Follow us on Twitter

Improving bus efficiency

I have noted in previous posts how many of our long-haul bus routes duplicate the rail network, leading to needlessly wasted money on subsidising buses and trains that compete with each other for passengers. Running all those Howick and Eastern buses right past Panmure station, Ellerslie Station, Greenlane Station, Remuera Station, Newmarket Station, Grafton Station to terminate right next to Britomart station seems a rather pointless exercise in my opinion.

The bus network itself has a large number of inefficiencies – generally resulting from its “branching” structure. A good example of this is to look at buses that pass through Newmarket – fed from a number of parts of Auckland: red being south (4xx routes generally), blue being the southeast (5xx routes for NZ Bus and all Howick & Eastern services), green from Manukau Road (3xx services) and orange from Remuera Road (6xx routes): The overlapping of routes over one another means extremely high frequencies – although often not in a way that’s aligned at all. So you might have no buses for 15 minutes then have seven come along one after the other. It seems that the main cause of this ‘branched’ network is the design of Auckland’s street network , which funnels one road after the next into bottlenecks like Newmarket (or the Harbour Bridge), rather than providing a series of parallel streets on a grid, where routes could be distributed more evenly. You sort of see more of a grid in the western part of the isthmus, but even those parallel roads (Manukau, Mt Eden, Dominion, Sandringham, New North & Great North) come together in bottlenecks of their own – like the top end of Symonds Street.

To get an idea about the number of routes and services that converge along this section of Great South Road and Broadway, I had a dig through MAXX timetable information to come up with the following: I’m not really sure whether Great South Road between Greenlane and Manukau Road generates demand in and of itself to justify a bus every 2 minutes at peak times, or that Khyber Pass generates demand to justify a bus every 40 seconds. I also think the fact that more than 50 different routes pass through these points in a given day is likely to mean there’s little co-ordination between timetables to avoid the vast platooning of buses. And remember, this is in a part of Auckland that’s pretty well served by rail too. Perhaps it’s no surprise that Remuera and Greenlane are two of the least used stations on the network.

In short, the branched route system leads to vast over-provision of services in inner areas, because of the overlapping of so many routes. Now this is great if you’re travelling between town and the hospital for example, but is it really the wisest spend of our transport dollar to be providing such an extremely high number of buses (most of which aren’t express services, and parts of the route have poor priority measures like through Newmarket, so this is no Northern Busway) here? Would we prefer to see that money spent on a high-frequency cross-town services? Cheaper fares? A higher farebox recovery rate so less reliance on subsidy? The low seat utilisation of buses crossing Grafton Bridge seems to reinforce my thinking on this matter.

Of course the main reason why we have so many buses travelling along the inner part of the routes shown above is for capacity (notwithstanding the low seat utilisation). A pretty large chunk of Auckland is served by these routes, which means a lot of people rely on them. But at such high levels of demand, one would think that shifting more people onto trains starts to be a more logical thing to do – especially if the train trip can be much faster for the traveller. For Howick and Eastern passengers headed for the city, transfering to the train at Panmure is going to cut a massive chunk off their travel time, even if they need to wait 5-10 minutes for the train and even if they then need to walk 5-10 minutes from Britomart to their destination. If we can make better use of the capacity that the rail network has (especially post electrification), then we shouldn’t need to run anywhere near this many buses between Greenlane and Grafton Bridge.

So how might we do things differently? As a basic structure, I’ve often thought about the following being a useful start:

