Potential for mitigating greenhouse gases through expanding public transport services: A case study for Gauteng Province, South Africa

Transportation Research Part D 32 (2014) 57–69

Contents lists available at ScienceDirect

Transportation Research Part D journal homepage: www.elsevier.com/locate/trd

Potential for mitigating greenhouse gases through expanding public transport services: A case study for Gauteng Province, South Africa Steffen Bubeck ⇑, Jan Tomaschek, Ulrich Fahl Institute of Energy Economics and the Rational Use of Energy (IER), University of Stuttgart, Hessbruehlstr. 49a, D-70565 Stuttgart, Germany

a r t i c l e

i n f o

Keywords: Public transport Greenhouse gas emissions Mitigation costs BRT Rapid rail Megacity

a b s t r a c t South Africa’s Province of Gauteng is a fast growing megacity region including the cities of Johannesburg and Tshwane. Increasing population and prosperity lead to a steadily growing energy demand and thereby increasing greenhouse gas (GHG) emissions. One third of the province’s final energy consumption comes from the transport sector, dominated by motorized individual transport. Due to the limited financial resources to fund public transport initiatives, the most cost-effective means to reach the GHG mitigation targets are intended, without jeopardizing the economic growth. Recently, a bus rapid transit (BRT) system (Rea Vaya) and a rapid rail link (Gautrain) have been introduced to enforce the public transport system. In this paper, we investigate planned and possible future network expansions of the BRT and the Gautrain in terms of transport performance, costs of network expansions and GHG mitigation potential. Based on a trip rate model, we show that extensions of the current network can increase passenger numbers significantly (between 320% and 660% between 2013 and 2040 depending on the framework conditions). However, despite these expansions, the modal share of the BRT and the Gautrain in total passenger-kilometres travelled remains below 4% until 2040. This results in a decrease of cumulated GHG emissions of less than 1% until 2040 and relatively high GHG mitigation costs (4948–30045 ZAR2013/t CO2e). Nevertheless, a better integration of all public transport systems can increase the attractiveness of the services, which can result in a higher modal shift from private cars and thereby higher GHG emissions reductions at lower costs. Ó 2014 Elsevier Ltd. All rights reserved.

Introduction South Africa is the largest economy in Africa and has experienced strong economic growth since the end of apartheid in 1994, showing annual growth rates of up to 5.6% (IMF, 2013). Hand in hand with this economic growth goes the need for energy services and an increase in greenhouse gas (GHG) emissions. South African carbon emissions increased from 250 Mt CO2 in 1990 to about 350 Mt CO2 in 2010 (IEA, 2013). These figures correspond to per capita emissions of about 7–8 t CO2/capita, which is a figure comparable to industrialized countries such as the United Kingdom or Germany. Gauteng Province is the economic centre of South Africa contributing about one third of the national GDP. Growth rates have historically been higher than in the rest of the country (in average about 0.3%/a higher between 1996 and 2008 (StatsSA,

⇑ Corresponding author. Tel.: +49 711 685 87855; fax: +49 711 685 87873. E-mail address: [email protected] (S. Bubeck). http://dx.doi.org/10.1016/j.trd.2014.07.002 1361-9209/Ó 2014 Elsevier Ltd. All rights reserved.

58

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

2014; GPG, 2013)). GHG emissions allocated to Gauteng were about 127 Mt CO2e in 2009, of which about 25% can be attributed to transport activity (Tomaschek et al., 2012b). The economic dominance has caused steady migration to the province. Gauteng currently holds more than one fifth of the nation’s population, contributing only about 1.4% to the total land area (GCIS, 2011). This trend is forecasted to continue in future, making it possible to double Gauteng’s population and increase its GDP by more than 250% (constant prices) between 2010 and 2040 (Wehnert et al., 2011; DPTRW, 2013). The Province has realized its importance for South Africa and wants to play a leading role for the country’s climate protection strategy. The transport sector was proposed as one of the key starting areas. The expansion of the public transport system is seen as one promising option to mitigate GHG emissions (DLGH, 2010). The potential and costs of suitable measures have, however, not yet been evaluated. The public transport system of Gauteng has historically been mainly built on minibus-taxis as well as on a few scheduled bus and train services. Important expansions were the introduction of a bus rapid transit system (BRT) in Johannesburg (Rea Vaya) in 2009 and a rapid rail system for Gauteng (Gautrain) in 2010. The construction of the Rea Vaya BRT in Johannesburg is divided into different phases and began operation with phase 1A in August 2009 with a trunk route of 25.5 km. In addition, phase 1B (18 km trunk route) started operation in October 2013. The long-term BRT network shall comprise 330 km of trunk routes and provide BRT access to about 80% of Johannesburg’s population (DOT, 2009). The construction of the Gautrain rapid rail system started in September 2006 and the first phase began operation in June 2010, connecting the CBD Johannesburg with the international airport in the east of the Province. A second phase, completed in August 2011, linked Rosebank Station in Johannesburg and Hatfield Station in Pretoria. The current system network is of 80 km length. Future plans outline comprehensive expansions (DPTRW, 2013). However, a first evaluation shows that the aims of both projects were not achieved in terms of passenger numbers. For example, the Rea Vaya BRT phase 1A was planned with passenger levels of approximately 136000 pass./day (Grütter, 2011). Conversely, the actual passenger numbers reached only 38000 pass./day in February 2012 (Ngcobo, 2012), which corresponds to a level of achievement of about 28%. A similar statement has to be made for the Gautrain, where about 52000 daily passengers trips were performed in October 2013 (Venter, 2013a), whereas an official plan estimated about 108000 pass./ day (Venter, 2012b). To show and to evaluate the likely performance of the future BRT and the rapid rail system in Gauteng and to determine the potential for mitigating greenhouse gases, this paper aims to:    

Estimate the achievable passenger volumes of already planned as well as of possible further phases. Assess the mitigation potential for greenhouse gases. Evaluate the costs of possible further phases and their GHG mitigation costs. Give policy recommendations on how to set future public transport initiatives in order to achieve the overarching political aim of reducing GHG emissions.

All monetary figures in this paper are given in South African rand (ZAR) with the value of the year 2013. The exchange rate is 1 USD2013 = 9.5 ZAR2013 (FED, 2013). To calculate the net present value of future investments, we use a real discount rate of 8% according to DOE (2011).

