Introduction

In 2023, immediately prior to the UN COP28 meeting, the United Nations (UN) Environment Programme issued a chilling assessment of the state of the world’s environment1. Against targets to limit global warming to 1.5 °C2,3, they predict that current activities will result in global greenhouse gas (GHG) emissions 55% higher than the levels required to meet this target. Despite a significant slowing down in the rate of increase, GHG emissions globally will still rise by a predicted 1.1% in 2023, a record high4. At this rate of increase, the world has a mere seven years from 2023 before the chances of limiting global warming to 1.5% are less than 50%4.

The only way to bridge the gap is by taking ‘immediate and unprecedented mitigation action in this decade’ (p28)1 by totally eliminating GHG emissions5. These mitigation actions generally focus on decarbonising sectors of the world economy that are responsible for the majority of GHG emissions, principally the energy and agricultural sectors, together with consideration of the impact of ‘multi-parametric’ factors such as aviation, economic growth, digitalisation and the development of eco-technologies6. Successful mitigation requires moving energy production to renewable sources as well as enhancing both existing and novel carbon capture technologies7.

Whilst actions to reduce GHG emissions require action from national and supranational entities, policy makers8,9 also recognise that a range of diverse stakeholders, including local communities, should all play an active role in climate mitigation. Their actions can assist in meeting a target of reducing the average emissions per capita, for example, in Europe, to 2.4 tonnes by 203010 from 7.7 metric tonnes in 202111. University scientists and academic societies form one of these stakeholder groups. They are hypermobile, responsible for considerable GHG emissions12, and have been classed as a medium priority13 in efforts to decarbonise their activities. Indeed, hypermobility in the field of sustainability research has been described as an ‘environmental curse’14 which is at odds with the implications from the findings of climate studies carried out by those same scientists.

Academic activity includes scientists acting as ‘knowledge brokers’15 with significant influence in society. Universities and academics facilitate relationships, interaction and engagement so that the results of their scientific endeavours can be applied by policymakers. One of the traditional methods used by universities and academic societies to broker such knowledge has been the conference. Conferences are social spaces16 whereby researchers, educators and policymakers come together from different countries to network, gain and disseminate new knowledge, foster collaboration including for article writing and funding bids17, spread good practice and reduce knowledge and clinical practice inequalities.

Face-to-face meetings and conferences16 often require participants to travel internationally to achieve their social goals. For example, in 2019, business travel emissions, including academics travelling to conferences, comprised approximately 17% of Stanford University’s total carbon footprint and 22% of their external emissions footprint18. In one detailed Belgian case study, conference attendance was responsible for 35% of a researcher’s total life cycle carbon footprint in the production of their PhD19. As noted by Leochico and colleagues, ‘large scientific conferences aimed at enhancing health-related knowledge, ironically, often implies leaving huge carbon footprints, a practice that may seem contrary to what healthcare professionals are advocating’20 (p1).

One solution—exclusively online, hybrid, zonal/local, or other ‘Nearly Carbon Neutral’ events21—will lead to the most dramatic reductions in GHG emissions22. Indeed, many academics adapted to the travel and face-to-face interaction restrictions during the 2020 COVID-19 pandemic by using online platforms for teaching and research. One study that modelled a continuation of these behaviours post-pandemic showed that GHG emissions from academic activities could be reduced by almost 50% compared to pre-pandemic behavioural norms23. However, a lot of academics think that these behaviours are only suitable for international administrative meetings24. Despite optimism from some technology advocates25, many academic participants do not believe that virtual communications can achieve the networking and educational objectives of scientific conferences and educational events21,24,26,27. Participants at conferences report benefitting greatly from face-to-face interactions, and by informal social connections made outside the scheduled programme. This belief is so entrenched, that in one study, when presented with three choices—transportation mode shifting (from aviation to ground-based), moving to virtual conferencing, abstention from involvement—it was transportation mode shifting that was rated most acceptable by respondents24. However, despite these views, in an age of low-cost airlines, rail is often a more expensive alternative to flying. In a recent analysis28, 79 of 112 routes analysed were more expensive by rail than by aeroplane, by as much as 10–30 times for routes between major European cities, providing a significant disincentive to transportation mode shift, especially for young, early career researchers and teachers.

