Australia: Pitfalls of Metro Envy

The Public Transport Users Association of Australia has a great press release and analysis on the need to re-think the ideal of a “metro,” and to question why Australians should wait decades for them.

DSCF1486 The major cities of Australia (Sydney, Melbourne, Brisbane, Adelaide, Perth) all have rather similar urban transit networks, dominated by a product that might be called European-scale commuter rail: trains with overhead-catenary power running mostly on a dedicated network.  Within the first 40 km or so of radius, these networks usually run at frequencies of 15-30 minutes, but there’s a huge surge of additional service during the peak commute period to handle the flows of commuters into and out of the central city.  Further out, many of the lines branch, or introduce diverse stopping patterns, which have the effect of increasing longer-distance speeds but reducing frequencies further, at least at minor stations.

Brisbane citytrain A crucial feature of these networks is that they are (a) almost totally radial, offering few if any substantial opportunities to travel between non-CBD centres, and (b) designed to cover a huge range of urban densities with the same basic product.  There’s no difference — in fleet, station facilities, staffing, ticketing, etc — between a train that you’d ride 3 km to a dense inner-city suburb and one that you’d ride 50 km or more to low-density suburbs on the outermost edges of your city.  There is often a difference in frequency — it’s usually lower further out as routes branch or introduce stopping patterns — but not always.  In some cases, such as the Inner West line in Sydney, the denser inner city stations are served by just one stopping pattern at a relatively low frequency, while most trains fly through these stations nonstop aiming for destinations further out.

As commonly happens with commuter rail, the government focus is so much on the peak-period commuter that it often misses out on opportunities to do other good things with its trains and infrastructure.  Density is high enough to support high all-day demand in the inner city (roughly comparable to the extent of each city in 1945) and at key regional centres further out.  But this demand is everywhere-to-everywhere, not just to the CBD, so it requires (a) good connection opportunities throughout the network, especially between trains and buses and (b) high enough frequencies that these connections are worth the wait.  Many of the major inner-city urban rail lines in Australia could easily run every 5 minutes or better if there were a commitment to fund operations and build connection opportunities.

For a variety of political and cultural reasons, many of these opportunities have not been pursued.  Instead, in Sydney and Melbourne especially, we’ve been encouraged to long for something called a “metro.”  Many Australians have been to Europe and often to some of the East Asian megacities.  The word “metro” is meant to refer to the high-frequency, high-capacity rail transit, usually underground, that laces the dense cores of those cities.

It’s been easy to jump from those desires to the notion that since Australia doesn’t have metros now, it needs to build them.  But Bowen’s work in Melbourne (and our own work on the Sydney Morning Herald inquiry) are pointing out that our cities already have a network of grade-separated rail lines covering the areas of European density, and that the quickest way to get a “metro” level of mobility is simply to run these lines much more frequently.

There are some barriers to this (notably the two-man crews still required on Sydney trains) but most would be easy to address if there were a conceptual shift.  The conceptual problem is that as long as our rail bureaucracies understand the peak commute to be their primary product, they will continue to care about running time more than they do about frequency.  Running time matters more than frequency only for relatively long trips (because we tolerate longer waits to go longer distances) and for rigidly scheduled trips such as classic commutes (because we select a particular schedule trip to use).  Everywhere else — throughout the inner city all day, for example — people experience frequency as maximum wait time, and are not willing to wait long to travel short distances.  Even a 15-minute frequency is on the outside of tolerable if you’re just going 3 km or so.  If you just miss one, take a cab or even a bus, and you may reach your destination sooner.

DSCN1552Having said all that, I do like driverless, fully-automated metros, which sever the link between frequency and labor cost and thus make it possible to run every 4 minutes even at midnight.  They have done wonders in Vancouver, not just for mobility but for urban form.  Sydney had a proposed driverless metro that would have broken ground this year, but the Government cancelled it in February 2010 after three years of confusion and acrimony.

In Sydney’s case, the process was too fast.  A long driverless metro line from the city to the northwest suburbs was suddenly announced in 2007, using a corridor that had never been identified for rapid transit in any land use plans.  When that proved too ambitious, it was hacked back to a tiny metro that barely got out of the CBD, which proved to be too disruptive to the life of the city and yet not long enough to do much good.  From the beginning, too, the media treated the metro as a political attack on the existing rail system and its unions.  That wasn’t entirely untrue, but it was certainly exaggerated, and while inter-agency spite can be a powerful political force in planning, most people agreed that it wasn’t a very good basis for a multi-billion-dollar infrastructure investment.

