SUSTAINABLE FREIGHT TRANSPORT IN SOUTH AFRICA : DOMESTIC INTERMODAL SOLUTIONS

Due to the rapid deregulation of freight transport in South Africa two decades ago, and low historical investment in rail (with resultant poor service delivery), an integrated alternative to road and rail competition was never developed. High national freight logistics costs, significant road infrastructure challenges and environmental impact concerns of a road-dominated freight transport market have, however, fuelled renewed interest in intermodal transport solutions. In this article, a high-level business case for domestic intermodal solutions in South Africa is presented. The results demonstrate that building three intermodal terminals to connect the three major industrial hubs (i.e. Gauteng, Durban and Cape Town) through an intermodal solution could reduce transport costs (including externalities) for the identified 11.5 million tons of intermodalfriendly freight flows on the Cape and Natal corridors by 42% (including externalities).


INTRODUCTION
Domestic intermodal solutions have been proposed as a key element to address South Africa's freight logistics challenges with specific reference to improving economies of density and size, increasing access for marginalised rural economies and sustainably lowering logistics costs (DoT, 1998(DoT, , 2005; CSIR, Imperial Logistics and University of Stellenbosch, 2009;Van Eeden and Havenga, 2010).
Intermodal services in South Africa are, however, still commonly understood to denote the movement of import and export containers -the concept of 'domestic' intermodal services does not really exist.As such, the development of intermodal technology has kept pace with the demands of shipping companies, but developments in domestic intermodal services have been neglected.
According to De Witt and Clinger (1999), domestic intermodal services are 'a significant and critical factor in the execution of supply chains'.Peetermans and Sellnick (2010) confirm that in Europe there is an increasing trend towards companies delivering both domestic and international services, with the number of companies providing both services increasing from 45% to 57% of total intermodal service providers between 2005 and 2009, and their TEU (twenty-foot equivalent unit) weighted market share increasing from 68% to 80%.Over the same period, intermodal operators providing only domestic services have on average grown their businesses; the opposite occurred with companies providing only cross-border intermodal services.
A regional study (in North Dakota, USA) conducted by Berwick (2001) demonstrated the benefits of domestic intermodal transport, including a reduction in overall transport costs, an increase in economic productivity, a reduced burden on highway infrastructure, higher returns from public and private infrastructure investments, reduced energy consumption, and increased safety.Brown and Hatch (2002) highlight the potential for intermodal freight transport to become the core of America's long-distance freight transport market, and even foresee a role for intermodal in middle-and short-distance markets.Strong growth in domestic intermodal freight is also reported for other countries, such as Italy and Germany (Silborn, 2008:42), which is significant as Germany holds a dominant position (almost half) in the European domestic intermodal market (Woxenius and Bärthel, 2008:27).One of the key driving forces behind this growth in intermodal freight transport is that intermodality allows each transport mode to utilise its core strength in building supply chains that are on the whole more efficient, cost-effective and sustainable.
The European Union's confidence in an optimal balance between road and rail freight transport is evidenced by its budget of €450 million for the period 2007-2013 for the Marco Polo programme.New projects that shift freight from road to rail, sea or inland waterways are co-funded during the start-up phase before the projects become profitable.
The programme aims to free Europe's roads of an annual volume of 20 billion tonkilometres (tonkm) of freight, the equivalent of more than 700 000 trucks a year travelling between Paris and Berlin (European Commission, 2011).
Similarly, in South Africa policymakers have expressed the desire for a modal shift and a domestic intermodal solution (DoT, 2005 and2011:7-8;The Presidency, 2011).Private partners have also expressed an interest in participating in such a solution through both direct investment (Naidoo, 2011) and co-operation using licensed technology such as RoadRailer (Grove, 2005: 19).In fact, as far back as 1995 in a survey conducted by Havenga (1995), 70% of 46 road hauliers indicated a willingness to explore piggybacking further with the railway.The road hauliers believed that the service would benefit insurance companies due to lower claims, reduce road maintenance expenditure and save transport costs; and that a joint venture with the railway could benefit all parties.The respondents were asked to provide numbers of specific trailers that they were willing to ring-fence and commit to a service, and 1 298 trailers were identified.However, four years later, Jorgensen (1999) reports that 'this potential, already successfully implemented in the Americas, Europe and Australasia, has unfortunately, not been realised in Southern Africa'.A contributing factor is that, in order to enable this shift, freight flows must be identified that exploit the core strengths of both rail and road, yet the case for domestic intermodal solutions has never been clearly and unequivocally made.The research presented in this article aims to rectify this situation.
In the next section, intermodal freight transport is defined.This is followed by observations from the intermodal freight transport experience in the USA and Europe to distil key guiding principles for South Africa.A description of the research strategy to build the case for domestic intermodal freight solutions is provided, followed by a discussion of the research results.The last section concludes and recommends a way forward.

