Guanwen Zhang

UAL has 161 staff members including librarians and support staff. As the second largest research library in Canada, it offers a collection of over 7 million titles and 10 million volumes in both print and digital formats. UAL’s digital collections are vast and diverse, including Education and Research Archive (ERA), research data repositories (Dataverse), Audio and Video collection (Aviary), and digitized books, newspapers, maps and other special collections.

Before July 2025, UAL had an on-premise cloud acting as the backbone of its digital preservation practices. That cloud consists of 6 physical servers residing at three separate data centers. After July 2025, the private cloud had been decommissioned. UAL embraces a number of ways of accomplishing its digital preservation goals, which include Internet Archive (IA), community-owned Ontario Library Research Cloud (OLRC), Aviary Platform (commercial SaaS), and LOCKSS and CLOCKSS.

Looking at just their web archiving activities as listed in the table below:

They needed to calculate the following:

a) Total storage (D) in TB, in this case D=42.1TB

b) Power consumption per TB, this depends on the storage types: disks, and tapes etc. We take the average number P=5W/TB

c) Data center efficiency (PUE): This varies from region to region. For North America, we take PUE=1.5. (https://journal.uptimeinstitute.com/datacenter-energy-efficiency-by-region/, 2020)

d) Define the duration of data activity (T), usually, we use one year as a time unit for carbon footprint estimation. T=365 days * 24 hour/day

e) Carbon Intensity: For example, the carbon intensity of Canada is 174gCO₂/kWh, and North America is 374 gCO₂/kWh (https://ourworldindata.org/grapher/carbon-intensity-electricity, 2024). We assume the Carbon Intensity (CI)=400gCO₂/kWh=0.4kgCO₂/kWh

The estimate of carbon footprint for a single year is thus:

Estimated Carbon Emission = D*P*PUE*T*IC=42.3 * 5 * 1.5 * (365*24) * 0.4 =1112 kg CO₂/year

Mikko Tiainen

CSC - IT Center for Science Ltd. is a company entrusted with special state assignment and owned by the state of Finland (70% of the shares) and Finnish higher education institutions (30% of the shares). Among other things, it provides a national digital preservation service for all Finnish cultural heritage institutions and academic institutions (https://www.digitalpreservation.fi/). The Finnish Digital Preservation Service has 7.5PiB capacity with 3+1 preservation copies where +1 is LTO8 tape dark archive, 3 preservation copies are spinning disk, LTO9 and LTO8 tapes. Around 4PiB currently in use. Since 2022 the Finnish Digital Preservation Service Team has published its annual carbon footprint figures.

Carbon footprint mitigation tasks that have been carried out since 2022 include:

• Relocating data center so that PUE of the used data center is better (1.6 > 1.05)

• Change the electricity purchase agreement to a green one (Hydropower)

• Upgrading spinning disks into helium filled 16TB models which has lower power consumption (40% reduction) and lower in-body carbon footprint (50%) than earlier spinning disk models (Seagate change to green electricity)

The carbon footprint optimization of Finnish DP service is a yearly optimization target. They published an Excel tool to calculate their annual carbon emissions. The current carbon footprint for the Finnish Digital Preservation Service is 21 kgCO2ev / TiB/ year.

Some recent work has been published as a poster from iPRES 2024 - A Generic Carbon Footprint Model. This poster describes a generic model for calculating the footprint in cases where hardware manufacturers do not report the embodied carbon footprint figures for their hardware.

Catherine Jones

The EDC is a capability of the UK Energy Research Centre and is based in the Science and Technology Facility Council/UK Research and Innovation. It is a small service with three staff members and a collection of research data, grey literature, and information on relevant energy funded projects. For the project described below they estimated the carbon impact as 932 kg CO2e per year.

The EDC supports researchers who are exploring energy and has a remit to hold research outputs for the long-term, but would also like to minimise the environmental impact of their service.

In 2021 they ran a project to investigate the impact of their policies on the energy consumption of the service - the public report is available here: An investigation into the impact of service policies on the energy consumption of the Energy Data Centre

In this project they used DPC’s Rapid Assessment Model and The National Archives DiAGRAM tool to set preservation service aspirations and identified collection risks. They then used an Intelligent Platform Management Interface (IPMI) to measure energy use on physical servers and thus the impact of routine jobs run to implement policy decisions. As a result of this work they decided to adjust the frequency of some of the jobs and to review the resilience and redundancy of the service.