Literature review Providing efficient transport systems is a central issue in fast growing urban areas such as in emerging and developing countries. Furthermore, the reduction of transport-related GHG emissions is a key challenge, which can be reached through various strategies. Nakamura and Hayashi (2013) give an overview of strategies and instruments for low-carbon solutions in urban transport. They find that the effectiveness of different GHG mitigation measures is highly dependent on the development process of the considered city and the types of urban land-use transport systems. To rank the cost-effectiveness of GHG mitigation measures in the transport sector, GHG mitigation costs can be considered (e.g. Kok et al., 2011; Lutsey and Sperling, 2009; Corbett et al., 2009). Kok et al. (2011) give a review on methodological approaches to evaluate the cost-effectiveness of measures for GHG mitigation in the transport sector. Lutsey and Sperling (2009) use a technology oriented, static, bottom-up approach in order to analyse the greenhouse gas mitigation potential of various technology options for the transport sector of the United States. Furthermore, alternative fuels were part of their analysis. They came to the result that many transport oriented mitigation options are available, which show low or even negative GHG mitigation costs, such as increases of vehicle efficiency through engine optimizations, reduced rolling resistance and aerodynamic improvements. Changes of the modal shift or expansion of the public transport system, however, were not part of their study. Such expansions represent another measure to reduce transport related GHG emissions aiming at a shift from private to public vehicles, which was inter alia analysed by Stanley et al. (2011). Also, the mitigation costs of an increased public transport share are the subject of different studies (e.g. Dedinec et al., 2013; Wright and Fulton, 2005). Wright and Fulton (2005), for example, examined GHG mitigation costs for the introduction of a BRT system considering the construction costs of network infrastructure. They did, however, not calculate the achievable modal share but assumed a modal share related to the introduction of a BRT assuming a modal split for a fictitious developing-nation city. They postulate that a diverse and

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

59

integrated package of measures promoting a shift towards low-carbon modes is the most cost-effective measure to achieve emission reductions. Numerous methodologies are available to estimate the modal split in a city (or country). Very simple analysis forms build on extrapolating historical travel demand development often based on indicators such as economic development or population growth. More complex analysis take into account possible decisions of travellers considering their travel needs and socio-economic characteristics (Wermuth, 2005). The relationship of living conditions and income to transport mode choice has, furthermore, been analysed and shown by numerous studies for many countries in the world (e.g. Kenworthy, 2003; Srinivasan and Ferreira, 2002; DPTRW, 2013; Mokonyama and Venter, 2007; Earnhart, 2003). Different methodologies are available to estimate the passenger performance of public transport expansions. Trip rate models can be applied to evaluate trip generation related to new stations in an expanded network (Balcombe et al., 2004). These models assume that trips generated by a new station are a function of the population living in the catchment area of this station. Trip rates are defined as the ratio of trips generated by a station and the number of inhabitants living within the catchment area of that station. They are determined on the basis of different existing stations and transferred to new stations in areas similar to the existing station. Trip end models (e.g. Blainey, 2010) or direct ridership models (e.g. Gutiérrez et al., 2011) represent an enhancement of trip rate models taking additional variables into account. These comprise socio-economic data of the considered area (e.g. proportion of employed inhabitants) as well as other variables such as the schedule of adjacent public transport services. The models are calibrated on the basis of data from existing stations (Balcombe et al., 2004). For Gauteng Province, a travel demand model is available which incorporates socio-economic aspects (Tomaschek et al., 2012a; Tomaschek, 2013). This model uses locally available travel surveys, which were used to derive homogeneous groups of travel behaviour (i.e. groups of persons which show similar travel characteristics) for Gauteng. These groups were built on the characteristics of car availability (as a function of household income) and employment. However, this model is built on surveys conducted before the introduction of the Rea Vaya and the Gautrain and therefore does not take their current network and possible future expansions into account. The impact of the introduction of the existing Rea Vaya BRT phases 1A and 1B on GHG emissions is analysed by Grütter (2011). The author estimates the GHG emissions reduction based on planned passenger volumes, assuming a modal shift from other transport modes. He finds decreasing GHG emissions, air pollution and increasing quality of life. The impact of the further development of the Rea Vaya or the Gautrain on GHG emissions has not been examined yet, nor has been the evaluation of the strategic option of public transport expansion through mitigation costs. Venter (2013b) analysed the success factors for formalizing the minibus sector and align it to the BRT. The author concluded, inter alia, that it is crucial to ensure that the BRT is financially attractive for its users and that the BRT can provide opportunities for the minibus-taxi industry. In this paper, we both determine the future modal share and evaluate the GHG mitigation costs of public transport expansions in a fast-growing megacity region whose economic power impacts on large areas of Africa. Material and methods Fig. 1 gives an overview of the employed method in this paper and the relevant data used. Starting from actual passenger volumes, we derive passenger volumes in an expanded network through a trip rate model, applying GIS calculations of the population in the service catchment area. From the passenger volumes, we calculate the future transport performance assuming an average trip distance. Subsequently, starting from a reference forecast of the modal split for Gauteng, we derive an adapted modal split including the Rea Vaya BRT and Gautrain network expansions on the basis of a modal shift from other modes to the Rea Vaya and the Gautrain. The adapted modal split leads to different well-to-wheel GHG emissions compared to the reference scenario, which we calculate taking into account the occupancy rates and the emission factors of each considered transport mode. Finally, we determine the mitigation costs of the public transport expansions based on their infrastructure costs, their vehicle costs and the delta in fuel costs between the adapted modal split and the reference scenario. To distinguish between possible passenger performance developments of the Rea Vaya BRT and the Gautrain, we look at different scenarios. In addition, we carry out a sensitivity analysis of the most relevant assumptions, i.e. the modal shift and the occupancy rates. Travel demand and vehicle data In this study, we use the available figures for a baseline development of travel demand in Gauteng until 2040 (Table 1) (Tomaschek et al., 2012a; Tomaschek et al., 2012b; Tomaschek, 2013). This reference scenario does not include infrastructure expansions and the actual increase in travel demand of the Rea Vaya and the Gautrain is only due to population growth. The assumed occupancy rates of the considered transport modes are shown in Table 1. To quantify their impact on the results, we vary the occupancy rates of the Rea Vaya and the Gautrain by ±30%. We use the energy consumption data for the different vehicles shown in Table 1, on the basis of energy balances for Gauteng (Tomaschek et al., 2012b). The average energy consumption of all vehicles (except rail-based vehicles) in the fleet in 2013 was about 11.2 litres of gasoline equivalent in terms of energy content (lge) per 100 km or 3.6 MJ/km. In future, we

60

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

Reference scenario (without infrastructure expansions)

Data

Scenarios A and B (with infrastructure expansions) Actual passenger volumes of the Rea Vaya and the Gautrain

Population in service catchment area of current network Actual trip rates (scenario A)

Increased trip rates (scenario B)

