Broadly speaking, there are two distinct types of trains in Auckland: the “Diesel Multiple Units” that we bought second-hand from Perth in the mid-1990s and the locomotive hauled carriages that were bought second-hand from the UK and then significantly refurbished, and now get hauled around by locomotives leased off KiwiRail. Some of those locomotives are 40-50 years old, while some of the Perth DMUs also date back to the 1960s (the ones without air-conditioning, known as the ADKs). (Note to rail nerds, I haven’t forgotten about the SX). Below are a couple of carriages from an SA train:

Up until around the time Britomart opened the rail network was pretty much solely operated by the ex-Perth DMUs. The impending patronage boom that Britomart was to bring (and obviously has brought) meant that additional capacity was required, and the SA/SD trains have provided that over the past few years. When there are debates about whether or not to sell the Port of Auckland, it’s worth remembering that the dividends on profits from the Ports are pretty much what has paid for the purchase and refurbishment of all the SA/SD trains that now form the majority of Auckland’s rail rolling stock.

But that’s enough history, the point of this blog post is to look forward. In the near future, it seems that we will have the last of the SA carriages coming online within the next couple of months to add capacity to the southern and eastern lines, now that their platforms have been lengthened to accommodate six carriage trains. This was outlined in a recent Auckland Transport media release:

Longer trains are being added to the rail network to help cater for the increasing popularity of Auckland’s public transport, which was up eight per cent for the year to 30 April.

From 17 July trains will use five and six carriages on the southern line, following the completion of platform extension works. Six carriage trains began operating on the Western Line last September. Longer trains allow more passengers on each service.

It’s worthwhile to note that under the current plans, these will be the last additional bit of rail rolling stock capacity that will be added to the network until electrification in 2013/2014. Unless we can find some ‘stop-gap’ measure to get more trains (or longer trains) on the network, my understanding is that from July this year until the new electric trains are operational in 2013/2014 we will have to manage with the same number of trains. An interesting prospect if rail patronage continues to grow at 10-15% a year. This issue was noted in Auckland Transport’s April business report:

I‘m still yet to quite figure out how the further optimisation will work in early next year. Not only will this include the improvement of peak time frequencies on the Western Line from a train every 15 minutes to a train every 10 minutes, but it will also include the introduction of trains to Manukau Station – presumably achieved by extending all the current ‘short-runner’ services to Otahuhu all the way down to Manukau before terminating them there. With the longer running time between Manukau and Britomart compared to between Otahuhu and Britomart, there will clearly be an increased number of trains required. I’m starting to think that having six carriage trains could be a pretty short-lived exercise – as from next year the carriages will need to be distributed to a larger number of trains.

As I have discussed previously, the concept of having fare differentiation between peak time rail travel and off-peak travel is a good one. Shifting some of the “peak of the peak” into times just before and after the main crush of passengers means that you can use your existing rolling stock more efficiently and effectively. Adding off-peak services is pretty easy as you don’t need more rolling stock and you don’t need more track capacity – you just need to work the system at the peak level for a bit longer. Getting 15 minute inter-peak frequencies on weekdays, longer peak-time frequencies in the evenings and at worst half-hour frequencies on all lines at weekends would be easily and quickly achievable without having to purchase any more rolling stock. Having hourly weekend frequencies on the Western Line, and no trains past Henderson on Sundays, it just downright stupid – as the Western Line passes near five large shopping centres (city centre, Newmarket, St Lukes, New Lynn and Henderson) and could be hugely popular on the weekend.

But even with fare differentiation and better off-peak services, I think by the time we get close to the rollout of the new electric trains on the network things are going to be pretty squashed. Which is why I have found the never-ending delays to the electric train procurement process so utterly infuriating.

The order of electric trains (well, my understanding of it) includes 35 three-car electric multiple-unit trains, plus a number of electric locomotives. The locomotives are necessary for the obvious reason that we have a large number of SA/SD trains around Auckland at the moment and not only would it be stupid to get rid of them when they have a lot of life left, but also that we don’t have enough money to purchase sufficient EMUs to operate the network alone. My understanding is that at peak times the EMU trains will generally be “paired up” to form six carriage trains, and will operate on the Eastern and Western lines. Single three-carriage trains will operate on the Onehunga Line, while the Southern Line will be served by the current SA/SD trains, but pulled by new electric locomotives rather than the ancient diesel locos that pull them along at the moment. Because there will be a lot of SA carriages available, my hope is that the current platform lengthening exercises being undertaken on the Southern Line will provide for eight-car trains to be operated.

