CBD rail capacity myths: Loop tunnel usage, Stations served, the European solution

In this blog post I hope to address a few myths around Melbourne’s rail system that I’m seeing floating around.

Train loading at Flagstaff, 5:50pm

The Loop tunnels have hardly any trains!

I’ve heard from a couple of sources in the past week (one on mainstream radio) the claim that nothing needs to be done about rail capacity in the CBD, because trains only run in the tunnels every 10 minutes or so.

It might be true in off-peak hours, but is certainly not true in peak, when most tunnels have a train every 3 minutes or so.

Looking at evening peak, the hour 5:00-5:59pm, Loop trains departing Flinders Street:

Clifton Hill tunnel 5:03 5:07 5:10 5:15 5:20 5:23 5:27 5:31 5:36 5:41 5:46 5:50 5:53 5:59
Caulfield tunnel 5:00 5:06 5:09 5:12 5:15 5:18 5:22 5:24 5:27 5:30 5:35 5:38 5:41 5:44 5:47 5:50 5:53 5:56
Burnley tunnel 5:01 5:03 5:07 5:10 5:13 5:15 5:17 5:20 5:23 5:26 5:31 5:33 5:36 5:39 5:43 5:48 5:51 5:56
Northern tunnel 5:02 5:04 5:07 5:10 5:13 5:19 5:22 5:24 5:27 5:30 5:33 5:36 5:39 5:42 5:44 5:47 5:50 5:53 5:59

(Trains departing Flinders Street running direct have been excluded, of course.)

The single biggest gap is 6 minutes, and the Clifton Hill tunnel has a few 5 minute gaps (See: PTUA on capacity for Doncaster trains), but for most of the hour, gaps of about 3 minutes are the norm.

If more trains are to run — and they need to, because some lines are very crowded during peak — something has to be done.

Potential upgrades include:

  • Measures to speed up dwell (loading) times at stations — such as trains with more doors, indicators to show which carriages of an approaching train are less full, wheelchair “humps”, or where they aren’t possible, platform staff to help with wheelchairs
  • Higher-capacity trains — including more efficient seating layouts to fit more people aboard
  • Running more trains direct to/from Flinders Street, not via the Loop — already the case for Werribee and Sandringham trains, and some Craigieburns and Frankstons and others in peak. For minimal conflicts at junctions, and maximum legibility of the system, all services from particular lines would run direct (see below)
  • Signal upgrades — planned for the Dandenong line; remembering that the highest capacity signalling involves retrofitting the trains as well, so it can be a tad expensive
  • More tracks — this is what the government’s Melbourne Rail Link and the older Metro Rail Tunnel plans offer, in conjunction with much of the above, to separate out Melbourne’s rail network into 6 independent groups of lines

Other measures include boosting off-peak and shoulder-peak services to encourage more people to travel outside peak hours if they can, and even pricing changes such as off-peak fares (or schemes such as Early Bird — rumoured to be being phased-out from 2015) to encourage this.

Another crowded train

But train X won’t serve station Y!

This isn’t a myth — it’s already a reality, though the ALP has fallen into the trap of claiming Frankston trains won’t serve Richmond (and the sporting precinct) under the Coalition plan. That’s not quite right — Frankston trains will stop at Richmond, but only after running via the CBD.

It’s true: under both rail tunnel plans, some lines will serve fewer CBD stations than they do at present.

Under both plans, the Sandringham and Glen Waverley lines won’t serve the City Loop.

Under the Melbourne Rail Link plan (backed by the Coalition), Frankston and the Camberwell lines won’t serve Flinders Street, but will stop at the other CBD stations (as well as Richmond). Dandenong and Sunbury won’t serve the underground stations, but will stop at Flinders Street and Southern Cross. (The Coalition tends to play this down in their rhetoric.)

Under the Metro Rail Tunnel plan (backed by Labor), Dandenong and Sunbury trains won’t serve Southern Cross, Flagstaff or Parliament, but will stop at (well, under) Flinders Street and Melbourne Central. (The ALP’s web site doesn’t seem to mention this when criticising the Coalition’s proposal.)