Population in service catchment area of expanded network Future passenger volumes of the Rea Vaya and the Gautrain Average trip distance Future transport performance of the Rea Vaya and the Gautrain Modal shift from other transport modes* Modal split for Gauteng 20112040 (reference scenario)

Adapted modal split Occupancy rates* and emission factors

GHG emissions in transport sector (reference scenario)

Adapted GHG emissions in transport sector Additional infrastructure, vehicle and fuel costs GHG mitigation costs

Fig. 1. Method followed. Parameters marked with an asterisk (⁄) have been varied in order to test for sensitivity.

assume that the fleet efficiency of all road vehicles increases by 1%/a from 2013 to 2040 according to IEA (2012b) and Edwards et al. (2011). Potentials of the infrastructure expansions To assess the future transport performance of the Rea Vaya and the Gautrain and modal split changes due to these systems, we consider three effects. First, the transport performance increases due to the expansion of the infrastructure, with new neighbourhoods having access to the transport systems (expansion effect). Second, it increases as the attractiveness of the services grows due to new destinations being served within the network and due to further integration with other modes (attraction effect). Third, the transport performance increases due to population growth (population effect). We calculate the expansion effect as in trip rate models, i.e. trip generation is determined by the population living within the service catchment area around new stations. The catchment area is defined as the area from which inhabitants have access to the public transport station, which we distinguish depending on the public transport mode. Since there is no projection for spatially differentiated socio-economic factors (such as income distribution or employment) available, these are not taken into account. Based on average travel distances in each transport system, we derive the transport performance from the generated trips. We calculate the trip rates used to estimate future transport performance assuming passenger numbers as in 2013. The trip rate for each system is used for new stations independently from the spatial location of the station, as statistics for

61

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69 Table 1 Basic assumptions considered for the calculation. Mode of transport

Bus Passenger car Minibus-taxi Motorcycle Passenger rail Gautrain Rea Vaya Non-motorized Total/average a

Occupancy rate [pkm/vkm]a

28 2.6 10 1 169 282 (±30%) 47 (±30%) 1

Transport performance in Gauteng [109 pkm/a]a

Well-to-wheel GHG emissions [Mt CO2e/a]

Energy consumption [MJ/vkm] a

2013

2020

2030

2040

2013

2020

2030

2040

2013

2040

10.2 76.7 41.0 0.2 14.1 0.7 0.5 5.1 148.4

11.3 89.1 45.4 0.2 15.6 0.8 0.5 5.6 168.5

12.6 106.2 50.5 0.2 17.4 0.9 0.6 6.0 194.4

14.0 125.3 55.7 0.2 19.2 1.0 0.7 6.4 222.6

0.58 13.02 2.75 0.03 0.31 0.01 0.02 0.00 16.72

0.61 14.16 2.84 0.03 0.34 0.01 0.02 0.00 18.01

0.61 15.33 2.85 0.03 0.38 0.01 0.02 0.00 19.25

0.61 16.41 2.85 0.04 0.43 0.01 0.02 0.00 20.36

12.2 3.3 5.1 1.5 46.2 39.5 15.4 0.0 3.8

9.3 2.5 3.9 1.1 46.2 39.5 11.8 0.0 2.9

Source: Tomaschek et al., 2012a; Tomaschek et al., 2012b; Tomaschek, 2013.

passenger numbers on a station level are not available. In scenario A, we assume this trip rate to be constant for future expansions (referred to as ‘actual trip rate’ in the following). Subsequently, in scenario B we use a trip rate calibrated on planned passenger volumes for the future expansions. The increase of the trip rate can be explained by a growing attractiveness of the services (attraction effect). In addition, the introduction of a mass transit station can facilitate an increased population density around this station. This trip rate is referred to as ‘increased trip rate’ in the following. To model the expansion of the network of the Rea Vaya, we divide it into different phases. In addition to the existing phases 1A (operating since 2009) and 1B (operating since 2013), we assume three further phases of the full phase 1 and take into account the long-term BRT network as shown in DOT (2009) and DPTRW (2013). The modelled evolution of the network is shown in Fig. 2, including the start of operation and the length of the trunk route network. The average travel distance in the Rea Vaya network in phase 1A was 27.1 km and is kept constant for the subsequent phases. The actual trip rate in phase 1A with 38000 pass./day (Ngcobo, 2012) is 3.68 trips/100 population, i.e. every 100 people living in the catchment area generate 3.68 trips a day. The increased trip rate assumes passenger volumes of 136000 pass./day as planned (Grütter, 2011), corresponding to 13.18 trips/100 population. To determine the service catchment area of the potential future network of lines, we use the geographic information system (GIS) ArcGIS 10 from ESRI Inc. One factor in the implementation of trip rate models is the distance that is selected to define the catchment area. For the Rea Vaya network, we limit the catchment area to 1 km around every station based on the maximum walking distance of 15 min that 95% of all passengers walk to a bus station according to an analysis of the households travel behaviour in South Africa (DOT, 2003) and an assumed walking speed of 4 km/h. Starting from current population figures at the district level and a population forecast (population effect) (Wehnert et al., 2011), we estimate the number of the current and future population having access to the transport system. To calculate the cumulated well-to-wheel GHG emissions over the considered period for future expansion phases (2014– 2040), we assume a start of operation for each phase supposing a constant speed of construction. Phases 1C to 1E of the Rea Vaya are assumed to start full operation at once, whereas the last phase (long-term BRT network) is assumed to be implemented continuously between 2021 and 2040. For the Gautrain, we define one potential expansion according to Gauteng’s transport master plan (DPTRW, 2013). The expansion reaches Roodepoort and Soweto in the west, Boksburg in the east and Pretoria East in the north (see Fig. 3). We assume the start of operation of the expansion in 2016, corresponding to the construction time of the previous phases. We consider a larger catchment area compared to the Rea Vaya, since the arrival to the stations takes place not only on foot, by bicycle and feeder buses, but also by car and other public transport modes. We divide the total catchment area into three zones with an assumed radius of 1, 5 and 10 km (see Fig. 3) with linearly decreasing attractiveness. The actual trip rates, based on 52,000 daily passengers trips (Venter, 2013a), amount to 3.14, 2.31 and 0.83 trips/100 population for the 1 km, 5 km and 10 km zones respectively. The increased trip rates, assuming planned passenger volumes of about 108000 pass./ day (Venter, 2012b), amount to 6.52, 4.80 and 1.72 trips/100 population. The average travel distance in the Gautrain network in 2013 was 33.7 km and is kept constant for the expansion. The transport performance of the Gautrain feeder buses is calculated as a proportion of the transport performance of the Gautrain. This proportion was about 5.1% in 2012 and is maintained constant for future phases. GHG mitigation costs of the infrastructure expansions To evaluate the expansion costs, for the Rea Vaya and for the Gautrain we consider the capital costs (CAPEX) for the construction of the infrastructure as well as its fixed operation and maintenance (FOM) costs, the CAPEX and FOM costs of the buses and train sets and the change in fuel costs of all modes from 2014 to 2040. According to Wright and Hook (2007), the costs of a BRT infrastructure range between 5 and 143 million ZAR2013/km – highly dependent on the features of the BRT system such as elevated stations or pre-board fare collection. To consider the Rea Vaya features and the local nature of costing, we derive the infrastructure costs of the Rea Vaya from the actual costs