  • Cut every route south of Manukau City at Manukau, turning them into feeders to the new railway station and a “400″ route b.line between Manukau and midtown (not Britomart as the train goes there). Frequency could be something like once every 10 or 15 minutes during the day, perhaps a bit more at peak and a bit less in the evenings and at weekends. This would be a reduction from the “bus every 5 minutes” service level we have now along Great South Road south of Greenlane, but would still offer a nice level of frequency along Great South Road all the time (for trips to and from places not easily served by rail), with excess demand in the south being soaked up by rail services from Manukau.
  •  Cut every route south of Onehunga (from Mangere) at Onehunga, turning them into feeders to Onehunga station and a new “300″ route b.line service along Manukau Road between Onehunga and midtown (once again, not Britomart as the train goes there). Frequency could probably be every 10 minutes most of the time, perhaps every 5 minutes during peak as the Manukau Road corridor doesn’t duplicate the rail network (unlike Great South Road). This would be a reduction from the 37 buses in the AM peak, with Onehunga rail services probably needing to go to every 15 minutes (so therefore post City Rail Link for full implementation) to soak up extra demand. It would be interesting to do an ‘origin-destination’ study for trips from Mangere, as I’ve heard that only a tiny fraction are to the CBD.
  • Cut all Howick and Eastern services at either Panmure or Ellerslie station, turning them into feeders. I need to have a bit more of a think about whether it’s best to cut these at Panmure then run a “500″ b.line between Panmure and midtown along Ellerslie-Panmure Highway or whether to continue all Howick and Eastern buses to Ellerslie, allowing transfers to the southern line and Great South Road buses. Any Panmure to midtown b.line route would probably have similar frequencies to the 400 Great South Road b.line, as it would only be the Ellerlise-Panmure Highway section that didn’t either duplicate the rail network (and the 400 route) or was relieved by transfers to rail at Panmure.
  • Remuera Road buses need a drastic simplification – probably combining a great number of routes into a simple ’600′ b.line service. This would be a good route to run through the city to Wynyard Quarter or even to the North Shore, as it “joins” our system too late for transfers to be attractive.

I’m pretty sure by implementing this plan we would save money (nowhere near as many duplicating services in our inner sections), give faster travel times (thanks to rail being much quicker than buses from Panmure, Manukau and Onehunga stations), provide a more easily understood bus network (based around the 300, 400, 500 and 600 b.line routes compared to the 61 routes that current serve the area) and a bus network that offers more regular services during weekends and off-peak (money saved on cutting back duplicative peak time services could be reinvested in better frequencies at other times).

Downsides are obviously a much greater reliance on transfers – both bus to train and bus to bus. We would need integrated ticketing and zone-based fares for this to truly work. We would also need good timetable alignment and excellent infrastructure at key transfer nodes like Manukau, Onehunga and Panmure. I’m guessing that many people would choose to take advantage of the faster travel speeds offered by rail, meaning that increased pressure would be placed on the rail network. Perhaps for this reason more than any other, a shift to this type of bus network would need to be implemented incrementally – particularly built around the introduction of our electric trains (and later the City Rail Link, particularly to allow more trains to serve Onehunga).

I think overall the positives significantly outweigh the negative though.

29 comments to Improving bus efficiency

  • Marty McFly

    Agree with ending at Manukau & Onehunga. In the ‘good olde days’ most buses from Sth Auckland ended at (north) or started at (south) Otahuhu. They were all the 4X2 routes & you knew if you bus was one, eg 432, you had to transfer at the Otahuhu bus terminal. Whent he bus routes changed, eg Yellow bus lost the 42X & 43X to Cityline (Stagecoach) on i think the 28th august 1995, that when it changed & lots of buses started going to & from ‘Town’ that used to transfer.

  • dan carter

    admin for mayor!

  • What resistance might there be from bus companies? Does the structure of the subsidy relationships with the companies easily enable these changes? Are there upsides for the companies in these changes? You would hope that things are structured to encourage the companies to run fuller buses on more efficient journeys than just trundle a half empty old belcher all the way to town and mine a subsidy….?

    Good work, and bring on that integrated ticketing, eh? There is a bit of chicken and egg here isn’t there…. We need the rail network with better capacity, frequency, and reliability; the removal of transfer penalties for the whole region; integrated stations and stops; new bus routes…. Yet each change is needed to provide the case for the others….

    • I think that it could be made to work in a way that benefits the bus companies – although there is certainly likely to be large-scale resistance from them in the short-term. In the longer term, obviously bus patronage is going to continue to grow and therefore the companies have the chance of growing their revenue.

      • Matt L

        I think the key is to effectively keep the subsidies at the same level but change the way the routes work so while we would effectively cut back on the number of buses along places like GSR, it would be made up by higher frequencies elsewhere or by the introduction of new local feeder services. PTOM should also help a little as AT are be required to create whole route or set of combined routes with a full timetable and tender that out, if a bus company wants to run a service commercially they would need to operate the whole thing commercially for the entire day so that will help to reduce resistance.

  • Matt

    There might be enough services for buses every five minutes at Greenlane, but buggered if they actually run like that. I’ve waited 15 minutes in the morning post-peak shoulder for a bus on GSR north of the Harp of Erin. That’s unacceptable, in my book, for such a highly-serviced route. I shouldn’t have to wait more than 10 minutes at pretty much any time of day during the week, but by 10pm it’s down to services maybe every 20 minutes; and that’s assuming you can figure out which bloody stop to use around Britomart to get onto a GSR-bound service.

