This work is licensed under the Creative Commons Attribution 4.0 International License (CC-BY 4.0). To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.

Please cite this document as:

SustainableHECAP+ Initiative, “Environmental sustainability in basic research: A perspective from HECAP+”, 2023, available at: https://sustainable-hecap-plus.github.io/.

Individual endorsements of this document can be made at https://sustainable-hecap-plus.github.io/. To make an institutional endorsement, please email sustainable-hecap-plus@proton.me.

The acronym HECAP was adopted in the early stages of this initiative, otherwise in common use to refer to High Energy Physics, Cosmology and Astroparticle Physics¹. With the subsequent inclusion of key contributions from the Hadron and Nuclear Physics community, with whom HECAP share common research infrastructure and common challenges in the pursuit of improved environmental sustainability, this acronym was modified to HECAP+. This modification is also intended to emphasise that many of the issues highlighted in this document apply broadly to all members of the basic research community.

This document has been converted from LaTeX using Pandoc. The font used is Atkinson Hyperlegible to maximise readability (see https://tug.org/FontCatalogue/atkinsonhyperlegible/).

The original PDF version of this document is available at: https://sustainable-hecap-plus.github.io/.

Changes in version 2.0: Clarifications have been added in relation to the authorship of this document on the title page and in the Statement of Intent; the inset on personal emissions and the figure reproduced in full from Ref. [10] have been removed in favour of a text-based discussion of the elements of the study [10] that give relevant context to research-related emissions; the text beneath the right-hand chart of Figure 1.3 has been corrected to indicate per-researcher rather than per-capita numbers, and clarifications have been added to the caption to Figure 1.4 and the accompanying text in relation to “per-researcher” emissions; corrections and clarifications have been made in the Computing section in relation to the High-Luminosity phase of the Large Hadron Collider; a correction was made to the left-most bar of Figure 4.1; and the paragraphs on low-carbon nuclear energy and the discussions on energy import and export via HVDC cables in Section 3.1.2 and Case Study 3.3 of the Energy section have been revised. Other minor typographical and grammatical changes have been made.

Statement of Intent

This reflective document was developed by a small group of concerned members of the HECAP+ communities (see ), beginning as a grassroots initiative Striving towards Environmental Sustainability in High Energy Physics, Cosmology and Astroparticle Physics.

Its focus is not to stipulate the research that our communities should undertake, nor to debate its intrinsic value. Rather, it is intended to be a synthesis of current data, best practices, and research in climate science and sustainability, as applied to our fields to the best of our ability as physicists, and a reflection on the roles that our communities can play in limiting negative environmental impacts due to our research work and scientific culture.

The scope of the document is inspired by the holistic approach of annual environmental reports of major institutes [1], [2], which include emissions directly related to research and collateral emissions, such as from personal commutes and institutional catering. Any imbalance in its content, in part, reflects imbalances in the availability of reliable data and resources relating to the environmental impact of aspects of our communities’ activities. Redressing this imbalance will require input from across our communities, in particular to identify the technical challenges of limiting the environmental impacts of our current and future research infrastructure.

While this document is primarily framed from the perspectives of high energy physics, cosmology, astroparticle physics, and hadron and nuclear physics (HECAP+), much of its discussion applies to basic research more generally. Its broad scope is intended to provide a first step toward greater coordination across the community in efforts to address environmental sustainability, and it is hoped that it may serve as a useful reference for our and other fields.

Comments on this document are welcome. Please get in touch with us via the online platform at: https://sustainable-hecap-plus.github.io/, where you will also find the latest version of the document. Individual endorsement of this document can be made at: https://sustainable-hecap-plus.github.io/. For institutional endorsements, please email us directly at sustainable-hecap-plus@proton.me.

Thank you for taking the time to read this document.

Forewords

What should we do to limit the negative environmental impacts arising from our scientific community?

Productive discussions on this crucial issue are often hindered by a number of roadblocks. Climate change scepticism is fortunately not very frequent in our community, but the same cannot be said, for instance, of climate change whataboutism — “What about other communities who produce way more greenhouse gas emissions than us? Why should we make efforts, when others are not doing enough?”.

But even for those of us who recognize the importance of doing something, a general lack of data and detailed information often prevents us from aligning our actions with our ethical values — “Does my Carbon footprint increase more by flying to conferences a couple of times a year or by eating red meat a couple of times a week? By computing or by commuting?”.

This timely and thought-provoking paper, arising from the grassroots initiative Striving towards Environmental Sustainability in High Energy Physics, Cosmology and Astroparticle Physics, provides a much welcome reflection on how to remove these and other common roadblocks towards sustainability, and how to put in place concrete actions to limit our environmental impact.

The reader will find here a synthesis of current data, best practices, and research in climate science and sustainability, and a set of concrete recommendations that clearly show how we can empower individuals and the broader community to take direct action and responsibility for mitigating climate change.

I encourage every member of our community to reflect upon the contents of this document and to actively engage in the ongoing dialogue surrounding this crucial topic. I hope it will spark broader change and promote a culture of sustainability in our community and beyond.

The conclusions from the IPCC are clear: global heating and climate change are an existential threat to human civilization. All aspects of society need to follow the recommended guidelines and eliminate greenhouse gas emissions as quickly as possible. The research fields of particle physics, cosmology and astroparticle physics have a role to play in this great transition. Not only are the emissions associated with those fields relatively large compared to other areas of daily life, but as researchers and scientists we understand the science and we can use our creativity and ingenuity in finding solutions. By placing sustainability at the forefront of our scientific approaches, we will provide guidance to other research fields and convey to society the importance of this topic. It is our duty to act now such that future generations can also enjoy the wonders of exploring the secrets of the universe. We used to pride ourselves with being great innovators at the cutting edge of technology. Let’s be ambitious again and help tackle all of the challenges associated with a net zero economy. In the process we will improve our health and make our research fields more diverse and accessible to all regions of the Earth. As scientists ourselves we can not ignore the science, there is a climate emergency and everyone needs to play their part.

The climate crisis is one of the most pressing problems facing humanity today: the long-term survival of our species, and countless others, will be impacted critically by when and how we choose to address it. This is not an issue we can ignore or put off; action needs to be taken in the very near-term — the next few decades — if we are to avoid a critical rise in mean global temperature that is predicted to result in major changes to climate and weather patterns, the first signs of which we are already experiencing. Of course tackling such a monumental issue requires global collaboration and coordinated action by world governments. But as individuals, as well as via our memberships of teams, groups, institutes, laboratories and research organisations, each of us can, and must, take our individual responsibility for helping to ensure the vitality of our planetary environment and sustainability of its resources.

For those of us working in the scientific fields represented, this report provides a clarion call for action on sustainability. The report both draws attention to the environmental impacts of pursuing our scientific endeavours in a ‘business-as-usual’ fashion and, moreover, highlights actions that we can take to reduce significantly these impacts. In many cases these involve straightforward changes to behaviour and practice that will yield benefits almost immediately. In other cases more effort will be required and the benefits may be realised later. The point is: we owe it to ourselves to take action, and we can start now. This report is timely and it contains many constructive recommendations; the findings are widely applicable within the broader scientific community, and well beyond.

In the context of the climate crisis, the scientific community is part of the solution — through research, teaching and evidence-based policy advice. However — and this aspect has traditionally been much less in focus — the scientific community is also part of the problem, through the emissions that it produces through its own operations. This report illustrates starkly, how large that problem is: for some researchers, emissions caused by their work are not just somewhat higher than the current average global per-capita emissions or the per-capita budget to 2050, they are roughly an order of magnitude higher (see Fig. 1.4, p. 17). In the face of such data, inaction can cost the HECAP+ community i) the credibility and trust it enjoys both from policy-makers and society at large, ii) the enthusiasm for HECAP+ topics that turns young students (many of whom care passionately about solving the climate crises) into next generation’s researchers and iii) even its research freedom, because it is likely that societal pressure will increase as the climate crisis progresses and policy makers will ultimately step in to regulate carbon-intensive sectors much more strictly.

Yet, there is ground for hope. As the report illustrates, there are already best practise examples that others in the HECAP+ community can learn from. Beyond the existing examples, the report provides actionable recommendations at three critical levels: the level of the individual, the level of research groups and the level of institutions. These three levels are necessary to achieve a change in culture towards a climate sustainable research community. Such a change in culture will happen, if individual behaviours and framework conditions (often set at the institutional level) that determine norms and incentives, change. The recommendations in this report provide first steps that the HECAP+ community can take to achieve this cultural change. Everyone in the community should read about, reflect on and, wherever possible, implement such steps.

All of us have roles to play in combating the climate crisis that is bearing down upon us and the generations to come. Some of these roles are individual, such as our personal lifestyle choices, whereas others are collective, linked to our activities in society. As scientists studying fundamental physical laws, astroparticle physics and cosmology, our activities burden us with particular responsibilities. For example, the scales of our research facilities imply that we consume orders of magnitude more resources than is sustainable for the bulk of humanity, so we must strive to research as efficiently as possible. This implies minimising the wall-plug power and other resources consumed by our accelerators, experiments and data analyses. Moreover, the international nature of our research teams implies that we travel more than most, so we should strive to travel as sustainably as possible, e.g., by train, and as little as possible, e.g., by using teleconferencing tools such as Zoom or Teams. On the other hand, our research can provide humanity with valuable tools for living more sustainably. For example, the World-Wide Web, which CERN placed in the public domain 30 years ago, has enabled information to be shared and the planet’s business to be conducted more efficiently and therefore sustainably. Moreover, instruments that we develop can provide tools for monitoring and possibly mitigating the effects of climate change. Finally, through our international contacts we can both learn and disseminate best practice. Individually and collectively we have unique responsibilities and opportunities to follow the right paths: let us take them.

Climate change is the biggest challenge that humanity is facing today, and science is called on to help society in identifying measures to avert the environmental catastrophe that looms over our future. While science can provide solutions, we must recognise that scientific research is also part of the problem: our projects, infrastructures, computing facilities and work habits are energy-intensive and waste producers. More than ever, it is mandatory for scientists today to assess carefully the environmental impact of their activities and prove to society that even the most advanced scientific projects can be carried out in environmentally sustainable ways. This document lucidly presents the challenges, substantiating them with facts and data, and paves the way towards realistic and effective solutions. It testifies to the unrelenting commitment to societal and environmental problems deeply felt within the scientific community.

In the past century the ever increasing resource demands of humans have a devastating impact on the climate of our planet. The resulting heat waves, droughts, strong rain falls, violent storms, melting ice and rising sea levels are posing an existential threat to many people world-wide already today, and even more in the future. This situation demands action from all of us individually, and as groups and institutions, to do whatever we can to reduce (or ideally eliminate) the emission of \(\mathrm{CO}_2\)(equivalent) gases, and more generally to preserve the resources of the planet. This report illustrates well that scientists working in the fields of high energy physics, cosmology, astroparticle physics, hadron and nuclear physics are using on average more resources than what will be acceptable, and action is needed immediately. Many recommendations are proposed for individuals, groups and/or institutions. Some are rather easy while others are challenging and require core habits to change. Given the very international nature of the research performed in these fields, there is a large potential to propagate the actions to 100s of institutions in countries of all continents and thereby increasing the impact further. I very much hope this document will help us embark on the right path.

Many of us who study astrophysics and cosmology do so out of a sense of awe and wonder at the universe. That same sense of wonder should compel us to consider the impact of our work on our planet. We work in areas that are driven by data, and the data on climate change shows clearly how humans continue to impact the climate. This extensive report presents a summary of the ways in which our work contributes to increases in emissions of greenhouse gases, collectively through our international projects and individually through personal choices. It is a sobering look at our impact, and provides recommendations for how we can effect change collectively within our research communities, collaborations and institutions.

International collaboration drives progress in the large, complex projects that we undertake to unravel the secrets of the cosmos. These projects often involve significant infrastructure investment and require a large computing budget. On the largest scale, the report challenges us to examine how we can reduce the environmental impact of the projects as a whole. This will require from us a renewed prioritisation of energy-efficiency, and strategic thinking to balance our research needs with these time-critical actions. At an individual level, it suggests ways to reduce impact on the climate by considering how we can reduce our international travel, while considering solutions that are inclusive of all members in our collaborations, regardless of geographic location. This will again require creativity — but I am confident that as researchers who have been trained in solving difficult problems we can rise to the challenge.

The challenges posed by the human-made climate crisis, the largest, most rapidly developing and pervasive threat to the natural environment and societies the world has yet experienced, are enormous. They are created by ever-growing human consumption, bolstered by rising energy production and its inefficient use and waste. This energy is predominantly generated by the massive combustion of fossil fuels, predictably leading to the greenhouse effect we are experiencing. While fossil fuels are the dominant source of greenhouse gases, direct release of these gases also contributes to global climate change. As scientists we have the responsibility to transparently expose the extent to which we contribute to the problem, as discussed in this document, and help develop solutions. Among the biggest challenges for current experiments is the emission of greenhouse gases by detectors developed and built decades ago, in some cases using gases that were at that time thought to be environmentally-friendly alternatives to even more harmful gases. Energy efficiency throughout all our instruments, including considerations of their embodied carbon, is another important challenge. All these challenges require and receive our immediate attention. Future high-energy colliders and experiments will need to be ’green’: they will have to rely on decarbonated energy sources and employ detector technologies that avoid the use of greenhouse gases. I am confident that our field will meet these challenges and contribute to building a sustainable future.

Climate change does not necessarily threaten the survival of our planet in the bigger picture of the solar system and the universe; rather, it threatens our own survival and the survival of the bio- and eco-spheres that we rely on for our subsistence in this cosmos. We have no alternative planet which will provide for us, so we better not make our current one uninhabitable. Climate research has been clear on the effects of climate change, and slowly the rising sea levels, the disappearance of glaciers, the frequency of hundred year floods, the storms, droughts, and heat waves beat the message home: climate change is happening, human action is the cause, and we better counter-act it immediately. As fellow scientists from the areas of high energy physics, cosmology, and astroparticle physics, we are trained to understand and consider scientific results in our daily work. We know how to interpret statistics and draw conclusions from data. This document is a start to take the conclusions from climate research seriously and put them into action in our own fields of research. Collecting and reflecting on available results, together with recommendations for the implementation — from easy to hard — is an important first step, but only a first step. Let’s use this document and get started in transforming our field of research into a sustainable field of research — for the benefit of our planet and our own futures. The data has been clear for a while, now is the time to act on it.

The Intergovernmental Panel on Climate Change (IPCC) Working Group II in their Sixth Assessment Report (2022) underscored that “the science is clear. Any further delay in concerted global action [on climate change] will miss a brief and rapidly closing window to secure a liveable future”. We are also rapidly advancing towards the 2030 deadline to achieve the 17 Sustainable Development Goals: an international call to end poverty, protect the Earth and deliver peace and prosperity for all.