62

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

Phase 1A (reference scenario): 26 km Phase 1B (reference scenario): 44 km (start of operation: August 2009)

(start of operation: October 2013)

Johannesburg CBD

Phase 1C: 55 km

Phase 1D: 74 km

(modelled start: 2014)

(modelled start: 2015)

Phase 1E: 122 km

Long-term network: 330 km

(modelled start: 2020)

(modelled start: 2021-2040)

Legend Trunk Route Complementary Route Feeder Route Wards Johannesburg

Legend Trunk Route SPTN Network

Fig. 2. Modelled expansions of the Rea Vaya network. For each phase, the network of trunk, complementary and feeder routes is represented on the map of Johannesburg. The cumulated length of the trunk route network (segregated lines) is indicated for each phase, as well as the start of operation. The term ‘SPTN Network’, shown in the long-term network graph, refers to the so-called Strategic Public Transport Network, which is part of the City of Johannesburg’s Integrated Transport Plan 2003–2008. Sources: Own graph in ArcGIS 10 (GIS data source: Oliver, D., Manager GIS, City Power (Pty) Ltd, Johannesburg, personal communication, 10.06.2011) and DOT (2009).

in phase 1A which amounted to approximately 1.3 billion ZAR2013, i.e. 52 million ZAR2013/km (Venter, 2011). The costs for the trunk routes are calculated in detail based on their length, as they cause the majority of the construction costs due to

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

63

Pretoria Station (Tshwane)

OR Tambo International Airport Station Park Station (Johannesburg)

Legend Stations Network 2013 (reference scenario) Stations Expansion Railway Feeder Bus Lines Catchment Area 1 km Catchment Area 5 km Catchment Area 10 km Wards Gauteng

Fig. 3. Network of the Gautrain in 2013 and modelled expansion on the map of Gauteng. The catchment area is divided into three zones. The feeder bus lines are not modelled for the expansion, since there is no indication of their possible routes. The illustrated railway line does not reflect the actual route, but only the link between the stations. Source: Own graph in ArcGIS 10.

the segregated lanes and elevated stations installed on these routes. The costs of the complementary and feeder routes are calculated as a share of the costs of the trunk routes of 4.5% (COC, 2011). The construction costs are comparably low, as these routes use conventional bus stations and operation is performed on existing roadways. The capital costs of the buses are calculated considering the number of buses of future expansions (based on the length of the trunk routes and an assumed constant annual mileage) and their investment costs. We assume the costs of a BRT articulated bus to be 3.3 million ZAR2013 and of a BRT rigid bus 2.6 million ZAR2013 (Rea Vaya, 2011). Investment costs for a Gautrain 4-car-train-set are calculated as in Baumgartner (2001) and sum up to 292 million ZAR2013.

64

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

Wright and Hook (2007) indicate infrastructure costs of light rail transit systems between 124 and 380 million ZAR2013/ km. The infrastructure costs of the 2013 Gautrain network amounted to approximately 28 billion ZAR2013 (Venter, 2012a). From this figure we derive the specific costs per kilometre of rail network, which we assume as future infrastructure costs of 352 million ZAR2013/km. The fuel cost savings achievable by a higher modal share of the Rea Vaya and the Gautrain are calculated using the price for petrol and diesel in South Africa (DOE, 2013) and on the specific energy consumption of the particular vehicles (Table 1). In South Africa, the fuel price is compound of a basic fuel price (BFP), based on the import price of crude oil, as well as on additional margins, taxes and levies. To forecast the product costs we assume the BFP to be correlated to the world crude oil price as given in the IEA’s World Energy Outlook 2012 (IEA, 2012a), which states a crude oil import price of 145 USD2011/barrel in 2035. Furthermore, we assume the additional mark-ups for delivery and transport as well as wholesale mark-ups to be constant (in real values). In result, the considered petrol price increases from 0.24 ZAR2013/ MJ (7.92 ZAR2013/l) in 2013 to 0.29 ZAR2013/MJ (9.53 ZAR2013/l) in 2040. The corresponding figures for diesel are 0.23 ZAR2013/MJ (8.31 ZAR2013/l) and 0.28 ZAR2013/MJ (10.10 ZAR2013/l) in 2013 and in 2040, respectively. For the electricity price (0.22 ZAR2013/MJ in 2013), we assume a growth rate of 0.5%/a, which corresponds to 0.25 ZAR2013/MJ in 2040. Taxes and levies are excluded from our calculation, since we analyse the costs from a macroeconomic perspective. This contrasts with a microeconomic view point, where the prices paid by passengers (including all taxes and levies) would be considered. To evaluate the climate impact of the expansion of the Rea Vaya and the Gautrain, we assess the change in well-to-wheel GHG emissions from 2014 to 2040. First, we calculate the changed modal split compared to the reference scenario, i.e. the existing infrastructure of the Rea Vaya (phases 1A and 1B) and the Gautrain. We assume that every additional person-kilometre travelled on the Rea Vaya and the Gautrain is travelled less on the other transportation modes. This way, the total transport performance remains identical compared to the reference scenario. Table 2 shows the assumed modal shift from the existing transport modes, on the basis of a survey among Rea Vaya passengers (Grütter, 2011) and is also applied to the Gautrain. The percentage indicated for minibus-taxis, for example, means that every 100 additional pkm travelled on the Rea Vaya or the Gautrain lead to a reduction of 61 pkm travelled in minibus-taxis. To assess the impact of the assumed modal shift, we also calculate an alternative where we swap the percentage of the minibus–taxis and the one of private cars (see Table 2). This will result in a more optimistic result with a more positive impact on GHG emissions. Second, with the adapted modal split and the occupancy rates, we derive the vehicle mileage of the different modes. Finally, using specific emission factors, the change in well-to-wheel emissions of the considered greenhouse gases (CO2, CH4 and N2O) are derived from the vehicle mileage. GHG emissions are converted to CO2-equivalents (CO2e) based on the IPCC Third Assessment Report, considering a global warming potential (GWP 100) of carbon dioxide CO2 = 1, methane CH4 = 23 and nitrous oxide N2O = 296 (IPCC, 2001). The following formula shows how the well-to-wheel GHG emissions of transport mode i are calculated for a specific year.

