Atmospheric Environment 43 (2009) 6260–6270 Contents lists available at ScienceDirect Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv Global temperature change from the transport sectors: Historical development and future scenarios Ragnhild Bieltvedt Skeie, Jan Fuglestvedt*, Terje Berntsen, Marianne Tronstad Lund, Gunnar Myhre, Kristin Rypdal CICERO – Center for International Climate and Environmental Research – Oslo, P.O. Box 1129 Blindern, N-0318 Oslo, Norway a r t i c l e i n f o a b s t r a c t Article history: Received 30 January 2009 Received in revised form 13 May 2009 Accepted 14 May 2009 Transport affects climate directly and indirectly through mechanisms that operate on very different timescales and cause both warming and cooling. We calculate contributions to the historical development in global mean temperature for the main transport sectors (road transport, aviation, shipping and rail) based on estimates of historical emissions and by applying knowledge about the various forcing mechanisms from detailed studies. We also calculate the development in future global mean temperature for four transport scenarios consistent with the IPCC SRES scenarios, one mitigation scenario and one sensitivity test scenario. There are large differences between the transport sectors in terms of sign and magnitude of temperature effects and with respect to the contributions from the long- and shortlived components. Since pre-industrial times, we calculate that transport in total has contributed 9% of total net man-made warming in the year 2000. The dominating contributor to warming is CO2, followed by tropospheric O3. By sector, road transport is the largest contributor; 11% of the warming in 2000 is due to this sector. Likewise, aviation has contributed 4% and rail w1%. Shipping, on the other hand, has caused a net cooling up to year 2000, with a contribution of 7%, due to the effects of SO2 and NOx emissions. The total net contribution from the transport sectors to total man-made warming is w15% in 2050, and reaches 20% in 2100 in the A1 and B1 scenarios. For all scenarios and throughout the century, road transport is the dominating contributor to warming. Due to the anticipated reduction in sulphur content of fuels, the net effect of shipping changes from cooling to warming by the end of the century. Significant uncertainties are related to the estimates of historical and future net warming mainly due to cirrus, contrails and aerosol effects, as well as uncertainty in climate sensitivity. Ó 2009 Elsevier Ltd. All rights reserved. Keywords: Transport sectors Radiative forcing Temperature Historical emissions Scenarios 1. Introduction The impact of emissions from the transport sectors on climate has recently received increasing attention. Initially, the focus was on the aviation sector because its emissions are deposited at high altitudes initiating complex mechanisms in a chemically sensitive part of the atmosphere (Schumann, 1990; Friedl, 1997; Minnis et al., 1998, 1999; IPCC, 1999; Myhre and Stordal, 2001; Travis et al., 2002; Stevenson et al., 2004; Sausen et al., 2005; Stuber et al., 2006). Attention then turned toward the shipping sector (Corbett and Fischbeck, 1997; Capaldo et al., 1999; Lawrence and Crutzen, 1999; Endresen et al., 2003, 2007; Eyring et al., 2005, 2007). Many of the studies of impacts from various transport sectors focus on the chemical responses and changes in atmospheric burdens (Granier * Corresponding author. E-mail address: j.s.fuglestvedt@cicero.uio.no (J. Fuglestvedt). 1352-2310/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2009.05.025 and Brasseur, 2003; Niemeier et al., 2006), and some have considered the radiative forcing resulting from these burden changes (Eyring et al., 2007; Lauer et al., 2007; Fuglestvedt et al., 2008). In this paper we calculate the change in global mean temperature due to emissions from all the main transport sectors (road transport, aviation, shipping and rail) and we consider both emissions and responses from pre-industrial times up to present and scenarios for future emissions. There are four main mechanisms by which emissions from transport affect climate: (i) by emission of direct greenhouse gases, mainly CO2, (ii) by emission of indirect greenhouse gases, i.e. precursors of tropospheric O3 or gases affecting the oxidation capacity of the atmosphere, such as NOx, CO and VOC, (iii) by the direct effect of emission of aerosols or aerosol precursors, in particular black carbon (BC), organic carbon (OC), and sulphur compounds, and (iv) by the indirect effect of aerosols, which trigger changes in the distribution and properties of clouds. In addition, water vapour from aircraft engines triggers formation of contrails R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 and cirrus clouds, with the latter effect being strongly dependent on atmospheric conditions. Thus, there is not only a broad mix of species emitted from the transport sectors, but also a broad set of mechanisms by which climate is affected. These mechanisms operate on very different timescales (hours to centuries) and cause both warming and cooling. Previous studies focusing on radiative forcing from historical and current emissions have shown that (i) the shipping sector has had a cooling effect on global climate (Lauer et al., 2007; Fuglestvedt et al., 2008), that (ii) aviation causes a stronger warming effect than might be expected from its total emissions as a result of the many additional potentially strong effects (e.g. formation of contrails and cirrus) (Penner et al., 1999; Sausen et al., 2005), and that (iii) road transport is the dominating sector in terms of total net forcing (Fuglestvedt et al., 2008). In this paper we aim to integrate and apply knowledge about the various forcing mechanisms from detailed studies in order to calculate the historical development in global mean temperature based on estimates of historical emissions of a suite of gases, aerosols and aerosol precursors. We also calculate the development in future global mean temperature for four transport scenarios consistent with the IPCC SRES scenarios (both excluding and including future effects of historical emissions), one mitigation scenario for aviation (ACARE) and a sensitivity test scenario. We calculate the total net contribution from transport as a whole to historical, current, and future warming. The contributions from the various emissions and radiative forcing mechanisms are also quantified, as are the effects of uncertainties in the chain from emissions to temperature response. 2. Data and methods 2.1. Emissions Current and future emissions for road transport, rail, shipping and aviation are based on the QUANTIFY (www.pa.op.dlr.de/ quantify) emission inventory and scenarios, while the historical emissions are supplemented by results from previous studies (see Table 1 in supporting information (SI) for a summary of the references for the transport emissions). Future emissions until 2100 were compiled consistent with the four SRES storylines A1, B1, A2 and B2 (Nakicenovic et al., 2000). The emissions scenarios take into account growth in activity, changes in emission factors/fuel efficiency, and vehicle/fleet turnover consistently with each of the storylines. The modelling has been more rigorously carried out for 2050 compared to 2100 and shortterm projections reflect current legislations. Exact technologies and fuel mix for 2100 are difficult to assess for the given storylines and the estimates rely on a number of assumptions concerning the development and implementation of new technologies and the future fuel mix. Consistently with the SRES thinking, none of the scenarios assumes explicit mitigation to reach a climate target. 2.1.1. Emission references and methodology Current (year 2000) emissions from road transport are described in Borken et al. (2007). Past road CO2 emissions are taken from Fuglestvedt et al. (2008). Emissions of non-CO2 gases are estimated from their emission factors for pre-control technologies and historical CO2 emissions. Future emissions from road transport are taken from Borken-Kleefeld et al. (in preparation). Current and future emissions from aviation are taken from Owen et al. (2006). CO2 emissions prior to 2000 are estimated from energy statistics from the International Energy Agency consistent with Fuglestvedt et al. (2008). Emissions of non-CO2 gases before 1990 are from Lee et al. (2009). 