What can Sydney teach us about airport rail lines?

Mode share (prepared by ACCC)

There is little doubt that Melbourne Airport needs action to improve land-side access for passengers arriving and departing from the airport.

Many observers argue the solution is a rail line from the CBD to the airport. I think there’s a much bigger picture they’re missing. They would be well advised to look at the Airport Monitoring Report 2009-10, just released by the ACCC (see chart).

It shows that only 39% of trips to Sydney Airport are made by private car (on-airport parking, rentals and kerbside drop-off), compared to 69% for Melbourne Airport. Since Sydney has a train and Melbourne doesn’t, it’s tempting to conclude that a train is the answer to Melbourne’s woes.

However the ACCC’s report says that more people travel to Melbourne Airport by public transport (14% – all by bus) than is the case for Sydney Airport (12% – train and bus).

A key difference between the two airports is that taxis (incl ‘mini buses’) are far more popular in Sydney, where they account for 49% of all airport trips. The comparable figure for Melbourne is just 17%.

Part of the reason for this difference is taxis are more competitive in Sydney against cars and against the train – Kingsford Smith is 8 km from the CBD and hence is relatively central.  In contrast, Melbourne is 22 km from the CBD so taxis are not as competitive with either buses or cars (other reasons for the difference include more tourists at Sydney, as well as higher parking charges).

As I discussed last week, Brisbane’s airport – like Melbourne’s – is also located a considerable distance from the city centre. It might be that the location of both airports on the edge of their respective metropolitan areas – well away from the centre of gravity of population in both cities – is a key reason for their high private car use (and low taxi use).

Yet distance can’t be the whole explanation. The Brisbane airport train only captures 5% of trips and all up, public transport carries 8% of airport journeys. That’s considerably less than either Sydney or train-free Melbourne.

Given the experience of Sydney and Brisbane, it cannot simply be assumed that constructing a rail line from the CBD to Melbourne Airport will inevitably lead to a significant increase in public transport use – at the expense of cars – over and above the already substantial mode share enjoyed by buses.

Or more to the point, it cannot be assumed that the size of any mode shift resulting from the provision of a rail line to the CBD would justify a likely cost in the region of $1 billion.

I think there will inevitably come a time when the demand for movement between the airport and the CBD exceeds whatever extra capacity can be eked out of the bus system. That will require some higher capacity transport technology – probably rail-based – to supplement or replace buses on that route.

But the critical point is that even a rail line would still only carry a small fraction of all travel to and from Melbourne airport. Like the residents of Brisbane and Sydney, the great bulk of Melburnites – whose trips start or end all over the metropolitan area – would ignore the train and continue driving to the airport.

The key challenge facing Melbourne Airport isn’t any gross failing in the ability of buses to transport people to and from the CBD. We’re already doing reasonably well at that. Rather, it’s how to move growing numbers of people from dispersed suburban locations to the airport and back again.

It’s analogous to the transport problem presented by the CBD but the solution is far more difficult. The CBD is an order of magnitude larger than the airport and located at the centre of the metropolitan area, not the edge. The CBD wasn’t established in the post-war era – it has had generations of investment in radial train and tram infrastructure starting from the days when cars did not even exist.

What we need to do is think much more imaginatively, widely and deeply about how we’re going to manage rising demand for air travel and the associated increase in land-side movement to and from the airport. If we just focus on replacing CBD buses with a CBD train we’ll miss the much, much bigger transport problem looming ahead. Public transport will undoubtedly have a key role to play in addressing that problem but the form and scope it takes should be a function of the problem.

(I’ve discussed a rail line to the airport on a number of occasions before e.g. here, here and here),

30 Comments on “What can Sydney teach us about airport rail lines?”

  1. Mahyar says:

    Is there any evidence that air travel will increase though? My instinct tells me the opposite.

  2. Matthew says:

    I’d think it better to compare how many turn up by rail to cities that have done airport rail right, i.e treated it as part of the normal PT system and serviced it with frequent services at normal (non-jacked up) ticket prices. (Brisbane and Sydney fail on that. Adelaide’s J1 bus is better, or Perth Domestic)

    So how many turn up by train to Heathrow, or Copenhagen or Zurich? What about Schipol? How about JFK? or CDG?

    • Alan Davies says:

      Comparisons with Australian cities are preferred because they minimise a range of cross-country differences e.g.historical and cultural.

      Sydney and Brisbane underline the point that it is unlikely in our modern political culture that any rail line from the CBD to Melbourne Airport would be priced at a standard Zone 2 fare (except maybe for permanent airport workers, because that would be smart politics and wouldn’t cost a lot).

