory of these national results.
The effectiveness of the hybrid methodology studied is also demonstrated by comparing the case study results with UC's official reporting. While UC’s sustainability reports show relatively stable emissions (ranging from 5.34 to 6.59 kt CO2-eq/year between 2019 and 2023), this study reveals substantially higher values by incorporating comprehensive scope 3 accounting. This difference does not indicate inaccuracy in official reporting but highlights how traditional approaches may underestimate institutional CF. On the other hand, all the data used is public, so the UC CF could have been estimated in the latest reports. The case study results align with other Portuguese HEIs’ carbon intensities while providing unprecedented transparency in emission sources, activities, and emission factors. By leveraging existing institutional data through our hybrid framework, significant indirect emissions that traditional reporting overlooks were identified.
Despite benchmarking efforts, detailed information about inventory composition, specific activities, and emission factors was not publicly available for most Portuguese HEIs, revealing a lack of methodological maturity in carbon footprint assessment within the national higher education sector. This transparency gap limited the ability to conduct more rigorous comparative analyses and highlights the need for standardised reporting protocols that include detailed methodological documentation.
4 Discussion
Integrating the GHG Protocol and ISO 14064:2018 standards represents a significant methodological advancement for HEI CF assessment. While these standards are often applied separately, their combination provides complementary strengths: the GHG Protocol’s precise scope-based categorisation enhances comparability, while ISO 14064:2018's detailed inventory management enables comprehensive activity tracking. The hybrid approach offers distinctive advantages through its systematic handling of data gaps using economic indicators, improved accuracy through tiered data quality assessment, and flexibility in accommodating varying levels of data availability across institutions. Additionally, it provides enhanced transparency in emission source categorisation whilst effectively integrating process-based and economic-based methodologies into a cohesive framework.
The comparative analysis in Table 4 demonstrates that although recent studies have addressed CF in HEIs, no model has efficiently unified the leading standards in the field, along with a hybrid approach to assessing inventory quality. For instance, Herth and Blok (2022) used a mostly economic model (EEIO), limiting accuracy to direct activities (Herth and Blok 2022). Kiehle et al. (2023) used EEIO data only for some missing categories, as in the present study (Kiehle et al. 2023). Cano et al. (2023) applied the GHG Protocol in integration with ISO 14064:2018 but did not incorporate the hybrid model, which was detrimental to estimating indirect impacts (Cano et al. 2023). This study overcame these limitations by integrating both approaches into a validated framework, ensuring cross-standard categorisation accuracy and inventory qualification to highlight the primary sources of uncertainty and pave the way for future improvements.
In terms of practical implementation, this methodology offers significant advantages over previous approaches that often require extensive data collection infrastructure. It utilises readily available institutional data, provides clear guidance on data quality assessment, and enables progressive improvement in assessment quality. The methodology’s ability to utilise standard institutional reporting data makes it remarkably adaptable for other HEIs, potentially establishing a more comprehensive baseline for sector-wide carbon management. It created a framework for meaningful inter-institutional benchmarking.
This study emphasises methodological transparency through comprehensive documentation of data sources, emission factors, and calculation procedures in the Supplementary Material, enabling verification and replication of results. Despite the effort to develop the inventory, the data gathered for the case study is far from complete. The data for the nine GHG Protocol sub-categories evaluated allowed the best possible assessment to be determined. However, some relevant underreporting data is due to the lack of accounting for some activities, mainly in procurement. There are twelve other sub-categories whose information was unavailable, many of which may be significant (e.g. business travel and fugitive gases). In addition, one relevant limitation of this study was the high level of data aggregation since it was often impossible to discriminate between campuses, faculties, and buildings of UC. Thus, increasing inventory details in the future is essential to facilitate decision-making regarding monitoring and mitigation strategies.
While EEIO analysis provides a valuable scope 3 emission assessment, we encountered specific limitations in the case study application. For construction activities, EXIOBASE economic sector “Construction” aggregates diverse activities from basic renovations to specialised laboratory construction. This aggregation likely overestimated emissions for UC’s particular building profile, which included significant heritage building renovations with lower material intensity than standard construction. Conversely, for research equipment (categorised under “Retail trade”), EEIO models underestimated emissions from specialised laboratory equipment production, as these energy-intensive manufacturing processes are diluted within broader sector averages.
A key strength of this study is its transparent framework, which allows other HEIs, particularly in Portugal, to leverage this inventory for benchmarking. Improving data granularity, increasing the use of tier 3 factors, and refining scope 3 coverage will further strengthen the accuracy of future assessments, supporting more effective carbon management strategies in HEIs. Expanding institutional databases and improving supplier engagement will enhance data completeness. This methodological choice ensures that the inventory reflects national conditions while balancing accuracy and data availability. This work demonstrates that comprehensive CF assessment in HEIs is achievable using existing institutional data when structured through an appropriate methodological framework. The findings provide practical guidance for other institutions and a foundation for more standardised sector-wide carbon accounting approaches.
5 Conclusion
The CF methodology proposed in this research overcomes data collection limitations, facilitating a comprehensive assessment of GHG emissions from internal operations and supply chains. The inventory work highlighted that significant information is essential for calculations, though data about the current system and estimates were limited. Integrating EEIO analysis proved valuable in quantifying often hidden impacts (such as lack of data and respective impacts), revealing that every euro spent may contribute to the CF.
