* [Book: *How Bad Are Bananas?*](https://hackmd.io/@h8Wc2RnARLypUKWR8Znxig/bananas-book)
* [Learning about Green Software](https://hackmd.io/@h8Wc2RnARLypUKWR8Znxig/learning-green-software)
# Hack Day: supporting resources
## Websites and research papers to support carbon calculations

| Website/URL | Useful for | Notes |
| -------- | -------- | -------- |
| **[Research paper: Assessment of the energy footprint of digital actions and services](https://op.europa.eu/en/publication-detail/-/publication/d3b6c0a1-1171-11ee-b12e-01aa75ed71a1)** | All roles | Well-researched paper which provides estimations of common onlineactivities, such as email, file storage and video streaming
|**[Carbon Base](https://base-empreinte.ademe.fr/)** | All roles | Requires user to create a free account. In French, so needs translation tools. Covers emissions from a wide range of activities, including non-digital
| **[Email or letter?](https://tfs.uk.com/email-or-letter-blog/)** | All roles | |
| **[Carbon footprint of greeting cards](https://blog.dontsendmeacard.com/exeter-university-report-carbon-footprint-of-a-greetings-card.html)** | All roles | This data is fairly old (2014)
**[Unnecessary emails](https://company.ovo.com/think-before-you-thank-if-every-brit-sent-one-less-thank-you-email-a-day-we-would-save-16433-tonnes-of-carbon-a-year-the-same-as-81152-flights-to-madrid/)** | All roles |
| **[Emails vs. letters – research from Germany](https://www.researchgate.net/publication/344502004_Dialogue_Marketing_Ecological_Sustainability_of_Letter_and_E-Mail_in_Comparison_in_Germany/link/5f7ce71c92851c14bcb36a9e/download)** | All roles | A more thorough analysis than most |
| **[Awesome Green Software](https://github.com/Green-Software-Foundation/awesome-green-software)** | Developers, architects |
| **[Cloud Carbon Footprint](https://www.cloudcarbonfootprint.org/) and [demo site](https://demo.cloudcarbonfootprint.org/)** | Developers, architects | Open source tool for measuring the emissions from your cloud operations. The demo allows you to compare various providers and locations |
| **[Digital Decarbonisation.](https://digitaldecarb.org/)** [calculator](https://digitaldecarb.org/calc_it.html) and [data map](https://digitaldecarb.org/co2-data-map/) | Data analysts | Research project looking specifically at the environmental impacts of data. Second link provides estimates on the carbon footprint of data. The third link is what they refer to as a ‘heatmap’ for the analysis of the environmental impact of data through its lifecycle |
| **Datavizta.** [raw data](https://dataviz.boavizta.org/) and [main website](https://www.boavizta.org/en) | Developers, architects, data analysts | Rich source of data for the environmental impacts of all types of hardware, including laptops, monitors and servers. Includes lifecycle data regional carbon intensity estimates |
| **[Impact Engine Framework](https://github.com/Green-Software-Foundation/if)** | Developers, architects |
| **[Understanding the results of cloud providers' carbon calculators](https://boavizta.org/en/blog/calculettes-carbone-clouds-providers)** | Developers, architects, supplier chain managers | In-depth review of how each of the major public cloud providers estimate their emissions |
| **[Energy Efficiency across Programming Languages](https://sites.google.com/view/energy-efficiency-languages/home)** | Developers | In-depth study comparing how different programming languages compare in terms energy efficiency|
|**[Do I emit CO2 when I surf the internet?](https://www.energuide.be/en/questions-answers/do-i-emit-co2-when-i-surf-the-internet/69/)** | All roles | Belgian website giving General guidance on the emissions of various digital activities |
|**[Climatiq](https://www.climatiq.io/)** | Developers, architects | Large dataset, covering all types of carbon accounting. API can be used to provide estimates for more than 15,000 activities.|
|**[We’re getting a better idea of AI’s true carbon footprint](https://www.technologyreview.com/2022/11/14/1063192/were-getting-a-better-idea-of-ais-true-carbon-footprint/)**|All roles|AI startup Hugging Face has undertaken the tech sector’s first attempt to estimate the broader carbon footprint of a large language model.|
|**[The Carbon Footprint of AI and Deep Learning](https://www.analyticsvidhya.com/blog/2022/03/the-carbon-footprint-of-ai-and-deep-learning/)**|Developers, architects|Yet another article/research paper highlighting the serious environmental problems with AI|
|**[Digital Declutter for Businesses](https://www.wholegraindigital.com/digitaldeclutter/)**|All roles|A toolkit that can help you measure and reduce the digital carbon emissions of your business.|