This report, which calls for climate and sustainability actions within the international scientific community, is both timely and welcome. It provides clear, actionable recommendations that can contribute to the collective effort to deliver positive changes across six key areas: i) Computing, ii) Energy, iii) Food, iv) Mobility, v) Research Infrastructure and Technology and vi) Resources and Waste. Ensuring both our science and the way we live are as sustainable as possible is an incredibly important undertaking, and scientists need to lead the way, and “walk the [sustainability] talk”. The global reach of this report offers a substantial opportunity to reorient the scientific community along a more sustainable trajectory. I encourage all who read it to commit to delivering its aspirations and best practices.

Executive Summary

Humanity’s impact upon the world’s climate and ecosystems is now as unequivocal as it is extreme [3]. Averting this climate catastrophe must be a critical concern for all global citizens at this pivotal time in world history.

High Energy Physics, Cosmology, Astroparticle Physics, and Hadron and Nuclear Physics (hecap) research has direct impacts on the environment. Our research infrastructure, including accelerators, detectors, telescopes and computing resources, requires enormous power generation and, in many cases, contributes directly to greenhouse gas (ghg) emissions. Our work practices give rise to additional emissions, e.g., from procurement, business travel and commuting, and our industry generates various forms of waste that are harmful to the environment.

As scientists working in HECAP+ and related disciplines, our responsibilities to limit and mitigate our impact on the world’s climate and ecosystems are manifold. Our opportunities and training have given us the science capital to appreciate the evidence that has been collated over many years by climate and environmental science. We must use our unique and privileged platform to impel positive changes in, as well as educate and advocate on, environmental sustainability and the connected issues of social justice. Moreover, as a community focused on basic scientific research, we should be no less accountable for our impacts on the world’s climate and ecosystems than any other industry, and we should anticipate that our activities will come under increasing scrutiny from the public, governments and funders. We have moral and pragmatic obligations to act.

This document follows the holistic approach taken by several HECAP+ institutions in their annual environmental reports (see, e.g., Refs. [1], [2]) in assessing the environmental impacts of HECAP+ research across six areas: computing, energy, food, mobility, research infrastructure and technology, and resources and waste, also within the larger context of global emissions. Specific recommendations are made for each of these areas, but the overarching message is simple:

Assessing, reporting on, defining targets for, and undertaking coordinated efforts to limit our negative impacts on the world’s climate and ecosystems must become an integral part of how we plan and undertake all aspects of our research.

This requires urgent action at an individual level, at a group level (including research groups, collaborations and organising committees), and at an institutional level (including universities, research institutes, funding agencies, and professional societies). Moreover, it requires systematic positive changes in everything from our day-to-day activities and the ways we interact as a global community through to the design and running of the ‘big science’ infrastructure on which HECAP+ research depends.

It should be emphasised that the reduction of GHG emissions or other environmental impacts from any source identified in this report should be considered a priority by the community, whatever the comparative scale of these impacts. Carbon offsetting via legitimate providers (see Ref. [4]) should be seen as a last resort, used only once all other options for reducing the \(\mathrm{CO}_2\) equivalent (co2e) emissions have been exhausted and to offset any residual CO\(_2\)e output.

We urge all members of the HECAP+ and related communities to take individual actions and push for group- and institution-level changes that:

• Establish community-wide formal and coordinated efforts to assess and improve the environmental sustainability of basic research, which calls for standardised reporting and data sharing.

• Consider the environmental cost of computational infrastructure and algorithms in decision making and prioritise the development of common and reusable software solutions across HECAP+.

• Prioritise the use of sustainable and renewable energy to power our workspaces and research infrastructure; increase their energy efficiency and recovery, and energy storage capacity.

• Move towards plant-based catering at conferences and in cafeterias, immediately reducing the provision of carbon-intensive foods, such as ruminant meats and dairy products.

• Prioritise environmentally sustainable modes of transport for commuting where possible.

• Prioritise responsible business travel that balances in-person and online meetings, acknowledges the benefits of virtual and hybrid meetings for inclusivity, and considers the disproportionate impact of changes to travel culture on different groups, e.g., early career researchers and those who are geographically isolated.

• Mandate comprehensive life-cycle analysis for all proposed research infrastructure projects that critically assesses the environmental impact of all project stages, including design and approval, construction, commissioning, operation, maintenance, decommissioning, and removal.

• Prioritise environmentally- and socially-sustainable sourcing of raw materials for experiments and infrastructure.

• Propagate and expand the culture of “reduce, reuse, repair, recycle”, including the implementation of life-cycle awareness and end-of-life planning for hardware.

• Educate and advocate on issues of environmental sustainability and social justice, and engage more broadly with policy makers to push for wider change, e.g., the improvement and decarbonization of local transport infrastructure.

Outline

The aims of this document are:

• To improve awareness of the impact that high energy physics, cosmology and astroparticle physics, and hadron and nuclear physics (hecap) has on the environment.

• To provide suggestions and encourage immediate action on ways that we, as a community, can play our part in limiting further degradation of the world’s climate and ecosystems.

• To provide impetus for ongoing and collective discussions of how we can make positive changes to our community’s work practices, in terms of environmental sustainability and for the issues of social justice from which climate change and environmental degradation cannot be disentangled.

The aims are not to stipulate the research that our communities should undertake, nor to debate its intrinsic value.

The discussions are divided into seven sections. Sections 2 through 7 cover the topics of Computing, Energy, Food, Mobility, Research Infrastructure and Technology, and Resources and Waste. Each of these sections contains a set of recommendations, for individuals, groups and institutions, and these are followed by longer discussions that include case studies and best practice examples, which can be read independently of the surrounding material. Collated lists of acronyms and abbreviations, best practices, case studies, and figures and tables are included at the end of this document.

Section 1, Preliminaries, begins by acknowledging the climate crisis and the environmental impacts of HECAP+ research. It provides a summary of the United Nations (UN) Sustainable Development Goals (sdgs) and how these relate to HECAP+ research, and briefly reviews similar and complementary documents.

1 Preliminaries

1.1 Introduction

The 2021 report of the Intergovernmental Panel on Climate Change (ipcc) [5] is emphatic in its statements about the current status of the climate and the damaging impact that humanity continues to have upon it [3]:

"It is unequivocal that human influence has warmed the atmosphere, ocean and land. Widespread and rapid changes in the atmosphere, ocean, cryosphere and biosphere have occurred. […] Human-induced climate change is already affecting many weather and climate extremes in every region across the globe. Evidence of observed changes in extremes such as heatwaves, heavy precipitation, droughts, and tropical cyclones, and, in particular, their attribution to human influence, has strengthened since [the Fifth Assessment Report in 2014]."

It is also clear on the consequences of further inaction [3]:

"Global surface temperature will continue to increase until at least mid-century under all emissions scenarios considered. Global warming of 1.5C and 2C will be exceeded during the 21\(^\text{st}\) century unless deep reductions in \(\mathrm{CO_2}\) and other greenhouse gas emissions occur in the coming decades. […] Many changes in the climate system become larger in direct relation to increasing global warming. They include increases in the frequency and intensity of hot extremes, marine heatwaves, heavy precipitation, and, in some regions, agricultural and ecological droughts; an increase in the proportion of intense tropical cyclones; and reductions in Arctic sea ice, snow cover and permafrost."

Net global co2e emissions must be halved before 2030 to fulfill the Paris Climate Agreement. Without this, we are unlikely to meet the target of limiting global warming to 1.5C in order to avoid fatal tipping points in the global biosphere (see Figure 1) [6]. Pledged policy changes by nations party to the Paris Agreement, known as Nationally Determined Contributions, are insufficiently far-reaching, and “make it likely that warming will exceed 1.5 during the 21st century” [7] (original emphasis). Demand-side mitigation, including changes in infrastructure use and social and behavioural practices, can reduce global ghg emissions in end-use sectors by 40–70% by 2050 [8].

Oxfam’s recent publication "Confronting carbon inequality" [9] notes a strong correlation between GHG emissions and income level, with the world’s wealthiest 10% accounting for over half of cumulative global emissions (see Figure 2 for infographic). This income bracket, corresponding to an average annual income of over €34,000, was identified by the IPCC as having “the greatest potential for emissions reductions, e.g., as citizens, investors, consumers, role models, and professionals” [8], and includes many HECAP+ physicists (notwithstanding significant disparities in income within our communities depending on location and career stage). A recent meta-study from Lund University [10] concluded that the most impactful individual climate actions include: living car free (country-averaged range of 1.4 – 3 tco2e reduction per year); avoiding one transatlantic flight (1.6 \(\mathrm{tCO}_2\mathrm{e}\)); purchasing green energy (country-averaged range of < 0.1 – 2.5 \(\mathrm{tCO}_2\mathrm{e}\) reduction per year); and eating a plant based diet (country-averaged range of 0.4 – 1 \(\mathrm{tCO}_2\mathrm{e}\) reduction per year). Note that the numbers quoted were based on “average conditions in developed countries”, so miss the substantial differences between emissions levels in developing and developed nations. They also neglect the effects of future climate policy. Even so, they give indicative scales that place HECAP+ researchers’ work-related emissions in context, cf. Figure 4. The study also highlights the disconnect between the low- and moderate-impact measures that consumers are commonly encouraged to take, and the high-impact measures that require more significant lifestyle changes and pertain to more complex and nuanced issues.

Figure 3 presents a breakdown of 2016 global GHG emissions by sector, with the emissions at the European Organization for Nuclear Research (cern) for 2019, during the Large Hadron Collider (lhc) shutdown, shown as a proxy for research emissions.³ CERN, like other HECAP+ institutions, categorises its emissions by scope rather than sector, making a direct comparison difficult. Instead, we consider total per-researcher emissions. A similar per-employee metric is used by the French National Centre for Scientific Research (CNRS), to quantify its climate impact [11]; this is also a default output of the Labos1point5 research emissions assessment tool [12] (see also Best Practice [BP:l1p5]). Dividing CERN’s emissions equally among its seventeen thousand Users (researchers involved in CERN-based experiments), we obtain roughly 15 \(\mathrm{tCO}_2\mathrm{e}\) per researcher per annum, or twice the global average of 6.3 \(\mathrm{tCO}_2\mathrm{e}\). Note that this does not include personal household emissions for researchers, which may exceed their workplace emissions. It also incorrectly apportions emissions due to CERN’s smaller cohort of direct personnel, e.g., due to CERN-funded travel or personal computing equipment, to its entire User base.

For a fairer accounting, see Figure 4. The latter chart shows reported work-related emissions in each sector divided by the ‘true’ consumers of each resource, for researchers at CERN, as well as four other HECAP+ institutions: the Max Planck Institute for Astronomy (mpia) in Heidelberg, Germany, the Department of Physics (DPhys) at eth, nikhef in the Netherlands, and Fermilab in Chicago, USA (fnal). For raw data and details of underlying methodology, see Appendix 9. Note that these institutions are somewhat self-selected, counting among the minority that have published quantitative estimates of their environmental footprint. Since CERN is the only institution on this list that attempts a full accounting of its Scope 3 emissions, the numbers in this chart should be taken as indicative only, and caution should be employed in making comparisons across institutions on this basis.⁴ It is nevertheless evident that work-related emissions for many of HECAP+ researchers far exceed our remaining budget to stay within 1.5C of warming [3].⁵

These work-related emissions are due to choices we make, as individuals, collaborations, or institutions. They could be a direct consequence of HECAP+ research, such as the choice of detector design; computing setup or software pipeline for simulation and analysis; or how we collaborate or communicate the results of our work. Alternatively they could be peripherally related to the science we do: e.g., how we commute between our home and workplace, or the food we consume while at work; or how our offices are powered, heated and ventilated. Historically, many of these choices have been made prioritising cost or convenience over environmental and social impact. However, the rapid and systemic societal change needed to keep to our climate change goals requires system-wide engagement at all levels of academia. We can impel positive change throughout the academic research system by re-assessing these choices and how central they are to our primary function as scientists.

This process has already begun. Universities and other institutions are including sustainability in buildings planning (see Best Practice [BP:Nikhef]) and engaging with voluntary assessments of the environmental sustainability of their research facilities (see e.g., Refs. [1], [2], [19]–[21]). Examples focussed on the environmental sustainability of laboratories include France’s Labos1point5 (see Best Practice [BP:l1p5]) and University College London’s Laboratory Efficiency Assessment Framework (LEAF) initiative [23], [24], a standard being adopted across an increasing number of universities [25]. The LEAF initiative, which awards three levels of certification to participating laboratories, is structured around online tools that promote best practice, and aid calculations of impact and reporting, as well as additional resources and training opportunities for staff and students.

1.2 Previous and Parallel Initiatives

This document is focused on environmental sustainability and associated social justice issues of particular relevance to the activities of HECAP+. It is important to acknowledge the attention that these topics are rightly being given across our communities. This includes, e.g., in conference plenary talks and in parallel tracks devoted to sustainability, and equity, equality, diversity, inclusivity and accessibility. This section provides a brief review of other documents with similar and complementary focuses on environmental and wider social responsibilities.

1.2.1 ALLEA, Towards Climate Sustainability of the Academic System in Europe and Beyond

The All European Academies (ALLEA) Working Group on Climate Sustainability in the Academic System published a report in May 2022 [26], the aim of which is "to assess current practices and to critically examine current and proposed measures." The document urges stakeholders — either individual (researchers and students) or structural (universities, funding bodies, conference organisers, ranking agencies, and policy makers) — to know their roles and responsibilities toward a climate-sustainable academic system. After summarizing available data on GHG emissions from various stakeholders and reviewing the current practices aimed at reducing those emissions, the report outlines recommendations for individual and group stakeholders. Dimensions of social justice and equity are among the principles underlying all recommendations, as well as the opportunity for the academic system to be a role model in the matter. While all group stakeholders are advised to embed sustainability in their strategies, individual ones differ: students and academic members are encouraged to hold university management accountable, to demand divestment and to generate awareness. The importance of the development of an evidence base is emphasised, along with mix-and-match approaches to meeting formats. Finally, stakeholders are pushed to allocate funding to the decarbonization of the academic system.

1.2.2 Snowmass Contribution, Climate Impacts of Particle Physics

The report "Climate impacts of particle physics" [27], submitted to the proceedings of the US Community Study on the Future of Particle Physics (Snowmass 2021) focuses on facility construction, detector gases, computing, and GHG emissions from particle physics laboratories. The report highlights two key motivations for addressing the ecological and climate impacts of particle physics: (i) that the particle physics community has a moral obligation to do so and (ii) that its professional activities will be under increasing scrutiny from a number of stakeholders. The latter means that the community will be under increasing pressure to justify its carbon emissions against its relative size, compared to other industries, and its societal benefits.