Even if Sydney had proceeded with starting a driverless metro system, the vast majority of the city would have continued to depend on the existing rail product, and would have needed to see further investment in all-day frequency to make the existing lines more “metro”-like.  So now, in a way, Sydney has a clearer path.  We have the tracks, the wires, the stations, and the trains.  All we need is a new commitment to all-day service.  It’s not cheap, but compared to an all-new all-underground automated metro, it’s a bargain.

16 Responses to Australia: Pitfalls of Metro Envy

  1. Daniel April 7, 2010 at 5:01 am #

    Thanks Jarrett.
    One note: Melbourne’s St Kilda Road “Metro tunnel” proposal does not involve driverless trains, though it may bring Melbourne’s first implementation of Automatic Train Control. The plan is to route the existing Sydenham/Sunbury line into the tunnel, with a later stage pushing it through to connect to the Dandenong line.
    We didn’t really flag the need for connecting/feeder services, but you’re right, this is vital. In the Melbourne context, most suburban stations have park+ride facilities which fill up during morning peak hour, so better ways have to be found to feed into the train system for those beyond walking distance of stations.

  2. Jarrett at HumanTransit.org April 7, 2010 at 5:04 am #

    Thanks, Daniel. I’ve removed the incorrect sentence.

  3. mike April 7, 2010 at 7:07 am #

    I don’t get the obsession with the number of drivers. Almost all New York City subway lines require two-person crews and as far as I know no one refuses to ride it because of this.

  4. anonymouse April 7, 2010 at 10:26 am #

    I think a good comparison might be to Copenhagen, which for a long time relied primarily on a regional rail system called the S-Tog, which runs EMU trains on almost entirely separated tracks, and was an almost completely radial system, although there’s a also a western belt line. In the 2000’s, Copenhagen also built a couple of driverless metro lines that run across the city center perpendicular to the main S-Tog line and toward the south and southeast, which is not covered by the S-Tog. In the context of Sydney and Melbourne, any metro line would have to complement the existing regional rail network in a similar way, perhaps serving as a circumferential line, or running in a corridor not served by regional rail, or crossing the regional rail mainline in downtown. In Melbourne’s case, perhaps they could build some sort of light rail which would combine the existing tramway network with new tunnels/els.

  5. Alon Levy April 7, 2010 at 11:45 am #

    One example of a circumferential line would be Paris’s proposed RER project. Or possibly the Berlin Ringbahn, or even the not-entirely-circumferential Yamanote Line. Sydney might be able to afford such a line, running service branded as CityRail at high frequency.
    Sydney could also adopt Tokyo’s penchant for running local inner-urban rail service at high frequency in addition to longer-distance express service. The distances in question would be much shorter than in Tokyo because of Sydney’s smaller size, but the principle would be the same. Add two short-turn off-peak locals to your 15-minute main line and you have high frequency in the inner city.
    The two-person crew problem is brutal, but there may be ways around it on new lines. If I’m not mistaken, the circumferential RER is supposed to be driverless, even though the RER has two-person crews. (Yes, they’re brutal in New York, too – it’s just mitigated by high fares and high demand, which bring the farebox operating ratio close to 1, and a local government rich enough to cover the difference.)

  6. dejv April 7, 2010 at 1:26 pm #

    … though it may bring Melbourne’s first implementation of Automatic Train Control

    Aren’t you confusing Automatic Train Control with Automatic train operation? ATC just oversees the train if it is running within permitted limits while ATO actually drives the train. AFAIK, it’s pretty common even on metros with drivers to make minimum headways possible.

  7. Jason April 7, 2010 at 5:10 pm #

    Mike,
    The argument against 2 person crews is the way they make running high frequency trains at less-popular hours expensive. Infrastructure costs are sunk, and personnel are more expensive than fuel. Driverless trains allow higher frequencies.

  8. anonymouse II April 7, 2010 at 10:26 pm #

    Jarrett,
    In this post and the Herald report you’ve raised some important points about the potential to migrate to all day frequent services on our rail networks. The poor level of off-peak level of service to higher density, inner city communities reflect a bias to the white collar work commute from the suburbs. Half hourly service on a weeknight or weekend works against creation of sustainable, car free lifestyles and transport equity in the communities that already exhibit low car ownership. A prime example of this poor service in Sydney is the restaurant and bar strip at Newtown, which has a half hour rail service after 5:30pm on a Saturday.
    However, I think you belittle the recent push for a metro system in Sydney, likening it to bringing home a holiday souvenir from London or Tokyo. The outer section of the North West corridor has been identified for rail service since at least 1998 in Action for Transport and the inner section of the proposed metro route is one of Sydney’s busiest bus corridors. Creation of a metro system was not the project objective – at the time it was seen as the best way to provide rapid transport service to this corridor.
    I think a more relevant question for Australia’s cities is whether we should extend existing rail systems that suffer from historical legacies, or develop new corridors as metro lines. These new lines would not only improve travel times and create new transit-oriented corridors, but could also improve overall system reliability and reduce congestion on existing bus and rail lines. This question was brought into focus when the Herald Inquiry proposed commuter rail as the preferred transport mode for Sydney’s high density Eastern Suburbs.
    As to how to promote higher levels of frequency on our existing rail lines, I would argue that a metro line in Sydney would have raised the bar for public transport across Australia. Because of public transport’s public funding and near monopoly position, innovation in operations and customer service is rare. An 18 hour, 5-10 minute service frequency on one metro line would certainly have me asking for better service on existing rail lines. In the interim, I agree that we should keep pushing for all day transport service on our existing network that is an attractive incentive to leave the car at home, or even better, sell it.