These are the compromises you end up having to make as the rail system gets busier. Not every train can serve every station, particularly the underground Loop stations, which only have four tracks.

This process started in the 90s when Sandringham trains came out of the Loop on weekdays, and has continued since then, with Werribee and most Frankston trains, as well as Glen Waverley on weekday mornings.

Rather than have a mix of trains on each line running direct to Flinders Street and via the Loop, it’s better to have some consistency, and run some lines direct and some via the Loop, for several reasons:

  • It avoids problems with running inconsistent frequencies. If trains alternate between via the Loop and direct, you get very uneven gaps in the timetable, because the running times are so different. It also means many people wait longer than necessary for a train.
  • Consistency is less confusing — witness the daily Frankston timetable confusion between 4-5pm and 6-7pm when stopping trains run half direct, half via the Loop.
  • It means less conflicts at junctions, so fewer delays as trains wait for one another. This improves punctuality, and capacity of the network, allowing more trains to run… which is the point, remember?

To avoid big problems, connecting services need to run frequently, and interchange needs to be as simple and quick as possible, so people can still quickly get to their destination, even if it involves changing onto another train (or for that matter onto a tram).

#Myki gates at Flagstaff still not working

Why not the European solution? Terminate the trains at the CBD edge, and get people to change to a shuttle service?

In many big European cities, the suburban trains terminate at the edge of the city centre, and people have to change to a “metro” connecting train to complete their journey.

This makes sense in old cities, where in the mid-1800s, when the trains to the suburbs (and farther afield) were first built, and they couldn’t knock down vast areas of the central city to accommodate them, and they hadn’t figured out how to put them underground yet.

When they did start building underground railways, initially they were limited in tunnel size, so generally smaller trains were used. Hence the London “tube”, where the trains are quite cramped, and the tunnels only barely bigger than the carriages. So it’s common for people to come into the cities on larger suburban trains, and change to frequent metro services to get around the city centre.

London’s cramped stations and Underground trains — photo by Phil Ostroff on Flickr

But in Melbourne, and other Australian cities, the railways came as the cities were established, so our large central railway stations such as Flinders Street are already pretty central.

You really don’t want to have thousands upon thousands of people changing trains unless you have to.

It would be creating lots of problems, and solving none, to stop the suburban trains at Richmond, North Melbourne and Jolimont and make people change onto a short-distance CBD-only service. Providing adequate interchange and terminating facilities would mean you’d need huge expansion of those stations. And it would be a complete waste of most of the rail capacity and platforms at the existing CBD stations.

A variation might be running all suburban trains to Flinders Street, and having dedicated City Loop (circle) services. But again, you’d be needlessly making a lot of people change trains who don’t currently have to. And remember one of the reasons for building the Loop in the first place was to reduce pressure on Flinders Street with regard to passenger numbers. With recent growth, its subways and other pedestrian routes are under strain.

With modern engineering, newer European city railway tunnels have brought those larger suburban and longer distance trains directly into the central city: Paris’s RER is a good example of this, as is London’s Crossrail project now underway.

There are a lot of good things to admire and copy steal adapt from European railway systems, but making people change to reach the CBD is not one of them.

Update 17/6/2014: The anonymous Coalition blogger SpringStSource has quoted extensively from parts of this post in an article posted today. It’s worth a read, but I’m wary of the rhetoric from both sides on these issues.

By Daniel Bowen

Transport blogger / campaigner and spokesperson for the Public Transport Users Association / professional geek.
Bunurong land, Melbourne, Australia.
Opinions on this blog are all mine.

24 replies on “CBD rail capacity myths: Loop tunnel usage, Stations served, the European solution”

Sydney provides a closer to home example of productive line interchange options further out in the network, as shown in:

Both Melbourne and Sydney are now on track to rival London and New York, at least as they were in the late 1980s when I got to know them a bit, having driven in from other cities on my first visits but quickly learning that their underground rail was the key to getting around.

It’s not politically sensible to get too angry about state Labor’s knee jerk policies on the run at this point where the alternative is so far out of bounds, but I have long been open about my fear that a one term Liberal government would not teach Labor why it lost in 2010.

