TRANSPORT RESEARCH LABORATORY [© fl, InM? AA TITLE A study of high developing cities capacity busways G Gardner Overseas Centre Transport Research Laboratory Crowthorne Berkshire United Kingdom by in ..I'. ;"-, Gardner,G (1992). A study of high capacity busways in developing cities. In: Proceedings of Institution of Civil Engineers. Transport. August 1994. A study of high capacity busways in developing cities G. Gardner, BSc Proc. Instn Civ. Engrs Transp.,1992, 95, Aug., 185 -191 Paper 9822 E Many of the world's maj or cities face the ever-increasing problem of traffic conges- tion. Buses are forced to compete for road space with all other road users and hence frequently fail to deliver an acceptable public transport service. Numerous cities throughout the world are now planning rail-based metro systems to provide public transport. Because of the extremely high capital costs, however, metros will rarely be appropriate, particularly in developing countries. Modern bus systems, including busways, could provide a viable mass transit option. This Paper presents the main findings of a TRR1L (now TRL) research project on the performance of busways in developing cities. From a case study of eight such systems, capacity figures are presented which show that passenger flows of over 20 000 per hour per direction are possible and that even higher flows might be achievable given the right combination of design and operational features. A butsway transit network, with a dedi- cated track, traffic signal priority and similar stop-spacing to that of a metro is potentially able to provide a suitable alter- native to a rail-based system, at a fraction of the cost. With high-quality buses~and a good corporate image, such a system could be attractive to city authorities in many UK and European cities. Bus stop spacing and design are found to be the critical component of a high- capacity busway transit system, although the provision of overtaking bays, together with efficient ticketing arrangements, can greatly improve throughput. The advantages of busways for a developing country include their relatively low costs, with minimal foreign-exchange requirement, and much greater flexibility* and adaptability when compared to a fixed-route system. Introduction Many of the world's major cities face ever- increasing problems of traffic congestion. This can hinder economic growth and cause severe environmental damage. Buses, which are the dominant form of public transport in most developing cities, are locked in this congestion and hence-'-frequently fail to deliver an accept- able service. 2. Many countries have favoured the intro- duction of heavy rail metro systems. However, a recent Transport and Road Research Labor- atory (TRRL now TRL) study of metros in developing countries' found that although rail may be the only means of -carrying more than - 30 000 passengers per hour per direction, a city should exhaust all possible alternatives before opting for a metro, because of the extremely high capital costs. 3. Several cities have implemented innova- tive and modern bus systems, including busways, as an alternative mass transit option. Such systems have many potential benefits, but relatively little is known about how well they perform. 4. In 1989 TRRZL, with Traffic and Trans- port Consultants Ltd (TTC), initiated a study of busway transit in developing cities.' The study objectives were to review the performance of existing bus priority systems, to determine their appropriateness and scope for general application and to establish relationships between passenger demand, design features and operation. 5. The impact of bus priority on general traffic is dependent upon the specific layout of a city's road network and must be examined using standard transport planning techniques. This has been discussed elsewhere (for example by Holman et al.)' and was not included in the present study. 6. As many popular misconceptions exist over the capacity of buses relative to rail-based systems, a high emphasis was placed upon case study~measurement of-eight busway schemes (Table 1). To ensure accu.racy, all surveys were personally supervised by a member of the study team. Busway features 7. The concept of a bus-lane is well-known, being an area of road space reserved for buses only by the use of paint and signs. This gives buses priority over other vehicles leading to fewer delays, especially on the approach to junctions. A 'busway'(as shown in Fig. 1) includes some form of physical segregation. 