While the fact that transport has become the number one issue in the current Super City elections doesn’t surprise me, one thing that has surprised me a bit in recent times is the level of focus on getting rail out to Auckland International Airport. In particular, it was very interesting to see in a recent NZ Herald article, more people considered the Airport Line to be Auckland’s number one transport priority than any other project. Support for the Airport Line was even ahead of the exceedingly all-encompassing “improving the roading system” – by 23.5% to 18.4%. With both major mayoral candidates (Len Brown and John Banks) supporting the CBD rail tunnel project, Len Brown’s support and John Banks’ opposition to, Airport rail is shaping up as a reasonably defining difference between the two candidates. I must add that it’s a bit odd to see Banks now opposing both Airport Rail and a North Shore railway line when he trumpeted the idea of both to the Campaign for Better Transport back in July.

So what do we know about this Airport Line project? Is is likely to be economically feasible? Will it go bust? What previous work has been done to analyse its cost, its benefits, whether it should be light-rail, heavy rail or a busway? And how about whether it should link to the east, to the north, or both? Perhaps most importantly, what is a reasonable timeframe for the project – and what else needs to be done first, if anything, before such a project can go ahead? These are the questions that I will try to shed some light on in this post.

In terms of concrete background work already undertaken, there are a couple of things that are probably worth noting. The first is that as part of their Manukau Harbour Crossing Project, NZTA build the new section of the Mangere Bridge strong enough to carry a railway line across it in the future – even if they did so reluctantly and only because the Campaign for Better Transport took them to the environment court over the issue. I’m not exactly sure of the detail about where the line would go – presumably it would be slung underneath the main roading deck – but a potentially very significant cost (that of building a new bridge) has been avoided thanks to something highly unusual in Auckland – some future proofing.

The second thing that has happened so far in terms of previous work is a study undertaken by Beca for ARTA back in 2008. The study looked at alignment options, it looked at the question of busway, light-rail or heavy rail, it looked at whether the line should link to the east, to the north or both, and made a number of recommendations. It’s well worth a look through. In short though, the study recommended heavy rail, even though it was the most expensive option – because of the ability to link in with the rest of the rail system and also the ability to carry many more people than either a busway or light-rail. The busway option was the cheapest, but was considered to have minimal benefits over and above what we have now (plus there’s always the question of what the heck to do once all these buses get to Onehunga, it’s not like we can build a brand new busway across the isthmus easily). Light-rail was calculated to cost nearly as much as heavy rail, but only offer benefits similar to a busway: the worst of both worlds. This is shown in the table below: Three different heavy rail options were analysed: with access from the northeast (via Onehunga), access from the east (via Puhinui) and a loop that gives access from both. The preferred option was both, with the route shown in the map below (the aqua coloured line) – including the Avondale-Southdown line that could be constructed separately. Now of course such a significant project will not come cheap. The whole line shown above (including the Avondale-Southdown section) is estimated to cost in the region of $2.2 billion. The Onehunga to Manukau via the Airport section makes up around $1.4 billion of that total (why does every big transport project cost $1.4 billion???) So we are talking some serious money here – even if the airport paid for its own station and a reasonable chunk of the cost of building the line within its land boundaries. Vancouver Airport paid $300 million of the $2 billion construction cost of the Canada Line (Canadian dollars), so it seems reasonably likely there would be a case for the Airport to chip-in.

So what sort of “bang” do we get for our buck with this project? Well this is the difficult one to calculate. If we look internationally, there has been mixed success with lines to airports. Even in Australia, the railway line to Sydney Airport has under-performed, whereas the line to Brisbane Airport has just reported a massive increase in its profit: yes, it makes a profit. Given that Sydney is a much bigger city than Brisbane, and has by far the busier airport – the argument that “Auckland isn’t big enough to warrant rail to its airport” may not necessarily be relevant to the debate. The cost of different sections, and their cost-effectiveness, is analysed by the table below: Looking at the “Peak Section Demand” column, the numbers are fairly significant when you consider that at the moment around 6000 passengers arrive at Britomart during the AM peak. Also remember that each rail trip in the Auckland region generates $17 of benefits to road users, and the numbers start adding up. Furthermore, many of the trips most likely to be taken by rail: those for business travellers, would be the exact trips where the time savings, greater reliability and CBD destination would be most valuable. Other major users would be tourists, and I must say as a recent tourist myself I really did appreciate being able to catch a train from the airport into whatever city I was visiting (possible at times, not possible at other times). There is also, of course, always the question of what the wider benefits would be, does rail to the airport make Auckland a “proper” world-class city? What are the flow-on benefits of making travel between the city and airport easier for tourists, business-people and various others? I don’t know, but they could be quite significant.