6261 The historical and current emissions from shipping are taken from Endresen et al. (2007). The methodology used to estimate future shipping emissions is described in Eide et al. (2007). Emissions from inland waterways (Borken-Kleefeld et al., in preparation) have been added to the maritime shipping emissions. It is assumed that the past trend in emissions from inland waterways follows the maritime shipping emission trends. Emissions from rail are taken from Borken-Kleefeld et al. (in preparation) and Borken-Kleefeld (personal communication, 2009). Prior to 1970, the emissions are scaled with the CO2 emission from rail from Fuglestvedt et al. (2008). Emissions from electricity generation are included in the railinventory. Current emissions were estimated consistent with EDGAR (Olivier and Berdowski, 2001). The emissions are scaled with the electricity consumption for past years, which is identical to the data used in Fuglestvedt et al. (2008). Year 2000 emissions from electricity production in Olivier and Berdowski (2001) are scaled to 2050 and 2100 based on Nakicenovic et al. (2000). For both years we calculate the fraction of total regional electricity production used by rail and apply this fraction to the power generation emissions in 2050 and 2100. Finally we correct for energy efficiency and technology improvements based on specific model data from the IMAGE model (Bouwman et al., 2006). A scenario is developed which is consistent with the strategic research agenda of the Advisory Council for Aeronautics Research in Europe (ACARE) (www.acare4europe.org) targets for greening of air transport. It is described in SI Section 1 and some results are presented in Section 4.2.3. Future transport is expected to introduce biofuels, hydrogen fuel cells and electric powered vehicles. While direct emissions from these technologies/fuels are small or zero, emissions at the production stage still remain and can be higher than those from the use of fossil fuels in the vehicles. SI Section 2 describes the quantification of the indirect emissions. The climate impact of these indirect emissions is calculated in addition to the scenarios. 2.1.2. Historical and future development of transport emissions Fig. 1 shows the past and future development of emissions of CO2, NOx and SO2 for the different transport sectors and SRES scenarios. CO2 emissions from the transport sectors increased during the 20th century (Fig. 1A), and road transport contributed w16% to the total anthropogenic CO2 emissions (excluding land use change) in 2000 (Fig. 1B). The shipping and aviation sector each contributed w3% to the man-made CO2 emissions in 2000, while the contribution from rail was only 1%. Future CO2 emissions In all scenarios, future CO2 emissions from road transport will remain higher than from the other transport sectors. Emissions from road transport are highest in A1 where they more than double from 2000 to 2050, while the growth in the other scenarios is smaller (Fig. 1A). The emissions decline from 2050 to 2100 in all scenarios except A2 where the growth is 6%. The development in emissions is a result of the interplay of several of factors; the development in economy and population under the different SRES storylines leading to changes in transport volume, the mix of high- and low-carbon fuels used, the emission indices, and the different temporal and regional development for these factors. The aviation sector is assumed to remain dependent on fossil fuels until 2100, except in the B1 scenario where it is assumed that, by the end of the 21st century, 35% of the fuel used will be non-fossil or produced from hydrogen. CO2 emissions from aviation increase substantially from 2000 in all scenarios (Fig. 1A). Under A2, B1 and B2 emissions double until 2050, 6262 R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 Road B 30 3 Rail 2.5 2 Ship 1.5 Avia 1 0.5 0 1900 1950 2000 2050 CO2 emission, % of total CO2 emission, Pg C/year A 3.5 25 Rail 20 Ship 15 Avia 10 5 0 1900 2100 Road 1950 Year Road 10 Rail 8 Ship 6 Avia 4 2 1950 2000 2050 2100 2050 2100 Year D6 SO2 emission,Tg S/year NOx emission, Tg N/year C 12 0 1900 2000 Year Road 5 Rail 4 Ship 3 Avia 2 1 0 1900 1950 2000 2050 2100 Year Fig. 1. Emission (1900–2100) of CO2 (A), NOx (C) and SO2 (D) from road transport, rail, shipping and aviation. For future emissions scenarios: A1 (solid), A2 (dash-dot), B1 (dashed) and B2 (dotted). Fig. 1B shows emission of CO2 from the transport sectors relative to total man-made CO2 emissions (excluding land use change (LUC)). while for A1 the increase is higher. In 2100 the growth in B1 and B2 has slowed down (emissions in B1 even decline slightly from 2050) while A1 and A2 emissions more than double from 2050. In the shipping sector, oil is assumed to still be the predominant fuel in 2050 and 2100 in all scenarios. However, biofuels and gas in internal combustion engines and fuel cells are introduced. Shipping also utilizes non-emitting technologies such as wind, nuclear and solar cells. In all scenarios, the fuel consumption increases from 2000 to 2100. As a result CO2 emissions more than double from 2000 to 2050; in A1 the increase is almost four-fold (Fig. 1A). This increase continues, and in 2100 emissions in A1 are eight times higher than in 2000, while emissions in the other scenarios are about five times higher. (The emergence of new shipping routes is included in the scenarios). Emissions from the rail sector are small relative to the other transport modes. The freight transport volumes and emission factors are higher, while the electrification rate and introduction of energy efficiency improvements are lower in A2 than in A1. Hence, A2 has the highest CO2 emissions in this sector. Emissions increase throughout the century in A1 and A2, while in B1 and B2 there is a slight decrease. Fig. 1B shows the percentage contribution from transport to total man-made CO2 emissions. With the exception of road transport in scenario A2 and B2 and of rail, the contribution will increase or stay stable. only be a small fraction of the 2000 emission level. This is due to already available technology for conventional vehicles in the road sector, prospects for further development and the introduction of zero emitting vehicles and fuels. The penetration of new technologies is also quicker in the road sector compared to shipping and aviation due to the generally shorter vehicle lifetimes. For both the aviation sector and the shipping sector, NOx emissions will increase throughout the century (except for aviation in B1), but the increase is much lower than for CO2 since the increase in activity is assumed to be compensated by technology improvements. In the rail sector, emissions are highest and increasing in A2, while there is a decrease in the other scenarios. The future SO2 emissions remain substantially higher for shipping than for the other transport sectors. In scenario B1 and B2 the rapid reduction in SO2 emissions starts early in the 21st century, and in 2050 the SO2 emissions are 1/3 of the emission level in 2000. The decline in SO2 emissions in scenarios A1 and A2 starts after 2050, and in 2100 the SO2 emissions in all four scenarios are less than 1/10 of the emission level in 2000 (Fig. 1D). The scenarios span a range of possible developments and capture the effects of the expected sulphur reductions according to the regulations adopted by the International Maritime Organization (IMO). 