The impact of information and communication technology on GHG emissions: how green are virtual worlds?

Abstract

Information and communication technology (ICT) has the potential to either increase or decrease global anthropogenic GHG emissions. Through its various life cycle phases, ICT now represents a rapidly growing source of GHG emissions from the energy and manufacturing sectors, but through its use to avoid emissions in other sectors, in particular those of the transport sector, ICT also has the potential to significantly reduce emissions. Here we explore the potential of so-called ‘virtual world’ technologies to provide a substitute for physical meetings, highlighting the resulting GHG emission savings that can be achieved. These virtual world technologies exemplify the double-edged nature of ICT in climate change mitigation, as their use may generate significant direct emissions through energy use, while, at the same time, reduce emissions from other sources, such as business travel. Actual and potential use, of virtual world technology (VWT) in a nonrandom sample of four large organizations was assessed. For two of these organizations a quantitative assessment of the impact of VWT on net GHG emissions from a specific event was carried out. Subsequent to this, a qualitative assessment of the barriers to increased VWT was made for the other two organizations. We found that, in the two quantitative case studies, VWT was successfully used to substitute physical meetings with virtual ones and that the extent of avoided transport-related GHG emissions was several times greater than the direct emissions generated by the VWT in its use phase. However, our qualitative assessments of current and potential VWT use in the two additional organizations indicated the existence of significant technical and cultural barriers to more widespread use. We conclude that increased use of VWT can potentially lead to significant net GHG emissions savings – primarily via reduced business travel. However, there remain a number of barriers, such as low levels of user engagement, which must be overcome if uptake of VWT at the scale required for globally-significant emissions reductions is to be achieved.

Figure 1.  GHG emissions avoided by virtual world technology (in tonnes CO₂e by source) for the IT and services company conference.

Represents gross estimate, based on equivalent physical conference to the virtual conference.

Figure 2.  GHG emissions avoided by virtual world technology (in tonnes CO₂e by source) for the Imperial College conference.

Represents gross estimate, based on equivalent physical conference to the virtual conference.

The information and communication technology (ICT) sector underpins everything we do as an economy and society [1]. In Europe, the ICT sector contributes 6% of GDP, employs over 6 million people and is achieving annual growth rates of over 3% a year [2]. It is characterized by rapid innovation with a plethora of new products, devices and services being launched regularly.

New technologies have transformed core elements of both UK and global businesses and ICT has become pervasive throughout human society, with the latest estimates indicating that we surpassed one billion installed PCs globally in 2008 [101]. At current growth rates (˜12% per annum) this is set to double to over 2 billion by 2014 [101]. There has been even faster growth in the use of mobile phones and fixed landlines, as well as use of the internet generally. Some 60% of the world’s population now has a mobile phone (4.1 billion devices) and 1.5 billion people already access the internet [102,103].

This phenomenal growth in ICT has led to concerns about the impact it is having on the environment in general [3,4] and, due to ICT’s heavy energy usage, its associated GHG emissions in particular [5]. For instance, it is estimated that GHG emissions from ICT represent 2% of global anthropogenic emissions and that they will reach 3% of global emissions by 2020 [6]. Such growth would make GHG emissions from the ICT sector greater than those from aviation [7]. However, ICT also has the potential to reduce GHG emissions in other sectors, principally by substitution of more emission-intensive activities [4,8]. These include virtualization of products (e.g., CDs to MP3s), digitization of information (e.g., catalogues to websites), avoidance of business travel (e.g., flights to teleconferencing), reduced office space requirements and the shortening of supply chains [9,10]. Indeed, a recent report estimated that ICT can contribute to GHG abatement initiatives that could reduce business-as-usual global GHG emissions by 15% (˜7.8 Gt CO₂e) by 2020, that is approximately five-times greater than the emissions they are predicted to create by 2020 [11]. For these projections the most significant areas for global GHG emission abatement have been identified as being in the use of ICT for smart grids, buildings, logistics and motor systems [12].

ICT & the environment

Several studies have highlighted ‘the three-order effects of ICT’ as a conceptual framework for looking at ICT and its environmental impact [8,13,14]. Within this framework, first-order impacts deal with the direct environmental impacts of ICT use and the production of its supporting infrastructure (e.g., energy use of hardware production and use, and pollution related to disposal of waste hardware). Second-order impacts then deal with indirect impacts of ICT on economic structure, supply chain efficiency and production systems, while third-order impacts cover indirect effects such as lifestyle changes, stimulation of economic growth and the longer-term implications of increased efficiency [8].

A growing number of analyses of first-order impacts have shown ICT hardware releases significant GHG emissions across its life-cycle phases (i.e., raw material extraction, manufacture, use and disposal) [15,16]. However, through ICT’s second-order impacts, such as ‘dematerialization’ and ‘demobilization’, significant opportunities to reduce net GHG emissions also exist.

Reductions in net GHG emissions through the dematerialization of certain goods and services have already been demonstrated [17]. For example, a recent examination of the music delivery sector estimated that digital download of an album resulted in a cut of 40–80% in energy use and GHG emissions compared with its best-case physical equivalent [18]. In using ICT to replace physical travel with teleconferences and virtual meetings (demobilization) the climate change mitigation potential may be still greater, although the extent of this substitution and its impact on net GHG emissions depends on the mode in which it is used, the emission intensity of the ICT system and that of the activity it replaces [19]. A comparison of wireless teleconferencing with the GHG emissions impact of business travel for its physical equivalent reported that the teleconferencing option resulted in emissions one to three orders of magnitude lower than those arising from business travel [17]. As such, the growing availability and use of ICT for virtual meetings and conferences deserves further scrutiny in terms of the additional GHG emissions it creates directly and the emissions reductions it achieves indirectly through the substitution of business travel.

Virtual worlds

In this study we explore an example of the ‘double-edged’ nature of ICT in climate change mitigation by examining so-called ‘virtual world technology’ (VWT). A new and fast-developing technology, use of VWT is growing rapidly around the world [104–106] and requires millions of PCs, an extensive server infrastructure and substantial electricity supplies to support it. As such, a growth in VWT use may be expected to result in increasing GHG emissions. However, the widespread use of VWT also has the potential to radically reduce transport-related GHG emissions, especially those associated with business meetings and conference attendance.