One great irony of electrification is that it won’t actually result in any more trains being operated on the rail network at peak times – compared to what we’ll have in early next year. Until the CBD Rail Tunnel is constructed, Britomart can only handle around 20-21 trains per hour – which means six from each of the three main lines plus two from Onehunga. Of course electrification will enable the trains to be faster, quieter, smoother and longer – which will add capacity to the system – but until we build the CBD tunnel our ability to get more trains into the city at peak times is constrained.

Furthermore, the desire (and need) to operate the network as efficiently as possible will mean that the mixing of diesel and electric trains is likely to be avoided wherever possible – as the diesels obviously accelerate slower and would therefore “hold up” electric trains. So that means all trains from beyond Papakura and Swanson will only be shuttle trains to the end of the electric line. The newer of the Perth DMUs (known as the ADLs) should fulfill this function fairly effectively, and they could serve out to Hupai (or beyond if demand was there) on the Western Line and also potentially beyond Pukekohe on the Southern Line.

So in the period between electrification and the opening of the CBD rail tunnel (so say between 2014 and 2021 if we’re optimistic about the CBD Tunnel) the new electric trains will be used in combination with SA Trains being hauled by new electric locomotives, plus the use of some of the DMUs for shuttle services. Additional capacity could be added (if the trains were available) by running services between the Western and Southern lines directly, without having to have every train go into Britomart. I know that maximum loading points on the Western Line generally occur between Mt Eden and Grafton station as many Western Line users have their destination at either Grafton or Newmarket. So that’s a potential way of squeezing the most out of the system up until around 2021 when the need for the CBD Tunnel will be dire.

Now, if we look at rolling stock requirements post-CBD Rail Tunnel, things become rather interesting. In the CBD Tunnel’s business case a rather bizarre operating pattern was suggested: The option above requires a lot of additional trains, so an interim option was considered that would utilise the existing number of post-electrification trains: I thought both options were far too complicated, and suggested my own operating pattern:

One big spanner in the works of all this is the likelihood that the SA Trains won’t be able to operate through the CBD Rail Tunnel, because it’s too steep (and also something to do with fire-ratings). EMUs can generally handle steeper gradients than locomotives, because the train is being driven from more points – kind of like how a four-wheel drive vehicle provides more control on slopes than a two-wheel drive vehicle. Because the CBD Tunnel is going to be very much at the maximum end of track steepness, it seems that in all likelihood it will only be EMUs that can operate through it.

That gives us a bit of a headache about what to do with all our electric locomotives and SA/SD trains – that still will have a lot of life left in them come 2021. I wonder if Wellington would be prepared to swap some of its Matangi Trains for loco-hauled SA/SD trains? It’s an interesting possibility.

Share this


    1. Bris if you look at my operating patterns they are quite “metro like”. There are two lines that could run at up to a train every 5 minutes. One lines runs from Manukau to Onehunga the other line runs from Swanson to Papakura. They overlay in a couple of places, particularly through the CBD tunnel which is what you want as you want as many trains through there as possible to maximise the investment.

  1. Whenever I’m on the western in the afternoons there are a lot of St Peters kids getting on at Grafton heading west, so any service following the current western line and not using the CBDRL around school times is going to get used, or I guess these kids change at Mt Eden…..

    But are you saying that the electric locos also can’t cope with the tunnel?

    1. But are you saying that the electric locos also can’t cope with the tunnel?

      That’s my understanding.

      Post CBD Tunnel I would imagine we’ll see some west to south services at peak times, probably using the electric locos and the SA trains.