2.2. Models and methods Future non-CO2 emissions In the scenarios, emissions of non-CO2 gases decrease substantially throughout the 21st century in the road transport sector, for example as seen for NOx (Fig. 1C), and will in 2100 To calculate the global temperature change from the transport sectors a simple climate model (SCM) is used (Fuglestvedt and Berntsen, 1999). The model has been applied in several studies of R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 contributions to global warming and responses to emission change (Fuglestvedt et al., 1999, 2000, 2008; den Elzen et al., 2005; Rive et al., 2007). The SCM calculates global mean concentrations from emissions of 24 gases and the radiative forcing for 30 components (see SI Table 2) based on detailed input. The global mean temperature change is calculated by an energy-balance climate/up-welling diffusion ocean model developed by Schlesinger et al. (1992). In the SCM the atmosphere is represented by a single layer separated into a Northern and Southern hemisphere, while the ocean under the surface layer is split into 40 vertical layers in addition to the north/ south separation. The climate response is governed by the prescribed climate sensitivity, which encompasses the processes, including feedbacks, involved in the response of the climate system to a radiative forcing, and parameters which control the uptake of heat by the oceans (Harvey et al., 1997). The input parameters are based on output from more detailed GCM experiments. In this work we have used a climate sensitivity of 0.8 K (W m2)1, but have also tested the effects of other climate sensitivities and other input parameters; see Section 4.2.1 and SI Section 3. The development in global concentration of CO2 is calculated using a scheme based on Joos et al. (1996). The CO2 module uses an ocean mixed-layer pulse response function that characterizes the surface to deep ocean mixing in combination with a separate equation describing the air-sea exchange (Siegenthaler and Joos, 1992). It also includes changes in CO2 uptake by terrestrial vegetation due to CO2 fertilization. A feedback between atmospheric CO2 levels and CO2 uptake via changes in oceanic pH is included, while the feedback via ocean temperatures is not. For the other gases, standard values for lifetime/adjustment time are used. Indirect effects of CH4 on tropospheric O3 and stratospheric H2O as well as effects on its own adjustment time are taken into account. Radiative forcings for the well-mixed gases are parameterized using updated concentration-forcing relations from IPCC AR4 (Forster et al., 2007), while for short-lived components we use RF results based on detailed CTM and forcing calculations (see below). To calculate the climate impact of transport, we use the common method of removing all emissions from one transport sector at a time and then calculating the difference between this perturbed case and the reference simulation with all anthropogenic emissions (Penner et al., 1999; den Elzen et al., 2005; Fuglestvedt et al., 2008; Marais et al., 2008). The difference in the climate response is then a measure of the impact of that sector. For long-lived greenhouse gases (CO2, CH4, N2O) we subtract the emission trajectories for the transport sectors. For short-lived components we subtract precalculated radiative forcing (RF) over time for the transport sectors. Transport emissions have a spatial and vertical distribution that differs from that of the total global emissions. This distribution must be accounted for when atmospheric burden and RF of short-lived components from the transport sectors are calculated. To establish the global mean RF development, RF(t), for short-lived components from the transport sectors, reference radiative forcings (RFref) and emissions (EMref) for the year 2000 from more detailed studies (Sausen et al., 2005; Fuglestvedt et al., 2008) are used. These studies use global 3D chemistry transport models (CTMs) and detailed radiative transfer models to calculate the RF. SI Table 3 gives the RFref and EMref used in this study. The reference RF and emissions are scaled with emission trajectories (EM(t)) for the transport sectors: RFðtÞ ¼ RFref EMðtÞ EMref (1) For SO4 direct and indirect forcing we scale with the SO2 emission trajectories and for BC and OC we scale with BC and OC emissions, respectively. For the short-term ozone forcing, mainly due to NOx, CO and VOC changes, we scale with NOx emissions. 6263 Studies of impact of transport emissions on O3 and CH4 indicate that such linear scalings only introduce inaccuracies of less than a few percent (Hoor et al., 2009). NOx emissions reduce the lifetime of CH4 (and thus the concentration), which gives a negative RF (Fuglestvedt et al., 1996; Wild et al., 2001; Berntsen et al., 2005). Since the RFref for the CH4lifetime effect from more detailed studies is due to all NOx emissions up to 2000, the RF development cannot be calculated by using Eq. (1). Instead Eq. (2), which takes into account the effect of NOx on CH4 over time, is used: RFðtÞ ¼ 0 tX ¼t NOxðt 0 Þ RFref ð1 expð1=sÞÞexpð ðt t 0 Þ=sÞ NOxref t0 ¼ 0 (2) where s is the methane adjustment time (12 years), NOxref is the reference NOx emission in 2000 and RFref is the RF-CH4 at steady state corresponding to NOxref. The term RFref ð1 expð1=sÞÞ gives the RF after one year and the term expððt t 0 Þ=sÞ accounts for the decaying effect over time. By summing up these contributions we capture the effect of historical and future emissions of NOx on CH4. This implicitly assumes that CO and VOC changes are parallel to NOx. The O3 perturbation consists of two components. In addition to the short-term forcing mentioned above, there is also a longer-term perturbation, called primary mode (PM) O3 forcing, which results from changes in CH4 due to NOx emissions (Wild et al., 2001; Berntsen et al., 2005). The development for RF-O3PM is calculated in a similar way as for the CH4 effect and the RFref is calculated as in Berntsen et al. (2005). The changes in CH4 lifetime are from Fuglestvedt et al. (2008) and Sausen et al. (2005). For the sensitivity of O3 to changes in methane we use results from Ehhalt and Prather (2001). For aviation, the RF in 2000 for H2O and contrails (Sausen et al., 2005) are scaled by the CO2 emissions to give the RF over time. In the case of the RF from cirrus, we have accounted for the increasing overlap of aviation induced cirrus when air traffic increases. We follow the method by Mannstein and Schumann (2005) to calculate cirrus cover using air traffic data on a 1 1 grid for the year 2000 from Owen et al. (2006): d c ¼ cpot 1 ed* (3) where c is the aviation induced cirrus coverage, cpot is the fraction of potential cirrus coverage, d is the air-traffic density and d* is the airtraffic density leading to full cirrus coverage. The value chosen for cpot is 0.2 and d* ¼ 0.5 km (km2 h1) (Mannstein and Schumann, 2005). The air-traffic density on a 1 1 grid is scaled with the CO2 scenario emissions, and future cirrus coverage on a 1 1 grid is calculated. The RF for the year 2000 for cirrus (Stordal et al., 2005) is scaled by the global mean cirrus coverage to give the future RF. The difference in cirrus forcing between this method and a linear scaling with CO2 is up to 6% in the latter part of the 21st century. 3. Radiative forcing To calculate the temperature change due to transport, the historical and future RFs are established for each component and each sector following the description above. Fig. 2 shows global mean RF since pre-industrial times by substance, transport sector and scenario for 2050. Both historical and future transport emissions are taken into account in this figure (see SI Fig. 1 for the results for 2000 and 2025). For all substances, the RF from the rail sector is much smaller than from the other transport modes, and will not be described further in the following section. 6264 R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 Radiative Forcing Year 2050 700 Road 600 Rail Aviation Ship 500 mW/m 2 400 300 200 100 0 -200 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 A1 A2 B1 B2 -100 CO2 CH4 N2O SO4 dir SO4 indir O3 O3 LL BC OC H2O Contrail Cirrus Fig. 2. Global mean radiative forcing (mW m2) from pre-industrial times by substance, transport sector and scenario for 2050. Transport emission period: 1900–2050. CO2 has the largest RF, with road transport being the dominating sector. In 2000, the RF-CO2 from transport as a whole was w200 mW m2, and in 2050 it reaches 470–630 mW m2 depending on scenario. The RF from CH4 in Fig. 2 includes both the direct effect of methane and the indirect effect of NOx, CO and VOC on methane. The direct effect is small, less than 1 mW m2, due to low CH4 emissions from transport. The indirect effect is the reduced lifetime of CH4 due to the NOx induced enhancement of OH levels. This results in a negative forcing. As can be seen in Fig. 2, shipping is the largest contributor to the CH4 forcing, followed by aviation. For road transport the low effect of NOx on methane is due to the high background levels of NOx and the opposing effect of VOC and CO. The contribution from the aviation sector to RF-CH4 increases from 2000 to 2050 and reaches 19 to 33 mW m2, depending on scenario, in 2050. The RF from N2O is small. In 2050 it is w2 mW m2 for road transport and less than 1 mW m2 for the other sectors. As mentioned, the emission of NOx leads to a strong, short-lived positive RF-O3, but also to a negative, long-lived forcing through changes