This was a short term project with a specific goal in mind. However it was both successful in identifying energy use and sharpening policy. It was very time-consuming to analyse the data and the IPMI tool only records energy consumption in steps which can lag actual consumption so it was difficult to accurately record usage. The use of the RAM and DiAGRAM tools enabled the EDC to focus on their goals and priorities. This exercise has made them more environmentally aware and is now a factor in planning and strategy decisions.

Alicia Wise

CLOCKSS is a distributed international archive with physical servers located at 12 institutions in 7 countries. The size of their digital preservation holdings is roughly 250tb and it is estimated that their base line carbon impact is approximately 109,541 kgCO2e/year from the archive and an additional c.11 kgCO2e/year from travel.

CLOCKSS followed the DIMPACT method as this is widely used by publishers and other media organisations. This is also the method that now underpins this toolkit.

It took CLOCKSS about 18 months to gather data at 4 of their 12 hosting sites and estimations could be calculated from here. It was a very useful exercise, but also a heavy lift with a steep learning curve. This experience inspired them to really want to contribute to the development of this toolkit! You can read more about their experiences in this blog post. 

In late 2024 the DPC established the Carbon Footprint Task Force. This task force was set up in response to an emerging need within the community for clear guidance on how to calculate the carbon footprint of digital preservation activities.

Alicia Wise (Executive Director, CLOCKSS) agreed to chair this task force and helped shape the charge. DPC members with expertise or experience in this area were invited to put themselves forward to join the task force. A small expert group was formed, which met for the first time in January 2025, and a schedule of monthly meetings was set up. Initial meetings focused on introductory presentations, agreeing the task and ensuring we were all on the same page. In April 2025 we began to develop a set of headings and topics for an online toolkit, and subsequent meetings focused on writing and discussing challenges and issues as we encountered them.

We agreed a number of guiding principles to shape our writing of this toolkit:

• The toolkit should contain clear, practical and actionable advice and information.

• It should demystify carbon footprint calculations for digital preservation and empower readers to take action to calculate the carbon footprint of their own digital preservation activities.

• The advice within the toolkit should be applicable regardless of:

  • digital preservation systems/infrastructure in place

  • current experience level with carbon footprint calculations

  • type/sector/size/geographic location of organization.

The resulting toolkit was a collaborative effort, in which we were able to benefit from the experiences of all group members. It was shared for community comment and testing in March 2026 and more formally launched at the iPRES conference in September 2026.

A big thank you to members of the Carbon Footprint Task Force for sharing their challenges, knowledge and experience on this topic, and their hard work and good humour throughout.

The Carbon Footprint Task Force met monthly over Zoom for a period of 12 months

Members of the Carbon Footprint Task Force:

• Alicia Wise (Chair), Executive Director, CLOCKSS [ORCiD 0000-0002-7898-3428] – read more about carbon footprint calculation work at CLOCKSS

• Sharon Bayer, Director, Systems Deployment and Integration, Harvard

• Catherine Jones, Energy Data Centre Lead, UKRI/STFC [ORCiD 0000-0002-5112-835X] – read about environmental impact assessment at the Energy Data Centre

• William Kilbride, Executive Director, Digital Preservation Coalition

• Jenny Mitcham, Chief Digital Preservation Officer, Digital Preservation Coalition [ORCiD 0000-0003-2884-542X]

• John McMillan, Chief Operating Officer, Digital Preservation Coalition

• Laura Peaurt, Archivist (Digital Preservation), University of Nottingham

• Caylin Smith, Head of Digital Preservation, Cambridge University Libraries and Archives [ORCiD: 0000-0001-6340-5708]

• Paul Stokes, Subject Matter Expert (Digital Preservation), Jisc [ORCiD 0000-0002-7333-4998]

• Danielle Taylor, Digital Preservation Librarian, Indiana University [ORCiD 0009-0005-2395-4768]