To date, only a limited number of studies have examined the GHG emissions associated with VWT and its use. One recent study calculated that the avatar of a ‘high user’ of Second Life, a VWT used very widely for both leisure and business purposes, consumed some 1752 kWh of electricity per year [107]. This is similar to the 1884 kWh annual energy consumption of an average Brazilian citizen for all purposes. Taken in isolation, such intensive energy use for VWT represents significant related CO₂ emissions – assuming the 2007 US-wide grid emission factor of 0.59 kg CO₂ per kWh [108], the Second Life avatar would be responsible for more than 1 tonne of GHG emissions each year.

Business travel is responsible for 33–40% of global civil aviation traffic [20], with civil aviation as a whole being responsible for approximately 490 million tonnes of CO₂ per year [21]. While other studies have highlighted the potential GHG emission reductions achievable via virtual attendance of conference and business meetings by participants [17,22–25,109,110], none have, to our knowledge, quantitatively assessed the net impact of VWT use on such emissions, and the degree to which direct emission by VWT use may be offset by avoidance of business travel.

The primary objectives of the current study were therefore to:

	▪ Assess the impact of VWT use (specifically that of Second Life) on net GHG emissions for specific conference events, where estimated emissions from the virtual event could be compared with estimated emissions for its real world equivalent;
	▪ Investigate the current utilization of VWT by large institutions, their plans for future use and what barriers to greater use of VWT currently exist in these institutions.

Methods

Use of VWT by both academic and nonacademic organizations (as opposed to individuals) is still at an immature stage and access to meaningful data on ‘virtual conferencing technology’ (VCT) usage for meetings remains limited. Surveys suggest that the majority of business travel is to attend meetings and that, in large organizations with extensive geographical distributions, the majority of business trips are related to collaborations within the company [26,110]. Therefore, we aimed to assess VWT use for meetings by a small number of large (>1000 employees) organizations where business travel volumes are likely to be high. Our initial search for instances of VCT use for specific conference events yielded two cases: an international ‘IT and services’ company and at Imperial College, London, UK, where sufficient quantitative data were available for the required analysis (Second Life was the VWT used in both cases).

For our qualitative assessment of barriers to current and potential VCT use, neither the IT and services company nor Imperial College were deemed appropriate, due to limited access to information and views beyond those relating to the specific VWT-based conferences cases mentioned above. Therefore, two additional large organizations known to use VWT (again, both used Second Life), a global ‘oil and gas’ company and the University of Edinburgh, were selected for the qualitative assessment, based on our securing direct access to personnel who had in-depth knowledge of current and planned VWT use in these organizations.

Although Second Life was the VWT utilized by each of the organizations studied here, it should be noted that many other forms of VWT exist (e.g., Moodle, Blackboard and PlayStation Home) and that the GHG emission impacts of these other VWTs may be very different from those for Second Life.

▪ Quantitative case studies

Virtual conference: IT & services company

On 21 October 2008 this company began a 3-day conference in Second Life for 260 members of its Academy of Technology. The members of the Academy are drawn from the company’s 200,000-strong technical community and are responsible for providing technical leadership. It had become apparent to several members of the academy in 2007 that VWT were becoming much more than an interesting gaming environment or a social networking tool. A decision to mount a conference in Second Life was taken and the company then worked with Linden Labs, the owners of this VWT, to create 16 separate ‘regions’ within this virtual world. Importantly, these ‘regions’ were behind the company’s firewall, so as to create a secure location for conference attendees while maintaining access to other islands within the public VWT environment.

The regions included conference facilities featuring a reception, three theaters, several gardens, support library, green data centre and community gathering places. The conference had three keynote and 37 separate breakout sessions. The session topics were decided by an Academy panel and were selected from a total of 65 submissions.

The total investment cost for the company in creating and running the conference was approximately US$80,000. The company estimate that the business travel and conference venue cost savings they made were approximately $250,000 and that they made an additional productivity gain of approximately $150,000 due to the reduced conference travel time of its employees. However, the company did not assess the net GHG emissions reductions achieved by holding the conference virtually.

GHG emissions assessment

The GHG emissions arising from the IT and services company conference using VWT were estimated based on the GHG Protocol [111], a widely used standard for corporate emissions footprinting, and by using published emission factors for each of the key activities and locations. Under the GHG Protocol guidelines, grid electricity is defined as a ‘Scope 2’ emissions source, while all business travel, including flights, travel by car/taxi and overnight hotel accommodation, are regarded as ‘Scope 3’ emissions. Details of assumed emission factors and activity data are provided below. Estimated emissions for the VWT-based conference were then compared with those estimated for the equivalent physical conference.

Emission factors for air travel [27], travel by car [28], overnight hotel accommodation [27,29] and grid electricity for New York, NY, USA [108], were used in conjunction with estimated travel distances and delegate numbers to calculate the GHG emissions from the virtual conference and its physical equivalent. The global warming potential values used for estimating CH₄ and N₂O emissions as CO₂e were 25 and 298, respectively [112].

