Saturday, June 03, 2006

Measuring a transit system's development

The previous post described the progression from providing transport only for niche markets to highly developed integrated systems that cater for a wide variety of city and cross-regional trips.

One measure of a network's progress is the number of transfers that passengers make. Whether we like it or not, modern cross-suburban travel patterns increasingly require transfers. Since it's impossible to have an efficient network that eliminates transfers while maintaining good service frequency throughout, timetable co-ordination with easy interchange is extremely important.

Particularly in car-owning cities, a good measure of success is the extent to which people transfer, and whether such trips represent an increasing proportion of the total. If this is the case, we know that public transport is meeting emerging as well as traditional travel patterns.

The article 'Renaissance of an Urban Railway' in Transit Australia January 2006 by Alan Mortimer has some Transperth patronage figures (page 10). These cover the 1980-2005 period and show patronage rises following upgrades to the Perth network. The most significant of these are (i) the restoration of Fremantle line trains, (ii) suburban rail electrification and service increases, (iii) the northern suburbs railway, and (iv) bus service improvements elsewhere on the network.

Statistics provided in the article include 'Total system boardings' and 'Total journeys'. System boardings include transfers, whereas total journeys do not. Hence the number of transfers is the difference between these.

The relevant Transperth patronage statistics are as follows: (numbers in '000 pa)

Year - Total Boardings - Transfers - Percent transfers
1980-81 - 63660 - 11628 - 18.2%
1981-82 - 61206 - 11163 - 18.2%
1982-83 - 58482 - 10739 - 18.4%
1983-84 - 60273 - 14924 - 24.7%
1984-85 - 58762 - 14459 - 24.6%
1985-86 - 61666 - 15120 - 24.5%
1986-87 - 63944 - 12865 - 20.1%
1987-88 - 63524 - 12774 - 20.1%
1988-89 - 66275 - 13507 - 20.4%
1989-90 - 64045 - 12767 - 19.9%
1990-91 - 62278 - 12379 - 19.9%
1991-92 - 61466 - 12322 - 20.0%
1992-93 - 64500 - 13030 - 20.2%
1993-94 - 70409 - 17556 - 24.9%
1994-95 - 71878 - 17964 - 25.0%
1995-96 - 74013 - 21526 - 29.0%
1996-97 - 76438 - 22289 - 29.2%
1997-98 - 75566 - 21310 - 28.2%
1998-99 - 74273 - 20744 - 27.9%
1999-00 - 78250 - 22439 - 28.7%
2000-01 - 83696 - 24178 - 28.9%
2001-02 - 86069 - 24449 - 28.4%
2002-03 - 88121 - 24914 - 28.3%
2003-04 - 90578 - 25370 - 28.0%
2004-05 - 94985 - 26165 - 27.5%

Over this period the number of transfers shows an increase that is second only to the trebling of rail patronage since electrification.

The main growth in transfers came about during 1983-4 (when the Fremantle rail service was restored) and during 1993-4 when the northern suburbs line opened. The former may be because a higher service on one leg of the journey (in this case the train) can stimulate use of its feeders. In the second case, the reconfiguration of buses to feed railway stations rather than continue down the freeway forced some transfers. However since rail ridership growth (8 million) far outstripped the bus ridership decline (2 million), this change attracted new custom to public transport above and beyond a simple substitution of bus journeys by rail trips.

The cause of the decline in tranfers in the late 1980s is not known. However the period coincided with difficulties with the (then antiquated) suburban rail network. This is reflected in a 20% decline in rail patronage between 1986-7 and 1990-1. Since trips that involve tranfers are much more affected by service unreliabilty than those that don't, it is likely that this was a prime contributor.

The period since 1999 has seen a slight decrease in the proportion of transfers, though their absolute number continues to grow. This is largely because of accelerated growth in the total number of trips. Thanks to service improvements (eg Circle Route, regional route reviews, clockface scheduling and increased frequencies) the fastest increase has been in bus patronage (from 45 million in 1998-9 to 62 million in 2004-5). Though some extra people are tranferring to other services, an even higher proportion must be using the bus for local trips, which better services also encourage.

Looking at the bigger picture, it can be seen that the absolute number of transfers has doubled in the previous twelve years after a period of stagnation. Discounting the northern suburbs rail line impact the increase lessens to 50 percent. Nevertheless this can still be considered an achievement and a sign that Transperth is catering to a wider variety of destinations than its former CBD-focus.


leccy said...

Here's an interesting question that you might have an opinion on.

It's fairly well established that train services are 'more attractive' than bus services (even once frequency/speed/cost are taken in to account).

I'd be interested in knowing whether feeder buses count as bus trips, train trips or as something in between in this regard.

In particular, I'd love to know how far people are willing to travel on feeder buses to access rail stations.

I *think* there should be some pattern, but I don't know what it is.

What do you reckon?

Peter Parker said...

Hard to say.

On one hand you have the minus of the transfer but then you have pluses of the fast rail (for the bulk of the distance) and the wider range of destinations reachable compared to the previous freeway bus.

A possibility is to survey route 400 passengers to Scarborough. Do most of them get the bus all the way from WSBS or do they get the train to Glendalough and then transfer?

I suspect its more a matter of travel time rather than distance.

Especially on the northern line where buses have to negotiate slow loop streets.

The south (at least between the City and Murdoch) should be better, assuming Leach Hwy gets a 106-quality service. The main issue here is pedestrian access from homes to Leach Hwy/South St (given at least portions are controlled access highway) and pedestrian crossing facilities at EVERY bus stop.

Mees discusses this, pointing out that the main effect of Toronto's small rail system is that feeder bus trips are on average longer than somewhere like Melbourne (where few suburbs are more than 5km from rail). However one of the districts in his book has something like 24% modal share even though it had no rail.

So on this evidence long feeder bus trips could work, though ideally they should be straight and include traffic light priority and/or lanes.

The pattern for feeder bus travel would be like an arch shape, with the trailing edge dropping off a bit slower.

The bottom end would be limited by other access modes like walking and cycling. With a 30 minute frequency walking is a viable option up to about 2km, so many people would walk rather than bus. This would be down to about 1-1.5km when frequency is 15 min.

My own experience with the 67 tram (20 min at night) is there is still a good chance of beating it for walks of up to 10-15 min. And, thanks to quite close stops there is often a chance of boarding it if it does threaten to pass you.

Pedestrian conditions (including perceptions of distance, which are magnified by barren streetscapes, overpasses, street layouts freeway barriers, etc) may well influence this so that in Perth fewer will walk than in Melbourne's tram suburbs (Mees says urban form is more important for pedestrians than for PT).

The upper end is open-ended.

If we consider minutes rather than distance, I'd guess that 10-15 minutes for all access modes (whether walking, cycling, or feeder buses) is a fair upper limit for people with a choice.

However this is provided the trip is not 'excruciating', which I define as lots of stops, detours, waits and twists on all access modes. If it is then the max time may be shorter.

BTW, it would be interesting to see what proportion of Perth's population is within 5km of a station once the southern line is built.

My guess is that it would be close to 75%. NE suburbs around Beechboro, Forestfield/Kalamunda/Lesmurdie, parts of the coastal strip north and south of Fremantle and a large area between Mandurah and Waikiki would be the main areas to miss out.