8. A variety of layouts are possible, includ- ing 'lateral' busways in the nearside lane, and 'median' busways usually occupying the Written discussion closes 15 October 1992 G. Gardner, Senior Scientific Officer, Overseas Unit, Transport ResearchLaboratory GARDNER Table 1. Physical characteristics of busways surveyed City Location Length: Average Average Special km stop junction features spacing: spacing: m m Abidjan* Blvd de la Republique 1.27 400 160 None Ankara Besevler-Dikimnevi 3.6 310 410 None Belo Horizonte Av. Cristiano Machado 8.57 610 920 Overtaking at stops Curitiba Eixo Sul 9.5 430 430 Trunk and feeder Istanbul Taksim-Zincirlikuyu 2.27 310 410 None Porto Alegre Assis Brasil 4.5 580 410 Bus ordering Porto Alegre Farrapos 2.8 560 390 Bus ordering Sao Paulo Av. 9 de Julho/S. Amnaro 7.9 600 530 Overtaking at stops *All median systems except Abidjan (lateral) and Istanbul (mixed). central reserve of a dual-carriageway. A typical which renders simple vehicle actuation mnoper- busway might extend for 1-10 kin, incorpo- ative. rating several stops and junctions. Bus-stops 11. Busway transit is the term used to for median busways are in the centre of the describe busway schemes which include a road, having pedestrian crossings to ensure package of performance-enhancing operational safety. and design features to provide a full mass 9. In order to achieve high performance and transit system. A busway transit system might to realize the full potential of a busway, good have many of the attributes of a metro, that is. design, complemented by operational and man- fixed routes which are clearly named (for agement measures are necessary. With very example, 'central line', 'green line' etc.), dedi- high flows, management measures are also cated named station/stops and a corporate needed to ensure that buses at either end of the image for vehicles, timetables and publicity busway are fed into the normal street network, material. or into a terminal area with maximum effi- ciency. Busway performance 10. General traffic may require diversion, 12. On each of the case study busways, syn- but will usually travel alongside the busway, chronized watches and number-plate matching with crossing movements restricted using 'C' techniques at points 1-5 km apart, were used or 'Q' turns. In some cases, special detectors to measure speeds. Occupancies were measured enable buses to be given preferential treatment on a 7-point scale, with judgements, where at traffic signals. With very heavy flows, required, tending towards under-estimation. however, buses can arrive at a rate of one every Key indicators of the performance of the fifteen seconds, producing a constant 'call' busways are shown in Fig. 2. In addition, Bus stop ~~~~~~~~~~~~~~All Bus stop ~~~~~~~~~~~~~~traffic - _^ -^- ^- - - --- -~ - 0 1- _ m- - ^ -~ (a) Fig. 1. (a) Typical bus layout, Avenida 9 de Julho, Sao Paulo, Brasil; (b) typical bus- stop layout, Avenida Cristiano Machado, Belo Horizon te, Brasil (b) BUSWAYS IN DEVELOPING CITIES several bus-stops were selected for more detailed surveys; these are discussed in the section on performance of bus-stops. Case study results 13. Bus flows Peak hourly bus flows per lane per direction ranged from 91 -378 per hour in the morning peak and from 80-304 per hour during the evening peak. Maximum flows exceeded 200 per hour at five of the sites, and exceeded 300 per hour at two sites. This corre- sponds to a maximum recorded number of available passenger places of 39 400 per hour (during the morning peak, taking nominal bus capacities, not crush loading). 14. Passenger flows The maximum record- ed line-haul passenger throughput was 26 100 passengers per hour per direction (p/h/d) on Assis Brasil, Porto Alegre (during the morning peak when passengers at the busy city centre bus-stops were predominantly alighting). The highest evening peak passenger throughputs were recorded in San Paulo (20 300 p/h/d) where an overtaking lane at bus stops facili- tates high throughputs at acceptable speeds. 15. The highest recorded passenger throughput on a basic busway, (i.e. one without any special operational measures) was 19 500 p/h/d in the predominantly boarding direction in Abidjan, this occurred under condi- tions of extensive bus queueing and severe crush loading during the evening peak. 16. Bus speeds Average bus commercial speeds along the case study busways ranged 30 25 h- N from 12-0 to 24.6 km/h during the morning peak and from 8-0 to 29.3 km/h during the evening peak (Fig. 2). Bus-stop and intersection spacing, and the provision of special operating features, would appear to be the main influence on bus speeds. 