As a concrete example, the report focuses on the Future Circular Collider (FCC) — the proposed 100 TeV electron-positron (and later hadron) collider. The authors estimate that the construction of the roughly 100 km circumference tunnel alone would lead to \(\mathrm{CO}_2\) emissions at the level of a few hundred kilotons, several times more than other large US building projects. This corresponds to "per physicist" emissions 80 times larger than their estimated 1.1 tco2 per capita per year limit needed to keep global warming to less that 1.5C. The authors emphasise the significant impact of the GHGs used in particle physics detectors and for cooling, which can have warming potential that exceeds that of \(\mathrm{CO}_2\) by as much as four orders of magnitude (in the case of \({\rm SF}_6\)). The report then highlights a number of avenues for reducing the GHG emissions due to the electricity consumption of computing, which is pivotal to running and exploiting such facilities. It also discusses the additional emissions due to collaborative research activities, with a particular emphasis on taking careful steps to reduce air travel that capitalise on potential benefits for social justice while minimising unintended negative consequences for members of the community.

The report’s recommendations stress the need for reporting on planned emissions and energy usage for new facilities; standardised reporting of emissions across the sector, and community-wide engagement to tackle the negative climate impacts of particle physics research through dedicated research time.

1.2.3 Recommendations by the yHEP association in Germany

The young High Energy Physicists (yHEP) association in Germany published the "yHEP recommendations on improvement of environmental sustainability in science" [28] and its Addendum [29] in December 2020 and 2021, respectively. The documents, which were the result of proposals from the yHEP community, including HECAP, hadron and nuclear, and accelerator physicists, contain ideas for improving the environmental sustainability of basic research. They take a qualitative approach on a broad range of topics, including, but not limited to, travel, conferences, computing and infrastructure, resource management and financing, and green energy.

1.3 Impelling Positive Change

The aim of this document is to provide as comprehensive a discussion as practicable of the various impacts of HECAP+ research, from our day-to-day activities through to the large infrastructure projects on which our science depends. The discussions presented here have much in common with those of the documents described in Section 1.2. This document is, however, intended to have broad scope and, through case studies and best practice, to illustrate potential actions that can be implemented at individual, group and institutional levels to limit the impacts of HECAP+ research on the world’s climate and ecosystems.

However, if the HECAP+ community is to succeed in improving the sustainability of its working practices, then the environment and related issues of social justice must be recognised as integral parts of the planning and management of our research activities. With this in mind, we collect below a list of recommendations for structural changes to the organisation of our community, our training and our professional development.

These recommendations complement those listed in the discussions of specific sources of environmental impacts of HECAP+ research on which the bulk of this document focuses. Together, these provide concrete suggestions of ways in which the HECAP+ community can act to reduce its negative climate and ecological impacts, and address issues of social justice in line with the United Nations Sustainability Goals, discussed in the next subsection.

Recommendations – Impelling Positive Change

	Individual actions: Consider the environmental impact of work practices. Be proactive in seeking best practice. Make and model positive change in research activities. Drive positive group and institutional actions.
	Further group actions: Include critical assessment of the environmental impact of all activities during planning stages. Monitor, assess, report on and set targets in relation to the environmental impacts of research activities. Drive institutional actions, and encourage, support and incentivise individual actions, e.g., through training.
	Further institutional actions: Require funding applications to outline plans for monitoring, reporting and minimising adverse environmental impacts, and for ensuring that research is undertaken in line with principles of social justice. Allow flexibility in policies and procedures e.g., budget allocation, that enable environmentally sustainable choices to be made. Ensure that degree programmes include a focus on global citizenship, encompassing environmental sustainability and associated social justice implications. Acknowledge focus on environmental sustainability and social justice in the accreditation of degrees by governments and professional bodies. Encourage, support and incentivise individual and group actions, e.g., by considering them in professional development and appraisal processes.

1.4 United Nations Sustainable Development Goals

As a global research community, HECAP+ has an impact on society all over the world. We contribute to basic scientific knowledge, drive innovation and promote international collaboration. Our institutions are large employers, large consumers of goods and services, and a key resource in the training and development of national skills bases. This places our institutions in a position to influence policy decisions, drive investment in local infrastructure, and leverage wider improvements to social and environmental standards. For these reasons, the HECAP+ community is in a strong position to support the UN Sustainable Development Goals (sdgs), summarised in Figure 5.

The topics discussed in this document are meant to support a multiplicity of these goals, and we aim to signpost the influence of our work in all aspects. The goals are listed below, with examples of how each is impacted by the HECAP+ research community and its work. The SDGs are defined in UN resolution A/RES/70/1 in detail [31]. It is impossible to cover all aspects in this document, but the manifold impact of the HECAP+ community on sustainable development is clear from this non-exhaustive list.

2 Computing

Computing represents an integral part of basic research, being used for theoretical modelling, simulation (including lattice simulation), and data analysis. With increasing data sets and demands for accuracy, computing resource consumption is expected to rise. This poses concerns in the context of climate sustainability. Within hecap, e.g., the High-Luminosity phase of the Large Hadron Collider (hl-lhc), expected to be operational from the end of this decade, will require 5 to 10 times the computing capacity needed for the Large Hadron Collider (lhc), with data storage needs reaching about ten exabytes [32]. At the same time, some lattice quantum chromodynamics (QCD) calculations, applied, e.g., to studying heavy quark decays and anomalous magnetic moments, can be too expensive to pursue, even if approximately 10% of open-science super-computing in the United States is devoted to such studies [33]. Up to 88% of the electricity consumption of an astronomy researcher at mpia, shown in Figure 4, is due to (super)computing [19], and cern’s (now defunct) data centre in Hungary is responsible for a third of its electricity emissions when the LHC is not running [1].⁸

HECAP+ research infrastructure ranges from local and portable computing, to high-performance computing (HPC) and high-throughput computing (HTC)⁹ in centralised computing centres that — depending on the application — deal with large volumes of experimental data. As an international community, we also rely on communication technologies and the ability to move these large volumes of data around the globe. The infrastructure we use to do so, comprising hardware, the data centres within which the hardware is housed, and cloud computing resources used for data storage, contributes to our community’s energy consumption and the waste that our research generates. Furthermore, the energy efficiency of hardware is ultimately limited by the efficiency of the computer programmes that run on this hardware, making the ghg emissions of HECAP+ researchers dependent upon the choice of software architecture.

This chapter covers sustainability in procurement, and extending and optimising the life-cycle of computing equipment in Section 2.1, choice and optimisation of software in Section 2.2, and energy savings in data centres in Section 2.3. For a full discussion of sustainable sourcing in a broader context, as well as information on E-waste and its impact, see Section 7. A brief explanation of the life-cycle analysis used to estimate the cradle-to-grave environmental impact of infrastructure and technology can be found in Section 6. For other aspects of energy use, see Section 3.

Recommendations – Computing

	Individual actions: Make sustainable personal computing choices by considering the necessity of hardware upgrades, the repurposing of hardware, and the environmental credentials of suppliers and their products. Assess and improve the efficiency and portability of codes by considering, e.g., the required resolutions and accuracy. Assess and optimise data transmission and storage needs. Follow best practice in open-access data publishing, prioritising reproducibility and limiting repeat processing. Read the section on E-waste (Section 7).
	Further group actions: Right-size IT requirements and optimise hardware lifecycles. Schedule queueing systems with environmental sustainability in mind, so as to maximise the use of renewables, accounting for the geographical location of servers/data centres.
	Further institutional actions: Ensure that environmental sustainability is a core consideration when designing and choosing sites for large computing infrastructure, such as data centres, including, e.g., the availability of renewables, the efficiency of cooling systems and the reuse of waste heat. Proceduralise the repair, upgrade and repurposing of existing computing, the de-inventorising of personal equipment for leaving personnel or for donation, and the responsible recycling of retired hardware. Select cloud computing services for their carbon emission mitigation policies.

2.1 Hardware

When considering the future of sustainability in HECAP+, the hardware aspect of computing is of great concern. Hardware is both energy- and resource-consuming. The manufacture, transport, energy consumption, and disposal of each piece of hardware contribute substantially to the environmental footprint of the HPC that HECAP+ relies on to analyse large swathes of data.

Manufacture is the largest source of hardware GHG emissions, with primarily fossil-fuel-powered manufacturing chains contributing as much as 80–85% of lifetime emissions of a personal computing device [34], [35]. Moreover, production is notoriously resource-intensive [34], with the mining of the necessary metals and ‘conflict’ minerals responsible for a number of negative environmental and social effects. Improper disposal of substances found in computing equipment is also linked to environmental hazards and a variety of other risks. For an in-depth discussion, see Section 7.

One way to mitigate the impact due to production is by purchasing modular equipment, which allows for easy upgrades and repurposing of hardware. In fact, the extension of hardware lifetime has been increasingly demonstrated to have major benefits over upgrading to more efficient technology. A study by the University of Edinburgh Department for Social Responsibility and Sustainability [36] found that simply using 174 computer monitors for six years instead of four saved 33 tco2e, which, when incorporated into standard practice, would not only reduce purchasing costs, but would result in annual GHG savings of 380 \(\mathrm{tCO}_2\mathrm{e}\). It is also crucial that institutions be able and willing to support repairs. This applies in particular to personal equipment, e.g., laptops, which come with additional peripherals such as display, keyboard, and housing, as compared with HPC units in data centres.

Furthermore, prioritising suppliers that implement sustainable sourcing, including recovery of secondary materials, and manufacturing methods would partially mitigate the resource burden, as would enabling circularity and appropriate E-waste recycling. As one example, TCO certification [37] is the world-leading sustainability certification for IT products, such as those supplied by Lenovo, Dell, or Acer. TCO-certified compliance is independently verified both pre- and post-certification. TCO certification also covers data centre products, which could be given preference over uncertified ones for cluster computing. For more information on sustainable procurement, including some hallmarks of sustainability in raw materials supply chains, see Section 7.1. For further discussion of E-waste, see Section 7.2.

A secondary source of hardware emissions is energy consumption during its use [34], with the majority coming from processors, memory, and runtime of jobs. Processor upgrades and the optimisation of memory type can greatly reduce energy consumption. See Case Study [case:LHCb] for details of energy-efficient hardware purchase at the LHCb experiment at CERN.

It is important to ensure ‘energy proportionality’ in hardware use, i.e., that energy consumption is proportional to computing performance over the full range of applications [38]. Often, hardware designed to be most efficient at maximum performance load in practice spends most of its time idle, or performing less intensive computations. This can be addressed by, e.g., running jobs at high utilization rate on as few servers as possible.

Implementing parallelisation within processors can also reduce the number of processors needed, and by replacing central processing units (cpu) with graphics processing units (gpu), the energy usage can be reduced. For certain tasks relevant to HECAP+ applications, other even more specialised processors are available, such as Google’s tensor processing unit (TPU) [39]. This consumes less power than its predecessors, although it suffers from poor energy proportionality: at 10% load, it consumes almost 90% of the power it would consume at 100% load [40].

However, it should always be tested whether parallelisation does reduce the overall energy usage of a task, as an increase in energy consumption per second could counteract the benefits of reduced runtime. Another aspect to take into consideration when implementing parallelisation is the particular application, and its requirements in terms of memory, scalability, and data access. Reference [41] discusses these issues in the context of the Worldwide Large Hadron Collider Computing Grid (WLCG). It also gives suggestions for power-aware software applications and scheduling that could reduce power consumption. Some of the advocated changes are software specific and are further detailed in Section 2.2. The Green500 list [42] ranks the most energy efficient high-performance computing systems. The GHG emissions of the computer centres that house them, however, depend critically on their infrastructure. This aspect is further discussed in Section 2.3.

2.2 Software

Software is integral to the work of HECAP+. It underpins how the global HECAP+ community communicates, shares data, produces papers and graphics, and acquires, manages, processes and analyses huge amounts of data from experiments, observatories and simulations.

It is therefore pivotal that the software developed and used by the HECAP+ community is efficient in order to minimise CPU hours, and to facilitate data sharing and long-term reproducibility. This requires a balance to be struck between portability and optimisation for particular architectures. While not directly linked to environmental sustainability, initiatives focused on software sustainability in HECAP+, such as the Institution for Research and Innovation in Software for High Energy Physics (IRIS-HEP) [43] and the hep Software Foundation [44], may provide an important platform for accelerating the inclusion of environmental considerations in software development. Doing so is compatible with the FAIR principles [45] for scientific data management, that software (and data) should be Findable, Accessible, Interoperable and Reusable.

Much of the code used in HECAP+ computing relies on libraries and public codes. Experiments use general frameworks and software infrastructure provided by experts in the experiments. They can have a tremendous impact on the energy efficiency of the employed code and, in some cases, work to meet strict requirements posed by the computing environment. Decisions on the computing language employed can be crucial, with Fortran and C++ specifically suited for numerical calculations, whilst others prioritise convenience or readability over performance. Changes in processor architecture have been utilised through dedicated and collaborative efforts, leading to a factor of 2 improvement in the performance (and energy efficiency) of the reconstruction code of the ATLAS experiment at CERN [46]. Other examples of software improvement are recent changes to the software framework and architecture at LHCb (see Case Study [case:LHCb]) and improvements in a Monte Carlo (MC) generator core code, having led to an improvement in speed of a factor of 50 (see Best Practice [CS:SoftwareOptimisation]). In the case of cosmological analyses, it has been suggested that the Likelihood Inference Neural Network Accelerator (LINNA) can lead to efficiency increases that would save $300,000 in energy costs and around 2,200 tco2 in first-year Rubin Observatory’s Legacy Survey of Space and Time (LSST) analyses [47].

Sustainable use of software can also be encouraged at an individual level. The energy used in a job directly correlates with the memory assigned/available for a job, so mitigation by individuals can be easily implemented through assigning the correct memory used and by optimising code [48]. Further examples of conscientious use of software include limiting resolution or precision to that which is necessary, effective testing to avoid wasted CPU hours, good practice in data retention to avoid data loss and the need to rerun analysis or simulations, and scheduling CPU hours when a higher percentage of the local energy mix is from renewables.

2.3 Infrastructure

Even the most energy-efficient data centres are not environmentally sustainable if they are powered by carbon-based fuels [27]. However, provided energy from renewable sources is available, this can be easily addressed. Indeed, there are many advantages to doing so, owing to the flexibility of HTC. Inherent fluctuations in supply of electricity from renewables can be managed using a smart queueing system that runs jobs at times where electricity has a large renewable component, or directs them to data centres where this is the case. Moreover, a carefully managed HTC system can even help stabilise fully-renewable power grids in response to local imbalances in supply and demand: an instantaneous reduction in the CPU clock frequency by up to 60% ensures per-second grid stabilisation, and a similar technique can be employed on longer time scales, e.g., hourly, in response to changes in the carbon intensity of electricity (or equivalently, market price). For longer periods, with higher latency, this can also involve powering down nodes. The reduced work can be compensated by operating older hardware longer, but only when the electricity price is low. See Section 3.1 for further discussion of renewables-based grid infrastructure.

