How do world transit systems compare?

This is an eye-opener. First thing you need to do is go here where artist and urban planner Neil Freeman has prepared maps of many more cities at the same scale  (none from Australia unfortunately). What these maps show is the enormous variation in the scale and density of different systems. No doubt the artist has had to make assumptions about what to include and exclude*, but this is nevertheless very important information to bear in mind when comparing the performance of Australian rail networks with systems elsewhere.

I think Melbourne’s rail system is a leviathan compared to some of these e.g. the networks in Toronto, Vancouver, Amsterdam, Budapest and Brussels. Note the scale (5 km) beside the map of Toronto above. As a point of reference, it’s about 5 km as the crow flies from Flinders St station to Merri station on the Epping line, Hawthorne station on the Lilydale line and Newmarket station on the Broadmeadows line. Five kilometres is the boundary of the inner city.

Some cities however have much denser networks than Melbourne. The Paris system, for example, is diminutive in extent, but extraordinarily dense. I’m not well-travelled enough to know how many of these cities have extensive tram networks, but if you factor in Melbourne’s trams we also have a pretty dense network (albeit a slow one) within 10-12 km of the CBD.

There are important differences in the way different configurations operate. For example, a system like Melbourne’s with long average trip distances will generally require that more space is devoted to seating — thus reducing capacity — than is the case with systems where trips lengths are relatively short. So it’s not always as simple as “if they can do it so should we”.

I’m using broad terms like ‘extent’, ‘scale’ and ‘density’ to describe these systems, but if you were serious about comparing the various morphologies you would need to come up with a range of descriptors to describe the differences accurately.

If anyone can incorporate Melbourne’s train and tram networks into the diagram at the same scale I’d be happy to add it to this post.

* See points made by commenters below. Reinforces the point that making any sort of comparison is a fraught task and there will be compromises.


12 Comments on “How do world transit systems compare?”

  1. brisurban says:

    It depends. What would they look like if the buses were mapped? Toronto’s system might actually be “functionally” larger because it uses buses to the train stations. What’s needed is a map of service, not necessarily infrastructure.

    • Alan Davies says:

      Fair point. The question should’ve been “how do rail systems compare?”. But it doesn’t pretend to show frequency. Even the best transit maps showing all high frequency modes tend to look like a bowl of spaghetti – such a map of 30+ major cities would be a challenge to read (and arduous to put together!).

  2. Russ says:

    Worth noting Paris also has the RER system which extends deep into the suburbs. Actually Paris is interesting for that reason. One of the weaknesses of the Melbourne system is that it resembles the RER on the periphery, the metro in the centre, and serves neither market well – too slow to be effective as a regional distributor, too sparse a network and too low a frequency to be a metro, and in several places competing directly with a near identical tram service. We could do worse than extending the tram system to handle the metro connections and closing 150 or so stations.

  3. Jim Wells says:


    Someone rightly pointed out that Paris has the RER. It also has, as London has, suburban railways operating from stub end terminii. London has some heavy rail crossing the City and is going to get more.

    This really does distort comparisons like this.

    Trams should be associated with buses not railways unless they operate off street.


  4. rohan says:

    Yes they’re not exact comparisons of services – some have overground and regional as well as underground (though Berlin map seems to include both) and then there is light rail / trams, bits of system not used much etc. But siply in scale and reach I would expect our system to look some where between shanghai (without the complicated middle bits) and almost the reach of San Francisco.

  5. T says:

    It’s a very interesting comparison. I could add a few points about toronto’s system having lived there for many years.

    1. The map only shows the subway but fails to include the suburban rail line called GO train which runs express from downtown to the outer areas (which extend beyond the regions of the map) thereby allowing those traveling longer distances to avoid making all the stops. This makes the trip quicker and divides up the congestion. Toronto’s trains are never as congested as melbourne’s dispute higher patronage.

    2. The map doesn’t show the tram and bus network which is very well organized. Essentially if you took a chess board grid and laid it over top of the subway map those would be your buses or trams. They are used by patrons to get to their nearest subway station. All buses and trams begin and end their journey at a subway station.

