Provision of Travelway Space for Urban Public Transport in Developing Countries

Transcription

1 Provision of Travelway Space for Urban Public Transport in Developing Countries

2

3 Table of Contents Provision of Travelway Space for Urban Public Transport in Developing Countries...1 FOREWORD...1 I. TRAVELWAYS FOR PUBLIC TRANSPORT IN DEVELOPING COUNTRIES...2 A. Introduction...2 B. Characteristics of travel demand...3 C. Low cost mass transit options...8 D. The impact of public transport segregation...22 E. Planning considerations...25 F. Strategy for development...34 BIBLIOGRAPHY...36 ENDNOTES...39 II. CASE STUDY: BUSWAY IN ANKARA AN INTERMEDIATE, LOW COST ACTION TO IMPROVE PUBLIC TRANSPORT A. Urban transport in Ankara...39 B. The development and performance of a busway in Ankara...44 C. Lessons to be learned...53 D. New developments in Ankara mass transport...55 REFERENCES...55 ENDNOTES...55 III. CASE STUDY: THE BRAZILIAN EXPERIENCE IN PLANNING, IMPLEMENTING AND OPERATING PUBLIC TRANSPORT ON SEPARATED ROAD AND LIGHT RAIL TRAVELWAYS A. Introduction...56 B. Medium capacity public transport modes...59 C. Lessons from the Brazilian experience...69 REFERENCES...75 ENDNOTES...78 IV. CASE STUDY: PROVISION OF SEPARATED TRAVELWAYS FOR PUBLIC TRANSPORT IN METRO MANILA, THE PHILIPPINES A. Urban development and transport...78 B. Development of LRT system...87 C. EDSA bus lanes...94 D. Lessons from the case studies ENDNOTES i

4 ii

5 Provision of Travelway Space for Urban Public Transport in Developing Countries United Nations Centre for Human Settlements (Habitat) Nairobi, 1993 The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the secretariat of the United Nations concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Mention of firm names and commercial products does not imply the endorsement of the United Nations. FOREWORD The demand for public transport services is growing steadily in the large cities of developing countries. This is most evident along the transport corridors linking ever expanding suburban areas to main concentrations of employment and services, in particular, to city centres. Sustainable urban transport development depends to a great extent on the capacity and quality of services along these corridors, not only for obvious economic and environmental considerations, but also with regard to social equity, as these corridors largely determine the mobility of the urban poor who are pushed by land forces to urban fringe areas and suburban localities. Most developing country cities with a population exceeding 2 to 3 million already have at least one heavily loaded corridor on which public transport should offer a capacity and speed of travel better than what buses operating in mixed traffic conditions could provide. Public authorities when seeking such alternatives usually focus on heavy rail transport systems such as metro. Metro systems are, perhaps, indispensable to support the development of very large cities but these are neither affordable nor a viable option in most intermediate cities. These cities will have to look for less expensive solutions using, as priority, already available public transport infrastructure. Providing public transport with partly segregated and exclusive travelway space increases both capacity and operating efficiency of any public transport system. Separated travelways not only ease or alleviate the interference from other traffic and thus the impact of transport congestion on public transport, but also allow the application of public transport operating regimes which further add to the capacity and quality of services. The objective of this publication is to promote public transport systems which will be significantly less expensive than metro and yet provide relatively high capacity and speed of travel by making use of segregated travelways. It highlights low cost mass transit options bus lanes, busway transit and light rail transit and attempts to assess their impact on transport users, transport operators and other beneficiaries in the light of the experience of a few selected developing country cities. The publication then outlines planning considerations that must precede investment decisions for the successful implementation of public transport segregation schemes. Finally, a strategy for mass transit development is outlined for the benefit of urban transport planners and decision makers. The publication has been prepared as a part of the ongoing effort of UNCHS (Habitat) to promote the development of public transport in developing countries. It responds to the concerns expressed in Agenda 21, adopted by the United Nations Conference on Environment and Development, on the need for sustainable transport strategies and actions as an integral component of the sustainable development of human settlements. We gratefully acknowledge the contribution of Mr. P.R. Fouracre of the Overseas Centre, Transport Research Laboratory (TRL), United Kingdom, in the preparation of this publication. The publication, without doubt, has benefitted from the findings of the research programme of TRL, funded by the Overseas Development Administration. The case studies have been contributed by Messrs T. Birgonul and H. Bayirtepe, both from the Middle East Technical University, Ankara, Turkey; by Mr. G.D. Esguerra, Director, Transportation Planning Service, Department of Transportation and Communications, the Philippines, and by Messrs L.A. Lindau and L.A. Dos Santos Senna, both from the Federal University of Rio Grande do Sul, Porto Alegre, Brazil. Their contributions are also gratefully acknowledged. 1

6 Elizabeth Dowdeswell Under Secretary General United Nations Centre for Human Settlements (Habitat) I. TRAVELWAYS FOR PUBLIC TRANSPORT IN DEVELOPING COUNTRIES A. Introduction Public transport is a growth sector in most developing country cities. A World Bank estimate puts the number of daily bus trips in 1980 at 600 million; this number is expected to double by the year Recent evidence from a limited number of cities suggests no change in this trend; bus passenger trips have been increasing at average rates of between 6 and 10 per cent per annum during the last decade. The increasing demand for public transport is being driven largely by population growth. This growth not only generates more trips, but with increasing city area, and hence longer trips, people become more dependent on public transport. Personal motorized transport is still beyond the reach of the majority of the people, and increases in real incomes are still likely to encourage greater use of public transport as well as higher car ownership levels. Indeed, in many African countries the economic down turn has been reflected in static, or even declining vehicle ownership levels, which is likely to increase further the importance of public transport in larger cities. Also contributing to the seemingly inexorable growth in demand for public transport are the locational patterns of the urban poor, resulting in a need to cater for long distance commuting at low cost. Changing life styles, more women in the labour force, and the youthful age structure of the society all contribute to the increasing demand for travel. To avoid public discontent, many countries maintain policies which actively encourage the use of public transport through, for example, subsidized fare levels and significant fare discounts for students and other privileged groups. The growing awareness of the impact of transport on the environment and on the consumption of energy is likely to enhance further the policies oriented towards the promotion of public transport which allow the movement of people in a much more energy efficient and less polluting manner than private motor vehicles. As cities grow in size, increasing attention is required so attention focuses on the mass movement of travellers along major transport corridors. Developing country cities today have populations in excess of 15 million, and these cities have to provide for million public transport trips per day; corridor flows can be in excess of 1 million passengers per day. By the year 2000, there will be several more of these mega cities, and many more "lesser" cities with populations greater than 5 million over 40 according to one estimate. Coping with this scale of demand requires a high capacity transit system. The recognized high capacity systems are suburban rail and metro. Suburban rail systems are important in only a handful of cities (for example, Bombay and Calcutta), and can not be easily integrated into the existing urban physical development structures. Metro systems can be sensibly justified, in economic terms, only in very special conditions. What is often overlooked in the development of mass transit capability, however, is the potential of existing road based public transport and trams. In developing country cities, both are associated with poor productivity, unreliability, long journey times and excessive over loading. Seldom are they considered as "mass transit" systems. Nevertheless, their performance can be substantially enhanced through the provision of reserved travel space, with exclusive right of way. This provides two main benefits: it removes the vehicle from the interference of general traffic congestion, and it provides the basis for introducing additional operational measures which can further improve performance. The track also lends a certain degree of image and permanence, which may be important in establishing the scheme with both users and land developers. 2

7 Modern busway transit systems and light rapid transit (LRT an advanced form of trams) operating on their own track, can offer relatively high capacity transit at moderate cost. These systems are likely to be able to cope with demands on many city corridors, and in the case of busways, have the added advantage that they capitalize on the existing work horse of urban transport. Though busway transit has been successfully deployed in a number of developing country cities, and indeed was largely developed in the major Brazilian cities, it has not attracted the level of interest which its undoubted advantages merit. On the other hand, while there are few tram systems in use in the developing world (and indeed many were discontinued in recent times), LRT has attracted a good deal of interest. This may be, perhaps, because of its "high technology" appeal or because it can be viewed as a worthy "second best" to a more expensive metro. Environmental concern is another factor in raising that interest. Clearly, buses and trams could contribute more to a city's mass transit needs through the allocation of reserved travel space for their operations. The purpose of this paper is to demonstrate this case fully and to give guidance on the conditions which would justify the dedication of road space for exclusive use of public transport. It is not the intention to provide a detailed technical account and guidelines for implementation; these can be found elsewhere. This paper is targeted at decision makers, agencies and planners who have the role of developing and implementing city transport policy. In targeting this group, the aim is to sensitize them fully to the concepts, opportunities and realities of these schemes. To some of this group, the idea that bus and light rail are viable and attractive alternatives to heavy metro schemes, in many situations, may seem fanciful; but in an increasingly difficult operating environment, and with limited financial resources available, these projects may well offer the best direction for the development of mass transit. The first section sets out the characteristics of city development which have an impact on transport and the way in which public transport has been made responsive to the growing demands placed on it. Priority measures are then discussed with reference to specific case study material; this draws attention to the types of priority measure, the resources needed, the likely impacts, and the factors which are needed for success. This material is used to develop the general guiding principles for adopting priority measures, and a strategy for promoting this policy within the structure of a city's development programme. B. Characteristics of travel demand 1. Outline To put the development of mass transit in some context, this section outlines the typical characteristics of travel demand in developing country cities. Daily trip rates per capita typically lie in the range Because incomes are low, the majority of trips are undertaken on either some form of non motorized transport (principally walking or by bicycle) or on one of the many forms of public transport. The latter is particularly important in providing for longer trip lengths. Trips to or from work and education are likely to be the majority of these non walk trips. Work trip movement is largely radial, focusing on the central area of a mono centric city. Third world cities generally exhibit less distinct travel "peaks", possibly because of the inadequacy of transport facilities to handle demand, but also because mid day commuting is common in some societies. The main factors influencing travel demand include: city structure, socio economic characteristics of the community and transport facilities available to the traveller. (a) Population 2. City structure At the aggregate level, changes in population have three important effects on travel. First, more people generate more trips. This fundamental effect of population growth is important. It is estimated that each additional thousand people in a developing country city is associated with an additional public transport trips per day. The second effect of population change concerns the development pattern of an urban area and the way in which the physical area of the city changes. Assuming there is no radical change in the city structure or density, more population implies expanding the city area and the likelihood of longer trip lengths. This has important implications for modal choice. Increasing city size puts more pressure on public transport because 3

8 travellers are unable to walk or cycle the longer distances involved. The final major impact of population change concerns the way in which the transport system adapts to urban growth and affects travel patterns. As a city becomes larger, transport operators tend to concentrate their resources on servicing general patterns of movement along major corridors, rather than trying to cater for all possible origin destination choices. These changes might" also be accompanied by a move away from flexible public transport services (like shared taxis) towards fixed route bus services, often employing large size vehicles. The result is more complex journeys, possibly involving interchange and lengthy walking and waiting times. Population growth appears to be slowing in the mega cities of the developing world, though rates of 4 per cent per annum are still common. The trend is still towards urbanization but, as in India, the growth seems to be most pronounced in smaller cities, while larger cities are dispersing over a wider metropolitan area and beyond. (b) Urban form A characteristic of many developing country cities, even some of the very large, is their mono centric spatial form. They may have retained this structure over time (and for much longer than their developed country counterparts) partly because transport and other infrastructure is spatially concentrated and partly because high income residents, who might be expected to lead any move towards suburbanization, live close to the city centre and have neither the incentive nor the opportunity to decentralize. The strong spatial concentration of transport comes about because of limited resources; developing country cities have a strong incentive to retain a compact, radial structure, which gives strong emphasis to low cost public transport corridors. Furthermore, a sizeable proportion of work trips in most developing cities are focused on the city centre with its commercial, service, government and retailing activities. Some light industry may also be centrally located, though main industrial sites are likely to be in the suburbs. The resulting impact on travel characteristics, particularly in medium to larger cities, has been the strong emphasis on radial movements towards the city centre. As an indication of the impact of city form on travel demand, table 1 presents estimates of total corridor flows for different city size and structure. In this analysis, cities are categorized as having one of three basic forms: circular, semi circular (where, for example, the city abuts on to the sea), and linear. They are also sub divided by spatial structure: mono centric (single, dominant central focus), uniform poly centric (with equal employment opportunities at both the centre and in a number of surrounding sub centres), and non uniform poly centric (where sub centres exist, but the centre dominates). For any given city size, corridor flows are highest in mono centric cities. Population (millions) Circular cities Table 1. Estimated radial corridor loadings for different city forms Corridor length (km) Non uniform, poly centric: trips per day (millions) Uniform, poly centric: trips per day (millions) > Mono centric: trips per day (millions) < <0.1 <0.1 <0.1 Semi circular cities > < <0.1 <0.1 <0.1 Linear cities >

9 < <0.1 <0.1 <0.1 Source: GOI, 1987 Table 1 can be used to make an approximate estimate of likely peak hour, directional flows. For example, a circular city will typically have 6 8 radial corridors and peak hour directional flows which are 6 per cent of total daily flow. Such a city with a population of 3 million would experience individual average peak hour directional flows of between 10,000 30,000 passengers (six corridors), or 8,000 20,000 passengers (eight corridors). It is difficult to be precise, but the likelihood is that once a city reaches a population of 2 3 million, it will have at least one corridor which will require mass transit facilities. This notion is supported by a conceptual approach to the estimation of how many central area jobs can be sustained by a road based transport system with no priority track. It has been estimated that an ordinary bus service, supported by some private traffic, could feed 250,000 travellers into a city centre during a 2.5 hour peak period; these, together with those workers resident in the city centre, would probably be generated by a city of about 2 million population. Larger cities would require a higher capacity transport system to support growth in employment at the city centre. Locational aspects are very important in the choice of transport mode. Commuters have to rely on public transport for two way journeys in excess of about 10 km. This characteristic is particularly marked in low income groups who have practically no other travel option. It is one of the ironies of development in third world cities that, as already noted, it tends to be the low income groups who live furthest from the city centre. Where high income commuters do travel long distances, they usually have access to modes faster than public transport and will accomplish their journeys speedily. 3. Socio economic characteristics The extended family system and high birth rates in developing country cities lead to an average size of household which is generally higher than in the developed world. Typical average household size is around 5 6 persons. Household size affects per capita trip rates: typically a 10 per cent increase in household size can be associated with only a 6 per cent increase in household trip making. The sex, age and role structure of household members are also likely to have an important influence on travel characteristics; activity analysis suggest that households have very different travel patterns at different life cycle stages. For example, an increase in the proportion of students and/or workers in the household will have a positive effect on both household and per capita trip rates. Women have a lower participation rate than men in both work and education; they are less likely than their western counterparts (or men in general) to travel long distances for employment. Where they do travel longer distances they are unlikely to use a bicycle or any motorized personal vehicle, except as a passenger. In some cities they travel in reserved compartments on public transport. Many women will not travel unaccompanied on any kind of business, apart from local shopping and taking young children to school. Age structure is important largely in respect of children and the retired. Pre school children are unlikely to make any significant trips, except in the company of elders. While all school children make school trips, their mode of travel may well be influenced by their age; young children will have only a short trip to a local school which can be accomplished on foot, while older children attending secondary school and colleges will inevitably travel further, possibly using some mechanized mode. There is likely to be a large increase in trip making as students progress from primary to secondary education age, and a rapid drop in trip making at the onset of old age. From the limited evidence available, household income does not appear to play a significant part in determining household trip rates in developing countries, though it is clearly a major determinant of modal choice. Much trip making must be a necessary part of life irrespective of income level; only households made up solely of the very poor, the unemployed or retired will not participate in these committed trips. Income is more likely to have an effect on trips associated with more leisurely pursuits, though these usually account for only 20 per cent of total trip making. 5

10 Income clearly affects the way in which people choose to travel. It sets the limit on their capacity to acquire a personal vehicle and also, given that trip making is relatively inelastic to income, it sets the limit on how much of a particular mode they can "consume" in order to achieve their desired level of travel. For example, it is quite common for low income commuters to switch their normal mode of travel from bus to walking towards the end of their pay period as money runs out. It is not unusual to find that 10 per cent of household income is spent on transport; sometimes the figure is as high as 15 per cent. At this level of expenditure, per capita incomes of between $US600 and $US1500 can just support fare levels of cents for a typical round trip of 10 km. (a) Infrastructure 4. Transport characteristics There is little understanding of the impact of infrastructure provision on travel demand in developing country cities. There can be little doubt that new infrastructure will generate new demand, but the scale of the impact is unknown. New infrastructure and equipment may also enable concentration of peak travel, a possibly unwelcome outcome! One likely major impact of a large scale investment in mass transit is that it will often emphasise and encourage existing travel patterns. This is because the investment is put in place to meet an existing high demand which is expected to grow. Mass transit lines constructed along main radial corridors will have the potential to feed many more commuters into a city centre than an existing road based system (i.e., with no reserved right of way). This has important implications for the development of the city structure; a radial mass transit scheme may well petrify a mono centric city development pattern. Some cities have advocated such schemes in a planned attempt to re structure the city, but there are few examples of the successful exploitation of transport in this way. It is generally believed that the level of infrastructure provision in cities of developing countries' is lower than in a counterpart industrialized city. This contention would help explain the paradox that despite low vehicle ownership in developing cities, congestion is as bad, if not worse, than that in developed cities which have far higher vehicle ownership levels. (b) Public transport Public transport is characterised in developing country cities by the wide range of vehicle types in use and services on offer. Despite the variety of vehicle types it is clear that most are employed in one of two main ways: either providing a bus like service with fixed routes and fares (for given trips) or a taxi like service where the route is determined by the hirer of the vehicle and the charge for the hire is metered or bargained. One important variation common in many cities is the shared taxi in which the first occupant determines the destination and other passengers, heading in the same general direction, are picked up en route each passenger paying a fixed fare. Most public transport is road based and is likely to remain so in the future. It has been noted earlier why public transport is likely to be important in any developing city and why its importance is likely to increase with city size. It has also been noted how the development of public transport may influence city development; in summary a radial mass transit system permits the city centre to develop. Continued dependence on more traditional public transport like cycle rickshaws may also have some impact on city development, just as will the provision of cheap subsidized buses to low income settlements at the city edge. Because of its importance in everyday life, public transport receives a good deal of attention from local government. This usually takes the form of strong regulatory controls, particularly in respect of fare levels and subsidies to public sector operators. While some authorities have also invested vast sums on metro systems, few have introduced priority traffic measures for the main carrier: road based public transport. Except on newer metro systems, service quality on all forms of public transport is generally poor, partly because fare levels are insufficient to meet investment needs. (c) Other modes Walking, as a mode of transport, is limited in range by both its speed and energy requirement. Few trips of more than 5 km are made regularly. Even so, walking is a major mode of travel and can account for between 20 and 40 per cent of trips, and even more if very short trips are included (travel surveys often exclude very short trips). Neither do these figures take account of walking associated with the use of other modes (i.e., 6

11 access to and from public transport etc.). The impact of walking on travel demand derives from the numbers of the population who are dependent on it; this group not only includes most housewives and children, but also anyone who has to walk to access another mode. The limitations of walking are therefore very powerful in the planning of the spatial location of local amenities, as well as the transport network. In cities where bicycles are widely owned, their use is impressive; for example, in medium sized Indian cities, cycles typically account for between 35 and 50 per cent of traffic on major corridors and 10 to 30 per cent of all trips. However, the bicycle is nut in universal use in cities of developing countries and even where it is used it may be barred, through social norm or male priority, to women riders (though not to women passengers). The bicycle extends the possible range of travel beyond the limits of walking to typically 10 km. Range is again constrained by speed and energy requirement. Furthermore, as with walking, few cities actively encourage cycling and its safety record is poor. Where the bicycle is common, it clearly has an impact on travel patterns; principally it allows low income male workers to commute from longer distances at very low marginal cost. In doing so, it frees these commuters from the more rigid radial patterns of movement which are often imposed by public transport networks. While personal vehicle ownership is typically on levels several times lower than in the industrialized world, growth in the number of motor vehicle is high and congestion in large city centres is as bad as in any developed city. Cars are mainly owned and operated by higher income groups, though there is undoubtedly a substantial fleet of company and government vehicles in use. The burgeoning middle income groups are also acquiring cars or motor cycles in increasing numbers. The impact of the private vehicle on the transport system seems to be immense, despite the fact that its share in modal choice is yet small. Most major urban transport infrastructure projects are designed to ease the flow of road traffic, a large proportion of which is made up of personal motor vehicles. Access to a motor vehicle confers a high degree of flexibility and range in travel. In some cities, petrol prices are artificially low, encouraging excessive car use, while in others the high costs of operating a car may restrict its use to non regular trips. Generally, however, without controls in its use, the growth in vehicle ownership must have an increasingly important impact on city development, possibly encouraging trends towards decentralization. 5. Summary Cities in developing countries present a range of development characteristics, dynamic growth patterns, transport infrastructure and operations, and social customs which defy all but the broadest generalizations. Even so, it is important to try to understand the processes and interactions which drive transport demand if transport planners are to contribute positively to the general debate about urban development. As cities become larger travel demand grows at a disproportionately higher rate and there is a greater dependence on public transport for travel needs, particularly from the urban poor. It is also evident that trip movements become focused on corridor travel feeding into the city centre; once a city reaches a population of about 2 3 million, corridor flows will have reached around 20,000 passengers per hour per peak direction. Corridors and city centres which have to handle this level of demand are prone to endemic and prolonged traffic congestion, because of the inadequate capacity of the infrastructure to meet both private and public vehicular demands. Public transport, potentially the most efficient carrier and that which serves the majority of travellers, cannot deliver an effective service in these conditions; journey times and waiting times are long, irregular and unreliable. Moreover, because of the poor productivity of buses, together with a low revenue earning potential, the financial position of operators is often weak. In these circumstances, the prospect for improved public transport is grim; operators cannot afford new investment when they cannot even afford the depreciation on existing stock. From the traveller's viewpoint the main concerns are reasonable access to activities in reasonable time and comfort, and at affordable cost. Even in the short term, transport planners and operators are struggling to achieve some semblance of satisfying these needs; and that at mounting cost as access to central area and congestion problems worsen with increasing city size. This situation can only get worse as cities grow and options for further infrastructure development are limited by finance and environmental concerns. Ultimately, if the transport system cannot respond to these pressures, then other land use developments may take control, leading to unstructured and diffuse city growth, and even the atrophy of the city centre. In order to sustain the growth of the city centre, and accommodate the associated high volumes of corridor travel, consideration has to be given to the controlled use of the infrastructure as a means of protecting the operations of public transport and hence making best use of limited resources. 7

12 C. Low cost mass transit options 1. Perspective Transitways are a means of controlling the use of road space so that public transport vehicles are segregated from general congested traffic conditions; one or more lanes of the road are reserved for the exclusive use of public transport. These transitways can accommodate fixed rail or road based vehicles, having many similarities in design whichever vehicle type is used. The method of segregation can be either permanent involving physical barriers, or open relying on good driver behaviour and enforcement of segregation rules. The latter include traffic management measures to improve bus flow (e.g., bus lanes and bus gates); these are prone to poor driver discipline and weak enforcement, and hence their effectiveness is not reliable. Physical separation of the travelway, because of its more permanent nature, provides more reliable protection for public transport vehicles: it also provides opportunities to achieve a significant increase in the productivity of the public transport vehicles using the travelway and hence to develop a low cost, high capacity mass transit system. In this section, the various options for travelways are described, starting with basic priority measures and then focusing on the physically segregated schemes (busway transit and LRT). Busway transit is a particular development of bus priority measures, which seek to optimize the output of bus technology through both infrastructure investment and improved operational procedures; in many instances it will require no substantial investment in equipment, since buses are in wide current use. LRT and other fixed rail "light" systems are not common in the developing world, and any investment scheme is likely to require a complete package of infrastructure and equipment. Even where LRT systems are in use in developing countries, the equipment is largely antiquated and of little value to a modern system. Inevitably, investment in LRT will be at a higher level than in a busway transit system. Before describing the options in detail, their general specifications are briefly compared with that of a typical heavy rail mass transit system. Table 2 sets out some key points of comparison. Here, separate columns are given for trams and LRT. on the grounds that they can be distinguished by the amount of sharing of their right of way with other traffic and, hence, there can be differences in specification, operation and performance. However, there are no hard and fast rules to distinguish at what point a tram becomes an LRT, and for the most part this paper will treat them as similar types of transit system. Right of way Av speed (km/h) Signalling Table 2. Main characteristics of mass transit options Busway transit Trams LRT Metro Physically segregated along key stretches Mainly shared with road traffic Physically segregated along much of route; some grade separated junctions Fully grade separated Visual, with some signals Visual, with some signals Visual plus signals Automated Platforms Low level Low level Low and medium level High level Pass. Transfer (per hour) <5000 <5000 <5000 > 10,000 Station spacing (m) Pass. capacity (thousands per hour per direction) Min. Headway (secs) Pass. per car Cars per train 1 1/2 2/4 6/8 Train length (m) (two car) (three car) (six car) 8

13 What will be clear from table 2 is that with the exception of the impact on the environment, there is little to choose between LRT and busway transit; both have medium capacity and can operate at speeds in excess of 20 km/h. Some authorities argue that busway transit can only achieve these levels of output during short periods, i.e., they cannot perform as consistently highly as LRT. This cannot be fully substantiated as the case study material will reveal. The table also illustrates how performance is affected by the level of sharing rights of way with other traffic and the spacing between stops; quite clearly, full segregation and long spacing between stations, confers on the metro the advantages of high speed and use of high capacity cars in large train formations. There are a few examples of fully segregated LRT (notably in Manila), but no known examples of a fully segregated busway transit system. It can only be speculated that the capacity of a segregated busway system would be in excess of 30,000 passengers per hour per direction. 2. Bus lanes and other bus priority measures The main feature of bus priority schemes is the separation of buses from other traffic (either at selected locations, like bus stops, or along running sections) through the use of "paint and signs" which indicate the rules of segregation. Bus priority measures critically depend for their effectiveness on disciplined driver behaviour, backed up as necessary by strong police enforcement. Junction related delays can be dealt with by spot priorities, examples of which are turn ban exemptions and bus gates. Turn ban exemptions permit buses to turn out of a particular road, where this movement is banned to other traffic. Bus gates (see plate 1) permit buses to turn into a particular road, where this movement is banned to other traffic. Short bus lane sections at junction approaches can allow buses to "queue jump" and bus activated traffic signal pre emption can reduce delays. However, while spot priorities are a useful traffic management measure, they cannot by themselves improve bus performance over whole routes. Painted bus lanes can give buses priority over long sections, provided they are respected. There are two main types of bus lane: with flow (see plate 2) and contra flow (see plate 3). In environments where road user discipline is poor, and police enforcement is weak, with flow lanes tend to be violated and are relatively ineffective; the effects of a bus lane scheme implemented in 1989 in Manila fall short of expectations for these reasons. In contrast, contra flow bus lanes tend to be self enforcing, since buses travel in the opposite direction to other vehicles. However, there are some indications that pedestrian/bus accident rates may be higher along contra flow than along with flow bus lanes, because pedestrians are unaccustomed to looking "the wrong way". With flow bus lanes can be constantly operated or only in specified hours when the demand for public transport is high. While the latter way of operation allows better use of road space in the off peak hours, it may make the enforcement of road user discipline more difficult. Plate 1. A bus gate providing buses with unhindered access to a main road: Hong Kong A traffic scheme may include both with flow and contra flow lanes, as well as spot priorities. Although one lane is usually provided in each direction for buses, two lanes may be provided where bus volumes are high, at busy bus stops (to allow buses to overtake one another) or on long uphill sections. 9

14 3. Busway transit (a) Outline The traffic violations experienced by with flow bus lanes can be overcome by physically segregating buses from other traffic by means of studs, kerbs or fences. A distinction is made between a bus lane and a busway as follows: A bus lane is essentially a "paint and sign" scheme where buses are separated from other traffic by road markings or separators, which dissuade but physically permit crossing by both buses and general traffic. A busway involves construction where schemes may be partially physically segregated from other traffic, for example in the vicinity of bus stops (e.g., by means of island stops) or may be fully segregated from other traffic by kerbs or fences. Plate 2. With flow bus lane: Bangkok 10