Despite the bloke at Sunbury on Saturday insisting that stopping freeway construction is an assault on the rights of (geographical) fringe dwellers, the only way Melbourne’s next doubling of population can be catered is by getting commuters and private visitors to inner areas out of their cars which will likely require at least a quadrupling of inner urban capacity. And that will require several new alignments through the inner core (CBD++) as well as a network which encourages alignment choice to be made further out, as the various earlier Metro routes did with choice points at Footscray and Caulfield, though losing choices at North Melbourne and Richmond currently enjoyed by Newport direct/through services.

While I’d rather see Mernda and one or two others nudged ahead of the start of the first inner capacity upgrade, that seems to no longer be a political option. However, that should not stop us starting to insist on much longer term planning for multiple inner alignments so that other prioritisation decisions can be made on a more informed basis.

thanks for busting myths – a great article today.
It would also help if people were told not to wear backpacks while standing in trains as this takes up the room of an extra person.

All very sensible. However there are a couple of points to make about extra capacity.

There is the option of bringing forward loop through-routing (connecting either the Burnley or Caulfield loop to the Northern loop so they run as a Flinders St and Southern Cross by-pass rather than a loop) to before any new tunnels are built (ruled out by the Coalitions plan).

Rather than extra platform staff for wheelchairs, platforms that allow them to board without ramps (like at Box Hill one island platform at Flinders St and parts of Perth) would allow even faster boarding and cost less.

I agree with most of the article, except for the claim that train carriages with more doors are better.
You’d think trains with more doors makes sense, but human behaviour suggests otherwise. The reality is that in even a sparsely occupied crowded carriage, every door has at least one or two passengers standing beside a door. They don’t get off the train; they just reduce passenger throughput for the door.
In older carriages with three narrow doors, this restricts flow to one person in or out a time (three people at a time).
Newer carriages on some lines, such as Dandenong/Pakenham, have two doors, but they’re significantly wider. Even when congested by selfish sods who loiter in the doorway and refuse to move down the carriage, these wider doors allow two people or and out at once (four people at a time). That’s an improvement on passenger throughput of 33%, which translates to potentially cutting dwell (loading) times by a third.
Reducing the number of doors but making them wider may not be intuitively obvious, but it’s actually a very smart design because it responds to (poor) behaviour of passengers. It also simplifies seating layout.

Does anyone know why they planned a ‘trial’ of High Capacity Signalling for the Sandringham line, then before its even started announced that they were bringing it in for the Dandenong line? Isn’t the idea of a trial to see how feasible an idea is, how workable, costs, problems etc? How can they condem one line to a certain system when they havent even trialed it on the line they said they would? If they haven’t done a trial in Melboune conditions, how do they know it will actually be of any benefit?

Kind of agree with you Damien. Trains are regularly delayed at platforms because of passengers ignorant behaviour. Akin to sheep, they all follow each other and they all want to get off at the door that is closest to the exit of their station. Along with the 90 other people exiting at the same time. Extra doors, extra capacity – will not do anything to help this. This problem can only be overcome by passenger behaviour and maybe a campaign to show people how long they are actually delaying trains for by being so selective on where they exit. They may be amazed to find that the 3rd door on the 4th carriage actually gets to the platform at the same time as the 2nd door in the 1st carriage..

“Legibility of the system” is a fairly bogus argument.

It may confuse some tourists , or infrequent users, or people straying from their usual patterns, that ( for example ) some Frankston trains “run direct” and some “run via the loop”.

The majority of users are not “confused” by this, and they know what they are doing, and would see an extra 20 minutes of their day, every day, wasted by “simplification”.

“Sydney is different because it has “only” four lines radiating out of it’s basin”.

How is four lines, so different from five lines ?

Crossrail is great. And it includes a station to the airport. What would Crossrail in Melbourne look like?

@enno – Legibility of the system is not a bogus argument. Sure, captive riders that rely on the system every day, and catch the same service the vast majority of the time will have no problem, but as you’ve acknowledged it is confusing to virtually everyone else. If you’re trying to attract choice riders, legibility is key. If an infrequent user decides they want to try riding the train to work and then find the system confusing and frustrating do you think they’ll be trying again the next day? Even amongst regular users patterns changing can cause problems: I’ve more than once boarded an Eltham or Hurstbridge train and ended up having to back track to Fairfield because I’ve not being paying attention to whether the service is express or S.A.S.; every single time I’ve done it I’ve also seen at least one other person cross the bridge and also wait for a train coming back the same way.