    There needs to be some serious rationalisation and coordination of bus services over the next couple of years, because the current situation is ridiculous for users and horribly expensive for AT to operate.

    • It’s useful to get some insights from people who use these services, so thanks for your thoughts Matt. I felt a little uneasy in proposing such radical change to an area where I’ve only extremely rarely actually used the bus.

      • Matt

        No, you definitely have the right idea. There needs to be a significant shake-up.

        My concern is that between Otahuhu and Newmarket the train services are probably not sufficiently regular to allow this to work well in a pre-CRL world, because of the Eastern/Southern service split. The train timetable for Ellerslie has two big gaps in the morning peak (0735-0750 and 0805-0824) that are pushing the very outside edges of what I would call appropriate for using buses as feeder services, especially since those holes are in periods that serve people wanting to get to work in the key period of 0800-0830.

        An adjustment of the timetable would be possible, but the flow-on effects could be very challenging to address in terms of impact on the Eastern Line and on the southern Southern Line.

    • The disparity of where to get on the bus is a point not in favour of services from Downtown (note I didn’t say Britomart) to Newmarket and beyond. Having to wait at a run-down stop outside Showgirls for Mangere and Great South Rd services is not a way to make the service attractive to the general population, and doesn’t look good in terms of how transport planners regard the population from that area.

      B.line services and rail transfers (with decent facilities at the transfer points) as you propose with pulse-timetables for outlying services timed to connect with the trains would make a much better, more attractive, more legible PT network.

      • Matt

        Agreed. A few times I walked from the H&E stop at Britomart to the bottom of Anzac Ave, just so that I could be at a stop which caught all the services heading in my direction. I remember one time I did that, and the “real-time” sign said it was going to be longer until the next bus than to the departure time of the train that I hadn’t wanted to wait for so had gone hunting for a bus. That was only in the mid-evening, too, not even late at night.

        The entire service schedule is diabolical if one wants to get out south, and I’m only on Main Highway in Ellerslie so it’s not like I want to go off the beaten track to an area that’s serviced by one bus an hour. I ought to have a regular smorgasbord of buses to chose from at nearly any hour of the day or night, but I don’t. And as you can probably tell, it hacks me right off. Between irregular train service schedules that might keep me waiting for 20-30 minutes after 7pm, buses that leave from seemingly-random spots all across the bottom of the city, and the horrible historical bus timetables, public transport from the city to home is a challenging experience outside peak times.

  • Riccardo

    Look to Perth, the experts:

    -integrated ticketing
    -pulsing (mandatory below 6/h freq)
    -purpose built bus/rail interchange, including slewing the track for a large centre platform with the buses in the middle (eg Kelmscott, or Moss Vale in NSW) or overline bus interchange with escalators direct from platform to buses
    -improved legibility, no crazy express stopping patterns (in PErth, stopping patterns are letter coded in the timetable)
    -finally, the combo must be quicker than bus alone. Sounds obvious, but frequently not achieved, net of waiting times and so on.

      • http://www.humantransit.org/2010/12/basics-finding-your-pulse.html

        Effectively it means aligning the times for transfers between services.

        • Matt

          Ah. Yes, we absolutely need to get that sorted. And soon.

          • Riccardo

            Melbourne has a real problem with lack of pulsing. Google melbourne on transit (Peter Parker). So many routes with 20 minute train and 15 minute bus, or 40 minute bus and 30 minute train on weekends.

            Amuses me how many transport bureaucrats don’t or won’t see this is as a problem.

            Even on a single line passing loop eg Upwey, the trains for both directions are simultaneously at the platform, the local buses should be there as well, having arrived 5 minutes before the trains did, and departing 5 minutes after. The area then goes quiet till the next pulse. Perfect legibility and predictability.

            But it requires bureaucrats to talk, and as we often see, doesn’t happen.

            Funny story in Melbourne as an aside.

            Golf Rd overbridge built in early 20th century, apparently too weak now for heavy vehicles. Nonetheless, local bus companies using it on licensed bus routes until one day when someone realised they shouldn’t be. The bus companies believed the bridge was OK for buses as it was the Transport Ministry who licensed the route, and they figured, since the Transport Ministry also owned the bridge, they would know if it was OK to do so. LOL! They crazily imagined that bureaucrats in government ministries talk to each other.