15 Plate 3. Contra flow bus lane: Bangkok A busway may be implemented as a traffic management measure, without complementary improvements to bus operations and management, but busway transit involves a package of such measures with the general aim of promoting high output. Thus busway transit includes a right of way for the exclusive use of buses, with at least one section of busway and some additional features like well designed bus stops, special operating methods (bus convoys or express operations) and efficient fare collection methods. Clearly defined routes with name's like "green line" or "circle line" can add to the image of the service. The earliest busway transit schemes were introduced in Europe in the early 1970s hut in the late 1970s and early 1980s a series of innovative busways was implemented in various Brazilian cities, many with the World Bank encouragement and assistance. Other examples of busways in developing cities are in Abidjan, Ankara, Santafé de Bogotà, Istanbul, and Lima; plans exist for others in Bangkok. Jakarta. Karachi, Nairobi and Shanghai. Other bus priority schemes (with flow and contra flow bus lanes, bus only streets and spot improvements) were also implemented in many cities. While some schemes were very effective many were ineffective due to enforcement difficulties and poor design. (b) Special operational measures A basic busway, comprising one lane for buses in each direction is essentially a traffic engineering measure. Its main shortcoming is that a bus boarding at a bus stop or a break down bus delay the operation of other buses on the busway as overtaking is possible only by using the lane for traffic running in the opposite direction. However, performance of this basic busway can be enhanced substantially by adopting various special operational measures in order to form a "busway transit system" (see table 3). Table 3. Special operational measures Busway transit = Busway infrastructure + Special operation measures Special operational measures include: bus overtaking facilities at stops; trunk and feeder operations; bus ordering (placing buses in the correct order at the beginning of a section); high capacity buses (e.g. articulated or double deck) off board ticketing; traffic signal techniques to give buses priority at intersections; bus dwell time management (to eliminate excessive delays at very busy bus stops); and guidance systems (e.g. O Bahn). 11

16 Where passenger demands are high, the provision of facilities to permit buses to overtake one another at bus stops can increase throughput and commercial speed considerably (see plate 4). This is because bus congestion is reduced and buses are no longer "trapped" behind one another in a single lane (as occurs with trams or light rail vehicles). Trunk and feeder operations also offer good performance. In this system, feeder buses collect passengers and bring them to a transfer terminal, where they transfer to line haul buses; some systems allow transfer without payment of an additional fare. Plate 4. Busway transit stop, with overtaking facility: São Paulo Early work in Brazil led to the development of a high capacity bus convoy scheme (COMONOR), in which buses were assembled at the beginning of a section in the order in which they would stop (rather like a train). Although not joined together, the buses started and stopped broadly in unison. Although initially successful, COMONOR was found to be too difficult to sustain. The system evolved into 'bus ordering" in which buses are allocated to one of three groups (A, B, or C). The buses arrive in random order at the beginning of a section and are marshalled into the preferred sequence, although not into strict convoys. This method operates effectively and can improve commercial speeds at high levels of passenger demand. Line haul capacity can be enhanced by the use of high capacity buses, whether articulated, double deck or with the use of bus plus trailer combination. However, passenger transfer capacity at bus stops is often the constraint on system performance, and door configurations and ticketing arrangements are often more important than bus capacity per se. Bus delays at bus stops can be minimized by collecting fares and issuing tickets prior to passenger boarding. Through ticketing of the type adopted in Curitiba (trunk and feeder buses) and São Paulo (bus metro tickets) also contribute to reduced boarding times. At bus stops where passenger volumes are very high, excessive bus dwell times can occur when too many passengers try to board incoming buses and block the doorways so that the doors cannot be shut. This problem can be minimized by assigning staff to control boarding. (c) Current usage of busway transit More than 40 busways exist throughout the world; many of these are in Brazil, where both Curitiba and Porto Alegre have five busway corridors of aggregate lengths 53.7 and 27.5 km respectively. Examples of individual busway transit schemes in cities of developing countries are shown in table 4. It should be remembered that the length of the busway relates to the section of reserved track, and not the route length over which buses using the busway are operated. (d) Bus design 12

17 The busways included in table 4 use mainly standard diesel engined vehicles of 10 12m length with a capacity of spaces ( in crush conditions). In São Paulo, the busways also operate with double deck and trolley buses. There is no known example of a busway which permits the operation of small vehicles, and, in general, the larger the bus used on the busway, the higher the capacity of the busway. However, there are limiting factors to this generalization: the number of exit/entry points to the bus is limited to two and possibly three doors (perhaps six in the special case of a bus trailer combination), whatever its capacity; large capacity double deck vehicles can encounter exit delays from the upper deck. Another limiting factor on most current bus designs is the floor height, which tends to restrict ease of access and egress. Bus design is changing in an attempt to improve on the problems of boarding and alighting. In one design direction manufacturers are producing buses with low floors; another innovation from Brazil, however, is to provide high platform access to specially modified standard buses which have wide metro like access (see plate 5). Location Abidjan: Blvd. de la Republique Table 4. Examples of busway transit in cities of developing countries Length (km) Average stop spacing (m) Average junction spacing (m) Special features Annual passengers (millions) none 95 Ankara: Besevler Dikimevi none 35 Belo Horizonte: Av. Cristiano Machado overtaking at stops 80 Curitiba: Eixo Sul trunk and feeder 45 Istanbul: Taksim Zincirlikuyu none 55 Porto Alegre: Assis Brasil bus ordering 130 Farrapos bus ordering 85 São Paulo: Av. 9 de Julho/S. Amaro (e) Design of busways bus ordering and overtaking at stops Busway track may be located along an existing or a new right of way. For an existing right of way, the bus track may be located in the centre of the road (median; see plate 6) or along the sides (lateral; see plate 7). Purpose built busways can comprise a dedicated at grade bus road, a dedicated right of way along a new road or an elevated busway Plate 5. Raised bus stands for faster boarding and alighting of bus passengers: Curitiba

18 Most busway transit schemes are physically segregated from other traffic along their entire length using kerbs, fences or heavy studs; a few have segregation only at island bus stops. Plate 6. Median busway using the central reserve of a dual carriageway: São Paulo Plate 7. Lateral busway using one half of dual carriageway: Istanbul In order to minimize disruption to busway operations, the number of roads crossing the busway is usually limited to main thoroughfares only. Kerbs or barriers may be placed to prevent traffic turning across the busway into or out of minor side roads. In such cases, side road traffic is restricted to right turn in/right turn out (right hand rule of the road) and "Q" and "G" turns are used to concentrate traffic on to a limited number of cross routes. Such arrangements have an impact on local access to adjacent properties, which needs to be considered very carefully. Bus stops are an important component of busway transit design and operations; delays can occur due to the interference of buses with one another, and due to interference between passengers awaiting different buses, as well as the time lost in boarding and alighting. Some very busy bus stops must be designed to handle in excess of 300 buses per hour and 5000 passengers per hour. To accommodate such numbers requires long bus stop areas, facilities for overtaking, with possible use of parallel bus bays. All bus stops are low level which incurs additional time penalties for boarding and alighting passengers who have to negotiate the bus 14

19 steps. As noted above, modem buses with low level floors improve the problem, as does providing as many access doors on the vehicle as is practicable. A high degree of traffic signal control is generally required in order to manage high bus and general traffic volumes, without excessive delays. However, where bus flows are high, there is a "constant" call for green signal time by buses and selective detection of buses may not be appropriate. Signal control can be used to aid bus movements in the following ways: Selective detection of buses to extend a green phase or to recall a green phase; Demand dependent stages (which enables a bus to call a stage which would not otherwise occur); Signal time biassing to favour a stream with a high proportion of high priority vehicles; "Gating" in order to manage queues in favour of high priority vehicles. It is important to organize suitable collector and distributor systems to feed buses on to the busway and to permit them to leave the busway, without undue congestion. The capacity of the collector and distributor systems should at least match the bus demand at the relevant locations. This can be difficult where one or more busways lead into a city centre; in such cases, special arrangements are needed to disperse high bus volumes into terminals or into a circulation system comprising bus roads or lanes. The use of a range of bus priority techniques in and around a city centre will usually be essential to enable a busway to function effectively. Where enforcement is expected to be a particular problem, due to poor road user discipline, physical and electronic measures are available to dissuade other vehicles from entering the busway. (f) Guided busways A "guided busway" is simply a transitway equipped with a guidance mechanism (tracks) which physically guides the bus enabling it to travel at speed in a relatively narrow right of way. The prime advantages of a guided busway compared with a conventional busway are: The track provides a permanent physical presence, which makes the system more "visible" to politicians and public alike; Where the right of way is severely constrained, or land values are high, guided buses can operate at high speed in a right of way about 1 metre narrower than that of a conventional busway; however, this advantage is lost at junctions (where capacity is usually critical) in the case of guidance systems which require an entry splay; The track "occupies" the right of way and makes violation by other vehicles extremely unlikely. The prime disadvantages are the additional cost compared with a conventional busway and the severance effect in urban areas. It appears that the prime locations for guided busways would be in suburban areas requiring high speed operations. A guided busway can offer broadly equivalent levels of service to LRT, but at much lower capital cost. It also has the advantage over LRT that the vehicles can leave the track and so offer door to door service over a wide catchment area, without enforced passenger interchange. (g) Busway transit performance The practical capacity of busway transit for various design characteristics are summarized in table 5. These estimates are based on surveys of the performance of existing busways, which use standard or higher capacity buses. It may be concluded that well designed and efficiently run busway transit systems can achieve consistent flows of 25,000 pass/hour per direction, at speeds of up to 25 km/hour. (a) Outline 4. Light rail transit As the term implies, LRT usually employs vehicles and track construction which are less substantial than a full metro. Some systems, including those in Manila and Istanbul, use lightweight vehicles on a system which has 15

20 an exclusive track and high platforms similar to many metros. Trams are a basic form of LRT which have limited rights of way over most of their route, sharing roadspace with ordinary traffic. Trams were first introduced in 1832 in the United States of America to overcome the poor road conditions experienced by horse drawn buses. This led to an improvement in ride quality, due to the use of steel rails; furthermore, because of the reduced friction, one horse was able to pull more passengers. Thus the tram was able to attract more passengers, and carry them more cheaply than horse drawn omnibuses. Table 5. Measured and estimated busway performance Description Example Measured peak hour flow: passengers per hour per direction. Basic busway no options Ankara, Istanbul, Abidjan Estimated practical capacity: passengers per hour. 7,300 19,500 5,800 18,100 Trunk and feeder Curitiba 9,900 13,900 24,100 Bus ordering Porto Alegre 17,500 18,300 8,200 14,700 Overtaking + express services Belo Horizonte São Paulo 15,800 20,300 14,900 27,900 Optimum combination of None 30,600 high capacity options Electric traction became practical around the turn of the century and encouraged further investment growth in a very short time. However, the great advances in automobile technology in the early part of the twentieth century led to bus design improvements which were not matched in the tram industry. The tram began to be seen as a down market form of transport. Although still important for mass transport of low income earners, its role was under attack; trams were seen as obsolete, and a cause of congestion. The onset of cheap diesel fuel, and the greater flexibility of bus technology saw the demise of many tram systems. In more recent times, however, some city authorities have taken a fresh view of trams and LRT. In post war Germany, and more recently in France, a desire to reduce imported fuel consumption has been used to promote light rail. Most recently the issue of environmental protection has played a significant part in decisions to build LRT in North American cities. (b) Current use of LRT Worldwide, there are more than 150 tram or LRT systems in use; the majority are concentrated in Europe and North America. Most systems (90 per cent) carry fewer than 20 million passengers per line per annum; the intensity of use of each line is typically below 2 million passengers per km per annum. Few developing country cities have a tram or LRT system. Those that do are listed in table 6. Two of these systems (Manila and Istanbul) have grade separated track and sophisticated signalling and control; the remainder are largely tram systems (see plates 8 and 9) with some sections (extensive in the case of Tunis and Cairo's Heliopolis and Alexandria's Ramel networks) of reserved track. All systems use lightweight cars, however. 16

21 Plate 8. Two car tram set of Cairo Transport Authority. Plate 9. Veteran trams still used in Hong Kong (c) Design of LRT Trams and LRT usually run on standard gauge track (1435mm); in the case of street operation, the rails lie flush with the road surface. Where the track is segregated from other road traffic, the tracks may be raised and laid on ballast, which allows for ease of maintenance. One of the characteristics of LRT which distinguishes it from a metro is that the minimum radius of curvature can be as low as 10m (though more commonly 20m), which allows the planning of systems in tight and sinuous rights of way. LRT can also work on steeper gradients (up to 8 per cent) than heavy metro systems. As with busway track, an LRT track can have a lateral or median position within the road alignment. The transit way can be delineated by "paint and sign" or by physical barriers like curbs, fences or studs. In the case of a median position of the track, special provision must be made for boarding and alighting passengers, who may have to cross the path of on coming traffic to access the tram. In a few cities, the concept of a shared transit way for both bus and LRT has been tried at special locations (bridges, tunnels and passenger stopping points); this requires a greater than minimum distance between tracks to accommodate the less precise nature of bus driving. 17

22 City Table 6. LRT and tram systems in cities of developing countries Population (millions) Alexandria: 3.5 Network length (km) Number of lines Passengers carried per annum (millions) Ramel Madina Anchan Ascuncion Cairo: 12.0 Heliopolis CTA Calcutta Campinas Changchun Dalian Guadalajara Istanbul Manila Mexico City Monterrey Rio de Janeiro Tunis Rolling stock is lightweight and usually powered by electric traction from overhead supply. The cars are operated singly or in trains consisting of two or three cars. Tram cars have 4 or 6 axles, while LRT may typically have articulated 6 or 8 axle cars, or multiple unit trains of 4 or 6 axle cars, or two 8 axle cars. Passenger loads can vary from 100/180 in tram cars to 250 in a modern LRT car. Again, as with busway transit, delays to LRT can be critical at junctions. Similar methods of giving priority are used: limiting the number of crossing roads; eliminating right turning (or left turning for right hand drive) traffic across the path of the tram; signal actuation by approaching trams; special signal phases for trams. Grade separation may be required at very busy junctions. Trams are most likely to have low level access (like buses) while modern LRT may well be designed for access from raised platforms. Some modem LRT cars have low floors which improve low level access. Delays at tram stops are minimized by providing several wide access doors for each car. In the nature of a fixed rail system, however, there is no possibility of one tram overtaking another. Delays to LRT vehicles can become cumulative, causing large irregularities in arrivals. (Headways of 16 min. have been recorded in case study systems having a nominal 3 min. headway.) This leads to further delays as passengers block doors whilst empty following trains wait to access the station. Most LRT will be operated under manual vehicle control, i.e., within the visual capabilities and judgement of the driver. The driver can close follow a preceding vehicle, and headways can be very low (though speeds are correspondingly low). Some high technology LRT schemes, in particular grade separated, will be run under some form of automatic train protection. This involves a sophisticated signalling system which controls the spacing between successive trains. (d) LRT performance There is little recorded information on passenger handling performance of LRT systems. Pre war Austrian and German streetcar systems were recorded as carrying between 20,000 and 25,000 passengers per hour per direction; but this was under exceptional circumstances of excessive over crowding, slow speed and absence of much competing road traffic. More typical for street cars are the figures shown in table 7, which records information for developing country systems. The Cairo CTA tram network has about 10 per cent of its length protected from other traffic, whereas the Alexandria El Raml network has 80 per cent protection. 18

23 Table 7. Measured LRT performance in cities of developing countries Description Examples Measured peak hour flow: passengers per hour per direction. Street car: low protection Cairo CTA 2,300 high protection Alexandria El Raml 11,600 LRT: high protection Tunis 9,300 grade separation Manila/ Istanbul 25,000 The LRT networks carry more passengers. In Tunis the majority of track is protected (see plate 10), and trams have priority control over signals at the at grade intersections. In Manila, the complete track is grade separated (see plate 11), and the performance of the system is correspondingly high. (a) Capital costs 5. Costs of the systems Out turn cost data for existing travelway schemes vary according to design standards, construction procedures, initial condition of the roadway, local inflation rates, exchange rate variations, and so on. Table 8 presents investment cost ranges for the main infrastructure and equipment components of the mass transit options. For comparative purposes, the equivalent cost ranges for heavy metro are included. Many of the infrastructure costs (particularly the labour component) can be contained within the domestic economy, but the cars, power and control equipment may well involve significant expenditure of foreign exchange. LRT and tram systems will incur higher off shore costs than busway transit, and may also need additional training and implementation programmes which depend on continuing foreign technical assistance and associated foreign exchange outflow. Plate 10. The new LRT in Tunis 19

24 Rolling stock: standard bus Plate 11. Manila's elevated light rail transit system Table 8. Travelway equipment and infrastructure costs (after Armstrong Wright, 1986) (costs in $US millions, 1993 prices) Busway transit Tram LRT Metro 0.12 per bus double deck bus 0.15 per bus articulated bus 0.20 per bus tram car 0.45 LRT car per car 1.2 per car metro car 1.5 per car Elevated structure 20 per km 20.0 per km per km per km Tunnel Segregated way per km per km per km per km per km Track per km per km per km Signals 0.02 per junction a 0.02 per junction a per km per km Power supply per km per km per km Stations/stops surface <0.05 per unit <0.05 per unit 0.2 per unit per unit elevated per unit per unit per unit per unit underground per unit per unit Transfer terminal per unit per unit Depot per unit per unit per unit per unit Workshops per unit per unit per unit per unit a Traffic signal costs shared with other road traffic. A typical at grade, partially segregated travelway is likely to cost between $US 1 and 2 million/km (1993 values), excluding vehicles and terminals, based on the following assumptions: No land acquisition would be required and existing road kerblines would not be moved; The existing median would be removed in order to allow busway construction; Existing road drainage would be adequate and would not be modified; 20

25 The existing road pavement would be adequate except in the transit stop area, where complete reconstruction would be required; No extensive diversion of public utilities would be required. Providing guidelines for the cost of an elevated structure poses greater difficulties than for an at grade travelway, due to the wide range of possible construction techniques, foundation conditions, bus stop treatments and other features. There is little direct experience of the cost of elevated busways since no extensive sections have been constructed, although several are under consideration. Based on British conditions, a representative elevated travelway for buses or trams is estimated to cost of the order of $US20 million per km (1993 values). Elevation need only be necessary where traffic capacity at selected junctions is critical. Clearly if transit stops can be accommodated at grade, considerable cost savings are possible. The costs of associated infrastructure will vary from place to place and will depend on local requirements. The cost of a footbridge might typically be in the order of $US80 100,000. Where a new and comprehensive transit system is to be implemented, new depot and workshop facilities will be required; this is certainly the case for rail schemes and will also be the case for busways where a purpose fleet is acquired. Again, costs depend upon many local factors, but a new depot for about 200 buses (or 100 tram cars) could cost in the region of $US6 8 million excluding land costs. A workshop and central stores facility might cost a similar sum, depending upon the scale of facilities required. If a trunk and feeder system is to be operated, transfer terminals will be needed along the main axes together with a terminal station at the end of each corridor. Costs depend upon many local factors, including standards, but could be of the order of $US0.5 million for a basic transfer station and $US0.8 million for a basic terminal station excluding land costs. Terminals may offer development opportunities and additional sources of revenue. The overall capital costs for a complete system are estimated in table 9. The more grade separation, tunnelling, use of heavy rolling stock and sophisticated control equipment, the higher the cost. Table 9. Capital costs of mass transit schemes (Costs in $US millions (1993 prices)) Bus lane Busway transit Tram LRT Metro Capital cost per route km. < Note: Includes rolling stock, except in case of bus lanes (b) Operating costs The key components of operating a transit system are labour, energy and replacement materials. Estimates of operating cost per passenger km are given in table 10. These costs include depreciation on equipment, but not on the initial infrastructure or any financial charges. Table 10. Operating costs of travelway systems (after Armstrong Wright, 1986) (Costs in US cents (1993 prices)) Bus on bus lane Busway transit Tram LRT Metro Operating cost per passenger km (c) Lifetime costs Lifetime costs of the different systems vary considerably, being heavily influenced by the extent to which track costs are included in the summation. Buses operating on bus lanes have little or no track cost, and their annual operating costs are likely to represent the major cost component (one year's operating cost being typically 25 per cent of the total investment cost). Over the lifetime of a bus lane scheme, operating costs may account for 85 per cent (or more) of total discounted costs. At the other end of the scale, the annual operating costs of a metro represent perhaps only 5 per cent of capital costs; over its lifetime, the capital cost component of total discounted costs for a metro is as much as 75 per cent, reflecting the massive initial costs in the early years of the project. (d) Financial performance 21

26 Little is known of the financial performance of low cost mass transit schemes. In the case of busways, the scheme's performance is usually subsumed within the total financial performance of the participating bus company; neither would it be normal for the capital costs of the track to be included in bus company accounts. Tram and LRT schemes fall into two groups: those which are well established systems using old technology (like Cairo and Calcutta) and the newly established systems using modern equipment (like Manila and Tunis). The financial performance of the old systems is generally poor. As an example, the Calcutta Tramway Company could only cover about 40 per cent of its operating costs exclusive of depreciation and interest. The modem systems present a better picture. For example, the Manila LRT is able to cover its direct operating costs; however, it makes only a small contribution to interest charges and loan repayment much of which is in foreign currency which has appreciated against the national currency. Very few public sector bus or rail services, if any, are able to rely entirely on direct revenue. An examination by the World Bank of 20 representative public sector bus operators revealed that on average only 62 per cent of total costs (operating costs, depreciation and interest) was recovered from fare box revenue and other direct income. For urban railways the figure is even lower and generally in the region of 30 per cent. Private sector bus operators (who service the majority of bus networks) are usually able to survive on fare box revenue alone; an examination by the World Bank of 33 large representative cities found that fare box revenue was, in effect, the sole source of funding for over 75 per cent of bus and minibus trips. The funding of bus services solely by users is possible because the majority of users accept a low standard of comfort and safety in travel, and labour costs of bus operation are low. 6. Summary The great advantage of busway transit, in particular over metros, is in its flexibility: the ability to change alignments relatively quickly in response to changing demands; the ability to implement progressively as demand increases or as funds become available; the ability to implement piecemeal projects in key areas and the ability to penetrate development, not necessarily where the main right of way exists. Perhaps most important of all, the development of busway transit builds on the city's existing wealth of experience in bus operations. One of the main disadvantages of busway transit, however, is that its implementation requires the active cooperation of the highway authority, the licensing authorities, the police and bus operators. Such cooperation can be difficult to achieve. LRT and tram systems are more expensive to construct than busway transit, particularly where the latter does not include the acquisition of a new bus fleet. The great advantages of a modern rail system over busway transit is in their image and general environmental "friendliness". Image is clearly important to decision makers and there is little doubt that it would weigh heavily in a straight choice between the two. The image of a busway transit system could be improved if the bus fleet was modernized and given a special livery; but usually the bus is cast in the role of the traditional and somewhat hackneyed transport, which has little to offer the modern world. LRT is seen as an example of modern technology which can contribute to civic prestige; it invites political support as a high profile gesture towards tackling urban transport problems. The great disadvantage of trams are their inflexibility (in route network) and the need to provide a complete track system before trams can be operated. D. The impact of public transport segregation 1. The users The majority of beneficiaries of busway transit in cities of developing countries are likely to be the users of existing public transport. In the industrialized world, there has been no evidence of any major switching to bus from private modes, as a result of the introduction of priority measures. The evidence for switching to LRT is less clear; it is known, however, that the patronage of metros in the developing countries has been drawn almost exclusively from the users of existing public transport, or through new trip generation effects. There are strong conceptual grounds for believing that most users of private vehicles in developing country cities are unlikely to be attracted to the use of public transport; these travellers come mainly from high income groups, who will value comfort and convenience of personal transport very highly. Generation effects may be substantial; for a number of schemes in industrialized countries, generated travel has constituted over 15 per cent of patronage, and typically the new transit scheme has contributed to a 3 per cent per annum growth in use of public transport. 22

27 Many earlier studies have attested to the level of user benefits which result from bus priority measures. Typical time gains in European and North American cities, measured over the length of the scheme, range between 20 and 50 per cent. Similar observations have also been noted in Singapore, Bangkok and Porto Alegre; in the latter, journey times were reduced by 29 per cent. Small improvements in regularity have also been noted and there should be some improvement in the quality of travel, particularly if the investment includes new rolling stock which is clean and comfortable. 2. Transport operators Very little quantitative work seems to have been done to assess how bus operators benefit from bus priority schemes. The benefits are likely to occur either through reduced fleet size required to service a route with the same headway, improved output per unit of cost and reductions in operating costs which are speed related. While it is known that bus speeds have been improved as a result of building busways (for example, by over 40 per cent along the bus corridors of Porto Alegre), the corresponding improvement in bus output and operating costs is not known. In Belo Horizonte, bus fuel consumption was reduced by almost 20 per cent as a result of using busways, but this is an isolated case of reporting. Even so, it should be assumed that bus and tram operators must benefit to some degree from the enhanced operating environment brought about by provision of travelways. Where frequency and reliability in the service is improved, in response to the priority scheme, new passenger trips may be generated. It is also conceivable that a busway scheme may generate opportunities for scale economies which might not otherwise be feasible (e.g. use of high capacity vehicles). 3. Non users The impact of bus priority schemes on other road users could negate the value of the scheme, and there is no doubt that user savings have been outweighed by non user disbenefits for some bus lane projects. This is usually traceable to some technical design feature which could be modified, or to the fact that bus flows are too low. It is likely that bus flows in excess of 60 per hour (or public transport passenger flows in excess of 10,000 per hour) could always warrant a reserved track. Interestingly, in both Singapore and Bangkok the improvements in bus speeds, resulting from bus priority schemes, were complemented by improvements in the speed of other traffic, a pattern which can probably be attributed to improved driver behaviour resulting from the segregation of vehicles with widely differing characteristics. Travelways are often promoted on the basis that they can contribute to relief of city centre traffic congestion through encouraging a modal switch from private to public transport. The evidence for success in this objective is, unhappily, not strong; most users of new modes have changed from another transit mode (or in the case of busways, their bus simply switches from an unreserved to a reserved track within the same right of way). In cases of some rail transit schemes this change has been forced on the traveller through the elimination of existing transit services to protect the new system. Even where there may have been a switch from private to public transport, the improved traffic conditions on the road network will quickly induce new car traffic to emerge. However, there are reasonable grounds for supposing that travelways could have some influence on the spread of traffic congestion. With increasing car ownership and use, city centre traffic congestion reaches what has been described as the threshold of the intolerable; it cannot get any worse, and assuming all traffic engineering measures have been exhausted, can only spread more widely, rather than more deeply. New roads to access the city centre may improve the situation, but there are limits to what can be achieved, simply because the land is not available and the resulting environmental damage is likely to be too great. A mass transit system, making the best use of the existing road system, provides the capacity needed to access the city centre, without the associated penalties of road building. In providing greater access, the mass transit system helps to reduce the spread of traffic congestion. In particular, the improvement of public transport can make the restraints of the use of private motor vehicles more politically acceptable, and such restraints may become necessary as well for arresting congestion as for environmental reasons. However, the environmental impact of public transport segregation will require also detailed assessment in the light of scheme characteristics and local circumstances. Travelways, by their nature, provide a high speed track in built up areas where pedestrian activity will be intense. The resulting severance, safety, noise and air pollution effects all warrant particular attention. Severance effects can be minimized, and safety enhanced, by suitable urban design and by the provision of adequate pedestrian crossing facilities. Some travelways have been designed so as to minimize the interaction of pedestrians and vehicles; but pedestrian crossing points are inevitably necessary (if only to access the travelway), as is interaction with 23