20 extra minutes? Hardly. Changing trains and Richmond or South Yarra might add an extra 3 or 4 minutes during peak, maybe up to 10 outside.

I’d also argue it is a lack of legibility that leaves so much of our bus network under used.

The worst example of this type ‘terminating the trains at the CBD edge’ in Australia at present. Is the proposal in Newcastle. Australia’s 7th largest city has plans to sever the heavy rail line and install a light rail. People in Australia should be annoyed at governments who have no foresight and are reducing services in our major cities and not planning for the future. I would recommend people try to stop this occurring in Newcastle.

In response to Roger’s comments about backpacks – this is really annoying, and is also a particular hazard for short people who get knocked on the head when the backpack wearer turns around (as they invariably do). My solution is to tap the wearer on the shoulder and politely ask them to place it at their feet where there is more room. They usually obey!

Regarding Damien’s comment about people selfishly loitering in doorways, I’d just like to point out that in a very packed train it is easy to get trapped down the aisle and not be able to get out at your station, particularly the case if you are getting out before the end destination. The solution is to stay in the doorway area. People who are going to the end do need to move down though.

I’d also like to mention the annoying habit of people sitting on the aisle seat and leaving window seats empty, thereby forcing people who want to sit (and free up standing space) to crawl over legs and bags. Please move along.

The fixed wheelchair ramps at stations are an excellent addition as, not only do they speed up train loading but they also prevent the driver from forgetting to get the ramp out at the right station, something that happens fairly often and they allow wheelchair users to be able to change their mind about where they want to get off, something that most people take for granted but is very difficult without fixed ramps. However there are some issues that mean staff assistance is still required at times. One issue is that some wheelchair users are unable to open the train doors themselves, especially on the older Comeng and Siemens trains where the door needs to be pulled open. Also at some stations the gap is still too big for some chairs. This typically happens on curved platforms where the ramp can not line up exactly with the train for example Mooroolbark station city end. This has the danger that wheelchairs can (and have) get stuck between the train and the platform as the gap looks small enough but is just slightly too big.

Terminating trains at the city edge could also cause issues for wheelchair users. I know of many who avoid Richmond station as not only are the station ramps too steep but the gap between the train and the platform is particularly large making the ramps to get off the train steeper than normal. I myself often need to go from Burnley to Caulfield groups and try to allow extra time so I can change at Flinders Street rather than Richmond. This however adds 20 to 30 minutes to my trip. I note that there is an article in today’s Age about upgrading Richmond station. I will believe it when I see it.

One of the problems, which you’ve probably blogged about somewhere already, is the generally unspoken assumption that Melbourne’s growth will remain city-centric. Paul Mees had a good point in “A Very Public Solution” that traffic congestion is its own limiting factor. That sort of applies to PT too. So instead of catering for CBD businesses who want their staff/customer catchment to extend from Pakenham to Werribee, and Frankston to Craigieburn; why not prioritise services to link middle and outer suburban centres, concentrically rather than radially? In time, the jobs and trips should move to adapt to the new opportunity.
I’m playing devil’s advocate – clearly it would only work as a long term solution to the current problems, and getting cars off roads/more space on trains is important for more reasons than just improving people’s hellish commuting experience (eg air pollution). Really, the solution is probably that we need both. But I think at some point, a responsible plan for Melbourne’s transport is going to have to question the reactive, city-centric development model that the city has pursued by default. There has always been some rhetoric about suburban activity centres, eg in Melbourne 2030, but it’s always played second fiddle to roads and the CBD.


I think you have a point there, intra suburban transports can be hard, but if Melbourne follows patterns of other world cities it will become increasingly important as work and leisure patterns change. The Overground in London is a good example of a moribund service that doesn’t even touching the CBD being rehabilitated to serve these new patterns.