  2. A couple of things
    1. I thought the plan was for 38 EMU’s? (or at least up to 38), we are also meant to be getting 13 electric locomotives for the SA sets
    2. All the things I have seen indicate that the electric loco hauled trains would be 6 cars long so would actually have less capacity than a 6 car EMU as it would be about 24m shorter.
    3. I think that by 2021 the SA carriages are going to be pretty worn out, first remember that they are about 40 years old now and even though they have been refurbished some are starting to show their age already. By 2021 I think they would have needed replacing anyway.
    4. While electrification doesn’t really increase the number of trains it will substantially increase capacity. As an example, a standard 4 car SA set has about 260 seats and is about 80m long (not including the loco). With a tight squeeze you can probably fit about 500 people on a normal SA train. The specification for the EMU’s was 3 car EMU’s at 72m long but holding about 250 seats plus greater standing capacity. A 6 car EMU which will be pretty standard should have about 500 seats with enough space for another 200 standing comfortably (and more if people are squeezed). This represents almost a doubling of capacity which based on current trends will probably take 4-5 years fill.
    5. I did my own calculations on how many trains would be needed and came out pretty close to what is being pursued. We will need about 9 trains (18 EMU’s) to run 10 minute frequencies on the Western line, about 7 (14 EMU’s) for the Eastern line to Manukau, 2 EMU’s for Onehunga which assuming we get 38 EMU’s leaves 4 of which we probably want 2 set aside for regular maintenance and 2 as spares in case one breaks down. We will need about 10 loco hauled trains for the Southern line so that also fits in with the numbers, one spare and two being regularly serviced.
    6. If people are wondering why the Western will solely get EMU’s, it is due to the fact that EMU’s tend to be much better at handling things like hills and curves, something the western line has a lot of.

    1. That’s an easy thing to change around. Very few places use DC anymore so the typical electric set up is for AC. The Matangis have had to have their electrics modified to cope with DC but that could easily be changed back.

      (Well, that’s what I’ve heard from people who should know).

      1. That doesn’t mean they could run easily in Auckland, I think Wellingtons platforms are lower so the low floor unit wouldn’t be level with platforms up here. Also more importantly just because it is an EMU, doesn’t mean it would work in the CBD tunnel, I think part of the reason for going for 3 car EMU’s is to also maximise the power available as you would have 2 powered cars and one trailer compared to one of each like the Matangi’s are. I think it would just be much easier just to order a second batch of EMU’s at the same time as the tunnel that replace the SA’s and allow for frequencies to be boosted (the idea would be they could start coming online just before the tunnel opens).

      2. I think you have the stoty the wrong way aroud regarding DC to AC conversions. To convert a DC unit to 25KVAC you need to find room for a weighty transformer ,rectifiers and lots more insulation. The resulting DC can then be fed in as if it had come from a DC contact wire. To go from AC to DC you just remove the aforementioned components. The current Oscar trains in NSW were to be designed to be upgraded from 1500VDC to 25KVAC for future alterations of the overhead supply on the Sydney-Newcastle run. This provision was cancelled when one of the costs was expressed as some 12 seats per 4 car set.

        1. But they dont have apostrophes 😉

          AC motors or not they will still need transformers, rectifiers, different main breakers etc.

  3. If Auckland-Hamilton-Tauranga has been electrified by 2021 the SA sets and electric locomotives could be used to run intercity services (after suitable refurbishment and adding toilets).

  4. Although the Eastern line would definitely have the demand for 10 minute frequencies off peak, Would Onehunga even in 2020? My understanding was that as it is single tracked it could only take up to 4tph.

    1. Six trains per hour would require double-tracking. Ideally by 2020 we would be starting work on the Airport Line – which is obviously where the blue line would continue to in the longer run.

      1. True. Coming to think of it having only 4 trains an hour on the Onehunga and 6 on the Eastern line would require a 10,20,10 frequency pattern which would be confusing for passengers.

  5. Miggle, for Auckland-Tauranga to be electrified by 2021, the project would need to be authorised in the next 2-3 years. For something that has already been ruled out because of the near billion dollar price tag (for very minimal benefits), it’s highly unlikely to happen.

    My prediction is Pukekohe will be wired up after about 20 years, and beyond Pukeohe in a timeframe nobody alive today needs to worry about.