Estimations of emission reductions, due to avoided air travel, travel by car/taxi, and overnight hotel accommodation, were made as follows:

	▪ The company supplied data covering the office locations that all attendees at the virtual conference were from (Supplementary Table). It was assumed, based on previous physical meetings, that a suitable venue for the equivalent physical meeting would have been the company’s facility in Somers, NY, USA and that delegates staying overnight would stay at a company-designated hotel in Westchester, NY, USA;
	▪ The avoided air miles from city of departure to JFK or La Guardia airports in New York were calculated using an online air miles calculator [113] for all those attendees that would have taken a flight. Emission factors used for air travel were based on the estimated travel distance (i.e. whether short- or long-haul), with short-haul flights designated as those <1600 miles and long-haul flights being those of >1600 miles. Economy class air travel was assumed for all delegates, based on the company’s latest travel policy guidelines, giving emission factors of 93 and 83 g CO2e per passenger km for short- and long-haul flights, respectively [27]. Note that, were all delegates to travel only by business class, the corresponding emission factor would rise to 140 (short-haul) and 242 g (long-haul) CO2e per passenger km; as such the estimates of GHG emissions from delegate flights provided here are deemed to be conservative;
	▪ For US attendees, who would be expected to drive to the equivalent physical conference (those within 150 miles), distances to the venue were calculated using Google Maps [114]. For those within 100 miles of the venue it was assumed they would drive directly to the venue each day. For those where the distance was greater than 100 miles (but less than 150 miles), the distance from their location and the conference hotel was calculated. Hotel accommodation for 3 nights was assumed for these delegates and a daily 40 mile round trip from the hotel to the conference location was included. Two hotel-to-conference travel scenarios were assessed, one assuming no car-sharing of delegates (‘no car share’) and the other assuming all delegates traveled each day by chartered bus (‘chartered bus’). Car or taxi travel GHG emissions were based on an emission factor of 0.419 kg CO2e per mile (0.262 kg CO2e per km), representing the average passenger vehicle in the US [28]. Chartered bus travel GHG emissions were based on an emission factor of 0.27 kg CO2e per mile per passenger (0.169 kg CO2e per km) [29]. GHG emissions arising from hotel accommodation were estimated using an estimate of emissions per delegate of 32.68 kg CO2e per night [30];
	▪ For those attendees who would have flown to the equivalent physical conference, it was again assumed that they each required hotel accommodation for 3 nights. It was assumed that these attendees would have made a 40 mile return journey by car or taxi from their homes to the airport, an additional journey totaling 80 miles for the taxi to the Westchester hotel and back to the airport in New York and a further 40 mile round trip (either ‘no car share scenario’ or ‘chartered bus’ scenario) to the co-location at Somers for each day of the conference (total per delegate of 240 miles). Emissions arising from this car, bus or taxi travel, and those for hotel accommodation, were estimated as described above;
	▪ To assess the potential impact of a proportion of the delegates who attended the virtual conference choosing not to attend its physical equivalent, we assessed an additional scenario, the ‘avoidance’ scenario, whereby 20% of those delegates required to travel long distance would chose not to attend. All other emission factors and assumptions remained as described previously.

Estimation of VWT emissions

GHG emissions produced by the use of VWT for the virtual conference was estimated using the following data and assumptions:

	▪ The number of dedicated servers used to support the virtual conference was 14 and the power usage in wattage per server, including air conditioning, was 250 W. It was assumed the servers ran for a full 5 days at 24 h per day. It was also assumed that the overwhelming majority of attendees remained within the company’s secure part of Second Life and so utilized only the dedicated servers during the conference. For the electricity used by the servers, the 2007 emission factor for grid electricity in New York city (320 CO2e/kWh) was used [108];
	▪ The number of PCs used was assumed to be 264 (based on the number of attendees) and the power usage per PC was 120 W. Daily usage at an average rate of 12 h per day for the 3 days of the conference was assumed, based on the server and PC usage statistics for the conference provided by company;
	▪ An estimate of the embodied GHG emissions in PCs, servers and programming was not included, as the current assessment was focused on the emissions arising from the use phase. Similarly, the emission calculations for air and car travel, and for hotel accommodation, are for the use phase and do not cover the embodied emissions arising from manufacture or construction;
	▪ Possible energy use within the equivalent physical conference venue (Somers, NY), including that of PC or laptop computer use by delegates, was also excluded.

Note that financial costs for the virtual conference and estimated avoided costs of its physical equivalent were calculated and provided by the IT and services company.

Virtual conference: Imperial College

On 3 December 2008, Imperial College invited scientists to what was claimed to be the world’s first virtual conference on climate change. This 1-day conference took place on the ‘Elucian Islands’, home of Nature Publishing Group in the virtual world of Second Life[115]. The virtual conference was broadcast live in conference rooms at Imperial College and at Stanford University, University of Wyoming, University of Southern California and the University of Texas. Individual attendees included researchers from 11 universities and three research laboratories in the USA and Europe. Again, while this conference was ostensibly held to reduce travel and hence GHG emissions, no analysis of the actual emissions avoided was made at the time.

GHG emission estimates were made using the same methodology as that employed for the IT and services company case study described above, unless otherwise stated.

Emission factors & travel assumptions

Imperial College supplied data covering the number of attendees at various locations. It was assumed that an equivalent physical meeting, had it taken place, would have been at Imperial College. Hotel accommodation for 1 night in a central London location, within walking distance of Imperial College, was therefore assumed for all overseas attendees.

The air miles from city of departure to Heathrow airport, UK, were calculated for all those attendees that would take a flight as described previously. Economy class was again assumed for all flights. It was also assumed that those delegates who would have traveled by plane had a 40 mile return journey by car or taxi from their homes to the airport, but used public transport to get from Heathrow airport to their hotel in central London. The GHG emissions arising from car or taxi travel were estimated as described previously, utilizing the emission factor for an average US passenger vehicle [28]. Emissions arising from the public transport component in London were regarded as negligible and were not included in the calculations. Attendees of the equivalent physical conference that were already based at Imperial College were assumed to have zero travel emissions.

As the virtual conference was streamed live to several universities it could be viewed by people without the need for each of them to create their own ‘avatar’. Thus, the number of individuals attending this virtual conference (62) was actually greater than the number of avatars present (44). Of the avatars who attended the conference, 12 did not register, so details on their physical location were unavailable. It was assumed, therefore, that the number of these that would have traveled to the equivalent physical meeting was in the same proportion as those for which location details were available, that is, 48% would have traveled to the equivalent physical conference. It was also assumed and that the avoided travel-related emissions for these unregistered attendees was the average of that calculated for registered attendees.

Overall, of the 62 people that attended the virtual conference hosted by Imperial College it was assumed that, had an equivalent physical conference taken place, 27 people would have traveled by air to the meeting in London, while the remainder would be able to attend without traveling any significant distance. Again, an additional ‘avoidance’ scenario was assessed whereby 20% (12 delegates) of those required to travel long distance would choose not to attend the equivalent physical conference.