• Elisabeth Thurlow, Digital Preservation and Access Manager, University of the Arts London [ORCiD 0000-0002-6630-6235]

• Mikko Tiainen, Development Manager, CSC - IT Center for Science Ltd. - read about carbon footprint mitigation tasks at CSC – IT Center for Science

• Guanwen Zhang, Senior Digital Preservation Specialist, University of Alberta [ORCiD 0009-0000-9119-8232] – read about the carbon footprint of web archiving activities at the University of Alberta

• Tamara van Zwol, Preservation Coordinator, Dutch Digital Heritage Network [ORCiD 0009-0000-9119-8]

The toolkit was published in draft in March 2026 with an open invitation to the community to actively test and provide comment and feedback. As well as accepting comments via an online form and within a Google doc, a Carbon Calculation Support Group was formed, which met monthly for a period of six months and worked together on carbon footprint calculations. This group was able to highlight areas of the toolkit where the advice and guidance needed to be improved, whilst working on their own case studies. Many improvements have been made to the toolkit in response to feedback from this group and our thanks go to all participants. Find out more about the Carbon Calculation Support Group and explore the case studies.

Alicia Wise, CLOCKSS

Introduction

The CLOCKSS digital archive protects scholarship for future generations. We are a community-governed collaboration of global research libraries and academic publishers. We provide long-term digital preservation services for academic content published in any language or subject area.

CLOCKSS preserves content in 12 archive nodes located at leading research libraries around the world. We preserve a wide array of published material (e.g. journal articles and book titles plus associated data, images, software, videos, and web archives) and an increasing array of library digital collections (e.g. digitized special collections, e-theses and dissertations, institutional repository content).

Why we got involved in developing and testing the DPC’s Carbon Footprint Toolkit

Digital preservation depends on infrastructure running continuously for decades, so we wanted to understand what that actually means in environmental terms. Rather than looking only at CLOCKSS as an organisation, we took an ecosystem-level approach, building on work undertaken through the DIMPACT project and subsequently helping develop a shared approach through the Digital Preservation Coalition.

What we chose to focus on

We weren't simply trying to produce a number. We wanted the data to help us make better decisions about how and where we preserve content, and ultimately understand the carbon footprint associated with preserving each participant's content.

What we achieved

We collected detailed data from four of our 12 preservation nodes and used this as the basis for modelling the footprint of the wider archive. This gave us a practical first estimate and, importantly, confidence, skills, and a methodology we can reuse as our infrastructure changes.

Screenshot of the summary tab from the Carbon Footprint Calculation Tool. The spreadsheet provides a framework for calculating the carbon footprint of different areas of digital preservation activity and the summary tab brings it all together.

Lessons learned

Measuring our carbon footprint changed our understanding of where the biggest opportunities for reducing impact lie. For example, it surprised us to learn that integrity checking is our biggest source of emissions, and that one year of organizational travel generates as much carbon as one month of running the entire archive. It also showed us that carbon accounting can't be a one-off exercise: equipment, hosting locations, and services change over time, so the calculations need to evolve too.

Next steps

• Making carbon footprint calculation part of ongoing operational responsibility rather than a one-off project.

• Factoring environmental impact into future infrastructure and preservation-node decisions.

• Asking equipment manufacturers for embodied-carbon information before purchasing new hardware.

• Enabling participants to include the carbon impact of digital preservation in their own carbon-footprint reporting.

Jemma Singleton – Newcastle University Library

Introduction

Newcastle University Special Collections and Archives, UK provide access to historic physical and digital materials for research and teaching purposes to both internal and external audiences. To date, there hasn’t been a consolidated understanding of the carbon footprint of digital activities designed to aid research purposes. The recent activities to evaluate two types of software infrastructure (an in-house vs externally supplied product) provided an opportunity to compare the carbon footprint of each system. In the future it is hoped that this methodology can be applied to the complete workflow activities around digital preservation within an archival context.

Why we got involved in testing the DPC’s Carbon Footprint Toolkit

I got involved with this project because I was interested to understand how carbon generation practices occurred, and at what scale carbon is generated in a number of digital processes that occur as part of our digital preservation activities. I was also fortuitously involved in an time-bound internal project (evaluating two digital library research platforms) that could be enhanced by participation in the carbon calculation working group.