  6. Josh, one thing that hasn`t been debated about train patterns is when the CBD tunnel is built why not actually start a loop service with Western trains stopping at Mt Eden, Southern at Newmarket and Eastern at Britomart with clockwise and anti-clockwise loop services in a loop taking in Britomart, Aotea, K Rd, Newton, Mt Eden, Grafton, Newmarket, and Parnell? We all know central loop lines operate successfully in cities around the world. Maybe that would be a better way to use our capacity while having very high frequency (5 mins) for the loop which would mean that people could still get to their location in or near the CBD quickly. With a loop line because the regular line services are not going all the way to Britomart (except Eastern line trains) for example their trip time would be shorter and they could be able to turnaround quicker and do more trips. Eight stops/stations I would aslo venture is enough for a loop line to be a viable option.

  7. One small point could our Train COntrol even cope with a 5 min frequency. I would think this would not allow sufficent spacing between trainsets.

    1. The new signalling system currently being installed allows for 4 minute frequencies on all lines and even greater on some sections.

      1. Serious question: how much improvement in operations would this require? I know that signalling faults are part of the current problem, but they really do struggle with current peak frequencies. An increase on this is obviously a step further.

        1. Once electrification is completed we will basically have a completely new rail network, with the only thing really remaining from pre-2003 being the rights-of-way and a few of the station buildings. So the system should work really well.

          I think that we wouldn’t need many more trains to run both my lines at 10 minute frequencies, which means five minute frequencies at points on the network where they overlap. I think over time you could push that to 5 minute frequencies both ways on both lines. Thats a train every 2.5 minutes – well within what’s run overseas.

        2. The system we are getting is pretty modern and all signals and points motors are being replaced as part of the project so the issues of the last few years shouldn’t exist going forward. They are putting in a system that has been designed for higher frequencies even though we may not need it for a few years which means there will be much less barriers to doing this when we need to. My understanding is the system has things like automatic train control, in cab signalling and signal overrun protection.

  8. Ron, if they can do it in other paces around the world, why not Auckland? Of course it`s going to cost money but if/when they build the CBD tunnel that`s going to cost money anyway, so hopefully they could build it with first class signalling and also further upgrade the signalling on other parts of the loop at the same time. Even if they could do say 7min frequencies that would still be enough I think. Kyoto`s subway for example has 7-8min frequencies.

  9. Will EMU’s have three doors/six channels per side? This is an obvious capacity enhancement at the cost of a few seats (through reduced station dwell time).

    You have to make sure that doors are wide enough to be two useful channels. Melbourne’s Comeng cars have three narrow doors a side (about 1450mm opening from memory), effectively three channels which of course defeats the purpose.

    1. Which is to say that if we’re not going to buy sufficient electrics to run a tunnel, we shouldn’t have one.

      This is pretty much the case in Auckland and New Zealand. We start doing something good, but because nobody with power really cares about getting it right, we get the “cheap” option which ends up causing far more difficulty and cost in the long run.

      1. The cost of the tunnel included more EMUs such that we would have an entirely EMU fleet, also the new trains coming in a couple of years will be the ones running through the tunnel. You also seem to forget that the SA sets will be pulled by electric locomotives come electrification with the current diesel ones released back to Kiwirail and either used for freight duties, as the shuttles trains or scrapped I guess.

    2. There has been and never was any plan to run diesels in the tunnel, there will be EMUs running in there.

      1. Yeah but not electric locos? Seems like a limit on the network, where will they run once the tunnel is in?

        1. They can run North-West for instance or South to Britomart – I agree it’s a limitation, which is why the plans for the CBD tunnel include new EMUs, which unfortunately is something Joyce has latched onto to try and claim the costing is gold-plated.

  10. Also they hiring enough drivers and conductors _now_?

    Because if we get to a situation in 12 or 24 months when capacity is constrained by staffing issues, I am going to figuratively, non-violently, kill someone.

    1. Okay, so perhaps that wasn’t my most constructive comment ever. But there are a lot of things that are entirely predictable, and if they’re not done for whatever reason (funding constraints, disorganisation) it starts to get a little ridiculous.

    2. AT knows how many trains they’ll have in 12 months, basically not many more than we currently have, so I don’t see why staffing issues would affect us in 12 months anymore than now….