    I can see though that it must be easier to manage with a smaller train system. But thats not the only problem. Melbourne has not good tram bus train linking system. You can’t just jump on a bus to the nearest train station and then catch a train and even where you can you end up waiting 20 min for the bus and then 20 for the train. Way too much of Melbourne’s system involves waiting. To put it in perspective Toronto’ system doesn’t even have a schedule. People there have no concept of having to catch a certain bus or train. You just leave whenever you want and you know you won’t ever be waiting much. Maybe Melbourne should stop building new train lines and just focus on getting the current system working properly. Melbourne can do it I’m sure

  6. jack horner says:

    Russ: ‘We could do worse than extending the tram system to handle the metro connections and closing 150 or so stations’

    I used to think this, but don’t any more.

    ‘Close 150 stations’ implies priority to speeding the trip of the long distance commuter to the city.

    Keep them open, with focus on frequency and connectivity rather than speed, gives priority to improving anywhere-to-anywhere public transport with transfers (subject to integrating bus services with them appropriately). Particularly transfer trips from the rail served outer suburbs to the inner area other than the CBD.

    Where you have a congestion-free right of way, why not use it for as many purposes as it can handle? Think of a short, frequent, high powered train, stopping all stations to Box Hill or Moorabbin or Sandringham, as the next cousin to light rail (I’d be interested in any comment on how their marginal per distance running costs differ). Add an express service to Ringwood or Frankston and think of that as more akin to the Paris RER.

    Run the short lines like Perth’s Fremantle and Midland lines. These ‘heavy rail’ lines probably carry fewer people than a number of Melbourne’s individual tram lines. But if their operating costs are not much different (?), and the infrastructure cost is sunk, is this a problem?

    • Russ says:

      Jack, there is always a trade-off between speed (and therefore total carrying capacity) and access in a transport system. If you run the numbers in a scenario where all patrons have to walk parallel to the line to reach the stop, the optimal stop-spacing is a function of total travel distance. For short trips (less than a mile) it is around 400m. For long 10-20km trips it is around 1-2km. And using either spacing for sub-optimal trips can be extremely slow. But if there is a bus/tram/train running along the same route with 400m spacing and 10min frequency then the optimal spacing for long trips is 3-4km.

      That is effectively what an express route is. But running many purposes on one line makes it harder to do those purposes well. Express trains are very limited for several reasons: 1) the total time saved is limited by trains in front and behind, so a line running more than 12 trains an hour isn’t gaining anything; 2) it makes wait times uneven (a problem on congested lines) as the express leaves a few minutes after the non-express followed by a long gap; 3) most of the gains from running an express are lost on increased wait-times for skipped stations; 4) it can reduce reliability as problems cavalcade (this is particularly true for triple lines).

      For a section of a line like Box-Hill->Ringwood, modifying the 109 to start/stop at Box Hill, Nunawading and Ringwood stations (short route lines are much more reliable and easier to coordinate), such that it stopped 2-3 times more frequently than the train, and getting the train to run to only those three stations would generally be faster for both short trips over that section and long trips originating in that section (not to mention a lot faster for trips that traverse the section).

      In two dimensions, your ideal choice is between a fairly uniform grid (like Toronto) or a nodal network with backbones. Grids work very well, but they also presuppose the city has a uniform density. Since that isn’t the case – and in theory we are planning to make it less true, by building on the existing spoke and hub network – a nodal network would be a better choice.

      • brisurban says:

        Skip-stop is an option that does not decrease line capacity but increases speed overall.

      • jack horner says:

        I’m more sympathetic to the idea of a Box Hill Ringwood tram extension becoming a rail feeder at intervals along the corridor. I wouldn’t be sympathetic to asking a 10-15km tram line to become the main form of commuting to the city for people from closed rail stations.

        But there would surely be significant difficulties in getting the tram from the parallel main road to the chosen interchange stations (and similarly in other possible examples such as Upfield and Epping lines).

        Whether you can fit a metro style train service (stops 1.5km) in the same corridor of course depends on the case. In Melbourne mostly you could given the triple tracks to Box Hill and Moorabbin. Not ideal, since you can’t have a balanced contrapeak service, but possible.

        You could fit 16 per hour to Dandenong with two skip stop patterns in which alternate trains skip different sets of about 3 stops. Similar in principle to the Perth Midland Fremantle timetable.

        I’d prpobably be sympathetic to closing some stations on the Upfield and Epping lines.

  7. jack horner says:

    correction to above: “Whether you can fit a metro style train service (stops up to 1km) and a commuter style (stops more than 1.5km) in the same corridor…’

Leave a Reply

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

You are commenting using your 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 )

Google+ photo

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

Connecting to %s