As I detailed a few days ago, transport academic Paul Mees – in his most recent book – outlines why it is essential to have a “transfer-friendly” public transport system – using the example of Squaresville. Mees notes that in conventional transport planning, transfers are treated as barriers, whereas if you take a “network effect” approach to public transport planning – the transfer actually becomes an opportunity. Transfers are the means to link what would otherwise be a collection of individual routes. Of course trasnfering is inconvenient, but if we minimise the inconvenience (through integrated ticketing, timetable alignment and so forth) then the benefits of a true network effect should significantly outweigh the costs of the transfer inconvenience.
There’s a challenge to being able to create a network effect, as many commenters on my previous post pointed out. We just don’t have the money to add all these extra routes, so how can we introduce something of a network effect without destroying the existing network. This is where we have to start making trade-offs, looking at what routes on the existing network need to be redirected or replaced by other routes that will support a true network. Do we want more routes but lower service frequencies (meaning a shorter walk to the bus stop or train station, but a longer wait when you get there), or so we want the opposite: fewer different routes but higher frequencies?
Mees picks up on this:
“…the central challenge is to provide sufficiently high occupancies to support high system-wide service levels, on cross-suburban lines as well as radial lines, and to low-density as well as high-density areas. This challenge is met by offering a sparse, but high-quality, network comprised of relatively few lines operating at high service levels.”
This is basically the opposite of the situation you have in Auckland at the moment. Auckland has an enormous number of bus routes, with the aim seemingly to provide a service within a few hundred metres walk of each and every house in the city – but not to worry about whether that service comes every 10 minutes, or whether it only runs once a day.
A sparse network concentrates services, allowing higher frequencies and longer operating hours. It is also simple and stable, and thus easier for passengers to understand. Ease of understanding, or legibility, is not regarded as important in traditional public transport systems designed for regular commuters or ‘captive’ patrons, as it is assumed people will use the same services every day and become used to any quirks or complications. But region-wide networks are for everyone: regular users, occasional travellers, people visiting unfamiliar parts of the city, hikers and tourists. They must be stable and comprehensible, just like a road system. The model is networks like the Paris Metro and Zurich trams system, rather than the bewildering tangle of low-quality lines in cities as diverse as Auckland, Canberra and Manchester.
I would certainly agree that, in Auckland, public transport works a million times better for your regular commute – where you know the particular bus that you should take, where it will go and how long it will take – than it does for random trips to the mall at the weekend or other non-commuting activities. The system is too complex, the frequencies are too low and the services are too unreliable for other trips.
So what does this mean? How should we change our system to make it work more like a network?
This means that a public transport network should be comprised of fixed lines that follow the same routes, with the same stopping patterns, at all times. Special routes that only operate at peak hours, and separate night or weekend networks, should be avoided. If services are added at peak period, or thinned out at night, this should be done without disrupting the basic line pattern. A good example is provided by the suburban rail system, or S-Tog, of Copenhagen. There are seven main lines, and on most a mixture of express and stopping trains operates. The stopping patterns and even the departure times are the same all day long, every day of the year. On most routes 10-minute services are provided on weekdays and during shopping hours on Saturday; frequencies drop to 20 minutes at other times by the simple process of deleting every second train. Only two lines have additional trains in peak period, and these are slotted between the regular services without breaking the basic pattern. The entire system-wide map and timetable takes up a single, letter-sized sheet of paper.
I have long stated that I think it’s crazy for us to have separate Saturday and Sunday timetables – so crazy that it almost seems to be an intentional attempt to make public transport confusing and put us off using it. I think it would also make a lot of sense for the weekend timetables to be the same as the weekday off-peak timetables. So you would have your “core timetable” with peak services simply added on top of it.
But what about route planning? Particularly when it comes to planning bus routes, the temptation is always to depart from the principle of simplicity because you can. You can run a bus down most roads, so therefore there is the temptation to try and get your route as close to as many people as possible, even if that results in long, circuitous routes. Mees says that this is the wrong approach:
“…bus planners should design routes as if they were operating trams or trains, with simple, direct structures and as little duplication and overlap as possible. This involves… the ‘one section one line’ principle. Each corridor is provided with a single service, closely spaced and overlapping lines are avoided because, as the Toronto Transit Commission (TTC) says, ‘parallel routes… split the potential demand resulting in many routes competing for the same passengers and no route attracting enough demand to warrant a high frequency service.’
When a number of routes converge on a single corridor, the same principle can be applied. While in theory, 20 bus routes running hourly down a joint corridor means a service every three minutes, in practice it means bewildered passengers. A single line running every five minutes would use less resources but provide a better service.
We see this route convergence in a few places in Auckland. Perhaps most notably, on Great North Road and Great South Road. Theoretically there should be a tonne of buses along these routes, but in reality chances are they’re operated by a variety of different companies – so your ticket won’t be valid on half the buses going through. Furthermore, it’s also quite likely that because of no alignment between routes, there will be 15-20 minute gaps, and then five buses will come all at once. Pretty hopeless really.
So simplicity is the key here in my opinion. Looking at the buses that run in my corner of Auckland: Jervois Road routes, Richmond Road buses, Williamson Ave services and Great North Road to Pt Chevalier services, it does seem as though there’s quite a significant amount of route duplication and overlap. As a result of there being so many different routes serving this corner of Auckland, all services run at pretty rubbish frequencies: generally a bus every 20-30 minutes at off-peak, evenings and weekend times and not too much better during peak periods. Following the guides detailed above, it seems altogether possible that we could probably cover this area with two services: a CBD to Pt Chevalier via Ponsonby, Herne Bay and Westmere route: and a CBD to Westmere route via Ponsonby and Richmond Road. Combining the resources of four routes into two would mean that it would be quite feasible to run services at 10 minute frequencies, seven days a week – putting these routes up there with Dominion Road as some of the best serviced in Auckland. That would mean no need to check timetables, and surely attract a significant amount more patronage without really making people walk much further to catch the bus (as places losing services like Great North Road and Williamson Ave are already well covered by longer-distance routes).