GHGi ¼

TP i  EC i  EF j ORi

GHGi = annual well-to-wheel GHG emissions of transport mode i [g CO2e/a] TPi = annual transport performance of transport mode i [pkm/a] ORi = occupancy rate of transport mode i [pkm/vkm] ECi = energy consumption of transport mode i [MJ/vkm] EFj = well-to-wheel GHG emission factor of energy carrier j [g CO2e/MJ] i = transport mode {bus, passenger car, minibus-taxi, motorcycle, passenger rail, Gautrain, Rea Vaya} j = energy carrier {petrol, diesel, electricity} The emission factors amount to 131.9 g CO2e/MJ for petrol and 131.7 g CO2e/MJ for diesel as in Tomaschek et al. (2012c). For electricity provision, we consider a GHG emission factor of 291 g CO2e/MJ (i.e. 1048 g CO2e/kWh) (Telsnig et al., 2013). To make the calculation manageable, we do not analyse various sources of fuels and electricity. Instead we consider an average electricity and fuel mix of today for the GHG emissions of the energy provision. To specify the GHG mitigation costs of the public transport expansions, we apply the following formula, considering the CAPEX and FOM costs of the infrastructure, the public transport vehicles as well as fuel costs over the period 2014 to 2040. Table 2 Modal shift (previous mode used by Rea Vaya and Gautrain passengers). The percentages indicated refer to the transport performance in pkm.

a

Mode of transport

Modal shift (as observed)a (%)

Modal shift (alternative) (%)

Minibus-taxi Rail and motorcycle Passenger car Bus Non-motorized transport

61 18 10 8 3

10 18 61 8 3

Source: Grütter (2011).

65

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

P2040

P  CAPEX FOM CAPEX FOM P  þIFOM þV CAPEX þV FOM þ F Þ ðIt þIt þV t þV t þ i F t;i Þa ðICAPEX t t t t i t;i ref

t¼2014

Discounted GHG mitigation costs : GHG MC ¼

P2040 P t¼2014

ð1þrÞðt2014Þ i GHGt;i;ref



P

i GHGt;i;a



GHG MC = discounted GHG mitigation costs [ZAR2013/t CO2e] ICAPEX = annualized CAPEX of public transport infrastructure expansions [ZAR2013/a] VCAPEX = annualized CAPEX of public transport vehicles [ZAR2013/a] IFOM = annual FOM cost of public transport infrastructure expansions [ZAR2013/a] VFOM = annual FOM cost of public transport vehicles [ZAR2013/a] Fi = annual fuel costs of transport mode i [ZAR2013/a] GHGi = annual GHG emissions of transport mode i [t CO2e/a] r = discount rate [%], i.e. 8% a = adapted modal split for scenarios A/B (incl. network expansion of Gautrain and Rea Vaya) ref = reference modal split (Gautrain and Rea Vaya infrastructure as in 2013) i = transport mode {bus, passenger car, minibus-taxi, motorcycle, passenger rail, Gautrain, Rea Vaya} t = year We assume an economic life of the BRT and the Gautrain infrastructure of 50 years and annual FOM costs of 4.5% of the annualized CAPEX. Results Potentials of the infrastructure expansions The resulting additional transport performance for both the actual and the increased trip rate is shown in Table 3. The application of the actual trip rate (scenario A) results in a relatively low patronage growth of about 370% until 2040 for the full expansion of the Rea Vaya. The reason is that phase 1A covers areas of high population density (the centre of Johannesburg and Soweto) and therefore the actual trip rate is relatively low. With the increased trip rate assuming passenger volumes as planned, the passenger numbers of the Rea Vaya can increase to a significant degree. In result, the patronage for the long-term network in the year 2040 is more than 2.5 times higher than under actual trip rates. Despite the far less extensive network expansion, the Gautrain can increase passenger numbers by about 280% until 2040 based on the actual trip rate, since the expansion reaches densely populated areas such as Soweto. The increased trip rate leads to a rise in transport performance by a factor of 1.5 in the year 2040 compared to the actual trip rate. Table 4 shows the resulting modal split with all expansions of the Rea Vaya and the Gautrain being realised. Despite the remarkable relative growth that is possible for both systems, the Rea Vaya and the Gautrain together reach a modal split of only between 1.6% and 3.2% in 2040, depending on the underlying trip rate. Passenger cars and minibus-taxis remain the dominant modes with a cumulated modal share of about 80%. Well-to-wheel GHG emissions and GHG mitigation costs Considering the actual transport performance of the Rea Vaya and the Gautrain with the existing infrastructure in 2013, the impact on GHG emission mitigation is relatively moderate. The achieved reduction in GHG emissions in phases 1A and 1B of the Rea Vaya amounts to about 14690 t CO2e in 2013. The operation of the Gautrain with its actual passenger numbers leads to a GHG emissions reduction of about 25600 t CO2e in 2013. In result, both systems together reach a GHG emissions reduction of less than 0.3% in 2013, compared to a transport sector without the Rea Vaya and the Gautrain.

Table 3 Transport performance of the different phases of the Rea Vaya and the Gautrain. Phase

Rea Rea Rea Rea Rea Rea

Vaya Vaya Vaya Vaya Vaya Vaya

Year of full operation

1A 1B 1C 1D 1E long-term network

Gautrain network 2013 Gautrain expansion

Additional transport performance in year of full operation [103 pkm/day]

Cumulated transport performance in year of full operation [103 pkm/day]

Scenario A: actual trip rate

Scenario B: increased trip rate

Scenario A: actual trip rate

Scenario B: increased trip rate

2011 2013 2014 2015 2020 2040

1031 188 141 279 585 1528

1031 188 504 997 2092 5470

1031 1269 1438 1746 2485 4659

1031 1269 1809 2842 5186 12004

2013 2016

1793 1541

1793 3198

1793 3459

1793 5116

66

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

Table 4 Travel demand (modal split in% of pkm) for the full expansion of the Rea Vaya and the Gautrain network. Mode of transport

Scenario A: actual trip rate

Bus Passenger car Minibus-taxi Motorcycle Passenger rail Gautrain Rea Vaya Non-motorized Total