28 other traffic at grade junctions and along unprotected rights of way. Evidence from transit schemes in industrialized countries suggests that accident rates per vehicle km on light rail are a little higher than those for buses, but on a passenger km basis light rail is generally safer than buses. On street noise and air pollution effects of busway transit can be minimized through the use of modem, LPG (liquid petroleum gas) or CNG (compressed natural gas) powered buses, or electric powered trolleybuses. 4. Urban development The essence of a city centre is that it is the most accessible point from both within and without the city. This superior accessibility is important for many activities, and in particular for those central functions which serve a wide area and/or need a wide labour market: head offices, central government offices and legal institutions, financial institutions, media firms, theatres, department stores etc. and all the supporting organizations (catering, hotels etc.) that exist to serve these central functions. The fortunes of the city centre are at risk if the public transport system proves inadequate in supporting these central functions. This is because the vast majority of commuters to any big city centre depend almost exclusively on road based public transport for access. If the city centre becomes congested, then its relative accessibility may suffer, because the public transport system cannot perform effectively. As a result, new central functions will be discouraged from locating in the city centre and old established ones may start to drift away. Clearly, there is an intimate and vital relationship between the well being of the city centre and its public transport system which should never be overlooked. For this reason, it is becoming increasingly apparent that the development of urban transport in the major cities may be reaching a stage where priorities have to be imposed, and mode choice has to be managed to the advantage of public transport systems. Once the limited supply of road space feeding the city centre is exhausted, the only possible relief would seem to be through the development of a mass transit system which makes best use of the available road space. Apart from promoting the performance of public transport, and thereby contributing to the healthy growth of the city, travelways may have their own intrinsic developmental impact on a city. Mass transit schemes have sometimes been proposed to enhance or encourage new city development and/or renewal. For example, it is reported that the development of the LRT in Manila has played a key role in shaping the urban development of the metropolis, triggering the redevelopment of the traditional centres of business and trade, and encouraging commercial growth along its route. this impact of travelways is not fully understood and has not always worked, in particular where planning controls on urban development are weak. Generally, if a city has a buoyant economy then a mass transit system can contribute to and accentuate that condition by removing any accessibility constraint; on its own, however, the mass transit system can do little. Thus ideally the travelway should be developed in unison with other on going major developments within the city. A number of mass transit schemes have managed to capture some of these developmental benefits for their own financial gain. This has been achieved through the commercial development of the air space above terminals and interchanges; these revenues can contribute to both the capital cost of the structure and/or to general income. 5. Other impacts Public transport is often used by people who do not have access to private motor vehicles and a high proportion of bus passengers tend to be children, old people and women. This means that improvements to transit services can have important social impacts. For example, suitable bus services can offer mobility to women who may not otherwise have access to motorized transport, and can increase their access to work opportunities, and to educational and social activities. In developing countries, the foreign exchange requirement of a proposed investment can be an important criterion in the selection of a technology. Busway transit offers considerable scope for civil engineering construction by local contractors and, where a local assembly or body building industry exists, a substantial part of bus costs can be incurred locally. As noted earlier, LRT is more dependent on imported technology, and hence a requirement for foreign exchange to pay for rolling stock and, in the case of advanced systems, signalling and power equipment. This could be off set in some countries by its use of locally generated electric power rather than imported diesel fuel. 6. Economic evaluation An economic analysis of any travelway project should try to take account of all the impacts which have been discussed. Many of these impacts are clearly difficult to quantify. A busway scheme is likely to improve bus commercial speeds and reliability, and therefore the potential benefits are typically: journey time savings to 24

29 bus passengers (including the value of increased reliability), and bus operating cost savings (including a possible reduction in fleet size). In general, the majority of benefits are likely to be associated with time savings at junctions. However, the analysis should also take into account changes in journey times and operating costs for other road users, especially if some reassignment of traffic is anticipated. Depending upon local geometry and traffic flows, introduction of a busway may increase or decrease the capacity available for general traffic, particularly at junctions, and detailed junction analyses are required to estimate these effects. There are no definitive studies of the economic viability of busways, but as noted earlier, bus lanes carrying more than 60 large buses per hour per direction, are likely to yield a good economic return. Post evaluation studies of the economic worth of advanced LRT have been undertaken in connection with the Tunis and Manila systems. In both examples the economic rate of return was estimated at just under 12 per cent. Finally, no study has examined the crucial issue of the developmental benefit to the city centre of a travelway scheme. It is a very complex issue since it raises questions about the city structure and its efficient growth. These are questions which go somehow beyond the bounds of transport planning, and pose major conceptual and technical problems of analysis. If the continued growth of the urban centre is an urban development objective then the travelway scheme can be considered as a major positive contribution to achieving that end. E. Planning considerations (a) Planning context 1. Scope of planning A primary consideration in the development of any travelway scheme is that it forms part of an overall strategy towards public transport development within the city. The policy is likely to be one which encourages the development of public transport facilities as a means to provide the greatest and most efficient access to the city centre. (If the policy is one of laissez. faire, travelways are unlikely to receive any strong support and public transport will inevitably suffer severely from the congestion caused by other road users.) Cost constraints and demand will help determine which of the low cost options (buses on bus lanes or busways and LRT) will best serve the purpose. Given the performance figures noted earlier, busway transit and LRT are likely to be suitable in a variety of locations, typical examples being: The main corridors of medium sized cities, where public transport travel demands are up to about 20 25,000 passengers/hour/direction; The secondary corridors of large cities, to complement rail mass transit; Outer city suburbs, to structure newly urbanizing areas. Medium capacity travelway systems can be viewed as a first step towards a "heavy" mass transit system. In this incremental approach, a right of way is secured which can be used initially for a busway or LRT, and subsequently up graded to a metro system. The idea of such a transitway has an appeal to urban "planners which dates back to the beginnings of modern urban design practice. Clearly, there would be problems of continuity of service during an up grading exercise, and there are likely to be technical problems associated with the change, but the basic concept of reserving land for long term transport development has considerable merit. Planning should be concerned with both current and future travelway needs, and in the context of future needs the transport planner should be working very closely with his urban planning counterparts. If future travelways can be identified at an early stage, the opportunity arises for protecting rights of way before major development encroaches. In the planning and design of any new river crossing or grade separated interchange, provision should also be made for the possible future installation of a travelway. For example, the new river crossings in Istanbul all have provision for a transit lane in the median, which could accommodate either bus or LRT. The identification of future rights of way should be based on sound transport analysis, rather than conceptually pleasing ideas. Travelways should only be necessary where it can be shown that traffic demands will become high enough to warrant a protected way. 25

30 Existing and available rights of way present the planner with a soft option for travelway development; again, however, there is no substitute for detailed demand analysis. Several metro systems have suffered from under utilization because their alignments (along a disused rail track) do not adequately follow demand needs. Equally, a number of suburban ring railways have been up graded, but with no apparent success in attracting passengers; the alignments are critically at fault. In planning either a busway or LRT system, much attention will have to be given to wider operational and organizational arrangements as well as the design of the physical infrastructure. Both are important components for the successful implementation of the transit system; issues like integration, funding and management require a realistic planning approach as to what is achievable. Too many metros have not lived up to the ambitious expectations of planners who have failed to appreciate the difficulties of organizational and institutional change. (b) Allocating roadspace Where a transitway is to be inserted into an existing right of way, difficult issues arise over the allocation of roadspace between the conflicting demands of different road users. In many cities, there is insufficient roadspace to meet the unconstrained demands of all road users and it is necessary to have a demand management policy to guide the allocation of roadspace. In cases where existing roadspace is limited, its partial allocation to a transit way may be justified because: It can carry 20 25,000 passengers per hour and direction, whereas a lane used by cars only is unlikely to carry more than 2 3,000 passengers per hour and direction at normal occupancy levels; It may be easier to divert cars rather than buses to alternative routes; It may be more cost beneficial to allocate existing roadspace to buses and to construct additional roadspace to be used by all vehicle types, rather than to construct the infrastructure required for a high capacity rail system. The width of a busway lane should never be less than 3m (a bus being typically 2.5m in width). For most two lanes busway configurations with design speeds of 50 60km/h, the track width is likely to be 7m. Including separators, a median busway would occupy 8m of available right of way along running sections, and 11m at staggered bus stops. A median tram running section (two way) can have a narrower width of only 6m, but will in all likelihood have approximately similar cross section dimensions to a busway. Evidently, each lane of travelway will occupy more than the space formerly occupied by one vehicular road lane. A 24m wide road might lose three or four road lanes for the provision of two travelway lanes. A right of way which can accommodate a busway system can invariably accommodate a tram system. A high specification LRT would need some changes, including the provision of high level stops and possible grade separation at junctions. If such up grading is considered to be a distinct possibility at some time in the future, then the right of way should be examined with this in mind from the outset. The land take for LRT stations could be considerably more than that for a busway or tram system. In Manila, LRT platforms are 100m in length and 3.5m wide; there is also an associated concourse area for sale of tickets and passenger dispersal. The problem of land take in a confined space has been solved in Manila by elevating the complete system. (a) Capacity concepts 2. Defining the scheme The technical literature contains many ambiguous references to the "capacity" of alternative transit systems (metro, light rail, bus etc.). Such statistics usually refer explicitly or implicitly to "line haul capacity". However, it is also important to consider "passenger transfer capacity", i.e., the maximum number of passengers who can board and/or alight at a stop/station during a given time period since bus stop/station capacity is usually the limiting factor in a transit system. For busway transit in particular maximum line haul throughput decreases as passenger transfer demands increase. Line haul capacity is a variable and it is not possible to quote a single "capacity" figure for a transit system without reference to the demand for passenger transfer and station/bus stop capacities. 26

31 Applying the traditional concept of line haul capacity to bus systems is also doubtful because the number of "available passenger places" on buses passing along a bus track does not correspond to the maximum passenger throughput achievable. Without a trunk and feeder system, passengers will only board buses serving selected routes. In practice, it is impossible to fill every bus to capacity because of the imbalance between the number of empty spaces on a bus and the number of passengers boarding at each stop. In some cases, a bus leaving a bus stop will have empty passenger places while in others, the bus may be full and leave some passengers waiting at the stop. This variation in loading from one bus to another imposes practical limits on average load factors (the ratio of the number of passengers to available passenger places). The available evidence suggests that without special operational measures, it is difficult to achieve an average load factor in excess of per cent without severe overcrowding on many buses. Furthermore, since boarding and alighting times are substantially different, maximum line haul passenger throughput is higher in the predominantly alighting direction than in the predominantly boarding direction, for a given demand for passenger transfer. For planning and design purposes, it will generally be critical to examine the predominantly boarding direction. Unless bus entry to the busway is controlled, bus arrivals will be irregular and may lead to substantial peaking. TRL surveys of existing busways suggest that the peak 5 minute flow can be between 1.3 and 3.0 times the hourly rate. (b) System planning parameters Table 11 re emphasises the main factors which will influence the capacity of a travelway. The most important is probably the degree of segregation between the travelway vehicles and other traffic. 1 The other critical components' of a transit way system which could constrain its capacity may be a running section (link), a station or bus stop, a junction or, in the case of busway transit, the collector/distribution system. In general, running sections are unlikely to be a bottle neck. Average bus headways of 4.5 seconds have been measured over short periods for bus platoons travelling along busways in Abidjan and Porto Alegre, corresponding to 800 buses/hour. Similarly for trams, the feasible frequency of operation is high; as noted earlier, pre war trams in Vienna and Stuttgart were operating at about 20 seconds intervals, offering a line haul capacity in excess of 30,000 spaces per hour per direction. Right of way characteristics: Table 11. Travelway capacity determinants road cross section degree of physical segregation from other traffic Junction design and control horizontal and vertical alignment road surface characteristics Transit stop characteristics: overtaking facilities spacing number of bays platform storage area passenger information platform height Vehicle characteristics: 27 vehicle size and capacity existence and control of doors number, location, width and use of doors number and height of steps floor height maximum speed

32 acceleration and deceleration rates Operating characteristics: route structure and scheduling driver behaviour fare structure and ticketing trunk and feeder bus ordering Passenger characteristics: passenger demand at stops passenger distribution by time of day behaviour General traffic conditions: volume and nature road user discipline encroachment In many cases, the capacity constraint on a system will be a single stop or station. The interactions between passengers, vehicle and driver characteristics, and stop layout are complex. Again, "capacity" is variable and dependent upon passenger behaviour, arrival time patterns and many other factors. Another important factor which influences performance is stop/station spacing. For a busway, for example, each additional 100m between stops will increase average operating speeds by about 2 km/h. Junctions influence the flow of vehicles along an at grade travelway and consequently affect both capacity and speed. Traffic signal controlled junctions (including pedestrian crossings) reduce the amount of green time available to the travelway and impose delays to some vehicles. These delays reduce vehicle and passenger throughput, and average commercial speed. In some cases, a particularly busy junction may be the critical bottle neck; this is the case on Farrapos busway, Porto Alegre, where the junction adjacent to the central area "controls" the flow of buses off the busway and into the city centre during the morning peak. In most cases, however, junction capacity will be greater than that of the most critical transit stop. For busway transit, the choice of route structure between trunk and feeder and an "open" system affects achievable load factors and other performance indicators. Trunk and feeder operations permit high line haul flows to be achieved, particularly with the use of high capacity buses, but at the cost of enforced passenger interchange at expensive transfer stations. Open systems can serve a wider direct catchment area (without interchange), but the irregularity of bus arrivals and dwell times limits sustainable performance levels. An LRT system is de facto a trunk and feeder network (though the feeders may often leave much to be desired). (c) Design, construction and implementation Table 12 summarizes the key components in the design and implementation of travelways. Heavy mass transit systems have a history of severe cost over runs and delays, some of which can be attributed to poor project organization and execution. The construction of an at grade separated travelway may not incur such major problems as occur in tunnelling and cut and cover construction for a metro, but some of the lessons from these exercises are worth recording. Construction along an existing, and probably congested road, will inevitably generate some traffic chaos, which will demand efficient management through the coordinated action of police, transport operators, highway authorities and others. Utility companies must collaborate in the identification and relocation of any services (telephone, gas, electric, water, sewerage) which interfere with the construction of the travelway; the identification of existing services from current records is notoriously haphazard, and estimates of the amount of effort and money required to relocate utilities is always too optimistic. Project implementation: land acquisition Table 12. Travelway design considerations 28

33 structure designs(elevated ways, grade separated crossings, ramps, depots, garages, stations, terminals, maintenance facilities) track specifications (road or rail, separators) equipment and rolling stock (power supplies, vehicle types) junction layouts and control (signalling specifications) contract and tender procedures project organization and execution Operational implementation: institutional establishment (status, location and accommodation) establish organization and operating procedures staff recruitment trials staff training Inadequate funding is another source of delay; construction companies will rarely want to bear any financial risk (except in the case where they are cofunding the project), and will quickly cease work if funds become tardy. Funding may become a source of concern because committed budgets are exceeded (because of agreed design changes or unforseen cost escalations because of, for example, unexpected service diversions), for purely political reasons (a change in administration which has no commitment to the project), or for technical financial reasons (for example, a change in foreign exchange rates which has a major impact on the value of foreign debt, or any sudden change in government finances which has an impact on public sector borrowing and the availability of government equity). Work may also come to a halt simply because of shortages in critical materials (like cement and steel). Land acquisition may be more difficult and time consuming to enact than expected. Land titles may be difficult to establish, and compulsory purchase may not be as smooth as would be hoped. In some countries court procedures and settlements can be extremely protracted. Operational implementation requires careful planning and attention to detail. The status of the organization needs defining in terms of objectives and responsibilities (see section (c)), and its administrative location and accommodation need identifying. Once appointed, new management needs to set in hand operating procedures and rules. Staff must be appointed, and trained; equipment must be accepted after trials. These are complex tasks which need good project management for timely and successful completion. One very good reason for adopting the low cost option of constructing a travelway (instead of a metro), is that it is a far less risky scheme than a metro. There are fewer things that can go wrong in its construction and implementation. (d) Regulatory framework, organization and management Regulatory control of the public transport sector, in some form, will be the norm in all cities. This can take the most stringent level of controlling vehicle numbers, routes and fares, as well as quality and safety; at the other end of the spectrum, only quality and safety may be prescribed. Transit ways present local governments with an additional control, that of access to the facility. This arises since they are most likely to be the owner of the LRT track and there are yet no known examples of the busways being privately owned. In the case of busways, the track authority (central or local government) may or may not have ownership control of the vehicles. This may influence the way in which access is granted. Access could be limited to certain routes, vehicle types or operators. A monopoly bus fleet (which will often be in the public sector) will obviously have sole access, as in Abidjan. In many Brazilian cities, including São Paulo, access is permitted to both private and public sector operators, reflecting their combined contribution to public transport output. In other cities, like Istanbul, preference is given to the public sector operator over private operators, particularly where the latter operate small vehicles. In some cities, such as Curitiba, both public and private sectors operate services as part of an integrated network, with a common fares policy, and with colour coding of vehicles according to function. 29

34 LRT is invariably a monopoly, with both track and vehicles belonging to the same owner which is almost universally in the public sector. There are some exceptions as to ownership: the new system in Manchester was designed, built and is now operated by a private consortium comprising the constructors (civil and electromechanical) and the principal local bus operator; the Manila LRT, though owned by Government, was managed under contract by a private concern. This arrangement was subsequently abandoned following adverse auditing criticisms. LRT and tram systems in the public sector are either part of a larger public transport service (for example, the Cairo Transport Authority, which is responsible for both buses and trams) or an individual company (like Calcutta Tramway Company). The more recent LRT developments have often been conceived within a bus company, and subsequently been born as an independent organization (for example, the Société du Metro Léger de Tunis grew out of the Société Nationale des Transports the bus operator in Tunis). There is much to be said for creating an independent organization which is not tied to bus working practices; different working skills, arrangements and conditions will be necessary for the new industry, and these may best be contained within a new organization. At the same time, links should be maintained in order to develop integrated networks. The track authority must necessarily be responsible for organizational and managerial arrangements for the effective running of the transit way. In the case of a busway this will include the provision of common services like management of the bus track (such as the provision of a tow truck to deal promptly with breakdowns) and maintenance of the track, bus stop facilities and traffic control devices. What is less obvious in the case of busways, however, is who should have responsibility for: Fare collection and ticketing, including off board ticketing and management of season tickets or travelcards, where the revenues are being shared between several bus companies; Driver training for the particular conditions associated with high intensity operations; Facilities and staff to undertake bus ordering, where appropriate; Provision of reliable and up to date passenger information; Supervision to limit bus dwell times at busy stops in order to avoid excessive delays and service disruptions. These are necessary operational features for the effective development of busway transit as a mass transit option. Whether these functions are to be performed by several agencies or by a single busway transit agency, specific arrangements must be made. Particular weaknesses which occur in practice are the fragmentation of responsibilities between the highway authority, bus operators and the police. This tends to lead to inadequate maintenance of busway infrastructure and to a lack of "track management" where there are several competing operators. In the case of LRT, the problem is less evident because track authority and operator are usually the same body. A final aspect of the organization of the travelway concerns the issue of integration. The general aim of integration is to focus as many passenger trips on the travelway as possible; this would be particularly important in the context of "closed" systems, i.e., trunk busways and fixed rail schemes, where the vehicles are constrained to the travelway network. This concept can take several forms: the physical integration of different modes (through, for example, common location of transfer points); the provision of feeder services by one mode to another (i.e., a trunk and feeder service which may employ two different modes); the elimination of parallel competing services (usually provided by different modes) in order to "protect" the travelway investment. In the latter case the network may be reorganized to provide feeder services to the travelway service. Service elimination and feeder service development are prone to both passenger and operator resistance; in the former case passengers have to make an enforced transfer, which can incur both time and cost penalties, while in the case of operators the interference with their routes and services may affect their profitability. Some incentive to passengers can be given by using a "through ticket" system, whereby the user pays only one fare for the whole journey and is not severely penalized (in cost terms) for the transfer. For this system to work, however, requires acceptance between operators as to the sharing of joint revenues; this can be difficult if the operators of the feeder service and trunk service are not one and the same. Few integration schemes in developing countries have worked well, a notable exception being the Curitiba busway transit system. 3. Funding projects 30

35 Although city authorities are not always in a position to be able to identify objectives for public transport, it is generally understood that funding is a problem, and mat fare revenues are often insufficient to cover operating costs, let alone replacement of existing fleet (this is particularly the case with public sector operators). The lack of funding for urban transport stems mainly from a number of inter related factors: The low income of users; The stagnation of national economies; Severe cutbacks of public spending; Stiff competition between sectors; Steep growth in outstanding debts; Decline in both official and private lending. Because of me difficulty of raising public funds, many governments are now looking to the private sector for urban transport finance, while still relying heavily on international and bilateral assistance. The funding of a system may often comprise several of the sources indicated in table 13. What is apparent from this table is that even where governments are not directly responsible for providing the funds, their involvement in the investment may still be critical through the provision of guarantees, rights to repatriation of funds, foreign exchange allocations, availability of development loans, etc. (a) Funding by users Direct income to the transport operators is mainly in the form of fare box revenue but sometimes includes small amounts of income from advertising, contract work and rentals from property rights. Fares are almost invariably controlled by government, and are often kept at an artificially low level in order that low income communities have access to public transport. This revenue is often insufficient to cover costs and must be supplemented by subsidies or loans, though some operators have to resort to other means (for example, non payment of taxes, delaying payments to staff and creditors, non provision for depreciation). In general, user revenues are insufficient to finance any capital spending on travelway schemes. Table 13. The main sources of public transport income Source Issues Revenue Fare box (user) Fare level User affordability Subsidy Loans Equity investment Govt. (b) Capital subsidies Development (property rents) Govt. general taxation hypothecated tax (betterment, employment, fuel etc.) tax incentives land Grants from aid donors (capital) Commercial (banking) sector domestic foreign Suppliers credits Govt. Aid donors Private sector operators constructors developers Cost revenue disparity Risk Access to capital Equity (rich or poor?) Level of subsidy and community benefit Productivity and cost drift Developmental benefit Industrial (donor) benefit High risk Availability to small operators Foreign exchange availability Exchange rate uncertainty Govt. Guarantees Developmental benefit Industrial (donor) benefit High risk Govt. Guarantees Repatriation of profits Exchange rate uncertainty Community benefit Most public sector bus and rail operators in developing countries rely heavily on both operating and capital 31

36 subsidies. Very few private operators receive direct subsidies but may benefit from infrastructure provided at below cost or at no cost. In some places subsidies are applied directly to certain types of passengers, for example, in the form of concessionary fares for students. Subsidy payments are a device for channelling non user money into transit schemes in order to achieve some socially beneficial purpose from the public transport system. The money can be raised from general tax revenues, but increasing interest is being given to both hypothecated tax and betterment tax as a way of tying the costs of the improved transit more closely to those non users who either benefit most from its development, or impose those social costs which justify the need for the development. In the case of betterment tax, the land and property owners who benefit from increased rents and property values, attributable to the transit development, would be liable to a tax on these revaluations. In the case of a hypothecated tax, private vehicle users might be taxed in some way (through an additional fuel tax, for example), with the specific aim of raising revenue to promote the transit system. In this latter case, although the private vehicle users incur a loss, it is intended that overall society incurs a larger benefit from the transit development which the user of a private vehicle can also use. The United States has a long tradition of hypothecated taxes. These are often introduced following a public vote, and specify, for example, that a certain percentage of the income from a sales or income tax will be used only for the benefit of public transport. In France, public transport in the larger cities is financed from a dedicated tax on employers of more than 10 workers. In Germany there is a fuel tax (Mineralölsteuer) which can be used for capital (but not operating) costs of transport projects. In Seoul, purchasers of new cars buy a subway bond which generates capital funds for the Seoul metro. There are few examples of the implementation of betterment tax; betterment is more often recouped by encouraging the active participation of a developer in the financing of a scheme (see below). There is a danger that developers could be deterred by high taxes or requirements to contribute to the costs of a rapid transit line, and hence decide to relocate in adjoining municipalities or in another town. Governments generally fund operating subsidies with the primary objective of providing the public with satisfactory services at affordable fares. While subsidies undoubtedly keep fares down, experience in the developing countries is that they often lead to inefficiency and increasing costs. Rather than being satisfactory, services are inclined to deteriorate under these circumstances, particularly in the face of growing demand. Inevitably they require more and more subsidies, which often cannot be met by limited national or local government budgets. Operating subsidies are also given in the expectation that car users will be induced to use public transport, and hence contribute to reduced congestion. Experience has shown that these benefits expected from subsidies are often disappointing and difficult to sustain. The use of capital subsidies as a means of funding urban transport infrastructure, rolling stock and equipment is widespread throughout the world. It does not raise the same objections usually associated with operating subsidies, and may, in fact, stimulate funding from other sources. Capital Subsidies are usually justified on the basis that they will permit the provision or improvement of systems with widespread economic and social benefits. Often these benefits stretch beyond the immediate users and it may be difficult to capture from all the beneficiaries their share of the costs. Most of the Brazilian busways, with widespread benefits, receive capital subsidies for this reason, as do a number of rail based systems. A danger with capital subsidy is that it can be applied to a prestige project for political reasons unsupported by proper financial and economic appraisal. Often in these cases, more appropriate, albeit less impressive options, are rejected. Capital subsidies can play an important part in capital intensive systems in which the interests of the private sector are being sought. In particular, they encourage private money to be forthcoming to fund the element of the cost that could be recovered from users in the form of tolls and fares. Capital subsidies also reassure investors of the commitment of governments to the success of proposed projects. They may be in the form of capital grants, government equity, and land or other assets at below market value. One or more of these devices has played an important part in a number of project financing packages. (c) Private equity A very considerable amount of private funds is invested in the vast fleets of private buses, minibuses, para transit vehicles and taxis. However, the individual investments are comparatively small and funded from a variety of formal and informal sources: commercial banks, development banks, unions, various types of entrepreneurs, loan sharks, and family savings. The majority of private operators are very small enterprises with few assets and thus find it difficult to meet criteria for commercial bank loans. Sometimes unions, cooperatives or route associations, have fared a little better in obtaining commercial bank loans for vehicles. 32

37 This is because commercial banks prefer to deal with corporate companies with fixed assets, but these are few and far between in private sector bus operations. Small operators also have difficulties in obtaining suppliers' credits for much the same reasons. A number of governments have attempted to overcome this problem by the provision of funds, including foreign exchange, specifically earmarked for small transport enterprises. However, little of this funding seems to have trickled down to urban bus operators. A notable exception is in Brazil where very substantial government lending, channelled through the National Development Bank, is a major source of funding for private bus companies. In a number of cities, transit systems, in particular bus companies, are funded and owned partly by public authorities and partly by private enterprises. Usually such companies operate Under a contract with a government ministry. The government's obligations are to approve timely fare increases in line with rising costs, and to provide a subsidy to cover concessionary fares and a share of capital costs. On its part, the company usually finances and supplies the rolling stock and other equipment and is required to operate a specified quantity of buses and to meet agreed performance targets. Experience with joint venture bus companies indicates that while superior services may be provided, these are rarely affordable to the majority of users. As a result, there is a strong risk of heavy operating subsidies. The funding of the more capital intensive transit systems, such as metros, LRT, and segregated busways, has in the past been the prerogative of the public sector. This has often been with the assistance of overseas governments prepared to support the purchase of equipment and rolling stock from their own manufacturers. In the expectation that metros would be financially viable, some cities have financed their metros mainly with foreign loans supported with only a small amount of local lending. But with results well below expectations, most metro authorities are currently heavily in debt. In many cases the principal debt is carried or guaranteed by central government (or has had to be taken over by central government because of the appalling damage inflicted on local government finances). Loans for public transport generate little or no financial return or foreign exchange and have contributed most strongly to the developing countries' debt crisis. In recent years there has been a growing interest by governments in obtaining private equity financing for expensive urban transport systems. Some governments have sought to raise private capital to support public sector transit systems against property or land development. Developers, who hold title to the land through which the system passes, have been invited to make a contribution either as a direct payment (before or after the line is built) or by entering into partnership in constructing the system. Alternatively, the local authority can acquire land and use the revenue from its sale and/or development to finance the transit system. It has been found in a number of examples that developers prefer to pay for specific parts of the system such as extensions, connections to stations or road access, rather than contribute to the general overall cost. It has also been found that developers do not want to be urban pioneers; they hesitate about making, a commitment to invest in land around new rapid transit lines until they see construction actually proceeding. Payments from developers could form a useful contribution to the funding of a rapid transit line, but that they are not likely to form a major part of the "up front" capital for the project. One of the most successful ventures has been the development of air space above the stations and depots of the Hong Kong Mass Transit Railway. Taking advantage of the enhanced land values in the vicinity of the railway, the transit corporation has entered into joint ventures with property developers. The developments have been undertaken at the cost and risk of the corporations' joint venture partners, with the corporation taking a share of the profits, generally about 50 per cent. As a result, both commercial and residential developments have been undertaken with proceeds of sale making a significant contribution of about 10 per cent of railway construction costs. In addition, recurring rental income from retained commercial facilities makes a contribution of about 5 per cent to the corporations' revenue. Busways and LRT also lend themselves to land development opportunities of this type. Another equity financing mechanism is through Build Operate and Transfer (BOT) concessions. BOT provides for private enterprises (concessionaires) to design, construct, finance and operate a transit system usually for a period of between 15 and 25 years. At the end of the agreement period, the assets are transferred to the government. To stimulate interest in such projects and to reassure interested bidders of their commitment, governments may bear a share of the costs, or may contribute other assets such as land. The concessionaires are expected to bear most of the risks of construction, including delays, and the commercial risks of operation. In return for providing the systems, the concessionaires are allowed to retain fare box and other income derived by the systems. They are often very interested in any development opportunities, such as building rights over and around stations, mentioned above. The terms suggested by interested BOT bidders reflect the likely viability and risks involved. In particular, BOT enterprises usually press to have full 33