      • Matthew says:

        Ah yes let’s repeat Sydney and Brisbane’s mistakes, rather than learn something from somewhere that does it better. With that kind of thinking Brisbane would copy Melbourne’s Bike Share mistake with a scheme requiring helmets. Oh they have. Perhaps we could get a zinc bulk carrier like goes up the Derwent to crash into the Sydney Harbour Bridge.

        Yes I’m being flippant. But really do you think we’re all so dim to put in second best when we could go for best practice? On the evidence of what we already have I’d guess you’d win the argument.

        My version of Fantasyland would be so much better to live in.

      • Sam says:

        I think comparisons to Canadian cities are worth while their cities are similar to ours (largely car dependent and suburban) but unlike us they still manage some well thought through land use planning (Vancouver) as well as well integrated and effective public transport (Vancouver and Toronto).

        On Matthew’s point below Australia does have a habit of not doing things properly whilst claiming ‘worlds best practice’ this city does need an injection of foreign expertise.

  3. rohan says:

    It is a lovely idea, but yes the bus is fine for the moment. But I would like to know why it would cost $1 bill. If it were a stop on a new fast rail sydney-canberra-melb train, would that help? Or it was a train that ran from say the Dandenong line all the way through. Chicago would be in interesting comparison (though passenger numbers are huge) – its a normal price part of the system train that takes 40 mins. Runs down a freeway I seem to recall.

  4. Joseph says:

    “I think there will inevitably come a time when the demand for movement between the airport and the CBD exceeds whatever extra capacity can be eked out of the bus system.”

    I don’t follow your logic here. You make the point that airport traffic is expected to double by 2030. Currently the Skybus timetable has a bus every 10 minutes. To keep proportions the same implies a bus every 5 minutes by 2030. Why would this be difficult to achieve?

    • Matthew says:

      10 minutes piffle. I’ve stood in line for over 30 minutes on a Monday morning waiting for a Skybus only to not fit on that one and have to wait for the next. Then the joy of sitting in a snarl on the Tulla freeway getting shouted at by some bimbo on the video screen telling us how wonderful the DFO shops are. Rubbish service, I say.

      It’s good to see there is a new bus to Broadmeadows station. I’d be on that every time.

    • Alan Davies says:

      I was thinking more about car and truck traffic growing so that the peak period extended and an increasing proportion of buses took longer. But good point, ‘eke’ is the wrong word.

  5. brisurban says:


    Brisbane Airtrain is an asset to our city. A taxi would easily set you back $50. We have buses that leave from Roma Street, but I never use them, I prefer to take the train.

    The problem is the high fares, the frequency of service and scope of hours stopping at 8pm at night. Combined with building a competing parallel road, I feel the real mistake lies not actually in building the infrastructure (there are many successful rail lines to airports) but rather having it done as PPP rather than have it integrated into the rest of the rail network, with the same fares like any other train line.

    Oh, and the other thing- it surcharged me when I used my GoCard and transferred from my train. It does this to everyone who uses GoCards. The line itself was rather cheap to construct at $220 million at the time, which works out to be 27.5 million/km at the time (year 2000?).

  6. Good article Alan.

    I’d be interested in seeing what it might cost to build an underground bus interchange right below the terminals.

    If it was done it should obviously be constructed in a way that easily allows a train terminal to be added when a train line does become more viable.

    Currently I believe all buses terminate at Terminal 4, however looking at satellite photos I can’t see where this happens as I can’t see anything that looks much like bus bays.

    Frequent smart buses departing from a range of northern suburbs as well as some frequent buses from North Melbourne Station would probably catch a fair chunk of people travelling from the northern and western suburbs. Those coming from the eastern suburbs may as well head via the CBD, as avoiding it probably will not save them too much time.

    I’m sure the bus interchange would be expensive, but given that at some point it would double as a train terminal that will almost certainly be built anyway at least it’s money that will be spent at some point anyway.

    The fact that it might “cut the walk” from the carpark, it might end up luring quite a few people that would otherwise drive. Although I still believe the price should be no more than a standard Zone 2 fare, despite your arguments otherwise.

    If the aim is to shift people away from other modes there need to be incentives, the easiest incentive to implement is a financial incentive.

    • brisurban says:

      I have to say I’m not convinced by this ‘dispersal’ argument. What city doesn’t have dispersed trips?

      The “squaresville” example Dr Paul Mees refers to in Transport for Suburbia is for a model city which has 100% dispersal and random trip origins and destinations.

      That is the worst-case scenario possible; I would expect real-world cities to therefore be a transport task easier than that.