Scenario B: increased trip rate

2013

2020

2030

2040

2013

2020

2030

2040

6.8 51.7 27.6 0.1 9.5 0.4 0.3 3.4 100.0

6.7 52.8 26.6 0.1 9.1 0.8 0.5 3.3 100.0

6.5 54.6 25.5 0.1 8.8 0.8 0.7 3.1 100.0

6.2 56.2 24.5 0.1 8.5 0.8 0.8 2.9 100.0

6.8 51.7 27.6 0.1 9.5 0.4 0.3 3.4 100.0

6.6 52.7 26.0 0.1 9.0 1.2 1.1 3.3 100.0

6.4 54.4 24.7 0.1 8.6 1.2 1.6 3.0 100.0

6.1 56.0 23.5 0.1 8.2 1.2 2.0 2.8 100.0

Table 5 Cumulated costs, GHG emissions reduction and GHG mitigation costs from 2014 to 2040. Infrastructure Vehicle capital Fuel cost savings costs [106 ZAR2013] costs [106 ZAR2013] [106 ZAR2013] Scenario Scenario A: actual B: increased trip rate trip rate

Phase

Rea Rea Rea Rea

Vaya Vaya Vaya Vaya

1C 1D 1E long-term network

532.3 872.5 1335.1 2118.9

Rea Vaya Phase 1C to long-term network Gautrain expansion

297.8 488.2 747.0 1185.4

32.1 56.7 72.8 57.6

115.0 203.0 260.4 206.1

GHG reduction [103 t CO2e]

GHG mitigation costs [ZAR2013/t CO2e]

Scenario Scenario Scenario Scenario A: actual B: increased A: actual B: increased trip rate trip rate trip rate trip rate 35.5 65.3 102.9 108.1

127.1 233.5 368.2 386.7

22465 19983 19534 30045

5625 4957 4948 8011

4858.8

2718.4

219.2

784.5

311.7

1115.6

23606

6089

15766.1

5117.6

533.9

1108.2

783.4

1626.0

25978

12162

Table 6 Reduction of cumulated GHG emissions from 2014 to 2040 compared to the reference scenario [%]. GHG mitigation costs [ZAR2013/t CO2e] are indicated in parenthesis. Phase

Scenario A: actual trip rate

Scenario B: increased trip rate

Occupancy rate + 30%

Occupancy rate –30%

Alternative modal shift

Occupancy rate + 30%

Occupancy rate  30%

Alternative modal shift

Rea Vaya 1C–1E

0.04% (20189)

0.06% (14058)

0.01% (91238)

0.12% (6028)

0.14% (5069)

0.20% (3382)

0.03% (24622)

0.44% (1158)

Rea Vaya long-term network

0.06% (30045)

0.09% (21142)

0.01% (129715)

0.19% (9421)

0.22% (8011)

0.31% (5541)

0.05% (35666)

0.68% (2280)

Gautrain Expansion

0.15% (25978)

0.16% (24512)

0.14% (29199)

0.31% (12640)

0.32% (12162)

0.34% (11447)

0.29% (13734)

0.64% (5724)

For the assessment of the future expansions compared to the reference scenario (with the Gautrain and Rea Vaya infrastructure as in 2013), the considered costs, the GHG emissions reduction and the resulting mitigation costs are shown in Table 5. Assuming the actual trip rate being applicable for future expansions (scenario A), the well-to-wheel GHG emissions reduction of the Gautrain is about 2.5 times higher than the one of the Rea Vaya (with the long-term network being realised). For the increased trip rate (scenario B), the difference between both systems is less considerable (factor 1.5), as the actual and the planned passenger numbers of the Rea Vaya differ more than the same figures for the Gautrain. The BRT has far smaller infrastructure and vehicle costs than the Gautrain, but also lower fuel cost savings. For the actual trip rate, we find that the mitigation costs of both systems are in the same range. For the increased trip rate, the mitigation costs of the Gautrain are about two times higher than the ones for the Rea Vaya. In order to test the results for sensitivity, we vary different parameters (ceteris paribus). Assuming an occupancy rate of the Rea Vaya increased or decreased by 30%, the GHG mitigation for the Rea Vaya increases by about 42% and decreases by about 77% respectively, both for the actual and the increased trip rate. For the Gautrain, these figures amount to +6% and 11% respectively. The most sensitive parameter is the modal shift. If the passengers shift mainly from minibus–taxis as observed (Table 2), the GHG reduction is relatively low due to the high occupancy rate of minibus-taxis and thereby their low specific emissions. Altering the modal shift towards a higher change from private cars with low occupancy rates (Table 2), the GHG reduction increases by about 210% for the Rea Vaya and about 100% for the Gautrain. However, the relative reduction of cumulated GHG emissions compared to the reference scenario, shown in Table 6, remains below 1% for all expansions and all parameter variations.