## Carbon intensity tools for websites
The following is a list of websites that try to estimate the carbon footprint of a web page or website.
• https://www.websitecarbon.com/
• https://ecograder.com/
• https://ecoping.earth/
• https://greenframe.io/
• https://globemallow.io/#api

## Browser extensions

### Website reporting extensions

* **Beacon** (https://digitalbeacon.co/). Developer tool to evaluate the environmental impact of a web page.
* **Globemallow** (https://globemallow.io/#home). Grades a web page on its environmental impact.
* **GreenIT Analysis** (https://chrome.google.com/webstore/detail/greenit-analysis/mofbfhffeklkbebfclfaiifefjflcpad). Adds a Green Software analysis tab to the developer tools.
* **Website Footprint** (https://chrome.google.com/webstore/detail/website-footprint/jlpnojjijmliogpegigbllcnpckflnik). Very simple plugin that measures the emissions for a web page and gives it a grade.

### Browsing habits extensions
* **CO2RD** (https://ecord.info/). Can be edited to use the UK grid to calculate emissions.
* **Pebble** (https://chrome.google.com/webstore/detail/eco%E2%82%82rd/abgangmjmkjbpimkgnlbbhkellcekegj). Can be edited to uses the average European grid to calculate emissions.
* **WeDeex** (https://chrome.google.com/webstore/detail/wedeex/ojlagggckhpedblhemgjhecbggnibale). Uses the French grid to calculate emissions. "HOW IS THE POWER GENERATION QUALITY INDICATOR CALCULATED? *In every area where we operate, we collect real-time production data from grid operators. We then analyze the quantities provided by each production type and compute them into a real-time carbon emission rate from power production, expressed in kgCO2/kWh. This real-time emission rate is then compared to the past and near future to determine if the current situation is good or not, and if you should try to delay your consumption or on the opposite, take the opportunity to charge your batteries or run a washing machine.*"

## Going beyond carbon
Those working in sustainability can sometimes be accused of carbon ‘tunnel vision’. This is a very fair criticism. While carbon emissions are arguably the most important area to focus on (at least from a human perspective), true sustainability must look much wider. Here are some other things to take into consideration which you may wish to use as KPIs:

* **Use of raw materials, especially so-called ‘conflict minerals’**. ICT uses vast amounts of rare earth minerals. How are you ensuring that you service is minimising or eliminating their use?
* **Use of water**. ICT and datacentres in particular are huge consumers of water. Technologies such as Blockchain, AI and Bitcoin are especially noted for their extravagant water consumption, primarily for cooling. How are you ensuring that you service is minimising water consumption? Both this and the use of raw materials are related to the [9 Planetary Boundaries](https://greenly.earth/en-gb/blog/ecology-news/all-you-need-to-know-about-the-9-planetary-boundaries), of which we have now exceeded 6.
* **Life-cycle assessments or LCAs**. In most cases, the embodied carbon from ICT is the primary source of emissions. But responsible disposal of these devices is critical once they have reached end-of-life within your service. How will you manage and improve LCA?
* **Understanding of the waste hierarchy (reduce/reuse/recycle)**. The ‘waste hierarchy’ has those 3 terms in that order. Recycling must be seen as an absolute last resort. This also helps support the move towards a circular economic model. How does your service understand and deal with the waste hierarchy?
* **Circular economy principles and environmental stewardship**. Referring back to the UN SDGs and the Social Value Model, is there a way to demonstrate how your service aligns to these?
* **User satisfaction**. Last but not least, how will you way up environmental considerations against user needs? Are there conflicts or synergies?

## Further guidance

* **The Technology Code of Practice**. Point 12 outlines the GOV.UK position on sustainability in ICT. See https://www.gov.uk/guidance/make-your-technology-sustainable
* **Government Functional Standard**. The government functional standard is to set expectations for the management of digital, data and technology in government. Section 8.7 references sustainability. See https://www.gov.uk/government/publications/government-functional-standard-govs-005-digital/government-functional-standard-govs-005-digital-html
* **W3C Web Sustainability Guidelines, 1.0.** A brand new initiative, covers business, design and technology aspects of digital services. See https://w3c.github.io/sustyweb/. Condensed version: https://w3c.github.io/sustyweb/glance.html and quick reference: https://w3c.github.io/sustyweb/quickref.html
* **Green Software Maturity Matrix**. A new initiative by the Green Software Foundation. A self-evaluation tool to help organisations and teams understand their level of maturity in this area. See https://maturity-matrix.greensoftware.foundation/
* **The intro guide to digital eco-design**. French-led project for designers, developed in part by the team which helped write the W3C Web Sustainability Guidelines. See The intro guide to digital eco-design. See https://eco-conception.designersethiques.org/guide/en/