38 control of fare schedules and other tariffs, and for protection from "unfair" competition by other public transport. They also call for protection from foreign exchange and security risks, and other risks, such as delays, that are seen to be beyond their control. BOT concessionaires are usually consortia of bankers, engineers, construction companies, equipment and rolling stock suppliers, public transport operators and possibly others. Because of the many factors involved including legal, political, social, financial, economic, environmental, engineering and management, BOT negotiations and contract preparations can be protracted, sometimes spread over several years. Delays can be due to many issues, such as the contribution and part to be played by government, including the guaranteeing of loans, and the terms for the provision of land. While a number of urban bridges, tunnels and expressways have been constructed as BOT projects, no transit system has yet been completed using BOT, although several are currently in the pipeline. At advanced stages of development are the BOT schemes for metros in Ankara and Bangkok. (d) International and bilateral aid In comparison with other sectors, urban transport has received only a modest amount of international aid. Nevertheless, in the case of World Bank lending, the largest source of international aid, this has amounted to over $US2000 million in the last 20 years. This amount was in support of projects costing in total $US5000 million. Most of this lending is for urban road construction, maintenance and improvements. Bus services, including vehicles, facilities and priority schemes accounted for 12 per cent. Rail systems received only 3 per cent. This lending pattern broadly reflects the sentiments of the World Bank Sector Policy Paper on Urban Transport, which promotes the idea of making existing facilities more effective, building up institutional capabilities and giving the private sector more support. Sadly, little Bank funding has found its way through to travelway schemes; and even those money which have been spent on bus priority measures (a favourite target) have not always proved effective in the longer term. For example, the with flow bus lanes of Tunis are largely disused and the busways of Abidjan are partly abandoned. For aid to be effective, perhaps much more thought needs to be given to promoting a comprehensive package which includes the establishment of an organization, the provision of vehicles and the construction of track. Public transport systems in developing countries receive a considerable amount of bilateral aid, much of which is tied to donor country industrial interests. The LRT systems in Istanbul, Tunis and Manila all received aid packages tied in to European equipment suppliers. Many bus fleets in the public sector are similarly financed, though as with Bank lending, there are no examples of a comprehensive busway package in which both bus track and buses are funded. F. Strategy for development 1. Setting objectives A key question facing any city authority is what son of transport system does it want. To answer this question objectively requires further answers to the wider issues of what type of city form and life style are being promoted, and an understanding of the ways' in which transport investment and development can be used to help achieve these aims. These are difficult questions to answer and few city authorities can provide a clear statement of transport development objectives. Some authorities might even abrogate such issues to the market place, and allow transport to develop in its own way without any outside interference. This, however, is not a practicable approach when city corridors reach the size which demands investment in mass transit; or, indeed, if the un fettered transport system encourages city development in an unforseen and unacceptable way. The planning of urban development is widely accepted and practised throughout me world, and it is sensible to incorporate transport, an important component of urban life, in that process. Aims and objectives for the development of urban transport should be established within the context of the aims for general urban development. These should set out the general requirement of public transport (what level of service is being aimed for) and the guiding principles for meeting this requirement. Targets like travel provision for the urban poor and women, energy conservation, and acceptable levels of environmental nuisance can clearly be specified as part of the objective. 2. Institutional management A body should exist which has a clear responsibility for the development of urban transport planning and development to meet the aims and objectives. This organization must be staffed by a professional cadre of transport planners, analysts, economists and engineers with sufficient financial support to undertake the necessary continuing surveys and analysis (using contemporary software). They must work closely and interactively with urban planners, contributing significantly to the urban development plan. 34

39 3. Project formulation Mass transit developments have to be planned in great detail; they (particularly advanced rail based systems) can consume vast amounts of money, and are prone to capital cost over runs and patronage shortfalls. One of the great virtues of a phased approach to mass transit development is that investment is closely geared to demand trends. The project should be planned in the context of the wider developments in city transport, and the impact that the scheme is likely to have on travel and transport. Thus, there will be a need to consider ancillary developments like the need for integration, traffic management, demand management etc. 4. Fund raising Sources of funding should be identified at an early stage in project development. With trams and LRT schemes, there will be no shortage of deals on offer from consortia of civil engineers and equipment suppliers. Most of these offers will be supported by aid, in some form. While these finance packages from aid agencies and manufacturers of industrialized countries seem readily available for LRT projects, there has been little encouragement for busway transit projects in the past. This may be because city authorities have not been convinced of the worth of busway transit. It is also the case that busway transit does not have a single promoter in the way that LRT has; there is no precedent for bus manufacturers to supply the bus track in the way that LRT suppliers have to. Thus, bus manufacturers are really only interested in selling their vehicles. And yet because the costs of a busway track are relatively low, and a very attractive way of enhancing the performance of an existing bus fleet, busways may appeal to aid agencies which use developmental returns as their main criteria for project appraisal. There is some irony in the fact that bus manufacturers who cannot win aid support (on an aid and trade basis) to sell their buses to cities in developing countries might get more support from their government aid agencies if they developed a total busway transit package. Some consideration might be given at central government level to the establishment of a transport development agency which would have the role of raising finance for transport development and channelling it through to transport operators in both the public and private sectors, and to local government. The money could be on lent by this organization at normal rates, but could provide for loans which commercial banks find too risky. It could also give authorities some additional control over transport investment: for example, there could be conditions on the size of vehicle purchased by an operator. 5. Implementation The implementation of a travelway scheme is a major project requiring coordination between various city institutions (for example: transport operators, police, city engineering office, service providers, and traders) and the contractors and suppliers. For effective and efficient implementation, one authority should represent city interests and have the power to effect necessary temporary regulations (for example, control of road traffic). Consideration must be given at an early stage to the management and operation of a travelway: whether a new organization needs to be created; the operating procedures and rules; the method of integration with other components of the urban transport network. Training is an important component of the project implementation, particularly where new technologies or operational practices are being deployed. The funds, available time and the organisation for this must be put in hand well in advance of project commissioning. 6. Future development Once an initial mass transit scheme has been installed there will be a continuing need to review its performance and appraise ways for its improvement, extension and even replacement. Experience from metro and recent LRT developments suggests that, despite financial setbacks, most cities would want to expand their networks. Much experience will have been gained from an existing system, and there will be many learning costs which can be forgone in future development. It is also the case that a new line will improve the overall performance of the network. Even so, decisions should be based on rigorous technical, financial and economic analysis within the framework of the policy objectives. 7. Role of government 35

40 Government in particular, amongst the many contributors to the well being of public transport, has an all pervasive influence; at both national and local levels, government has wide powers and opportunities to be involved in public transport development. As already indicated, government is likely to be involved in funding arrangements in support of fares revenue: this through direct subsidy and loan, or indirectly through loan guarantees or regulatory controls on operations. But it is never clear just how much government support is either necessary or good for the financial viability of the sector. What perhaps is most important, however, is that government ensures that a strategy is developed and enforced: that objectives are determined; appropriate institutional organization is put in place; projects are properly formulated and appraised; funds are available which can either be repaid by the project or can be shown to be economically worth investing in the project; that the project implementation is well managed. To achieve this end, government might set up a central agency, perhaps using Brazil's EBTU as a role model. BIBLIOGRAPHY General references (applicable to several sections): Armstrong Wright, A., Public Transport in Third World Cities. TRL State of the Art Review 10 (London, HMSO, 1993). Asian Development Bank, Review of the scope for Bank assistance to urban transport (Manila, Infrastructure Department, Asian Development Bank, 1989). Bushell, C., and P. Stonham (eds.), Jane's Urban Transport Systems (London, Jane's Information Group, 1991). Economic Commission for Africa, Urban Transport: a Baseline Assessment and Strategy for the Second United Nations Transport and Communications Decade for Africa (Addis Ababa, 1991). Institution of Civil Engineers, Moving People in Tomorrow's World; Proceedings of Conference organised by ICE, October, 1986 (London, Thomas Telford, 1987). PTRC, Urban Transport in Developing Countries (London, PTRC, 1991). Rogers, L.H.(ed.), International Statistical Handbook of Public Transport (Brussels, UITP, 1986). Thomson, J.M., Towards Better Urban Transport Planning in Developing Countries. World Bank Staff Working Paper No. 600 (Washington, D.C., The World Bank, 1983). World Bank, Urban Transport: A World Bank Policy Study (Washington, D.C., The World Bank, 1986). References for section B Barrett, R., Urban Transport in West Africa, World Bank Technical Paper No. 81 (Washington, D.C., The World Bank, 1988). Barrett, R., Nigeria: Urban Transport in Crisis (Washington, D.C., The World Bank, 1991). Fouracre, P.R., and J. Turner, "Travel characteristics in developing cities". Paper presented at the Sixth World Conference on Transport Research, Lyon, July Fouracre, P.R., D.A.C. Maunder and G.A. Banjo, "Urban public transport as a function of city size; the case of Nigeria", in: New Perceptions and New Policies. Urban Transport in Developing Countries (Paris, CODATU, 1988). Godard, X., "Politiques de transport urbain en Afrique sub Saharienne. Synthèse de six études de case", Series: Transport, Transfert, Developpement (Paris, CODATU INRETS, 1991). Government of India, Report of the Study Group on Alternative Systems of Urban Transport (New Delhi, 1987). 36

41 United Nations, Patterns of Urban and Rural Population Growth (New York, 1980). United Nations, Estimates and Projections of Urban, Rural and City Populations, (New York, 1985). World Bank, Urban Policy and Economic Development: a World Bank Policy Study (Washington, D.C., The World Bank, 1991). World Bank, World Bank Development Report 1991 (Washington, D.C., The World Bank, 1991). References for section C Allport, R.J., "Appropriate mass transit for developing countries", Transport Reviews, vol. 6, No. 4, pp Armstrong Wright, A.T., Urban Transit Systems Guidelines for Examining Options, World Bank Technical Paper No. 52 (Washington, D.C., The World Bank, 1986). Armstrong Wright, A., and S. Thiriez, Bus Services: Reducing Costs, Raising Standards, World Bank Technical Paper No. 68 (Washington, D.C., The World Bank, 1987). Eastman, C., "Improvements to the street tram system in Cairo", Proceedings of the PTRC Transport and Planning Summer Annual Meeting (London, PTRC, 1987). Fouracre, P.R., and G. Gardner, "Mass transit in developing cities: the role of high performance bus systems", Proceedings of the Institution of Mechanical Engineers Conference: The Expanding Role of Buses Towards the 21st Century, March, 1992 (London, IME, 1992). Gardner, G., P.R. Cornwell and J.A. Cracknell, The Performance of Busway Transit in Developing Countries, TRL Research Report, RR 329 (Crowthorne, Transport Research Laboratory, 1991). Gardner, G., "The performance of light rapid transit in developing countries", TRL Project Report (Crowthorne, Transport Research Laboratory, in preparation). Gray, P., "Public transport planning for Karachi", Proceedings of the PTRC Transport and Planning Summer Annual Meeting (London, PTRC, 1990). JMP Consultants Ltd/Chulalongkorn University, Bangkok Traffic Research Project. TRRL Contractor Report. CR 84 (Crowthorne, Transport Research Laboratory, 1988). Lindau, L.A., "Bus priority systems in Brazil: from theory and practice", Proceedings of the PTRC Transport and Planning Summer Annual Meeting (London, PTRC, 1987). Marler, N.W., The Performance of High flow Bus Lanes in Bangkok. TRRL Supplementary Report, SR 723 (Crowthorne, Transport Research Laboratory, 1982). NATO, "Bus priority systems", CCMS Report No. 45; NATO Committee on the Challenges of Modern Society (Crowthorne, Transport Research Laboratory, 1976). Simpson, B.J., Urban Rail Transit An Appraisal. TRL Contractor Report CR 140 (Crowthorne, Transport Research Laboratory, 1989). TRL, Busway Transit Video, TRL Video V254 (Crowthorne, Transport Research Laboratory, 1991). Vuchic, V., Urban Public Transport Systems and Technology (New Jersey, Prentice Hall Inc., 1981). References for section D Belda, R., "Les raisons du success du Metro de São Paulo", Proceedings of the Institution of Civil Engineers Conference: Rail Mass Transit for Developing Countries, October, 1989 (London, Thomas Telford, 1990), pp

42 Benitez, B.N., and O. Gonzalez Gomez, "The Mexican Experience", Proceedings of the Institution of Civil Engineers Conference: Rail Mass Transit for Developing Countries, October, 1989 (London, Thomas Telford, 1990), pp Dalvi, M.Q., "Calcutta Metro", Proceedings of the Institution of Civil Engineers Conference: Rail Mass Transit for Developing Countries, October, 1989 (London, Thomas Telford, 1990), pp Dans, J.P., "The Metro Manila LRT System, its future", Proceedings of the Institution of Civil Engineers Conference: Rail Mass Transit for Developing Countries, October, 1989 (London, Thomas Telford, 1990), pp Fouracre, P.R., R.J. Allport, and J.M. Thomson, The Performance and Impact of Rail Mass Transit in Developing Countries, TRRL Research Report RR278, (Crowthorne, Transport Research Laboratory, 1990). INRETS, "Evaluation des metros dans villes latino americaines", Rapport pour le Ministère de la Recherche (Paris, INRETS, 1987). Marchand, L., J. Veinberg and F. Guittonneau, "Rail based urban transit systems", in French Railway Review, vol. 1 (1983), No. 5, pp Sperling, D., "Caracas metro a luxury?", in Transportation Research Record (Washington, D.C., Transportation Research Board, 1981), No. 797, pp Walmsley, D.A. and K.E. Perrett, The Effects of Rapid Transit on Public Transport and Development, TRL State of the Art Review No. 6. (London, HMSO, 1992). Walmsley, D.A., Light Rail Accidents in Europe and North America, TRL Research Report RR 335, (Crowthorne, Transport Research Laboratory, 1992). References for section E Allport, R.J., "Transport concessions study Thailand, a case study", OECD Conference. Seville 1991 (Paris, OECD, 1991). Augenblick, M. and B.S. Custer, The Build, Operate and Transfer (BOT) Approach to Infrastructure Projects in Developing Countries, World Bank Working Paper (Washington, D.C., The World Bank, 1990). EBTU, Manual Técnico Tratamento preferêncial ao transporte colectivo por ônibus (Brasilia, Empresa Brasileira dos Transportes Urbanos, 1982). Gutman, J.S., "Financing urban transport: constraints and options for developing countries", Proceedings of the International Union of Public Transport Conference, Singapore 1988 (Brussels, UITP, 1988). Henry, E., "L'entreprise bresilienne d'autobus urbains: un modele?", in: Recherche Transports Securité (Paris, INRETS, 1991), No. 31, pp Lethbridge, N., "Private financing of public transport infrastructure", Proceedings of the PTRC Transport and Planning Summer Annual Meeting (London, PTRC, 1990). Mandon, B., "Bilan d'une decennie de politiques de transports a Dakar", in: Recherche Transports Securité (Paris, INRETS, 1991), No. 36, pp Menckhoff, G., "Technological and ownership options of public transport in India", paper presented in Seminar on Metropolitan Transport in India, organized by Times Foundation of India, New Delhi, Mitric, S., Crisis and Recovery: Urban Public Transport in Morocco, Transportation Research Record 1297 (Washington, D.C., Transportation Research Board, 1991). RATP, Autobus en site propre (Paris, Régie Autonome des Transports Parisiens, 1977). Simpson, B.J., Urban Rail Transit: Costs and Funding, TRL Contractor Report CR 160 (Crowthorne, Transport Research Laboratory, 1990). 38

43 Situma, L., "Problems of public transport operations and planning in Zimbabwe", Developing World Land Transport (London, Grosvenor Press International, 1988), pp TRL, Design Guidelines for Busway Transit, Overseas Road Note 11 (Crowthorne, Transport Research Laboratory, 1993). Walmsley, D.A., and M.W. Pickett, The Costs and Patronage of Rapid Transit Systems Compared with Forecasts, TRL Report RR 358 (Crowthorne, Transport Research Laboratory, 1992). Won, J., "Bus cooperative systems in Korean cities", Transportation Quarterly, vol. 40, (1986), No. 2, pp ENDNOTES 1. It may be argued that the apparent advantage of rail transit over conventional bus systems is largely due to differences in the degree of segregation rather than to inherent technological differences. II. CASE STUDY: BUSWAY IN ANKARA AN INTERMEDIATE, LOW COST ACTION TO IMPROVE PUBLIC TRANSPORT 1 A. Urban transport in Ankara 1. Population growth and transport demand Ankara the capital, although not the prime city of Turkey has experienced a population growth which raised the number of inhabitants from less than 100,000 in 1930 to more than 2.5 million in 1990, increasing thereby the share of the city in the country population from slightly more than 0.5 per cent to almost 4.5 per cent (see table 1). The annual average growth rate of the population 2.5 per cent in the years may seem low in comparison with that observed in cities of developing countries, yet it is high enough to bring about considerable increases in the demand for passenger transport. Apart from the population growth, the changes in the pattern of physical development and in the people's mode of travel the latter from trips made by common carriers to trips by individual transport modes also contribute to the city problem in transport. 2 Years Population of Turkey (thousands) Table 1. Population growth in Turkey and Ankara Percentage of urban population Population of Ankara (thousands) Share of Ankara in total population Source: SIS, Population Census Statistics There were above 3.8 million person trips made in Ankara in 1985, and the average trip generation rate was 1.72 trips/person, out of which 1.16 trips/person were made by motorized transport. Trips by common carriers amounted to 54 per cent of all person trips and to almost 80 per cent of motorized trips (EGO, 1992). It is estimated that in the years , the city population increased by 13 per cent and there was a similar increase in the number of trips made by common carriers. However, trips by individual transport modes grew by as much as 30 per cent and expanded their share in motorized trips by about 2 per cent. The modal split in travel varies among particular parts of the city. In trips to the city centre, the share of trips by common carriers is higher than the city average, while walk trips prevail within residential areas. 39

44 Table 2 presents the modal split of passenger trips made by motorized transport in 1990 and the number of vehicles serving those trips. The concentration of motorized trips in the peak hours is higher than usually observed in cities of developing countries. About 26 per cent of such trips are made between 6 a.m. and 9 a.m., and 20 per cent between 5 p.m. and 7 p.m. This, of course, adds to difficulties in meeting the transport demand. Table 3 presents the structure of trips by trip purposes as it was established in the year It should be noted that trips in other purposes than work and education have a large share in trips by common carriers, meaning that people can afford the cost of such non obligatory trips and common carriers provide needed services. 2. Supply of transport services Passenger transport is provided by both public and private operators. It is dominated by road based transport modes. The public sector is represented by bus services offered by the Bus Department of the General Directorate of Electricity, Gas and Bus Operation (EGO) an agency of the Municipality of Ankara Greater City and also by suburban railways which, however, play a marginal role in urban transport. Faced with the shortage of funds, the public sector has been unable to supply services adequate to the growing transport demand and, in consequence, its share in the services provided by common carriers is shrinking. It is, nevertheless, the public sector that serves not only profitable routes, but also distant areas of low development density, and offers transport services all over the day on a regular basis. Private operators prefer routes where the demand is high and do not care much about the frequency of services in off peak hours. Private operators are very flexible in adjusting the supply of services to effective demand. Making a satisfactory profit, they have a considerable potential to expand the scope of services but are constrained by the regulations binding the public transport industry. Mode of travel Table 2. Modal distribution of motorized passenger trips Number of vehicles Number of passengers Percentage of passengers Municipal bus Private bus Minibus Service bus Suburban train Common carriers in total Taxi Official and private cars Other individual modes Individual transport in total Grand total Source: EGO, 1992 (a) Municipal buses Municipal buses only 973 in number by the year 1990, operate on as many as 226 routes of which 31 are express service lines. The frequency of service is low and, thus, municipal buses have only about a 37 per cent share in trips made by common carriers. The increases in the EGO'S bus fleet, related to the increases in urban population, are shown in figure 1. This figure indicates that not only is the number of buses in the EGO inventory low, but also that a large number of buses are out of service. The fleet is old and the average bus age reaches 11.4 years per cent of buses have completed the seven year period of economic operation and require considerable maintenance expenses to remain usable. The EGO attempts to replace the worn out buses and as pan of this effort made also with the aim of reducing air pollution in Ankara new articulated buses which operate on natural gas have been put into service in Table 3. Distribution of trips by purposes (percentage) Mode of travel Purpose of travel 40

45 Work School Other purposes All purposes (N) Travels by common carriers Motorized travels in total Total travel, including walking Source: EGO, 1992 (b) Private bus services In 1980, the inadequate supply of transport services by the public sector led to opening of the market to the operation of private buses. The original policy pursued in the regulation of private bus operation was that these buses should provide services complementary to those of the EGO buses, rather that as rivals. Private buses were expected to serve the passenger demand originating mainly in distant residential areas and, in the first arrangement, were assigned to routes characterized by relatively low demand. Figure 1: Annual Growth of EGO Buses Supply (Per people) In view of the lack of positive response on the part of potential private operators to such a policy, certain changes had to be made in the regulations. Through a series of concessions made by the public authority, privately operated bus services have been established with 200 buses operating on 18 routes. The fleet of private buses consists, however, of models of buses from as early as 1971 and the average age of the fleet was 9.2 years already in 1985 (EGO, 1987 and 1992). Nevertheless, private bus operators have an almost 8 per cent share in trips served by common carriers. (c) Minibus (dolmus) services Dolmus a sort of paratransit service has been operating in Ankara for 40 years. In 1990 this type of services was provided by 1884 minibuses owned by small private operators which serve 31 fixed routes without observing any specified time schedule. It is believed that minibuses offer much more frequent services and are more rapid and comfortable than buses. They have also the advantage of running on steep roads with a small radius of road curvature. Thus, they present a suitable mode of transport for the tough topographic conditions of Ankara. However, owing to the low capacity of a minibus, they add to traffic congestion on certain routes and are neither energy efficient nor considered suitable from the environmental point of view. The minibuses compete successfully with EGO and private buses. They enjoy an abundance of passengers, compared with their capacity, although they operate generally along the same routes as buses. In 1990, they had the same share of the passenger market as municipal buses (EGO, 1987 and 1992). 41

46 (d) Service buses Government institutions are given the right to provide their staff with transport services by their own buses with the permission of the Municipality. Service buses operate only a few hours a day, thus, their capacity is little utilized. Through a decree issued in 1979 these buses were handed over to the EGO, with the exception of buses belonging to military institutions, universities and some other institutions. This allowed the EGO to start the operation of express bus services. However, as the transport conditions did not improve considerably, the governmental institutions bought or hired buses from private companies and doubled their fleet in a couple of years. The size of the bus fleet owned by these institutions is not exactly known. A survey conducted in July 1985 put their number at about 2000 in Ankara. This number increased to 3200 in the year 1990, so that it exceeded three times the number of EGO buses, but service buses transported only 16 per cent of passengers using common carriers (EGO, 1987 and 1992). (e) Taxi services Figure 2. Urban transportation investments decision making process Although not considered as common carriers throughout this paper, taxi services play an important role in urban transport in Ankara. There are about 8000 taxis licensed within the boundary of the Ankara Municipality. They carry approximately 152,500 passengers a day. Their share in meeting the daily demand for motorized passenger transport in Ankara was 13.7 per cent in 1970 and 19.2 per cent in After 1976, the share of taxis declined, amounting to 10 per cent in 1980, 6.6 per cent in 1985 and about 5.0 per cent in 1990 (EGO, 1987 and 1992). This decline was mainly attributed to the growth of private car ownership. 3. Institutional set up in urban transport management The responsibilities for the development, maintenance and operation of transport in Ankara are split among many authorities at national, provincial, municipal and local levels. This can be seen in figure 2, which depicts the decision making process concerning investments in transport infrastructure, and in table 4 which gives information on the division of responsibilities in the management of transport system operations and in traffic control. 42

47 The central role in all the above mentioned aspects is played by the organs of the Municipality of Ankara Greater City. Local governments at the district level have a responsibility limited to the construction and maintenance of roads with less than three traffic lanes. They do that within a budgetary allocation approved by the metropolitan Municipality. They are not entrusted with the provision and management of public transport. On the other hand, the development and management of transport system components of national or provincial importance are in the hands of the Central Government. Action being taken by this level of government has to be incorporated into Ankara transport development and management plans. The latter are, moreover, subjected to approval by the central authorities and, for example, the EGO investment plans and specific development proposals need approval by the Ministry of Energy. Table 4. Urban Transportation Systems Operational Aspects Traffic Management Control Transportation Systems EGO Bus Private Bus Service Vehicles Minibus (Dolmus) Taxi Private Public Involving Institutions at System Establisment Stage Parliament Ministry of Internal Affairs G.C. Municipality Assembly UKOME UKOME G.C. Municipality Assembly Governor (Optional) Prov. Traffic Com. (Optional) Decision of related institution Ministry of Transportation G.C. Municipality Assembly The Public Organization UKOME G.C. Municipality Assembly Governor (Optional) Prov. Traffic Com. (Optional) UKOME Governor (Optional) Operating Routes UKOME EGO UKOME Urban Trans. Com. (Advisory) Prov. Traffic Com. Governor Prov. Traffic Com. G.C. Municipality Assembly UKOME not applicable Fare Levels UKOME EGO UKOME Prov. Association (Advisory) not applicable Vehicle Conditions Traffic Department G.C. Municipality Assembly EGO Traffic Department G.C. Municipality Assembly EGO Traffic Department G.C. Municipality Assembly Traffic Control/Enforcement Level of Service Provided EGO G.C. Municipality Police not applicable Traffic Traffic Departmen Traffic Departmen Traffic Departmen not applicable not applicable not applicable Traffic Departmen UKOME Prov. Association (Advisory) UKOME Prov. Association (Advisory) Traffic Department G.C. Municipality Assembly G.C. Municipality Police G.C. Municipality Police Private Car not applicable not applicable not applicable not applicable 43 Traffic Departmen Traffic Departmen