in CH4; RF-O3PM. Fig. 2 shows that the RF-O3 is, in absolute magnitude, larger than the RF-O3PM. The aviation and shipping sectors are the main contributors to the positive short-lived RF from ozone. From 2000 to 2050 the forcing from aviation increases from 30 mW m2 to 38–82 mW m2, while the forcing from shipping increases from 33 mW m2 to 36–62 mW m2. Similarly to the indirect CH4 effect, the RF-O3 from the road transport sector is greatly reduced from 2000 to 2050 due to a rapid reduction in NOx emissions. The RF-O3PM is larger for shipping than aviation. This is due to the larger change in the ozone primary mode lifetime for ship emissions compared to aviation; 5.2% and 1.3% respectively (Fuglestvedt et al., 2008). Sulphate aerosols are formed from emissions of SO2. The aerosols can affect the climate directly through scattering of solar radiation or indirectly through changing cloud properties, both having a cooling effect on climate. Shipping is the transport sector with the largest SO2 emissions, and is the dominating contributor to direct and indirect SO4 forcing. In scenarios B1 and B2, the SO2 emissions from shipping decrease rapidly in the beginning of the 21st century (Fig. 1D). Both the direct and indirect RF from SO4 in 2050 are therefore 22% (in B1) and 32% (in B2) of the corresponding year 2000 RF. In scenarios A1 and A2, the emission of SO2 has increased since 2000, and hence the direct and indirect RF is enhanced. Aircraft SO2 emissions increase up to 2050 in all scenarios, and the SO4 direct forcing from this sector is enhanced. In the B1 and B2 scenarios the aviation sector dominates the SO4 direct forcing in 2050. The RF from BC and OC is largest at the beginning of the century. In 2000 the RF–BC from transport was w16 mW m2. Strong reductions in BC emissions from the transport sectors after 2000 lead to the strong decrease in RF. The RF from water vapour in Fig. 2 is the net RF due to reduced methane concentration in the troposphere, which, in turn, reduces methane oxidation in the stratosphere and the resulting water vapour, and the direct emission of H2O from aviation. The contribution is largest for the shipping sector (11 to 7 mW m2 depending on the scenarios) due to the high NOx emissions. For aviation the two different effects (direct emission and reduced methane oxidation) compensate each other. The high-altitude emission of water vapour from the aviation sector leads to RF mechanisms in addition to the increase in stratospheric water vapour, namely contrails and cirrus formation. The activity in the aviation sector increases in all four scenarios, and the RF from cirrus and contrails increases from 2000 to 2050. The RF from contrails increases from 10 mW m2 to 20–36 mW m2 and the forcing from cirrus increases from 30 mW m2 to 60–105 mW m2. The RF in 2050 from these mechanisms is 3–4 times larger for A1 and twofold larger in the other scenarios compared to the year 2000 value. The uncertainty in the cirrus and contrails forcing is large. IPCC (Forster et al., 2007) gives a 90% confidence interval ranging from 0.003 to 0.03 mW m2 for contrail distribution in 2005. Stordal et al. (2005) gave an interval ranging from 10 mW m2 to 80 mW m2 for the RF from cirrus. The uncertainty in RF calculation is taken into account when the temperature change due to the transport sectors is calculated. The effect of aircraft soot on cirrus was recently investigated in two studies. Penner et al. (2009) found a change in the ice number concentration due to aircraft soot emissions leading to a large net negative RF of 160 to 120 mW m2. However, the effects of the additional soot on cloud occurrence and cloud fraction was not included. This was investigated in a study by Liu et al. (2009). Here, aircraft soot was found to give a net negative cloud forcing only if soot acts as an ice nuclei (IN) at a threshold RHi of 120–130%. When a higher threshold RHi is required, the change in cloud forcing was positive. The results from these two studies have the potential to significantly change the overall picture of radiative forcing of aviation. More studies are needed on this topic and due to the large uncertainties related to the role of soot in ice nucleation and the very differing results in these first studies, we do not include these effects in our estimates. CO2 has a long adjustment time and the forcing is thus largely dependent on the emission history. If the emission prior to 2000 is R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 excluded from the analysis, the RF for CO2 from transport in 2050 is 111–119 mW m2 less, depending on scenario, than when the historical CO2 emissions are included. For the other components with shorter lifetimes, emissions prior to 2000 do not affect the RF in 2050. 6265 included. However, here we want to make comparison across the sectors on an equal basis. Therefore we also calculate the global mean temperature change due to only future emissions (i.e. perturbation start year 2000 and end year 2100). 4.1. Temperature change due to historical and future emissions 4. Development of global mean temperature Fig. 3A shows the temperature increase for the control simulation for the four SRES scenarios. We calculate a total net temperature increase of 0.76 C from pre-industrial times to 2000. In 2100, the total temperature change since pre-industrial times is between 2.5 and 4.5 C using the best estimate from IPCC of 0.8 K (W m2)1 for the climate sensitivity. The best estimate for global mean temperature increase from 1980–1999 to 2090–2099 in IPCC AR4 (Meehl et al., 2007) varies between 1.8 and 3.4 C for the four scenarios. Different time periods are considered in the above estimates and our results are in good agreement with the IPCC estimates when a temperature increase from pre-industrial times to present day of w0.75 C is accounted for. Fig. 3B and C show absolute and percentage contributions from the transport sectors to total net man-made temperature change. Road transport contributes most to historical temperature change. In 2000 the temperature change due to road transport is 0.08 C, corresponding to 11% of total anthropogenic temperature change. Even though the CO2 emissions from road transport have increased In this work we present two different perspectives on the evaluation of global mean temperature change due to transport. First we calculate the temperature change due to historical and future transport emissions (perturbation start year 1900 and end year 2100). This gives the complete picture of how transport contributes to global mean temperature change over time. However, including historical emissions when calculating the contribution to future climate change from transport is less relevant for policy making than considering future emissions only. The transport sectors have very different emission histories and past emissions thus affect the future temperature change to a varying degree. For example, road transport has a long history and large CO2 emissions which strongly impact the future contribution to temperature change. Aviation, on the other hand, is a relatively ‘‘young’’ sector with more short-lived emissions, and the historical emissions thus impact the future temperature change much less. As long as one compares different scenarios or mitigation measures within one sector the full history can be 4.5 A1 A2 B1 B2 4 3.5 B 3 2.5 2 1.5 0.4 0.3 0.2 Ship 0.15 0.1 0 −0.05 2000 2050 −0.1 1900 2100 Avia 0.05 0.5 1950 Rail 0.25 1 0 1900 Road 0.35 Temperature, K Temperature Control Simulation, K A 1950 2000 Year 2100 15 Road D 25 0.8 A1 A2 B1 B2 0.7 Rail 5 Ship 0 Avia −5 Temperature, K 10 20 0.6 15 0.5 0.4 10 0.3 5 0.2 0 0.1 −10 −5 0 −15 1900 1950 2000 2050 2100 −0.1 1900 Temperature, % of total Temperature, % of total C 2050 Year 1950 2000 2050 −10 2100 Year Fig. 3. A: Total temperature change due to historical emissions and future emissions following the four SRES scenarios: A1 (solid), A2 (dash-dot), B1 (dashed) and B2 (dotted). B: Temperature change due to emissions from transport sectors in the period from 1900 to 2100. C: Temperature change due to transport sectors relative to the total temperature change. D: Absolute and percentage contribution from the transport sector as a whole to total net man-made temperature change. Climate sensitivity used in simulations: 0.8 K (W m2)1. 