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

67

Conclusions and policy recommendations In this paper, we analysed the possible performance of future expansions of a BRT and of a high-speed rail system in terms of their passenger volumes, GHG mitigation potential and GHG mitigation costs. The study region Gauteng province is the socio-economic hub of South Africa expecting rapid economic and population growth. Moreover, the provincial government plans to significantly reduce the GHG emissions especially in the transport sector, which made this region an ideal study area for our analysis. To do so, we combined and adjusted available methods for calculating travel demand, infrastructure costs and GHG mitigation costs. Moreover, we extended those methods where it seemed beneficial, for example by including vehicle fuel costs into our calculation. However, we also had to make some simplifications as some data were not available for the study region of Gauteng such as spatially distributed assumptions on future employment figures or income distribution. The analysis showed that it seems reasonable that planned new phases as well as additional future expansions for the BRT and the high-speed rail can significantly increase passenger volumes. Whereas the current system shows a passenger performance of about 90000 pass./day for the BRT (Rea Vaya) and the rapid rail link (Gautrain), a full system could provide service for more than about 320000 pass./day assuming a constant trip rate based on actual passenger numbers. The analysis showed furthermore, that it is necessary to increase the passenger attraction of the Gauteng Rea Vaya and the Gautrain, in order to make these public transport systems reach a higher share in total transport volume in the province. If trip rates as initially planned for both systems were reached, the passenger volume for the full system would be increased to 660000 pass./day, which equals a growth of 106% in comparison to the assumption of a trip rate according to actual passenger numbers. However, despite the relative growth, the modal share of both systems in 2040 remains below 4% of total road and rail passenger transport. In terms of the GHG mitigation potential, we found that both systems, the Rea Vaya BRT as well as the Gautrain rapid rail link, cannot achieve significant reductions in total transport well-to-wheel GHG emissions. The overall cumulated well-towheel GHG mitigation potential found was between 1.1 and 2.7 Mt CO2e for the full expansion of both systems between 2014 and 2040 (depending on the assumed future trip rates). However, it has become obvious that it is important to change the target group when implementing future phases. Whereas the current systems mainly compete with minibus-taxis, further expansions should focus on mainly attracting motorized individual transport. The corresponding alternative modal shift would result in an increase of well-to-wheel GHG mitigation potential to 2.7–7.0 Mt CO2e and furthermore reduced GHG mitigation costs from about 10000–25000 to about 4000–11000 ZAR2013/t CO2e, depending on the assumed trip rate. Other transport related GHG mitigation measures such as an increased use of alternative fuels are found to show lower mitigation costs. Thus, expanding the public transport infrastructure is not a cost-effective GHG mitigation option but rather an improvement of public transport service quality. The mitigation costs for the Rea Vaya BRT are significantly higher compared to figures given in Wright and Fulton (2005), who calculated mitigation costs of a fictitious BRT implementation to 66 USD or 627 ZAR2013/t CO2 assuming a higher modal share of 5% for the BRT, not including costs for fuel and vehicles. In order to achieve higher trip rates and to increase the GHG mitigation potential, we recommend that the future transport system of Gauteng should better integrate the various public transport modes, including minibus-taxis. Those minibuses provide a flexible and energy efficient system which could ideally serve as feeder and distribution system for the BRT, for the Gautrain and for commuter buses. Thus, the minibus system should be standardized to ensure safety, security, reliability, availability as well as quality standards and to make it available for broader parts of the society. Moreover, existing policy and regulations should be implemented with due cognizance of the role of minibus-taxis on greenhouse gas mitigation efforts. Possible routes could be defined in a stakeholder process to develop an integrated ticketing system for the whole public transport network, where departure and arrival times as well as service frequency need to be coordinated. An accompanying information campaign could demonstrate the advantages of the new systems (e.g. shorter travel times, lower costs, easier use and environmental benefits) to the public society and would thus increase the acceptance, too (refer e.g. to TCRP (2007) for an overview of available research findings on BRT passenger attraction, and to Wright and Hook (2007) for an analysis of stakeholder processes and communication strategies for BRT implementation). Modal split models take socio-demographic factors into account and they are thus appropriate for a use in developing countries with rapidly changing living conditions. However, a comprehensive data base is necessary to perform such models. In this paper, we presented a modified method to evaluate the effect of expansions of public transport systems on passenger volumes and GHG mitigation. As this method is restricted to available data for Gauteng, we had to include some limitations. Thus, the considered relationships between passenger volume and network expansion are largely linear. So are the investment costs for the modelled interventions. In addition, a non-linear growth of the trip rate due to network expansions could be considered. Additionally, we assume a constant transport performance for reasons of simplicity. However, there are studies that found that generally a switch from private car to public transport reduces the total pkm of the passengers. This is due to the fact that typically the travel time is fixed and not the travelled distance. The quality of the results could, furthermore, be improved by using genuine input figures, e.g. for willingness to switch modes as well as more detailed ridership figures for the province. These could for example be derived by regularly surveys of the traffic participants. The analysis has shown that public transport systems can contribute to GHG mitigation if well implemented. However, other aspects go obviously in hand with the provision of transport services such as convenience, time savings or safety,