Furthermore, airport travellers would only be a part of the potential patronage of the line shown in the map above. The airport area has over 10,000 workers – one of Auckland’s greatest employment concentrations. The line would also offer a rail option to those living in southwest Auckland, whose current bus services are some of the worse and most complex in the city.

However, one thing that I cannot stress strongly enough is that we have to build the CBD rail tunnel first. Britomart is nearly at capacity, as by the time the Western Line’s peak frequencies are increased to one train every 10 minutes Britomart will be at capacity. This means at peak times you could only run a train to the airport every half an hour (basically as an extension of Onehunga Line services). I think this would be insufficient, particularly considering future increases in rail patronage.

So do I support rail to the airport? Well I would like to see some further information – both in terms of analysing its projected use a bit more, but perhaps more interestingly, an analysis of the wider benefits to Auckland of having such a link. I’m fairly confident the numbers would come out to be reasonably good, so I do support the project. The big question to answer though is “when?” and “what priority should this have?” In my opinion the CBD rail tunnel has to come first. Not only is it essential to make the airport line operational, it is probably a more justifiable project. Airport Line would be number two for me though – and I see no reason why we shouldn’t aim to have it under construction within the next 5-10 years, and certainly complete before 2025.

Share this


  1. Of course the CBD loop is the priority, even Banks says he is in favour of it, although he does insist on referring to it as a ‘train set’. But if through some crazy alignment of the planets the the Southwestern loop [airport line] were finished first services could run direct to the Western line which would mean a change at Newmarket for some Britomart bound passengers. Less than ideal but do-able. The South so desperately needs real transit action and, for me a great benefit of the this line even being discussed seriously is that it makes the next step over in the South East so much clearer. It would help head off talk of ‘lite’ rail there or possibly even more motorways. The planning and protecting for these routes needs to proceed urgently and can surely run concurrently. I think this line is a winner; Manukau City Centre has always sat apart from its hinterland, that’s why it hasn’t really worked, and Mangere is sadly cut down the middle by the motorway, this line will bring functionality to these undervalued parts of South Auckland.

  2. Slight correction:
    Option 2B was rail from Otahuhu to the Airport, not a single link from Onehunga to the Airport.
    Option 2A was a single rail link from Onehunga to the Aiport but it did not pass the first stage of evaluation (thus why it is not in the weighted table of attributes above).

  3. edit: in my last post weighted should have been unweighted.

    Also, the first table you used to show that “Light-rail was calculated to cost nearly as much as heavy rail, but only offer benefits similar to a busway: the worst of both worlds.” does not deal with any costs, the table you are looking for is at the top of page 23.

    1. Those are operating costs on page 23 (and light-rail comes out worst again). That particular statement was from memory…. I remember being somewhat surprised by the “worst of both worlds” outcome for light-rail.

  4. @ Admin

    I don’t think the costs on page 23 are just operating costs. They are annual costs which I would guess include the annualisation of capital costs component listed on table 8, page 22.

    I wasn’t disagreeing about the outcomes, as you point out both tables have busways and light rail at the bottom. Rail to the airport is definitely the better option.

  5. We need to learn for the roads lobby. Build the rail link first if that is what’s easiest to get off the ground. Then the CBD rail loop will become an obvious priority project to complete and eliminate congestion on the rail network.

  6. Even without the CBD tunnel I don’t see why it wouldn’t be possible to have trains running from out West to the airport via Newmarket (thereby bypassing Britomart) as well as trains from oit South to the airport via Penrose (thereby again bypassing Britomart). This would allow for more frequent airport trains, prevent the overloading of Britomart and provide a more convenient option for people from West and South to get to the airport without the need to change. You could of course still have the half hourly trains currently terminating at Onehunga continuing to provide a direct Britomart to Airport service.

  7. Admin, according to the Statistics NZ there are 22,000 jobs in the airport precinct, split roughly 50/50 between the airport itself and the industrial estate to the north (both of which would have stations on the line). This shows the importance of having regular pricing on this line, most of the day to day users will be commuter workers who aren’t going to pay $15 dollars each way!

    Actually I wonder if two airport stations would be a good idea, one at the terminals and a second basically for workers at the eastern end of the area. There are some big workplaces closer to the Puhinui Rd end which are a good 1.5km from the main terminal building.

    I would also suggest that this line would be possible before the CBD tunnel. If we assume the two slots from the Onehunga line get rolled over then that is only 2 to 4 more slots needed per hour at Britomart, which may be quite possible once the signalling improvements have been made. Furthermore, a success on the airport line would almost guarantee the CBD tunnel was built, but you cannot say the reverse is true.

  8. Although I don’t agree that a busway would be a good option, the report uses an average speed for a busway of 20-50kmph (and the number of proposed stations seems to be the same as rail) which appears quite underestimated compared to, for example, the Northern Busway. What would the average speed on the Northern Busway be?