Many academics are acutely aware that they should take action to reduce their environmentally damaging activities29. As much of our knowledge on the climate emergency comes from the work of scientists employed by universities, one would imagine that academics, particularly climate scientists, are one of the communities best placed to take the lead in reducing GHG emissions from their activities30. In common with other academics, such scientists are indeed well aware of their responsibilities and accept that they should advocate for climate change mitigation31. They are also aware that they will rapidly lose public credibility if their own behaviours are in sharp contrast to their scientific work32. However, these same researchers find it very difficult to actually make such radical changes to their own behaviours29,33. Academics need a comprehensive and theoretically coherent strategy to enable and motivate them to enact such behavioural changes.

Balancing social good with environmental harm is a core challenge that must be faced by academic societies before 2030. This issue has been addressed by several authors34,35,36,37, including a published ‘Cercedilla Manifesto’35, and in a review of 11 articles which assessed the GHG emissions produced by academic conferences20. Strategies that academic societies could take to reduce their GHG emissions range from moving conferences entirely online, through hub and spoke and hybrid conferences, to face-to-face venue choices and on-site strategies, including sustainable food procurement choices, sourcing power from renewable energy companies, and so forth. Whilst these works are helpful in highlighting potential strategies that can be taken by event organisers or employers36, they are not focussed on the key issue of individual participant behaviour change. Despite some claims that conference travel by academics is unrelated to their productivity38, conference attendance is so deeply embedded in academic culture that whilst organisers must make corporate changes, it is also necessary for individuals to then take advantage of the opportunities brought about by these changes.

This individual behavioural conundrum is well understood in healthcare. Many effective treatments and health interventions provided by health systems are not adopted by those who might best benefit from them. Type 2 diabetes, for example, can often be prevented and reversed by lifestyle choice, but despite this knowledge, many with the condition or at risk of developing diabetes do not take up or maintain these proven strategies and programmes39,40. Likewise, handwashing by healthcare professionals is known to reduce infection rates in hospitals, yet compliance by doctors, nurses and other health workers can be less than 50%41. Change is hard to do and even harder to maintain.

As a consequence, behaviour change strategies within the related field of ‘implementation science’42 have been developed to reinforce desired changes in behaviour. One influential model is ‘COM-B’ which postulates that for individuals to change their Behaviour, they need the Capability to do so, along with the Opportunity and personal Motivation43. Illustrating this model is the ‘behaviour change wheel’, in which specific strategies, including policy action, are identified to boost capability, opportunity and motivation, thereby enabling the desired behaviour change. The use of COM-B has been shown to be effective in many situations, for example, in increasing physical activity in survivors of cerebrovascular accidents44. Although potentially suitable to address the observed ‘motivation-impact gap’ in helping people to make pro-environmental choices45, the COM-B model has not yet been applied to the difficult task of changing academic behaviours associated with attending international conferences, educational events and networking meetings. In this article, we take an explicit behaviour change perspective on the problem of academic conference GHG emissions. Our aim was to estimate the environmental impact of two exemplar in-person academic events for PhD students held in Ghent, Belgium and Lisbon, Portugal, and, using the COM-B behaviour change framework of capability, opportunity and motivation, to identify strategies that could be applied by organisers and participants to reduce this impact. We wanted to know: (1) what are the estimated GHG emissions in carbon dioxide equivalent attributable to two differently located European summer schools for PhD students; (2) what factors are the main contributors to these emissions; (3) what are the differences in financial costs to student and faculty participants of attending differently located events; (4) what strategies from the COM-B framework on behaviour change can be mapped against sources of GHG emissions from academic events to potentially reduce such emissions?