Another source of GHG emissions associated with computing is the construction and operation of the large data centres within which IT equipment is housed. Although emissions due to construction can be significant, particularly if concrete is used, our focus in the remainder of the section will be on cooling the facilities and equipment, which is responsible, on average, for almost one third of facility power use. A judicious choice of location for the centre can minimise these energy costs, by provision of a cooler external environment, or other means to cool efficiently. Proximity to a large body of water, e.g., could make water cooling an attractive option. Care must be taken, however, to ensure minimal disruption to the natural environment. Waste heat from the data centre can also be reused to heat nearby infrastructure. For examples of best practice in data centre design and construction, see Best Practice [BP:CSCS], Best Practice [BP:Prevessin], and Best Practice [BP:GreenITCube]. For more information on energy-efficient LHCb computing infrastructure, see Case Study [case:LHCb].

3 Energy

The operation of experimental equipment and computing facilities at large-scale research centres has a significant energy footprint. In addition, energy is required for the construction and disassembly of infrastructure, for heating and cooling buildings, and for transport of goods and people. To comply with the Paris Agreement, future facilities must be effectively climate-neutral, and this presents a significant challenge for hecap.

Particle collider experiments are particularly power-hungry, with the lhc at cern being a prominent example. With its particle accelerators, detectors and extensive infrastructure, CERN consumes up to 1,300 GWh of electricity annually, of which 55% is due to LHC operations [1], [58]. CERN plans to significantly increase the scale of its installations in a push towards higher energies and intensities. Doing so responsibly will require a concerted effort to minimise power consumption and increase the energy efficiency of the infrastructure, and a careful analysis of how to source the remaining energy needs sustainably.

CERN receives most of its electricity from the French grid, which is currently characterised by a high share of low-carbon nuclear power, suppressing its electricity-related \(\mathrm{CO}_2\) emissions in comparison with other facilities (see Figure 4).¹¹ However, taking into consideration the decreasing share of nuclear power in the French grid over the last 15 years, as well as the wider common European electricity market, where fossil fuels account for 37% of electricity production on average [59], the outlook is more worrying.¹²

It is important to place the energy needs of HECAP+ research infrastructure within the context of the world’s necessary and rapid transition to zero-carbon energy sources. Global primary energy consumption in 2019 was approximately 160,000 TWh (equivalent to an average power consumption of 18,000 GW), around 80% of which comes from \(\mathrm{CO}_2\)-emitting fossil fuels [61]. Moreover, demand is rising, primarily due to the growth and industrialisation of emerging countries. A 50% reduction of \(\mathrm{CO}_2\) emissions by 2030, as stipulated in the Paris Agreement in combination with the latest ipcc scenarios (see Figure 1), will create a huge global energy gap, as shown in Figure 6 [62].

Many experimental technologies such as \(\mathrm{CO}_2\) capture and storage (CCS) will not be viable for large-scale implementation within this short time frame [66]. Filling the energy gap with solar, wind, and nuclear power requires upscaling existing facilities by more than an order of magnitude within seven years, a not inconsiderable task. Therefore, substantial increases in energy efficiency will be indispensable. Even tripling the output of existing solar, wind and nuclear installations by 2030 (which may itself be unrealistic), bridging the fossil energy gap would still require a 40% global efficiency increase compared to today. Substantial energy savings requires systemic changes in technology and behaviour, such as transitioning from combustion engines to electric motors, from gas heating to heat pumps and from cars to rail. The global situation is therefore likely to result in energy becoming scarce and expensive in the coming decades, with the potential to directly limit our capabilities to conduct energy-intensive experiments and data analysis in basic research.

Section 3.1 elaborates on the wider context of global production of low-carbon energy and focuses on potential sources of sustainable energy for HECAP+ research infrastructure, as well as energy savings and recuperation in Section 3.2. Saving energy through structural and organisational changes are described elsewhere: see Section 2 for computing, Section 5 for mobility, and Section 7 for procurement.

Recommendations – Energy

	Individual actions: Save energy in all ways practicable, e.g., by avoiding unnecessary heating or cooling of workspace, and by turning off electrical items when not in use. Read the sections about computing (Section 2) and mobility (Section 5).
	Further group actions: Ensure that energy efficiency is a major focus in experimental design, and prioritise technologies that minimise consumption and maximise energy recovery. Monitor, report, and assess energy usage with the aim of reducing consumption and resulting emissions. Read the section on research infrastructure and technology (Section 6).
	Further institutional actions: Ensure that energy efficiency is a major factor in the renovation of existing estates and the design and construction of new infrastructure. Prioritise moving to sustainable and renewable energy sources via both local generation, and energy import and export. Collate and publish energy usage and emissions statistics, stratifying by source, e.g., heating, experimental infrastructure, computing, transportation, and procurement. Lobby for environmentally sustainable energy policy.

3.1 Low-Carbon Energy

Transitioning the energy demands of HECAP+ research to \(\mathrm{CO}_2\)-neutral sources will likely require a mix of sources: solar, wind, hydro, geothermal, and nuclear power, many of which will be strongly location dependent. Despite their relatively low cost (see Figure 7), the geographical limitations of renewable energy, combined with the challenge that significantly increasing the share of nuclear power presents, (see the paragraph on nuclear power for details), will make a rapid transition to carbon-neutral energy difficult for HECAP+, and large-scale transmission or import of sustainable energy may offer a solution. Alternatively, efforts could be made to site future facilities near abundant sources of renewable energy, which could have the additional benefit of contributing to developing economies over a longer period. The Synchrotron-Light for Experimental Science and Applications in the Middle East (sesame) project  [67] is one example. World-leading research centres like cern, with their history of cooperation across political and ideological boundaries, are uniquely placed to spearhead such initiatives.

3.1.0.1 Solar

Solar energy is abundant, and near-universally available. Its intensity depends on latitude, with the highest efficiencies in the deserts of the sun belts north and south of the equator. According to a 2021 report by the Carbon Tracker Initiative [70], populating an area of only 450,000 km\(^2\) with solar panels would be sufficient to satisfy global energy demands. This corresponds to an area the size of Morocco, two-thirds that of Texas, or 4% of European landmass.¹³ According to the IPCC [66], “The global technical potential of direct solar energy far exceeds that of any other renewable energy resource and is well beyond the total amount of energy needed to support ambitious mitigation over the current century.” However, large-scale adoption of photovoltaic (PV) panels poses concerns due to resource use, particularly the energy-intensive production of silicon used to produce the panels, and end-of-life waste generation (see Section 7 for more information). These impacts can be partially mitigated by material recovery in PV cell recycling, and their reuse [71].

Solar panels are easily retrofitted onto existing infrastructure, and their price has dropped by almost 90% in the last two decades. Unfortunately, solar power can be unavailable when it is needed most: at night and during winter (in countries at higher latitudes), leading to a need to increase its efficiency and storage capacity. An overview of energy storage technologies can be found in Section 3.1.1. See Case Study [case:solar] for a study of the implementation of in-house solar power at CERN.

3.1.0.2 Wind

By comparison with solar energy, wind energy is more sensitive to local conditions. In Europe, competitive locations for wind energy, with costs below per kWh, are primarily offshore, and are concentrated along the coasts of the North and Baltic Seas, the Bay of Biscay, and the English Channel [72]. Landlocked countries, such as Switzerland or Austria, are generally less suited for production of energy through wind turbines. Producing 25% of, e.g., Swiss energy demand from wind power would require populating 100% of its agricultural farmland with wind turbines (although this does not preclude growing crops beneath the windmills). By contrast, fulfilling a quarter of Danish or Estonian energy demand would require less than 4% of the farmland [72].

3.1.0.3 Hydroelectric

Water power is even more reliant on local conditions, such as high flows or water volumes and large altitude difference, which naturally limits its applicability. However, the energy output of hydroelectric plants is steady, and can be adjusted to demand very quickly, making it a good complement to other renewable sources. It can also be used for energy storage, see Section 3.1.1. The largest hydroelectric capacity is in China, which produces almost 30% of the global hydroelectric power [73], thanks to its large projects in the Yangtze River valleys.

Mega-dams, however, constitute a large intervention on the natural environment, and consequently come with associated risks, such as landslides, earthquakes, and destruction of habitats, and can themselves be a source of the potent ghg methane when flooded flora rots. The Three Gorges dam in particular has been controversial both domestically and abroad [74]. In arid areas, or during periods of drought, which are expected to become more prevalent due to climate change, hydroelectric power may be in competition with agricultural needs, and climate change may jeopardise future yields of existing dams.

While the potential for marine power generation from ocean currents, tides, waves, and gradients in salt and temperature (collectively known as Ocean Energy Technologies) is huge, there is no technology currently mature enough to produce marine power at large scale [75].

3.1.0.4 Geothermal

Geothermal energy is a stable source of renewable energy. It consists of residual heat from the time when the Earth was formed and of heat newly produced inside the Earth due to radioactive decay of hot elements in the mantle, by tidal forces due to the Moon and Sun, or by friction along tectonic plate boundaries. Although it has low intensity compared with solar energy, its technical potential of about \(1.4 \times10^6\) TW-years is around three times total global energy consumption [76].

The most easily exploitable is the ‘shallow’ geothermal energy stored in the upper few metres of the Earth’s surface. This can be employed to provide space heating or cooling for buildings and urban areas using buried pipes containing a circulating fluid as a heat exchanger, and a geothermal heat pump [77]. The low thermal conductivity of the ground limits the total amount of geothermal heat that can be exploited and depends strongly on rain and ground water. Modern geothermal heat pumps use the ground as heat storage and not so much as heat source. The ground heat that is extracted in winter is regenerated in summer by using the heat pump for cooling the building.

Geothermal power generation, however, requires higher temperatures. Easily accessed only in areas of volcanic activity, and along plate boundaries, much geothermal power-generating potential is locked up below common drilling depths, where the rock is less porous. Enhanced Geothermal Systems (EGS) [78] induce porosity by fracturing deeper, hotter rock using high-pressure water injection, to allow for fluid circulation. This hydrothermal ‘fracking’ has attracted significant controversy as, in addition to the injection of toxic chemicals into the Earth, which can then pollute nearby groundwater sources, it brings a risk of induced seismic activity if unwittingly carried out near a ‘locked’ dormant fault, and was thought to be responsible for triggering a magnitude-5.4 earthquake in the South Korean city of Pohang in November 2017 [79], [80].

3.1.0.5 Nuclear

Nuclear power is a source of low-carbon electricity. Nevertheless, an energy source is only termed sustainable if it does not carry any significant long-term risk for future generations. This definition of sustainability based on the Brundtland Report [81] has also been adopted by the International Atomic Energy Agency (IAEA) [82], which argues for the ‘weak sustainability’ of nuclear power.

The share of primary energy from nuclear power has been decreasing on all continents except Asia over the last two decades, and has fallen below 4% of global primary energy production [83]. According to the IAEA, nuclear reactors have a median construction time of 93 months [84], not including planning and permissions. See Case Study [cs:nuclear_gap] for an estimate of how many nuclear power stations must be constructed to cover global energy needs.

Safety, security and climate resilience of the reactors, and availability of fuel, as well as storage of spent fuel, are important challenges. The exact form these take is crucially dependent on future technological developments, to which HECAP+ research contributes directly. Today, several new reactor types are being developed, which promise to have additional safety features, an efficient use of more abundant isotopes and less long-lasting nuclear waste. HECAP+ research can also contribute to non-proliferation efforts, see e.g., Ref. [85]. Bringing developing technologies to maturity and commercial viability takes time, and in the near term the IPCC does not assess favourably the mitigation potential for nuclear energy, see Figure 7.

Case Study : Local solar power at CERN
In-house solar power production is not sufficient to cover the full needs of a huge laboratory such as cern. Nevertheless, it can make up an important contribution to foster a fast transition to renewables.¹⁵ Research centres are often characterised by the many flat rooftops. These rooftops make excellent locations for installing solar photovoltaic (PV) panels.

Using publicly available tools provided by the Canton of Geneva [89] and the Swiss Federal Department of Energy [90], it is possible to estimate the solar potential of these rooftops. Similar public tools are now available for most countries, provided by local governments or Non-Governmental Organizations (NGOs). Figure 8 shows part of the main CERN site as taken from the Geneva solar cadastre [89]. Buildings in red are classified as “optimal” for their orientation towards the sun. The large rectangular building in the middle is assembly hall 157. The cadastre lists an estimate of 392 MWh per annum of electricity generation for the south-west half of this 2,055 m\(^2\) roof, with the other part capable of producing an additional 335 MWh per annum. CERN has 653 buildings with a total roof area of 421,000 m\(^2\),¹⁶ which amounts to approximately 80 GWh annual electricity generation potential. A comparison with the electricity consumption in 2019 of 428 GWh [1], when the lhc was not in operation, shows that around 18% of CERN’s basic (non-LHC) electricity demand could be produced locally with solar power.

Using the cadastre, the cost for electricity from rooftop PV for CERN can be estimated to be fixed around at /MWh for the next 30 years. This cost is well below current wholesale market spot prices in France (>/MWh), but also below the average price over summer 2021 (> /MWh) [91].

3.1.1 Renewable grid infrastructure

For global power grids to rely more heavily on intermittent sources of renewable energy, such as wind and solar photovoltaic, assessed as having the highest climate mitigation potential to 2030 by the IPCC (see Figure 7), much of the existing grid infrastructure and controls will need to be updated to smooth out fluctuations in supply and demand [92]. Grid inertia, which acts as a short-term buffer in fossil-fuel grids in periods where electricity demand outstrips supply, is significantly lower for intermittent sources, compromising the stability of the grid. ‘Smart grid’ infrastructure must provide peak-shifting capabilities and fast frequency response to stabilise electricity supply despite the lower intrinsic inertia of the grid. It will also need to draw upon a novel and expanded energy storage capacity to bridge longer-term gaps between supply and demand [93] (for further discussion, see below), as well as the capability to regulate bi-directional flow of electricity, well-suited to the distributed generation of intermittents. Inverter-based resources with low inertia are ideally suited to near-instantaneous response [92]. Electronic control of this response, coupled with developing ‘grid-forming’ technologies, which allow inverters to emulate a traditional grid’s stable frequency, as well as automated demand-side response to voluntarily disconnect non-critical loads momentarily, have the potential to transform our existing networks [92].

Existing solutions have been utilised successfully in several fully renewable island microgrids, and on a larger scale in the Electric Reliability Council of Texas (ERCOT), the smallest of the three power grids in the United States, which achieved 58% instantaneous wind penetration in 2019 [92]. Scaling up these solutions requires further research, although several highly cited studies argue for the feasibility of 100% renewable-based grids world-wide at low cost, eschewing any fossil fuel or nuclear energy component (for a comprehensive review, see Ref. [94] and also Ref. [95]).