  9. Thats average speed including stopping at stations. Albany Station to Britomart is 16km is 28 minutes, or about 35km/h. Even just the full busway section Constellation to Akoranga takes 6km in 7 minutes, or 51km/h.

  10. Again the issue with the busway is what to do with all the buses at Onehunga? Does everyone have to transfer to trains for the 25 min journey to the city or do the buses then need to sit in traffic for the 1 hr journey that a bus takes to get into the city. The advantage of rail is that it just needs to be built to Onehunga and can then feed directly into the existing RTN which is not the case for a busway.

    I also completely agree that an airport train costing $15 each isn’t a good proposal.

  11. Thanks for sharing this. A few questions…

    Table 8 shows total annualized costs for the preferred option 2D of $207 million, after deducting revenues of $4.9 million. I’m not sure if I reading this right… does this mean fare income will account for 2.3% of the actual costs of the system and the remaining 97.7% will be paid by ratepayers over the 40 year amortisation period? (in which case comparisons with Sydney and Brisbane would presumably not be very meaningful)

    If that’s the case, is there a good reason for even having fares? Perhaps just to prevent excessive demand (if there is such a thing)?

    How does $207 million compare with the expected total rates income of the super city?

    Table 6 shows $19.90 net operating cost per passenger. Comparing that with annual net operating cost in Table 8 ($34.5 million) shows that the per passenger figure is evidently based on a projected 1.73 million trips/year. If that’s the case then the total annual cost of $207 million represents $119.65/trip, while revenue of $4.9 million represents $2.83/trip. That would make the gross cost $122.48/trip.

    Maybe I’m misinterpreting or miscalculating – these numbers look frightening. Even $19.90/trip operating cost seems high to me.

    Can you tell us what the basis and source is of the $17/trip benefit that you cite?

    1. The $17 per trip benefit table comes from NZTA’s economic evaluation manual. In terms of cost-effectiveness, I don’t really know how detailed Beca’s analysis really was, hence my call for further information.

    2. Mark as I read it the study calculates 33.5 m in maintenance and operation costs annually, for all of the proposed system Avondale to Onehunga, Penrose to Onehunga, and Puhinui to Onehunga. And expects 1.73m journeys p/a. hence a cost per ride of around $19. Capital costs excluded. Seems high to me. I wonder what is included in that 33.5m? And how accurate the journey quantum is looking now, two years on with well over 10% p/a compounding growth on the system as it is- pre electrification and even double tracking. Surely it is safe to assume that extending the reach of the network will increase uptake outside of any natural growth as the system becomes more and more useful. But still this shows how hard it is to mount a financial case for this kind of investment, the benefit is largely on the economic side.

    3. BECA acknowledge in their report that their construction costs are only rough estimates accurate to within +/- 30-50%. And of course the projected demand is just an estimate as well and might increase.

      But if my calculation of $119/trip ratepayer subsidy is correct based on those estimates, and even if the $17/trip benefit were a reasonable estimate, the margin is so huge that I can’t see that any reasonable adjustment to the assumptions is likely to change the basic conclusion… building this system, in my opinion, and again if my interpretation and calculation is correct, would simply be scandalously irresponsible.

      I’m not an expert. I may simply be misinterpreting what the report says, but that’s why I posted the analysis so someone can correct me if I am. So far I’m not hearing anyone say that my understanding is fundamentally wrong.

  12. i suspect the airport won’t want rapid transit to the airport in any form. car parking is one the most lucrative parts of their business and they would not want that to be threatened.

    1. I think it depends on whether the airport can make more money from other uses, like the hotel that is being built on land formerly used for parking. Plus issues like airport traffic contonuing to grow, there’s a point where it would be in the airport’s interest to have the railway line.

  13. All of these surveys have shown surprisingly high support for airport rail (oh to be like National and be able to poll and run focus groups every week!).

    I can see the logic that you campaign for the airport rail line (because it so strongly appeals to the public) as a WAY to get a whole package of improvements to the rail system over 10 years (e.g., CBD rail, Avondale to Onehunga etc etc). Would be a massive campaign but sure the CBT could handle it 🙂

  14. NZTA build the new section of the Mangere Bridge strong enough to carry a railway line across it in the future – even if they did so reluctantly and only because the Campaign for Better Transport took them to the environment court over the issue

    For the record, CBT didn’t take Transit to the environment court over this issue. Our issue was the new motorway south of the bridge didn’t make any room room for a potential rail corridor. Through mediation, we achieved an eastward shift of the motorway at Mangere bridge so now in theory a rail corridor will fit on the western side of the motorway. However this covers only the section from the bridge to Walmsley Rd in the south. After that who knows.

    Thanks for raising the BECA study – I’d forgotten about that.