In my personal experience getting between suburbs on the South Morang and Caufiled lines south of Bell streets is very hard on transit, the infrequency of buses can make going via the loop more attractive depending on where you are. I’m sure there are other anomalies but giving people options to avoid the loop may help relive pressure, and encourage patterns that maximise the city and its infrastructure.

It is a shame that the outer circle is gone, and just how busy it is as a bike path shows the value of east-west transport links in the inner north. There isn’t decent east west Bus Service until you get above Bell street and meet the 903. Normamby/Moreland and Separation/Arthurton/Blythe really need good frequent service. and While I’m fantasising about join dup transport thinking, how about a sky bus that stopped at North Melbourne, Clifton Hill and Richmond.

It’s interesting that the Overground didn’t really come to fruition until there was a unitary authority responsible for transport policy across the London region; the Mayor, GLA and TfL. Melbourne has no such holistic body, state and federal government are and should be concerned with wider issues.


I think what you say about through running is very pertinent. I think there is a good argument for joining services on opposite sides of the loop and only running via Flinders or via the Loop. Naturally it will disadvantage people but given the groupings its possible to split services and exchange passengers e.g. South Morang via the Loop to Upfield and Eltham/Hurstbridge via Flinders st to Cragieburn. (of course this is the worst example to choose because of the flat junction west of Jolimont, but you get the idea)

question – would more aggressive stopping patterns allow for more trains to pass through the system? When I went to Perth I was surprised to discover that they seem to embrace the idea that alternate trains on the same line stop at (usually) every second station … so you can have more running trains at once on the same line (as I understand it) … passengers will need then to make sure they’re on the correct train for their station, but wouldn’t that be useful to push more people through the system overall?

Al wrote:
“Trains are regularly delayed at platforms because of passengers ignorant behaviour. Akin to sheep, they all follow each other and they all want to get off at the door that is closest to the exit of their station. Along with the 90 other people exiting at the same time. Extra doors, extra capacity – will not do anything to help this. This problem can only be overcome by passenger behaviour and maybe a campaign to show people how long they are actually delaying trains for by being so selective on where they exit.”

@Al, while I agree that it’s a problem that passengers tend to cluster around the train doors that are closest to the station exit, I think you’re being too harsh on them. It’s not necessarily ignorant, sheep-like behaviour, and an education campaign won’t necessarily change their behaviour.

Many passengers have a very good reason for needing to exit the station as quickly as possible. Often it’s because they’re trying to connect with a bus or tram, which, as you probably know if you’ve ever tried to do the same, isn’t going to wait around until all the train passengers have left the station at a leisurely pace and hopped on board. It’s very frustrating to alight from a train and sprint out of the station just in time to see the bus close its doors and pull away from the kerb. Even more so if you know the next bus won’t be for another half hour or even more. If you experience this often enough on your daily commute, you too will be trying every strategy you can think of to shave off a few seconds between train and bus – one of which is to stand next to the train door which is closest to the station exit.

This problem is made worse if there’s a road between the station and the bus stop. One of the stations I use is next to a busy arterial road. The train line crosses that road, with boom gates. The bus stop is on the other side of the road, and there’s no pedestrian crossing. Train travellers wanting to catch that bus need to wait for a break in traffic before they can cross the road. Traffic is often so heavy that they wait for a long time, often long enough to see the bus pull up, stop, and take off again, while they remain frustratingly stranded on the station side of the road. What I try to do when I use this station is to hurry out the exit, along the path adjacent to the railway line, and across the road while the boom gates are still down and traffic is stopped. If you leave the train by the doors closest to the station exit, there’s just enough time. Any further away, and you have no hope.

All the education campaigns in the world are not going to change this behaviour. What is needed is a change to infrastructure that makes it unnecessary in most cases to stand close to a particular door.

I would suggest, as a start, rethinking the idea that passengers need to be funnelled out of a single exit at each platform, or two at the most. To me this seems a relic of the days when the stationmaster stood at the exit collecting tickets. You needed the single exit to make sure people weren’t trying to slip past the stationmaster and evade the fare. Now that we have Myki readers, and can in theory install them right along a platform’s length (whether at an exit or not), it’s not necessary any more. (Or if there is a reason, could someone please enlighten me, because I can’t think of one.) So, wherever there is just a fence dividing the platform from the street, and both platform and street are at a similar level, I suggest getting rid of the fence and installing ramps and/or steps if need be, so that the passengers can exit anywhere they please.