48 Sub urban Rail Trucks (Cargo) Buses of Neighbouring Municipalities Traffic Department Inter urban Transport Highway Railway Airway UKOME Ministry of Transportation TCDD Individual or Institutional decision Neighbouring Municipality Ministry of Transportation TCDD Ministry of Transportation THY Ministry of Transportation UKOME Ministry of Transportation TCDD Governor Prov. Traffic Com. Regional Traffic Dept. UKOME Governor Prov. Traffic Com. Gen. Directorate of Highways TCDD THY Traffic Department Traffic Department UKOME TCDD TCDD TCDD Determined by market forces Neighbouring Municipality Governor TCDD Ministry of Transportation THY Ministry of Transportation Traffic Department Traffic Department Neighbouring Municipality Ministry of Transportation Ministry of Transportation Ministry of Transportation G.C. Municipality Police G.C. Municipality Police Neighbouring Municipality Ministry of Transportation Ministry of Transportation Ministry of Transportation Plans and projects for the development of urban transport come from different departments and agencies of the Municipality of Ankara Greater City and from the private sector. UKOME the transport coordinating centre established at the metropolitan level coordinates these plans and projects and exercises the following duties: (a) Approves the overall transport plan and investment proposals, the latter being subject to further evaluation by metropolitan organs; (b) Drafts local regulations concerning urban transport for the approval by relevant authorities; (c) Guides the implementation and monitoring of traffic management schemes; (d) Approves service plans and operating arrangements for passenger transport; (e) Sets timetables and fare levels for passenger common carriers; (f) Advises on the impact of the overall urban transport plan; (g) Advises public organizations on staggering working hours with the aim of easing transport problems. The implementation of UKOME recommendations concerning investments in transport infrastructure is in the hands of AYKOME a municipal department in charge of coordination and implementation of all infrastructure development within metropolitan boundary, including central government projects, and an administrator of the Infrastructure Investment Fund. Traffic control and its enforcement are the responsibilities of the Traffic Department and Municipal Police, supervised by the Provincial Transport Committee. Permits and licenses to operate transport services are issued by appropriate organs of the Municipality of Ankara Greater City, following the guidelines established by UKOME. Traffic Department Traffic Department Traffic Department Ministry of Transportation Ministry of Transportation B. The development and performance of a busway in Ankara 1. Location of the busway Ankara was the first municipality in Turkey to develop a busway as an intermediate, low cost means of improving mass transport in the city central area before a rail transit system could be built. This example was 44

49 then followed by the provision of exclusive bus lanes in other cities, including a busway in Istanbul. There are two main transport corridors in the city of Ankara. One of them runs in a north south direction between Aydinlikevler and Çankaya and links two important business centres at Ulus and Kizilay. The second, on which the busway was constructed in 1978, is situated in an east west direction and connects the districts of Dikimevi and Besevler (see figure 3). Two business centres Kizilay and Maltepe have developed along this corridor, apart from universities, high schools, government offices, entertainment centres and a huge number of shopping and banking facilities. Large residential areas are also adjacent to this corridor. Kizilay, where the two corridors intersect, is me most important business centre in the city. The busway was inserted in the middle of a dual carriageway, a six traffic lane artery divided by a median of 1 m minimum width. Contrary to prevailing experience in the development of busways, it serves traffic in the city inner area and not between such an area and distant suburbs. 2. Background of the project Under mixed traffic conditions, the services provided by municipal buses along the artery linking Besevler and Dikimevi districts were characterized by low speed and irregularity. Buses were losing patronage not only to private cars but, first of all, to dolmus services provided since 1961 by vehicles of me station wagon type. Dolmus offered a frequency of service, and a level of comfort much higher than that on buses, but at lower fares. They, however, contributed a great deal to traffic congestion, not only by their number, but also by me way they were operated. Figure 3: Location of the Besevler Dikimevi exclusive busway 45

50 In 1970 a transport study conducted by SOFRETU (Société Française d'etudes et de Réalisation de Transport Urbans) proposed the development of a light rail transit system for the Besevler Dikimevi corridor. The traffic survey conducted by SOFRETU reported that 7335 passengers were passing the survey point at Demirtepe in the direction to Kizilay during the morning peak hour. The total number of transit passengers and vehicles recorded in both directions, as well as the modal split in passenger traffic, are presented in tables 5 and 6 below. Table 5. Results of the SOFRETU survey on the number of EGO buses and passengers carried (Survey date: March 1970) Survey points To Kizilay 7.45 a.m a.m. From Kizilay 5.30 p.m p.m. Buses Passengers Buses Passengers Kolej Demirtepe Table 6. Results of the SOFRETU survey on the number of dolmus and passengers carried (Survey date: March 1970) Survey points To Kizilay 7.45 a.m a.m. From Kizilay 5.30 p.m p.m. Dolmus Passengers Dolmus Passengers Kolej Demirtepe As the development of an LRT system required overcoming many difficulties, the construction of a busway was proposed in the late 1970s as an intermediate, low cost solution. The main objectives of the project were the following: (a) Releasing the bus from the traffic congestion created by cars; (b) Increasing the number of bus passengers and hence decreasing the social and economic costs of transport; (c) Giving better service to those users with. fixed incomes who are dependent on bus transport. The EGO expected that with the development of me busway the number of bus passengers would double and 25 per cent of passengers using private cars and dolmus services would be transferred to municipal buses. 3. Busway design The busway extends over a distance of 5.5 km and is 7 m wide. Designed as a median busway, it has one traffic lane in each direction with no provision made for overtaking. Lanes devoted to buses are separated from those serving general traffic (two lanes in each direction) by raised islands at the bus stops, and 1.5 high fences along the mid block sections. There are 13 bus stops along the busway at an average distance of about 450 m, but they are unevenly distributed, depending on the location of the intersections with other roads, so that the distances between bus stops vary from 240 m to 760 m. Bus stops are located about m before traffic stops lines at intersections, and are adjacent to pedestrian crossings. In order to minimize the effect of the iron fences on the cross movement of pedestrians between the two sides of the road, at grade crossings and overpasses have been provided. Two overpasses have been constructed at Izmir and Selenic Avenues which were transformed into a pedestrian zone. All intersections are controlled by traffic signals, although police intervention is often required at Kizilay and some other places during rush hours. Efforts were made to coordinate traffic signals and provide for a "green wave" along the busway, but without much success. Traffic signals systems prevent, however, turn movement collisions between buses and other vehicles with a priority given to the movement of buses. 46

51 There is no access from side roads to the main road between main intersections. All these side roads are one way off the main road. 4. Implementation of the project The busway project was implemented in two stages. In the first stage a section Dikimevi Kizilay was constructed within a period of 90 days. In another 60 days the section Kizilay Besevler was completed. The works were carried out by the Construction Department of Ankara Municipality and financed from the municipal budget. The costs of project implementation are shown in table 7. Cost components Cost in TL Design and labour costs 19,000,000 Fences 11,000,000 Resurfacing of the road 5,309,206 Traffic signals 831,200 Total 36,140,406 In October 1978, $US 1 = TL (a) Traffic flows Table 7. Costs of busway construction 5. Busway performance The provision of the busway brought about an evident increase in the number of passengers carried by EGO buses along the corridor. A traffic survey conducted in May 1981 showed that these buses were used by the following numbers of passengers during the morning peak hour between 7.30 am and 8.30 am. Dikimevi Kizilay direction: 5796 passengers Kizilay Tandogan direction: 5828 passengers Tandogan Kizilay direction: 4719 passengers Kizilay Dikimevi direction: 2602 passengers A better service provided by EGO buses, on me one hand, and the growing congestion on me carriageways parallel to me busway, on the other hand, led to the abolishment of station wagon dolmus services in In a further development, the movement of minibus dolmus was also restricted on the section of me corridor served by the busway. By 1989 minibus dolmus operated up to Tandogan and Kolej, but not on the central section of the artery, in spite of the inadequate capacity of EGO buses. In tables 8 and 9, the results of a traffic survey conducted on the corridor during the peak hours of a working day in May 1991 are presented. The survey revealed further increases, in the number of passengers travelling by EGO and private buses. Those buses had already become overcrowded, in particular, at the morning peak hours on the section between Iccebeci and Kizilay where the average bus occupancy far exceeded 100 passengers. This made it impossible for many passengers to board the bus or increased bus boarding time significantly. While the number of buses operating on the busway was still relatively low, bus stop operation had become chaotic, thereby affecting the flow of buses and increasing their waiting time at bus stops. Traffic congestion was also getting high on the traffic lanes designated to general traffic on the corridor. (b) Bus travel speed After the opening of busway operation, time savings ranging from 5 to 10 minutes per bus trip were noted in the surveys conducted on EGO buses travelling along the corridor. The comparative results of a "before and after" survey are given in the table 10. It should be noted that the data in table 10 refer to routes over which the busway was only a part; nevertheless, it was the provision of the busway on the most congested section that brought about the shortening of travel times. 47

52 In May 1991, a survey was carried out to establish bus travel time on the busway at the morning and evening peak hours. The survey provided information on bus waiting times at particular bus stops incurred by both traffic signals and the boarding and alighting of passengers. It allowed the establishing of the journey time between successive bus stops. Tables 11 and 12 contain the results of the survey, including the observed minimum and maximum values of measured parameters as well as average values calculated for a series of bus runs. Table 8. Number of vehicles and passengers recorded on the busway at the morning peak hour 7:30 a.m. 8:30 a.m. A: Number of vehicles Sections EGO and private buses Taxi Private car Others Minibus dolmus Total Iccebeci Hukuk Kurtulus Kolej Kolej Mithatpasa Mithatpasa Kizilay Kizilay Demirtepe Maltepe Sanayi Bak Tandogan Besevler Besevler Tandogan Sanayi Bak. Maltepe Demirtepe Kizilay Kizilay Mithatpasa Mithatpasa Kolej Kolej Kurtus Hukuk Iccebeci B: Number of passengers Iccebeci Hukuk Kurtulus Kolej Kolej Mithatpasa Mithatpasa Kizilay Kizilay Demirtepe Maltepe Sanayi Bak Tandogan Besevler Besevler Tandogan Sanayi Bak. Maltepe Demirtepe Kizilay Kizilay Mathatpasa Mathatpasa Kolej Kolej Kurtulus Hukuk Iccebeci Table 9. Number of vehicles and passengers recorded on the busway at the evening peak hour 5:30 p.m. 6:30 p.m. A: Number of vehicles Sections EGO and private buses Taxi Private car Others Minibus dolmus Total Iccebeci Hukuk Kurtulus Kolej Kolej Mithatpasa Mithatpasa Kizilay

53 Kizilay Demirtepe Maltepe Sanayi Bak Tandogan Besevler Besevler Tandogan Sanayi Bak. Maltepe Demirtepe Kizilay Kizilay Mithatpasa Mithatpasa Kolej Kolej Kurtus Hukuk Iccebeci B: Number of passengers Iccebeci Hukuk Kurtulus Kolej Kolej Mithatpasa Mithatpasa Kizilay Kizilay Demirtepe Maltepe Sanayi Bak Tandogan Besevler Besevler Tandogan Sanayi Bak. Maltepe Demirtepe Kizilay Kizilay Mathatpasa Mathatpasa Kolej Kolej Kurtulus Hukuk Iccebeci Table 10. Results of a "before and after" study of EGO buses travel time along the busway corridor Survey date Average travel time at various hours of the day (minutes)

54 Bahcelievler Dikimevi route , Monday, A , Monday, B , Monday, C , Tuesday, A , Tuesday, B , Tuesday, C Emek Dikimevi route , Monday, A , Monday, B , Monday, C , Tuesday, A , Tuesday, B , Tuesday, C Source: EGO, 1992 A: Before opening of the busway B: After opening of the busway (according to bus service timetable) C: After opening of the busway (according to survey records) The average travel time for end to end travel on the busway was about 22 minutes for each direction, meaning that the average travel speed was about 14.5 km/h. However, different proportions of bus waiting time and bus running time were revealed with respect to the direction of traffic. In the Besevler Dikimevi direction, average waiting times at bus stops amounted to only 29 per cent of the travel time, while in the opposite direction they represented as much as 48 per cent. The explanation for this difference is that in the first case many overcrowded buses skipped me bus stops alighting passengers on the mid block sections (note, the minimum waiting times of 5 sec. only recorded at several bus stops), thereby increasing journey time between bus stops, while saving time at bus stops themselves. The tables point to Kizilay bus stop as critical to busway operation. Average waiting time at that bus stop was much longer than at any other and the above described bus operation practice made journey time between Kizilay and Mithatpasa bus stops long. Table 11. Recorded waiting times of EGO buses at bus stops and elapsed journey times between successive bus stops (Dikimevi Besevler direction, May 1991) Bus stop No. Bus stop Distance between bus stops (metres) Wailing time at bus stop (seconds) 3 Dikimevi Journey time between successive bus stops (seconds) Min Av. Max Min Av Max Iccebeci Hukuk Kurtulus Kolej Mithatpasa Kizilay

55 Demirtepe Nokta Maltepe Sanayi Bak Tandogan Besevler Total busway In tables 13 and 14 average sectional and mid block speeds are presented as they were calculated on the basis of average waiting and running times. Sectional speed is related to the time which elapses between the arrivals of buses at two successive bus stops, i.e., the time which includes both the waiting time at the bus stop and the running time on the following mid block section. Mid block speed is related to time spent on the mid block section only. The average sectional speed was about 16 km/h in both directions, while mid block speed amounted to 23 km/h in the Besevler Dikimevi direction, and to 29 km/h in the opposite direction. Again the explanation for this difference is largely the discharging of passengers in mid block sections. It was, therefore, concluded that there was an insufficient supply of buses, and their overcrowding significantly hampered the operation of the busway, with the chaotic transfer of passengers at the bus stops being another cause of delays in the operation of buses. Table 12. Observed waiting times of EGO buses at bus stops and elapsed journey time between successive bus stops (Besevler Dikimevi direction, May 1991) Bus stop number Bus stop name Distance between bus stops (metres) Waiting time at bus stop (seconds) 15 Besevler Journey time between successive bus stops (seconds) Min Av. Max Mm Av. Max Tandogan Sanayi Bak Maltepe Nokta Demirtepe Kizilay Mithatpasa Kolej Kurtulus

56 Hukuk Iccebeci Dikimevi Total busway Table 13. Most likely average sectional and mid block speeds of EGO buses (Dikimevi Besevler direction, May 1991) Bus stop No. Bus stop name 3 Dikimevi 4 Iccebeci 5 Hukuk 6 Kurtulus 7 Kolej 8 Mithatpasa 9 Kizilay 10 Demirtepe 11 Nokta 12 Maltepe 13 Sanayi Bak 14 Tandogan 15 Besevler Distance between bus stops Average sectional speed (km/h) Average mid block speed (km/h) Average Economic evaluation of the busway project There is no information available on any economic evaluation of the Ankara busway made by or for the public authorities. A partial cost/benefit analysis was made by one of the contributors to this paper (Birgonul, 1981). This analysis was made for the 15 year period of expected busway operation and accepted a 14 per cent discount rate for calculations. As the capital costs of the busway those presented in table 7 above were included plus the costs of resurfacing the busway every five years and renovating the fences. As running costs, the city average bus operation and maintenance costs per bus/km were accepted. As benefits, the fare revenues and time savings for passengers were considered. Such elements as environmental impact or comfort in travel were beyond the analysis. Also the impact of the busway on the condition of general traffic in the corridor was not evaluated. The analysis showed that the operation of buses would not be financially viable for the operator and that the calculated economic benefits would riot equal the costs of busway provision and operation. However, the analysis was a partial one and had only limited information value. Its results were, first of all, affected by a fare level divorced from the real costs of services. 52

57 Table 14. Most likely average sectional and mid block speeds of EGO buses (Besevler Dikimevi direction, May 1991) Bus stop No. Bus stop name 15 Besevler 14 Tandogan 13 Sanayi Bak 12 Maltepe 11 Nokta 10 Demirtepe 9 Kizilay 8 Mithatpasa 7 Kolej 6 Kurtulus 5 Hukuk 4 Iccebeci 3 Dikimevi Distance between bus stops Average sectional speed (km/h) Average mid block speed (km/h) Average C. Lessons to be learned 1. Response to transport demand The busway was provided at a time when the bus flow and passenger turnover were both at the lowest limit to justify the segregation of bus transport from general traffic. Its construction resulted from the need to improve the efficiency in operation of the EGO buses, the number of which was far below the needs of Ankara thus giving way to an intense operation of station wagon dolmus services. These services, based on the "hail and ride" manner of operation of a large number of vehicles, contributed considerably to traffic congestion and also affected the operation of buses. Providing buses with a separated travelway allowed significant increases in bus speed and raised the competitiveness of municipal buses with regard to other modes of transport. As the number of cars travelling along the artery was not yet very high, the reallocation of a part of available travelway space from general traffic to exclusive use of buses, perhaps benefited both public and individual transport. It should be noted that the provision of the busway was considered as an intermediate solution before a rail transit system could be built on the corridor. However, as congestion in general traffic increased, while the services provided by buses operating on the busway improved, the decision was taken to abolish station wagon dolmus services. This, and the overall growth of demand for mass transport, brought about increases in the passenger flow along the corridor, 53

58 unhappily not accompanied by an adequate increase in the number of buses. Overcrowding of buses and the unorganized operation of bus stops made the performance of buses on busway deteriorate; waiting times at bus stops increased, platoons of buses started to appear in mid block sections and the speed of travel decreased to a level much below that recorded on busways worldwide. Complementary arrangements in managing the operation of busway and its redesign became necessary. They are not being undertaken as the Municipality has already embarked on the project to replace the busway by an LRT scheme. In summary, the busway fulfilled its role as an intermediate solution to a transport problem but, owing to shortcomings in its design and operation, it was far from reaching the potential effects a busway might produce. The increase in the passenger flow on buses from about 5000 in 1981 to 8700 in 1990 was, perhaps, the main achievement. 2. Location of the busway The busway in Ankara was developed on a section of an artery passing through the urban central area, while usually busways are built to link suburban areas to city centres on corridors where the increases in the mid block speed of a uniform flow of buses following mainly similar routes can bring most of the desired effects. In the case of Ankara, the location of the busway motivated by the need to ease congestion has had the consequences of short bus stop distances inherent to city centre conditions, changes in the number of buses passing particular sections of the busway (as numerous buses turned into or from intersecting roads), and large numbers of passengers boarding and alighting at each bus stop. It made the bus stop operation crucial for the performance of the busway, while mere was not much to gain from the improvements of the mid block operating conditions. Unhappily the "on line" bus stops arrangement and the lack of effort to organize a bus flow in convoys (which was difficult owing to the bus routes' configuration) considerably limited the effectiveness of the busway. The provision of the busway as a temporary solution allowed, however, the creation of a public transport right of way in a very sensitive area which can now be used for me development of LRT. 3. Busway design Conceived as a temporary solution, the busway in Ankara had to be a low cost solution, thus, the scope of civil engineering works was kept at a minimum level. Moreover, the location of the busway in the urban central area almost excluded the changes in the existing right of way along the artery. For these reasons, the busway design provided for "on line" bus stops without an opportunity for bus overtaking and without constructing bus loading bays. Bus stop passenger transfer capacity soon became the main problem in busway operation, aggravated by the insufficient supply of buses. In most cases, passenger boarding time limits me capacity of a bus stop and, in consequence, of the whole busway. The boarding time is as short as 0.8 sec/passenger in smooth boarding conditions but it increases up to 5 sec/passenger when a case the bus is already crowded. In Ankara, passengers can enter the bus through the front door only for their ticket validation to be supervised by the driver. They alight by rear and/or middle doors with an average speed of about 1 sec. However, when the alighting passengers face the crowd of other passengers waiting on the platform to board the bus, alighting can become a problem. Providing for more than one bus stand at heavily loaded bus stops and organizing buses in ordered convoys would considerably improve the flow of buses even at "on line" bus stops (TRRL, 1991). The experience of Ankara confirmed me benefits of constructing a busway in me median of a artery, although it required the passenger to cross general traffic lanes when boarding and alighting from buses. The collision of buses with other vehicles coming from or going into side streets has been avoided and access to adjacent properties has not interrupted. 4. Public perception of the busway With the implementation of the Besevler Dikimevi busway in Ankara, mass transit became popular among the citizens. With this very first application in Turkey, both public authorities and passengers have realized the fact that busway transit could be efficient and effective if special priority measures are given to it. This spirit of mass transit gained momentum after the exclusive busway and was passed to other cities in the country, where either exclusive busway or bus lanes have been implemented. In October 1988, the Mayor of that period decided to close the exclusive busway on the grounds that the iron fences of the busway are not aesthetic and also because of the expectation that with the opening to traffic of the newly constructed Çelal Bayar Boulevard, the Besevler Dikimevi corridor would not be as congested as 54

59 before. As a result of this decision the Tandogan Besevler section of the exclusive busway was closed. The reaction of the public and press to this closure was very negative and the Chamber of Architects even brought the subject to court. Finally, with the decision of the court, the closure of the Besevler Dikimevi exclusive busway has been rejected and the already closed portion (Tandogan Besevler) has been transformed into a bus lane scheme. This event alone shows that there exists great public participation in and cooperation with the exclusive busway and that passengers are satisfied with the performance of busway transit although its performance could be further improved with alternative dispatching policies and traffic management techniques. D. New developments in Ankara mass transport The transport studies executed by SOFRETU (EGO, 1992) revealed that by the year 2010 a high capacity LRT system would be needed on the Besevler Dikimevi and other corridors. A network of 54 km of LRT lines has been proposed and me build operate transfer deal was concluded for the construction of a line which would replace the busway, perhaps by At the same time work commenced on the full scale metro along the north south axis. The two systems will intersect at Kizilay in 1996, when the metro is expected to be completed. REFERENCES EGO, Municipality of Greater Ankara, General Directorate of EGO Transportation Survey, vol. 1, "Ankara Urban Transportation Study". EGO, Municipality of Greater Ankara, General Directorate of EGO (in Turkish) Transportation Survey, vol. 1, "Ankara Light Rail Transit System". Ozdirim, M., "Otobus Oncelikli Kent ici Ulasim Sisteminin Ankara'da Uygulanmasi", 1, Toplu Tasim Kongresi, pp Birgonul, M.T, "Economic evaluation techniques of urban transportation alternatives", unpublished M.Sc. thesis, Civil Engineering Department of Middle East Technical University. T.R.R.L The Performance of Busway Transit in Developing Countries Transport and Road Research Laboratory, Research Report 329. ENDNOTES 1. This paper is based on the contribution by Dr. T. Birgonul and Mr. H. bayirtepe. Civil Engineering Department, Middle East University, Ankara, Turkey. 2. For the distinction between common carriers and individual transport modes adopted in this paper refer to the categories of transport modes given in table 2. III. CASE STUDY: THE BRAZILIAN EXPERIENCE IN PLANNING, IMPLEMENTING AND OPERATING PUBLIC TRANSPORT ON SEPARATED ROAD AND LIGHT RAIL TRAVELWAYS 1 55

60 A. Introduction 1. Brazilian urban transport at large Brazil is a large country with visible regional differences in economic development, urbanization, personal incomes, culture etc. Most of these differences have occurred over the last 35 years when a drastic change in Brazil's previous development pattern has been observed. Within this period, total population has more than doubled, today reaching some 145 million inhabitants. The population living in urban centres has jumped from 19 million to 80 million; out of this total, 40 million reside in nine metropolitan areas (Rio de Janeiro, São Paulo, Porto Alegre, Belo Horizonte, Recife, Salvador, Curitiba, Fortaleza and Belem). About 90 million motorized trips are made every day in Brazil's 117 urban areas with a population of more than 100,000. However, such trips constitute only less than a half of total person trips. Walking trips over a distance of above 500 m amount, for example, to 25 per cent of total trips in São Paulo and to as much as 70 per cent in middle sized cities (Wright and Sant' Anna, 1989). Out of the total number of motorized trips being made in the above mentioned urban areas, about two thirds are accounted for by public transport; 56 per cent by bus (and in a very minor portion by trolleybus) and 6 per cent by suburban railways and metros. About 32 per cent of motorized trips are made by cars. The car ownership rate in Brazil already amounted to 77 cars/1000 inhabitants in 1988 (MVMA, 1990) and was much higher in metropolitan areas. In the metropolitan region of São Paulo, me number of cars exceeded 2 million in 1987, car ownership rate was above 140/1000 inhabitants and 43 per cent of households possessed a car. The inner areas of metropolitan regions showed, presumably, even higher car availability since, contrary to the tendency observed in developed countries, the most affluent segments of the population use to live near to city centres and, certainly, within the limits of the key city of a metropolitan region. In urban areas, road transport accounts for about 90 per cent of people's motorized travels and is also essential for the movement of goods. As traffic congestion increases, efficient use of available road space becomes the crucial issue in transport policies. Also the impact of road transport on the environment gives rise to considerable concern. Car traffic is the main cause of road congestion and road accidents. Buses and cars generate most of the air pollution and noise (Wright and Sant' Anna, 1989), and air pollution often exceeds the level considered tolerable by human beings (UNEP, WHO, 1988). Rail based systems, specifically built for or adapted to urban passenger transport, are available in a few metropolitan areas. Metro operates in São Paulo (Belda, 1989) and Rio de Janeiro (Barat, 1989), while suburban railways have been developed in Belo Horizonte, Recife and Porto Alegre (Lindau and Rosado, 1989). Although there are plans to develop rail transport in many cities, buses are, and will be for a long time, the main public transport mode, owing to the relatively low costs of the provision and operation of bus services. For example, studies on transport systems for Curitiba indicated mat the proportion of investments required for the provision of 1 km of busway, light rail line and metro would be as 1:10:100, respectively (Via Urbana, 1991). The cost of public transport services are already high, especially for the low income population groups. By 1985, me urban poor were spending up to 12 per cent of their wages on transport tickets (Ministerio dos Transporte, 1985). Moreover since they are usually forced to live in distant suburban areas where housing is relatively cheap, while a significant proportion of jobs is concentrated in central districts, the urban poor also bear significant costs in the time they spend on travel. In 1985, the average commuter in Metropolitan São Paulo was spending as much as 2.5 hours per day on travel by public transport. In 1985, the Federal Government introduced the so called vale transporte (a bonus partly paid by employers, the main objective of which was to subsidize public transport users) to reduce the share of transport expenditure in the budgets of low wage workers. This policy reduced the users' perceived costs 2 of transport and changed users' attention from the costs to the quality of public transport service. A recent survey made by ANTP (1991) identified the public transport attributes highly valued by travellers in São Paulo. In that study, 62 per cent of the users asked for an increase in the supply of services; 15 per cent for greater comfort in travel; 13 per cent for safety improvement and 10 per cent for lower fares. Under the item "increase in supply", 35 per cent of respondents asked for more vehicles to be available and better reliability of service 3, 14 per cent for the shortening of waiting times, and 13 per cent for improving the accessibility to their destinations. 2. Development of public transport in Brazil 56

61 The evolution of the urban transport sector in Brazil has been strongly linked to changes in its economic and land use development contexts. Table 1 summarizes the evolution of transport supply in Brazil from the end of me nineteenth century on. History shows that the Federal Government did not interfere in decisions related to urban public transport until the middle of the 1970s. Until then, state and local governments had the full responsibility for planning and financing public transport but were discriminated against by the existing tax structure which was orientated towards accumulating the financial resources in the hands of the Federal Government. The creation of the Empresa Brasileira de Transportes Urbanos (EBTU), in 1976, was the first action of the Federal Government aimed at formulating a national urban transport policy. From the mid 1970s to the late 1980s, decisions about public transport were increasingly concentrated at me federal level. This process can be linked to the beginning of an economic crisis and growing income disparities in the country. It should be noted that in the period from 1964 to 1983, Brazil was ruled by the armed forces, and that this coincided with a substantial increase in Brazil's external debt and the loss of power by local authorities. However, it was in that period that metros and new suburban rail systems became a reality in Brazil. Even then, transport actions were largely directed not to the sector itself, but towards such objectives as employing low wage workers and reducing oil import bills. Indeed, it was the concern with the cost of petroleum fuels that brought about the development of the PROALCOOL, i.e., a national programme for developing the use of sugarcane alcohol as an energy source alternative to gasoline. By the mid 1980s, the strategic decisions on transport were in the hands of federal authorities for such areas as energy, finance and the foreign exchange. The strategy for financing public transport was also drastically modified, and institutions such as EBTU and RFFSA/CBTU (Rede Ferroviaria Federal Companhia Brasileira de Trens Urbanos) became involved in the search for external funds which were then allocated to programmes of often doubtful priority (Fagnani and Cadaval, 1988). In consequence, the relationship between debt and investments, which in the case of EBTU was around 0.5 per cent in 1980, reached 71 per cent in 1985 and 538 per cent in Table 1. Evolution of transport supply in Brazil Period End of the nineteenth century Public authority mainly involved Dominant transport mode State government Electric trams Operator Type of contract Economic context and urban development Private enterprise Concession (**) Primary economy Export Local government Trams Small vehicles (pick ups) Private enterprise Individual operators Local government Buses Small private enterprise Local government Buses Small private enterprise Concession Permission Provisional permission Permission Provisional permission Provisional permission Consolidation of central areas Import substitution process Exponential growth of suburbs Industrialization process Suburban and urban growth Economic miracle Motorization Quick urban growth Federal Buses government Metros State government Big private enterprise States Conditional permission Monopoly Beginning of the economic crisis Metropolization 57