DEFINITION OF INTERMODAL FREIGHT TRANSPORT
To ensure a common understanding of the concept of intermodal freight transport the following definition for intermodal is accepted: 'Intermodality is a characteristic of a transport system that allows at least two different modes to be used in an integrated manner in a 'door-to-door' transport chain.In addition, it is a quality indicator of the level of integration between different transport modes' (European Commission, 1997).
Integration requires the promotion of efficient interconnection and interoperability between networks and modes.'Interoperability' refers to harmonised interfaces between transport systems and a high level of service.'Interconnection' refers to the physical interconnection of networks through new design and expansion of the transport infrastructure and interchanges (Keller, Tsavachidis & Hecht, 2000).
In the case of domestic intermodal freight transport in South Africa, the transport modes specifically refer to road and rail freight transport.The process of intermodal transport consists of short-distance road feeder services to an intermodal terminal in a logistics hub where freight is consolidated into main-line block trains running the length of the corridor to a destination terminal.From the destination terminal, it is transported to distribution centres or end destinations via road transport.

OBSERVATIONS FROM THE INTERMODAL FREIGHT TRANSPORT EXPERIENCE IN THE USA AND EUROPE
Intermodal rail transport is the fastest-growing rail traffic segment in the USA, increasing from 3 million trailers and containers in 1980 to 11 million in 2010 (a compound annual growth rate of 5%).Most intermodal rail traffic consists of consumer goods, and approximately 40% is contributed by domestic intermodal rail traffic, reflecting the vital role railroads play in the domestic economy (Association of American Railroads, 2011).Case (2009) states that intermodal traffic is 'holding its position (in the USA) in a falling transportation market' and that this is proof that 'such services are now structurally integrated in the transportation market'.In Europe (OECD member countries), the number of TEUs on rail tripled between 1980 and 2008 to 16 million TEUs (a compound annual growth rate of 4%) (OECD, 2010).
Intermodal traffic in the USA surpassed coal traffic as far back as 2003, and is projected to be rail's best hope of recovery after the recession (Kolstad, 2009).Woodburn (2008:442) also cites the decline in 'traditional' rail transport industries and the concurrent development of consumer goods markets as a driver of the substantial intermodal freight growth in the United Kingdom.Yevdokimov (2000:6-9) provides an extensive analysis of intermodal's advantages and echoes Berwick's (2001) regional findings (mentioned in the introduction) on a macroeconomic scale, i.e. increased volumes on the transport network resulting in economies of density, expansion of the network and resultant economies of size, reduction in logistics, and increased access to input and output markets.The European Intermodal Research Advisory Council (EIRAC, 2005) confirms that intermodal transport will enable Europe to cope with increasing transport demands, improve the environmental impact of transport, and enhance its competitiveness.
These advantages are better understood if intermodal transport is regarded as a general purpose technology (GPT) which is typically characterised by statistically significant spill-over effects to other areas of the economy (Laaksonen, 1999;Yevdokimov, 2000).Yevdokimov (2000) states that, in this view, intermodal transport is viewed as a two-way improvement of economic productivity, by improving both the current operational functions of the system and expanding these functions.He demonstrates that a once-off 10% increase both in the frequency of transport and transport network expansion due to intermodal transport resulted in a permanent increase in annual economic growth, gradually increasing and reaching a peak of 3% per annum after 15 years and settling over the long term at a 0.4% increase in economic growth per annum.The GPT nature of intermodal is also alluded to by Brown and Hatch (2002), who suggest that intermodal 'may be the most efficient and socially beneficial means of providing freight capacity'.
Positive environmental effects were not included in Yevdokimov's analysis (2000:18).The addition of this factor to future models should further illustrate the positive spin-offs from intermodal transport.In Europe, for example, rail-road intermodal solutions reduce carbon dioxide emissions by 55% (or by 1.8 million tons per year) and save 29% of energy usage compared to a road-only solution, with estimated annual environmental savings of about €180 million (European Commission, 2008).According to the Association of American Railroads (2011), railroads are on average four times more fuel efficient than trucks, with 7 gallons of fuel required to haul one ton of freight coast-to-coast in America via rail, compared with 28 gallons via road.This resulted in railroads, while almost doubling their freight volume between 1980 and 2010, using virtually the same amount of fuel as in 1980.
The Association states that shifting 10% of long-distance road freight to rail would save more than a billion gallons of fuel per year, and annual greenhouse gas emissions would fall by more than 12 million tons.
Despite the benefits already being experienced, a number of areas that require attention in order to expand the use of intermodal transport have been identified internationally.
The continuous development of optimal transport connections and interoperability standards among and between modes and logistics hubs are vital for the development of the logistics network of a region.Policymakers should also align regulations and infrastructure plans at all levels in order to enable efficient transport connections and stimulate integrated logistics strategies.Effective IT platforms and improved education and training on intermodality are critical, as is the requirement for all modes to operate at agreed levels of efficiency with concomitant shared liabilities (EIRAC, 2005;Site & Salucci, 2005;Vasiliauskas & Bazaras, 2006).
The fact that South Africa's national railroad has a historical infrastructure investment backlog and that the above-mentioned factors all require attention in South Africa, point to significant challenges locally.However, there is also the opportunity to learn from international experience and, through collaborative planning and development, to bypass many of the challenges caused by disjointed developments.The existence of long, dense, standardised flows between a few, highly concentrated origin and destination pairs also significantly enhances the prospects of success, as is illustrated in the research results.