3.1.1.1 Energy storage

The feasibility of pure renewable-based grids is crucially dependent on an increased energy storage capacity to smooth out fluctuations (on timescales ranging from diurnal to seasonal) in the supply of intermittent renewables and demand [93]. While the cost of Li-ion batteries have plummeted 40-fold in the past 35 years [96], and notwithstanding the implications of resource extraction (see, e.g., Section 7), their development has focused on short-duration portable energy storage, as driven by needs of the electric vehicle industry. Projections show that in order to minimise the costs of a net-zero energy system, storage capacity must increase by almost an order of magnitude, from 160 GW (and 9 TWh total capacity) today, to 1.5–2.5 TW (85-140 TWh) globally by 2040 [93].

Most existing and planned storage capacity is in pumped storage hydropower, a mechanical form of storage where water pumped into a reservoir at high elevation turns a turbine as it flows to one at lower elevation [93]. As well as being geographically limited, however, these open-air reservoirs suffer the same environmental problems as other hydropower projects (see discussion above), and are similarly subject to the vagaries of the climate. A new promising approach to pump storage is the use of undersea bowls that are evacuated to store energy which is restored when they are filled up with water again. An even simpler approach is to build a large ring wall inside of a deep lake, and empty and refill the internal area using pump-turbines. This way, pumped-storage hydroelectricity does not require a separate upper and lower lake, and a single lake is sufficient. Defunct open pit mines, large natural lakes and even the sea allow for new opportunities to install large-scale storage devices with less environmental impact [97], [98].

Interesting alternatives include storing energy as compressed air; or as latent heat in, e.g., aluminium alloy [93]. However, there is need for significant investment in the energy storage sector to bring new ideas, including novel mechanical, thermal, electrochemical and chemical storage methods to commercial viability [93]. For more details on capacity and market-readiness of promising long-duration energy storage methods, see Ref. [93].

3.1.2 Energy import and export

The uneven geographical distribution of sources of renewable energy leads to the question of whether large-scale import and export of renewable energy could be a cost-effective way of closing the energy gap. For Europe, detailed studies have shown that energy import by cable, as well as by chemical energy carriers, have comparable or lower costs compared to local energy harvesting [99].

Technical options to transport electricity over long distances have improved significantly in the last decades. In South America and China, projects to transport electricity over more than 2,000 km by Ultra High Voltage Direct Current (UHVDC) lines are already operational [100]. The Viking Link UK-Denmark project provides another example [101]. For a summary of all current and planned European transmission projects, see Ref. [102].

This technological progress opens up an alternative to the traditional import and export of chemical energy: direct import of renewable electricity, e.g., from Northern Africa to Europe [97], [103]. An excellent potential for electricity generation by solar and wind power, large unused tracts of land, and existing energy trade partnerships for fossil fuels make North African countries ideal export partners for procuring electricity from sustainable sources. Indeed, the Xlinks Morocco-UK Power Project [104] aims to connect a solar and wind energy facility in Morocco’s Guelmim Oued Noun region to the UK energy grid by 3,800 km HVDC sub-sea cables by 2030.

While the import and export of energy is a promising solution on the technical and economic level, constructing wind or solar farms, e.g., in the sun belts of Africa, to then export the power to Europe involves geopolitical and social considerations. Resource and person-power extraction from Africa to the benefit of Europe and America has a long, reprehensible colonial history. It is of the utmost importance to make fair power trade agreements between the continents that ensure strong integration into local communities and include the local population in the planning and implementation of such projects and related infrastructure. In this way, a win-win situation for all stakeholders should be ensured, and well-planned cooperation has the potential to act in a geopolitically stabilizing way in line with the 16th and 17th UN sdgs ("peace, justice and strong institutions" and "partnership for the goals"). If achievable, equitable import and export of renewable energy could be considered as part of a catalogue of solutions to cover future global energy needs.

The HECAP+ communities have a record of successful collaboration between nations, and they could therefore be important players in proving the viability of such partnerships. One possible scenario for the transmission of solar energy is detailed in Case Study [case:desert].

Case Study : CERN-LINK — Clean power from the desert
The hecap community, and cern in particular, has a long history of effective cooperation across geographical and socio-political boundaries, in the pursuit of science. CERN brought scientists from East and West together during the Cold War, and Arabic and Israeli people together for the sesame project, the first accelerator laboratory powered by solar energy from the desert [67]. This makes CERN ideally placed to spearhead a project to import energy from countries rich in renewable energy sources, and transport it across international boundaries. This type of spin-off could help cover CERN’s energy needs, while also reinforcing the idea that fundamental research has the potential to solve problems outside its immediate purview in new and innovative ways.

The technology for long-distance energy transmission, in the form of High-Voltage Direct Current (HVDC) lines, is commercially available, and is being increasingly employed to transport energy from renewable-rich to renewable-poor regions. Prominent examples include the Viking Link UK-Denmark project, which broke ground in July 2020 [101], as well as the planned XLinks Morocco-UK power project [104].

Even so, it is of paramount importance that fair power trade agreements be put in place to ensure mutual benefit to all stakeholders, including the local communities hosting the energy-harvesting infrastructure. It is also important to acknowledge that additional environmental considerations are required when planning and implementing a project of this type, in terms of minimising the impacts on local ecosystems, as well as the marine environments across which the underwater cables would be installed.

A scenario for connecting, e.g., Morocco, Algeria or Tunisia to Southern France, Spain or Italy by sub-sea cable is plausible from a technological point of view (for a detailed feasibility study, see Ref. [106]), and could be employed for HECAP+ applications (see Figure 9). Costs are estimated to be around /kWh for a year-round power supply of 3.6 GW in the daytime and 2.2 GW at night [97], [103], [107]. This estimate includes infrastructure costs for generating the electricity, buffer storage and transmission line costs. Feasibility and cost estimates agree well with those for previously proposed commercial projects [104].

Electricity import on this scale would exceed the power needs of CERN, and surplus power could be returned to the European electricity grid to power other research institutions and universities that join the initiative. Southern France is well-suited to the role of import terminal for electricity due to its pre-existing grid infrastructure, as well as its proximity to CERN and other major research institutions.

3.2 Energy Saving and Recuperation

A first step in reducing energy usage is energy monitoring, which will allow us to assess where improvements are needed. The best energy saving measures will be individual to each location and facility, making it hard to recommend specific actions here, although insulating buildings and ensuring that the heating/ cooling systems are maximally efficient are universally applicable measures. For an example of energy recuperation in the context of basic research, see Best Practice [BP:DESY_recycling].

Most of the energy budget for many high energy experiments is due to the accelerators and detectors. Initiatives to reduce their energy use are many and varied. Relevant references for detectors are collected in Section 6, see also the discussion on energy saving in the LHCb experiment at CERN in Case Study [case:LHCb]. A particularly impressive example of energy-efficient accelerator design is the Energy Recovery Linac Test Accelerator (CBETA) [108], based in Cornell. This accelerator saves energy, both by recovering the energy of the bunched particles to accelerate the next batch, and by using permanent magnets to guide the particle beam. See Best Practice [BP:EnergyRecoveryAccelerator] for more details and Case Study [case:wakefield] for energy savings using plasma wakefield acceleration technology.

4 Food

Over a quarter of global ghg emissions comes from food production [111], and a quarter of this food is lost in the supply chain, or discarded by consumers [112]. Plant and animal agriculture also has extensive and profound negative impacts on the environment through land use, freshwater use and pollution, and terrestrial acidification.

A recent article in Science argued that limiting warming to 1.5C will not be possible without “ambitious changes to food systems” even if fossil fuel emissions are immediately halted [113]. The most impactful change reported in the study is a global switch to the healthy, plant-rich diet recommended by the EAT-Lancet commission [114], which is one that can be implemented immediately, and on an individual level. Supplementing this with measures such as reducing food waste and increasing efficiencies in food production could result in a net carbon-neutral food system by 2100 [113]. While sourcing sustainably-grown food and ‘eating local’ can have a positive impact on food-related emissions (of which transportation is responsible for 6%, see Figure 10), the largest impact can be achieved by reducing the consumption of high-methane emitters, such as beef, lamb and dairy [115]–[117], OWID-Sustainable?.

However, choices related to the food that we eat are deeply personal and often loaded with cultural and social significance. As such, it is important to acknowledge that changes to food systems will be a gradual process and will not have a ‘one size fits all’ solution. Their equitable implementation will require cross-disciplinary analysis of the implications of such changes for all stakeholders. This includes producers and the local populations to which they belong, and steps must be taken to ensure communities are empowered and resilient to multiple overlapping pressures, from climate change and markets [118]. The devastating impact of the quinoa boom and bust on pastoral communities in Bolivia provides one well-documented example [118], [119]. Moreover, such changes must account for global disparities in wealth, and the variations in availability and access to food sources, to avoid further cementing geographic inequalities in diet. Notwithstanding the care that these factors necessitate, a significant proportion of the HECAP+ community is in the privileged position to be able to reduce their consumption of animal-derived food products and minimise food waste.

Recommendations – Food

	Individual actions: Reduce consumption of animal products, especially those that result in the highest emissions, e.g., ruminant meat, and dairy. Minimise food waste.
	Further group actions: Prioritise plant-based options in conference catering, and optimise service method to reduce food waste.
	Further institutional actions: Incentivise the consumption of plant-based products at on-site restaurants by increasing their variety and quality, and subsidising their cost. Highlight the environmental impact of food choices through service layout and labelling. Minimise food waste by providing multiple portion sizes and donating unused food. Read section on waste (Section 7) and limit food-service waste e.g., through industrial composting of biodegradable food containers.

4.1 Food Production

Figure 10 reveals the overall environmental impact of food production. The agriculture sector uses 70% of the world’s fresh water reserves and has caused eutrophication¹⁷ of most of the world’s oceans and freshwater. It is responsible for large-scale deforestation and habitat loss [115], [120], [122], resulting in an historic low in mammalian biodiversity, with total mammal biomass dominated by humans and their livestock [121].

Animal agriculture is responsible for just over half of GHG emissions from the food sector, due to direct emissions from livestock and fisheries, land use, and production of crops for animal consumption.¹⁸ It accounts for three-quarters of global agricultural land use, while providing just a fifth of the world’s calories, and under 40% of its protein supply [115], [120], [122]. The over-use of antibiotics in animal agriculture is partially responsible for the development of antimicrobial resistance in "superbugs" [124], and may be a risk factor for the emergence of new zoonotic diseases [125]–[127]. Furthermore, there is substantial evidence linking high intake of red meat to an increased rate of heart disease [128].

Shifting consumption away from animal products to a more plant-based diet would significantly reduce both the environmental and healthcare costs of food systems. The potential annual reduction in GHG emissions from eliminating different food groups from our diet is shown in Figure 11. Beef, lamb, and dairy, responsible for the largest cumulative global emissions, are also among the highest emitters per 100 grams of protein, see Figure 12. In addition, animal products are generally more expensive than plant products [129], as well as being less inclusive of people with dietary restrictions or preferences, due to religion, lifestyle choices, food allergies or intolerances.

4.2 Food Service

Several universities and other institutes for higher education have implemented measures to limit or eliminate consumption of animal-derived foods and reduce food waste, including eliminating red meat from their cafeterias, increasing the quality and variety of plant-based options, changing the cafeteria layout, and modifying default meal options and food labelling [132]–[135]. By way of illustration, we quantify the GHG savings due to replacing beef with alternative sources of protein in the weekly menu of cern Restaurant 1 in Case Study [case:CERNR1].

At conferences and workshops, the primary purpose of any food served is to create additional opportunities for attendees to mingle and discuss. As such, conference organisers enjoy more leeway to make sustainable food choices the default option for these short-term, small-scale events. Best Practice [catering_bp] contains two examples of successful physics conferences with plant-based catering. They highlight, among other things, the importance of institutional partnerships with plant-friendly caterers, and organisers should push for these if they do not already exist. For further discussion on sustainability at conferences, see Section [sec:ConferenceWaste].

Case Study : Sustainable catering at CERN Restaurant 1
cern Restaurant 1 (R1) serves an average of 2,000 meals per day [136]. It offers five hot meal options daily, and has recently overhauled its menu options to include a larger variety of vegetarian and plant-based options, including at least one plant-based main course. We assume each of the five mains are chosen with equal likelihood, and neglect cold food options, such as salads and sandwiches.

In, e.g., the week beginning 27 June 2022, beef, fish and seafood were each served three times as the primary component of the meal, veal once, poultry five times, and fish twice. We assume that these distributions are representative of a typical weekly menu at R1 and that the beef originated from beef-herd cows. The GHG emissions of the various forms of protein are shown in Figure 12.

Substituting each gram of protein from beef with a gram of protein from chicken or fish reduces emissions by 440 g co2e. Assuming a serving contains 20 g of protein¹⁹, substituting all beef meals at R1 with farmed fish or chicken would result in a reduction of its annual carbon footprint by 528 tco2e.²⁰ This corresponds to approximately 260 return flights between London and New York. Both the emissions savings and the overall environmental impact would be even greater if plant-based substitutions were made.

Instituting one weekly meat-free day (taking as a benchmark a day where one beef, one fish, and one poultry meal were served, and replacing them with one tofu-based meal and two pulse-based meals) would result in a reduction of 735 \(\mathrm{tCO}_2\mathrm{e}\) annually, where the bulk of the savings comes from the beef replacement.

5 Mobility

Mobility constitutes a significant portion of the emissions of a hecap researcher. This includes short daily commutes between the home and the workplace, and longer-distance business travel (see Figure 4).

Transport accounted for almost a fifth of total global emissions in 2016 [142], and is the sector that saw the highest growth in pre-COVID years [143]. Demand for car, rail and air transport is expected to continue to increase over time with increasing global population and income levels.

Unsurprisingly, self-powered mobility, such as walking and cycling, are the most carbon-efficient means of transportation, with train travel being the next best. A quantitative comparison between these and other options requires further details to be specified, such as the distance travelled, the fuel efficiency of the vehicle used and the number of passengers carried, and the underlying electricity mix for the country of travel. In the UK, for instance, driving alone in a medium-sized petrol-fuelled car yields smaller ghg emissions than air travel for distances shorter than 1,000 km, whereas flying in economy class beats driving over longer distances [144] (data taken from Ref. [145]).²¹ For a detailed comparison of emissions due to various forms of transport within France, see Figure 13.

When and how we travel, however, are not always free choices, being constrained by existing transport infrastructure, local geography, our research, finances, and caring responsibilities. Universities and HECAP+ institutions, with their large and progressive workforce, can help tip the balance in favour of the more environmentally sustainable option with a judicious combination of policy, incentives, on-site infrastructure and advocacy.

Our current societal infrastructure is set up to facilitate individual travel by car, to the detriment of a large part of the population. Universities, as large employers with a relatively progressive workforce, have the potential to act as instigators of change in this. Making public transport and cycling the preferred options when possible will increase demand for these more sustainable forms of transport and thus encourage cities to improve the infrastructure for them.