6266 R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 relative to the total man-made CO2 emissions (excluding land use change, Fig. 1B) in the 20th century, the proportion of temperature change due to road transport relative to total temperature change declines in the latter part of the 20th century in our simulations. This is due to the increasing effect of other sectors than transport. Furthermore, the shipping sector contributes to cooling in the 20th century, mainly due to SO2 emissions. The contribution to temperature change in 2000 from shipping is 0.05 C. Emissions from the aviation sector started later than those from the shipping and road transport sectors, and contributed 4% to the total anthropogenic temperature change in 2000 (i.e. 0.03 C). The contribution from the rail sector was 1%. Fig. 3D shows both the absolute and the percentage contribution from transport as a whole to total net man-made temperature change. The transport sector contributed to global warming in the 20th century, with the exception of a period from 1900 to 1950 where a small negative temperature change is seen, mainly due to shipping. Relative to total temperature change, transport contributed an increasing fraction to the warming in the latter part of the century, to 9%, or 0.07 C, of the total temperature change in 2000. According to our calculations and given the emission scenarios used, transport will continue to contribute to global warming in the future (Fig. 3B), and road transport remains the largest contributor in all scenarios. The temperature change due to aviation increases strongly in the A1 scenario. This scenario assumes very rapid economic growth, low population growth and rapid introduction of new and more efficient technology. However, for aviation the high growth in demand outweighs the technological improvements, and emissions are highest in the A1 scenario (Fig. 1A). The cooling from shipping decreases in the future due to reduced SO2 emissions. In scenarios B1 and B2 the temperature change due to the shipping sector will be positive in the mid 21st century (Fig. 3B). In A1 and A2 the reduction in SO2 emissions from shipping is introduced later (Fig. 1D), and the reduction in negative contribution to temperature change takes place close to 2100. The temperature change from rail traffic remains small. In 2050, the transport sector as a whole contributes about 14% to the total temperature change in scenarios A1 and B2, 16% in A2 and 10% in B1. At the end of the century, the contribution is between 9% and 20% for the different scenarios, showing a growth in scenarios A1 and B1 (Fig. 3D). 4.2. Temperature change due to future emissions In Section 4.1 the whole emission history is included when contributions from the transport sectors to temperature change are calculated. This section shows the contribution to temperature change in the 21st century due to future transport emissions alone in order to facilitate a comparison across sectors without the influence of very different emission histories (see beginning of Section 4). The start of the emission perturbations is now year 2000. Fig. 4A and B show the temperature change in 2050 and 2100 due to emissions from the transport sectors in the period 2000– 2100 for the four scenarios. The best estimates for road transport are 0.13–0.18 C in 2050 and 0.19–0.35 C in 2100 depending on scenario. For aviation the best estimates of temperature change are 0.07–0.10 C in 2050 and 0.10–0.28 C in 2100. For the rail sector, the future temperature change is small, w0.007 C in 2050 and w0.015 C in 2100 in scenarios A1 and A2. The road sector is the largest contributor to temperature change in all scenarios, but in the A2 scenario the contribution to temperature change from aviation is nearly equal at the end of the century. In the A2 scenario, the aviation sector undergoes a rapid growth and there are no technological improvements assumed after 2020, while the temperature change due to road transport has a slower rate of increase. This is due to low growth in the road sector following low GDP growth and large improvements in fuel economy. Shipping makes a negative contribution to temperature change at the beginning of the century in all scenarios due to the SO2 emissions. The contribution to temperature change still switches from negative to positive in all scenarios except A2, but this occurs later than when the whole history is included. This is due to the lower contribution from CO2 when historical emissions are removed. Excluding historical emissions gives 22–28% (i.e. w0.05 C), depending on scenario, less temperature increase from road transport in 2050 than when the whole emission history is included. For the aviation sector, the temperature change is reduced by 7–11% (i.e. w0.008 C). In 2100 the difference is 11–16% for road transport and 2–7% for aviation. This shows the effect of pre-2000 emissions on post-2000 temperatures and how the magnitude of this effect varies between the different transport modes. 4.2.1. Uncertainties In Fig. 4, which shows the temperature change in the years 2050 and 2100 for the transport sectors for the scenarios, the uncertainties at one standard deviation level is included. The establishment of the uncertainties in the RF calculation follows the method in Fuglestvedt et al. (2008). The uncertainties in atmospheric modelling (atmospheric dispersal, removal, and RF), with references, are given in SI Table 4. Uncertainties in emissions are not included, since the emission scenarios themselves span a range of possible developments and magnitudes. The uncertainties in RF are combined with the uncertainty in the climate sensitivity, as shown in Fig. 4A and B. IPCC 2007 stated that the equilibrium climate sensitivity (ECS), i.e. the global average surface warming following a doubling of carbon dioxide concentrations, is likely to be in the range 2–4.5 C, with a best estimate of about 3 C. However, RF from transport is likely to continue to increase through the coming century. It may therefore be more relevant to use the ‘transient climate response’ (TCR, defined as the globally averaged surface air temperature change at the time of CO2 doubling in the 1% yr1 transient CO2 increase experiment) in the establishment of the uncertainty in the climate sensitivity used in our simulations. IPCC gives a 90% confidence interval ranging from 1 to 3 C, with a best estimate of 2 C, for the TCR based on climate modelling (Meehl et al., 2007). We use this confidence interval to calculate the standard deviation for the climate sensitivity, which is then 30% of the best estimate of the TCR. The uncertainty in the contribution to temperature change is assessed by combining the uncertainty in RF for different components and the uncertainty in the climate sensitivity by using a Monte-Carlo approach; for further description see SI Section 5. Fig. 4A and B show the absolute temperature change due to the transport sectors with total uncertainty (i.e. RF and climate sensitivity) for 2050 and 2100, respectively. The uncertainty in the temperature increase due to the transport sectors is larger than the difference between the emission scenarios. However, ranking the transport sectors according to their temperature impact is not possible based on Fig. 4A and B since the uncertainty in the climate sensitivity is not independent across all sectors. To be able to perform a ranking, we also calculate the uncertainties in temperature change due to uncertainty in RF alone, i.e. assuming equal climate sensitivity for all sectors (Fig. 4C and D). From comparison of the two top figures with the two bottom ones it can be seen that the uncertainty in RF dominates for aviation. For road transport on the other hand, the uncertainty in temperature increase is mainly related to the uncertainty in climate sensitivity. This is due to the fact that the well known RF from CO2 contributes most to temperature increase from road transport, while for aviation temperature change has large contributions from R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 A 0.50 B Temperature change (K) Year 2050 Uncertainty in climate sensitivity and RF 6267 Temperature change (K) Year 2100 Uncertainty in climate sensitivity and RF 0.50 A1 0.40 A2 0.30 B1 B2 0.20 A1 0.40 B1 B2 0.20 0.10 0.10 0.00 0.00 -0.10 -0.10 -0.20 A2 0.30 -0.20 Road Aviation Rail Ship Temperature change (K) Year 2050 Uncertainty in RF C 0.50 Road D Aviation Rail Ship Temperature change (K) Year 2100 Uncertainty in RF 0.50 A1 0.40 0.30 0.40 B1 0.30 B2 0.20 A1 A2 B1 B2 0.20 0.10 0.10 0.00 0.00 -0.10 -0.10 -0.20 