17. In the three city-centre sites, where stops and junctions occur frequently, average speeds were around 11 km/h. On the suburban busways, where longer distances exist between stops and intersections, averages of around 21 kmph were achieved. The suburban busways also tended to have special operating features, and work is currently in progress, using multiple regression techniques, to esti- mate the relative influence of these factors. Special operational measures Study results 18. The study revealed that certain physical and operational characteristics are linked to busway performance. These include the follow- ing. 19. Trunk and feeder services With trunk- and-feeder services, very short headways are used on the trunk routes on the busway, with other services rerouted to feed terminals at the end of the busway, rather than travelling through to the city centre. 20. As all passengers use the same trunk service, all passengers board the next arriving bus, each bus fills up and very high load factors (that is passengers carried/nominal capacity) can be achieved. The limited number of routes, however, does result in enforced interchange for many passengers using the m Passengers: 000s per hour per direction 1~ Speed: kmph Suburban City centre / 7I~ Sao Porto Paulo Aliegre(i / 7 PortoAllegre(ii) Curitiba 7 / 71 Abidjan Ankara Istanbul Fig. 2. Summary of busway performance 20 h- 15 k 5 0 / BeloHorizonte to
- GARDNER Fig. 3. Median butsway with a very high volume bus -slop, Ankara, Turkey feeder buses, or a limited choice of destinations. 21. In Curitiba, these terminals are enclosed, such that tickets are purchased at the entry to the terminal, rather than on the bus. This reduces boarding times and provides free transfer for those interchanging inside the closed area. 22. COMMONOIR or bus ordering COMMONOR is a technique which involves assembling buses at the start of the busway into a sequence corresponding to the route and stand order at individual bus stops along the b's-way.' Buses then proceed along the busway in a manner similar to a train; boarding of buses at all stands takes place concurrently, thus co-ordinating the time lost through decel- eration etc. and reducing the clueueing time associated with loading buses at the first stand holding up all others. 23. COMMONOR was not in full operation Table 2. Average bus travel times (in seconds) by bus-stop Category Moeate 66o Balighing Very high 108 44 High 143 Moderate 52 n/a during the present study, but the less formal bus ordering system which existed in Porto Alegre, whereby buses are assembled in a regular order, but not necessarily grouped into complete convoys, was associated with better performance than might otherwise have been expected. Performance of bus stops 24. Figure 3 shows a very busy bus stop in Ankara which handles more than 4000 pas- sengers in the peak hour (this is more than many stations on the London Undergi-ound). As buses usually have only one or two doors avail- able to boarding passengers, passenger move- mnents at stops have a large influence on line-haul performance. As the numbers of boarding/ alighting passengers increase, so buis dwell times at stops increase, and this causes capacity limitation. Study results 25. Bus stop surveys A 1number of key bus stops were selected for detailed investigation. rThese were mostly in the case-study busways, but additional surveys took place at stops which had particularly interesting features, such as in Singapore (Fig. 4) and H ong Kong. The times of arrival at the approach and of exit from the stop were noted, the times that the doors opened and closed were taken, and the number of passengers boarding and alighting were counted on a sample basis. 26. Passenger flows and travel times Three categories of bus stops were identified in the BUSWAYS IN DEVELOPING CITIES study, according to the number of passengers handled (a) 'very high volume' stops, typically at city centre locations, with either boardings or alightings greater than or equal to 2500 per hour (b) 'high volume' stops, typically in local centres, with maximum boardings or alightings less than 2500 but greater than or equal to 1000 per hour (c) 'moderate volume' stops, with both board- ings and alightings less than 1000 per hour. (These all represent extremely high volumes when compared to most European or American bus-stops.) 27. Travel time was defined as from the moment the bus arrived at the queueing area, until it cleared the last bus stand. Large varia- tions were found at each site: mean travel times varied from 26-203 s and loading/unloading (i.e. door open to door close) times varied from 11- 109 S. 28. As shown in Table 2, although some of the variation in travel time was explained by passenger volumes, the presence of overtaking facilities was also important. Between 15 and 69 per cent of bus time spent in the stop area was not associated with passenger movement, being mainly due to queues of buses on the approach to the stop and to traffic controls. 