Exploring bus transfer behaviour in metropolitan Melbourne

Although passengers dislike transferring between routes, quality transit systems need to facilitate easy transfers to provide competitive city wide access to the private car. This paper reviews bus transfer behaviour in Melbourne, Australia and aims to understand factors which enable transfer behaviour with a particular focus on buses. The research also explores if high transfer rates can be associated with \u27network effects\u27; high ridership associated with frequent services/ simple (grid) networks. Around half of bus travellers make transfers – mostly to rail. Transfers between trams/buses are low overall, however in inner/CBD areas where trams run they dominate bus transfers. Younger people, secondary students, males, commuters and those using periodical and full fare tickets (like commuters) have higher bus transfer rates. Middle and older age groups, off peak travel, shopping trips and concession ticket holders have lower bus transfer rates. Because commuters transfer more, weekday peaks have high transfer rates. Analysis of bus route types shows that transfers are higher for more frequent and longer distance routes, those with simpler (straighter) route design, commuter based services and those which operate in networks which require transfers. Schedule coordination with rail was shown to increase transfer rates but not by much. Analysis of transfer route pairs demonstrated a modest (R2= 0.25) though significant relationship between the volume of transfer trips and service frequency. Analysis of the best or highest frequency transfer route pairs showed that high volume transfers tend to occur where one of the routes has a frequency of 10 minutes or better. This is an important finding because current bus network plans are for a 15 minute based grid route network. These findings suggest that if major routes in the grid network have 10 minute headways or better then a significantly higher share of transfers would result. Analysis of very high bus transfer sub-networks showed a concentration of transfers around a grid network with high frequency service near the CBD. A comparative analysis found that this area has high service frequency, network density and residential density. The analysis concluded that this transfer behaviour was highly consistent with the „network effect‟ but this cannot act as conclusive proof the effect exists since high transfer rates can also be explained by the nature of travel patterns in this area. In simple terms the „network effect” though intriguing remains an unsubstantiated theory which informs good practice but should be treated with caution when applied in the real world