Results

Distances travelled

In Ghent 2018, 122 participants attended for between three to 12 days each. According to the assumptions detailed in our methods section, 111 (90.1%) of participants would have used air travel for some of their journey (Table 1). Similarly, in Lisbon 2018, there were 137 participants, of whom 130 (94.9%) would have flown. The average return distance participants would have travelled by air to Ghent was 1,565.85 km, (SD 1,087.37) and by total rail journeys, 347.34 km (235.75). Corresponding figures for Lisbon were 3,904.88 km, (SD 1,693.27) for flights and 117.79 km (159.31) for rail.

Table 1 Estimated travel mode and distances in km to individual events and in total
Full size table

GHG emissions

The mean per participant GHG emissions for travel (Table 2) would have been 248.80 kg CO2e (SD 166.16) for Ghent and 593.90 (255.62) for Lisbon, to which air travel contributed the greater proportion compared to rail (95.05% for Ghent and 99.30% for Lisbon). Average on-site GHG emissions were 85.34 kg CO2e (SD 39.49) per participant for Ghent and 129.80 (64.69) for Lisbon. On-site emissions contributed 25.55% of the GHG costs in Ghent and, although higher in Lisbon (driven by higher emissions profiles for the hotels provided to participants and the energy-intensive teaching rooms), only 17.94% of the total GHG emissions due to the overall GHG emissions being inflated by higher emissions from travel. In total, the GHG emission cost of the Ghent event was 40.77 tonnes CO2e and Lisbon 99.15 at an average per participant of 334.14 (SD 173.34) kg CO2e for Ghent and 723.70 (262.49) for Lisbon.

Table 2 Estimated GHG emissions in kg CO2e for individual events and in total
Full size table

Financial costs

The total financial cost of travel and accommodation combined for the Ghent event was a mean of €825.01 (SD 303.08) and for Lisbon €761.07 (286.14) (Table 3). Flight costs to Lisbon were higher than Ghent, but were outweighed by higher hotel and rail costs for the Ghent venue.

Table 3 Estimated current travel and on-site financial costs in Euros for individual events and in total
Full size table

Discussion

We estimated that the total GHG emission cost of annual academic summer schools for European PhD students was either 41 tonnes or 99 tonnes per event, representing a per participant mean of 0.334 (SD 0.173) or 0.724 (SD 0.263) tonnes CO2e. The main driver of these emissions and their variation was the distance travelled from the participants’ home base to the venue location. On average, participants had to travel around 2.5 times further to the venue on the geographical periphery of Europe compared to the one situated more centrally. Given the time involved in using ground-based travel options, a mean distance of almost 4000 km would have required most participants to choose air travel, at a GHG emission rate 4.3 times that of rail. Although making a much smaller contribution overall, the geographically peripheral location was in a part of Europe where the venue GHG emissions were also higher, which increased the difference between event emissions.

To put the GHG emissions into perspective, for these events, the daily GHG footprint per person per day of attendance was 0.04 (Ghent) and 0.11 (Lisbon). In 2021 the European Union annual per-capita GHG emission footprint was 7.7 tonnes or 0.02 tonnes per day11. Therefore, every person attending would have a GHG emission footprint of two to 5.5 times higher than the daily average for 2021. Compared to the European Commission’s 2030 target of 2.3 tons of CO2 per person per year10, these figures are either 5.7 or 16 times higher per person per day. Or to put it another way, in total these events generated the equivalent of either 5.5 or 13.5 European citizens’ annual 2021 GHG emissions, representing 16 or 40 citizen’s annual emissions against the 2030 target. In citing the above figures, it is worth remembering that the per-capita emissions profiles above are averages across a population, and that individual emissions are skewed towards elites, including academics that travel to conferences, and those living in more wealthy European countries. As such, our estimates may actually contribute to existing ‘polluter elite’ figures, which are likely to be higher than the averages we have cited.