* **Planet Centred Design**. A new proposal by Defra to support those designing and building digital services. See https://defradigital.blog.gov.uk/2023/03/06/how-being-planet-centric-can-really-make-a-difference-when-designing-services/ and https://drive.google.com/file/d/161GLM3MiuLmehr4zR6yfmFRV67ON4ml1/view
* **General Digital Services Ecodesign (RGESN)**. Official guidance from the French government on designing digital services sustainability. In French only. See https://ecoresponsable.numerique.gouv.fr/publications/referentiel-general-ecoconception/
* **Step by Step Towards Sustainable AI**. High-leve guidance from algorithmwatch.org on planning, developing and implementing responsible AI. Includes environmental considerations. See https://sustain.algorithmwatch.org/en/step-by-step-towards-sustainable-ai/
* **Project RAISE (Responsible AI Strategy for the Environment) from the Global Partnership on AI**. This group has produced several papers on the impacts of AI on the environment. They also wrote the 2021 guidelines *Climate change and AI: Recommendations for government action* which is worth looking at. See https://www.gpai.ai/projects/responsible-ai/environment/
* **The Tarot Cards of Tech**. Lastly, something a little bit fun. If you are stuck for ideas or have a creative block, consult this deck. See https://tarotcardsoftech.artefactgroup.com/

## Glossary

**Carbon forecasting and timeshifting**: in countries such as the UK with a distinctly mixed energy grid, the emissions generated by using electricity can vary dramatically throughout the day or week, based on environmental and market conditions. A bright and windy day will typically mean that emissions can be very low, while a dull and windless day will mean that the same kWh produce far higher amounts of CO2. Carbon forecasting uses the predicted weather to look for ‘green windows’ of time. The https://github.com/bbc/carbon-minimiser is a good way to get started with this in the UK, and the Green Software Foundation has also produced the https://github.com/Green-Software-Foundation/carbon-aware-sdk which works at country level.

**Carbon intensity**: the carbon cost or producing electricity, based on the local energy mix. In the UK, this is typically a mixture of gas, nuclear, wind and solar. In western Europe, countries with high intensities (over 400g per kWh) include Poland, Czechia and Germany, all of which rely on coal for much of all of their power generation. Countries with very low carbon intensity include Iceland, Norway and Sweden, which use energy sources such as hydro, geothermal or nuclear. See https://app.electricitymaps.com/map for an interactive map/tool for a global view and https://www.carbonintensity.org.uk/ for a UK regional view. Both services have very useful APIs to support activities such as demand shifting.

**Demand shifting**: in Green Software development and GreenOps, there is the opportunity to locate servers or activities in regions where the carbon intensity is lower. This could mean moving some or all of your activity in a country or region with a cleaner grid, such as Sweden or Scotland. It could also mean using carbon forecasting (for example using https://github.com/bbc/carbon-minimiser) to schedule some activities when your local grid is greener.

**Embodied carbon**: the carbon emitted when creating a physical object. For the purposes of this exercise, we look closely at the carbon that is emitted when creating a device such as a smartphone, monitor, laptop or server. For consumer devices, the embodied carbon makes up the vast majority of the overall emissions, as much as 80% in most cases. For servers, the usage can make up a far bigger percentage of the emissions, but this depends on how the server is utilised. As embodied carbon is such an important number, this is why devices must be kept in operation for as long as possible. Getting a new phone every year or a new laptop every three years is not acceptable. See https://dataviz.boavizta.org/ and https://github.com/rarecoil/laptop-co2e for detailed breakdowns of different device types and manufacturer data. Understanding embodied carbon is critical for measuring your Scope 3 Emissions.

**Energy scopes**: the internationally recognised Greenhouse Gas Protocol uses the term ‘emission scopes’ to be used for reporting purposes. Broadly speaking, Scope 1 includes the burning of fossil fuels within an organisation’s operations. This would include operating a vehicle fleet or heating a building with gas-powered radiators. Scope 2 is the direct use of purchased electricity within an organisation’s operations. For a typical government department, this would include lighting, air conditioning and the power to run laptops and printers. Scopes 1 and 2 are generally easy to measure, just check your utility bills. Scope 3, on the other hand, is much, much bigger and more complex. For most organisations, Scope 3 makes up about 80% of emissions. In the broad definition, Scope 3 includes all the other activities needed to run a business or organisations, such as employee commuting, cloud hosting, purchased goods and the like. Scope 3 emissions fall into 15 subcategories, which can be found at: https://www.compareyourfootprint.com/difference-scope-1-2-3-emissions/. Government departments are obliged to report on Scope 3 emissions to Defra quarterly in the STAR report. The reporting of Scope 3 will become mandatory in the EU (see https://www.thomsonreuters.com/en-us/posts/esg/csrd-esg-regulations/) and was recently passed into law in California (see https://www.theguardian.com/us-news/2023/oct/09/california-carbon-emissions-law). It would seem likely that similar legislation will follow in the UK.