      • Alan Davies says:

        I think there’s an important distinction to be made between PT being ‘viable’ in low density areas (which might mean a mode share of just a few percent) and PT capturing a really large share of all travel from those suburbs. Like the CBD, the airport will ultimately need PT to have the lion’s share of travel in order to function efficiently, given projections of future demand.

  7. brisurban says:

    I’ll just point out for the benefit of readers of your blog:

    1. The Brisbane Airtrain is a profitable operation
    2. It was built for $220 million at no cost to the taxpayer.

    I’m not familiar with the details for Melbourne, you have a much bigger city and I haven’t seen details of the planned rail route. So I will have to read up on those.

    Many of the points you raise appear to have been raised by BusVic: http://www.busvic.asn.au/database/files/BusSolutions3.pdf

  8. Eli Gescheit says:

    When you have parents who are dependent on their cars for travel, I find it difficult to understand how their children will embrace public transport. In my opinion, education on sustainable transport should be encouraged at schools. Proposals to improve public transport networks, such as providing designated wide bicycle paths in the CBD, do not address the reality that Sydney residents are unwilling to give up travelling in cars.

    For example, I live in a suburban street in northern Sydney and the majority of the cars parked on the street have P plates on them. So as a new generation of drivers are getting comfortable with driving, the city roads are even more congested and there are less opportunities to improve the public transport. I use public transport personally and anticipate that without proper education and incentives, Sydney will only suffer more headaches and more State government planned projects that are never delivered.

  9. MattyQ says:

    The other important aspect to keep in mind is that the airport line in Sydney also serves as another needed route into the city from the southwest, and although this could have been achieved much more cheaply by sextuplicating Tempe to redfern, it gives the airport line another role rather than just being an airport transit link. Green square is also supposed to be Australia’s largest urban renewal, although it is proceeding at a snails pace

  10. […] to the ACCC, even at the two Australian airports that have direct train services to the CBD – Sydney and Brisbane – rail’s mode share is no higher than 10%. Melbourne Airport also has attributes […]

  11. Paul Grgurich says:


    Whether or not rail becomes a viable solution or not will depend on whether the line is looked at as a stand alone service or is incorporated into the broader metropolitan transport strategy.

    Brisbane’s airport’s rail line is a failure because 1) it is a dead end service 2) it is actually slower than taking a taxi 3) the CBD train station is not in the business district

    Sydney’s airport rail line is a failure because 1) it is clogged with commuters from outer suburbs 2) the cost is considerable and with 2 passengers it is certainly cheaper to take a taxi 3) the state of Sydney’s rail system is appalling and business passengers simply don’t want to use it. The rail service does have the advantage of being reasonably integrated into the greater rail network and the city stations do circle the CBD. Also despite Sydney airports proximity to the CBD, the congested road network does infact make train travel considerably quicker on almost all occasions.

    Melbourne does have an advantage here, in that it can learn from the others mistakes. The cost of the rail line will be significant. I think $1B is pretty light on – it will be double that. Unfortunately if the State looks at this project using a DCF analysis, we will end up with the Sydney model and be stuck with a cost per ticket that no one will pay. Unless you accept that the upfront cost is to be written off – much like the new line out to Geelong – the project will fail. The ticket prices need to be a standard zone 2 cost.

    The next issue that needs to be tackled is one of integration. Part of this problem has been solved by the construction of the new high speed line out to Geelong. This will see Airport trains travel on a new dedicated line as far as Sunshine. From there the line will need to be built along the Bendigo/Sunbury line past Albion, up the goods line corridor and then follow the already reserved land into the Airport. Here is where the Government can make positive decision. If they extend the rail beyond the Airport to Sunbury – all Bendigo trains can be re-routed via the airport. This will free up the Sunbury line for commuter trains and improve efficiency and capacity. Some Airport services would terminate at the airport and some would pass through on their way further out.

    The next issue is at the city end. Successful European Airport services don’t terminate at the CBD – they pass through it and this should be no different. If you want suburban families to take the train – then there is no point ending the train in the CBD. It need to keep going out to Dandenong, Pakenham, Camberwell etc – and you need to provide secure parking at these suburban locations.

    On the negative side, Melbourne’s freeway system is quite advanced – despite obvious congestion issues – and this makes commuting to the CBD by road, quite efficient and probably from a time perspective it will often have an edge over rail – no matter what the speed of the train. And yes Melbourne covers a very large urban area and there will always be significant areas of the city that just are not adequately serviced by rail.

    It is reasonable to target approximately 25-30% as the desired number of trips that could be undertaken by rail. Whether this justifies $2Billion in expenditure is a different question. It would definitely be a better investment than the $4.5B on the Geelong line or the completely farcical Avalon rail link!