A2 -0.20 Road Aviation Rail Ship Road Aviation Rail Ship Fig. 4. Temperature change due to future emissions from the transport sectors (perturbation start year 2000) in 2050 (A) and 2100 (B) with uncertainty in both climate sensitivity and RF, and in 2050 (C) and 2100 (D) with uncertainty in RF only. Climate sensitivity used in simulations: 0.8 K (W m2)1. the uncertain forcing from contrails and cirrus. For shipping, uncertainty in RF dominates in 2050, while uncertainty in climate sensitivity becomes more important in 2100. In one of the scenarios the uncertainty range includes both positive and negative values for the temperature change (scenario B1 in 2050 and A2 in 2100). When only the uncertainty in RF is included (Fig. 4C and D), the temperature increase in 2050 from aviation, even with the maximum value in the uncertainty range, remains smaller than that from road transport. In 2100, the temperature increase from aviation can be larger than from road transport in scenarios A1 and A2. Correlations between the sectors exist also for the uncertainty in RF, but this is likely of less importance for these two sectors since the contribution from cirrus and contrails to temperature change from aviation is not present for road transport. To illustrate the contribution to temperature change from some forcing mechanisms with large uncertainty, several sensitivity tests are performed. For example, the temperature increase in 2050 due to aviation is reduced by w47% if the cirrus effect is excluded. A high uncertainty is also related to the indirect aerosol effect. Lauer et al. (2007) suggested a substantially higher indirect RF from shipping than the reference study used in this paper (Fuglestvedt et al., 2008). Lauer et al. (2007) calculated the indirect RF from shipping using three different emission datasets. Furthermore, they used a general circulation model (GCM) and included both the first and second indirect aerosol effect, whereas the study by Fuglestvedt et al. (2008) only considered the first. The results in Lauer et al. (2007) vary between 0.19 and 0.6 W m2 for the year 2000. If we use their sensitivity (RFref/EMref) for the SO4 direct and indirect effects, the temperature change from shipping in 2050 is 0.04 to 0.33 C compared to 0.01 to 0.08 C with the sensitivity from the reference study. 4.2.2. Indirect emissions In addition to the emissions directly from vehicles during operation and propulsion, emissions are generated during the fuel production and distribution stage. We calculate the temperature change due to these indirect emissions in addition to the scenarios, using the same method as described in Section 2.2. In 2050, the contribution to temperature change from indirect emissions from road transport is between 5% and 11% of the contribution from the direct emissions depending on scenario. For shipping and aviation the contribution to temperature change from indirect emissions is small compared to that from the direct emissions. For the rail sector the contribution is minor since we have chosen to consider electricity generation as part of the direct emissions. 4.2.3. Effect of mitigation The transport emission scenarios, as well as the SRES scenarios, do not take mitigation of climate change into account. In our study we use the ACARE scenario for aviation to test how a strong mitigation will affect the contribution to climate change for this sector. The ACARE scenario assumes ‘‘excellent fuel efficiency’’ and larger NOx reductions than in scenario B1. The ACARE target is achieved in 2020 and continuing improvements beyond 2020 are assumed. The ACARE aviation scenario results in a 0.05 C temperature increase in 2050, 20% less than the temperature increase in the B1 scenario (SI Fig. 2). 4.2.4. A scenario sensitivity test Starting between 2010 and 2030, depending on the scenario, NOx and other air pollutant emissions from road transport decrease strongly (Fig. 1C). This decline is justified by an increasing level of road transport emission regulation in all world regions driven by 6268 R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 current and future European and US legislation. However, there is uncertainty related to the enforcement of regulations as well as to the development and deployment of future technologies. Here, a ‘‘delayed reduction’’ scenario for NOx, CO, NMVOC, BC and OC emissions from road transport is introduced. All emissions after 2030 are a factor 1.5 higher than in the original scenario and the decline is less abrupt. Using this emission scenario, the temperature increase due to road transport is increased by 4% (0.007 C) in A1 in 2050 and 0.5% (0.001 C) in 2100 relative to the original road perturbation. The emissions in this ‘‘delayed reduction’’ scenario differ less from the original scenarios towards 2100, and the difference in temperature increase is smaller in 2100 than in mid-century. Furthermore, the components affected are short-lived forcing agents with RFs of opposite signs and the early perturbation therefore has little influence on the temperature in 2100. The exception here is the negative RF from the CH4-lifetime effect. CO2 is not changed in this test, since the emission scenarios applied here span a range of possible developments for CO2 emissions through the 21st century. 5. Discussion There are a number of non-linear relations in the climate system. Some of these are accounted for in our model framework by using results from more sophisticated and complex studies. There is however no single correct way to handle non-linear behaviour since this will depend on the chosen perspective. If one is interested in the total effect of a sector, then this can be studied based on the assumption that all other sectors and emissions remain unchanged. One may also be interested in changes at the margin, which would be relevant for assessment of small and medium scale emission reductions. Because nonlinear processes control some substances (e.g. O3 and OH), the RF of ozone and methane cannot be scaled exactly to obtain the effect of marginal changes reflecting realistic short-term mitigation measures. Many of the short-lived components have effects that depend strongly on the location of emissions, e.g. NOx (Fuglestvedt et al., 1999; Wild et al., 2001; Berntsen et al., 2005; Derwent et al., 2008). For the NOx–O3 relation for road transport we have scaled the O3 forcing for this sector from Fuglestvedt et al. (2008). Ozone production also depends on the CO and VOC emissions. We have scaled the RF with global emissions, which can be done if the geographical distribution of the emissions is assumed to remain unchanged. This represents an approximation since the distribution is likely to change in the future. For cirrus and contrails the calculation of RF includes scaling with fuel use (i.e. CO2 emissions). Additionally, the cirrus and contrail formation depends on the meteorological conditions. If there are changes in flight paths the RF can be different. Many of the RF agents studied here are controlled by processes that are sensitive to changes in climate (e.g. water vapour or temperatures); for example, a change in OH levels will affect chemically active climate gases such as O3 and CH4. We have assumed a constant background atmosphere and climate. The couplings between chemistry and climate are potentially important and need to be considered in more detailed studies. Due to important feedback processes in the climate system, such as increased atmospheric CO2 due to rising ocean temperatures and the release of CH4 from melting permafrost, current conditions and assumptions may not remain valid in the future, which might affect our results. A larger fraction of anthropogenic CO2 will stay airborne if climate change is accounted for. Friedlingstein et al. (2006) found that by the end of the twenty-first century, this increase in CO2 varied between 50 and 100 ppmv for the majority of the models applied in the study (and 20 and 200 ppmv for two extreme models). The higher CO2 levels led to an additional climate warming ranging between 0.1 and 1.5 C. Accounting for this effect would increase the climate impacts of transport emissions, but is not included in this study. In our calculations we have not included efficacies for the various forcing agents. Efficacy is defined as the ratio of the climate sensitivity parameter for a given forcing agent to the climate sensitivity parameter for CO2 changes. In other words, different