Estimation of VWT emissions

For the electricity used by servers in this instance, the 2007 emission factor for grid electricity in California, (310g CO₂e per kWh) was employed [108]. Three dedicated servers were used to support the conference – two servers to support the ‘Elucian islands’ and one server to host the software that enabled voice communication. Power usage of 250 W per server was assumed, each running for 24 h a day for 2 days. The number of PCs in use was 44 (based on the number of avatars attending the conference as distinct from the number of real people) and the power usage by each PC was 120 W for 5 h per day (including 3 h for presentations and 2 h for poster sessions). Again, only use-phase energy use and emissions were estimated, with embodied emissions of equipment and software development excluded.

Financial costs for the virtual conference were provided by Imperial College. The avoided costs of its physical equivalent were estimated using an assumption that each delegate who would have had to travel to the conference from outside of London would incur travel costs of $850 (based on a return economy air fare for the average ‘visiting’ delegate) plus one night’s hotel accommodation at $100 per ‘visiting’ delegate. We assume that no additional conference fees would have been charged for the physical equivalent compared with the VWT-based conference, and that any related food or drink costs incurred by delegates would have been the same for both the VWT and the physical conference. No financial estimate of additional productivity savings, due to time saved by delegates not having to travel to the conference, was made.

▪ Qualitative case studies

The qualitative methodology employed was principally that of one-on-one semi-structured interviews. The objective of these interviews was to assess current usage of VWT technologies, identify to what extent these technologies are already reducing business travel and to explore the barriers to greater use of VWT in the future.

To assess current usage of VWT by the oil and gas company and the University of Edinburgh a questionnaire was developed and used as a discussion guide in one-to-one discussions with relevant personnel within the two organizations (Supplementary Questionnaire). The questionnaire covered the use of VWT, why they had chosen Second Life and what factors were behind their decision to explore the technology. It also aimed to elucidate the extent to which environmental considerations were a factor, with specific questions on the precise use of VWT in the organization and its frequency. Finally, it enquired as to whether the organization had implemented any formal measurement of savings in terms of either GHG emissions or business travel as a consequence of using VWT, what the benefits and challenges had been in its use, and to what extent did the organization envisage extension of VWT use in the future.

These interviews took place during July and August 2009, either by a teleconference call in the case of the oil and gas company or, in the case of the University of Edinburgh, a meeting in Second Life itself.

Results & discussion

▪ Quantitative case studies

Virtual conference: IT & services company

A total of 264 delegates registered for the virtual conference – 60 from the Americas, 49 from Europe and 55 from Asia (Supplementary Table). For its physical equivalent, and based on their starting locations, it was estimated that 39 delegates would have traveled by short-haul aviation covering a total distance of 45,318 km. A further 181 (59 from within the USA and 122 from overseas) would have traveled by long-haul aviation, covering a total distance of 2,569,554 km.

Estimated delegate car travel and taxi transfers totaled some 99,888 km, with hotel accommodation for all those delegates requiring it being 690 nights in total.

The gross estimate of GHG emissions avoided by the use of VWT, in comparison to the equivalent physical meeting, totaled 282 tonnes CO₂e. Delegate flights constituted the largest source for avoided emissions at 84% of the total. Of these, the bulk (213 tonnes CO₂e) arose from long-haul aviation. Travel by car or taxi represented 14.8% of total avoided emissions (‘no car share’ scenario), while avoided GHG emissions from hotel accommodation comprised 8% of the total (Figure 1).

The substitution of the ‘no car share’ scenario, for hotel to conference venue travel with that of the ‘chartered bus’ scenario, reduced related GHG emissions from 3.85 tonnes CO₂e to 2.48 tonnes CO₂e, a reduction of <1% in overall GHG emissions.

The ‘avoidance’ scenario, whereby 20% of those delegates required to travel long distance to the equivalent physical conference chose not to, yielded a reduction in overall GHG emissions of 55.6 tonnes CO₂e.

Estimation of VWT emissions

The electricity demand and GHG emissions arising directly from the use of VWT (Second Life) for the IT and services company’s virtual conference were estimated using the data shown in Table 1, with PC power representing the bulk of demand and emissions due to the large number of units used.

The bulk of the GHG emissions arising from the use of VWT in this case were due to the energy consumed by PCs (73% of the total), with the balance coming from the servers. Overall then, the GHG emissions arising from this virtual conference constituted less than 0.2% of the emissions estimated for the equivalent physical conference. Per capita, emissions for the 264 delegates attending a physical equivalent of this virtual conference were approximately 1 tonne CO₂e, compared with only 0.002 tonnes CO₂e per capita for the virtual conference (Table 2). The net avoidance of CO₂e emissions achieved by the IT and services company in holding this conference virtually therefore totaled 281.5 tonnes.

As stated previously, the company had calculated that its costs for running the virtual conference were $80,000 (which included set-up costs but excluded on-going maintenance costs of the server infrastructure). This compared with estimated travel costs of $250,000 that would have been incurred by the physical conference, and a further estimate of $150,000 in additional productivity savings, due to time saved by those employees not having to travel to the conference.

Virtual conference: Imperial College

Of the 62 delegates for the Imperial College, virtual conference, the 27 delegates who would have had to fly to the physical equivalent were predominantly from North America and would have required long-haul flights (Table 3).

The gross estimate of GHG emissions avoided by the use of VWT, in comparison to the equivalent physical meeting in this case, totaled 34.8 tonnes CO₂e. This is a much lower figure than that estimated for the IT and services company conference and reflects the smaller number of participants. Delegate flights again constituted by far the largest source for avoided emissions, at more than 95% of the total. This very high percentage stems from the conference being of short duration (requiring only one night’s accommodation for traveling delegates) and from the fact that many of the delegates were already based in the host location (Imperial College) and so their car or taxi travel emissions were effectively zero. Travel by car or taxi therefore represented just 1.4% of total avoided emissions, while avoided GHG emissions from hotel accommodation comprised 2.8% of the total (Figure 2).