Obviously there are some stations where increasing the number of exits would be difficult and expensive. Essendon springs to mind, because of its island platform. But even there, it’s not impossible. There are three subways under Essendon station, but the platforms can only be accessed by the central one. Build ramps/steps joining the platforms to the north subway, install Myki readers there, and all the passengers who are going to the car park, the tram, or heading to Mt Alexander Rd or beyond, would use that exit. Build more ramps/steps joining the platforms to the south subway, and all the passengers who are heading to Buckley St and beyond would use that one. People wanting to catch a bus, and Rose St shoppers, would continue to use the central subway. That would take a huge load off the central subway, and train passengers might have less need to cluster around the central doors of the train.

Certainly at this station, increasing access would be expensive. But at other stations it would involve little more than taking down a cyclone wire fence and installing some more Myki readers. Give passengers enough exit points from each station, and you may find the problem of clustering around particular doors diminishes to the point where it no longer affects train dwell times.

@Damien 26 May 2.47 on train doors:

Doors should be clearly distinguished from ‘door channels’. A channel should be at least 700mm wide *and* not obstructured by people leaning against the windbreak partitions.

The Siemens and Extrapolis doors are wide enough for two channels in theory but still get blocked by people leaning against the partitions, which are not set back far enough. Their flow rate could be improved simply (thus reducing dwell time), by sacrificing a few seats to move the partitions.

RE: moving to simpler operating patterns that may require more transfers to reach some inner city stations:

The proposals that I described here –

– have significant advantages for transfers in the inner arear, viz:

1. Run the ‘cross city line’ from Werribee to Dandenong instead of Frankston (Frankston uses the Caulfield loop), with up-up-down-down running to Caulfield and a flyover at Caulfield (this is necessary to avoid conflicting movements):

This gives same direction cross platform interchange between Flinders St and Melbourne Central trains at Richmond 3-4 and 5-6. It also expedites running longer trains from Werribee to Dandenong, which is probably the most cost effective way to increase capacity where it is most needed in the medium term.

2. Join the Northern and Caulfield loops to run from Craigieburn to Frankston via Melbourne Central (proposed in stage 4 of the PTV 2012 Network Development Plan, though it would not be compatible with the current Melbourne Rail Link via Fishermans Bend). Then Sunbury runs through to Glen Waverley.

This would desirably include a flyover west of Nth Melbourne and a new platform 7 at Nth Melbourne. The platforms at Nth Melbourne become: 1: disused (or maybe Vline), 2-3: Craigieburn – Frankston via Melbourne Central and Sunbury – Glen Waverley via Flinders St; 4-5 ditto the other way; 6-7 Werribee/ Dandenong

This gives same direction cross platform interchange between trains from Craigieburn and Sunbury to Flinders St and Melbourne Central.

“why not prioritise services to link middle and outer suburban centres, concentrically rather than radially?”

The reason this is so hard, is density and accessibility at the destination ( i.e. workplace ) end.

It is very difficult to create enough workplace density at suburban locations, to put enough workplaces within practical walking distance of the station, or to provide “feeder buses” at an adequate service level to provide an overall transport experience competitive with driving there in your own car.

To get from your home to the station, you have the options to
(a) walk there
(b) catch a bus there
(c) “kiss and ride”
(d) “park and ride”
(e) cycle.

At the workplace end of your train journey, (c) (d) and (e) are not generally available ( unless you can take your bike on the train ).

@enno — yes. it’s not ideal. However, PT planning is currently reactive, in the sense that it assumes this is the case and thereby perpetuates that situation, by only seriously catering to 9-5 city centre commuters.
The method I’d suggest would be to look at suburban destinations like schools, universities, shopping centres and the few office precincts that exist (so not only workplaces), and work out how best to ensure that people can get to them by PT and walk/cycle instead of by driving.
This would be most easily achieved, perhaps, by buses for a lot of current locations. I don’t think it’s that hard. The orbital “smart bus” service seems to have been taken up fairly well.

Comments are closed.