  • SteveC

    while there is the germ of a good idea here and a real need for rationalisation, departure spreading on clockface timetables and bus/train integration, we also need to keep in mind the different markets that bus and train cater for, even when apparently operating in parallel

    buses have a much finer grain stopping pattern than train and allow people to make shorter journeys with boarding and alighting much closer to their origins and destinations, while rail is well suited to the longer journeys that become tedious by bus, plus buses run along activity corridors, when rail stations (particularly the older ones) tend to be more remote from activities

    so together, the two complementary modes probably remove more cars from the road than either singly

    • Matt

      Yes, absolutely true, but Auckland has lots of bus routes that are focussed on getting people into the CBD, even when there is a rail service that runs parallel to much of the same route. A perfect example is Papakura-Britomart, a route which neatly duplicates the Southern Line and takes considerably longer.

      Feeder buses that did circuits past two or three train stations (I sketched out a suggestion in a comment a few weeks back: one does Remuera, Greenlane, Ellerslie; another Ellerslie, Penrose, Onehunga) would provide the bus functionality without the waste of time and resources associated with running them all the way into the CBD. They would also provide much better levels of service to residents near the train stations but outside the convenient walking catchment.

      • SteveC

        Matt,

        if you look closely at the 47* routes, parallel can only really be used as a high level description, the 47* group serves different micro catchments along Great South Rd, for example, until the Manukau Spur opens, Manukau City Centre is a considerable walk from Homau or Puhinui (are trains from the south going to divert into and out of Manukau? honest question, this Shore boy doesn’t know the answer)

        this is an example of what I mean by different, complementary markets

        • Steve, the answer to your question about Manukau train schedules is here:
          Draft March 2012 train timetable summary.

          Under Josh’s proposal the 47x buses, which presently only diverge once they reach Papakura to serve around there, would only run from Manukau south – their northern terminus would be Manukau Station. North of Manukau, 47x users would switch to a Great South Rd b.line service that runs from Manukau north via places like Otahuhu, Penrose, Greenlane, Newmarket, or use the train for a quick trip in to town – whichever option best suits their needs.

        • Matt

          There’s no reason those various 47x services can’t loop around their existing off-GSR catchments and then drop the passengers at a train station (or two, or three). That’s the idea behind my loop feeder suggestion. A bunch of buses doing relatively small circuits (the two I sketched out would be about 20 minutes) will allow many more passengers to be carried than a few buses doing very, very long ones. Most buses, as Josh has observed more than once, don’t get filled up until they get near to the city: South buses fill up as they come through Ellerslie/Greenlane, they don’t fill up through Papakura/Manukau. Running mostly-empty buses for long distances is inefficient in terms of fuel litres per passenger and also driver hours per passenger.

          Obviously the transfer penalty needs to be removed before this is viable, but once that happens a complete reorganisation of how we operate buses to complement trains in Auckland will become possible.

      • Ben

        @Matt, your feeder bus circuit past two or three stations with an example of stations you gave above has another benefit too. Emergency rail bus replacements when the rail system is blocked due a breakdown or a points failure and these circuit buses can offload the rail passengers, go past the block and load up the passengers on another train on the other side of the blockage. No need to scramble and get buses as the buses are already there, just might need a few more to help out if there are A LOT passengers to be moved. Good example of the bus feeder circuit system would be Penrose to Onehunga when Onehunga trains get cancelled.

  • You’re right Steve, which is why there are still b.line frequency bus routes along all the corridors. What I’m mainly doing is shifting a lot of the peak time to CBD demand bulge onto a faster mode with more capacity (rail). We would still have a bus every minute or two going past the hospital in each direction at peak times, just not seven in a row platooning and generally not carrying many passengers.