The ‘avoidance’ scenario, whereby 20% of those delegates required to travel long distance to the equivalent physical conference chose not to, yielded a reduction in overall GHG emissions of 6.9 tonnes CO₂e.

Estimation of VWT emissions

The electricity demand and GHG emissions arising directly from the use of VWT (Second Life) for the Imperial College virtual conference were estimated using the data shown in Table 4. Unlike the IT and service company VWT conference, server power here represented more than 50% of demand and emissions, with PCs representing 42%. This reflects the base-level requirement for running a virtual conference with this VWT (Second Life), irrespective of the number of user PCs.

Overall, the GHG emissions arising from this virtual conference (less than 20 kg CO₂e) constituted less than 0.1% of the emissions estimated for the equivalent physical conference, much less than that estimated for the IT and services VCT conference and primarily a result of lower user numbers and a shorter conference duration.

Per capita, emissions for the 62 delegates attending a physical equivalent of the virtual conference were approximately 0.56 tonnes CO₂e, compared with 0.0003 tonnes CO₂e per capita for the virtual conference (Table 5). Although the per capita emission estimate for all delegates attending the equivalent physical conference was 0.56 tonnes CO₂e, more than half of the attendees were from London, and so their associated conference travel was effectively nil. Therefore, we have also calculated the avoided emissions per ‘visiting’ delegate at 1.29 tonnes CO₂e.

The net avoidance of GHG emissions achieved by the IT and services company in holding this conference virtually therefore totaled 34.8 tonnes, with use of VWT in this instance effectively negating all the GHG emissions that would have occurred had the equivalent physical meeting been held.

As stated previously, Imperial College had calculated that its costs for running the virtual conference were approximately $5000. This compared with our total estimated travel and accommodation costs of $25,650 that would have been incurred by the 27 ‘by air’ attendees of the equivalent physical conference. Clearly, these estimated costs would likely be borne by the individual delegates rather then the conference organizers and so, in this case especially, financial costs may represent a significant disincentive to attendance of the physical conference by overseas delegates.

▪ Qualitative assessment of VWT usage

Current & planned use of VWT by an oil & gas company

Current and planned usage of VWT by this company was assessed using one-on-one semistructured interviews. This company stated that it was already using “a rich array of technologies across its divisions, divided into two broad areas of collaboration tools, both synchronous and asynchronous” [116].

Synchronous tools enable real-time communication and collaboration in a same time–different place mode. These tools allow people to connect at a single point in time. Asynchronous tools enable communication and collaboration over a period of time through a different time–different place mode. These tools allow people to connect together at each person’s own convenience and own schedule.

Connection tools used by the company currently include mobile phones, web conferencing, live meetings and video conferencing facilities to support collaboration across the group’s activities. Asynchronous tools include discussion forums, team wikis and blogs. This already rich provision of collaboration tools could help to facilitate the further use of VWT by the company.

The company began to consider use of VWT in late 2007. They had a specific interest in using VWT in three specific areas: intercompany meetings, conferences and learning. All had the common aim of reducing business travel expenses. The company had identified that, “a significant proportion of our travel costs are in moving our own personnel to and from our training and learning centres. We have a long-term workplace learning strategy that aims to bring the training to the learner and not have the learnee having to travel.” As a result, it has a long-term workplace learning strategy that aims to provide ‘learning to the learner’ and avoid learners having to attend a learning centre.

Environmental considerations had some impact on their decision to explore VWT. Senior management were keen to both reduce travel costs and the associated GHG emissions, although use of VWT was perceived by the business as being aimed primarily at the expense issue. A key consideration made was that, whatever technology was chosen, easy replacement by other VWT systems should be possible if required. The company would also wish to be able to extract any ‘data assets’ they had input into the chosen VWT.

In part, the company’s interest in VWT was driven by a belief that it could provide real value through its experiential learning and user safety aspects, which were not available with other technologies. The example quoted was of the handling of sour gas, a form of natural gas with high concentrations of hydrogen sulfide [117]. This is a byproduct of oil and gas exploration and at high concentrations is lethal to humans. To date, training the company’s engineers in the proper procedures for handling sour gas has been both expensive in time and travel. It is also very difficult and dangerous to do in a live setting. Use of VWT can replicate much of the experiential learning (i.e., what not to do in a virtual oil field setting) without human health risks. Reduced travel costs and related GHG emissions in this case are an additional benefit.

In terms of the other two areas of focus (inter-company meetings and conferences) the company found that meetings using VWT did not compete well with existing technologies, such as web and video conferencing, or with face-to-face meetings in terms of user engagement, stating, “if you are bored being in a video conference then you will just as easily get bored in a virtual conference.” The company also recognized that part of the value of conferences is the informal networking between colleagues and so, to date, are unsure how such informal networking can be properly replicated in a virtual world. As such, the company does not envisage much future application of VWT in these areas.

However, for conferences they are seeing benefits in what they call ‘extended conferences’. Here a physical conference format is still used, but the presentations are also streamed by live video into a virtual conference centre at which a much larger global audience can then attend via their office PCs. This has significant advantages over the alternative of streaming the video to their existing video conference centres, as it requires no travel for participants to the video conferencing centres and requires no scheduling of video conference suites. It does require the creation of secure conferencing facilities within the virtual world, but once created these facilities are reusable.

Overall, this oil and gas company see most opportunities for VWT in the employee learning area, leveraging its ‘serious gaming’ aspects. Extensions of existing conferences to add more value and reduce travel costs and GHG emissions are also viewed positively. There is an emerging view within the company that even when the current economic crisis is over, higher oil prices and concern over the company’s GHG emissions will ensure that the severe travel restrictions on all business travel currently in place will remain. In that context, greater use of VWT to replace a significant amount of in-house learning events represents an opportunity to both meet the learning strategy and reduce travel costs and GHG emissions.

A key barrier to increased use of VWT by this company is that of technology. They would like to bring Second Life behind the company’s firewall for reasons of security, allowing them a private space to use it that is separate from the more public space of Second Life today. However, they face many technical challenges in doing this, ranging from differences in protocols through to making voice-over internet protocol (VOIP) technology work on the company’s legacy PC infrastructure. Such technological barriers may be preventing many large businesses from being able to deliver a scalable, robust VWT infrastructure, which can then deliver the promise of an improved learning environment at lower travel costs and with reduced GHG emissions.