29. Overtaking at stops With very high flows of buses, queues can build up at stops, and all buses will then travel at the speed of the slowest until there is an overtaking opportunity. 30. All bus stops surveyed which had over- taking facilities, had lower overall delay times than thoste without. The innovatory parallel stands bus stop in Singapore (Fig. 4), for example, had very much lower delays than comparable sites in Turkey. 31. Overtaking also permits the intro- duction of limited-stop and express services, making overtaking bays at stops one of the most cost-effective measures to improve capac- ity and commercial bus speeds under normal circumstances. (The problem found in the UK, of buses being denied re-entry to their lane was not found to exist, as can be seen in TRRL's video.)' 32. With overtaking facilities, a bus can leave as soon as boarding is comijlete, loading of several routes occurs concurrently and, with very short headways, supply can be matched closely to demand. It is possible, therefore, that the capacity of an efficient overtaking bus stop can be higher than other alternatives which are usually thought to be of higher capacity, but which do not normally have overtaking, such as trolley-bus, guided bus and even light rail. Fig. 4. Jnnovatory parallel-bay bus-stop, Singapore Further work will be needed to confirm this, since reliable case study data are not yet avail- able. 33. Boarding and alighting times Boarding and alighting times were surveyed at a selec- tion of stops; these varied considerably from one city to another. For example, average boarding times for a typical group of ten pass- engers ranged from about 19-41 s. As shown in Table 3, board ing times per passenger where free entry onto the bus is permitted were lower, at around 1 s, than those where fare collection restricted entry at about 2 s per passenger. 34. Lost time per bus (i.e. time when doors were open but no passengers were moving) appeared to be fairly constant at around 10 s, although this was greater at the very long bus stops (some were up to 70 m long), and at the very busy bus stops such as in Ankara (Fig. 3). 35. Alighting times measured were in the range 0.4 to 0.9 s per passenger, confirming the well-known fact that boarding times per pas- senger tend to be longer than alighting times, typically about double. Busway advantages and disadvantages 36. This study has shown that busways are capable of carrying high passenger flows at acceptable speeds. As shown in Fig. 5, the capacity of the busways studied compares very favourably with many of the metros studied by Fouracre' particularly when capital costs are taken into consideration. In addition, the GARDNER Table 3. Passenger boarding times by city and fare collection arrangements City Lost Time/ Entry Fare time: passenger arrangementst collection S* ~~~S method Abidjan 10-3 0-9 Free entry Turnstile Bangkok 9-8 1-2 Free entry Conductor Belo Horizonte 5-2 1-5 Free entry Turnstile Sao Paulo 8-6 1-3 Free entry Turnstile Ankara 23:0 1.8 Driver supervised Paybox Hong Kong 13-1 1-7 Driver supervised Pay driver Istanbul 9-3 2-3 Driver supervised Paybox Singapore 8-4 2-2 Driver supervised Pay driver *Lost time represents average delay to all buses, irrespective of number of passengers boarding. t Driver supervised entry requires single-file boarding, otherwise full width of door available for passenger entry with fare collection on-board. advantages of busways for the city authorities are Self-enforcement: Because a busway physically segregates buses from general traffic, the priority for buses does not need to be enforced by a strong police presence. Flexibility and diversity: Since buses can join and leave a busway anywhere, routes from all over the city can use the busway for all or part of their journey. Passengers from a wide catchment area can therefore benefit from a faster service without having to transfer to a faster vehicle, as would be required with a fixed-track system. Affordability: An at-grade busway along an exist- ing right-of-way is likely to cost US $400 000 - 1 000 000 per km (end- 1989 values), depending upon the need for utility relocation and other local factors.Since busways can be provided with locally avail- 'ig. 5. Capacity of ietro and busway ystems with capital 9S ts 0 0) m able labour, materials and vehicles, the foreign exchange requirement for 'hard' currency is mini- mized.Scope for Incremental Development: Sections of even a few hundred metres of Busway can be useful (whereas rail transit needs a depot and a signifi- cant route length before it can attract passengers). Busway Transit can also be enhanced step-by-step (e.g. by adding grade separation at critical intersec- tions; introducing off-bus ticketing etc.) as and when finance permits. Existing Experience: busways enhance the use of buses, the predominant public transport mode in most cities, and can draw upon the wealth of .experience and knowledge of bus operation which already exists.' 