Looking ahead to 2030 and beyond, rapid European renewable and efficiency energy investment to meet the EU 2030 GHG reduction targets10 will significantly reduce on-site emissions from power usage (at least 30% according to some estimates)46. However, even if sites could be classed as ‘carbon-neutral’, the contribution of on-site emissions to the overall GHG emissions profile of the academic events analysed here is at most 26%. Therefore, such changes, albeit very welcome, will still leave 74% or more of GHG emissions present, reinforcing the overwhelmingly dominant contribution of aviation GHG emissions to the overall GHG emissions profile. Given the lack of decarbonising options for aviation, reducing kilometres travelled by flying will be the most effective strategy to reduce future academic event GHG emissions.

Financial costs were quite similar between the events. Although flight costs were more for Lisbon, Ghent had higher costs due to the price of hotels chosen by the event organisers and higher airport to venue rail costs. Whereas Lisbon has its own airport linked by efficient urban transfer systems, Ghent requires additional train travel from the nearest two airports to hotels. Nonetheless, the difference was only around €60 per person on average out of a mean total of between €761-€825 per event.

The results of our analysis show that for individual academic face-to-face conference participants to make less-polluting choices, event organisers must provide them with the structural opportunity to do so. Here, opportunity is defined as “all the factors that lie outside the individual that make the behaviour possible or prompt it.” (p4). For face-to-face scientific events, venue location and sustainable on-site components are the ‘Opportunity’ element of the COM-B behaviour change model43.

Notwithstanding utilising digital-first technologies47 and advocacy from digital enthusiasts48, organisers may believe that totally virtual events will not deliver against the event objectives. Therefore, they must do all they can to facilitate participants who attend face-to-face events to reduce their GHG emissions, mainly through transportation mode-shifting. Our findings demonstrate that to reduce GHG emissions from face-to-face events, event organisers must first choose venues that are as close to the home base of as many participants as possible in order to provide them with the opportunity to reduce their overall GHG travel emissions. Organisers must also find places where there are rich alternative ground-based transport options such as high-speed electrified train lines and good onward connections. In terms of venue characteristics, progress on renewable energy investment to meet the EU 2030 GHG reduction targets10 is not equally distributed across different countries49, so choosing countries that have a greater proportion of renewable energy production compared to fossil fuels as event hosts will reduce the contribution of sites to future event GHG profiles. Choosing superior energy efficiency performance accommodations that source energy from renewables will reduce event GHG emissions. So, too, will organisers providing meals from plant-based sources only, including seasonal and field-grown vegetables, given these have significantly lower GHG emissions compared with meat and dairy products50. As part of the application process, organisers may contract with ethical and explicitly sustainable travel agencies that provide bookings for travel and accommodation with lower GHG emissions characteristics.

Opportunity, however, is insufficient. Even with these structural opportunities being present, as noted in other research33, the majority of participants will, of necessity and habit, likely default to current standard travel, hotel and dietary choices with a significant GHG emissions profile26. Two other factors are necessary to change individual behaviour: capability and motivation. Unless participants’ capabilities and motivation to make more environmentally sustainable choices are enhanced, it is likely that many will not do so.

Capability is defined as “the individual’s psychological and physical capacity to engage in the activity concerned. It includes having the necessary knowledge and skills.”43 (p4). Notwithstanding the need to ensure that events are accessible to all, we would assume that the majority of participants will most likely have the physical capability to undertake a less GHG-emitting ground-based journey. However, even if theoretically willing to do24, they may not have some psychological capabilities, particularly in terms of knowledge about scheduling, routes and connections. Exactly how to make lower emissions choices may be a barrier to behaviour change given that it is much more difficult to put together a rail or other ground-based journey across international borders than to book a flight. Individual train company websites may not link connecting trains, and routes may appear opaque. In order to address this knowledge deficit, organisers can remedy knowledge gaps by providing resources to participants during the application process. A number of websites exist (e.g. www.seat61.com) that provide extremely detailed guidance on train-based travel. These can be made available for participants to use when planning their journeys, addressing the psychological capability deficit.

