A comparison of direct and trunk-feeder configurations for bus rapid transit systems

Abstract

Bus Rapid Transit (BRT) systems have gained popularity worldwide as a cost-effective alternative to more expensive urban rail systems, carrying around an estimated 33 million passengers each weekday (https://brtdata.org/ ). In South Africa, several BRT systems are either in the planning stage, detailed design, or construction, with only a few being operational (Ackerman, 2015). When planning BRT operations, planners need to decide when to use feeder or direct routes to supplement the trunk routes: this takes into consideration that trunk routes cannot be built to be within walking distance of large catchments of people. This research aims to explore the strengths and weaknesses of two BRT-based network types: trunk-feeder (buses operating inside and outside the BRT trunk corridor are segregated and operate independently) and direct (buses operating outside the trunk corridor can enter and leave it, providing additional services in the corridor). The Rea Vaya BRT system has both 'trunk-feeder' and 'direct' networks in operation and is used as a case study for this research. Rea Vaya routes have three classifications: trunk, complementary, and feeder routes. Trunk routes (T) use dedicated median-exclusive busways only. Complementary routes (C) use a combination of normal mixed traffic roads and dedicated median-exclusive busways. Feeder routes (F) start and end at Rea Vaya trunk stations using normal mixed traffic roads. The approach for the study is empirical and evidence based. The activities of the research are to: • develop a list of observable indicators to compare trunk-feeder and direct BRT networks; • collect data on indicators for trunk, feeder, and complementary routes; • analyse the data using different analytical tools; and • make direct versus trunk-feeder network recommendations for BRT systems in South African cities. Data collection is from four sources: station surveys, on-board surveys, ticketing information, and system data sourced from the operator. In this study, five key indicators (reliability, saturation, speed, load factor, & operating costs) are identified in guiding the comparative analysis. This led to the formulation of five hypotheses to be tested and make reasonable recommendations. According to analytical studies, the case for a trunk-feeder network rests on economies of density where it is cheaper per passenger to operate larger trunk buses on the main streets with high demand. For Rea Vaya, it is cheaper per passenger to operate trunk and feeder routes compared to the complementary routes. This saving is because of using larger vehicles (18m articulated buses) on the trunk corridor to achieve more capacity and costs are spread over a larger passenger number. However, the costs are highest for the trunk routes because of increased cycle times (and long routes), and increased fleet size requirements. From a cost perspective, trunk routes work best for densely populated areas but not over long distances. Literature suggests that the number of transfers that a trunk-feeder configuration require creates several operational inefficiencies and slower commercial speeds due to considerably higher dwell times (DTs). This is not entirely the case for Rea Vaya BRT system. While the trunk and feeder routes have longer dwell times than the complementary routes, the vehicle operating speeds for the trunk and feeder buses are higher than that of the complementary buses. The average vehicle operating speed for trunk buses is 30 km/h; for feeder buses, it is 25 km/h, and for complementary buses, it is 20 km/h. This is because the complementary buses are operated on major arterials with high levels of congestion before joining the trunk corridor. It can be concluded that the potential time savings of complementary routes through avoiding transfers does not materialise as it is more than offset by the slow vehicle speeds on mixed traffic routes. Overall, the results indicate a mixed view with regards to direct and trunk-feeder BRT networks in a South African context. While direct networks have an advantage of avoiding transfers, they are also found to be competitive in terms of headway reliability, maintaining low dwell times at the stations and having a high load factor (during peak only and consistent with the high peak to base ratio observed in South Africa).