The Four Key Environmental Factors of ICT: Energy, Carbon, E-waste and Water

Almost every element of the ICT lifecycle – from material acquisition to product disposal – affects the environment. In this article we look at the four major areas of impact: Energy consumption, greenhouse gas (GHG) emissions, electronic waste (e-waste) generation and water consumption. We also discuss ways your organization can measure and reduce each of these.

The Rise and Rise of Energy Consumption

Energy is used throughout the ICT product and service lifecycle, mostly as electricity but also in other forms such as gas and vehicle fuel. Many countries rely heavily on non-renewable energy sources, which is putting increasing pressure on a finite global resource. ICT’s substantial energy consumption also produces GHG emissions, which are regarded as the primary cause of climate change.

Industry estimates suggest ICT is responsible for 2–3% of annual global GHG emissions1– similar to the aviation industry’s contribution. The figure below shows the likely increase of ICT-related GHG emissions from 2002 to 2020, and the relative split of emissions across the three major elements of ICT: data centers, voice and data networks, and end-user devices.

ICT emissions

This graph shows that end-user devices such as PCs and laptops contribute just over half of all ICT-related emissions, with data centers and networks responsible for around one-quarter each. This ratio is forecast to remain similar in the foreseeable future. However, there are some predictions that ICT-related emissions could rise further to 4–6% of global emissions by 20202,  reflecting increasing demand for ICT, especially in growing economies such as China and India.

As a result of this inevitable surge in ICT-related GHG emissions, the industry is focusing most of its sustainability efforts on reducing emissions by improving energy efficiency. However, this is not always a simple equation, as we’ll explain below.

What is the difference between carbon, CO2 and GHG emissions?

The terms ‘carbon’, ‘CO2’, ‘carbon dioxide’ and ‘GHG emissions’ are often used interchangeably, but not always accurately. Carbon is the element that combines with oxygen to produce carbon dioxide (CO2). There are a number of greenhouse gases, of which carbon dioxide is the most abundant but not the most damaging in terms of its global warming potential (GWP).

Other greenhouse gases are more damaging per volume – for example, the GWP of methane is around 25 times that of carbon dioxide – but less abundant. When referring to the actual amount of greenhouse gases generated by a process, the conventional term ‘CO2e’ (carbon dioxide equivalent) is used, which weights all greenhouse gases involved relative to the GWP of carbon dioxide (=1). For example, an organization’s servers might generate 45 tons of CO2e each year.

A Question of Scope

When calculating and reporting the GHG (or CO2e) emissions attributable to a process, product, service or organization (referred to as ‘carbon footprinting’) it is essential to consider their scope, or more precisely, which of the three scopes – as outlined below – to include. The Greenhouse Gas Protocol defines direct and indirect emissions as follows:

Direct GHG emissions are generated from sources that are owned or controlled by the reporting entity. Indirect GHG emissions are produced as a result of the reporting entity’s activities, but occur at sources owned or controlled by another entity.

The Greenhouse Gas Protocol categorizes these direct and indirect emissions into three scopes:

  • Scope 1: All direct GHG emissions.
  • Scope 2: Indirect GHG emissions from the consumption of purchased electricity, heat or steam.
  • Scope 3: Other indirect emissions, including the extraction and production of purchased materials and fuels; transport-related activities in vehicles not owned or controlled by the reporting entity; electricity-related activities (for example, transmission and distribution losses) not covered in Scope 2; outsourced activities; and waste disposal.3

For ICT and more generally, most carbon footprinting to date has focused on Scopes 1 and 2. This is because these emissions sources are the easiest to control, have the most data available, and are most affected by legislation and standards. However, organizations are increasingly reporting against Scope 3 as well, as they become more aware of their indirect (or ‘embedded’) emissions, and as governments set new standards.

This growing focus on Scope 3 emissions is changing business expectations. For example, organizations are paying more attention to the sustainability performance of their supply chains, and are demanding greater transparency in their suppliers’ carbon footprint claims and reporting.

Organizations are also increasingly conducting full ICT lifecycle emissions accounting. This allows them to make more informed decisions when they refresh their technology. For example, replacing existing desktop computers with more energy-efficient equipment may seem to be environmentally beneficial (over, say, three years), but if the company takes into account embedded emissions, the total GHG emissions might be greater than if the assets were ‘sweated’ for another year.