  • on duplication i would like to through NEWLYNN into the debate- duplication at its best
    brand new transport center, good design but ask anyone who travels through it by bus will you what a nightmare it is. the bus has to negotiate 6 lights none have a priority bus-light (previously it was 2 and the terminus had a priority light) the streets are so narrow and the bus stops can only hold 8 buses, this is just a joke, recently our lovely ambitious Dy Mayor Ms.hulse demolished the building to make way for a shopping/parking center. I have great respect for her and her work but the council has to practice what they preach- PT PT PT
    My next one is LINK- if we have good bus lanes and effective transfers why do we need link s s s s…… i just think it is a desperate attempt by AT to please ratepayers.
    at the moment it is a farce- the city link will be more crippled with less patronage and longer loading times withe the 50c charge. its just a joke what streets it has to go through its so unreliable. when a complaint is raised, obviously to MAXX, Promptly taken down and get a reply that it will be passed on to the operator, the operator very promptly responds and places the blame on the traffic with all the apologizes, its just a JOKE JOKE. and nobody fault. we are at fault to use the PT.
    i recently read the AKT blog on how well the link is doing , REALLY have you been on one, it costs 1.80 i will even pay 10.80 if the link has a dedicated bus lane priority lights and reached to my appointments in time the word i am looking is RELIABLE, i have tried it. my car is quicker really i am not joking it is a fact the buses go through the same lanes i travel and i dont stop to pick up passengers and give information.
    who really came up with the buzzer for back door —i read the post earlier what an inconsistency how do they justify not to
    have it on other buses-
    one thing the link has done good is to the DEPOT where i live(gowest) in the west we have some of the EURO buses which Helen Clark inaugurated (the older link buses) they are more reliable , they run late but still turn up as they haven’t broken down, ( they still have the noisy, bumpy Nissan- the links -grey colored ones) and i am sure AT will come up with more plans like this benefiting NZ bus and metrolink and the biggest winners will be us in the west as we will get all the used ones of metrolink

  • chris

    To go back to the comment that greenlane and remuera are among the least used stations – could this, in particular with greenlane – also have anything to do with the lack pedestrian connection over the motorway? There’s only that revolting 3 lane roundabout that i can see.

    in terms of integrated ticketing: check out Switzerland. I was investigating buying a flight from london to zurich, i think, and with the plane ticket they also gave you options for onward rail tickets and then following buses to get exactly to your destination. Including timetables. and allowances for transfers between modes. Absolutely brilliant.

    http://www.swisstravelsystem.ch/en/

    • There is a separate pedestrian bridge just south of the Greenlane rotary. However, yes I agree that one of the main reasons Greenlane station is used so rarely is because it’s so incredibly hidden. It’s almost as difficult to find as the Broadway entrance to Newmarket station, or Te Mahia station generally.

      I’m not quite sure why we go to such lengths in Auckland to hide our railway stations.

      • Matt

        It’s not just hidden, it’s poorly connected to the surrounding residential areas. Getting from the north-eastern side of Greenlane East is a hard slog, either walking all the way down to the crossing at Ascot and then walking back (a distance that is nearly the entire 800m walking catchment, all on its own), or risking life and limb to dash across the road.

        From the north-western side, coming through from Marewa Rd requires another lengthy detour to the nearest crossing or an even-more-dangerous dash across Greenlane East.

        From the south-eastern side, it’s just a really long walk to the nearest residential area.

        And from the south-western side, a walking track tying together Walpole (steps up from next to the rail bridge), Adam and Derry, and then connecting with the track access road that comes in from Greenlane East, would increase the walking catchment significantly and quite cheaply. This is the area that’s best-connected to the station, but it’s still rather disconnected by the need to walk all the way up to Greenlane East in order to get access.

        I can’t help but wonder if the total lack of access to Remuera station from the northern end doesn’t play a big part in its lower passenger numbers, too. If you’re north of the station, it’s a loooooooooooong walk up to Market Road in order to get at the sole station access. This is especially frustrating when you’re around Marei Road, with a lovely motorway overbridge that doesn’t connect you in any way to the rail corridor over which it also passes.

  • LucyJH

    one thing I do wonder about this is whether you would be creating quite an inconvenience for students at Auckland Uni and AUT? I would imagine that a lot of the people who catch buses in Auckland from the West, South etc into the middle of town are university students. Right now they have very frequent services which leave from right outside uni (or 5 minutes up the road). It sounds like you’re proposing that instead they would either have to a) take a much less frequent bus service or b) walk for a good 15-20 minutes from Britomart to get up to uni. I realize that maybe it would svae them time overall to take the train but I think for many people such a long walk would be a significant deterent (rain, it ruins your makeup if you get hot and sweat in summer, you can’t wear high heels and walk even moderate distances, obesity makes walking difficult etc). Maybe the new Parnell station will help – not sure exac6tly where it will be in relation to uni. But (as I understand it) the walk from there to uni will be steep (you have to walk up, out of Grafton Gulley), will take at least 10 minutes, and be almost completely uncovered thus exacerbating the rain problem.

    • Matt L

      I guess with integrated fares and free transfers one could always catch one of the frequent buses that travels up Symonds St, not ideal having to transfer a few times but if the frequency is high enough, which it should be on that route and it is easy with all buses having a stop in the same place then it shouldn’t be to bad

Leave a Reply