62 Local government Informal transport Individual operators No contract process Income disparities 1988 today Local government Bus (*) Source: Brasileiro (1992). Metros Informal transport Projects of modern tramway Big private enterprise State government Individual operators Public consortium Process of passing to the local authority Payment by production Social and economic crisis Urban growth Growth of satellite towns (**) Note that the term "concession" means the right of way that the Government gives to an enterprise to operate a system or part of it, according to conditions previously defined. The term "permission" means that the Government gives a license for operation without pre defining conditions. The Constitutional reform of 1988 introduced a discussion concerning the legal limits of action by the Federal Government in the field of urban transport. The involvement of the Federal Government is now limited to the definition of national transport policy, formulation of strategic plans for the development of urban transport and establishing of basic transport regulations (Lima, 1989). The Federal Government closed down the EBTU which for several years had played a key role in the development of urban transport, and the local governments became again responsible for public transport in urban areas. The current period is characterized by a deep social and economic crisis. Fast urban and suburban growth creates the need for considerable investment in public transport; however, very little is being made due to the lack of funds. The Federal Government is trying to enhance private investments in public transport but, so far, without much success. Very little capital, if any, has arrived from international institutions since the closure of the EBTU, inter alia, because of the local authorities' lack of experience in dealing directly with the international funding organizations. Table 1 also reveals the importance of Brazilian private bus operators. Private bus enterprises tend to operate under permission or concession conditions. Fares are fixed by transport authorities at the level which expresses a compromise between the cost of service provision and passengers' budget constraints. When fare collection does not cover actual costs, subsidies are demanded by bus operators. In order to overcome financial deficits, new deals have been recently made between public authorities and private operators. In some cities, public authorities contract the services of private operators. Fares are passed to the public authority which pays the operators for the amount of services they actually provide (Severo, 1991). Fares may, however, remain separated from the actual costs of services rendered, owing to social reasons. 3. Policies towards improving Brazilian public transport There has been a clear tendency for policies that give privileges to public transport. Among the key actions promoted by federal authorities are the following: (a) Subsidizing public transport fares through the vale transporte; (b) Building light rail transit systems (or "modern tramways", as they are usually called in Brazil) on many transport corridors in several cities; (c) Improvement of existing bus systems through the adoption of sophisticated techniques aimed at achieving better performance by bus transport. The first policy has already been presented in previous sections; it is orientated towards meeting the needs of low income groups. The second proposal (to build LRTs) is still at an early stage of implementation. Indeed, medium capacity 58

63 systems based on rail technology have been proposed with a certain degree of "euphoria and fashion" (Boletin, 1992) as an option for solving transport problems in Latin America. It should also be noted that LRT is not a new modal option 4. The third proposal considers bus transport and assumes that it is possible to improve its capacity through different measures such as a high level of travelways segregation, operation of buses in convoys, better on board fare collection systems, improved lay out and operation of bus stops, and a wide implementation of bus actuated traffic signal systems. Indeed, bus operation along properly designed segregated lanes offers a much better level of service than in mixed traffic conditions (Lindau and Willumsen, 1988). According to Wright and Sant' Anna (1989), there is no need in terms of capacity to implement rail based systems on the corridors on which passenger flows are smaller than 30,000 passengers/hour/direction or can be split among several roads. The basic question nowadays is: what is the place for medium capacity systems within the urban transport network and public transport segmented market? Medium capacity systems may be defined as those capable of carrying flows varying from 12,000 to 30,000 passengers/hour/direction (Boletin, 1992). By following this definition, several technologies can be identified as capable of attending medium flow corridors. Among them is bus technology, provided buses are given a separate right of way and there is an appropriate regime of bus operation (Lindau, 1983; and Gardner et al, 1991). Indeed, the potential capacity of different transport technologies heavily depends upon how have they been applied and operated. In general terms, any road or light rail based system designed to cater for medium size transport demand must include a proper combination of vehicle characteristics, type of right of way and regime of operation. Some other aspects, such as the type and location of the critical stops or the type of traffic control system, have also to be considered if medium flow capacities are to be achieved. The decision on the adoption of one or another public transport priority measure must be based on specific criteria which, in turn, are related to me hourly and daily volumes of vehicle and passenger flows (Lindau, 1983). In terms of right of way, a mix of me following options is usually considered in either bus or LRT projects: (a) Exclusive right of way total segregation through tunnels or elevated roads for the exclusive circulation of bus/tram vehicles; (b) Segregated right of way this may result in the introduction of exclusive median lanes, such as those in Curitiba, São Paulo, Porto Alegre and Goiania, where at grade intersections prevail and no physical barriers exist for pedestrians crossing the road; 5 (c) Reserved right of way part of the road is reserved for buses or trams. The so called conventional bus or tram lanes may be delineated by painting the road; 6 (d) No right of way the circulation of buses or trams occur in mixed traffic conditions, with transit vehicles competing for space with other vehicles. B. Medium capacity public transport modes 1. Choosing between busway and LRT options The advent of the PNTM Plano Nacional de Transporte de Massa (EBTU, 1988) generated many controversies within the community of public transport professionals. Contrary to all previous Brazilian experience with planning and operation of high capacity bus transport systems (Szasz, 1978; and Lindau, 1987), PNTM 7 strongly recommended LRT technology as the only one capable of meeting medium size passenger flows. However, as has already been said, bus and light rail systems may both the achieve capacity and speed levels needed to serve such flows, all depending on the adoption of a proper technical and planning measures. In the discussion on medium capacity systems, which figures strongly on the agenda of professionals' meetings in Brazil, the supporters of the light rail option emphasize the following: 59

64 (a) The low level of generated noise and absence of air pollution; (b) High capacity, claimed to be reaching 35,000 passengers per hour in one direction; (c) Considerable improvement of comfort in travel; (d) Growing popularity of LRT throughout the world and its image as a new mode; (e) Contribution of the systems to urban revitalization and redevelopment; (f) Low costs of construction and operation, in comparison with metro; (g) Generation of jobs in rail and construction industries. They see light rail systems as a natural step to be followed in developing urban transport after the modernization of bus systems has been accomplished. The opponents or, rather, defenders of bus transport argue that: (a) A light rail system is an isolated entity, while busways can be easily integrated into the bus based public transport in Brazilian cities, do not need modal interchange facilities, allow attainment of high capacity by using available rolling stock, and are flexible in operation and expansion of the transport network; (b) Cost of busway development is much below that of LRT, while achievable capacity is similar, since no LRT system in the world reached the theoretical capacity of over 30,000; (c) Transport operators in Brazil are familiar with bus technology and the management skill in operating busways has been highly developed in the country; (d) Buses are produced locally and can be run on alcohol fuels. 2. Light rail systems in Brazil The experience of developing countries with the construction and operation of light rail systems is still modest and this may be explained, among other things, by the lack of funds to implement the systems. Brazil is no exception to this general trend. From the 1920s to the 1960s, trams were an important public transport mode, and in 1930 tram systems were operated in 40 Brazilian cities (Gonçalves et al, 1993). With the advent of mass motorization after the Second World War (see table 1), trams started to lose in competition with buses and minibuses. The result was an almost total eradication of trams in the late 1960s. One of the very few remaining schemes is the old tram operating in Santa Tereza city in the State of Rio de Janeiro. The first Brazilian system with some similarity to the today's LRT was the pre metro of Rio de Janeiro inaugurated in In practice, this single line pre metro system operates as line 2 of the metro of Rio de Janeiro. According to the original project, the pre metro system is to be replaced by a full metro. The pre metro operates along 7.9 km of exclusive right of way with only two intersections at level. Its articulated electrical coaches provide space for 300 passengers. In 1986 a consultants' report suggested the implementation of an LRT system in Salvador, Bahia. The National Mass Transportation Plan (PNTM), issued by the EBTU in 1988, recommended the implementation of LRT lines on 50 corridors of 24 Brazilian cities before the year The implementation of the LRT option proposed by PNTM would imply a need for investments valued about $US4.4 billion, with $US2.4 billion destined for infrastructure works and some $US2 billion allocated to the acquisition of vehicles. PNTM relied heavily on the expected financial assistance of such institutions as the World Bank and the Brazilian National Bank for Economical and Social Development BNDES. However, so far, there is a lack of such support, since both mentioned institutions are rather concerned with the need to control Brazil's public deficit. PNTM assumed also large participation by the private sector in the development of Brazilian LRT systems which did not yet occur. Moreover, the Constitutional reform of 1988 limited the action of the Federal Government in the field of public transport, while the closure of the EBTU deprived the industry of a driving force for the implementation of LRT's in Brazil. It should be mentioned that most of the arguments presented by PNTM in favour of LRT schemes were the same which could be used to support the case of busways, i.e., flexibility to adapt to variations in transport demand, possibility of gradual upgrading, to operate on steep gradients and along short radius curves etc. In particular, there is little justification based on practice or simulation work to back up the introduction of 60

65 LRT systems along the existing busways. The main argument for replacing busways by LRT is that of the need to improve the nearby environment by reducing drastically noise and air pollution levels, in particular around stops attended by close to 500 buses during peak hour. The other reason for replacing busways by LRTs is the improvement of comfort inside transit vehicles. PNTM emphasised that the development of LRT systems would not only improve the quality of public transport, but also help the Brazilian rail industry. This industry could benefit from the economy of scale by producing about 2500 vehicles needed for LRT schemes recommended by PNTM (Gonçalves et al, 1993). However, the plans for building LRT systems are dependent on foreign investments and this type of financing is usually linked to the acquisition of equipment and technology from the financing country. According to Gonçalves et al (1993), despite the difficulties regarding availability of funds for financing the implementation of LRT systems, the action of lobbies of manufacturers and consultants brought about preliminary studies for 15 Brazilian cities. Table 2 presents some basic information about the proposed LRT systems. City Table 2. Proposed LRT schemes in Brazil Length (km) Total cost ($ millions) Cost/km ($ millions) Campinas Curitiba Brasilia Belo Horizonte Goiânia Rio de Janeiro Manaus São Gonçalo Santos Maceio João Pessoa Blumenau Salvador 49 1, Natal Vale Paraiba Total ,070.5 Av: 12.7 Source: Gonçalves et al, 1993 The majority of planned LRT corridors are radial, linking suburban areas to city centres. In general terms, these systems aim at meeting a travel demand of about 20,000 passengers per hour and direction. Proposed types of vehicles tend to be single or double articulated, 30 m long, with individual capacities of about 300 passengers. Apart from Goiânia and Curitiba, the remaining LRT systems intend to employ few at grade crossings with urban roads. The significant variation in costs is basically related to the possibility of using an existing right of way which allows a substantial reduction of the land expropriation costs. By the end of 1992, only the first, 8 km long line of LRT in Campinas was ready for commercial operation. The line was built in anticipation that two main transport corridors, Avenida Anhanguera on an east west orientation, and Avenida Goias in a north south direction, would be saturated with buses as transport demand on these corridors was expected to reach, by the year 2000, volumes of 26,700 and 20,200 passengers per hour in one direction, respectively. The system will use coaches similar to those operating on the pre metro line in Rio de Janeiro and will provide services from 6.00 a.m. to 9.00 p.m. Its development was financed by the State of São Paulo via FEPASA (Ferrovias Paulistas S.A.) the owner of a former freight railway right of way into which the LRT track was inserted. FEPASA, as the holder of the concession to operate the system, intends to do it through a contracted private operator. FEPASA will collect the fares, set at the same level as for the bus system in Campinas, and pay the contracted private operator for the amount of effectively rendered services. The contractor will cover the costs of fleet maintenance, cleaning and security services. Further extension of the Campinas LRT system is dependent on obtaining extra funds. 61

66 In Salvador, according to Gonçalves et al (1993), the work has been stopped due to lack of funds. Where the exclusive right or way has already been provided, articulated buses are used in a sort of provisional operation. The original project for LRT in Brasilia has been gradually altered and the system will operate on a fully separated right of way with energy supplied by a third rail. Therefore, the system resembles rather that of a metro or a suburban railway and should now be more properly defined as a metro. The system is to be inaugurated in April 1994 and will have a network totalling 40 km, out of which 9 km will be underground. About 50 per cent of investments will be financed by BNDES with the remaining funds coming from the Federal Government and the Government of the Federal District (Revista Ferroviaria, 1992). Curitiba presents another example of an LRT system still waiting for implementation in the consequence of the lack of funds (Revista Ferroviaria, 1991). The development of an LRT is seen in Curitiba as a way to improve the quality of public transport and the environment, as well as the means to building the image of the city. Curitiba has often been shown as a model for urban development in the context of developing countries and, certainly, the local population ranks improvement of the environment at a high rate. It is believed that, by replacing the existing busways by a large LRT system, the need for metro may be eliminated, taking into account the fact that Curitiba's main arteries, along which busways are located, are sided by parallel roads with spare capacity for providing bus services complementary to LRT, if required. Local authorities in the city expect to set a deal with private investors in a way that private and public sectors will each be responsible for covering about 50 per cent of the total costs of the system. The private investor, probably the current bus operators, would have to bear the costs of LRT fleet and other equipment, while the Municipality would provide the right of way, construct stations and tracks, and help to find funding sources for the implementation of the whole system. Its cost is estimated at $US250 million, out of which $US45 million is for me acquisition of the rolling stock (Via Urbana, 1991). Current efforts are, however, placed on the replacement of articulated buses used in the city busways by bi articulated buses with a capacity of 300 passengers per bus, including 56 seated. It is understood that the natural evolution of busways based transport is to increase the size of buses before introducing an LRT system. Moreover, the busway system in Curitiba is of a high quality and provided with interchange terminals where passengers change the buses without paying a second fare. 3. Bus priority systems in Brazil While the development of LRTs is still at an early stage in Brazil, the country has long experience in catering for heavy passenger demand along urban and suburban corridors, either by metro and suburban railway systems or by high capacity bus priority schemes. Indeed, Brazil has made some pioneering work in improving the performance of bus transport. (a) Bus priority It is commonly accepted that flexibility, speed, cost, and capacity of the services provided are the main differences between the road and rail based public transport modes. Most of the studies attempting a better understanding of the potential capacity of different urban transport modes usually indicate that the cheap and flexible, but slow, low capacity small bus stands at one extreme of the scale, whereas the sophisticated metro lies at the other. The environmental criteria point to the same hierarchy. Indeed, if buses are sharing road space with other traffic, the obvious result is a quick deterioration in the service level because of traffic congestion. The spectrum between an ordinary bus and a metro may be filled by a variety of alternative modes, ranging from buses of various size to other than metro forms of rail transit (Lindau and Willumsen, 1988). The type of right of way and the regime of operating a mode are the fundamental factors to be considered here. Many Brazilian authorities have adopted various priority measures to protect bus operation from traffic congestion, providing thereby the passengers who use this mode with higher speed in travel and achieving a more efficient use of expensive road space and bus fleet. The main measures to create bus priority schemes include: (a) provision of separate travelways in the form of with flow and contra flow bus lanes or busways; (b) granting buses the priority at traffic signals; and (c) permitting buses an entry right or turns prohibited to other traffic. Many successful schemes result from a combination of these and other measures. Brazil has developed a unique experience in designing innovative high flow bus priority schemes. This experience is particularly suited to overcome congestion problems associated with the semi chaotic behaviour of numerous buses arriving almost simultaneously at critical bus 62

67 stops where large numbers of passengers wait to board. Under such conditions the introduction, for example, of a single conventional bus lane may generate more benefits to non bus traffic than to buses themselves. (b) Choosing the priority measures The criteria for choice of bus priority measures range from general "rule of thumb" criteria to more detailed indications derived from studies on me implemented schemes, test track experiments and computer models (Lindau and Willumsen, 1988). Lindau (1985) presents a critical review of some of those criteria. Table 3 contains a summary of minimum flow and capacity criteria for the introduction of different types of bus lanes along urban arteries. As is shown in the table, some options may provide fairly high capacity. Bus lanes situated in the centre of a road (median bus lanes) may cater for flows of over 21,000 passengers per hour and direction. Brazilian experience in this regard is quite extensive. In attempting to overcome some of the shortcomings experienced in earlier high flow bus schemes, each particular corridor must be studied in detail. Microcomputer assisted design methods, such as BUSWAY (Lindau, 1988), BLISS (Iunes and Willumsen, 1988), IRENE (Gibson et al, 1989) and the expert system developed at University College, London (Tyler, 1991), can help achieve the best possible design of bus priority scheme prior to the field implementation. Table 3. Minimum flow and capacity criteria for design of bus priority schemes Type of separate travelway Minimum peak flow (buses/hour/dir) Capacity buses/h/dir pass./h/dir With flow bus lane Contra flow bus lane Double bus lane Median bus lane Source: Lindau and Willumsen, 1988 BUSWAY simulates the operation of buses along high flow, exclusive bus lanes situated in the centre of a carriageway in a high level of detail. It is currently being adapted to simulate also the operation of trams along tracks situated in the centre of a road. BUSWAY is particularly useful for testing the combinations of alternative geometric (changing location of traffic signals, bus stops etc.) and operational conditions (buses ordered in convoys, non convoy operation etc.) under identical traffic input settings. BLISS is orientated towards simulating the operation of buses along with flow, kerb side bus lanes. It uses a platoon dispersion model to represent the behaviour of priority and non priority vehicles at different points of the road. BLISS is being extended to simulate bus stop operation in detail, this being once main limitation for its use in developing countries. IRENE simulates macroscopically the behaviour of buses approaching an isolated bus stop. IRENE focuses on the analysis of me performance of a bus stop subjected to different levels of passenger demand, circulation of buses of different sizes and various forms of bus operation. In the United Kingdom, Tyler (1991) is investigating factors of impact on the location and design of bus stops, as well as the impact of the implementation of a bus priority scheme upon other traffic in the corridor. He developed an expert system mat helps design high capacity bus priority schemes. (c) Strategy for the adoption of bus priority measures The need for selecting one or a set of measures for establishing bus priority may be characterized by the willingness to: (a) Increase the public transport capacity of service and productivity; (b) Reduce travel times along congested corridors; (c) Reduce energy consumption, especially of fuels derived from petroleum; (d) Reduce the use of private cars. In order to analyse the critical points in bus circulation, it is necessary to collect a set of data following the 63

68 steps indicated below: (a) Define a study area, usually confined to a sector of a city or a transport corridor; (b) Identify the roads used by buses in the study area; (c) Collect traffic information on the peak and off peak, performance of buses and general traffic (volumes, travel speeds, delays etc.); (d) Indicate locations with bus circulation problems, such as (EBTU, 1982): Speeds lower than 20 km/h; Excessive delays at traffic signals; Difficulties to reach bus stop stands due to irregular parking; Excessive number of passengers boarding/alighting buses; Traffic congested road segments. Basically, bus priority measures can be divided in two sets: (a) Measures to facilitate bus circulation in the mixed traffic conditions; (b) Measures to establish separate travelways for buses. Types of measure Traffic circulation devised for buses Table 4. Bus priority measures and their effects Route extension Bus speed Exclusive conversion A U A Number of steps Effects Bus waiting time Number of traffic signals Bus delays at traffic signals Location of bus stands A A Distance between bus stops Dividing bus stands Paving bus routes A A S Kerb bus lane U M Median bus lane U M Contra flow bus lane S U M Busway U M Bus only road S U M Adjustments in traffic cycle Adjustments in percentage of green light Bus actuated signals A A Extra bus only signal A A Removing signal for A buses Source: Adapted from EBTU (1982). The letters indicate the recommended period of application/incidence: S = Sometimes; A = Always; M = Mainly at peak hours; U = Usually; L = always and a lot A The first set of measures attempts to improve the performance of buses without imposing penalties on other vehicles, while the second set includes the possibility of reallocating the existing travelway space from general traffic to an exclusive use by buses. A A A A 64

69 Tables 5 and 6 present a list of alternative bus priority measures. They aim at helping the designer in the selection of adequate measures for solving bus circulation problems. (d) High flow bus priority schemes: practical examples In 1982, the EBTU published the first Brazilian state of the art study on bus priority schemes. Apart from reviewing the fundamental concept of bus priority schemes, the publication described the schemes introduced in Porto Alegre, Goiânia, Curitiba, Belo Horizonte, Rio de Janeiro and São Paulo. Since then, some priority schemes have been discontinued, the majority consolidated and a few even superseded by more sophisticated measures or technologies. This section focuses on the practical results derived from the implementation of bus priority schemes on several transport corridors in Brazilian cities. Particular attention is given to cases in which significant benefits were achieved. (i) São Paulo In 1932, São Paulo had only 400 buses (Mandon, 1991). This number was compatible with the city's built up area (approximately 150 square kilometres) and the population of 1 million (Armstrong Wright, 1986). By 1962, the city's size had reached 750 square kilometres and urban population exceeded 4 million. In 1980, the metropolitan area was 1400 square kilometres and the population counted more than 12 million. In 1990, there were 5000 public transport vehicles operated by private enterprises and almost 2800 buses run by CMTC, a state owned company 8 (Mandon, 1991). The market is highly regulated by public authorities. Simultaneously, drastic changes have taken place in the distribution of workplaces and residential areas, making the people strongly dependent on public transport. Bus priority schemes were introduced in the late 1970s to alleviate the ever increasing demand for public transport. In 1981, 74.6 km of roads were covered by such schemes, ranging from contra flow lanes (Pereira, 1980) to comprehensive schemes such as that one implemented along the Rangel Pestana Celso Garcia corridor where, in 1979, more than 30,000 passengers were carried by a flow of 500 buses during the afternoon peak hour (Szasz, 1978). Type of measure Traffic circulation devised for buses Exclusive conversion movements Location of bus stands Distance between bus stops Dividing bus stands Paving bus routes Table 5. Bus priority measures and their adverse impacts Need strong control Affect access to car parking Impacts Affect car traffic safety Affect pedestrian safety Imply land use changes A A U S A S Kerb bus lane A A A Median bus lane S A A Contra flow bus lane A A L A Busway A A S Bus only road A A A Adjustments in traffic cycle Adjustments in percentage of green light Bus actuated signals A S Extra bus only signal 65 S S S S U S

70 Removing signal for buses Source: Adapted from EBTU (1982). The letters indicate the recommended period of application/incidence: S = Sometimes; A = Always; M = Mainly at peak hours; U = Usually; L = always and a lot Table 6. Costs of the implementation of bus priority measures Type of measure Traffic circulation devised for buses Exclusive conversion movements Location of bus stands Distance between bus stops Dividing bus stands Paving bus routes Kerb bus lane Median bus lane Contra flow bus lane Busway Bus only wad Adjustments in traffic cycle Adjustments in percentage of green light Bus actuated signals Extra bus only signal Removing signal for buses Source: Adapted from EBTU (1982). Cost of implementation Very low Low High Very high The concept of bus operation in ordered convoys, depicted in figure 1, was first implemented along Av. 9 de Julho (Szasz, 1978) in order to organize the flow of at 300 buses per peak hour along a with flow, kerb side bus lane. On a 4 km section of the avenue with five signal controlled intersections and six bus stops, and with 1200 passengers boarding at the critical bus stop during the peak hour, average afternoon peak hour bus speed increased from 10 to 18.5 km/h (+85 per cent) once the operation of ordered convoys consisting of six buses was introduced. In the case of the Rangel Pestana Celso Garcia corridor, the bus stands on the critical bus stops had to be divided. Parallel pedestrian platforms located at the kerb and in the middle of the avenue were equipped with six loading stands each, which made it possible to accommodate simultaneously convoys of up to 12 buses. Along the critical 2.7 km long section that holds five divided stations and 17 signal controlled crossings (nine for pedestrians), average bus speeds jumped from 6.1 to 9.8 km/h (+60 per cent). In an effort to overcome saturation problems, new solutions have been implemented along the Santo Amaro 9 de Julho corridor (Kato, 1986). The so called "trunk express line" consists of a median bus lane with transfer terminal or terminal bus stops built along it to allow integration with feeder lines operated by conventional buses (see figure 2). On a 11.8 km section of the median lane, 16 out of the total of 19 bus. stops are provided with an extra traffic lane to allow bus overtaking manoeuvres. Full scale services began in 1987 when three different trunk lines started to operate. Each line serves a group of stops. A bus operating line 2 does not stop until it reaches the second group of stops and returns soon after serving the last stop in its group. Finally, a line 3 express bus travels along the total length of me scheme, i.e. from upstream to downstream terminals. Average speed increased from 14 to 23 km/h (+62 per cent), while average passenger waiting time at bus stops during the peak hour dropped from 12 to 2.5 minutes. Bus operation improved considerably even before the scheme was fully implemented (bus lanes were only partially completed and traffic signals not coordinated). Travel time by bus from Santo Amaro to the city centre was reduced from 80 to 50 minutes, and this was achieved at the expense of only 10 minutes of additional travel time by cars. A worldwide busway survey study conducted by TRRL (Gardner et al, 1991) ranks this 66

71 corridor as probably the best among all. Nevertheless, some further measures are still required in order to improve its efficiency. Figure 1. The convoy concept Source: Department of Transport and Institution of Highways and Transportation (1987) Figure 2. High capacity bus system Source: Department of Transport and Institution of Highways and Transportation (1987) An exclusive median trolleybus lanes were implemented along the 33 km separating Jabaquara from Sao Mateus (Seção Transporte de Massa, 1986). The scheme's operational centre has computers and a radio system to control the fleet of 141 trolleybuses, 55 stops and nine terminals. The system serves 250,000 passengers a day (15,000 peak direction flow) with average operational speed of up to 25 km/h. São Paulo authorities have just proposed the implementation of bus priority schemes on six other transport corridors. By using the existing infrastructure, these new schemes will integrate the intercity bus lines, the suburban rail and metro systems, and the local bus routes. The effort will require combined actions of the São Paulo State Government, metropolitan authorities and the private sector. The latter will be responsible for supplying the vehicles, infrastructure and terminals. (ii) Curitiba In Curitiba, several so called structural radial routes started to be built in 1974 as a part of an urban development package aimed at organizing the growth of the city. Each structural radial route is formed by three separate roads running in parallel. The middle road consists of an exclusive two lane busway flanked on either side by a one way service road. Each of two outer roads is also a one way, multi lane road. There are 50 kilometres of exclusive busways built along five arterials with 15 different routes covered daily by 250 buses. They constitute a backbone of an integrated mass transit systems encompassing also circular inter district bus routes and feeder lines which provide access to terminals located at the starting points of 67