RESEARCH METHODOLOGY
An initial approach to determining the target market for domestic intermodal solutions in South Africa was proposed by Van Eeden and Havenga (2010), identifying intermodalfriendly flows based on uniform, dense freight flows on long-distance corridors, derived from South Africa's national freight flow model (Havenga, 2007).
Subsequent to this analysis, more in-depth research and stakeholder consultations were conducted to develop a high-level business case for intermodal-friendly traffic.The research was also expanded to quantify the potential cost savings of an intermodal solution versus a road-only solution for the identified traffic.In addition, the potential impact of a domestic intermodal solution on externality costs and carbon dioxide emissions was also analysed.This subsequent research methodology is much more detailed and robust, reflecting the deeper analysis and segmentation principles required to initiate the process of building a business case for domestic intermodal, building on the strategic position put forward by Van Eeden and Havenga (2010: 263).The research methodology is depicted in Figure 1.The approach to the national freight flow model and corridor freight flows has been welldocumented and, as it is not the focus of this article, the reader is referred to Havenga (2007 and2010) for a detailed description.The next step is to identify rail-friendly freight, followed by extracting intermodal-friendly freight as a subset thereof.The potential cost savings of shifting freight to intermodal solutions are subsequently quantified.Next, the research approaches for each of these steps are discussed in turn.

Identification of rail-friendly freight
In South Africa, rail's core strength has traditionally been the transportation of bulk commodities, because this market segment enables the optimal utilisation of rail's genetic technologies and economic fundamentals.
Railways create and develop market opportunities by taking advantage of their genetic technologies, i.e. bearing, guiding and coupling.Bearing supports axle-loads of up to 40 tons (and therefore density) but at relatively low speed, while guiding refers to the wheelon-track differentials (and therefore speed of movement) but with relatively low axle-load.When bearing and guiding are added to coupling, this translates into long trains with high volumes (Van der Meulen, 2007).Bearing, guiding and coupling in optimal combination therefore support heavy intermodal or double-stack container trains, attaining axle loads of up to 32 tons at 120 km/h.This provides volume throughput opportunities, but also slightly less price-sensitive markets.
The key economic fundamentals that apply to rail are freight uniformity and long-distance line density.Freight uniformity refers to the standardisation of freight to facilitate handling and transport, such as bulk and palletised commodities (hereafter referred to as unitisable commodities).
Line density refers to the volume of traffic per kilometre of railroad, expressed as 'tonkm per route kilometre'.'Tonkm' is the standard unit for measuring freight transport, as it takes into account both tons and distance travelled (Chasomeris, 2003:133).Given rail's high fixed cost, higher density means that the cent per tonkm cost of a railroad will decrease with each additional tonkm of activity over the same track length.Dense corridors are ideal for rail or intermodal transport, as the density creates economies of scale because of the large volumes of tonkm generated (De Bod & Havenga, 2010:98;Havenga, 2010).For the purposes of this article, it is assumed that the core corridor network of the current rail system is in place and that an economic business case exists to keep it in place.(The current debate in South Africa is not about the corridor network, but revolves around the branch line network, and it has been shown that even in some of these cases, the lines could be profitable [Simpson & Havenga, 2010:252]).Flowing from this assumption, any additional freight will cause an improvement in density (i.e.tonkm per route kilometre, as discussed above), thereby reducing costs.Even so, for commercial purposes, a cut-off density of 100 000 tons per year (a train per week) is used to identify rail-friendly flows.
Distance is an important consideration in identifying intermodal freight ('d' in Figure 2) (Button 2010: 339), because of the low distance sensitivity of rail transport versus road transport.The DfT's case study confirms cost savings and environmental benefits for both routes, even over these short distances.
The results from South Africa's national freight flow model indicate that most long-distance transport of unitisable commodities in South Africa occurs over distances longer than 500 km (Figure 3).Therefore, an assumption of distances further than 500 km for rail-friendly unitisable commodities was deemed to be conservative and was adopted.Rail-friendly freight is, therefore, defined as freight flows between dense origin-destination pairs of 100 000 tons per annum (a minimum of a train per week) over distances greater than 500 km.
The portion of this freight that is suitable for intermodal transport is subsequently identified.Brown and Hatch (2002) assert that 'rail intermodal's economic value and contribution to the economy resides primarily in long-haul corridors' and highlight typical freight, mostly fastmoving consumer goods (FMCG).These are products that are sold quickly and generally consumed on a regular basis, as opposed to durable goods such as kitchen appliances, which are replaced over a period of years.FMCG product categories comprise food and dairy products, pharmaceuticals, consumer electronics, packaged food products, household products, beverages, and the like.