Energy use and GHG emissions

Reducing energy consumption almost always means reducing GHG emissions. The energy might be used as electricity for desktops, servers and other equipment (producing Scope 2 emissions), or as other forms of energy such as the fuel for vehicles transporting equipment or support engineers (producing Scope 1 emissions). Electricity is the dominant form of energy used throughout the ICT lifecycle.

Depending on the power-generation method, the carbon intensity of electricity – known as the ‘emissions factor’ and normally expressed as kilograms of CO2e per kilowatt hour of electricity (kg CO2e/kWh) – varies considerably. For example, the emissions factor can be almost zero for geothermal, wind and hydro energy sources, or more than 1 for older coal-fired power stations.

Most countries generate electricity from a variety of sources and use a ‘grid average’ to estimate GHG emissions from grid electricity. The grid average varies greatly by country depending largely on the ratio of fossil to non-fossil power sources used. For example, due to its heavy reliance on coal, Australia has a grid average of around 1 kg CO2e/kWh. Germany and the UK, which use a mix of power-generation methods, have grid averages of around 0.5, while Iceland’s is almost zero because it only uses renewable sources such as hydro and geothermal energy.

The figure below illustrates how the disparity in energy sources in four European countries causes a large variation in the amount of use-phase GHG emissions generated by a desktop PC.

Fujitsu PC data

Because of the variations in carbon intensity between countries, the impact on GHG emissions by reducing ICT energy use also varies considerably. This disparity can have major positive or negative impacts when an organization moves components of an ICT system from one country to another.

The most obvious example of this is the Cloud model, in which a data center hosting an organization’s ICT services may be in a different country. Migrating to the Cloud, therefore, could mean a reduction or increase in GHG emissions for that organization, depending on the difference in emission factors between the countries.

Embedded emissions: The hidden impact

Embedded (or embodied) emissions are those incurred before a product is used. For ICT products, this typically means when the raw materials are extracted and manufactured (which also normally involves importing components from other countries to assembly plants), and all the activities necessary to deliver the equipment to where it will be used.

Embedded emissions have traditionally been ignored when carbon footprinting ICT products and services. This is mainly because it’s been difficult to obtain reliable data, but also because many organizations believe embedded emissions are insignificant compared to use-phase emissions.

But we now know that for much ICT equipment, embedded emissions make up a substantial share of the total lifecycle emissions. For example, a 2010 Fujitsu study found that 40–50% of the total lifecycle GHG emissions of a desktop PC can be produced before it is first used.4  (This figure is lower for servers because they are normally powered almost continuously over a longer time period than PCs.)

Can you cut emissions without reducing energy use?

Increasing the energy efficiency of ICT infrastructure is usually a major objective in organizations’ environmental strategies because it also reduces costs. However, organizations can reduce their carbon footprint by using less fossil fuel–based electricity – either onsite (such as solar panels on a data center) or offsite (for example, a renewable tariff purchased from an electricity utility).

The renewable cost–benefit calculation depends on a number of factors including government incentive schemes, the payback period of onsite installations, the availability of renewable tariffs from power utilities, and how renewables are treated in any carbon-tax reporting scheme. As a result, the financial benefits of increasing renewable energy use can be much less clear-cut compared with the benefits of reducing energy usage – if they exist at all. Where there is a net cost of using renewables, organizations that see brand value in environmental responsibility may consider it worth paying.

E-waste: More computers, bigger disposal problem

Around 40–50 million tons of e-waste is generated annually, including 30 million computers thrown away in the US and 100 million mobile phones discarded in Europe.5  The US and China are the two biggest producers of e-waste, disposing of around 3 million tons and 2.5 million tons respectively each year.

E-waste will rise rapidly over the coming decades as the use of ICT increases, and consumers and businesses constantly invest in newer and better (and often cheaper) technology. The current slowdown in some parts of the global economy is unlikely to make more than a small dent in this trend.