37. One of the main disadvantages of busways are that being a mixture of highways, traffic and public transport, their implementa- Maximum passenger flow per hour per direction: 0O0s 0 10 20 30 40 50 60 70 Hong Kong (131) Mexico City (15) Sao Paulo (80) Cairo (12) Rio de Janeiro (66) Manila (40) Santiago (36) Singapore (37) Pusan (2 1) Porto Alegre (10) Tunis (23) Calcutta (41) AS. Brasil Sao, PauloAbidjanFarapos Belo HZ Istanbul CuritibaAnkara 80 90 ZZXJZZZYJYAZZZZ$JYZZ/777/ZZ/ZA'777 """'/// 77y77.17 77 7 777 77 / / ZZZZ Figures in brackets show capital cost per kilometre of metros (millions US$, 1986), from Fouracre. Comparable busway costs in region of 1 million US$ plus the cost of buses and terminals. BUSWAYS IN DEVELOPING Cl tion requires CO-operation from a number of separate institutions which is not always easy to achieve. 38. Busways can also be criticized for taking road space away from cars. For schemes with low passenger demand, as in the UK, careful transport planning evaluation would be necessary, on a case-by-case basis, to ensure optimum benefits. However, this study concen- trated on corridors with extremely high pass- enger flows of up to 26 000 persons per hour per direction. Under these conditions, since a busway can carry between five and ten times more passengers than a general traffic lane, there would seem to be an overwhelming case, on technical grounds, for providing space for buses at the possible expense of other traffic. 39. Similarly, for such high passenger flows, a bus with a well-maintained diesel engine is clearly superior in environmental terms, than the equivalent number in private cars. A comparison between mass transit options, which will investigate the total environmental advantages and disbenefits of electric and diesel alternatives, is to form the follow-up to the present study. 40. Perhaps the main disadvantage of busways is that they are perceived as being an 'outdated' and 'unclean' technology. Irrespec- tive of the potential demonstrated in this study, the worldwide demand for rail-based mass transit systems continues unabated, with (except perhaps in those UK cities which 'failed' to gain a light rail transit system) little sign of an active lobby for the busway option. Conclusions 41. The TRRL study has shown that given the right combination of operating features, it is possible for segregated busways to provide a hig~'ltapacity mass transit system. 42. For maximum throughput, care should be taken with all aspects of the design, but in particular with the layout of the bus-stops which should, ideally, include overtaking facili- ties. 43. Operational measures such as bus order- ing or trunk-and-feeder services can further enhance performance. 44. As demonstrated in Fig. 5, the pas- senger carrying performances of busways com- pares very favourably with that of metros, and at substantially lower cost. 45. Against all of the arguments in favour of busways, the popularity of metros has shown that it is not necessarily technical performance alone that determines the choice of public transport system, and the,'image' of the service can be critical. Acknowledgements 46. The Busways study team consisted of Dr Philip Cornwell and John Cracknell of Traffic and Transport Consultants, and Geoff Gardner and Phil Fouracre from TRL. 47. This work forms part of the programme of the Urban Transport and Traffic Manage- ment Section of the Overseas Unit (Unit head: J. S. Yerrell) of the Transport Research Labor- atory, and is published by permission of the Director. Crown Copyright. References1. Fous.AcRE P. R., ALLPORT R. J. and THOMSON J. M. The performance and impact of rail mass transit in developing countries. TRRI, Report 278, 1990. 2. GARDNER G., CORNWELI. P. R. and CRACKNELL J. The performance of busway transit in developing cities. TRRLI Research Report No. 329. Crow- thorne, 1991. 3. HOLMAN S., SANG A. and WILLUMSEN L. G. Com- puter assisted design of bus priority schemes. PTRC 19th Summer Annual Meeting, 1991. 4. SZASZ P. A., DE CARVALHO MONTANs L. and GODOY A. R. COMONOR: COMbois de ONibus 0Rde- nados naes avenidas Rangel Pestana e Celso Garcia, Boletim Tecnico da GET no 22, Coin- panhia de Engenharia de Trafego, Sao Paulo. 5. TRRI, Busways. Video produced by TTC and Herts Media for TRRL, (amateur format), TRRL, Crowthorne, 1990. 6. CORNWIELL P. R. and CRACKNELL J. A. The Case for Busway Transit. PTRC 18th Summer Annual Meeting, 1990.