Identification of intermodal-friendly freight
The key driver of density (one of rail's economic fundamentals described previously) is the unitisation of cargo.This requires a large storage footprint (e.g.iron ore is stockpiled, and containers are stacked while goods within containers are palletised).Distribution centre to distribution centre (DC-to-DC) traffic for redistribution is naturally densified around a few corridors, but has no intermodal potential if it cannot be unitised (through palletisation) and 'connected' from shelf to shelf between these DCs.For the purposes of isolating intermodal-friendly transport from rail-friendly freight flows, the concept of 'unitisable' was therefore narrowed to 'palletisable' in order to ensure that only freight that can be easily packed on pallets and stacked in containers was identified.In order to identify freight that could be described as 'palletisable', three workshops were conducted with industry experts and the commodities from the freight flow model (Havenga, 2007) were classified into two groups, i.e. 'palletisable' and 'non-palletisable'.This classification is reflected in Table 1.Finally, the concept of terminal density was added as a special consideration for intermodalfriendly transport.Dense terminals lead to the standardisation of equipment and processes, which drives down costs (Kreutzberger 2008:153), and also moves 'd' (refer to Figure 2) to the left, i.e. reducing the distance at which intermodal transport becomes more costeffective than road-only transport.In this case, two types of flows were considered: longdistance flows between metropolitan areas (as 'low hanging fruit', or traffic that can switch easily) and flows from large manufacturing installations in rural areas to metropolitan areas.

Cost analysis
The cost of road freight transport is obtained from South Africa's national logistics cost model.The logistics cost model research approach and 2008 results are detailed in Havenga (2010), and the 2009 results published in Havenga, Pienaar and Simpson (2011).
The calculation of road transport costs is driven by weight in tons and distance travelled, and involves the summation of all the different cost elements of road transport on a specific route (including items such as licence fees and toll fees).The different cost elements of road transport are determined by the vehicle type.The vehicle type, in turn, is determined by the commodity type, network typology (i.e.whether it travels on corridors, in rural areas, or in metropoles) and route of travel.
To determine the current transport cost of palletisable freight, the total weighted cost of these flows between distribution centres was calculated.This was translated into cent per tonkm data and compared to the current actual cent per tonkm data for similar freight received from the national railroad.The potential cost savings experienced on rail due to the introduction of an intermodal solution can be attributed to an increase in density (i.e. a more favourable relationship between fixed and variable costs, where more freight is available to absorb fixed costs), i.e. a shift down the Harris (1977) curve, depicted in Figure 4. (System density savings represented by logistics service providers taking over some marketing, customer service and terminal handling costs have not yet been established).