RF mechanisms will yield different temperature responses. Both the geographical and vertical distribution of the forcing can have significant effects on efficacy (Forster et al., 2007). Ponater et al. (2005), Grewe et al. (2007) and Grewe and Stenke (2008) presented efficacies for aircraft RFs from model studies. Efficacies specific to other transport sectors are not established. There is also a large spread in efficacies presented in IPCC AR4 (Forster et al., 2007), especially for BC. Differences in efficacies between the various forcing agents may modify the calculated contributions of the various components and the sectors. However, Marais et al. (2008) performed a sensitivity analysis for the effect of different efficacy values on the integrated temperature response and found only a relatively small effect. The transport sector contributes to climate change by means of several forcing agents not covered by the Kyoto Protocol, most notably tropospheric O3 driven by NOx, CO, and VOC. The negative forcings caused by SO2, OC, and ozone precursors are also excluded. Therefore, a ‘‘Kyoto perspective’’ does not capture the full climate impact of transport, especially for the shipping sector due to its large contribution to SO2 and NOx emissions. Furthermore, the Kyoto Protocol does not cover emissions from international aviation and shipping. If we look at the effect of including only the Kyoto gases in the perturbations (i.e. only CO2 since the contribution from N2O and CH4 is negligible), the temperature increase due to road transport remains similar to the case in which all emissions are included, only w4% and w0.5% smaller in 2050 and 2100, respectively. This is due to the large reduction of nonKyoto gases at the beginning of the century in this sector and the dominating contribution from CO2. For the aviation sector, the difference is larger; excluding non-Kyoto components significantly reduces the calculated temperature increase. This is the case even if the cirrus effect is excluded from the standard aviation perturbation. In line with earlier studies, the temperature response from shipping is positive when only the Kyoto gases are included and is of the same magnitude as the effect of aviation. In addition to the RF effects considered in this study, other indirect processes have been proposed. These include reduction of the albedo of snow and ice by BC deposition (Warren and Wiscombe, 1980; Hansen and Nazarenko, 2004; Flanner et al., 2007), the so-called semi direct effect of BC on clouds (Ackerman et al., 2000), and indirect effects on cirrus clouds by particles from aviation (Penner et al., 1999). All these effects are currently given a low or very low level of scientific understanding by the IPCC (Forster et al., 2007). Emissions of CFCs and HFCs from mobile air-conditioning (MAC) are not taken into account in this study. CFC-12 (from MAC) has lead to a direct forcing of 170 mW m2 in 2005 and an indirect RF of 34 mW m2 due to stratospheric ozone depletion while the RF from HFC-134a was 5.5 mW m2 (Uherek et al., submitted for publication). 6. Conclusion This work has shown that there are large differences between the transport sectors in terms of sign and magnitude of their effect on climate, and also with respect to the contributions from the R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 long- and short-lived components and thus the temporal characteristics of the climate response. Although we have analysed the contributions both historically and for future scenarios to both RF and global mean temperature, the results agree with the general conclusions obtained by Fuglestvedt et al. (2008) and Berntsen and Fuglestvedt (2008) for RF and temperature response to historical and current emissions. We find that, since pre-industrial times, the transport sector as a whole has contributed 9% to the total net man-made warming in 2000. The dominating contributor to warming is CO2, followed by tropospheric O3. Emissions of SO2, mainly from shipping have, on the other hand, had a cooling effect. By sector, road transport is the largest contributor to warming (11%). Aviation has contributed 4% and rail 1% to the total net anthropogenic warming in 2000. Shipping has caused a net cooling up to 2000, contributing 7% to the net man-made temperature change. We have also calculated future effects of transport for four scenarios consistent with IPCC SRES scenarios. In 2050 we find that the net contribution from the transport sector as a whole to total man-made warming is approximately 10% in scenario A2, 14% in A1 and B2 and 16% in B1 when historical emissions are included. In 2100 the contribution reaches w20% in scenarios A1 and B1. In order to facilitate a comparison across the transport sectors without the influence of their very different emission histories, we calculate the future temperature increase due to only post-2000 emissions. In all scenarios road transport is the dominating contributor to warming. The temperature increase resulting from this sector is between 0.13 C (A2) and 0.18 C (A1) in 2050 for a climate sensitivity of 0.8 K (W m2)1. Furthermore, the contribution from rail is 0.006–0.008 C in 2050, and the net effect of emissions from the shipping sector is negative, ranging from 0.01 C to 0.08 C. The contribution from aviation varies between 0.07 C and 0.1 C. The absolute values of temperature change depend strongly on the chosen climate sensitivity. The relative temperature change, however, does not, and will remain approximately the same for different choices of climate sensitivity. Throughout the 21st century road transport remains the dominating sector, followed by aviation, and the warming from these transport modes increases up to 2100. The contribution from rail remains small. The net effect of emissions from the shipping sector switches from cooling to warming towards the end of the century in three of the scenarios. This is due to anticipated reductions in SO2 emissions from shipping. Significant uncertainties are related to our estimates of both historical and future net warming from the transport sectors, mainly due to cirrus, contrails, and direct and indirect aerosol effects. However, we find that the conclusions that (i) road transport provides the largest contribution to warming with CO2 and O3 as the dominating warming agents, and that (ii) the response from shipping switches from initial cooling to warming on a longer time scale, remain robust. Acknowledgements We thank Lynn Nygaard, Jens Borken-Kleefeld and colleagues in QUANTIFY for valuable comments and discussions. This research was supported by the European Union’s Sixth Framework Integrated Project QUANTIFY Contract No 003893 and the Norwegian Research Council. Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atmosenv.2009.05.025. 6269 References Ackerman, A.S., et al., 2000. Effects of aerosols on cloud albedo: evaluation of Twomey’s parameterization of cloud susceptibility using measurements of ship tracks. Journal of the Atmospheric Sciences 57 (16), 2684–2695. Berntsen, T., Fuglestvedt, J., 2008. Global temperature responses to current emissions from the transport sectors. Proceedings of the National Academy of Sciences of the United States of America (PNAS) 105 (49), 19154–19159. Berntsen, T.K., et al., 2005. Response of climate to regional emissions of ozone precursors: sensitivities and warming potentials. Tellus Series B-Chemical and Physical Meteorology 57 (4), 283–304. Borken-Kleefeld, J., et al. Global transportation scenarios: emissions of direct and indirect greenhouse gases in the long run, in preparation. Borken, J., et al., 2007. Global and country inventory of road passenger and freight transportation: fuel consumption and emissions of air pollutants in 2000. Transportation Research Records 2011, 127–136. Bouwman, A.F., et al., 2006. Integrated Modelling of Global Environmental Change: an Overview of IMAGE 2.4. www.mnp.nl/en/themasites/image/index.html. Capaldo, K., et al., 1999. Effects of ship emissions on sulphur cycling and radiative climate forcing over the ocean. Nature 400 (6746), 743–746. Corbett, J.J., Fischbeck, P.S., 1997. Emissions from ships. Science 278 (5339), 823–824. den Elzen, M., et al., 2005. Analysing countries’ contribution to climate change: scientific and policy-related choices. Environmental Science & Policy 8 (6), 614–636. Derwent, R.G., et al., 2008. Radiative forcing from surface NOx emissions: spatial and seasonal