The final element of the COM-B behaviour change model is motivation, defined as “all those brain processes that energise and direct behaviour, not just goals and conscious decision-making. It includes habitual processes, emotional responding, as well as analytical decision-making.”43 (p4). Flying to academic meetings might be classified as a habitual process, which, despite opportunity and capability elements, will not change unless people are motivated to do so24,26,27.

As well as facilitatory factual information to increase capability, organisers can provide participants with morally based impact-orientated information. Indeed, providing data on the environmental impact of aviation has been shown to increase behavioural intentions to reduce air travel26. Direct feedback on the GHG emissions consequences of different travel choices can act as an emotional and persuasive reinforcer of different actions. This strategy has been incorporated into policy at Griffith University, Australia, after a study showed that academics were supportive of such a strategy51. Such emotional information can motivate participants to engage with the factual information provided in the capability component of our behaviour change model, and from the element of opportunity to behave more sustainably.

Equally, explicit emission information on the accommodation and food choices provided by organisers can sensitise participants to more environmentally sustainable behaviours specific to the event and in general. This can be accompanied by a published environmental sustainability policy that accompanies application forms. Once again, event organisers can embed this information into the application process so that participants are able to compare and contrast the results of their individual travel, accommodation and food choices. Tools are available online, for example, https://travel-footprint-calculator.irap.omp.eu/home.html52; https://travelandclimate.org/53 that allow individuals to compare and contrast different choices and the impact these have on GHG emissions.

One significant barrier to individuals making more sustainable ground-based travel choices is cost. In our data, participants average flight costs were only €60 more to travel to Lisbon and back than Ghent, despite them travelling 2.5 times further to Lisbon than Ghent. We do not have comparable data on ground-based/rail journey costs, but budget airlines usually outcompete rail companies on price28 by between 10–30 times for equivalent distances. Even if academics are willing to spend longer travelling by ground-based modes, they are less willing to pay more to do so24,26, particularly if they are not recompensed accordingly by their employer, or indeed if employers insist that individuals take the cheaper option. Financial incentives apply across both capability and motivation behaviour change domains. Event organisers can use financial incentives as a powerful behaviour change strategy by making ground-based travel modes more financially attractive, and by ensuring that participants are not out of pocket. Reduced registration fees or direct subsidies for participants travelling more sustainably can act as a motivator, but can also strengthen capability to make such choices for early career academics on limited budgets. In order to estimate the actual organisational costs of doing so we will be using our data in a subsequent analysis, to be reported later, on modelling additional costs based on participant distances and time to these venues.

We provide a model in table form (Table 4) and as a diagram (Fig. 1) to demonstrate that a combination of opportunity, capability and motivational strategies can be put together to help scientific and academic organisations reduce their GHG emissions profile for face-to-face events. We also provide the EANS sustainability policy as a supplementary file (S1 EANS Sustainability Policy). Our paper here is unique as, unlike other papers that have collated and then listed potential ideas for change20,34, including in the ‘Cercedilla Manifesto’35, we have taken the next step by placing these strategies in a coherent and widely utilised behaviour change framework. This framework distinguishes between, and then combines, structural change from organisations (such as in the ‘Fly Less Toolbox’54) and individuals’ personal agency. We have categorised these levers and triggers using the intervention definitions from the COM-B behaviour change framework43. Interventions include environmental change, enablement, education, incentivisation and persuasion. The combination of strategies will be key. No one single action by organisers and/or participants will be sufficient in of itself.