Development of VWT within this company therefore remains at the ‘proof of concept stage’ and as yet they have no specific measures in place for tracking actual business travel savings or indeed the associated GHG emission savings associated with VWT use.

▪ Current & planned use of VWT by the University of Edinburgh

The current and future use of VWT by the University of Edinburgh was assessed by a one-on-one semi-structured interview in Second Life with their ‘Virtual Worlds’ advisor. The University of Edinburgh began using VWT in 2006 and, along with many other educational institutions, made Second Life their virtual world of choice. They began with an acre of virtual land on ‘Campus Island’ before building an Island named ‘Holyrood Park’ in 2007. It is used to support 130 students who are studying for an MSc in e-learning [118]. This is a distance learning course and its student cohort is comprised of teachers from all levels, independent consultants and academics. The technology provides a virtual meeting space where the students feel present and embodied – evidence from student feedback indicates that this sense of presence is a major issue for online distance learners and one of the key benefits they get from using VWT.

Another aspect of VWT that is relevant for this specific MSc course is the topic of online identity. This can be readily explored in Second Life by giving students an opportunity to experiment with their choice of Avatar when they change age, gender and race. Second Life is not the only VWT in use at the University of Edinburgh. Technologies such as ‘OpenSim’, ‘There’ and ‘Active Worlds’ are each used for the purpose of learning, teaching, research and collaboration. Virtual worlds are also used as a simulation tool, as an exhibition space, as a learning space and as an assessment space.

The total number of VWT users at Edinburgh is currently approximately 350, a sizeable VW community, but still small in relation to the university student population of approximately 25,000. While several of the above activities do represent opportunities for reduced travel and associated GHG emissions, these are not their principle aims. As such, this university is currently making no specific measurements of travel expense or emissions savings due to the use VWT.

Despite the lack of a GHG emissions focus in the current use of VWT within the university there is a very strong emphasis on climate change mitigation across the university, with a commitment to both measure its carbon footprint and to take steps to reduce its overall GHG emissions in the future.

Future plans for the increased use of VWT at the University of Edinburgh centre on its use to support online distance learning for students, with considerable investment currently being made in the conversion of traditional Edinburgh-based courses to e-learning equivalents. The primary incentive is that of increasing access to potential students who are unable to reside in Edinburgh for their studies, but the potential GHG emissions reductions achievable through this modal switch to online distance learning may also be of significant interest to this and other large higher education institutions. It should be noted that an increase in distance learning using VWT may not result in a decrease in absolute GHG emissions. Where online learning provision leads to additional students rather than substitution of on-campus students, overall GHG emissions would increase.

▪ Quantitative case studies

GHG emissions & financial benefits

We have demonstrated that use of VWT (in this case Second Life) has the potential to radically reduce the GHG emissions associated with business travel for meetings and conferences. Even without a marketized price for GHG emissions, such virtual meetings also have the potential for significant cost savings.

In the case of the IT and services company, the use of a virtual rather than physical conference was estimated to have resulted in a significant return on their $80,000 investment. This return included hard travel savings of at least $250,000 and soft savings in estimated productivity increases of another $150,000. This level of return excludes the additional revenue this company may accrue in service fees for assisting other companies to build VWT behind their own firewalls.

The current study also suggests a paradox, in that acceptance of VWT into mainstream use was assumed to be by overcoming their perception as ‘games’ or entertainment platforms and yet, in both the oil and gas company and Imperial College cases, it is the ‘serious gaming’ aspect of the technology (e.g., an in-built intelligent game engine), which holds the most promise for future VWT use and development.

In both of the virtual conferences studied here, the avoided GHG emissions were several orders of magnitude greater than the emissions caused by the use of the VWT itself. In the case of Imperial College, while the overall emissions savings are more modest relative to those for the IT and services company, the estimated emissions saving per ‘visiting delegate’ was 1.29 tonnes CO₂e and so comparable to that for the IT and services company virtual conference (1 tonne CO₂e per delegate). Clearly, in making a direct comparison between a virtual conference and its physical equivalent, the assumption that exactly the same delegates would attend both is significant – the very fact that the virtual conference avoids the need for long-distance travel may serve to encourage overseas participants who would not otherwise have attended. This issue is of particular importance for the Imperial College case study, where the additional costs of travel and accommodation would likely be borne by the delegates themselves.

Our simplistic ‘avoidance’ scenario highlights the large impact any significant decrease in overseas or long-distance delegate number would have on overall GHG emissions from an equivalent physical conference, when compared with a virtual one. However, the reduction in delegate numbers would have to be >95% in both case studies examined here before the associated GHG emissions would fall to levels of similar magnitude to those of the virtual conferences.

Similarly, our exclusion of embodied emissions in ICT equipment and software development for the virtual conferences, and in buildings, furniture, presentation equipment and staffing for their physical-world equivalents, overlooks potentially very large sources of GHG emissions. However, for the more simplistic analysis of direct energy usage by VWT conferences and avoided direct transport emissions for conference delegates employed here, the potential of VWT to reduce GHG emissions is clear.

A key factor underlying the differences in GHG emissions identified here for VWT-based conferences versus their physical equivalents was that of air-travel. It is interesting, albeit somewhat speculative, to examine how a change in the assumed mode of transport may affect this difference. For the IT and services company VWT conference, for example, a physical equivalent where those participants could have (i.e., those within 1600 miles of the venue) utilized intercity rail travel, rather than air travel, would alter the corresponding emission factor for their travel from 93 g CO₂e per passenger km (economy-class short haul aviation) to approximately 116 g CO₂e per passenger km (intercity rail [29]). Such a change in mode would therefore tend to increase, rather than decrease, the extent of avoided GHG emissions resulting from VWT use. Clearly, were these delegates to switch from business-class air travel to train travel, then the resulting change in emission factor (from 140 g to 116 g CO₂e per passenger km) is likely to result in decreased emissions. The class of travel, as well as the mode, is therefore an important consideration in making such comparisons. For car and taxi travel, another significant source of avoided GHG emissions, the assumption of ‘no car sharing’ is likely to be an overly pessimistic one, as some delegates traveling from the same starting locations may well share a car to reduce costs. However, our comparison of the ‘no car share’ and ‘chartered bus’ scenarios for hotel to conference venue travel indicate that the reduction in total GHG emissions achieved by the latter would be relatively minor (<1% of overall GHG emissions arising from the physical equivalent of the IT and services company virtual conference).

