Investigating the Carbon Footprint of a University - The case of NTNU Hogne N. Larsen a,*, Johan Pettersen a, Christian Solli a, Edgar G. Hertwich b a MiSA AS, Beddingen 14, NO-7014 Trondheim, Norway b Industrial Ecology Programme, Norwegian University of Science and Technology (NTNU), NO-7491 Trondheim, Norway a r t i c l e i n f o Article history: Received 7 March 2011 Received in revised form 5 October 2011 Accepted 6 October 2011 Available online xxx Keywords: Carbon Footprint Environmental extended InputeOutput analysis Public services Universities a b s t r a c t In this paper we apply an Environmental Extended InputeOutput (EEIO) model to calculate the Carbon Footprint (CF) of the Norwegian University of Technology and Science (NTNU). Results show that the CF of NTNU is very significant with an average contribution of 4.6 tonnes per student. In particular, the purchase of large amounts of equipment and consumables for scientific use is found to be an important contributor to this. In the paper we also derive per-department CFs, enabled by the standardized structure of the financial accounting system used by the university. Results show large variations in the CF of the different faculties. Social Science and Humanities have a significantly lower CF per student compared to Natural Science, Engineering, and, in particular, the Faculty of Medicine. The single most important CF contributing input to the university is, however, allocated to the property department regarding the use of energy and other building related activities. Also, the CF structures of the depart- ments/faculties show large differences that indicate that different mitigation strategies are needed. �� 2011 Elsevier Ltd. All rights reserved. 1. Introduction There has been an increasing focus on the evaluation of the environmental performance of businesses, organizations and governmental institutions as a means to focus environmental management efforts and track development over time (Huang et al., 2009 Wiedmann et al., 2009 Berners-Lee et al., 2010 Lenzen et al., 2010). One type of institution where there has been a specific focus on sustainable achievements is universities. This focus is high- lighted by a specific conference (Environmental Management for Sustainable Universities, EMSU (Ferrer-Balas et al., 2010)) and several rankings (e.g., the Engineering Education for Sustainable Development, EESD) on the environmental performance of universities. Most of these initiatives have quite broad scope: the role of universities in creating knowledge, integrating sustainability in educational and research programs, and the promotion of environmental issues to the society (Lozano, 2010 Stephens and Graham, 2010 Waas et al., 2010). However, a few more specific scientific analysis regarding carbon (Baboulet and Lenzen, 2010 Klein-Banai et al., 2010 Thurston and Eckelman, 2011) and ecological (Venetoulis, 2001 Wood and Lenzen, 2003 Conway et al., 2008 Klein-Banai and Theis, 2011) footprinting of universi- ties are available. Furthermore, a wide range of CF inventories for universities, often applying bottom up collection of data in combination with fixed CF intensities from online carbon calcula- tors, have emerged.1 Note that most of these studies are not directly comparable to the one applying EEIO modeling, as they only include selected indirect, scope 3, contributions. The Norwegian University of Technology and Science (NTNU) is the second largest university in Norway. It consists of two main campuses, covering most of the activities, and is located in the city of Trondheim. More than 20 000 students and 5500 employees are divided into seven faculties and 53 departments. As the name indicates, we find a strong focus on science and technological education at this institution. In 2005, the NTNU administration introduced an environmental program based on the ISO 14001 guidelines and identified four target areas: energy, transport, waste and procurement. For the three first target areas, data now exist that can be used for deriving indicators on the environmental performance. For emissions relating to procurement, however, no calculations have been made. There are also other potential gaps in the environmental programs, perhaps the most important of which is the emissions related to buildings (construction, maintenance and other inputs in operating a building, besides energy). One indicator introduced in GHG accounting is the Carbon Footprint (CF) (Wiedmann and Minx, 2007 Weidema et al., 2008 Peters, 2010). As a measure covering all direct and indirect GHG * Corresponding author. Tel.: ��47 91 73 09 52. E-mail addresses: hogne.n.larsen@ntnu.no, hogne@misa.no (H.N. Larsen). URL: http://www.misa.no 1 Due to initiatives like the Americal Colleage & University Presidents��� Climate Commitment. Contents lists available at SciVerse ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro 0959-6526/$ e see front matter �� 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jclepro.2011.10.007 Journal of Cleaner Production xxx (2011) 1e9 Please cite this article in press as: Larsen, H.N., et al., Investigating the Carbon Footprint of a University - The case of NTNU, Journal of Cleaner Production (2011), doi:10.1016/j.jclepro.2011.10.007

emissions, more specifically scope 1 (direct process emissions), scope 2 (indirect emissions from the purchase of energy and scope 3 (other indirect emissions caused by the purchase of goods and services) according to the GHG protocol (WRI and WBCSD, 2004), the CF indicator has proven effective as a suitable measure in a wide range of studies, ranging from global (Peters and Hertwich, 2008 Hertwich and Peters, 2009 Davis and Caldeira, 2010), regional (Peters and Solli, 2010), national (Peters and Hertwich, 2006 Wiedmann et al., 2010) to the sub-national level (Wier et al., 2001 Druckman and Jackson, 2009 Larsen and Hertwich, 2010b Lenzen and Peters, 2010). Environmental Extended InputeOutput (EEIO) modeling (see e.g., Minx et al. (2009)), has proven to be the most promising methodology in calculating CFs on the scales indicated (Peters, 2010). In this paper we investigate the CF of NTNU using EEIO modeling, covering all aspects of the university���s activities. We investigate the total CF, benchmark results, and identify structural contributions to the total. We further investigate the CF by refining the results to the different contributing departments/faculties. The