    • Alan Davies says:

      Thanks Paul, you make some interesting points. My immediate reaction:

      First, the fact that both of Australia’s two airport rail lines are failures for various reasons suggests to me the idea simply isn’t that robust. Two out of two is not good. Yes we can learn from their (known) mistakes, but at what cost? And we can’t learn from the mistakes they didn’t make!

      Second, I agree a max Zone 1-2 fare would generate more patronage, but it would also generate more subsidy. That’s not an argument for a rail line – the first question that suggests is: should Skybus fares be Zone 1-2? I’m doubtful about the warrant for subsidising airport fares, as I said here. I think there are higher priorities for taxpayers money than replacing Skybus with a train, at least at this time.

      Third, I doubt the extra patronage generated by linking with the Bendigo line would be sufficient to offset the additional cost. Bendigo travellers bound for Southern Cross might not appreciate the extra distance. As the Regional Rail Link won’t be electrified, airport services wouldn’t be able to go beyond Southern Cross (or even use the city loop, making an airport service a bit like Roma St).

      Fourth, 25-30% mode share of all airport trips is a massive number and probably could justify an immediate start on a rail line, but note that Brisbane and Sydney get nothing like that. Even work destinations east of Punt Rd don’t get that sort of mode share despite subsidised fares and a very rich train and tram system.

  12. Paul Grgurich says:


    Most infrastructure projects are subsidised – it’s just by how much. This is really a philosophical debate as to whether you will subsidise the construction of a road or a railway. A tollway such as Eastlink was heavily subsidised as the Government purchased all the land and that cost $100’s of millions. In Sydney it costs an average of approximately $500/m2 to acquire land (slightly less in Melbourne and Brisbane). This means a 10km rail corridor costs $100M in land and a road corridor $250M. Then you have to build something.

    All roads are subsidised, they are either free (100% subsidised), or, when you include the cost of acquiring the land, you find that the toll offsets only 40-60% of the total cost. Rail is no different.

    The value in linking the Bendigo line is not to gain patronage from Bendigo residents, it’s to free up the line on the city side of Sunbury for commuter trains only. The current line is overcrowded and the Bendigo trains are stuck behind the commuter trains. The government is facing two choices, build an extra two rail tracks adjacent to the current ones, or, if the airport line is built, then extend the line to Sunbury. The previous government was working through this analysis. Your point about the Bendigo train having to cover extra distance is not really an argument. The distance change is immaterial, but the speed that the train can travel will double.

    I am not sure where you are going on your point about electrification. The current commuter trains (electrified), country trains (diesel) and goods trains all run on the same tracks. The country train to Geelong just keeps going through Southern Cross and Flinders St Station and becomes the Traralgon Train on the Caulfield line. The Bendigo train runs on the same track as the Sunbury train. They can also run on the city loop if need be. Extending the Airport train beyond the CBD is imperative for its success as people need to be able to park their car at the station and hop on the train – or at least not have to catch a taxi too far.

    The Avalon airport service will most certainly will be diesel service. It’s likely that they will use the same trains for both – it’s cheaper.

    As to the % of train travel – 25-30% is the level they should be aiming for. I do agree that if they can’t convince themselves that they can achieve this, then forget it. Oslo is an example where the train service is 35% of the total. It’s as much a mind set as anything.


    • Alan Davies says:

      Paul, I don’t get the relevance of your rail vs road argument in the context of the airport rail debate. My point is that the cost of replacing Skybus with a train is not warranted at this time. The actual, real cash required from the State budget for a train would be much higher than the implicit subsidy to Skybus for roadspace (Skybus also contributed to the cost of freeway works). I acknowledge that there’ll be subsidies, my concern is about the size, not the principle.

      I also don’t get your argument that an airport-Bendigo line link would “free up the line on the city side of Sunbury”. Isn’t the RRL going to provide a new set of dedicated VLine tracks from Sunshine to SXS?

      I acknowledge your point that VLine diesels can indeed run from SXS to FSS. But that still leaves the question of finding enough train paths to run both airport diesel and electric commuter trains on the Pakenham line. And if the airport diesels were to run on the loop, what would make way?

  13. Paul Grgurich says:

    My comment on road vs rail is just to emphasis that we are just choosing what to subsidise. Agreed that the cost of rail would be $2B, rail cant compete if looked at on that basis. I dont believe any rail line could.

    Wrt Sunbury – I am not certain of this now we have a new Government- but you may be correct.

    Wrt rail congestion. Fair point – Stage 2 of the submission to IA would be required, which would see country trains have a dedicated line as far out as Westall station.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s