The findings highlight the critical role of scope 3 emissions, which, on average, accounted for approximately 78.37% of the emissions at UC. This study demonstrated that UC’s CF is more than three times higher than the previously available official numbers (without scope 3 emissions), emphasising the necessity of including indirect emissions from procurement, transportation, and other value chain activities. However, this was already expected as most HEIs do not account for scope 3 emissions as systematically as the one proposed in this work. Despite the estimation of carbon sinks, they were residual compared to the emissions in this case.
The hybrid approach presented in this study demonstrates significant potential for adaptation across different institutional contexts. For smaller institutions with limited resources, the methodology could be simplified by focusing on key emission hotspots identified here (construction, commuting, energy) while using sector averages for less significant categories. Conversely, research-intensive institutions might expand the methodology to include detailed laboratory equipment and specialised research infrastructure inventories. With more detailed data, several methodological improvements become possible, such as the following: (i) building-level energy monitoring would enable space-use efficiency analysis and targeted interventions; (ii) comprehensive travel surveys could provide deeper insights into commuting patterns, enabling tailored transportation policies; (iii) detailed procurement records with product-specific information would allow more precise process-based emission calculations instead of sector-level EEIO estimates. As data availability improves, the methodology could evolve from the current hybrid approach toward more comprehensive process-based assessment, potentially reducing uncertainty while maintaining system completeness.
The research also identified limitations and challenges. Refining the model through sensitivity analyses and scenario projections will enhance its robustness, yielding more reliable CF assessments. Increasing the inventory quality and conducting further sensitivity studies are essential next steps to improve the model's reliability. Comprehensive assessments should also account for other flows, with high activities, to accurately balance the mass and energy system.
On the other hand, the validation could be implemented for different HEIs in the national context to help them with their decarbonisation plans. In conclusion, addressing these research limitations will significantly advance environmental knowledge and sustainability in HEIs. By adopting refined and context-specific methodologies, HEIs can achieve more accurate, reliable, and auditable CF results, promoting informed decision-making and enhancing environmental performance.
Declarations
Competing interests
The authors declare no competing interests.
Use of Generative AI
During the preparation of this work, the author used Claude to assist with language editing, formatting suggestions, and structuring the manuscript. After using this tool, the author reviewed and edited the content as needed and took full responsibility for the content of the published article (Claude Sonnet 4).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Cano N, Berrio L, Carvajal E, Arango S (2023) Assessing the carbon footprint of a Colombian University Campus using the UNE-ISO 14064–1 and WRI/WBCSD GHG Protocol Corporate Standard. Environ Sci Pollut Res 30(2):3980–3996. https://doi.org/10.1007/s11356-022-22119-4CrossRef
Helmers E, Chang CC, Dauwels J (2021) Carbon footprinting of universities worldwide: Part I—objective comparison by standardized metrics. Environ Sci Eur 33(1). https://doi.org/10.1186/s12302-021-00454-6
Helmers E, Chang CC, Dauwels J (2022) Carbon footprinting of universities worldwide part II: first quantification of complete embodied impacts of two campuses in Germany and Singapore. Sustainability 14(7). https://doi.org/10.3390/su14073865
Herth A, Blok K (2022) Quantifying universities’ direct and indirect carbon emissions – the case of Delft University of Technology. Int J Sustain High Educ 24(9):21–52. https://doi.org/10.1108/IJSHE-04-2022-0121CrossRef
International Organization for Standardization (2018) ISO 14064–1:2018 Greenhouse gases—part 1: specification with guidance at the organization level for quantification and reporting of greenhouse gas emissions and removals. https://www.iso.org/obp/ui/en/#iso:std:iso:14064:-1:ed-2:v1:en. Accessed 25 July 2024
International Standard Organization (2018) Specification with guidance at the organization level for quantification and reporting of greenhouse gas ( ISO 14064:2018—Part I) (Vol. 2018)
Kiehle J, Kopsakangas-Savolainen M, Hilli M, Pongrácz E (2023) Carbon footprint at institutions of higher education: The case of the University of Oulu. J Environ Manage 329. https://doi.org/10.1016/j.jenvman.2022.117056
Li Z, Chen Z, Yang N, Wei K, Ling Z, Liu Q, Chen G, Ye BH (2021) Trends in research on the carbon footprint of higher education: a bibliometric analysis (2010–2019). J Clean Prod 289. https://doi.org/10.1016/j.jclepro.2020.125642
Osorio AM, Úsuga LF, Vásquez RE, Nieto-Londoño C, Rinaudo ME, Martínez JA, Filho WL (2022) Towards carbon neutrality in higher education institutions: case of two private universities in Colombia. Sustainability 14(3):1–24. https://doi.org/10.3390/su14031774CrossRef
Paredes-Canencio KN, Lasso A, Castrillon R, Vidal-Medina JR, Quispe EC (2024) Carbon footprint of higher education institutions. Environ Dev Sustain 26(12):30239–30272. https://doi.org/10.1007/s10668-024-04596-4CrossRef
Robinson OJ (2017) A universal method for accounting greenhouse gas emissions from the activities of higher education institutions using a hybrid life-cycle approach. University of Southampton.
Schoenaker N, Delahaye R (2017) Environmental input-output analyses for the Netherlands. 05121, 1–30
Valls-Val K, Bovea MD (2021) Carbon footprint in higher education institutions: a literature review and prospects for future research. Clean Technol Environ Policy 23(9). https://doi.org/10.1007/s10098-021-02180-2
World Business Council for Sustainable Development, & World Resources Institute (2004) The greenhouse gas protocol: a corporate accounting and reporting standard.
Yue D, Pandya S, You F (2016) Integrating hybrid life cycle assessment with multiobjective optimization: a modeling framework. Environ Sci Technol 50(3). https://doi.org/10.1021/acs.est.5b04279