What we chose to focus on

Our main focus was to establish carbon generation values of a cloud-based vs on premise platform infrastructure. A secondary focus was to see if these two different ways of digital hosting has knock on carbon footprint implications for digital activities concerning metadata management, networking, administrative user and end user input.

What we achieved

The achievements of this process has been the generation of a carbon footprint comparison between two digital research systems. In particular, the carbon generation values between on premise vs in cloud storage. This may influence the internal evaluation of said systems and decisions about sustainable implementation choices. This case study has demonstrated that carbon calculation tools can be implemented at a wider university level as a guide to make appropriate digital platform choices.

Further achievements include: an awareness of how carbon calculations are affected by networking digital data through cloud-based servers at an increased capacity to on premise hosting. The eventual carbon calculation for on premise hosting still needs to be finalised which will affect eventual summary totals. A surprising result of analysing carbon generation activities revealed that non-digital administration of short-term projects (such as project-based travel) can really add to the carbon generation weighting. This finding will hopefully encourage colleagues to think carefully about the necessity of their travel for collaborative projects.

There has also been a sense of personal achievement in critically analysing appropriate sources of carbon generation data from multiple parties.

Lessons learned

My main lesson learned from participation in this group is an awareness of where the biggest carbon generators are within digital activities. This means that opportunities can be identified to reduce carbon generation or substantiate appropriate carbon generation in the most sustainable way.

Further lessons have identified areas of industry opacity or lack of knowledge with regards to carbon generation.

In a more practical sense, I have learnt that information about resource requirements, especially for digital networking and download traffic, can be accessed using the ‘inspect’ function on websites.

Next steps

• Promote an analysis of potential environmental impact into future digital preservation operational and policy decisions.

• Conduct a carbon calculation baseline for special collections digital preservation ingest, processing and accessibility workflows.

• Investigate awareness, or improve communication, with internal and external IT service providers about knowledge and access to embodied carbon information.

• Promote increased organisational awareness about utility of the DPC’s Carbon Footprint Toolkit to aid strategic planning.

Dr Eva Samaras and Elise Bradshaw, University of Melbourne

Introduction

The University of Melbourne’s Digital Preservation Program is a small team within the Library. We provide guidance to staff and researchers on strategies, actions and tools for long-term digital preservation. In 2020 we implemented a vendor-hosted digital preservation repository system to safeguard our digital collections, including cultural collections and research. The software as a service solution (SaaS) is fully hosted on Amazon Web Services (AWS) cloud infrastructure in Australia. All digital assets within the solution are managed in both hot and cold AWS storage.

Why we got involved in testing the DPC’s Carbon Footprint Toolkit

In 2025 we conducted a literature scan about environmental sustainability in digital preservation and presented some findings at the Australian Society of Archivists Conference. Sustainability is a core principle in our decision-making, so we welcomed the DPC’s Carbon Footprint Toolkit as a complementary resource to support our program’s environmental objectives and begin quantifying its environmental impact.

What we chose to focus on

In testing the Carbon Footprint Toolkit, we were focused on the cloud carbon figures in the calculator spreadsheet (the In cloud tab of the spreadsheet) and wanted to explore whether we could obtain carbon figures from our vendor. Additionally, we wanted to assess the carbon footprint of transferring data across the University network to AWS infrastructure during ingest.

What we achieved

We contacted our system vendor and asked if they could provide a report from the AWS Customer Carbon Footprint Tool (now known as the AWS Sustainability Console). They then provided us with a CSV export which provides running carbon figures for each month that encompass compute and hardware (Amazon EC2), storage (Amazon S3) and distributed infrastructure (Other). The AWS export includes columns with two types of figures:

• Market-Based Method (MBM) (AWS' default view): Calculates emissions based on the specific electricity contracts and renewable energy certificates (RECs) purchased by AWS. Best for: Tracking your footprint against AWS’s progress towards 100% renewable energy and carbon-neutral goals.

• Location-Based Method (LBM): Calculates emissions based on the average carbon intensity of the local electricity grid where your AWS resources are physically located. Best for: Understanding the true environmental impact of your usage on the local power grid, regardless of AWS's external energy purchases.