**ESG reporting**: Environmental, social and corporate governance. Our home planet is not in a healthy state. ESG reporting includes all legislation aimed at holding governments, institutions and corporations to account with regards to environmental and social issues. This now includes a wide array of things, such as the United Nations Sustainable Development Goals, Net Zero targets, Greening of Government Commitment, the Social Value Model and recent Corporate sustainability reporting legislation passed by the EU. DWP is committed to Net Zero by 2050 at the latest, and for all government departments abd ALBs, the Defra STAR report is the primary ESG reporting mechanism.

**GHG**: Green House Gas. CO2 is obviously the most well-known, but other GHGs include methane, nitrous oxide and sulphur dioxide, to name just a few. To make things easier, the abbreviation CO2e is often used, which stands for CO2 equivalent, where the warming potential of the various GHGs can be compared in a way which all can agree upon.

**GreenOps**: building on the job titles ‘DevOps’ and ‘FinOps’, ‘GreenOps’ is the role within a software team that looks at optimising the deployment and running of software (typically in the cloud) based on GreenSoftware principles. GreenOps is well aligned to FinOps, in that both typically look at reducing cloud usage, but GreenOps takes this one step further by including carbon as a cost.

**Location vs. market-based reporting**: carbon intensity can be reported on in two ways. With location-based reporting, the emissions are calculated by using the carbon intensity for the local or national grid. In this way, emissions can be calculated quite accurately. Market-based reporting allows companies to buy renewable energy certificates, for example from countries such as Norway. In essence, this is a form of offsetting, and the accuracy of this approach is a topic of debate. The Sustain.Life website recently published an article on this: https://www.sustain.life/blog/purchased-electricity-emissions-accounting-market-and-location-based. Most cloud providers use market-based calculations, which is why they can claim to be powered on ‘100% green energy’. The UK Government and Defra’s STAR report use a location-based approach, as the reliability and accuracy of market-based approaches can be difficult to verify. See https://www.gov.uk/government/publications/greenhouse-gas-reporting-conversion-factors-2023 for the conversion factors used by GOV.UK. The SCI also uses location-based reporting for their calculations.

**Net Zero**: the internationally agreed and legally binding term used to describe how organisations and countries will reduce their emissions by 2050 in order to keep as close to the 1.5C of global warming, the agreed ‘safe’ limit to sustain human and much of plant and animal life. For many organisations, this must mean producing zero or near zero emissions. See https://netzeroclimate.org/what-is-net-zero-2/ and https://energysavingtrust.org.uk/what-is-net-zero-and-how-can-we-get-there/ for further information.

**Offsetting, carbon-neutral**: although popular up until recently, these terms have fallen out of favour due to their vulnerability to ‘greenwashing’. Recent EU legislation has in fact banned the use of terms such as ‘carbon neutral’ (see https://carbonmarketwatch.org/2023/09/20/carbon-market-watch-welcomes-eu-ban-on-carbon-neutrality-greenwashing/). Instead, organisations must explain how they will reach Net Zero. For certain hard-to-decarbonise industries such as the manufacture of steel and cement, offsetting can be used as a last resort.

**Rebound effect**: the tendency for gains in efficiency to be outpaced by increases in demand. This can be observed in areas such as transport, where widening roads typically increases traffic, or in digital where, improving efficiency only increases demand, quickly cancelling out the efficiency gains. Sufficiency can be seen as a strong counterpoint to efficiency. Also see Wirth’s Law.

**Social Value Model**: this act was passed into UK legislation in 2020, and is composed of 5 themes which should be taken into account when awarding government contracts. Theme 3, Fighting climate change: Effective stewardship of the environment, is now 10% of the decision-making process when awarding large ICT contracts.

**Software Carbon Intensity specification (SCI)**: a new ISO standard for measuring the environmental impact of software produced by the Green Software Foundation. The SCI uses location-based reporting combined with embodied carbon to calculate the impact of a piece of software. See https://github.com/Green-Software-Foundation/sci/blob/main/Software_Carbon_Intensity/Software_Carbon_Intensity_Specification.md for details.

**UN SDGs**: the United Nations Sustainable Development Goals. These are the 17 internationally agreed goals which were adopted globally in 2017. They include many goals related to sustainability, and are often referred to within ESG reports.

**Wirth’s Law**: we are all probably familiar with Moore’s Law on computer performance. Wirth’s Law observes that software is getting slower more rapidly than hardware is becoming faster. It’s a good argument for sufficiency.