  9. Jarrett at HumanTransit.org April 7, 2010 at 11:11 pm #

    @ anonymouse II: I think we agree.
    Not sure what you mean about the Herald report’s recommendation on Eastern Suburbs. That recommendaion was for a South East rail line, but we concluded it would be most efficient if it could eventually flow through to a new Harbour crossing, which would have required it to be able to share tracks with Cityrail. I’m personally agnostic about it, but I do think a high-frequency south east line is needed, and I have been brought around to support the Herald’s proposed alignment (via Central rather than via Taylor Sq). It’s also a line of citywide interest because of the massive event venues it would serve at Moore Pk.
    If future study or changed circumstances led us to conclude that we don’t need the South East line to share track with Cityrail, I’d have no problem proposing driverless metro there. Note too that we did propose driverless metro for the West Metro corridor, just not as a top priority.

  10. bzcat April 8, 2010 at 3:34 pm #

    Speaking as a visitor (albeit a frequent one ) to Australia’s big cities, I think it’s definitely a bit of “metro envy” for Melbourne to want a new system. I found the streetcars to be pretty useful within the CBD. Can’t speak for its utility outside the CBD but commuter rail is probably better for suburban users anyway.
    Sydney is kind of interesting case… The main problem I had with CityRail when I tried to use it is that the fare structure is definitely not “metro” like… For example, a trip from CBD to the airport cost about $15… for a 10 minute ride! Maybe some of this is the high cost of 2-man crew but I think this is where the arguments for a proper “metro” system vs. high frequency urban rail service needs to start… JR Yamanote line does not charge commuter fare if you want to ride it for 10 minutes from Shinjuku to Shibuya.

  11. Alon Levy April 8, 2010 at 8:38 pm #

    Do other 10-minute trips cost $15, too? I’m asking because it’s routine for transit agencies to gouge airport riders.

  12. Jarrett at HumanTransit.org April 8, 2010 at 9:35 pm #

    Sydney Airport Line fares include a surcharge that supposedly repays the private partner who was involved in the line's construction.  (The operations are regular Cityrail operations, but the construction funding was privatized).  These reflect a brief period of radical privatization that probably will not recur.  Normal fare for going the same distance would be around AUD 5.00 (which is not much less than US$1 these days)

  13. Joseph E April 8, 2010 at 9:53 pm #

    In Sydney, regional rail service every 30 minutes in the evening is considered low quality? Those of us in Los Angeles can only dream of such high frequencies throughout the day.

  14. Daniel April 11, 2010 at 8:59 pm #

    dejv: “Aren’t you confusing Automatic Train Control with Automatic train operation?”
    Unless I’m woefully misreading the terms, then no, I meant ATC. I’m not aware of any plan in Melbourne to introduce driverless trains.
    Joseph E: “In Sydney, regional rail service every 30 minutes in the evening is considered low quality?”
    Put it this way: if you just miss a train, how far can you travel by car in 29 minutes? In many cases, you might be able to make the entire trip in that time.
    30 minute frequencies may be okay for trips of 50+km, especially if the travel time is quick, but for short to medium distance trips, isn’t competitive. It’s not going to get people out of cars.
    And if that’s at night, it just makes it worse if it means waiting in the dark.

  15. Tony April 12, 2010 at 6:57 am #

    And to add to Daniel’s observations, Melbourne of course also has a surface level ‘metro’ in the form of its tram network. Albeit one that could use a few more circumferential connections – though they could be adequately provided at low cost with tram-like bus services instead of trams.
    Unfortunately, even though the trams benefit from segregated ROW in the inner city area, they are forced to battle through traffic signals that operate on clockwork cycles designed to favour cars. (For whatever reason, Australian cities have some of the longest traffic light cycle times in the world.) Melbourne could really experience a revolution in inner-city transport if trams got a traffic priority scheme like the one Zurich has.

  16. Geoff March 6, 2012 at 12:16 am #

    Here in melbourne, people are begining to use public transport especially to the cities and the dependency on public transport is highly increasing. But the most annoying thing is the waiting time.