Making transfers easier

Last week I wrote quite a long post about “cross-town routes” and how important they are in connecting up the suburb-to-CBD routes so that you can create a network allowing people to travel from any part of the city to any other part of the city without having to travel into the CBD and then out again. The critical part of this was the use of transfers – where people start their route on one bus (or train) and then transfer to another in order to finish off their trip.

Now transfers are seen as something to generally “avoid at all costs” in the current structure of Auckland’s public transport system. There are a number of reasons for that, as I will explain in the moment, but overall it’s generally considered to be an enormous pain in the neck to have to make a transfer. Therefore, our current bus routes are generally designed to enable us to avoid transfers as much as possible by creating long and windy routes that attempt to serve all possible destinations, and by having an enormous number of different routes. The consequence of having so many different long and confusing routes is that we spread the resources very thinly, and therefore the frequencies that can be offered are generally pretty pathetic.

Clearly, transfers are annoying. However, if you look at overseas cities with successful public transport systems you’ll find that transfers happen all the time. People don’t think twice about changing lines on the London Underground or the Paris Metro, around 70% of people using Toronto’s subway system arrived at the subway station on a bus, while the majority of people using Perth’s train system did the same. So clearly, while transfers are annoying, they’re certainly not something that kills your system’s popularity altogether. The key is to make the transfer as painless as possible. There are three important elements to this in my opinion:

1. Make the ticketing of the transfer simple and seamless.
2. Make the timing of the transfer as least annoying as possible.
3. Make the physical process of the transfer as easy as possible.

Integrated ticketing, and particularly time-based ticketing (where you buy an unlimited number of trips within a two hour period within a certain distance), is critical to accomplish the first element. At the moment Auckland is utterly useless at this – in that every trip you have to pay again and again and again, and if you switch from bus to train (or vice-versa) you can’t use your same pass, your daily pass might be valid on some bus companies but not others…. it’s horrific. However, fortunately that is being sorted out – and we should have integrated ticketing…. well, some time in the next few years. Shifting Auckland from a “stage-based” fare system to a “zone-based” system will also be a critical part of making transfers easy – as that will enable the ‘time-based ticketing’ I have outlined above.

In order to get the second part right (and not have situations like we’re finding at the current Newmarket train station where there’s no alignment whatsoever between the timetables of southern line trains and western line trains – to make transfering easy) the wait between getting off your first service and getting on your second service must be minimised. I would suggest a maximum waiting time of 5 minutes, otherwise I imagine that people are simply not going to bother with the transfer. There are a couple of ways to do this: either by carefully aligning arrival and departure times (easier for trains services than buses, which are affected by traffic) or ensuring that your frequencies are high enough that it doesn’t really matter whether the timetables are aligned or not. I would suggest that a service every 10 minutes is probably the minimum level of service if we’re not going to worry about trying to align timetables. Pulse timetabling is a bit of a combination of the two, where you have a number of services all arriving around the same time – to make life relatively easy for transfering, but not dependent on one service linking up. I would suggest that ideally we would be able to run services frequently enough to not worry about aligning timetables or having pulse timetables (the London Underground solution), but if that’s not possible then doing pulse timetabling or timetable alignment.

Which brings us to the third necessary part of making the transfer easier – the physical process of transfering. In a rail context, the ideal situation is a “cross-platform transfer”, where you simply get out of your first train, wander across the platform and get on your second one. This might only take 20 seconds if the train is waiting there for you. I think either Hong Kong or Singapore (or both) was smart enough to make all (or almost all) their subway line interchanges into cross-platforms transfers, and it works incredibly well for them. In a bus to rail, or rail to bus, context I would say it’s critical to get the buses as close to the train station as possible. If possible, it would be great to keep the buses within the “fare paid” zone to keep things really simple.

For bus-to-bus transfers, things are a bit trickier. Often you’re going to have a transfer point at a big intersection, which means that to change buses you’ll have to cross many sets of traffic lights and it’ll take forever. A solution Vancouver operates is having bus stops both before and after the intersection, to minimise the number of streets you have to cross over. This seems like a sensible solution. Alternatively, it might make sense for one of your routes to take a small detour into a specified “transfer zone”, like a bus station or simply a stretch of route used by the other service so that effectively a “cross-platform” transfer is possible. Let’s have a look at how that might benefit one of Auckland’s future transfer nodes – Point Chevalier.

In the map below, Point Chevalier shops are shown along with the main existing bus routes that run through it. There are a great number of routes that I’ve indicated in blue as Main West services. Almost all of West Auckland’s buses pass through this point, so there’s excellent frequencies (particularly once we have integrated ticketing as Ritchies, NZ Bus and Urban Express services all pass through here). The green line indicates what I consider to be Auckland’s most successful cross-town route – the 007. The red and aqua lines indicate buses that serve Unitec and Pt Chevalier beach. The aqua route – the 043 – is a pretty pathetic service that runs hourly during the day only on weekdays.

In a previous post, I explained why I think the 045 route should probably be altered quite radically so that it is combined with the 005 route. Basically, by combining these two routes (effectively by extending the 005 along Meola Road to Pt Chev) you could create one service with excellent frequencies rather than two services with fairly rubbish frequencies. Furthermore, as there are already a million buses going along Great North Road, it would hardly miss that 045 service. The 043, as I note in the image above, is pretty rubbish and could probably be done away with altogether if it was easier to transfer between the “Main West Routes” and the 007 Crosstown Route. So ultimately, we may well end up in a situation where there are only two routes going through Pt Chevalier: an obvious east-west route (all the West Auckland buses) and an obvious north-south route (the 007, with much much better frequencies).

Now if this were to be implemented there’s an obvious problem to overcome: making is physically easy for the transfer to be made. Integrated ticketing can be sorted out, the 007 is likely to form part of a future Quality Transit Network so we would have high enough frequencies to ensure the waits when transfering were acceptable. But to complete the job we need to ensure that the physical process of transfering is easy. And unfortunately, that’s currently not the case. In the maps below I show the location of each existing bus stop, and the process of transfering from each service to the other one. As the table below shows, each transfer involves a walk of around 200m plus the crossing of up to two main roads:

So once again I think “how could we make this easier?” A 200 metre walk, plus waiting for up to two phases of lights to change in your favour, might take around 5 minutes and would certainly put a lot of people off transfering between services in this location I think. It’s also not particularly obvious the connections between the two sets of bus stops, while the location of the hugely bus intersection makes it difficult to shift them closer to each other. In order to find a solution it’s necessary to step back a bit, and in my opinion look at whether we perhaps shift the location of the 007 route a bit. The map below shows how this route could be altered in such a way as to provide much more simple transfers between services at the Point Chevalier shops: The red stars show where the two bus stops are, and as there’s already a pedestrian traffic light crossing between the two, each possible transfer would be just a very short cross of the road (if that) away from each other). Now of course there are a couple of drawbacks of this option.

1. By making the green crosstown route zig-zag down Moa Road you do lose some of its legibility and a trip along it would take longer than if it went straight down Pt Chev Road.
2. You’d need to install traffic lights at the corner of Moa Road and Great North Road. Interestingly enough, the 007 used to somewhat follow the route shown above, but because of the lack of traffic lights trips were really unreliable and it was a bit of a mess.

Overall though, I think the advantages would outweigh the disadvantages here. Because we are near the end of the cross-town route (it would fully end at Pt Chevalier beach), I think there would be a lot of transfer passengers and perhaps not so many people continuing right through from south to north. So the advantages of an easier transfer may outweigh the disadvantages of a slower trip.

Share this