We decided to use the LBM figures as they seemed most aligned with the purpose of our investigation. We totaled the LBM figures for the previous 12-month period and then converted from tonnes to kilograms. We then entered the total figure into the DPC Carbon Footprint Calculation Tool In Cloud tab.

In addition to the cloud figures, we reviewed the total volume of data ingested into the preservation system then calculated a daily average in megabytes which was entered into the ‘Acquisition, Transfer and Ingest’ section of the Networking tab of the Carbon Footprint Calculation Tool.

Lessons learned

As a SaaS customer, we lack granular, activity level breakdowns for our actions in the preservation system. Requesting a report from our vendor’s AWS tool proved the most practical way to obtain usable cloud footprint data. The AWS figures revealed that to date, our compute and hardware infrastructure generates more carbon than our storage requirements.

Next steps

As the DPC Carbon Footprint Calculation Tool also includes an option to record travel, we have decided to track the carbon footprint of any inter-state or international travel we undertake for digital preservation-related conferences and events. With a process now established to obtain carbon statistics from our vendor, we will request these figures on a regular basis and use them to inform program decision-making. We will also include carbon footprint data into our regular reports to Library management, enhancing visibility of the environmental impact of our digital preservation work and supporting our ongoing commitment to more sustainable practices.

Kate Roberts (with input from Preservation and Property Operations colleagues at Ngā Taonga; Dan Eady, Luke Tearle and Steve Hutchinson)

Introduction

Ngā Taonga Sound & Vision is the national audiovisual archive for Aotearoa New Zealand. It is the archive for the country’s broadcast TV and radio heritage, as well as New Zealand made content on film and audiovisual formats, from the 20^th-21^st C, deposited by film production companies and producers. Ngā Taonga is a charitable trust which is core funded by Manatū Taonga Ministry for Culture and Heritage. It is not an archive within a broadcast agency, nor a government documentary heritage institution with an audiovisual component to its collections. It holds deposits mostly on behalf of owners, including the New Zealand Government. Our Strategic Plan prioritises the collection, management, preservation and supply of heritage audiovisual content in digital form, while also continuing to hold a legacy sub collection of film and television physical artefacts and documentation.

This case study describes the use of a method for calculating the carbon emissions of the electricity consumption used to run a system whose core function is the ingest of digital preservation bundles, their subsequent storage to LTO tape, and restoration and transcoding of digital files for supply. The method involved using a Carbon Calculation Toolkit (Excel workbook) developed by the DPC Carbon Calculation Support Group (v0.6) in 2026.

The scope of the calculation omits any fabrication or transport of the equipment to the facility and does not include the components related to staffing, or the disestablishment or destruction of the equipment at its end of life.

The key intention is to show the annual impact, in terms of greenhouse gas emissions, of running the digital preservation infrastructure, as a start for characterising, measuring and comparing all collection functions carried out at Ngā Taonga in terms of carbon emissions. The context for undertaking the calculation was the recent work done by our collegial institutions the National Library of New Zealand and Archives NZ to reduce the energy spent on physical collection storage, through better HVAC management and introduction of solar power generation.

Why we got involved in testing the DPC’s Carbon Footprint Toolkit

Ngā Taonga have been associate members of the DPC since 2022 and became aware, through networking, of the carbon footprint calculation work being undertaken to support members. Kate Roberts joined the group testing the calculation spreadsheet tool, in order to become more informed about the climate change impacts of digital preservation, at a time when data centre proliferation for AI processing is causing public concern, increasing the scrutiny of data management practices in public institutions.

What we chose to focus on

We decided to restrict the inputs for the calculation to the electrical power spent running our digital preservation equipment infrastructure, at this point, expressed in kgCO2e/year. This kept the study focussed on the functions carried out in one server room and one digital archive tape library and the cooling and humidity control running in those spaces. This could be augmented in future by also undertaking the calculations for the cost of running equipment used to transfer analogue formats to digital and to manage supply of bulk files to customers. A nominal amount could be added for the cost of people to support the system, although there is currently no permanent team structure dedicated entirely to the digital preservation effort.