Nevertheless, efforts to limit emissions resulting from travel must be balanced against legitimate needs for mobility: the establishing and maintenance of close collaborative relationships, sharing of research outputs, individual exposure and career development, and travel to research facilities. Changes to our travel culture and policies must be implemented so as to benefit and, at the very least, not to worsen barriers to inclusion, by avoiding the disenfranchisement of members of our community such as early-career researchers, members of our community from the Global South or those otherwise geographically isolated.

Recommendations – Mobility

	Individual actions: Re-assess business travel needs, using remote technologies wherever practicable. Choose environmentally sustainable means of transport for daily commutes as well as unavoidable business travel, amalgamating long-distance trips where possible.
	Further group actions: Define mobility requirements and travel policies that minimise emissions, while accounting for the differing needs of particular groups, such as early-career researchers or those who are geographically isolated. Re-assess needs for in-person meetings, and prioritise formats that minimise travel emissions and diversify participation by making use of hybrid, virtual or local hub participation, and optimising the meeting location(s).
	Further institutional actions: Support environmentally sustainable commuting by improving on-site bicycle infrastructure, subsidising public transport and providing shuttle services. Disincentivise car travel where viable alternatives exist, facilitate car pooling, and provide on-site charging stations. Incentivise the reduction of business travel, e.g., by implementing carbon budgets with appropriate concessions. Ensure unavoidable travel is made via environmentally sustainable means through flexible travel policies and budgets, and the use of travel agents that offer multi-modal itineraries. Employ carbon offsetting only as a last resort. Remove any requirement on past mobility as an indication of quality in hiring decisions. Lobby for improved and environmentally sustainable local and regional transport infrastructure.

5.1 Commuting

Changes in commuting patterns are typically affected by life circumstances, including changes in education, employment and residence [146]. The viability of environmentally sustainable mobility, like walking, cycling and taking the train, depends crucially on characteristics of the home and workplace locations, including the distance between them, and their local environment. These properties are seen to influence the relative importance of commuting and business emissions for different HECAP+ institutions.

For example, cern, fnal, and the Department of Physics at eth have wildly different \(\mathrm{CO}_2\) emission profiles due to personal transportation. While emissions due to commuting were roughly equal to those for business travel at FNAL, commuting outweighed business travel for CERN, and conversely, business travel swamped commuting emissions for ETHZ. This reflects the unique environment and characteristics of each these research centres:

• ETHZ is located in an urban centre and is well connected to the local public transport. In 2008, only 1,700 tco2e were recorded for commuting, with 7.5 to 10 times larger emissions attributable to business travel (using numbers from 2006–2012). It is clear here that the emissions per capita for all staff (including researchers) are significantly smaller than those for FNAL or CERN.

• FNAL and CERN have more rural settings, with a 77% majority of CERN employees commuting by car from France. FNAL’s commuter emissions [22] of about 6,000 \(\mathrm{tCO}_2\mathrm{e}\) are approximately on par with business travel emissions,²² whilst CERN quotes 5,836 \(\mathrm{tCO}_2\mathrm{e}\) of commuter emissions compared to 3,330 \(\mathrm{tCO}_2\mathrm{e}\) business travel emissions for its approximate 4,000 staff members. The small amount of travel emissions compared to emissions from commuting reflects to some extent the status of CERN as a scientific centre, to which other members of the community are expected to travel and where travel is easier to avoid, also because the experiments are located at CERN.

Whilst ETHZ, FNAL and CERN face different boundary conditions, all three of them, and HECAP+ institutions in general, should aim to reduce emissions from commuting, even if these contribute to their overall budget to a different degree.

This reduction requires an interplay of institutional and individual actions: while institutions cannot force employees to choose more environmentally sustainable commuting habits, they can incentivise them through various measures, from the availability of bicycle-friendly infrastructure, such as showers and secured/covered parking, to financial incentives for greener transportation. They can also allow employees to avoid long commutes by formalising telecommuting options, which have become more normalised since the start of the COVID-19 pandemic, and use their standing to push local authorities towards better public transit/cycling/carpool infrastructure. Individuals and groups can, on the other hand, push for these actions at the institutional level. Table 1 collects some means by which ETHZ, FNAL and other academic and HECAP+ institutes promote ‘green’ transport. An estimate of the emissions per distance of different forms of transport in France is presented in Figure 13.

5.2 Business Travel

A global scientific endeavour such as HECAP+ will mandate some amount of long-distance travel, e.g., to experimental sites, or to build close working relationships. However, the current academic culture, which rewards hyper-mobility, is neither environmentally sustainable, nor equitable to all scientists. Visa rules and prohibitive long-haul travel costs can make participation in conferences extremely challenging, especially for researchers from the Global South. Moreover, the freedom to travel can be heavily restricted for people with disabilities, health impairments or caring responsibilities. For example, the burden of childcare is still unequally distributed, and this burden falls predominately on female shoulders [158].

Emissions from commercial aviation is a long-recognised problem, contributing 2.5% of \(\mathrm{CO}_2\) emissions and 3.5% of ‘effective radiative forcing’ (a closer measure of aviation’s impact on warming as explained in footnote 21) [159] in 2018. Note that the majority of these emissions derive from the one-tenth of the world’s population that can afford air travel. Almost all HECAP+ scientists belong to the 4% of the population taking international flights, and many fall within the 1% classified as the most frequent flyers [160]. These statistics highlight the inequalities inherent in travel emissions.

More troubling is that global aviation statistics belie the significance of business travel emissions for many HECAP+ researchers, which are comparable to, and in some cases even dwarf, their direct and indirect emissions (see Figure 4). This is clearly in tension with the push to net-zero emissions, particularly given that we do not expect the aviation sector to decarbonise at the same rate as the rest of the transport sector [161], [162].

Emissions related to conference travel have been studied in detail and dominate conference-related emissions [163], [164], contributing annual emissions 30% larger than the total annual transportation emissions for Geneva (720 kilotonnes \(\mathrm{CO}_2\) [165]). However, the \(\mathrm{CO}_2\) emissions for a single conference trip amount to about 7% of an average individual’s total \(\mathrm{CO}_2\) emissions [163]. This might be even worse for HECAP+ researchers, for whom frequent trips to experimental sites and meeting venues to undertake international collaborations are common. See also emissions estimates for business travel of members of the LHCb experiment at CERN in Case Study [case:LHCb].

Discussions about reducing business travel are highly charged, as active engagement with other members of the scientific community is integral to scientific practice. Any changes that we make to HECAP+ travel culture have to be considered in the context of other aspects of our working practices, such as hiring decisions, where any curbs on travel may, e.g., disproportionately impact early-career researchers. At the same time, the reprioritisation of business travel and a move toward a greater share of virtual/hybrid formats can have a positive impact both on the climate and on inclusivity.

For necessary travel, sustainable alternatives to air travel should be prioritised where possible, keeping in mind that the increased travel time and costs of sustainable travel as compared with air travel could make this choice difficult for researchers with caregiving responsibilities, or limited travel budgets. In Case Study [case:traveltoCERN], we compare emissions, travel time and cost of different modes of travel to CERN, from various starting points within Europe, for CMS Week in January 2022. HECAP+ institutions and funding bodies are beginning to implement more sustainable travel policies, including travel top-ups for green travel; we highlight two examples in Best Practice [bp:Erasmus] and Best Practice [bp:DESY_travel].

If the community is to rethink this travel culture and move toward more hybrid/virtual modes of engagement, we must recognise that these require additional planning to maximise engagement, which amounts to much more than streaming the in-person event format (see Case Study [cs:CfH]). It is also important to appreciate that virtual participation requires an internet-ready device and stable connection, and devices with which to connect, which may not be universally available in lower-income countries. A possible remedy for this might be the concept of hub conferences, where the conference has several locations spread globally (see, e.g., Ref. [166]). In Case Study [case:ICHEP], we study travel emissions and participation in the context of the last 5 instances of the International Conference for High Energy Physics (ICHEP) conference, and assess the reduction in emissions from optimising the conference location, moving to a hub model, or hybrid/virtual forms of attendance.

Case Study : Comparative study of travel emissions for ICHEP conferences (2012–2020)
Based on the study of the annual meetings of the American Geophysical Union (AGU) in Ref. [147], and the methodology and software tools employed therein, we undertake a survey of the past five editions of the ICHEP with the aim of assessing the ghg emissions of conference travel to ICHEP, as well as the (geographical) diversity of participants.

The International Conference for High Energy Physics (ICHEP) is a biannual conference with a large and steadily growing participation, of order 1,000 researchers, and a location that alternates mainly between Europe, America and Asia. We study the 5 most recent instances, with locations in Melbourne, Australia (2012); Valencia, Spain (2014); Chicago, United States (2016); Seoul, Korea (2018); and Prague, Czech Republic (2020, fully virtual).²⁴

5.2.0.1 Methodology

Participant details were taken from the Indico conference system registration pages [169]. The departure location for each participant was assumed to be the city of their affiliation, save for cases where it was clear that the participant was based in Geneva, as is often the case for members of lhc collaborations. Direct travel to and from the conference was assumed. Distances were calculated as the great-circle distance using coordinates obtained with Nominatim from the OpenStreet Map data base. Rail, car or bus travel was assumed for all journeys with distances of less than 400 km, with air travel assumed for longer distances. ‘Short-haul’ was defined as travel distances of less than 1,500 km; distances up to 8,000 km are ‘long-haul’; and longer distance still were classified as ‘super long-haul’.

	AGU Fall Meeting 2019	ICHEP Melbourne 2012	ICHEP Valencia 2014	ICHEP Chicago 2016	ICHEP Seoul 2018	ICHEP Prague 2020 (virtual)
2-7 Number of participants	24,009	764	966	1,120	1,178	2,877
GHG emissions per participant [kg \(\mathrm{CO}_2\mathrm{e}\)]	2,883	8,432	1,902	2,699	2,648	0

Table 14 shows the average GHG emissions per participant for the ICHEP editions alongside those for the 2019 AGU Fall Meeting for reference. With the exception of the 2012 Melbourne edition of ICHEP, the per-capita emissions were significantly lower for ICHEP, which is a “travelling” conference, as compared with the stationary AGU Meeting, which always takes place in San Francisco. This indicates that moving a conference series between continents naturally reduces the travel-related emissions as participants tend to wait for the conference to be held near them to make the trip. Comparing the geographical distribution of home institutes for each conference reinforces this conclusion. Note that ICHEP Melbourne (2012) was the first and only ICHEP conference taking place in Oceania.

The emissions for two typical ICHEP conferences, one in Europe (Valencia) and one in Asia (Seoul) are displayed as a function of travel distance in Figure 15. A large fraction of attendees at the Seoul conference had to fly super long-haul, giving rise to the majority of the emissions. Emissions for the remaining half of the attendees was nearly negligible. This was not the case for Valencia, where as many attendees travelled short haul or less. It is also clear that the bulk of the emissions is due to long-haul or super long-haul air travel.

Reference [147] investigated possible optimisations of the conference location for the given participant distribution in order to reduce emissions.²⁵ Note that this is a slightly artificial construction because of the basin of attraction phenomenon discussed above, where participant distribution is self-selecting, based on the conference location. Unlike the AGU example, where moving the conference location to the middle of the country, rather than on a coast, significantly decreased the travel-related emissions, we found that the ICHEP locations were already pre-optimised, and further optimisation yielded at most a 10.2% reduction of GHG emissions. (The real outlier again was Melbourne, where the majority of participants had to fly super long-haul, and for which a 70.7% reduction would be achievable given the same participants by changing the location). If the location was optimised using participants from all 5 ICHEPs, the optimal location would be close to Amsterdam.

Further emissions reductions are only possible with a hub-based conference, and mandatory virtual participation above a certain distance from the hubs. Reference [147] trialled hubs in Chicago, Seoul and Paris, with virtual attendance for all participants with origins greater than 2,000 km from the hubs. Having found that Chicago, Seoul and Paris were not far from the optimal locations for the respective ICHEP conferences, we did the same, for the total ICHEP participation over the 5 conferences. Simply using a 3-hub model can reduce the carbon footprint of the conference to around 15–35% of a traditional one. Adding compulsory virtual participation for more distant participants reduces the carbon footprint further by 5–15% of a traditional conference with 10–25% of the participants attending virtually. As a test case, and without any prior optimization, we chose Rio de Janeiro, Johannesburg, and Kolkata as alternative hubs. This, however, increased virtual participation to 95%¸ mainly due to the strong European participation in hep and the remoteness of Johannesburg from Western Europe. Switching Paris for Johannesburg reduced the footprint to about 10% of the nominal one, with 40% of participants attending virtually. While the virtual fraction is still relatively high, it might be acceptable in a bid to include more remote HEP communities (like Melbourne), while keeping the emissions low.

Finally, one might expect a fully virtual conference to be more inclusive than in-person ones, especially for underserved participants, such as those with care-giving responsibilities, limited travel funding, or visa problems. We studied this by classifying participants by the human development index (HDI) [170] of their country of affiliation, and dividing them into four categories (low, medium, high and very high HDI).²⁶ The share of participants in these categories for each of the ICHEP conferences is shown in Figure 16. Indeed, in addition to enjoying the largest number of participants (by a factor of 2), the virtual ICHEP in Prague had the largest proportion of participants from countries with high or medium human development index, although it was not clear how much of this increase was due to its virtual nature, as opposed to a steady increase in physics participation from high and medium-HDI countries. There was virtually no participation from low HDI countries in any of the ICHEP conferences studied.

Case Study : Cosmology from Home
This case study is, in part, adapted from the Cosmology from Home website [173].
Cosmology from Home (CFH) [173] is an “online by design” conference series that has been run on a yearly basis since 2020. It exploits the advantages of digital communication to accomplish things that have no analogue in traditional conference formats, while staying true to the dynamic and social nature of traditional conferences.

CfH is spread over two weeks, with two days per week dedicated to plenary and parallel talks. Talks are pre-recorded and shared with the conference participants ahead of the start of the conference. Other than a strict time limit, CfH places few constraints on the format of the talks. Participants are expected to watch the talks before the scheduled online discussion events. The scheduled discussions last a maximum of three hours per day to mitigate online fatigue, but multiple sessions can be scheduled to accommodate different time zones.

Other days are reserved for themed discussions, which are proposed by the participants. Topics have included content related directly to cosmology, to scientific research more generally (i.e., technical or computational aspects), and social aspects, such as inclusivity and outreach in science.

The live discussions are hosted in virtual conference venue spaces to create a social atmosphere in which participants can wander around and join conversations. Suitable conference spaces are, e.g., Sococo [174], Welo [175] and Gather [176]. In case the number of participants is such that the conference spaces are too small to host everyone at once, break-out rooms of a suitable video-conferencing platform can be used. In the specific case of CfH so far, this has been Zoom [177].

CfH is coordinated asynchronously through an online message board, and a large part of the conference is dedicated to asynchronous text-based discussions. These discussions can start before the official opening of the conference and continue long after the conference has finished. In previous years, CfH used Slack [178]. (Alternatives include MatterMost [179], Zulip [180], Microsoft Teams [181] or Discord [182].) Discussion channels are grouped according to the scheduled live-talk discussion sessions (see below). Participants can discuss points in dedicated threads in these channels. Once a given discussion has grown sufficiently and has branched out into different sub-discussions, it can be given its own dedicated channel.