These results extend beyond previous analyses that focus on very heavy usage of Second Life[107]. They indicate that the use of VWT can radically reduce GHG emissions associated with conference travel and attendance, implying that the more widespread substitution of business travel by VWT has potential as a significant emissions mitigation strategy for businesses and for the global transport sector as a whole. However, direct extrapolation of the two quantitative VWT (both Second Life) case studies reported here to all organizations, large and small, would not be appropriate, as the precise impacts on net GHG emissions are likely to vary greatly between organizations, mode of VWT use and the activities that VWT is substituting. Similarly, all of the cases examined here utilized Second Life as their VWT, but there are numerous other forms of VWT (e.g., Moodle, Blackboard and bespoke in-house VWTs) that organizations may make use of. The net GHG emission impacts arising from use of these different VWTs will inevitably vary and so case-specific assessments are required if precise impacts are to be quantified.

▪ Qualitative case studies

Barriers to widespread uptake identified in this study centre on access to the required VWT, its alignment with the needs of the specific business or institution and on the quality of the user experience – potential loss of the informal networking benefits of physical meetings remains a significant obstacle in our view. Incentives for increased use of VWT are arguably the most important determinant of future uptake. While the GHG mitigation benefits appear substantial, travel cost and time savings tend to be the leading incentives for VWT use at the current time. For specific uses, such as training and distance education, the experiential learning capabilities of VWT hold great promise for both educational institutions and businesses [31].

One key additional barrier to more widespread use of VWT and other ICT substitutions for conference travel may be that of frequent flying having become normalized for many institutions and individuals [32]. Where physical attendance at a meeting or conference and any associated air travel are seen as ‘perks’ by attendees, resistance to substitution by VWT is likely to be very high. Indeed, recent surveys of frequent flyers indicate that some business trips are made because of the related leisure activities rather than for the meeting itself [32]. In such cases VWT would fail to provide an attractive substitute.

In the four cases studied here, VWT was being used differently and had different drivers for its use. None cited GHG emission reduction as the primary driver, although for the Imperial College case study their virtual climate conference did have GHG emissions savings as one of its objectives.

For the oil and gas company, reductions in both business expense and GHG emissions were implicit in their use of VWT. It was also apparent that, for the two large businesses studied here, business travel expense has become an area of concern for their senior management and that they are actively introducing polices to reduce travel costs. This growing concern reflects the finding by Arnfalk [25,107] that internal meetings (including training) form a large part of business travel expense in large global corporations and that ways must be found to reduce unnecessary travel. The critical issue, of course, is that just because a particular piece of technology, in this case VWT, has the potential to avoid business travel on a large scale, does not in and of itself mean that it will.

Further research is required to better quantify the net impact of VWT on GHG emissions in other situations (e.g., for smaller organizations). Assessment and comparison of embodied GHG emissions, although challenging, will also help elucidate the true climate change mitigation potential of VWT and ICT more generally. Additional research is also required to account for future changes in the emissions intensity of key variables, such as electricity generation and transport modes. For instance, were grid electricity emission factors to radically decrease in the coming years but transport emission factors remain the same, then the GHG emission benefits of travel substitution by VWT are likely to increase.

Conclusion

The substitution of business travel by the provision of virtual meetings appears a fruitful area for climate change mitigation efforts, given that business travel is a key driver of aviation emissions growth and contributes disproportionately to the economics of modern air travel. We find that, while there is very strong evidence that VWT can deliver significant travel and GHG emissions savings, its broad scale take-up and use faces significant hurdles. Barriers to its greater use centre on technical, cultural and organizational issues, with additional research into the way ICT is currently deployed, used and integrated both in the ‘enterprise’ and ‘personal’ contexts being needed.

The very pervasiveness of ICT requires that we better understand its impact on human behaviour both inside and outwith the workplace. We also require greater insight into how the often highly structured IT environments within enterprises can be made to work better with the slightly unstructured way staff and consumers increasingly use ICT outside the workplace. In our view it is in a better understanding of these third-order effects [10,119] that the undoubted potential of VWT to make a positive contribution to resolving the climate change challenge can be made real, and its current ‘double-edged’ nature resolved.

Future perspective

The technical, cultural and organizational barriers to using VWT within businesses are not trivial. On the one hand, the innovation that created VWT has led to their rapidly growing use in the ‘personal’ world of computing, as have many of the most recent innovations, for example, Google Earth. On the other hand, leveraging these tools inside a corporation’s existing IT infrastructure presents many technical obstacles, as this study and other studies have highlighted. Many of these obstacles stem from the organizations own infrastructure, IT policy and architecture.

Historically, IT was seen as a great enabler that created efficiencies and made things happen faster. Today, many major innovations and initiatives in companies are slowed down or stopped because of the technological barriers to implementation. A ‘business as usual’ approach is therefore unlikely to lead to a rapid and widespread increase in the use of VWT or other innovative new technologies within businesses.

A more radical reappraisal of the way IT is used in the ‘personal’ and within the ‘enterprise’ context is required. In our view, these two worlds should be brought closer together – the ‘gaming’ aspects of the former being more readily adopted by the latter. In doing so we can perhaps realize the cost and environmental benefits that ICT can offer more rapidly and more sustainably. For example, the IT and services company’s use of voluntary virtual community to learn about VWT has enabled them to develop expertise in this area much faster than other IT companies, and so take a leadership position.