72 busway and at strategic transfer points throughout the city. The terminals permit the transfer and distribution of passengers to multiple destinations. The buses travel along the busways with a median speed about 20 km/h and reach the sectional speed up to 60 km/h in noncommercial areas. High capacity articulated buses, specially designed to serve the system, are being introduced and headways are as short as 1 minute during peak traffic hours when the demand reaches the volume of 10 15,000 passengers per hour and direction. The integrated mass transit system serves about 1.2 million passengers daily and the demand has grown by 30 per cent over the six years since the system was established. An integrated fare system enables the user to pay a single fare for travel on any of the 15 routes operated on busways and on inter district conventional bus routes, although not on the feeder lines (Lerner 1991). (iii) Porto Alegre Porto Alegre is the capital of the southernmost state of Brazil. In 1991, the population of me Porto Alegre Metropolitan Region counted 3,250,000 and there were 474,000 cars registered in the region. Apart from suburban railways, more than 2800 buses served public transport, and out of this 1450 the city of Porto Alegre itself. Experiments with bus priority schemes started in As a result of several studies, some 30 km of median bus lanes were built along five radial roads. The development of busways in Porto Alegre was largely influenced by the experience of Curitiba and São Paulo. The concept of a structural radial route was combined with mat of ordered bus convoy operation on the 3.3 km long Farrapos 1 corridor. On the 8 km long Benito Gonçalves João Pessoa corridor the "trunk feeder" operation was tried. Erico Veríssimo corridor was introduced to influence the land use development pattern. The case of the Benito Gonçalves João Pessoa corridor illustrates the need for a comprehensive approach to the development of bus priority schemes. As has been mentioned, a trunk feeder operation was attempted on this corridor. However, contrary to what was achieved in Curitiba, Porto Alegre constructed only a few of the interchange terminals out of those proposed by the general plan. Worst of all, it was decided to start me operation of the trunk feeder system on a single radial corridor with only one interchange terminal located at the city limits. The passengers did not perceive many advantages in changing from conventional feeder buses to articulated buses operated on me busway, apart from an initial fare which was lower than the flat bus fare charged elsewhere in the city. They have, moreover, to face the inconvenience of the transfer without benefit of changing to buses that would connect them to other radial corridors. Once the fare on the feeder lines was set at the same level as for the remainder of the bus system, passengers forced the public authorities to finish with the system and re establish the former regime of operation. The Assis Brasil busway (4.6 km), located along a radial corridor, connects the city centre with outer suburbs. Buses of several sizes operate along this busway, especially during the peak hours. Some local adaptations, such as trailers 9 towed by conventional buses, run together with conventional buses inside the busway. In spite of me operation of six bus convoys, the Assis Brasil corridor nears saturation level during afternoon peak hours, and some new design efforts are required to improve the bus routing system. The Farrapos 1 and Assis Brasil corridors provide good examples of how planning and practice can diverge if me institutional aspects of busway operation are not appropriately considered. Transit authorities wanted to have a single bus company operating large articulated buses along median lanes, with lines of conventional feeder buses converging at transfer terminals. Private operators did not agree with amalgamating part of their companies. The initial operation plan had to be totally reviewed before operations started. Bus stations had to be rebuilt to accommodate a much larger flow of buses. The ordered convoy concept (COMONOR) was introduced, with lines divided into different groups according to their frequency and destination (Secretaria Municipal dos Transportes, 1980). Until the advent of concurrent suburban rail services, in 1984, and consequent reduction of total bus flow, some of the bus lines operating along the Farrapos 1 corridor could not be accommodated inside the median bus lane. Table 7 displays some data on bus and passenger flows as well as on bus operating speeds recorded in the Farrapos and Assis Brasil corridors in various years. It shows that bus operating speeds improved in comparison with those observed before busways had been developed. This was achieved in the period of time in which traffic grew up in the city, congestion increased and general traffic speed tended to decrease. However, the current speeds are below the level of 25 km/h at which they were assumed to be according to the original busways development plan. The speed improvement recorded in 1985 was attributed to the 68

73 installation of detectors and the vehicle actuated traffic signals (Branco, 1985). Peak hour bus passenger flow increased in the Assis Brasil corridor and stabilized in the Farrapos corridor, in spite of the introduction of suburban rail services in that corridor. It should also be noted that the accident rates in both corridors dropped by 19 per cent between 1978 and The initial investments required for building this busway network in Porto Alegre averaged $US1 million (1982 dollars) per kilometre (EBTU, 1983). At that stage, not much was done for improving the pavement of roads, especially close to bus stops where acceleration and deceleration manoeuvres impose severe strains'. Later on, in the Farrapos 1 corridor me busway road surface had to be totally replaced by a pre stressed concrete structure. C. Lessons from the Brazilian experience 1. Highlights The overview presented above of actions taken by the federal, state, metropolitan and municipal authorities in Brazil over the last three decades reflects the lack of consistent urban transport strategies for the country as a whole. Nevertheless, some of me measures which have been introduced gained a fair amount of support and recognition even at international levels. Among them, the following have to be emphasised (all have been described in detail in previous sections of this paper): Year (a) The creation of EBTU as an agency oriented towards defining national public transport policies, an experience which is currently being transferred to other countries; (b) The development of the PROALCOOL programme to reduce dependence on imported petroleum fuels; (c) The implementation of vale transporte, the bonus for transport users aimed at compensating the low wages of many workers, that has drastically influenced me preferences of public transport users; (d) The introduction of measures devised to give the bus priority over the general traffic, including the provision of busways and the development of the concept of bus convoy operation (COMONOR). Table 7. The results of "before and after" studies on two Porto Alegre busways Bus flow a Passenger flow Speeds (km/h) b Daily total Peak b Daily total Farrapos Peak b Buses Cars Operational scheme Mixed traffic conditions Exclusive median bus lane, several bus lines outside bus lane 1985 n.a. n.a. n.a. n.a Exclusive median bus lane 1987 n.a. 272 c n.a. n.a n.a. Exclusive median bus lane, few bus lines outside bus lane 1990 n.a. 304 n.a n.a. Exclusive median bus lane, few bus lines outside bus lane Bus flow a Passenger flow Speeds (km/h) b Assis Brasil Mixed traffic conditions Exclusive median bus lane, few bus lines outside bus lane 1985 n.a. n.a. n.a. n.a Exclusive median bus lane 69

74 1987 n.a. 240c n.a. n.a n.a. Exclusive median bus lane, few bus lines outside bus lane 1990 n.a. 260 n.a n.a. Exclusive median bus lane, few bus lines outside bus lane Source: Lindau (1987), and Gardner et al (1990) n.a. = not available a Total flow, inside and outside median lane b Afternoon: 6 7 p.m. (direction CBD suburbs) c Bus flow in exclusive median lane d Afternoon: 7 8 p.m. (direction CBD suburbs) This paper focuses on reviewing the Brazilian experience with bus and LRT technologies directed towards attending medium size passenger flows. There are at least three good reasons for further exploring the medium capacity transport modes operating on at grade segregated travelways (Lindau, 1992): (a) Few cities in the developing world can afford me costs of transit technologies, such as metros, without drastically compromising social investments in other areas; (b) The majority of peak hour flows along main corridors of a metropolitan area stand well bellow me 57,000 or 80,000 passengers/hour/direction recorded by metros in São Paulo and Hong Kong, respectively; such levels of demand usually result from the concentration on a single metro line of the passengers who would otherwise travel along two or more corridors; (c) Public transport systems that use the existing roads can be improved as they have the potential to carry 20,000 or more passengers per hour and direction, while their capital costs are substantially lower than those of the metro. There are strong indications that both, LRT and busways, have not been fully explored options for serving medium traffic corridors, even though there are several schemes implemented in the developing and me developed world. This is particularly the case of Brazil, where many busways have been introduced, but very little has been achieved in terms of implementing LRTs. Brazilian experience with light rail based technologies is still at its early stage. Plans have been developed for several cities, but no relevant practical contributions are yet available. Perhaps the main explanation for this are the budgetary constraints. Most of the attempts for solving public transport problems in Brazil depend on the availability of external financial resources. Indeed, this may be a major problem in the development of LRT systems, because most external loans are linked to the acquisition of equipment from countries which are financing me project, and this may not be the most economical solution. The most important contribution of the Brazilian experience derives from the implementation of busways. Brazilian experience in this field has been utilized by other countries. Therefore, farther analysis focuses on several aspects that must be carefully understood before progressing with the design of a busway scheme. 2. Designing busways with the Brazilian experience in mind There are several ways of granting buses a priority over other urban traffic. Several studies indicate that median bus lanes, or busways a solution tested and improved in several Brazilian cities present the best potential for improving bus performance along medium flow passenger traffic corridors in cities of developing countries. There is a wide range of factors to be considered when planning a busway. For example, one of the main challenges in proposing any public transport priority scheme is that of finding a compromise between the interests of different segments of the society, i.e., private and public transport users, pedestrians, local businesspeople, residents and the population in general. In designing a busway, one normally uses a combination of manual techniques, computation models and empirical methods. The large number of variables involved in the process the majority of them of a stochastic nature makes a formal analysis a demanding task. Thus, the use of computer models becomes essential for achieving a good project; some of the available models were briefly reviewed in a previous section of this paper. 70

75 Busways need fringe islands (or a central island together with other siding elements) that allow for the physical segregation of the buses from other vehicles in the traffic stream. The central or side islands are also important for protecting pedestrians who cross the road. The islands and siding elements tend to be extended up to the stop lines, just before the traffic signals at the intersections. One of the key advantages deriving from this type of busway is the auto control of right of way violations imposed by the heavy flow of circulating buses. There is also little interference from lorries serving commercial precincts along the corridor. Figure 3 shows a typical busway, highlighting the key elements to be considered in any successful design: (a) The busway width that may vary from two lanes (one for each direction of bus flow), each wide approximately 3.5 metres (total = 7.0 metres), to three lanes which allow overtaking manoeuvres for buses running in the same direction; these manoeuvres are usually restricted to some critical stops when trunk feeder systems are adopted; (b) The traffic signal control system that may be actuated by vehicles travelling along the main roadway, by buses circulating along the busway, or by mixed traffic arriving at the intersections from the crossing roads; thus, signal offsets may be determined as a function of nearby intersections (i.e., green waves) by, for example, an area wide control of traffic signals (Robertson and Vincent, 1975) or even be randomly set, as is usually the case of many developing countries. Although constantly referred in the literature as a way to improve speed, grade separated interchanges are seldom applied to busway schemes. (c) The ordering station that allows buses to be set in ordered convoys before they enter the busway; the ordering station may be constantly working or its operation may be restricted to peak hours, all depending on the flow of arriving buses. The ordering station may be located at one or both ends of the busway, along the main roadway or at side roads. (d) The bus stops that allow the simultaneous servicing of several buses during boarding/alighting operations; to reduce the boarding time, bus stops may have bus stands specified to direct the group (or queue) of passengers waiting for a particular bus within the convoy. Bus stops may be located in the middle of the block, before, or after the intersections, depending on which location is more convenient for the passengers; the length of a bus stop depends on the size and the number of buses to be simultaneously serviced; (e) The pedestrian crossings (their location may or not coincide with bus stops) that usually tend to be zebra crossings provided with traffic signals. Source: Lindau (1988) Figure 3. A typical busway The combination of these elements' allows for different bus performances along the busway. A good project would always rely on a proper evaluation of alternative designs prior to field implementation. In Brazil, busways were usually developed as "turn key" projects. However, some of them evolved further from their original concept. For example, the 9 de Julho busway in São Paulo has been redeveloped from a 71

76 two lane, two way busway into a three lane scheme. 3. Operation of a busway (a) Critical direction: city centre suburb In Brazil, big cities are usually surrounded by dormitory cities where low income workers live. As the inhabitants of dormitory cities tend to work in the centre of the metropolitan area, they undertake trips by public transport on every working day. To met the demand for such trips, busways were usually built along radial transport corridors linking the city centre to me suburbs. During morning peak hours, boarding passengers are dispersed among several bus stops located among the suburban roads, outside the busways. These passengers tend to alight at bus stops located inside the busway, close to the city centre. During the afternoon peak hours, the situation is reversed, i.e., passengers tend to board at bus stops close to the city centre (usually located inside the busway) and alight at bus stops near to the end of a bus route. As the passenger transfer capacity of bus stops and the time spent by buses at bus stops largely determine the capacity of a busway as a whole, and as the boarding of buses takes more time than alighting, afternoon peak hour traffic bound to the suburbs is critical for busway operation. Public and private operators exercise a strict control of fare collection. Fares may be paid inside the buses or through previously acquired tickets. In the particular case of Brazil, a large proportion of travellers use me vale transporte as the form of payment. Fare control is usually performed by a fare collector seated near the boarding door. A turnstile is usually used for recording the number of passengers and so as to prevent fare evasion. Therefore, each bus door has a single function: either for boarding or alighting passengers. The closer the turnstile is to the boarding door, the better the control is; on the other hand, turnstiles located near boarding doors increase passenger boarding times. (b) Trunk feeder system Radial bus lines bound to city centres collect and distribute passengers travelling from the suburban areas. Along suburban roads, buses usually run in mixed traffic conditions. The access to the city centre is normally provided by roads with several lanes in each direction. Bus stops along these roads are served by various bus lines and there is a heavy flow of buses, especially during peak hours. In order to reduce some of the inefficiencies generated by the heavy flows of competing buses, the so called trunk feeder services are introduced. In a trunk feeder system there are usually two terminals, one located at the city centre and the other located at the end of the busway. The one located at the end of the busway is used as the interchange terminal for buses arriving from the suburbs. A typical trip for a passenger using a trunk feeder system may consist of boarding a feeder bus (a minibus or a medium capacity bus) in the suburbs during the morning peak period, alighting at the interchange terminal, and then, transferring to a bus mat operates only on the busway (articulated or high capacity bus). In the afternoon, a suburban passenger leaving work boards the first high capacity bus departing from the central terminal or arriving at her or his bus stop along the busway and travels to the end of the busway where he or she transfers to the bus that will take her or him home. If overtaking is allowed at some points along the busway (usually at critical bus stops), then it is possible to operate different trunk services (see figure 2). For instance, trunk buses may serve a set of bus stops, and then return to the terminal. However, this type of operation requires a busway with at least three lanes provided at critical bus stops, and this may jeopardize the remaining road capacity for other vehicles such as cars, lorries etc. The main advantage of a trunk feeder system comes from speeding up the boarding movements for passengers travelling in the outbound direction during afternoon peak hours. This is the critical direction and passengers may use the first bus arriving at the bus stop. The main disadvantage of such a system derives from the delays imposed on passengers who need to interchange. (c) Ordered bus convoy system An alternative to trunk feeder system are the ordered bus convoy systems, such as COMONOR (Szasz, 1978). COMONOR tends to avoid the inconvenience caused to passengers by the need to change a bus whilst allowing significant benefits in bus performance (Lindau, 1987). 72

77 For example, let us assume that at the suburban end of a busway bus lines can be divided between those whose origin/destination points are three different neighbourhoods, the first located in the north, me second in the east and the third in the west. Bus lines going to the north are named "A"; bus lines going to east "B" and those going to the west as "C". Given that me outbound direction is the critical one, let us further assume that, at some point close to me beginning of me busway, a bus ordering station is introduced, as in figure 3. The objective of a bus ordering station is to arrange buses into convoys before they enter me busway. According to our example, a convoy with six buses would be formed as: A A B B C C. Initially, the two buses named "A" would depart; then, the two buses named "B" and, finally, the two "C" buses. When arriving at me bus stops, buses "A" stop at me first two bays, men "B's" and "C's" at the remaining two stands (see figure 1). As buses use pre defined stands, all passengers wishing to travel, for instance, to me northern neighbourhood, wait near bus stand A. Bus operation in ordered convoys enables significant benefits to be achieved by reducing overall boarding time and thus increasing the bus stop capacity. However, there is still very limited understanding of traffic behaviour at heavily used bus stops. The Highway Capacity Manual (National Research Council, 1985) summarizes previous research efforts conducted in the United States by developing analytical equations to determine the capacity at near side bus stops as a function of passengers boarding and traffic signal settings. The model incorporates factors to compensate for dwell time and arrival variations, as well as one which represents me cumulative efficiency resulting from me varying number of possible bus stands. Studies conducted in São Paulo (Companhia de Engenharia de Trafego, 1979) established a few simple relationships where bus flow capacity at critical bus stops is expressed in terms of me total number of passengers boarding buses. Figure 4 attempts to summarize the existing knowledge in the field. Different bands are constructed based on the American and the Brazilian criteria. For the sake of comparison, data from field and theoretical studies conducted in São Paulo were fitted to all equations; for details concerning equations and parameters see Lindau (1987). The results produced by the two lower bands indicate that, under usual bus boarding conditions, it is difficult to expect bus stop capacity to be above 300 buses per hour in the case of 1000 or more passengers boarding per hour; note that the upper limit of the São Paulo band (before introducing convoys) is determined by chaotic peak hour boarding conditions. On me other hand, one of the virtues of figure 4 is to show the potential improvements to be achieved if buses travel and stop in ordered groups or convoys. Gardner et al (1991) concluded that the capacity of a busway operated by bus convoys (Assis Brasil in Porto Alegre) is equal to that attained by a trunk feeder system running on three lane busways in the Avenue 9 de Julho in São Paulo. 73

78 Source: Lindau and Willumsen (1988) Figure 4. Bus stop capacity It is relatively easy to operate buses in ordered convoys on a busway. Nevertheless, particular attention must be given to the analysis of conditions that justify such operation. For instance, the assembling of convoys involves initial delays mat may make it reasonable to run buses in a non ordered way during off peak periods. Other aspects to be carefully studied at the busway design stage involve me way in which the convoys are released and the practical limits to me size of the convoy (Lindau, 1988). Most of problems in me operation of busways are related to difficulties at the institutional level. For instance, trunk feeder systems should be operated by a single bus company, or, at least, by a pool of several bus companies operating in unison. Thus, one of the very early tasks of the public sector should be directed towards the amalgamation of existing companies. In the case of Porto Alegre, for example, me disagreement between private operators and public authorities in this regard caused the need for major modifications of a busway that was almost ready for operation. 4. Costs and benefits of busway development The analysis of busway costs and benefits is a difficult task in Brazil because of the lack of data. With rare exceptions, local authorities do not collect data on busway performance on a regular basis, and even information on implementation and maintenance costs is either difficult to obtain or not available at all. There is little evidence to conclude about public satisfaction. However, there are some items that are universally accepted such as users' benefits resulting from reductions in travel time and travel time variability. Indeed, field experience with Brazilian busway has demonstrated their potential to reduce travel times and accommodate a high flow of buses. (a) Costs of developing busways It is almost impossible to assess a posteriori the costs incurred by the development of busways in Brazil. Almost all busways were introduced along existing avenues by taking their median traffic lanes. In some cases, parallel roads had to upgraded so as to compensate the general traffic for the loss of travelway space allocated to the busway. Also, almost everywhere new traffic signal systems were installed and several intersections were redesigned. Intense bus operation requires considerable strengthening of the road construction. Experience shows that busways need specially designed pre stressed concrete carriageways because a bitumen pavement cannot sustain the load of heavy buses. Funds for the above improvements 74

79 came from various sources and at different periods of time, and due to the high inflation rate in Brazil, they are difficult to summarize in constant prices. The development of busways was always commissioned by local or state authorities, the latter being involved when a busway scheme extended over a municipality's limits and served a metropolitan region. The EBTU also participated in funding the projects. In fact, it played a key role in the development of busways. The EBTU staff and consultants were effective in promoting busway schemes and the organization helped a lot setting the agreements between local, state and federal agencies concerning financing of busways. The EBTU negotiated also the international funding of projects with IBRD. On the other hand, private bus operators did not participate in busway financing. (b) Impacts on traffic and urban development In general, busways helped organize the chaotic interface between the high flow of buses often exceeding 300 buses per hour and direction and other vehicles on arteries with three to four traffic lanes in each direction. Again, it is difficult to separate the impact of the busway alone on traffic conditions from the impact of all other interventions made on a transport corridor, such as the reconstruction of intersections or new traffic control systems. Some studies (Lindau, 1983) indicate that for the type of busway schemes designed for Porto Alegre, the sole introduction of median bus lanes would not significantly improve bus operation. No advantage in terms of system capacity and operational speed of buses would be derived if operational modification such as bus convoys, for example, were not introduced. Other traffic did not always benefit from separation from bus flow. With less travelway space available, congestion often increased and travel time rose. The provision of busways sometimes also required a reduction of the width of sidewalks. Little can be said on the impact of busways on land use development because of the lack of relevant studies. Perhaps, one of the few exceptions was a short study conducted in Porto Alegre by PROPUR/UFRGS in It focused on the impacts of the Farrapos and Assis Brasil busways, but covered only a short period after the beginning of busway operation. The study concluded that the implementation of both schemes was a substantial factor contributing to changes in the pattern of growth of different commercial activities. To our knowledge there was no study on the impact of the Erico Verissimo corridor on land development in Porto Alegre although the busway was built with the objective of influencing the land use pattern. Curitiba displays, however, a fine example of how transport and city can be developed in unison (Fouracre, 1975). Already in me 1960s, mass transit became an integral part of a comprehensive urban development strategy and soon started to be seen as a factor contributing considerably to land use characteristics. The development of the busway network enabled enhancement of urban growth along busway corridors, yielding higher population densities. The transfer terminals between feeder lines and busways encouraged the location of commercial and service enterprises, leading to the formation of mini centres away from the downtown area. With the support of firm land use policies, the balance in the distribution of housing, services and work places has been considerably improved. Only 45 per cent of me passengers who use busways are actually headed for the downtown area, while others meet their needs travelling on shorter distances (Lerner, 1991). REFERENCES ANTP (1991). Imagem dos serviços de transporte coletivo na grande São Paulo (São Paulo, Associaçao Nacional de Transportes Publicos). Armstrong Wright, A. (1986). Urban Transit Systems: Guidelines to Examining Options, World Bank Technical Paper 52 (Washington, D.C., The World Bank) Barat, J. (1989). "Rio de Janeiro mass transportation system: the role played by metro lines", in: Gutman, J.S., and Scurfield, R.G. (eds.), Rail Mass Transit (London, Thomas Telford), pp Belda, R. (1989). "Les raisons du succes du metro de São Paulo", in: Gutman, J.S. and Scurfield, R.G. (eds.), Rail Mass Transit (London, Thomas Telford), pp Boletin de los Transportes Publicos de la America Latina (1992). "Editorial: El lugar de los sistemas de media capacidad en las ciudades latinoamericanas" Associação Nacional de Transportes Publicos (São Paulo, ANTP), Año 1, no. 2, octubre/diciembre. 75

80 Branco, O.C. (1985). "Uma nova politica semaforica para corredores exclusivos de onibus", paper presented at me 5 th Congresso Nacional de Transportes Publicos, June. Associação Nacional de Transportes Publicos, Belo Horizonte, Brazil. Brasileiro, A.B.D. (1992). "Dinamicas Urbanas e evoluçao da oferta de transporte coletivo no Brasil", Proceedings of VI ANPET, Rio de Janeiro, Brazil. Companhia de Engenharia de Trafego (1979). "COMONOR Coordinated bus convoy", Proceedings of PTRC Summer Annual Meeting, July, Seminar G (University of Warwick), pp Department of Transport and Institution of Highways and Transportation (1987). Roads and Traffic in Urban Areas. Crown Copyright, U.K. EBTU (1982). Tratamento Preferencial ao Transporte Coletivo por Onibus (Brasília, Ministerio dos Transportes). EBTU (1983). Relatorio de conclusão do projeto EBTU BIRD (Brasília, Ministerio dos Transportes). EBTU (1988). Plano Nacional de Transporte de Massa (Brasília). Fagnani, E., and Cadaval, M. (1988). "A politica federal de transportes publicos nos años 80", Revista dos Transportes Publicos (São Paulo, ANTP), no. 42, año 10, pp , dezembro. Fouracre, P.R. (1975). The Development of Public Transport in Curitiba, Brazil, SR 197 (Crowthorne, Transport and Road Research Laboratory). Gibson, J., Baeza, I., and Beckett, R. (1989). "IRENE: un programa de simulación de paraderos", Actas del IV Congreso Latinoamericano de Transporte Público y Urbano, Habana, Cuba. Gardner, G., Cornwell, P.R., and Cracknell, J.A. (1991). The performance of busway transit in developing cities, Research Report 329 (Crowthorne, Transport and Road Research Laboratory). Gonçalves, A., Orrico Filho, R., and Simoes, R. (1993). "Implementation of LRT's in Brazil", Proceeding of VI Conference on Urban Transport in Developing Countries, Association CODATU. Janes Urban Transport Systems (1989) (London, Jane's Publishing Company Limited). Kato, S. (1986). "Linha tronco em canaleta central: corredor de trolleybus Santo Amaro 9 de julho", paper presented at the 4 th Congresso Panamericano de Ingenieria de Transito y Transporte, Sociedad Chilena de Ingenieria de Transporte, Santiago, Chile. Lima, I.M. (1989). "As resposabilidades do governo federal nos transportes urbanos na decada de 90", Revista dos Transportes Publicos, (São Paulo, ANTP), no. 44, año 11, junho. Lindau, L.A. (1983). "High flow bus operation on urban arterial roads", PhD thesis, Faculty of Engineering and Applied Science. University of Southampton. Lindau, L.A. (1985). "Criterios para la implementación de pistas solo bus en arterias urbanas de alto flujo", Actas del II Congreso Chileno de Ingenieria de Transporte (Santiago de Chile, Sociedad Chilena de Ingenieria de Transporte). Lindau, L.A. (1987). "Bus priority systems in Brazil: from theory to practice", Proceedings PTRC Summer Annual Meeting, Bath, U.K. Seminar J, pp Lindau, L.A., and Willumsen, L.G. (1988). "How far can bus capacity be stretched?" Proceedings IV Conference on Urban Transport in Developing Countries, Association CODATU. Lindau, L.A. (1988). "BUSWAY", unpublished technical paper. 76

81 Lindau, L.A., and Rosado, A.B.A. (1989). "The metropolitan railway transit system of Porto Alegre (TRENSURB): forecasts and reality", in Gutman, J.S., and Scurfield, R.G. (eds.) Rail Mass Transit (London, Thomas Telford), pp Lindau, L.A. (1991). "Bus priority systems in Brazil: from theory to practice", in Heraty, M., Urban Transport in Developing Countries: Lessons in Innovation (London, PTRC Education and Research Services), pp Lindau, L.A. (1992). "Sistemas de transporte urbano de media capacidade: uma analise comparativa enfocando as tecnologias VLT e onibus". Proceedings of VI ANPET, Rio de Janeiro. Lindau, L.A., Antunes, J.A.V., and Lima, L.A. (1993). "Applying basic quality and productivity techniques to a state owned bus transport operator", Proceedings of VI Conference on Urban Transport in Developing Countries, Association CODATU. Ministerio dos Transportes (1985). Subsidio ao Transporte Coletivo Urbano: Alternativas de Aplicação (Brasilia, DF). Mandon, B. (1991). "Enjeux du développement du secteur privé dans les transports urbains le cas de São Paulo (Brésil)", Memoire pour le Diplome d'etudes Approfondies d'economie du Développement, Université de Paris I Pantheon Sorbonne. U.E.R Economique 07. MVMA, (1990). World Motor Vehicle Data (1990 edition) (Detroit, Motor Vehicle Manufacturers Association of the United States), p. 38 National Research Council (1985). Highway Capacity Manual, Transportation Research Board, Special Report 209 (Washington, D.C.). Pereira, A.L.S. (1980). Projeto Brigadeiro: Faixa Exclusiva de Onibus no Contra fluxo, Boletim tecnico 25 (São Paulo, Companhia de Engenharia de Tráfego). PROPUR/UFRGS (1981). Transporte e Uso do Solo o Caso da Região Metropolitana de Porto Alegre (Programa de Posgraduacoao em Urbanismo, Universidade Federal do Rio Grande do Sul). Revista Ferroviaria (1991) (Rio de Janeiro, Empresa Jornalistica de Transportes), año 52, no. 10, outubro. Revista Ferroviaria (1992) (Rio de Janeiro, Empresa Jornalistica de Transportes), año. 53, no. 4, abril. Robertson and Vincent, R.A. (1975). Bus Priority in a Network of Fixed time Signals, TRRL Research Report LR 666 (Crowthorne, UK). Santin, E. (1989). "Trilhos em Goiânia", Revista a Construcaçao (Rio de Janeiro), no. 280 (fevereiro) pp Seção Transporte de Massa (1986). "Metro sobre rodas serve a periferia", Transporte Moderno, vol. 23, No. 10, pp Secretaria Municipal dos Transportes (1980). Estudo Complementar de Operaçao NOS Corredores das Avenidas Farrapos e Assis Brasil, Prefeitura Municipal de Porto Alegre (Porto Alegre). Severo, C. (1991). "Sugestôes para a formulação da politica de transporte de passageiros no Brasil onibus urbano", Revista dos Transportes Publicos (São Paulo, ANTP), no. 51, año 13, marco. Szasz, P.A. (1978). COMONOR: Comboio de Onibus Ordenado, Boletim Técnico 9 (São Paulo, Companhia de Engenharia de Tráfego). Tyler, N. A. (1991). The contribution of expert opinion to the design of high capacity bus systems. Computers and Structures, 40 (1), UNEP, WHO, (1988). Assessment of Urban Air Quality, Global Environment Monitoring System, United Nations Environment Programme, World Health Organisation, pp. 46 and 56. Via Urbana (1991), Año 1, no.2, novembro, Empres Journalistica dos Transportes, Rio de Janeiro. 77