In addition to scheduled talk discussions, CfH implements scheduled themed discussions. These are moderated, workshop-style discussions on cosmological themes that are of interest to the conference participants. The topics are suggested, voted on and chosen via the Slack workspace. People can connect to the session via the conference space, but the main session is hosted in Zoom break-out rooms.

The final live discussion format featured by CfH is composed of spontaneous and unscheduled²⁷ “informal” live discussions. The time allocated to the scheduled live discussions is limited. The informal, breakaway sessions allow the participants to engage in more detailed explorations of the topics brought up by the live and asynchronous discussions.

The final live component of CfH consists of social events and interaction formats, such as social games and casual get-togethers, which complement the scientific discussion sessions on all days of the conference. These activities aim to reproduce (at least partially) the evening interactions of in-person conference social events.

One of the main advantages of the CfH format is that, compared to an in-person conference, a much longer time can be spent debating the talks of the participants. This allows for a much greater dissemination and understanding of the research that is presented. Additionally, it removes the need for travel. This keeps the participation costs low and offers the potential for the conference to be carbon neutral or carbon negative (achieved, at present, through carbon offsetting of residual emissions). The online format also makes the conference accessible to researchers from all over the world, subject to the availability of a suitable device and a stable internet connection by which to connect. The geographical distribution of participants in the previous three CfH conferences are shown in Figure 17.

Additionally, participants can watch the talks according to their own schedules and personal obligations. Participants are able to pause and restart the talks, take time to digest them and to look up background material. They are also able to prepare and raise points for discussion in any of the conference environments. The asynchronous discussions can be tailored exactly to the schedule of the conference participants.

In the first CfH (2020), participants needed time to adjust to the format. This was to be expected, and the organisers actively encouraged participants to partake of the various aspects of the conference and actively modelled the expected social norms. The activity and enthusiasm of participation increased year on year. Speakers created innovative and accessible talk records, and participants regularly referred to these in the live and text-based discussions. Participants organised watch parties and impromptu discussions, made use of the various breakout rooms, and gathered in the virtual environments. The themed discussions proved to be particularly popular, and parallel sessions were often necessary to accommodate the high number of topic suggestions.

All participant feedback has been constructive and positive, and CfH has been well attended. The number of CfH participants were 255, 427 and 275 in 2020, 2021 and 2022, respectively [173]. The format is easily tailored to various topics: a conference featuring this format in HEP is expected to be run in late 2023.

6 Research Infrastructure and Technology

hecap research areas rely on big science infrastructure. These particle accelerators, large-scale collider experiments, observatories and associated buildings infrastructure have a lifetime environmental impact from cradle to grave. This is recognised in Section 8, “Sustainability considerations”, in the Accelerator R&D Roadmap of the European Strategy for Particle Physics [183]. It divides the topic into three aspects:

• Energy efficient technologies,

• Energy efficient accelerator concepts, and

• General sustainability aspects.

The first two focus on the biggest impact of accelerators: the energy consumption during their operation. One aspect focuses on the current technology and its energy efficiency, the other on the development of new accelerator concepts with smaller energy requirements. These topics are discussed in depth in other sections of the Strategy. The third aspect is more broadly defined and considers sustainability beyond energy [183]:

“A carbon footprint analysis in the design phase of a new facility can help to optimise energy consumption for construction and operation. For cooling purposes accelerator facilities typically have significant water consumption. Cooling systems can be optimised to minimise the impact on the environment. For the construction of a facility environment-friendly materials should be identified and used preferably. The mining of certain materials, in particular rare earths, takes place in some countries under precarious conditions. It is desirable to introduce and comply with certification of the sources of such materials for industrial applications, including the construction of accelerators. A thoughtful life cycle management of components will minimise waste.”

In the case of astronomy research, Ref. [184] argues that emissions due to research infrastructure dominate the carbon footprint of an astronomer.

A number of initiatives have already been formed to consider the manifold technical challenges of improving the environmental sustainability of research infrastructure and associated technologies. Three examples are listed below, (for others see Section 1.2 and references therein):

• The International Committee for Future Accelerators (ICFA) has the specific panel “Sustainable Accelerators and Colliders” [185].

• Every 2 years since 2011, the Energy for Sustainable Science at Research Infrastructures (ESSRI) workshop [186] takes place.

• Innovation Fostering in Accelerator Science and Technology (ifast) [187] is an EU-project in which the “WP 11 – Sustainable concepts and technologies” is aimed to increase sustainability. The current participating institutes are cern, desy in Germany, the European Spallation Source (ESS) in Sweden, the GSI in Germany, the Paul Scherrer Institut (PSI) in Switzerland, and the Science and Technology Facilities Council (STFC) in the United Kingdom.

Environmental sustainability is also being considered by individual experiments, and Case Study [case:LHCb] provides a summary of efforts by the LHCb experimental collaboration to assess and mitigate the environmental impact of both the experiment and work practices, more generally.

In this section, we consider the following aspects of environmental sustainability in research infrastructure: life cycle assessment (LCA), (carbon) accounting, and technological developments, particularly in the context of accelerator technologies and detector gases. While the discussions of technological developments provide concrete examples, the primary focus of this section is the need for critical life cycle analysis for all research infrastructure projects to assess and limit their environmental impacts. The impacts of mining and processing of materials is also considered in Section 7.1, wherein complementary aspects of the LCA are also discussed.

Recommendations – Research Infrastructure and Technology

	Individual actions: Seek out new innovations and best practice. Rethink how the impact of frequently-used equipment can be reduced, and reduce "over-design" by reassessing safety factors and other margins to reduce resource consumption. Read section on resources and waste (Section 7).
	Further group actions: Ensure that environmental sustainability is an essential consideration at all stages of projects, from initial proposal, design, review and approval, to assembly, commissioning, operation, maintenance, decommissioning and removal, using life cycle assessment and related tools. Engage with industrial partners who exemplify best practice and sustainable approaches. Appoint a dedicated sustainability officer to oversee project development, and institute regular meetings with a focus on environmental sustainability.
	Further institutional actions: Critically assess the environmental impact of materials, construction and the operational life cycle as an integral part of the design phase for all new infrastructure. Provide training opportunities, required tools and technical support to assess and improve the environmental sustainability of project life cycles. Recognise and reward innovations that minimise negative environmental impacts, regardless of revenue. Promote knowledge exchange on sustainability initiatives between groups and institutions, including decision-makers, designers and operators of projects, setups and infrastructure.

6.1 Accounting and Reporting

The methodology of a life cycle assessment can be used to analyse the environmental impact of resources used to build, run and decommission an accelerator, observatory or experiment, see Section [sec:sustainablesourcing] for further details. Such assessments have already been undertaken by a number of facilities, including:

• The European Southern Observatory (ESO) [188].

• The Giant Radio Array for Neutrino Detection (GRAND) Project, a multi-decade astrophysics experiment [189] — This led to a full issue of the Nature Astronomy Journal on climate change [190].

• The Relativistic Ultrafast Electron Diffraction and Imaging (RUEDI) facility at STFC Daresbury Laboratory  [191].

• The Compact Linear Collider (clic) is planning to conduct an assessment  [192].

• The ISIS-II project, the next generation of the ISIS neutron and muon source, is planning to conduct life cycle analyses for the project and various design options  [193].

There is currently limited availability of data on estimated emissions and resources consumption for basic research infrastructure, and, where it is available, its presentation is not standardised. This makes overall assessments of sustainability and comparisons of individual technologies challenging. Implementation of effective life cycle assessment across the HECAP+community could provide the impetus for standardised reporting that will provide the data needed for ongoing assessment of current and future technologies and research infrastructure projects, such as any future collider concept (see Case Study [case:FutureColliders]).

See Best Practice [bp:SiWafer] for a summary of life cycle assessment for a silicon wafer used in particle detectors, summarised from the ProBas library for life cycle assessment [194].

The Labos1point5 working group has proposed a standardised carbon accounting procedure and associated assessment tool for research laboratories. This programme is described in detail in Best Practice [BP:l1p5].

6.2 Technological Improvements

There are ongoing efforts across the HECAP+ community to reduce the environmental impacts of the technologies implied in research facilities. Two examples in the case of accelerator technologies are provided in Best Practice [BP:EnergyRecoveryAccelerator] and Case Study [case:wakefield], namely energy-recovery accelerators and plasma wakefield acceleration technologies. Further details of efforts by the LHCb collaboration are included in Case Study [case:LHCb]. A detailed discussion of the impact of detector gases is provided in the following subsection.

6.2.1 Gases

Significant quantities of GHGs are used for particle detection technologies and cooling systems across HECAP+. As such, they are used as a resource, but can escape the detector volume into the atmosphere and turn into potentially dangerous waste gases. Table 2 lists a number of GHGs, their chemical formulae, atmospheric lifetimes and global warming potentials (gwp). All of these gases are used either for cooling purpose or as active ingredients in gaseous detector systems, where they are often added to noble gases to improve detector properties, such as drift charge velocities or diffusion coefficients [211].

Whilst gases are used in a variety of detectors, such as time projection chambers, ring-imaging Cherenkov detectors or multi-wire proportional chambers, the main offenders in ecological terms are the detectors used in the muon systems of the LHC experiments, specifically resistive plate chambers (RPCs). This is due to their large areas of around 7,000 m\(^2\) in total for both the CMS and ATLAS muon systems, and the gas mixtures used to cope with the large event rates at the LHC, which often feature HFC-134a as a main component.

As shown in Figure 4, Scope 1 direct emissions made up about 25% of CERN’s carbon footprint in 2019. During the LHC Run 2 in 2018, it was about a factor of two larger. 92% of these Scope 1 emissions are related to the activities of the large LHC experiments [1], [212]. CERN and the LHC experiments are actively working on reducing this impact by continuously repairing gas leaks that are one of the main reasons for the large amount of waste gas. In addition, CERN has tested an HFC-134a recuperation plant showing an efficiency of close to 85%, which is to be installed in the detectors to reduce the environmental impact [212], and is actively researching alternative gas mixtures [213]. Future detector projects still plan to use RPCs (DUNE covering an area of about 860 m\(^2\) [214] and SHIP with about 100 m\(^2\) [215]) but are testing their prototypes also with alternative gas mixtures [216].

The gases responsible for about 80% of CERN’s Scope 1 direct annual GHG emissions are perfluorocarbons (PFC), hydrofluorocarbons (HFC) and sulphur hexafluoride (SF\(_6\)) in particle detection, and HFCs and PFCs for detector cooling. To put the emissions into context, CERN’s PFC emissions are roughly of the same size as the Swiss emissions [219] and only reduce by about 30% when there is no LHC run. For 2017 and 2018 (during LHC data taking), CERN’s SF\(_6\) emissions are about 5% of Switzerland’s, and of the same size as those of Luxembourg or Latvia [220]. The HFC emissions are 6% of the Swiss emissions, about twice the size of Luxembourg’s and a bit less than half of Latvia’s emissions, again looking at 2017–2018 data [221]. During 2020–2021, when the LHC, and more importantly its experiments, shutdown for upgrades and maintenance, SF\(_6\) emissions were down to about a third, while HFC emissions were down to 25%. All of these so-called F-gases have EU supply restrictions imposed on them since 2015 [218], effectively phasing out their usage. A 2022 regulation proposal aims to extend EU regulations, e.g., to reduce HFC usage down to 2.4% by 2048 compared to 2015 levels [222]. An additional significant factor in this proposal is the removal of certain sector exemptions for HFC usage, such as research. All of this could significantly impact the availability and cost of F-gases in the future and therefore affect the HECAP+ community, which should be reflected in the plans for ongoing and future experiments. Even independent of these EU regulations, the HECAP+ community should work with highest priority on abolishing problematic GHGs in existing and future detectors. Ideally, this would consist of replacing them with non-GHGs, or gases with low GWP, or gasless detector technology.

For cooling of the LHC-experiments, concrete plans are in place to upgrade the future detectors (Phase-II Upgrades) to CO\(_2\) cooling, reducing the total Scope-1 carbon footprint of the experiments during HL-LHC exploitation significantly [223], [224].

Solutions to the problem of gas use for particle detection are less straightforward. In some scenarios, low-GWP gases used as replacements for industrial applications are not suitable for detector applications and, as such, studies for alternative gases are currently ongoing. The difficulty in finding replacement gases originates from having to satisfy several factors: safety (non-flammable and low toxicity) and environmental impact (minimising GWP), while maintaining their detector performance (including preventing the ageing of the detectors, ensuring good quenching³¹ and being radiation-hard) [27].

Current gas mixture alternatives for particle detection are centred around tetrafluoropropene (chemical formula C\(_3\)H\(_2\)F\(_4\) and industrially referred to as HFO-1234ze/R-1234ze). The mixture has zero ozone-depletion potential and a global warming potential below 1, over a span of 100 years. Simulations of RPCs that operate with various gas mixtures, which include R-1234ze, have shown encouraging results, although further studies are still required [225], [226].

Long term, it is crucial to design future detector systems with gas GWP in mind. Consequently, it is essential that current state-of-the-art and future detectors are compatible with this and, if not, R&D is aimed at reducing the GHG emissions of such systems.

7 Resources and Waste

Half of the world’s ghg emissions, and over 90% of global water stress and biodiversity loss events, are due to the extraction and processing of raw materials [238]. Although most extracted materials are slated for the energy or agriculture sectors, the small fraction associated with consumption of goods and services is responsible for 18% of EU emissions [238]. Mitigating the climate impacts of the extraction, processing and trade of raw materials is a priority for the resilience of the EU [238], and it should also be a priority for the world climate agenda.

The generation of waste is a direct consequence of material consumption, and is aggravated by constraints in production, distribution, usage and repair, and disposal or recycling of consumables. Waste has severe impacts on life on land and at sea, often destabilizing local ecosystems. It also damages the global ecosystem by contributing to climate change. Accumulations and inefficient disposal of waste products can result in pollution of ground water and air, thus directly affecting the health of individuals and communities at a large cost to society in terms of disease burden and lives lost.

In an attempt to curb the footprint of waste generation, the concept of a circular economy has been proposed [239]. Any such proposal must be established in parallel with a will to reduce waste at source through sustainable procurement, repair and reuse, and used only as a transitional measure. Even a fully circular economy has some dissipation, and signatures of this energy waste need to be independently addressed and reduced [240]–[242].

Procurement accounts for almost two-thirds of annual emissions at cern [243], with a GHG footprint of the same order as its direct emissions in 2018, when the LHC was running [1]. Although not yet fully included in reporting by other HECAP+ institutions, the environmental cost of procurement is likely proportionately large elsewhere. Maximising the sustainability of the use cycle of resources should be a priority of the hecap community.