What we achieved

We established that we could roughly calculate the power usage over a year by taking a single reading of the amperage running through two PDUs connected to all of the relevant digital processing equipment in our server room, converting it to Watts and then calculating the equivalent annual consumption. While the load on the equipment will be subject to a lot of fluctuation due to ingest and restoration of varying file numbers and volumes, we have not added in a mean based on a range of min to max, to avoid building in more assumptions. In discussion with Steve Hutchinson, our Property Operations Manager, with a background in power station management, I opted to add in the power usage to cool the server room and to run a process cooler to manage both temperature and humidity in the LTO tape library room which acts as our digital archive. On Steve’s advice, I chose a proportion for this power use of 25% of the power to run the digital equipment itself and added it to the two ‘on premise’ lines in the spreadsheet, incorporated in the summary below. In the summary table below the power use in the server room is allocated to the Acquisition, Transfer and Ingest row and the power use in the LTO tape library is added to the Bitstream Preservation line.

We chose to use the international average grid carbon intensity in kgCo2e/kWh of 0.445 (included in the Carbon Footprint Calculation Tool), rather than using figures for New Zealand made available online because the reporting on carbon emissions tends to focus on national aggregated power consumption of key high-volume industries and we couldn’t tell whether the data on carbon intensity would apply for all power consumers. New Zealand does have a relatively high proportion of hydro-electric power, as opposed to fossil fuel generated electricity, giving it an overall score of 83% renewable energy and 0.0148 kgCO2/kWh, so this may mean the carbon footprint is naturally lower than represented in my calculations (ref: Carbon Intensity by Country – Estimated Grid Emissions Rankings | World Power Plants).

Screenshot of the summary tab from the Carbon Footprint Calculation Tool. The spreadsheet provides a framework for calculating the carbon footprint of different areas of digital preservation activity and the summary tab brings it all together. Users of the toolkit are encouraged to focus on the areas of most interest to them (or focus their initial calculations on an achievable goal) rather than trying to calculate everything at once.

Lessons learned

We learnt that digital preservation based on LTO tape storage contributes very little to national or global greenhouse gas emissions. This is because the digital preservation bundles are stored on LTO magnetic tape so they are not in a state of constant availability (and therefore power consumption) as they would be in disc storage. Based on the information gathered we are satisfied that we can show that digital storage running in a stable annual throughput without major ingest or file delivery activity, is very economical, in terms of power, compared to maintaining a physical vault running to magnetic tape storage specifications. Being able to express this in carbon emissions equivalent volumes over time helps in any future discussion about trade-off between physical and digital storage as long term strategies (putting aside the issue of impending degralescence of legacy tape formats).

We learnt that we can roughly compare our digital preservation and its power consumption/greenhouse gas emissions to that of the Finnish national digital preservation services whose annual carbon footprint was calculated in 2023 at 21,869 kgCO2e/year (equivalent to just under the annual emissions for 3 Finns per year). This is a useful comparative set of data for us given that the Finnish digital preservation storage solution is also LTO tape.

Next steps

It would be useful to undertake the work to define the annual carbon footprint of physical collection storage by volume , and to follow up the calculation of just the digital storage with a calculation for the analogue to digital preservation workstreams in film and audiovisual collection processing. It would also be useful to establish the cost of production, transport, replacement and disposal of the digital process infrastructure, especially considering how far we are from the location of the manufacturers of the equipment. We are currently procuring a second LTO tape library to act as our second site for security of the content, in another city in NZ, so our power use for digital preservation and procurement will also increase when we factor in back-up and migration activity.

Amanda Oliver, Wilfrid Laurier University & Julie Shi, Scholars Portal

Introduction

Laurier Archives and Special Collections, located in Waterloo, Canada, is on the shared traditional territory of the Neutral, Anishnaabe, and Haudenosaunee peoples. The archives collects material with an emphasis is on Wilfrid Laurier University, the environmental conservation movement in Canada, the Evangelical Lutheran Church in Canada and its Eastern Synod, music in Kitchener-Waterloo, and the history of Waterloo Region in analogue and digital formats.