variations. Climatic Change 88 (3–4), 385–401. Ehhalt, D., Prather, M., 2001. In: Houghton, J.T. (Ed.), Atmospheric Chemistry and Greenhouse Gases in Climate Change 2001: the Scientific Basis. Cambridge Univ. Press, Cambridge. Eide, M., et al., 2007. Ship Emissions of the Future. Technical Report No 2007-1325. Det Norske Veritas, Høvik, Norway. Endresen, O., et al., 2003. Emission from international sea transportation and environmental impact. Journal of Geophysical Research-Atmospheres 108 (D17). Endresen, O., et al., 2007. A historical reconstruction of ships’ fuel consumption and emissions. Journal of Geophysical Research-Atmospheres 112 (D12). Eyring, V., et al., 2005. Emissions from international shipping: 1. The last 50 years. Journal of Geophysical Research-Atmospheres 110 (D17). Eyring, V., et al., 2007. Multi-model simulations of the impact of international shipping on atmospheric chemistry and climate in 2000 and 2030. Atmospheric Chemistry and Physics 7, 757–780. Flanner, M.G., et al., 2007. Present-day climate forcing and response from black carbon in snow. Journal of Geophysical Research-Atmospheres 112 (D11). Forster, P., et al., 2007. Changes in atmospheric constituents and in radiative forcing. In: Solomon, S., et al. (Eds.), Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge and New York. Friedl, R.R., 1997. Atmospheric Effects of Subsonic Aircraft: Interim Assessment Report of the Advanced Subsonic Technology Program. Reference Publication 1400. National Aeronautics and Space Administration, Washington, D.C. Friedlingstein, P., et al., 2006. Climate-carbon cycle feedback analysis: results from the (CMIP)-M-4 model intercomparison. Journal of Climate 19 (14), 3337–3353. Fuglestvedt, J., et al., 2008. Climate forcing from the transport sectors. Proceedings of the National Academy of Sciences of the United States of America (PNAS) 105 (2), 454–458. Fuglestvedt, J.S., et al., 1996. Estimates of indirect global warming potentials for CH4, CO and NOx. Climatic Change 34, 405–437. Fuglestvedt, J.S., Berntsen, T., 1999. A Simple Model for Scenario Studies of Changes in Global Climate: Version 1.0. Working Paper 1999:02. CICERO, Oslo, Norway. Fuglestvedt, J.S., et al., 1999. Climatic forcing of nitrogen oxides through changes in tropospheric ozone and methane; global 3D model studies. Atmospheric Environment 33 (6), 961–977. Fuglestvedt, J.S., et al., 2000. Climate implications of GWP-based reductions in greenhouse gas emissions. Geophysical Research Letters 27 (3), 409–412. Granier, C., Brasseur, G.P., 2003. The impact of road traffic on global tropospheric ozone. Geophysical Research Letters 30 (2). Grewe, V., et al., 2007. Climate impact of supersonic air traffic: an approach to optimize a potential future supersonic fleet – results from the EU-project SCENIC. Atmospheric Chemistry and Physics 7 (19), 5129–5145. Grewe, V., Stenke, A., 2008. AirClim: an efficient tool for climate evaluation of aircraft technology. Atmospheric Chemistry and Physics 8 (16), 4621–4639. Hansen, J., Nazarenko, L., 2004. Soot climate forcing via snow and ice albedos. PNAS 101 (2), 423–428. Harvey, D., et al., 1997. An Introduction to Simple Climate Models used in the IPCC Second Assessment Report. IPCC Technical Paper II. Hoor, P., et al., 2009. The impact of traffic emissions on atmospheric ozone and OH: results from QUANTIFY. Atmospheric Chemistry and Physics 9, 3113–3136. IPCC, 1999. Aviation and the Global Atmosphere. Cambridge University Press, Cambridge. Joos, F., et al., 1996. An efficient and accurate representation of complex oceanic and biospheric models of anthropogenic carbon uptake. Tellus Series B-Chemical and Physical Meteorology 48 (3), 397–417. Lauer, A., et al., 2007. Global model simulations of the impact of ocean-going ships on aerosols, clouds, and the radiation budget. Atmospheric Chemistry and Physics 7 (19), 5061–5079. 6270 R.B. Skeie et al. / Atmospheric Environment 43 (2009) 6260–6270 Lawrence, M.G., Crutzen, P.J., 1999. Influence of NOx emissions from ships on tropospheric photochemistry and climate. Nature 402 (6758), 167–170. Lee, D.S., et al., 2009. Aviation and global climate change in the 21st century. Atmospheric Environment 43 (22–23), 3520–3537. Liu, X.H., et al., 2009. Influence of anthropogenic sulfate and black carbon on upper tropospheric clouds in the NCAR CAM3 model coupled to the IMPACT global aerosol model. Journal of Geophysical Research-Atmospheres 114. Mannstein, H., Schumann, U., 2005. Aircraft induced contrail cirrus over Europe. Meteorologische Zeitschrift 14 (4), 549–554. Marais, K., et al., 2008. Assessing the impact of aviation on climate. Meteorologische Zeitschrift 17 (2), 157–172. Meehl, G.A., et al., 2007. Global climate projections. In: Solomon, S., et al. (Eds.), Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge and New York. Minnis, P., et al., 1998. Transformation of contrails into cirrus during SUCCESS. Geophysical Research Letters 25 (8), 1157–1160. Minnis, P., et al., 1999. Global distribution of contrail radiative forcing. Geophysical Research Letters 26 (13), 1853–1856. Myhre, G., Stordal, F., 2001. On the tradeoff of the solar and thermal infrared radiative impact of contrails. Geophysical Research Letters 28 (16), 3119–3122. Nakicenovic, N., et al., 2000. Special Report on Emissions Scenarios. Cambridge Univ. Press, Cambridge, pp. 1–599. Niemeier, U., et al., 2006. Global impact of road traffic on atmospheric chemical composition and on ozone climate forcing. Journal of Geophysical ResearchAtmospheres 111 (D9). Olivier, J.G.J., Berdowski, J.J.M., 2001. Global emissions sources and sinks. In: Berdowski, J., et al. (Eds.), The Climate System. A.A. Balkema Publishers/Swets & Zeitlinger Publishers, Lisse, pp. 33–78. Owen, B., et al., 2006. New Aviation Scenarios for 2050, Paper Presented at Proceedings of the TAC – Conference, Oxford, UK, June 26 to 29. Penner, J.E., et al., 1999. Aviation and the Global Atmosphere. Cambridge Univ. Press, Cambridge, 365 pp. Penner, J.E., et al., 2009. Possible influence of anthropogenic aerosols on cirrus clouds and anthropogenic forcing. Atmospheric Chemistry and Physics 9 (3), 879–896. Ponater, M., et al., 2005. On contrail climate sensitivity. Geophysical Research Letters 32 (10). Rive, N., et al., 2007. To what extent can a long-term temperature target guide nearterm climate change commitments? Climatic Change 82 (3-4), 373–391. Sausen, R., et al., 2005. Aviation radiative forcing in 2000: an update on IPCC (1999). Meteorologische Zeitschrift 14 (4), 555–561. Schlesinger, M.E., et al., 1992. Implication of anthropogenic atmospheric sulphate for the sensitivity of the climate system. In: Rosen, L., Glasser, R. (Eds.), Climate Change and Energy Policy: Proceedings of the International Conference on Global Climate Change: Its Mitigation through Improved Production and Use of Energy. American Institute of Physics, New York, pp. 75–108. Schumann, U., 1990. Air traffic and the environment–background, tendencies and potential global atmospheric effects. In: Proceedings of a DLR International Colloquium, Bonn, November 15–16, 1990, Lecture Notes in Engineering, vol. 60. Springer, Berlin, p. 170. Siegenthaler, U., Joos, F., 1992. Use of a simple-model for studying oceanic tracer distributions and the global carbon-cycle. Tellus Series B-Chemical and Physical Meteorology 44 (3), 186–207. Stevenson, D.S., et al., 2004. Radiative forcing from aircraft NOx emissions: mechanisms and seasonal dependence. Journal of Geophysical Research-Atmospheres 109 (D17). Stordal, F., et al., 2005. Is there a trend in cirrus cloud cover due to aircraft traffic? Atmospheric Chemistry and Physics 5, 2155–2162. Stuber, N., et al., 2006. The importance of the diurnal and annual cycle of air traffic for contrail radiative forcing. Nature 441 (7095), 864–867. Travis, D.J., et al., 2002. Climatology: contrails reduce daily temperature range – a brief interval when the skies were clear of jets unmasked an effect on climate. Nature 418 (6898), 601. Uherek, E., et al. Assessment of transport impacts on climate and ozone: land transport, submitted for publication. Warren, S.G., Wiscombe, W.J., 1980. A model for the spectral albedo of snow. 2. Snow containing atmospheric aerosols. Journal of the Atmospheric Sciences 37 (12), 2734–2745. Wild, O., et al., 2001. Indirect long-term global radiative cooling from NOx emissions. Geophysical Research Letters 28 (9), 1719–1722. www.acare4europe.org. www.pa.op.dlr.de/quantify.