Externality costs
Externality costs include all non-charged costs and encompass emissions, accidents, congestion, policing and noise pollution.Total road and rail freight externality costs were calculated for the first time in South Africa's sixth annual state of logistics survey (Havenga, Van Eeden & Simpson, 2009), and refined in the seventh survey (Havenga, Simpson, Van Eeden, Fourie, Hobbs & Braun, 2010).
Externality costs are calculated to depict the actual cost to the environment in monetary terms.Externality costs, with the exception of emissions, are based on a study done by Jorgensen (2009).Jorgensen (2009:6-12)  Emission externality costs are determined by calculating emission generation per tonkm per mode.For road, fuel consumption is the critical variable when calculating the amount of carbon dioxide emissions per tonkm.For rail, a weighted combination of diesel consumption and electricity consumption is used to determine carbon dioxide emissions per tonkm.
The externality cost of the emissions is based on a flat externality cost of $30 per ton of carbon dioxide emissions generated, which is based on the average cost of offsetting carbon dioxide emissions.(Calculating the average carbon dioxide emission offset cost is a topic of research in itself and has generated many debates in this and other industries. Values as low as $10 and as high as $90 have been observed.The fact that exact levels are unknown at this stage, and that tax levels are under debate, makes calculating carbon dioxide emissions a precarious subject.McCarl and Sands [2007] use three scenarios, with $30 as the average.) With regard to the intermodal solution's emissions, the same approach is used as previously stated but only for intermodal-friendly freight.This is possible as the emissions calculated are mode specific and are calculated at a tonkm level.Savings are calculated as the amount of emissions not being produced by road freight due to the switching of all the potential intermodal freight to rail.South Africa's National Treasury is seeking to implement a carbon tax by 2012 to reduce greenhouse gas emissions (National Treasury, 2010).A tax of R75 per ton (in 2005 terms) of carbon dioxide is proposed, increasing to around R200 per ton (in 2005 terms) of carbon dioxide (the timeframe is not clear from the National Treasury discussion document).In this article, the potential reduction in tonnage of carbon dioxide emitted is demonstrated through a R165 per ton of carbon dioxide tax (this was the value used by the National Treasury in their feasibility studies) (National Treasury, 2010).Calculating the amount of carbon dioxide tax savings that could be achieved by using an intermodal solution to transport freight was determined by using the emission savings calculated and multiplying this with the tax rate of R165 per ton of carbon dioxide.The challenges in developing and implementing successful domestic intermodal freight solutions are manifold, as highlighted by international experiences, and should be approached realistically.However, as illustrated, the potential benefits of domestic intermodal solutions to the South African economy are clear -significant cost and emissions reductions in an uncertain energy and carbon offset world.The economic realities of the future and increased demand for freight transport also leave players little choice but to galvanise solutions in this area.Stone's (2008:246) view that domestic intermodal 'will be a significant part of the expectations of a new invigorated rail freight activity' is therefore supported by the findings presented in this article.

Total
The stakeholders in the potential domestic intermodal freight market in South Africa are highly concentrated and consist of three core groups, namely, Transnet, logistics service providers and shippers or freight owners, with the market concentration spread over a limited number of large freight owners.

Figure 3 :
Figure 3: Transport distances of unitisable commodities >150 km (based on results from South Africa's national freight flow model as described inHavenga, 2007 and 2010)

Figure 4 :
Figure 4: The economics of rail density (adapted from Harris, 1977) calculates accident costs based on a 2006 survey on the N3 highway between Johannesburg and Durban, congestion costs based on studies in Australia and Europe, and noise and policing costs based on Australian research.
corridor freight in South Africa amounted to 192 million tons (or 100 billion tonkm) in 2009.Twenty-seven per cent of freight shipped on these corridors (51 million tons) and 28% of tonkm (28 billion tonkm) represented intermodal-friendly freight.(The discrepancy between the ton and tonkm contributions indicates that this freight tends to move over longer average transport distances than other freight.)Themajority of this intermodal-friendly freight moves on the Gauteng-Cape Town corridor (Cape corridor) and the Gauteng-KwaZulu-Natal corridor (Natal corridor) (47% of tons, 59% of tonkm), resulting in a narrow description of a highly densified, easily targetable market segment (Figure5).

Figure 5 :
Figure 5: Relative size of the intermodal market for the Cape corridor and Natal corridor

Figure 6 :
Figure 6: Intermodal density change on the Harris curve

Figure 7 :
Figure 7: Current road cost versus potential intermodal cost for Cape and Natal corridor combined (per tonkm)

Figure 8 :
Figure 8: Current total road cost versus potential intermodal cost for Cape and Natal corridor combined

Figure 9 :
Figure 9: Reduction in tonnage carbon dioxide emissions due to an intermodal solution on the Cape and Natal corridors

Table 1 :
Classification of palletisable versus non-palletisable final consumption products