Table 4 Behaviour change strategies to reduce GHG emissions from face-to-face academic events
Full size table
Fig. 1: Behaviour change strategies to reduce GHG emissions from face-to-face academic events.
figure 1

The figure follows the model proposed by Michie et al’s ‘Behaviour Change Wheel’. At the centre of the figure are the core behaviour change categories (capability, opportunity and motivation). Surrounding these core concepts is an inner ring of behavioural targets (travel, accommodation, and food) and details of specific behaviour change actions. In the outer ring are the strategy categories that relate to these targets, specific strategies and the core concepts to help change behaviour. Darker colour denotes interventions by organisers only; lighter colour for both organisers and event participants.

Full size image

Finally, we must acknowledge that interventions to reduce GHG emissions by scientific organisations and individual academics attending events organised by such societies exist in a context55 dominated by an academic culture of collaboration and internationalisation. Universities and research centres are central stakeholders13 whose actions will determine individual behaviour. Universally, promotion and remuneration criteria set by universities include international collaboration, particularly for those wanting to progress to the professoriate56. As noted earlier, academics hold entrenched beliefs that face-to-face networking24,26,27 is the best way to foster the international collaborations necessary for their careers and are, therefore, incentivised to attend and organise global meetings and conferences to progress in their careers. As institutions of education and research, universities are increasingly being called to act as ‘key catalysts’ in driving shifts toward environmental sustainability through advocacy and activism30. One of these catalytic actions should be to adopt human resource policies that incentivise, rather than penalise, academics who choose carbon-friendly strategies such as restricting their international travel or taking long and expensive ground-based journeys to international events.

Therefore, whilst the strategies we outline are not all-inclusive, our model does provide an exemplar for other organisations to add to and amend. We have not detailed sustainable changes already made by EANS—for example, by hybridising structural teaching resource allocations for the summer schools, or by reducing the number of faculty attending events in-person. The strength of our model is the combination of structural and individual factors, operating via organisations’ actions and individual actors’ agency. Our framework is specifically designed around ways in which organising societies can enable, and participants make, sustainable personal choices to help reduce GHG emissions from academic meetings. We now intend to test this framework during 2024 and 2025 by asking event organisers to use the framework when organising their events, and then collecting participant and organisation-level data on framework implementation, participant behaviours and emissions outcomes to determine how successful it is in driving down GHG emissions. We would also like to invite other organisations to test the framework by implementing it and then collecting event GHG emissions data for their own conferences, meetings and events.

One caveat is that our study is based on secondary data from an organisational dataset and does not include individual survey data on travel and diet. We assumed that participants would take intra-country journeys by train and cross-border journeys by air. These assumptions are defensible, given data that Europeans are more likely to travel by air than by train for distances of over 300 km, and those with higher levels of education are more likely to travel by air than other modes of transport57. Nonetheless, our travel mode assumptions may not have been valid for everyone—some participants may have travelled by train or bus from other countries, especially to Ghent. However, even if we have overestimated the flight GHG emissions for Ghent, our conservative analytical position is analogous to an ‘intention to treat analysis’58 in medicine, in that it will have reduced the actual difference between the events. The GHG emissions from Ghent may have been lower still. Taking this conservative analytical position strengthens our main finding, that it is the distance travelled by participants that drives conference and meeting GHG emissions. Again, in the absence of direct participant survey data, whilst we have had to make necessary assumptions, we have also directly compared two very disparate venues using this detailed data, rather than modelling ‘what ifs’, allowing us to directly quantify the contribution of two sites’ location on emissions for both travel and on-site contributions.

One limitation was our inability to include journey distances from participants’ home to the nearest airport or train station as we did not have data on participant home addresses. However, this is likely to be only a small additional contributor to the overall GHG emissions picture, and an effect that is equally spread across both events, so is unlikely to alter our main findings. Likewise, we did not subtract participants’ computer use at home from our event figures, as we had no means to estimate home use. However, computer use contributed less than 0.02% of the total emissions profile of the event and so this reduction would not have reduced the event emissions by a significant degree.