The consequences of e-waste

  • Air pollution from processing hazardous and toxic materials can be dangerous to human health. This is a significant problem in countries not subject to international legislation such as the European Union’s Waste Electrical and Electronic Equipment Directive.
  • Dumping materials that can’t be recycled requires more land – an increasingly scarce resource in many areas of the world – which can’t be recovered economically. According to the US Environmental Protection Agency (EPA), only 15–20% of e-waste is recycled, with the remainder going to landfill or being incinerated.6
  • Recycling uses large amounts of energy, generating GHG emissions.
  • Some companies in countries with stringent e-waste regulations illegally ship e-waste to countries with few or no regulations. For example, the Environmental Investigation Agency found that much of the UK’s e-waste is illegally shipped to Africa.7

Reduce, reuse, recycle

The simple ‘reduce, reuse, recycle’ mantra has long served as best practice for organizations or individual consumers wanting to reduce their waste. It’s an excellent principle, but of course there can be complexity in making it a reality. For example, a product might need to be repaired, refurbished or reconditioned before it can be reused.

When considering investing in new ICT equipment, your first step should be to critically assess your needs in order to reduce demand. Can your business function just as well by extending the life of the equipment you currently have or by buying a smaller, lower-spec alternative?

Once a product reaches the end of its life, you might consider various ways to reuse it, for example using it for the same purpose but in a different part of your organization; repurposing it within your organization; or remarketing it (refurbished if necessary) to another organization.


WEEE Man, part of the Eden Project in Cornwall in the UK, is a striking and symbolic example of how e-waste can be reused for entirely different purposes. It represents the amount of waste electrical and electronic equipment (WEEE) the average British household discards in a lifetime.

Recycling generally involves breaking the product down into its constituent parts, and recycling the parts that have a further useful life – either as components for other ICT products or as raw material.

Remanufacturing has received growing attention in recent years due to its environmental benefits. It involves retaining as much as possible of the assembled structure of the product (as opposed to breaking it down for traditional recycling) and using it ‘as is’ when manufacturing a new product. This wastes less material and uses less overall energy.

ICT equipment manufacturers have a major part to play in reducing e-waste. For example, manufacturers can minimize their use of non-recyclable materials (biodegradable materials are generally beneficial to the environmental but recyclable is better) and redesign their products in a way that facilitates remanufacturing.

Reducing e-waste responsibly

As you can see, reducing e-waste at the organizational level can be complex, as there are many factors to consider and possible trade-offs to evaluate. Tackling e-waste at a national or regional level is even more difficult because it involves numerous stakeholders – including consumers, organizations, ICT hardware vendors and governments – that may have conflicting agendas.

Tackling the problem at the global level adds a new dimension of complexity, mainly due to the various e-waste legislation regimes and the huge disparity in living standards and cultural attitudes to waste. In very simple terms, developed nations have created the kind of throw-away habits that developing countries can’t yet afford. We could argue, then, that the disparity in e-waste legislation between developed and developing countries is just a result of the difference in attitudes; the richer countries need tougher legislation because their e-waste problem is much bigger – at least for now.

Water consumption: A rising global problem

Water scarcity is now a major global challenge, and is likely to become progressively worse due to climate change and the world’s rapidly expanding population. A lack of water will have different impacts around the world, due to uneven rainfall patterns and varying demand for water based on different living standards.

While the future scale and geographic distribution of water scarcity is still uncertain, it seems clear that reducing demand for water will be essential in most regions. It also appears that increasingly efficient water usage will be driven by a combination of financial, legal, political and technological measures.

Many environmental commentators believe water scarcity will be ‘the new carbon’ in terms of its global recognition as a major sustainability challenge and the urgent need to address it. However, whereas the impact of carbon is long-term, indirect, cumulative and global, water scarcity has local, direct and immediate impacts.

The water–carbon trade-off

Water scarcity and CO2e emissions are linked in several ways. For example, the energy sector is a massive water consumer (for instance, in power station cooling) and the water sector uses significant amounts of energy (such as electricity for pumping). This means reducing demand for one may reduce demand for the other. That said, this logic should not be stretched too far; there are situations where energy consumption can be reduced by using more water and vice versa.

For example, the amount of energy needed to cool data centers (mostly through air cooling) has traditionally been similar to the amount required to power the ICT inside them. While data center cooling efficiency has improved markedly in recent years, water cooling is seen as a more energy- and carbon-efficient method than air cooling because water conducts heat far better than air, and allows surplus heat to be more easily reused for domestic or commercial heating.

For these reasons, some industry experts predict a significant increase in data center water cooling over the next decade.8  This could mean increased water usage by the ICT industry but less energy consumption and related GHG emissions.