82 Wright, C.L., and Santana, J.A. (1989). "Problemas e perspectivas dos transportes urbanos na decada de 90", Revista dos Transportes Publicos, (São Paulo, ANTP), no. 44, año 11, pp , junho. ENDNOTES 1. The paper is an edited version of a manuscript contributed by Dr. L. A. Lidau and Dr. L. A. DOS Santos Senna, Escola de Engenharia, Universidade Federal do Rio Grande do Sul, Brazil. 2. Urban fare expenditures are limited by the Constitution to 6 per cent of the wage a worker earns. 3. A recent study (Lindau et al, 1993) has shown that higher reliability levels can be achieved by introducing new company wide quality control techniques. 4. The traditional trams operating in mixed traffic conditions, such as the ones so popular in Brazil in the past and still in use in East European countries, are also pan of the LRT family. 5. In terms of city division, the introduction of at grade median bus lanes causes little interference. 6. This measure does not impose physical restriction on other vehicles that may share the bus or tram lane during off peak periods. 7. Probably the last plan formulated before the extinction of EBTU; the main aspects of PNTM are presented in EBTU (1988); more details about PNTM are presented in following sections. 8. The CMTC fleet also includes trolleybuses and double deck buses. 9. These buses are locally known as Romeo and Juliet, and have a higher vehicle capacity than conventional buses. IV. CASE STUDY: PROVISION OF SEPARATED TRAVELWAYS FOR PUBLIC TRANSPORT IN METRO MANILA, THE PHILIPPINES 1 A. Urban development and transport 1. Socio economic profile Metro Manila is comprised of the City of Manila, three other cities and 13 municipalities and covers an area of 636 sq km within the central part of the Philippines' main island of Luzon (Figure 1). With an estimated population of over 7.9 million in 1990, Metro Manila accounts for 13 per cent of the total population of the Philippines. Its share in me generation of gross domestic product has increased from about 23 per cent in 1948 to 31 per cent in Despite me decentralization and regional dispersal policies of the Government, Metro Manila still dominates the economy and socio political activities in the country. As me National Capital Region, it is the seat of the National Government and the premier centre of business and trade. Metro Manila has most of the institutions of higher education, the main health facilities and banking establishments, and a major share in the national cultural events. It remains the popular destination of tourists and has emerged as an international convention centre. The concentration of people, economic activities and employment opportunities in Metro Manila has brought about complex urban problems, notably of housing and of an inefficient urban transport system. 2. Characteristics of urban transport (a) Overall transport demand 78

83 The demand for urban transport in Metro Manila was estimated at about 17.6 million person trips per day in Its phenomenal growth from 11 million trips in 1980 is attributed to the strong links of the metropolis with the rapidly developing fringe areas and the number of commuting workers and students. About 70 per cent of the transport demand is served by public transport modes (see table 1), with the ubiquitous jeepneys as me predominant mode. Bus and railway services, including the recently introduced light rail transit system (LRT) along two major traffic corridors, account for 34 per cent and 3 per cent of the public transport demand, respectively. The share of private vehicles has increased significantly from 26 per cent in 1980 to 30 per cent in

84 80

85 81

86 Figure 1. Map of the Metro Manila SOURCE: DOTC, 1989 Another indicator of transport demand is the number and type of road vehicles in use (see table 2). From 1980 to 1991, the number of motor vehicles registered in Metro Manila increased at a rate of 3.1 per cent per annum. However, since 1985, the total vehicle registration has been growing at an annual rate of almost 10 per cent. Table 1. Transport demand by modes Daily person trips (in millions) Modal share (percentage) Private vehicle Jeepney Bus 0.1 Commuter train Light rail Total Source: Department of Transportation and Communications (b) Road network, traffic flows and car ownership The road network of Metro Manila consists of about 4820 km of roads, including 1827 km of private roads. The major arterial roads follow a semi circular and radial pattern, but only three circumferential roads have been completed, the most important of which is C 4 or EDSA linking 10 radial routes (see figure 2). Total registered vehicles Table 2. Motor vehicle registration Philippines Manila Metro Manila Vehicle Share (percentage) Private car/taxi Jeepney Pick up/van Private bus Truck/trailer Motorcycle/tricycle Total Source: Land Transportation Office, DOTC 82

87 83

88 84

89 Figure 2. Metro Manila road network SOURCE: L.E.B.C. METRO MANILA CAMPS CONVERSION PROGRAM, PROS 1990 Many of the road sections have reached their capacity limits, particularly those within EDSA (see figure 3). It is estimated that nearly 40 per cent of the arterial roads are heavily congested, specifically during peak periods. Due to congestion, the average traffic speed is about 18 km/h, with jeepneys and buses having lower operating speed due to frequent service stops. Recent traffic counts indicate that private cars are the major cause of congestion on most Metro Manila roads. Traffic congestion is expected to worsen further since the pent up demand for cars is not matched by expansion of the capacity of major roads. SOURCE: TEC Figure 3. Traffic flows Car ownership and use have grown rapidly since the end of economic recession in In Metro Manila alone, where about 42 per cent of the national motor vehicle fleet and two thirds of all cars are concentrated, per capita vehicle ownership is more than three times the national average. The number of private cars is increasing at a rate of 13 per cent per annum. The average car occupancy was about 2.2 passengers in

90 (c) Public transport services The public transport services in Metro Manila are predominantly road based and provided largely by jeepneys and buses on primary and secondary routes, and by tricycles and pedicabs on feeder routes. Since the highly urbanized areas extend to neighbouring towns, short distance provincial bus transport also forms an integral part of Metro Manila's public transport system. Jeepney routes cover more than 610 km of Metro Manila roads, while buses operate only on about 350 km of the major corridors, with nearly 300 km being jointly serviced by jeepneys and buses. Jeepney routes are generally short and concentrated in radial roads, while bus routes typically passing through circumferential roads, such as EDSA are twice as long. Jeepneys are banned along most parts of EDSA, Roxas Boulevard and South Superhighway, which are designed to operate as high speed roads. The road based public transport is characterized by a low level of service in terms of travel speed and passenger waiting time at stops, and by the operation of dilapidated, pollution emitting vehicles, despite the government sponsored infusion of new buses in Table 3 provides the operational characteristics of jeepney and bus services. Urban rail services, currently limited in coverage and patronage, are provided by commuter trains of the Philippine National Railways and the LRT Line No. 1 along Rizal and Taft Avenues. The detail characteristics of the LRT is given in section B. (d) Organization of public transport Public transport in Metro Manila is mainly owned and operated by the private sector. Private operators, who lease their vehicles to drivers on a daily basis, own a majority of 38,000 franchised jeepneys, the rest being in the hands of transport cooperatives. Bus companies, which are mainly family enterprises, own and operate about 3100 buses. The minimum fleet size of such a company is 10 units. Number of routes Table 3. Jeepney and bus operations, 1991 Indicator Jeepney Bus < 15 km > 15 km Total Vehicle capacity (passengers) Commercial speed (km/h) Passenger trip length (km) Passenger waiting time at stops (minutes) Source: DOTC, Metro Manila Urban Transport Development Plan Project The public sector is represented by services provided by the Philippine National Railways, Light Rail Transit Authority and Metro Manila Transit Corporation. The last of these operates 300 buses and is soon to be privatized. The Light Rail Transit Authority, itself attached to Department of Transportation and Communications, operates and maintains the existing single, 15 km long, light rail line through its subsidiary Metro Inc. on a contract basis. The Department of Transportation and Communication (DOTC) an organ of the National Government is responsible for the formulation of transport policies and for the planning and development of transport infrastructure. The Land Transportation Franchising and Regulatory Board of DOTC is a regulatory agency for road based public transport, and it sets, inter alia, fare regulations. The Office of Transportation Planning Service within the DOTC coordinates the services provided by various public transport modes. It advises the Secretary for Transportation and Communications, the Land Transportation Franchising and Regulatory Board, the Philippine National Railways and me Light Rail Transit Authority on transport policies and planning matters. There is also the Investment Coordinating Committee, at Cabinet level, which sets priorities for transport investments and evaluates project proposals. The Committee is composed of top officials of the Office of the 86

91 President, National Economic and Development Authority and the Central Bank of the Philippines, as well as me Departments of Finance, Budget and Management, Transportation and Communications, and Public Works and Highways. 3. Improvements of public transport system Since the 1960s, strategic transport plans for Metro Manila called for the introduction of an integrated network of mass transit systems, with rail based systems providing the trunk line services and buses used as feeders to link fringe areas with rail transport stations. Efforts were also made to replace jeepneys by high capacity buses. Development policies focused on public transport and control of private vehicle use; however, the Government could not muster the political will to implement such policies. In spite of me above setbacks and the overriding constraint on the Government to finance the development of major transit systems throughout those years, two successful undertakings have been launched to demonstrate the cost effectiveness of public transport segregation, i.e. provision of separated travelways. These are the development of the LRT system and the introduction of bus lanes along EDSA. B. Development of LRT system 1. Project background In 1977, the most comprehensive urban development and transport study on Metro Manila, commonly known as MMETROPLAN, strongly recommended the construction of an at grade LRT system for the metropolis, consisting of a central area network and four lines radiating along Rizal Avenue, Espana/Quezon Avenue, Shaw Boulevard and Taft Avenue (see figure 4). The recommendation was made after a critical review of various proposals, including, inter alia, monorail and subway systems. Through exhaustive deliberations by me Investment Coordination Committee and the later initiatives of the newly created Ministry of Transportation and Communications (now called Department), the final blueprint of the first LRT line was developed in This 15 km long line on a north south orientation runs along the second most heavily travelled route in Metro Manila. Studies indicated that this line was the second best on economic grounds, but was chosen over the EDSA line for social considerations, since most of the passengers along Taft and Rizal Avenues belonged to the lower income group in need of public transport. It was also decided to elevate the entire line so as to avoid not only interference from surface traffic, but also the expensive acquisition of land and properties that would be otherwise necessary for the capacity of roads on these corridors to be maintained. 87

92 Figure 4. Proposed network of light rail transit In 1980, the Philippines and Belgium formalized an agreement granting capital assistance to finance the construction of the elevated line running from Baclaran in Pasay City to Monumento in Caloocan City (see 88

93 figure 5). The total project cost, including financial charges, was estimated at P2.5 billion ($US300 million at the 1980 exchange rate). The decision to implement this project triggered the creation of the Light Rail Transit Authority (LRTA) which assumed overall responsibility for the construction, operation, maintenance and further expansion of the LRT system in Metro Manila. A turn key contract was awarded to a Filipino Belgian group with a Belgian consortium being held responsible for providing the vehicles, track works and electro mechanical equipment, while their Filipino counterparts undertook the design of civil works and construction. Hard hit by the economic recession at the beginning of 1981 and by the inability of the Government to provide the required local counterpart funds (almost 60 per cent of the estimated project cost) for all civil works, utilities, right of way acquisition, and interest charges, construction was delayed by about 10 months with regard to the original contract. The south section (Taft Avenue) became operational on 1 December 1984, and the remaining section, extending across the Pasig River to the north, was opened for full line operation on 13 May To ensure efficiency and to prolong the life of the system, a 10 year management contract was concluded with a private entity: MERALCO Transit Organization, Inc. (METRO). However, following an adverse audit opinion of this arrangement, the contract was declared null and void. METRO became a wholly owned subsidiary of LRTA in July Figure 5. LRT line No. 1 89

94 2. System specifications LRT line No. 1 is fitted with standard gauge track (1435 nun) with flat bottom rails elastically fastened on twin block, reinforced concrete sleepers. Track structures supported by single columns are about 7 metres above street level (see figure 6). The minimum curve radius is 170 m in the main line and 25 m in the depot area. The maximum grade is held at 4 per cent. Each of the existing 64 light rail vehicles (see figure 7) consists of double articulated cars with eight axles, with two motor trucks (one at each end) and two trailer trucks under the articulations. The main characteristics of the cars are indicated below. Body length Total width Floor height Parameter Height, rail to roof Wheel diameter Number of double doors Width of doors Empty weight Table 4. Light rail vehicle specifications Value m 2.50 m 0.90 m 3.27 m 0.66 m 5 per side 1.50 m 41 tons Number of seats 81 Standing passengers 293 Total car capacity 374 Maximum speed, full load Maximum acceleration, full load Ave. service braking, full load 60 km/h 1.0 m/s2 1.3 m/s2 Emergency braking, full load 2.1 m/s2 Source: Light Rail Transit Authority There are a total of 15 stations and three terminals located either at major intersections or in commercial areas. The average distance between stations is 825 metres. Each station has two side platforms 100 m long, 3.5 m wide and elevated 80 cm above the track level. Stations have token operated reversible turnstiles at platform level. The electrical power required for me traction substations is supplied by the MERALCO 34.5 kv network. There are eight substations on the main line and one in the depot area at an average spacing of 1.85 km. Each substation contains two air cooled silicon rectifiers which convert the power supply to the, required 750 volts direct current. 90

95 Figure 6. LRT track structure Automatic light signals are used at each block section. In operation, two trains are always separated by one block section. The minimum theoretical headway is about 85 sec for the automatic block system, although the system operates with a train interval of 132 sec. At the three terminal stations (north, central and south), there are devices for centralized signalling with route control. A mimic diagram and control desk are located in me signal operator's room. The communication system is composed of a 240 line electronic private automatic branch exchange (EPABX) and UHF radio network for communication with train drivers. A public address system is installed in all stations and trains. Figure 7. LRT vehicle Repair and maintenance facilities are located next to the depot, close to the Baclaran terminal. They are designed for preventive maintenance and major repairs. Ten stabling tracks, each capable of accommodating 91

96 three trains (two car configuration) are also located in the depot area. Fare collection is designed for a flat fare operation of the system. Turnstiles at the entrance to the platform can be moved after dropping a token. Tokens can be bought at the stations or at authorized selling points. 3. Cost of the system The cost of the development of the line was estimated in 1980 at P2.5 billion, while the actual cost amounted in 1986 to almost P3.4 billion (about $US170 million at the 1986 exchange rate). Table 5 compares the 1980 cost estimate with the actual costs as they accumulated up to It was the devaluation of Philippine currency which mainly contributed to the cost overrun. (in million pesos) Particulars Table 5. Comparative project cost estimates 1980 and cost estimate 1986 actual cost Variance (percentage) Civil works Electro mechanical works Pre operating other costs Land acquisition/relocation costs Total Source: Light Rail Transit Authority 4. System performance (a) Operations As far as operations are concerned, the ridership on LRT has reached the volumes estimated in the project. LRT enjoys high utilization rates, with average daily patronage increasing from 191,400 in 1985 to 351,200 in 1990, at an annual growth rate of 12.9 per cent (see table 6). Since operation started in December 1984, about 860 million passengers have benefited from the system. It is estimated that about 60 per cent of the total passenger traffic use the LRT system along the Taft Rizal corridor. The zonal two fare system, prescribing the use of token coins at the first 12 stations north or southbound (Zones 1 and 2) and of PI coins at the last six stations (reduced to five in November 1990) to end terminals, was adopted starting 12 October In 1990, about 89 per cent of passengers boarded from Zones 1 and 2 using tokens, and the rest originated from Zone 3 using the PI coins. The current fare is P6 ($US0.24) for Zones 1 and 2. (b) Financial situation The financial performance has remained bleak since the beginning of commercial operation. With revenues being obtained in local currency and the outstanding loans denominated in foreign currency, and with the Government's equity contribution (mainly for the construction works) not released, me annual financial statements show negative bottom lines. The LRTA's operating margin from fare box revenues has been enough to cover direct operating expenses in the years from 1984 to In 1991, LRTA generated a revenue (net of refund and off station sales discount) of P513.3 million, while operating expenses totalled P336.3 million, thus the fare box ratio was 1.53 (1991 foreign exchange rate was $US1 = P26.6). However, LRTA still incurred a net loss of P567.4 million, representing interest expenses and foreign exchange losses. As of end December 1990, LRTA's total assets, liabilities, and capital, stood at P6.17 billion, P5.63 billion, and P543.2 million, respectively. Its debt to equity ratio rose from 4.8:1 in 1989 to 10.4:1 in Month Table 6. LRT ridership (in million passengers) Years January February

97 Much April May June July August September , October November December Annually Cum. Total Source: Light Rail Transit Authority 5. Project impact (a) Impact on traffic and public transport The introduction of the LRT system on one of Metro Manila's most travelled corridors has resulted in drastic changes in public transport operations. While most of the bus and jeepney routes around the LRT line have not changed much, despite earlier plans to reroute jeepneys to parallel streets, the composition of the traffic stream has been significantly altered. More specifically, there has been a marked decrease in the number of buses and jeepneys operating along and around the LRT route. Comparative "before LRT" (1979) and "after LRT" (1987) traffic data indicated the following: Reduction of about 45 per cent and just over 20 per cent of vehicle traffic along the southern and northern sections of the LRT Line No. 1, respectively; Drop of 55 per cent and 35 per cent in the number of buses and jeepneys, respectively; Increase in the number of private vehicles. These observations point to the significant diversion of passengers from the road based public transport to the LRT. Notably, the LRT line upgraded the quality and expanded the capacity of public transport along its route, which would be unthinkable if buses and jeepneys were mainly relied upon to transport the daily commuters. Peak hour travel times on the corridor improved from one and a half hours to around 30 minutes by LRT and one hour by jeepneys, respectively, by Table 7 compares the average travel times of the LRT and jeepneys along the 15 km route in Immediately prior to the full operation of the LRT, a comprehensive plan to restructure existing bus and jeepney routes was adopted. This was intended to minimize the expected adverse impact of the bus and jeepney operations and to ensure the revenues for the LRT. The plan involved the rerouting of jeepneys and buses to secondary roads and the creation of new routes to feed into the LRT stations. Eventually, the plan was never fully implemented, except for the newly opened routes. Many jeepney operators arid a number of bus operators transferred to some of the more lucrative new routes. The rest of the operators continued their services along the LRT routes to serve in between station locations and short distance passengers. Weekdays: Table 7. Comparative average travel times, 1990 (hours) Morning peak Afternoon peak LRT Jeepney Holiday and Sundays: LRT

98 Jeepney Source: LRTA (b) Impact on urban development The LRT has played a key role in shaping the land use development of Metro Manila. It has triggered the redevelopment of me decaying traditional centres of business and trade within the City of Manila. With its fixed route and concentrated generation of traffic, me LRT line has caused me commercialization of areas around me stations. Moreover, the economic losses experienced by businesses located between stations have been reversed and land values have continued to appreciate. 6. Extension of the system At me beginning of 1993, the Government approved me extension of the LRT system in Metro Manila. Line No. 2 (I phase) will connect the central area of Manila with Cubao district. An 11 km long line will cost about $US300 million financed by an OECF loan. It will be implemented by LRTA. Line No. 3 will run along EDSA highway and link Harrison, Pasay City, to North Avenue, Quezon City. The cost of this 18 km long line is estimated at $US850 million. A contract has been approved for a private consortium to implement me project under a Build Lease Transfer arrangement. DOTC is designated to operate the line, possibly under a management contract to a private firm. Both lines. No. 2 and No. 3, should be operational by the end of C. EDSA bus lanes 1. Project background In 1989, in response to the worsening transport crisis (i.e., a chronic shortage of buses) and the serious traffic congestion along the major arteries of Metro Manila, transport authorities identified various approaches and techniques to improve the flow of traffic on these corridors with special emphasis on the priority treatment of public transport modes. By the end of 1989, the Metro Manila Authority (MMA), upon the recommendation of transport related agencies, passed a metro wide ordinance establishing the bus lane scheme along EDSA (see figure 8). The introduction of bus lanes formed a part of the joint government/private sector collaboration, which was primarily aimed at improving traffic conditions along EDSA. EDSA is by far the dominant component of the urban transport system of Metro Manila. It is a 12 lane, limited access highway extending some 22 kilometres from me McArthur Highway at Monumento in the north to Taft Avenue in the south. Virtually enclosing the main part of the metropolis, EDSA was built as a semi circular route intersecting with seven of me main radial routes and crossing the boundaries of four cities and three municipalities. Its importance stemmed from me development of major activity centres in its proximity, notably: die Makati commercial area as the business and financial centre, not only of Metro Manila but of the entire country, and the recently established mega commercial complexes at Ortigas, North Avenue and Monumento. The future expansion and development of Metro Manila is firmly based on maintaining the transport efficiency of the EDSA spine. 2. Bus lane scheme description The bus lanes consist of dual traffic lanes adjacent to the road curbside on both sides of EDSA. These bus lanes are delineated from other traffic lanes by a continuous yellow line and appropriate signs and pavement markings. Under the ordinance, all vehicles other than buses, except emergency vehicles, are prohibited from entering or travelling on me bus lanes. Jeepneys, which are generally banned on EDSA, could use the bus lanes if their routes cross or pass a short section of the artery. In the case of private vehicles entering frontage properties or turning right at intersections, a broken white line is painted alongside the yellow line to indicate to the drivers that they may cross the yellow line in preparation for right hand turns. Likewise, buses are allowed to leave the bus lanes to pass traffic obstructions such as stalled or parked vehicles and road diggings. To complement the bus lane initiative, the police was put to task of strictly enforcing the bus lane ordinance. The presence of traffic law enforcers and the relatively high penalties for violations (minimum of P200 or $US8) were expected to result in a satisfactory level of adherence to the scheme. 94

99 The general location of the bus lanes, a typical cross section of the EDSA travelway, and frontage land uses along the bus lanes are presented in figure 9. The travelway of EDSA is divided along its entire length by a median strip and provides four lanes available for general traffic and two lanes for buses. At grade separated intersections, particularly those with underpasses as on Aurora and Shaw Boulevards, three lanes through the underpass structures are typically provided in each direction. Buses are normally required to use the service roads on each side of the underpass. Land use activities along the bus lanes vary considerably, ranging from low density residential areas with no direct connection to EDSA, to commercial establishments with access to EDSA. 3. EDSA traffic pattern Vehicle traffic volumes on EDSA (see figure 10) have been growing at the rate of 5 per cent per annum (from 248,400 in 1985 to 317,000 vehicles in 1990), which closely relates to the growth of vehicle ownership in Metro Manila. The road has a very high capacity and, for a time, could still accommodate further increases in traffic. However, me daily traffic now exceeds 140,000 vehicles on its busiest section near the Guadalupe area. Traffic congestion has become quite regular and severe at several sections, particularly close to Makati during peak hours. This situation led to the decision by the Government to construct new grade separation structures at troublesome intersections. Three more are to be completed by the end of

100 96

101 97

102 Figure 8. EDSA bus lanes locality plan and study route Passenger traffic (see figure 11) grew from 1.5 million to 2.3 million passengers per day during the same period. Overall passenger movement has increased at an annual rate of about 11 per cent, with public transport ridership increasing at a rate close to 9 per cent per annum ( ). 98

103 99

104 100

105 Figure 9. Bus lane/frontage land use interaction and typical EDSA cross section Vehicle and passenger data for 1990 indicated that public transport vehicles, while constituting about 8.5 per cent of the total vehicle flow, carried 63.2 per cent of the passengers. Average occupancy figures at selected points of EDSA were around 61 passengers per bus, 14 passengers per jeepney and 3 persons per private vehicle. Equivalent figures in 1985 for buses, jeepneys and private vehicles were 47, 13 and 2.2 passengers per vehicle, respectively. In terms of passenger throughput, the above figures highlight the importance of providing priority measures, such as bus lanes, for public transport vehicles. Figure 10. EDSA vehicular traffic growth, 1980 to 1990 ROOSEVELT AVENUE SCREENLINE STATION 101

106 GUADALUPE BRIDGE SCREENLINE STATION SOUTH SUPER HIGHWAY SCREENLINE STATION 102

107 ALL STATIONS COMBINED 4. Performance of the scheme (a) General comment The bus lane scheme was suspended from September 1990 to November 1992 in view of the traffic rerouting associated with the construction of grade separation structures (fly overs) at five major intersections of EDSA. The assessment of bus lane performance covers the period prior to its suspension. Figure 11. EDSA passenger traffic growth, 1980 to

108 ROOSEVELT AVENUE SCREENLINE STATION GUADALUPE BRIDGE SCREENLINE STATION 104

109 105 SOUTH SUPER HIGHWAY SCREENLINE STATION

110 ALL STATIONS COMBINED (b) Travel time characteristics There was a general perception that bus travel time along EDSA improved as a result of the introduction of the bus lane scheme. Bus operators observed that the number of round trips in any given day increased. This would tend to suggest a reduction in bus travel times along EDSA. 106

111 However, contrary to expectations, the results of "before and after" studies on end to end travel showed mat mere was an increase in travel times for buses travelling between Monumento and Tramo in me years from 1988 to 1990, even with me introduction of the EDSA bus lanes (see figure 12). The increase in morning peak period travel times was generally greater man mat in the afternoon peak. Delay times, mainly at intersections (see figures 13 and 14), increased during the same period. Delay time remained at around 35 per cent to 40 per cent of the total travel time for each direction of travel. The increase in delays was brought about by the excessively long signal cycles at intersections, ranging from four to eight minutes, and me disorderly loading and unloading of passengers at bus stops. During the long signal cycle, bus waiting lines build up so that all buses on the approach are forced to load and unload passengers just before the intersection, thereby reducing the effective discharge capacity of the bus lane approach during the signal green phase. Likewise, blocking of lanes within bus stops further diminishes bus lane capacity. It must be emphasised that 73 per cent of observed delays were attributed to the long signal cycle and delays at bus stops. Therefore, the performance of the bus lane operation should not be judged only on the basis of a comparison of travel times, but also of their components (e.g., running time, intersection signal delay, passenger processing time). In effect, the bus running speed represents mid block operating speeds resulting from the combined effects of bus flow volume, bus lane geometry, spacing of stops, and general effects of other vehicles within the bus lanes. The main impact of bus lanes along EDSA has been a reduction in mid block delays resulting from general traffic congestion. Bus running speeds averaged around 30 km/h and was fairly consistent during both morning and evening peak periods. Here mention should be made that before reopening of bus lanes for traffic in December 1992, bus stops for different bus routes have been separated to alleviate the formation of bus queues. D. Lessons from the case studies 1. Adequacy of planning and design consideration (a) Planning process The case studies on LRT and the EDSA bus lane depict contrasting approaches to urban transport improvement which address both traffic congestion and public transport capacity problems. The proposal to introduce a rail based mass transit was born out of a comprehensive urban development and transport study and follow up detailed studies. Various technical options were considered and the final system design, including route alignment, was selected after a thorough evaluation of alternatives taking a Metro Manila wide transport network viewpoint. On the other hand, the bus lane scheme evolved from a realization of an existing problem directly confronting the transport users, operators and government authorities. The development of bus lanes followed an operational planning approach, with the immediate objective of providing relief in the worsening urban transport situation. The bus lane planning process proved to be shortsighted in scope, with the apparent failure in identifying key success factors such as traffic law enforcement, bus driver behaviour, bus stop requirements etc. The bus lane scheme should have been developed as a package of supportive and complementary measures and not conceived as an isolated action. Figure 12. Summary of the results of EDSA bus lanes travel time 107

112 "BEFORE" "AFTER" Figure 13. EDSA bus lane travel time northbound bus travel time components 108

113 NORTHBOUND AM PEAK NORTHBOUND PM PEAK Figure 14. EDSA bus lane travel time southbound bus travel time components 109