We did not include an entirely online event as another comparator. Our data shows that the GHG emissions from computing required for such events is negligible and others have noted that ‘Nearly Carbon Neutral’ events would self-evidently be by far the best solution to reduce GHG emissions from academic events21,22. However, as noted earlier, individual academic beliefs and organisational contexts are firmly entrenched in favour of face-to-face meetings to achieve the networking and educational objectives of scientific conferences and educational events24,27, and are, therefore, likely to continue. Consequently, we decided to design a behaviour change model to decarbonise conference and networking behaviours that academics will continue to engage in for the foreseeable future.

The final and most obvious limitation is that we have only analysed the GHG contribution of European events attended solely by European academics. Intercontinental travel would hugely increase GHG emissions. The impact of significant numbers of participants travelling from outside Europe would probably wipe out the differential between the two sites. Europe is also well provided for with ground-based transport, a situation not enjoyed by other parts of the world, again limiting travel choices for participants in other settings. Nonetheless, the elements that drive GHG emissions for academic events remain pertinent wherever they might be situated, particularly for intra-continental or intra-regional travel.

To conclude, academic societies organising conferences to deliver their social, cultural and educational goals can find themselves in conflict with sustainable environmental objectives, as their activities generate excess GHG emissions. Our analysis of two such events demonstrates that venue location is the main driver of such emissions. Venues closer to the centre of the potential participant catchment area will incur far fewer overall GHG emissions than those where participants must travel further. Venues that have a good non-aviation travel infrastructure and sustainable energy, accommodation and food policies will also perform better. Individual participants attending such events retain considerable personal agency, and so building their capabilities and enhancing motivation must go arm in arm with organisations setting the context for behaviour change. If such societies continue to organise face-to-face events, persuading academic participants to make less-polluting behavioural choices requires action by the societies to increase participants’ opportunity, improve their capability and enhance motivation, all within the academic cultural context set by the university sector. Societies must use a mix of strategies, including environmental change, enablement, education, incentivisation and persuasion to drive down the GHG emissions profile of their events. The ‘COM-B’ behavioural change framework gives such societies a formal structure on which to base their actions and reduce their GHG emissions. We present this model and an example strategy document to the academic community as a service to help others reduce the GHG emissions caused by academic societies delivering their social and educational objectives through conferences and meetings.

Methods

We undertook a retrospective analysis of the GHG emissions from two international academic events for European PhD students run by the European Academy of Nursing Science (EANS), an exemplar of an academic society. EANS is a network of individual nurse scientists from universities across Europe (https://www.eansnursing.eu) that provides a forum for established and developing nurse researchers to meet, network and develop a European perspective to their work. We included all major sources of event-related GHG emissions, including participant travel, accommodation and food consumption. We calculated emissions from estimates of distances travelled, accommodation profiles and diet. We also calculated estimated financial costs associated with attendance. We then related these data to the COM-B behaviour change categories to determine strategy types that could be used to reduce emissions from such events.

Participants and setting

Participants were faculty and PhD students attending two academic events – summer schools for European PhD students organised by EANS—in 2018 in Ghent, Belgium and in 2019 in Lisbon, Portugal. These events were chosen as standard exemplars for such events and were the final two events organised by EANS for which we had data before the COVID-19 pandemic led to significant changes in organisational behaviour, albeit these changes were not determined for environmentally sustainable reasons. The summer schools are multinational learning environments designed to provide doctoral students with a common European perspective for their research, to learn advanced research methods and to enhance the opportunities for them to collaborate with other European researchers. Over 100 doctoral students and around 30 faculty attend each annual summer school, which lasts 2 weeks, with students attending between five and 12 days each.