While this study has shown that use of VWT is still at an early stage in the organizations assessed, tangible benefits in terms of reduced travel costs and GHG emissions have been achieved through its use. Our study supports previous research suggesting that aspects of VWT use, such as providing a sense of presence for users and enabling experiential learning without real world consequences, may be an important advance on other available distance learning technologies [120]. Many of the barriers to its wider use, such as the steep learning curve in using the technology, may disappear over time as more of the ‘digital age’ generation permeates the workforce, but wherever business travel is still seen as a ‘perk of the job’, substitution by VWT is likely to face continued resistance.

Table 1.  Power demand (kWh electricity) and GHG emissions (tonnes CO2e) arising directly from the IT and services company’s virtual conference in Second Life.

Hardware type	Number of units	Power demand (W per unit)	Hours used per day	Days used	Total power demand (kWh)	GHG emissions (tonnes CO2e)
Servers	14	250	24	5	420	0.134
PCs	264	120	12	3	1140.5	0.365
Total					1560.5	0.499

Table 2.  Relative total and per capita GHG emissions (CO2e) from the virtual and equivalent physical IT and services company conference.

GHG emissions source	Number of delegates	Total GHG emissions (tonnes CO2e)	GHG emissions per delegate (tonnes CO2e)	Estimated cost (US$)
Virtual conference	264	0.499	0.002	80,000
Equivalent physical conference	264	282	1.07	250,000 (travel only)

Table 3.  Breakdown of home locations of delegates, avatar numbers and estimates of number of delegates that would be required to fly to the equivalent physical conference and the resulting roundtrip flight distances to London Heathrow.

Home location	Number of delegates	Avatars	Number of delegates required to travel by air	Departure airport	Cumulative flight distance: long-haul (miles)	Cumulative flight distance: short-haul (miles)
Wyoming	2	1	2	Laramie	17,920
Stanford	2	2	2	Pal Alto	21,400
USC	2	1	2	Los Angeles	21,600
UT	8	3	8	Austin	78,400
Imperial	24	15
Albuquerque	1	1	1	Albuquerque	9960
Heidelberg	1	1	1	Frankfurt		812
Toronto	1	1	1	Toronto	7080
Alamo California	1	1	1	San Francisco	10,700
Helsinki	1	1	1	Helsinki	2300
Vancouver	1	1	1	Vancouver	9420
New York	3	1	1	New York	20,640
London	2	2
Cambridge	1	1
Nonregistered	12	12	6	N/a	50,058
Total	62	44	27		249,478	812

USC: The University of Southern California; UT: The University of Texas.

Table 4.  Power demand (kWh electricity) and GHG emissions (tonnes CO2e) arising directly from the IT and services company’s virtual conference in Second Life.

Hardware type	Number of units	Power demand (W per unit)	Hours used per day	Days used	Total power demand (kWh)	GHG emissions (tonnes CO2e)
Servers	3	250	24	2	36	0.011
PCs	44	120	5	1	26.4	0.008
Total					62.4	0.019

Table 5.  Relative total and per capita GHG emissions from a virtual and equivalent physical conference hosted by Imperial College.

GHG emissions source	Number of delegates	Total GHG emissions (tonnes CO2e)	GHG emissions per delegate (tonnes CO2e)	GHG emissions per ‘visiting delegate’ (tonnes CO2e)	Estimated cost (US$)
Virtual conference	62	0.019	0.0003	0.0003	5000
Equivalent physical conference	62	34.8	0.56	1.29	25,650

Information and communication technology

Comprises computer hardware (e.g., personal computers and servers), software and telecommunications equipment (e.g., handsets, network equipment and routers).

Dematerialization

Substitution of material infrastructure, services or physical products with information goods and virtual services, also known as ‘virtualization’ (e.g., substitution of music CDs by MP3 downloads).

Demobilization

Substitution of distance communication for travel; for example, replacement of a physical meeting and its associated transport use with a virtual meeting using the internet.

Virtual world technology

Technologies that employ computer-based environments to simulate a user’s physical presence in real and fictional worlds. Commonly used for virtual meetings, educational purposes and online gaming.

Avatar

Graphical representation of a computer user, often in 3D form in computer games and in online virtual world environments.

Emission factor

Measure of the amount of GHGs discharged into the atmosphere as a result of a specific process or activity (e.g., 0.5 kg CO₂ emitted per kWh electricity use). Commonly used in estimation of emissions budgets reported to the UNFCCC and often based upon default emission factors published by the IPCC.

Global warming potential

Climate forcing resulting from a mass unit of a GHG, relative to that of a mass unit CO₂ over the same time-horizon (usually 100 years), where the global warming potential of CO₂ is 1. Used to express a unit of any GHG in CO₂ equivalents.

Executive summary

Information & communication technology & the environment

	▪ The information and communication technology (ICT) sector is a key contributor to GDP growth and job creation, and a major enabler of improved efficiency and climate change mitigation opportunities across the economy.
	▪ The success of ICT and its pervasive presence in our lives has led to concerns about its environmental impact, particularly the GHG emissions associated with production, use and disposal, estimated at 2% of global anthropogenic GHG emissions and growing rapidly.

Virtual worlds

	▪ Through the increased use of ‘virtual world technology’ (VWT) there is the potential to ‘dematerialize’ many high-emission aspects of our lives, such as the substitution of conference and business travel by virtual meetings
	▪ VWT may also significantly increase GHG emissions through increased energy demand for servers and PCs.

Quantitative case studies

	▪ In two case studies of VWT use for large-organization conferences, avoided GHG emissions from conference travel far outweighed the direct emissions arising from the use of the VWT.
	▪ Delegate air-travel represented the dominant source of avoided GHG emissions through the use of VWT.

Qualitative case studies

	▪ Semi-structured interviews with personnel in two large organizations indicated that, while VWT offers a climate change mitigation opportunity, several barriers exist to its wider use.
	▪ Key barriers identified included the bespoke technical needs organizations have for VWT and the perception that VWT provides less ‘agency’ than face-to-face meetings.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript.

Supplementary data

To view the supplementary data that accompany this paper please visit the journal website at: http://www.future-science.com/doi/suppl/10.4155/CMT.11.62.