paper starts with a short introduction to the methodology, we then present the results of the analysis, and finally we discuss the main findings and potential for future work. 2. EEIO modeling InputeOutput Analysis (IOA) was introduced in the 1930s (Leontief, 1936). A few decades later work begun on adding envi- ronmental information, both internalized (Leontief, 1970) and as externalities (Ayres and Knese, 1969). Further developments included work by Dantzig (1976) and Miller and Blair (1985). The inclusion of environmental information gave birth to Environ- mental Extended InputeOutput (EEIO) based modeling. Recent developments are numerous on multi-regionality (Peters and Hertwich, 2008 Hertwich and Peters, 2009 Stromman et al., 2009 Tukker et al., 2009 Wiedmann, 2009 Wiedmann et al., 2010), hybridization (Treolar, 1997 Nakamura and Kondo, 2002 Suh et al., 2004 Suh and Huppes, 2005 Stromman and Solli, 2008 Lenzen and Crawford, 2009) and on sub-national levels (Lenzen et al., 2007 Larsen and Hertwich, 2009 Wiedmann et al., 2009 Lenzen and Peters, 2010). A thorough overview of the different IOA applications to environmental analysis is provided by Minx et al. (2009). For the purpose of calculating the CF of a complex enterprise such as a university, EEIO modeling (Munksgaard et al., 2005 Peters and Hertwich, 2008 Hertwich and Peters, 2009 Minx et al., 2009) has proven very useful (Wiedmann et al., 2009 Lenzen et al., 2010). There are several reasons why an EEIO based model was considered an appropriate calculation methodology in this work firstly, the focus on public services mandates a need to take into account non-physical flows. For this purpose standard LCAs are found insufficient (Suh and Huppes, 2002 Junnila, 2006). Secondly, the financial framework applied by governmental entities provides both a detailed and a standardized format suitable for EEIO modeling. Thirdly, EEIO modeling works effectively in providing good quality, timely estimates of reliable accuracy compared to the more detailed, time-consuming, LCAs. While EEIO modeling works very well to provide a complete overview on the contributions of the total CF, it performs less well on detail (in Norway limited to 58 sectors), improvement options (all products within a category are assumed to have identical CF) and deriving time series (although time-efficient to derive, also vulnerable to price variations). Further, most EEIO models are a few years old due to the time-consuming construction of increasingly complex models so that changes in production technology from year to year are not sufficiently captured. Several authors now apply combinations of EEIOAs and LCAs, termed hybrid-LCA (Heijungs and Suh, 2002 Suh and Huppes, 2005), to compensate for the weaknesses of both (Marheineke et al., 1998 Treloar et al., 2000 Stromman et al., 2006 Michelsen et al., 2008 Rowley et al., 2009 Bilec et al., 2010 Mattila et al., 2010). For the EEIO model used for the NTNU study, we hybridized the model for all Scope 1 and Scope 2 GHG emissions (WRI and WBCSD, 2004). The model is similar to that used for municipalities and counties, as described in Larsen and Hertwich (2009), and has further been refined (Solli et al., 2009) for the purpose of assessing the CF of central government entities such as universities. Capital is internalized and import fractions included (represented by the German technology). In total, 120 EEIO sectors are available in the model two scope 1 contributions (combustion of fuel and heating oil), two scope 2 contributions (the purchase electricity and district heating), 58 domestic EEIO sectors and the corresponding 58 EEIO import sectors covering all other purchases of goods and services (Scope 3) using the standardized NACE industry classification.2 Some key elements of the model are summarized in Table 1: The financial account for NTNU from 2009 was used in this analysis. It constitutes of more than 200 categories covering both purchases of goods, services, and investments. All elements that contribute to the CF are listed in Appendix B. For investments, the depreciation values in the accounts were used, obtained by dividing the CF per year equally over the economic life time expectancy of the different products. A key part of applying the model is matching the data from the financial account to the EEIO sectors in Appendix A. In most cases the matching causes no problems (for instance matching the purchase of publications to the ���Y22 publishing, printing and reproduction of recorded media��� EEIO sector, while in other cases more information on the composition of purchasing categories is needed. There is also the possibility of matching one purchasing category to several EEIO sectors. For instance, the demand for most types of equipment is quite aggregated in the financial account and therefore distributed over several EEIO sectors. When the matching is complete we constructed the total demand NTNU has on the economy. The data was then price adjusted to fit the 2005 EEIO data using the consumer price index. The numbers are further converted to basic price to fit the model, and the trade and transport margins (TTM) are distributed to the relevant sectors. Using EEIO modeling to calculate the inter- industry flows (supply chains) in combination with GHG emission intensities for each EEIO sector, we are now able to generate the complete CF of the university. Using the financial account as the Table 1 Key elements of EEIO model used. Key elements Description Comments Year of EEIO data 2005 Number of EEIO sectors 120 (4 �� 58��58) Process �� domestic �� imports Hybridized processes Fuel for transport, heating oil, electricity, district heating GHG gases included CO2, CH4, N2O, CO, HFC, PFC, SF6 For imports only CO2, CH4, N2O Treatment of imports Assumed to be represented by the German economy Other adjustments Price adjustments, Trade and Transport margins, Capital For Capital: only consider depreciation 2 NACE is derived from the French title ���Nomenclature g��n��rale des Activit��s ��conomiques dans les Communaut��s Europ��ennes��� (Statistical classification of economic activities in the European Communities). H.N. Larsen et al. / Journal of Cleaner Production xxx (2011) 1e9 2 Please cite this article in press as: Larsen, H.N., et al., Investigating the Carbon Footprint of a University - The case of NTNU, Journal of Cleaner Production (2011), doi:10.1016/j.jclepro.2011.10.007