Scholars Portal is a service of the Ontario Council of University Libraries, hosted through the University of Toronto Libraries (UTL), to provide shared technological infrastructure to university libraries in Ontario and across Canada. The University of Toronto is on the traditional land of the Wendat Nation, the Seneca, and the Mississaugas of the Credit.

Permafrost, our hosted digital preservation service, provisions member institutions with technical infrastructure and support to process their digital collections for preservation. This includes a dedicated Archivematica instance running on a virtual machine on servers at the UTL data centre and storage provisioned through our Ontario Library Research Cloud (OLRC) storage network. The OLRC is built on the open-source OpenStack Swift software, integrated with the bit preservation management application DuraCloud, and has five storage nodes at partner university libraries across Canada.

Laurier Archives and Special Collections has been a member of Permafrost since 2016.

Why we got involved in testing the DPC’s Carbon Footprint Toolkit

Since 2022, Laurier Archives and Special Collections has collaborated with Facilities and Asset Management and the Sustainability Office to measure emissions of archival storage spaces. This was accomplished by calculating m3 of natural gas and kWh of electricity used during a selected period, calculating greenhouse gas emissions using co-efficient factors, and converting to kg CO2e using Global Warming Potential. The current kg CO2e values for archival storage spaces on campus is comparable to driving an average gas vehicle for 14,000 km annually. These values are 0.13% of total campus scope 1 emissions and 0.10% of total campus scope 2 emissions. This work only captures emissions for archival storage spaces.

Laurier Archives and Special Collections wanted to help test the DPC’s Carbon Footprint Toolkit to develop a more fulsome understanding of the department's total carbon emissions. Scholars Portal wanted to help test the Carbon Footprint toolkit to understand how the tool may be used to support Permafrost members interested in the environmental impacts of digital preservation and to determine the impacts of shared technological services more broadly.

What we chose to focus on

Laurier Archives and Special Collections focused on gathering internal data to the institution to populate the on premise and website use tabs of the toolkit. From a service provider standpoint, Scholars Portal focused on gathering information about running Archivematica and the OLRC on virtual machines hosted on local servers. These two systems support institutions with addressing many activities within the Bitstream Preservation, Content Preservation, and Metadata Management functions.

Information about network traffic between Scholars Portal machines and between Laurier and Scholars Portal requires discussions with additional partners and will be explored in a later phase.

What we achieved

The systems team at Scholars Portal has created various dashboards and inventories that provide information about the virtual machines and physical hosts that Laurier’s Archivematica instance and the OLRC run on, and graciously shared access and time to allow us to gather data. This data does not account for all the moving parts involved in Permafrost, but provides a window into the hardware used, the resources assigned to and consumed by each VM and host server, and data about compute and storage use that forms the basis of initial emissions estimates for the service.

Figuring out how to map the data collected and the virtualized environment to the Carbon Footprint Calculation tool proved more challenging and required reading up on the exact meaning and purpose of various metrics and the differences between virtual and physical computing environments. On advice from the systems team, we also switched from focusing on physical hosts, which will invariably change as hardware is refreshed, to the VMs themselves, where the main activity and resource consumption is happening. When data is only available for the host server or service (in the case of the OLRC, which is used for other storage needs as well) as a whole, resource usage was also assumed to be proportionate to the vCPU assigned to the specific VM or storage used by the service or Laurier respectively.

Above all, this exercise has provided us with a better understanding of the components involved in processing and providing access to digital materials with Permafrost and the factors we should be considering when calculating emissions for digital systems and processes.

Lessons learned

Gathering the data required to measure the carbon footprint of digital preservation activities takes time. You need to explain the purpose and importance of this work to information technology partners and service providers, who may not immediately understand the what and why of this ask. We spent a significant amount of time explaining the project and what information we needed to test the Toolkit, with varying levels of success. Though we do not have all the answers yet, we now know what questions to ask and have a sense of the time and effort needed to gather this information.

Next steps

Our next step is to refine this process so it can be replicated by other Permafrost members. This will allow peer institutions to include carbon emissions estimates from digital preservation activities in their own carbon calculations. We also hope to bring this conversation to the Permafrost community to foster ongoing discussions on the environmental impacts of digital preservation more broadly.