Travel

Using our data on participants’ home city, we estimated travel distances in kilometres (kms). We applied a general assumption that participants travelling from countries different from the summer school venue would fly, and those travelling from within the same countries would take the train. Both Portugal and Belgium are small countries where it would take people much longer to fly from their home city in the same country to the event venues than take a train. For the Lisbon event, this assumption is likely to be entirely reasonable, since it is simply not possible to arrive in Lisbon from outside Portugal by direct train. For Ghent, we applied the same assumption apart from a small number of instances, for example, where the home base of participants (e.g. Maastricht in The Netherlands) was so close to the venue that no one would expect such a participant to choose air travel over rail. We estimated individual participant’s return flight journey distances between their nearest home country international airport and airports nearest to each summer school venue, using A to B figures given by the Great Circle Map https://www.greatcirclemap.com/. Again, we only departed from this principle in a very few instances, for example, participants from Lund and Malmo in southern Sweden would cross an international border to take a flight from Copenhagen not a Swedish airport, due to Copenhagen airport’s close proximity to that area of Sweden. We used www.skyscanner.net to identify available direct economy class return flights between the two airports identified above. Where a direct flight was not available, we identified flights with connections during 1 day of travel. For participants travelling within countries, we calculated those individuals’ return rail travel in km from their home base to the venue using Google http://maps.google.com road map distances as a proxy for rail distances, since actual rail distances are not readily available. For participants taking flights, we also calculated individual’s return rail travel in km from their home base and their departure airport, and their return rail journeys from the airport nearest the destination to the summer school venue using the same method. Where there were no trains available for these journeys, we calculated km using coach travel.

GHG emissions: travel

We calculated the GHG emissions of all travel in kg CO2e using the UK Government GHG Conversion Factors for Company Reporting 202259 https://www.gov.uk/government/publications/greenhouse-gas-reporting-conversion-factors-2022, multiplying the published conversion factors in this tool by the number of km for each individual’s journey. For flights, we used the per person per km short-haul economy conversion figures (0.15102 kg CO2e per km); for rail, the national rail conversions (0.03549 kg CO2e per km) and for coach 0.02733 kg CO2e per km.

GHG emissions: venues

We used published data on the environmental impact of services provided by the summer school venue, including hotel rooms at a bed and breakfast rate, lunchtime meals and several social events multiplied by the number of days of participant attendance. We excluded evening meal consumption because these were not provided by the site, would have been incurred by participants at home and would have varied greatly according to individual participant preferences and budgets. For hotel and meeting rooms, we used a web-based tool that uses data from the Cornell Hotel Sustainability Benchmarking index (https://www.hotelfootprints.org). For all meals, we calculated the GHG emissions based on the region of Europe where the summer school was held60. We used data on population dietary preference averages60,61 and assumed 10% of the sample would choose a vegetarian option with an average consumption of 700 kcal per person for lunches and socials, and 1200 kcal per person for the gala dinner. Finally, we calculated individual computing GHG emissions for students and faculty at 8 h per day at 0.48 kg CO2e per day (assuming 175 kg CO2e per year) plus wireless internet provision at 0.10 kg CO2e, multiplied by the number of event days attended per individual62.

Financial costs

As historical cost data were not available for travel, for all travel modes, we used 2023 costs in Euros for a predicted event in 2023 on dates similar to the events in question. Using the flight data identified for the calculation of km above, we used www.skyscanner.net to select economy class return fares for the flights or flights selected. For rail and coach travel, we consulted the national rail companies’ web pages for prices of current second-class return fares. For hotel costs, we were able to use the actual venue costs incurred. We multiplied the hotel costs by the number of days each student and faculty staff member spent at the summer school.

Analysis

For travel, we estimated the total event and mean per person km travelled, kg CO2e, and financial costs with standard deviations (SD) from the individual per person data. We summed the travel distances, GHG emissions and financial costs for each individual participant to calculate the total travel distances, kg CO2e and financial data by mode of travel per event. We included on-site kg CO2e data for each participant by number of days of attendance per person and similarly amalgamated these data. Finally, we combined all kg CO2e data to produce an overall estimate of the GHGs of each event and the contribution of each aspect of the event (travel, travel mode, on-site impact) to the overall emissions and financial figures.