Measuring the overall sustainability impact of water cooling, or even determining whether there is a net positive or negative effect, will depend on the relative importance placed on energy, carbon and water. This importance will vary by circumstances such as region, season and financial cost. Further, because water is used in many forms, each will have its own relationship with energy and carbon. For example, chilled water uses more energy than unchilled. Also, recycling water within the data center requires energy – though probably less than the energy used by a water utility to provide the same volume of drinking water.

The challenges in measuring ICT’s water demand

There are currently no reliable figures on ICT’s total global water usage. This is not surprising, as the theory and practice of ‘water footprinting’ of ICT (measuring water consumption from manufacturing through to product disposal) is about where ICT carbon footprinting was three or four years ago. Metrics and standards are being developed but there is very little real-world experience or data – for example data from ICT suppliers on the water used to manufacture their products.

It is fairly clear that the two main sources of direct water use are manufacturing (especially computer chip production) and data center operations, though there are no accurate figures available for these. Taking into account indirect water usage – especially the water used in generating electricity – may change the total water usage profile and show that a large proportion of water is consumed in the use phase. But the lack of data makes it difficult, if not impossible, for an organization to accurately calculate the water footprint of its ICT.

A new sustainability metric

One promising method for measuring the ICT industry’s water usage is The Green Grid’s water usage effectiveness (WUE) metric for data centers. WUE is a companion metric to power usage effectiveness (PUE) and carbon usage effectiveness (CUE). It provides a quantifiable basis that allows data center operators to consider various cooling options, and, in conjunction with the related PUE and CUE metrics, better analyze trade-offs in overall sustainability strategies; for example, energy consumption versus water consumption.

There are two variants of WUE. Similar to the direct (Scope 1) and indirect (Scope 2) definitions for carbon emissions, The Green Grid defines WUE metrics as follows:

  • WUE is a site-based metric that is an assessment of the water used onsite in data center operations. This includes water used for humidification, and water evaporated onsite for energy production or cooling of the data center and its support systems (similar to carbon Scope 1).

    (Annual site water usage)/(IT equipment energy)

  • WUEsource is a source-based metric that includes water used onsite and water used offsite to produce the energy used onsite. Typically this adds the water used at the power-generation source to the water used onsite (similar to carbon Scope 2).

    (Annual source energy water usage + annual site water usage)/(IT equipment energy)

Both WUE and WUEsource are expressed as liters of water per kilowatt hour of electricity (L/kWh).

It should be noted that WUE is concerned only with operational water efficiency – akin to carbon Scope 1 and 2 – and not with upstream or downstream (Scope 3) water usage. This is considered too difficult a calculation to make, in part due to the lack of available data. The Green Grid is continuing to develop the WUE metric and provides guidance on its calculations. We hope that in the coming years, data center operators will adopt this metric, and any similar water-related metrics, to increase their understanding of ICT-related water usage and provide a sound basis for reducing it.

Too much water is also a problem

As a result of climate change, we are seeing more frequent extreme weather events around the world, and climate experts predict that this will increase. Such extreme weather events include torrential rainstorms and subsequent flooding. While ICT can’t cause these problems, it can be affected by them. An extreme example of this was the massive and prolonged disruption to the worldwide hard-disk supply chain following the 2011 Thailand floods caused by tropical storm Nock-ten.

For further thought and discussion

  • When calculating the carbon footprint of your ICT operations, do you take into account all sources of emissions, including not just use-based sources but also those embedded in other stages of the ICT lifecycle?
  • Has your organization considered ways of extending the operating life of its ICT equipment; for example, by delaying technology refreshes or repurposing hardware?
  • Is your organization considering calculating its ‘water footprint’?
  1. Generally accepted estimate. Referred to in various sources, including Fujitsu white papers Is the Cloud Green? and ICT Sustainability: The Global Benchmark 2011.
  2. Ibid.
  3. Greenhouse Gas Protocol calculation tools.
  4. Life Cycle Assessment and Product Carbon Footprint: Fujitsu ESPRIMO E9900 Desktop PC, Fujitsu, 2010.
  5. Sthiannopkao S, Wong MH Handling e-waste in developed and developing countries: Initiatives, practices, and consequences, Science of the Total Environment, July 31, 2012.
  6. Statistics on the Management of Used and End-of-Life Electronics, US Environmental Protection Agency, March 2012.
  7. System failure: The UK’s harmful trade in electronic waste, Environmental Investigation Agency, May 2011.
  8. Water cooling vs. air cooling: The rise of water use in data centres,, accessed June 3, 2013.