68

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

which were not part of the analysis. We focused our analysis on well-to-wheel emissions but did not analyse different sources of electricity and fuel provision. Including alternative provision chains into further analysis would thus be reasonable. Moreover, pollutant emissions were not part of this study but might be important to consider in future studies. Acknowledgements This research is part of the Megacity Research Project ‘EnerKey’, which is a German-South African collaboration to develop an integrated energy and climate change concept for Gauteng province in South Africa. It is funded by the German Federal Ministry of Education and Research. References Balcombe, R., Mackett, R., Paulley, N., Preston, J., Shires, J., Titheridge, H., Wardman, M., White, P., 2004. The Demand for Public Transport: A Practical Guide. Report TRL593 of the Transport Research Laboratory (TRL), London. Baumgartner, J.P., 2001. Prices and Costs in the Railways Sector, Laboratoire d’Intermodalité des Transports et de Planification, École Polytechnique Fédérale de Lausanne, Lausanne. Blainey, S., 2010. Trip end models of local rail demand in England and Wales. J. Transp. Geogr. 18 (1), 153–165. City of Cape Town (COC), 2011. BRT Infrastructure in Cape Town, Conference Presentation, Southern African Transport Conference 11. to 14. July 2011 in Pretoria. Corbett, J.J., Wang, H., Winebrake, J.J., 2009. The effectiveness and costs of speed reductions on emissions from international shipping. Transp. Res. Part D: Transp. Environ. 14 (8), 593–598. Dedinec, A., Markovska, N., Taseska, V., Duic, N., Kanevce, G., 2013. Assessment of climate change mitigation potential of the Macedonian transport sector. Energy 57, 177–187. Department of Local Government and Housing (DLGH), 2010. Gauteng Integrated Energy Strategy, Johannesburg. Department of Energy (DOE), 2011. Integrated Resource Plan For Electricity 2010–2030, Revision 2, Final Report, Pretoria. Department of Energy (DOE), 2013. Petrol Price Archive, Pretoria. http://www.energy.gov.za/files/petroleum_frame.html, accessed on 22.10.2013. Department of Transport South Africa (DOT), 2003. The First South African National Household Travel Survey (NHTS), Pretoria. http://nasp.dot.gov.za/ projects/nts/Default.html, accessed on 12.06.2009. Department of Transport South Africa (DOT), 2009. National Transport Master Plan 2005–2050 – Gauteng Province, Phase 2: Analysis, Second Draft, Pretoria. Gauteng Department of Public Transport, Roads and Works (DPTRW), 2013. 25-Year Integrated Transport Master Plan, Draft Plan, Johannesburg. Earnhart, D., 2003. Do travel cost models value transportation properly? Transp. Res. Part D: Transp. Environ. 8 (6), 397–414. Edwards, R., Larive, J., Beziat, J., 2011. Well-to-wheels analysis of future automotive fuels and powertrains in the European context, Tank-to-wheels Report, Version 3c, Petten. The Federal Reserve System (FED), 2013. Historical Rates for the South African Rand, Washington DC. http://www.federalreserve.gov/releases/h10/hist/ dat00_sf.htm, accessed on 24.10.2013. Government Communication and Information System (GCIS), 2011. South Africa Yearbook 2010/2011 – Chapter 1, Pretoria. http://www.gcis.gov.za/ resource_centre/sa_info/yearbook/2010-11/chapter1.pdf, accessed on 18.03.2012. Gauteng Provincial Government (GPG), 2013. Provincial Economic Review and Outlook 2013. http://www.treasury.gpg.gov.za/Document/Documents/ PERO%202013.pdf, accessed on 16.06.2014. Grütter, J., 2011. BRT Rea Vaya Phase 1a and 1b, South Africa, VCS Project Document, Version 1.2, Reinach. Gutiérrez, J., Cardozo, O.D., García-Palomares, J.C., 2011. Transit ridership forecasting at station level: an approach based on distance-decay weighted regression. J. Transp. Geogr. 19 (6), 1081–1092. International Energy Agency (IEA), 2012a. World Energy Outlook 2012, Paris. International Energy Agency (IEA), 2012b. Technology Roadmap Fuel Economy of Road Vehicles, Paris. International Energy Agency (IEA), 2013. CO2 Emissions from Fuel Combustion Statistics, Paris. International Monetary Fund (IMF), 2013. World Economic Outlook Database. April 2013, Washington DC. IPCC, 2001. IPCC Third Assessment Report – Climate Change 2001, Cambridge. Kenworthy, J.R., 2003. Transport energy use and greenhouse gases in urban passenger transport systems: a study of 84 global cities. In: Proceeding of the Third Conference of the Regional Government Network for Sustainable Development. Notre Dame University, Fremantle. Kok, R., Annema, J.A., Van Wee, B., 2011. Cost-effectiveness of greenhouse gas mitigation in transport: a review of methodological approaches and their impact. Energy Policy 39 (12), 7776–7793. Lutsey, N., Sperling, D., 2009. Greenhouse gas mitigation supply curve for the United States for transport versus other sectors. Transp. Res. Part D: Transp. Environ. 14 (3), 222–229. Mokonyama, M., Venter, C., 2007. Forecasting household car ownership in South Africa: alternative models and future trends. J. South Afr. Inst. Civil Eng. 49 (3), 2–10. Nakamura, K., Hayashi, Y., 2013. Strategies and instruments for low-carbon urban transport: an international review on trends and effects. Transp. Policy 29, 264–274. Ngcobo, J., 2012. Operational Data of the Rea Vaya BRT. Director Operations, Transportation Department, City of Johannesburg, Personal Communication, 09.03.2012. Rea Vaya, 2011. Funding Rea Vaya buses, Johannesburg. http://www.reavaya.org.za/news-archive/august-2011/711-funding-rea-vaya-buses, accessed on 07.03.2012. Srinivasan, S., Ferreira, J., 2002. Travel behavior at the household level: understanding linkages with residential choice. Transp. Res. Part D: Transp. Environ. 7 (3), 225–242. Stanley, J.K., Hensher, D.A., Loader, C., 2011. Road transport and climate change: stepping off the greenhouse gas. Transp. Res. Part A: Policy Pract. 45 (10), 1020–1030. Statistics South Africa (StatsSA), 2014. Gross Domestic Product, First Quarter 2014. http://beta2.statssa.gov.za/publications/P0441/ P04411stQuarter2014.pdf, accessed on 16.06.2014. Transit Cooperative Research Program (TCRP), 2007. Bus Rapid Transit Practitioner’s Guide, Report 118. Transportation Research Board (TRB), Washington DC. Telsnig, T., Tomaschek, J., Özdemir, E.D., 2013. Assessment of selected CCS technologies in electricity and synthetic fuel production for CO2 mitigation in South Africa. Energy Policy 63, 168–180. Tomaschek, J., 2013. Long-term Optimization of the Transport Sector Addressing Greenhouse Gas Reduction Targets Under Rapid Growth – Application of an Energy System Model for Gauteng Province, South Africa. PhD Thesis, University of Stuttgart, IER, Stuttgart. Tomaschek, J., Dobbins, A., Fahl, U., 2012a. A regional TIMES model for application in Gauteng. In: Proceedings of the 31st International Energy Workshop (IEW), Cape Town.

S. Bubeck et al. / Transportation Research Part D 32 (2014) 57–69

69

Tomaschek, J., Haasz, T., Dobbins, A., Fahl, U., 2012b. Energy Related Greenhouse Gas Inventory and Energy Balance Gauteng: 2007–2009, Stuttgart. http:// www.enerkey.info/images/stories/intern/module1/ghg_inventory_gauteng_october_2012.pdf, accessed on 18.03.2012. Tomaschek, J., Özdemir, E.D., Fahl, U., Eltrop, L., 2012c. Greenhouse gas emissions and abatement costs of biofuel production in South Africa. GCB Bioenergy 4 (6), 799–810. Venter, I., 2011. Joburg to Roll Out R1,2bn Rea Vaya Phase 1B by 2012, Johannesburg. http://www.engineeringnews.co.za/article/joburg-to-roll-out-r12billion-rea-vaya-phase-1b-by-2012-2011-02-07, accessed on 22.03.2012. Venter, I., 2012a. R280m Bombela top-up likely as toll delay affects Gautrain numbers, Johannesburg. http://www.engineeringnews.co.za/article/gautrain2012-02-10, accessed on 18.02.2012. Venter, I., 2012b. Rosebank–Park Station Gautrain Leg to be Operational by April, Johannesburg. http://www.engineeringnews.co.za/article/gautrain-201202-03, accessed on 18.02.2012. Venter, C., 2013. The lurch towards formalisation: lessons from the implementation of BRT in Johannesburg, South Africa.. Res. Transp. Econ. 39 (1), 114– 120. Venter, I., 2013a. Gautrain Agency Mulls Buying 40 New Rail Cars as Ridership Breaches 50,000, Johannesburg. http://www.engineeringnews.co.za/article/ gautrain-agency-mulls-buying-40-new-rail-cars-as-ridership-breaches-50-000-2013-10-01, accessed on 10.10.2013. Wehnert, T., Knoll, M., Rupp, J., 2011. Socio-economic Framework for 2010 set of EnerKey Energy Scenarios – Summary of Key Figures, Berlin. http:// www.enerkey.info/images/stories/intern/module2/Documents/enerkey%20m2%20scenario%20assumptions%202010_may%202011.pdf, accessed on 07.10.2013. Wermuth, M., 2005. Modellvorstellungen zur Prognose. In: Steierwald, G., Künne, H.D., Vogt, W. (Eds.), Stadtverkehrsplanung: Grundlagen Methoden Ziele, second ed. Springer, Berlin, pp. 243–295. Wright, L., Fulton, L., 2005. Climate change mitigation and transport in developing nations. Transp. Rev. 25 (6), 691–717. Wright, L., Hook, W., 2007. Bus Rapid Transit Planning Guide, third ed. New York.