  11. By the way. Your solution is far simpler and easier to understand and better linking the 4 exisiting lines through britomart and the CBD tunnel. Like it.

  12. Quick question. Why did you opt to have a west/south line and an nehunga/east(manuaku) line rather than a west/east(manaku) and a onehunga/south line (in the direction hat the southern line goes first to newton, and onehunga first to parnell)?

    -Currently a west to east passenger needs to transfer and my option will give them a one seat ride, they would go via britomart anyway.
    – a west(large catchment) to south (large catchment) passenger will only require a single transfer to avoid the frequent stopping, local CBD traffic jammed tunnel. (with admins plan it would need 2 transfers). This is a case where transfer would save time over riding the whole tunnel via britomart.
    – red line would be much shorter, and hence (hopefully) more reliable

    -a bit more complicated pictorially, could be confusing for stations between new market and the beginning of the onehunga branch (covered by twice by the same line)
    – 2 transfers required to go from the west to onehunga (small catchment anyway)
    – with admins route larger trains could be put on on the high demand (many station) west/south line, and shorter ones on the lower demand onehunga-manuaku route. My plan would need roughly even rolling stock allocation as each line has a busy component.

    What have i missed?

  13. Hi Guys:

    Could you get me right on the calculation of the train capacity.

    At the monent at Britomart we are having 32 trains per 2 hour peak time during the week. For electrified train each could carry 800 passeger. Then the capacity of 2 hour Peak time for Britomart with the same frequency shall be 25000 . Is this correct?

    Then the annual peak time (4 hours) capacity shall be 2x25000x5x52=13312000

    This number seems not right

    Please help me to get it right if it is wrong.



      1. The problem we are having is that nobody knows what the real patronage is for the peak hours. I could not find it from either APB&B report or Auckland Transport committee report. We can only guess the ratio of the patronage between peak time and non-peak time. I do not think that was so difficult to generate such more meaningful data (distinguishing the number between peak and non-peak hours) from the authority. I feel there was something wrong with it.
        Anyway according to the survey results provided by Auckland Transport a typical whole day boarding number at Britomart is 10148. This is a typical weekday data. Let us use it as an average everyday’s data then the annual boarding number shall be 10148×365=3,704,020. Considering in and out number would be roughly the same so the total annual patronage for Britomart would be 7,408,040. As we have already worked out the total capacity of Britomart during 4 hours peak time (5 hours if afternoon peak hour starts at 3 – 6) is 13 million once the electrified system is in place. So we need to find another 7,000,000 people to make Britomart to be really crash crowded. Bearing in mind 7,408.040 is annual patronage for whole day including weekend and public holiday. And 13 million is the number for just 4 hours of the peak time. That is why I was afraid my calculation maybe wrong.

    1. The electrified system will have 6 tph on each of the 3 main lines and 2 tph on the Onehunga line, this gives us a total of approx 40 trains arriving at Britomart in a two hour peak. A single EMU has about 250 seats and based on Auckland Transports max loading target of 1.4 each EMU should have a maximum of 350 people (we know a trains can have more than this but to make a service attractive we don’t want overcrowding).

      We know the Onehunga trains will be single EMU’s and that the Eastern and Western lines will be mostly double EMUs. The plan for the Southern line is for it to be run by 6 car SA sets pulled by electric locos, they have a capacity of about 560.

      Based on all of that your estimate of 25000 capacity over the two hour peak is about correct (just you got to it a different way) and therefore the annual estimate of 13m peak capacity is roughly correct. The CRL study estimated that max patronage on the system would be about 22m however I suspect much of that will be off peak. Also remember the afternoon peak is longer but more spread out, it starts at about 3 when schools start finishing and carries on till about 6.

  14. I mean too low or too high? About a million a month then, in March 2011 the gross ridership for the whole system, and remember most journeys involve Britomart [one of the problems of have a Terminus system] was over 1 million. So already the demand is there, with electrification and integrated ticketing, and two more stations this shows that we will already be at capacity well before the CRL is built.

Leave a Reply

Your email address will not be published. Required fields are marked *