This section covers sustainable sourcing in Section 7.1, and reduction and treatment of waste, including E-waste, in Section 7.2. The use of materials in research infrastructure is also discussed in Section 6.

Recommendations – Resources and Waste

	Individual actions: Limit purchases and consider environmental credentials such as repairability and recyclability of products in purchasing decisions. Service appliances regularly; share, repair, reuse and refurbish to minimise waste; sort and recycle. Read the sections on computing (Section 2), energy (Section 3), food (Section 4), and research infrastructure and technology (Section 6).
	Further group actions: Adopt life cycle assessments and associated tools to assess environmental impact of all activities. Institute sustainable purchasing, usage and end-of-life policies in the management of group consumables, office supplies and single-use plastics e.g., in conference events (see also Section 7.2.3 and Best Practice [bp:PlasticFreeConf]).
	Further institutional actions: Prioritise suppliers instituting sustainable sourcing and operating policies, with a particular focus on the raw materials processing stage (see Best Practice [bp:SustainableRawMaterials]) and with the aim of creating demand for recycled (secondary) raw materials. Provide an institutional pool of infrequently-used equipment to avoid redundancy in purchasing. Proceduralise and prioritise repair of equipment, and enable through provision of tools and know-how. Assess waste generation and management for the design, operation and decommissioning of IT and infrastructure projects by right-sizing needs, establishing specific treatment channels for all waste categories, and setting recycling targets that include the recycling of all construction waste, see, e.g., Best Practice [bp:HERAShieldingBlocks].

7.1 Resources

HECAP+ research can be resource-intensive, particularly in the building and maintenance of the often large experiments that drive progress in our fields. These resources have an environmental impact over their entire life cycle, due to extraction of the raw materials used in their manufacture, their production and use, and their disposal once they become unusable or obsolete. Of these, the raw materials processing stage has been highlighted as having the greatest potential for emissions reduction (see, e.g., Figure 2 of Ref. [238]).

The extraction of raw materials has important and extensive environmental costs [244], mostly associated with the mining industry. Acid mine drainage is the overriding problem and is a serious threat to water resources. It results from water flow over ore creating sulphuric acid and leaching heavy metals from surrounding rock, thus contaminating groundwater and soil. Mining operations can also deplete water resources, particularly in regions of limited water supply, severely restricting the availability of water to local consumers. Fine particles and dust produced during mining operations and dispersed by winds affect air quality, and mining and its infrastructure leads to loss of agricultural land and even entire ecosystems through contamination or destruction of soil cover. Mining is the world’s largest producer of waste, with copper, zinc, bauxite and nickel mining generating the largest ratios of waste to mined metal. Disposal or storage of tailings, the waste products remaining after the extraction of valuable material from ore, is a major problem. These can be radioactive, and are sometimes illegally disposed of directly into rivers or seas. Even when stored ‘responsibly’ in tailings dams, incorrect geological siting of these dams, in tectonically active zones or regions of high rainfall, can lead to catastrophic loss of life and usable land [245].

Environmental sustainability aside, mining has a poor safety and human rights record [246], and is sometimes subject to dubious financing [247]. Mining of ‘conflict minerals’, such as tin, tungsten, tantalum and gold, used in mobile phones and other everyday products, are sometimes used to finance armed conflict [248].

Sustainability regulations, both externally imposed and voluntary, are slowly being incorporated into the raw materials supply chains (see,e.g., the Voluntary Principles on Security and Human Rights [249]), albeit slowly, and in an inconsistent and sometimes superficial manner [246], [250]. For examples of sustainability initiatives, in particular in relation to raw materials supply chains, see Best Practice [bp:SustainableRawMaterials].

An analysis of components used inside a smartphone and their impacts can be found in Ref. [251]. Smartphone manufacturer Fairphone, for instance, sustainably sourced 56% of 8 of the materials used in its phones in 2020, and have a set a target of fair sourcing of 70% of 14 materials by 2023 [252].

A ranked list of mined metals by overall environmental impact can be found in Table [tab:MetalImpact]. This table was taken from the EU Raw Materials Information System [253], with source data from Ref. [254]. Sourcing recycled metal from scrap produces significantly lower emissions. Secondary aluminium, for example, was reported by the European Aluminium Association to emit 95% fewer GHGs than primary production [238]. Other materials used in HECAP+ experiments (e.g., cobalt for magnets, rare earths for permanent magnets, niobium) are produced under very difficult conditions, with a high environmental or societal cost [255]–[257]. Formal discussions of their use and impact have already begun in the HECAP+ community, most recently at a workshop on Rare Earth Elements organised by ifast at desy [258].

7.1.1 Life cycle assessment

Best practices in sustainable use and disposal of resources begins with a life cycle assessment (LCA): a cradle-to-grave accounting of all the environmental impacts of a resource. As an example the ISO 14040  [259] and ISO 14044 [260] standards provide a systematic procedure for the analysis. Depending on the goal and scope of the analysis, the life cycle inventory comprises the quantification of all input and output flows. This includes raw materials, consumables, energy, products, waste, emissions, and groundwater and soil contamination. There are online tools and auditing agencies who provide help with the analysis, see, e.g., Ref. [194]. For an LCA of a silicon wafer used in particle detectors, see Best Practice [bp:SiWafer].

7.1.2 Sustainable sourcing

Purchasing policy can have a major impact on the environmental costs of procurement. The HECAP+ community should prioritise suppliers that implement sustainable thinking, sourcing, and operation. This could include voluntary provision of life cycle assessments for their products (see above), or certification of, e.g., proof of origin. Sustainability requirements on suppliers could also be incorporated into tenders and purchasing regulations, allowing these considerations to be weighed in tandem with cost in the tendering process. Since much of HECAP+ funding is public, purchasing regulations, which are influenced by funding agencies, an additional important stakeholder in this process, must be reassessed. For examples of best practice in sustainable procurement, see Best Practice [bp:SustainableRawMaterials]. A strategic approach to sustainable purchasing has been outlined in ISO 20400 [261].

CERN is in the process of defining a new environmentally responsible procurement policy, to be implemented in 2023 [15]. Key measures being considered include requiring sustainability certification from suppliers, with a focus on those with highest potential to drive sustainability issues [15]. For further sustainable procurement and waste policies being explored by CERN, see Best Practice [bp:CERNSustainableProcurement].

7.2 Waste

Around 3% of global GHG emissions is due to solid waste disposal, the organic component of which decomposes in wastewater and landfills, producing methane and nitrous oxide [13]. This is despite a 60% decrease in the amount of waste landfilled in the EU in the past two decades, due partly to an increased legislative focus on alternative treatment methods, such as recycling and composting, as well as more widespread landfill gas recovery [262].

The fastest-growing portion of EU waste output is E-waste [263]: powered products with electrical components that are discarded into the waste stream. In addition to releasing hazardous chemicals into the environment, improperly treated E-waste contributes to global warming through failure to recuperate valuable mined materials, and direct release of GHGs including refrigerants (see Figure 22 for statistics on global E-waste generation and disposal in 2019). Moreover, E-waste can often be shipped illegally to developing countries without infrastructure for safe recycling [264], [265]. Exploding demand for digital devices, fuelled by their fast obsolescence and difficulty to repair, has also given rise to a boom in the mining industry. This boosts the economy in resource-rich countries, where working conditions are usually unsafe and unpleasant, and causes deforestation and pollution [265].

Increasing the life-span of consumer electronics through more comprehensive right-to-repair legislation is an integral part of the EU’s strategy for a circular economy [263], [266]. According to a 2018 European Commission study, however, the cost of repair was the most frequent reason consumers chose to replace four common household electrical and electronic goods [267]. Right-to-repair advocates are pushing for more comprehensive and far-reaching legislation, including policies that will empower consumers to make simple repairs themselves, as well as government incentives to repair through, e.g., financial aid [268]. A trial scheme for the latter has already been implemented in France [269]. Unfortunately, today’s legislation is formulated piecemeal and is slow to take effect. In addition to instituting sustainable purchasing, use and de-inventorising/disposal policies for electrical and electronic goods (see Recommendations), HECAP+ institutions could encourage the sustainable use of personal electronic devices by making repair equipment freely available, with guidance from experts on a volunteer basis.

7.2.1 Plastic Waste

Plastic is a versatile product; efficient, cheap, stable, and infinitely mouldable. Its unique properties have resulted in our increasing reliance on it over decades, making it very hard to live without.

GHG emissions from production, use and disposal of conventional (fossil-fuel-based) plastics is significant, and growing. It is expected to increase to \(\sim\)2.1 tco2e per capita by 2040, accounting for almost 20% of the global carbon budget [270]. Less than one tenth of plastic waste produced to date has been recycled, with the remainder being landfilled or incinerated [270]. This plastic waste contaminates the natural environment. It is slow to degrade (biodegradable plastics included), releasing potentially harmful chemicals into the environment in the process, eventually breaking down to micro- and nano-plastic particles that infiltrate the food chain. Microplastics are also found in waste water, due to washing of synthetic textiles, and in cosmetic and personal care products [271]. The health impacts of ingested plastic waste are not yet completely understood, but they are thought to alter metabolic pathways and hormone signalling in animals.

A simple ban on single-use plastics, and plastics that are not at least 99% recyclable would greatly limit microplastic pollution.

7.2.2 Conference Waste

The amount of waste generated at conferences can be significantly reduced by replacing printed timetables and welcome packs with a well-designed conference app, see, e.g., Whova [272], and keeping conference gifts digital (e.g., e-vouchers/discounts for local restaurants or activities). Sustainable stationery should be distributed on a need-only basis, and banners, posters and name tags made plastic-free and reusable. For waste-minimising initiatives implemented at the plastic-free 2019 conference of the Australian Marine Society, see Best Practice [bp:PlasticFreeConf]. For sustainability concerns in conference catering, see Section 7.2.3.

7.2.3 Catering Tableware

A life cycle analysis by the UN Environment Programme concludes that “reusable tableware consistently outperforms single-use tableware in all the studies and across all environmental impact categories (with water use being the exception, because of washing). This type of analysis takes into account all the variables that affect the environmental impact of a product, from manufacturing to end-of-life treatment. The case for reusable tableware is strengthened in countries where renewable energy makes up a high proportion of the grid mix and where end-of-life treatment options are not well developed” [274].

In outdoor or remote environments or ‘pop-up’ events with no fixed catering facilities, where reusable tableware is impractical, single-use biodegradable tableware is preferable to other single-use tableware if it is industrially composted mixed in with food waste [274].³³

Unlike emissions due to reusables, which are dominated by their use phase due to repeated washing, the main impact of biodegradable tableware is due to its production. For conventional plastic, a significant role is also played by end-of-life management. Quantitative analysis of their relative emissions is thus strongly dependent on assumptions about manufacture and disposal, including the material demand. On the practical side, to minimise this impact when planning conference catering one should always choose the lightest-weight disposable tableware fit for purpose, preferably manufactured in a country with a significant proportion of renewables in its energy and electricity mix. For a comparison of emissions due to different choices of disposable catering tableware for pop-up catering, see Case Study [cs:conferencetableware].

8 Outlook

The central message of this document is clear:

Assessing, reporting on, defining targets for, and undertaking coordinated efforts to limit our negative impacts on the world’s climate and ecosystems must become an integral part of how we plan and undertake all aspects of our research.

Achieving this relies on individual-, group- and institution-level actions, and concrete suggestions are made in this document.

This is a call to reflect with humility on these suggestions, and this is a call to action. This is a call to consider the opportunities that reassessing the environmental sustainability of our work practices also offers for addressing systemic barriers to inclusivity and accessibility.

But it must be understood that this and similar documents are only the beginning: The hecap community must come together to secure a sustainable future for our fields.

Thank you again for taking the time to read this document.

Acknowledgments

This initiative was conceived at the workshop “Sustainable HEP”, which was organised by Niklas Beisert, Valerie Domcke, Astrid Eichhorn and Kai Schmitz, and hosted by CERN and held online 28 to 30 June 2021.

The authors acknowledge input received during: the Sustainable HEP Mini-Workshop, 18 January 2022; the Sustainability & Inclusion Panel Discussion at the 21st String Phenomenology Conference, hosted by the University of Liverpool, 4 to 8 July 2022; the Education and Outreach track of the International Conference on High Energy Physics 2022 (ICHEP2022), Bologna, 6 to 13 July 2022; the second Sustainable HEP workshop, 5 to 7 September 2022.

The authors thank Shehu Shuaibu Abdulassalam, James Alvey, Nicolas Arnaud, Gabriela Barenboim, Till Bargheer, Niklas Beisert, Micah Buuck, Enrico Cennini, Nikola Crnković, Valerie Domcke, Astrid Eichhorn, Daniel Errandonea, Stefan Fredenhagen, Elina Fuchs, Tetyana Galatyuk, Spencer Gessner, Clare Gratrex, Roberto Guida, Aaron Held, Tomas Kasemets, Stavros Katsanevas, Yves Kemp, Jinsu Kim, Ben Krikler, Valerie Lang, Paul Laycock, Roberto Lineros, Benno List, Jenny List, Daniel Maitre, Sudhir Malik, Beatrice Mandelli, Zachary Marshall, Pablo Martínez-Miravé, Chris Parkes, Michael Peskin, Ruth Pöttgen, Melissa Quinnan, Salvatore Rappoccio, Rachel Rosten, Jan Rybizki, Filippo Sala, Wayne Salter, Quentin Salvi, Sabrina Schadegg, Kai Schmitz, Kathrin Schulz, Jason St John, Lindsay Stringer, Denise Völker, Rodney Walker, Tien-Tien Yu, and Sebastian Zell for invaluable input.

The authors thank Jessie Muir for producing the individual, group and institution icons for the recommendations.

9 Supplementary Data for Figure 4

Tables [tab:ComparativeEmissionsData] and [tab:ComparativeEmissionsDenominator] contain the raw data that was used to produce Figure 4. Each set of data was taken from a publicly-available environmental report issued by (members of) the institution in question; the original documents are referenced below.

Our approach differs from existing estimates of the ghg footprint per researcher in the divisor used to compute this quantity. We shared the emissions per resource equally by the total number using that resource, whether it be total number of employees, or research staff, or in the case of large laboratories like cern, the number of Users, rather than using the same divisor throughout. For instance, while we divide the commuting emissions for each institute by the total number of employees, we assign the business travel emissions solely to the research staff, assuming the support staff have negligible long-distance travel. For concreteness, we have colour-coded the per-researcher estimates in Table [tab:ComparativeEmissionsData] by the denominators used in their computation, with the colour key provided in Table [tab:ComparativeEmissionsDenominator].

Authors

Endorsers

Individual endorsements can be made at https://sustainable-hecap-plus.github.io/.

To make an institutional endorsement, please email sustainable-hecap-plus@proton.me.

An up-to-date list of endorsers can be found at https://sustainable-hecap-plus.github.io/.