    I disagree that we have to build the CBD tunnel first (or at least it shouldn’t be a reason to delay planning and designation work on the airport line). Rail patronage is going to grow with or without the airport line, so the two shouldn’t be linked together. Not all trains need to lead to Britomart – if we had integrated schedules and an integrated ticket, then transfers at Newmarket (or even Westfield) should be straight foward.

  15. I also think the Puhinui link to the airport is going to be first off the block as well, just because the land designation is far more straight forward (this corridor can fit entirely in council or airport owned land)…

    1. The big problem with that is you would then have a line to the airport and the government and roading lobbys will say “well you wanted a line to the airport and you now have one you are asking for a second one”. I also think that the public desire for it would be reduce because there is something there, even if it isn’t ideal. Personally I would use the same technique that has been used on roads, build part of it then say “we have to build X line to complete the network”. It should also be done in stages, perhaps push for an extension of Onehunga to Mangere alongside the motorway, that could be done in 5-10 years and start to build patronage from the area then the calls will increase to “lets just get the last few kms finished”

      By comparision coming from Puhunui is all or nothing to the airport as there is nothing in between to use as a staging area.

  16. I would agree the Puhinui link looks tantalising to politicians because it is easiest and cheapest (“there you go, there’s your airport rail, now shut up”).
    However I think it would be a serious mistake and a failure to do that link either first or completely alone. The problem there is it would just be a line to the airport terminals: there would be no station at the airport industrial park (which has as many jobs as the airport itself), none of the three new suburban stations in the Mangere area, no upgraded Onehunga link, and no vastly improved network effect of linking the existing lines together.
    That means the line would have a maximum pool of only 10,000 jobs (two way) and 30,000 traveller trips (one way) to supply patronage. That equates to 25,000 return trips a day at the absolute most. It would require around 50% of everyone going to the airport for any reason to take the train to maintain a ten minute service across most of the day. Obviously that is pretty damned unlikely, but for an airport line to work it needs those quick frequencies and long operational hours. To me that makes it essential to include the other major trip generators on the line from the first instance so it can provide a viable, frequent service for 18+ hours a day.
    A further problem would be the fact that a Puhinui link would be a second branch off the southern line, leaving Auckland with what amounts to one line with three ends. Maintaining adequate frequencies to each of those ends would require either some ridiculous frequencies on the main line or some circuitous routing around the branches.
    IMHO the best solution is the full line from Onehunga to Manukau, ideally built in one go or staged with the Onehunga section first. As Matt suggests to do Puhinui first would end up an expensive boondoggle with low patronage that would probably kill any chance of spending any more on links via the airport.

    Matt, thats a pretty good idea, using the old motorway trick of cutting the salami up one small slice at a time. Cam, you mention there is a viable(ish) corridor as far as Walmsley Rd in addition to the bridge foundations being able to take a rail deck. So rather than the whole airport line perhaps the next big push should be for a 3km long extension of the Onehunga branch to Mangare Bridge and Favona? This gets rail across the harbour to two new stations, it might not necessitate a major upgrade of the onehunga branch (yet), but it would inevitably lead to further extensions to the airport.

  17. Yes, you’re right, a Puhinui only line could be a disaster… too many see this line as only about airport to Britomart when it must be the whole deal. Getting across to a Mangere Bridge Station under the already prepared structure of the bridge would not be too expensive but would be tricky-ish to sell. There’s space for a station and it is only after there that it all gets expensively constrained. But the cost benefit would still be high, remember they couldn’t even get the new Manukau station pushed just that little further to where it clearly needs to be so what are the chances of such an extension now?

  18. The northern busway was built alongside a motorway, and included a variety of re-jigged on ramps, new viaducts and cuttings etc so it should provide a reasonable model for the cost of this. Rail line tends to be a little cheaper than busway actually as it can be narrower without crash barriers or pavement treatments.
    The busway cost $21 million a kilometre for two way route, plus an average of $17 million each for the stations. Therefore the cost of a 3.3km Walmsley Rd extension with two new stations would be something like $100 million plus the cost of the bridge, a rebuilt Onehunga station and extension of the Te Papapa loop.

    So maybe $200 million for the capital works to allow a 15 minute frequency to Managre Bridge and Favona.

  19. I’m not surprised the Brisbane link is making a profit. I’m about to spend 6 hours in transit there with the family this weekend – looking up stuff to do to while away the time revealed that the fare is nearly 35NZD return! 25 minute journey…. ouch! The coffee in Brisbane better be worth it 🙂

  20. While I’m here – in the meantime everyone needs to think harder about how to make the 380 bus to Manukau City work better, when it doesn’t seem to get a lot of use out to the airport proper. A fifteen-minute frequency would be a good start – another would be running a shuttle just to Papatoetoe station where people could join the existing rail service without too much hassle. (In the past I found that a cost/time-effective way to get to the airport from the central city was a train to Middlemore + cab the rest of the way).

Leave a Reply

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