When the 2009 state budget actually committed funding to it, the price had dropped slightly to $562 million, but this was due to the initial figure including running costs, apparently for several decades. Even counting the duplication from Keon Park to Epping at the same cost as the extension, it was still five times higher than the per-kilometre cost of the Craigieburn project completed just two years ago.

A feature article in The Age in June highlighted the issue, with local activists delighted at the commitment but mystified over the price, the opposition claiming taxpayers are being dudded, and public transport advocates (such as myself) fearing that the high price will discourage governments from future rail extensions. The Department of Transport offered the explanation that the project was a “more holistic approach to scoping the expansion of the Epping line”, but apparently didn’t clarify this in any great detail.

We (the PTUA) subsequently met with the Department, and finally discovered the real truth behind the term “holistic”. It turns out the scope of work is much bigger than just the South Morang extension plus duplication from Keon Park.

The way it was described, it includes:

• Keon Park to Epping
  • duplication, obviously including track and overhead
  • two or three pedestrian grade-separations
  • station upgrades, and an additional platform at Thomastown
  • upgrades to 4 level crossings to latest standard
  • resignalling, including removal of bidirectional signalling on the existing single track (it conflicts with the proposed location of the second track), and re-signalling most of the rest of the line, almost down to Clifton Hill
  • stabling at Epping, with driver facilities to enable future changeovers to move away from Flinders St
  • extra substations
• Epping to South Morang
  • dual track, overhead. We were told the old alignment can’t be used without modification, as there are grade separations and other issues with it
  • signalling
  • Dalton Road grade separation. Due to proximity of Epping train maintenance facilities (which can’t be moved), relocation of local roads (which can’t be disconnected from Dalton Rd) and the nearest creek (which the track must get over), this is said to be a reasonably complicated component, on a similar scale to Springvale Road grade separation
  • 3 bridges over creeks
  • bike/ped path along rail route, and included in bridges
  • grade separation of Pindari Avenue and Civic Drive
  • a new substation
  • communications systems including radio towers
  • South Morang station, including bike, bus, car parking, and provision for further extension to Mernda
• Hurstbridge line
  • signalling upgrades on parts of line, to help harmonise frequencies to work better with the Epping/South Morang line
  • stabling at Eltham, including driver facilities, and which will require the moving of some existing trackwork

The scope of works goes some way to explaining the cost. Perhaps it doesn’t bring it down to the level of Perth’s Mandurah line, but at the very least it brings it back down to somewhere near Planet Earth.

It seems to makes some sense to include in the project scope upgrades that will help the rest of the Clifton Hill group run better. If things turn out to plan, in 2013 (just in time for the 2014 state election) the benefits should be felt not just to residents in South Morang, but also elsewhere along the Epping and Hurstbridge lines.

What is a real mystery is why the Department doesn’t publicise the true scope of the project. While the information is apparently no secret, as a number of groups have been briefed on the project breakdown, neither has it been made public.

Surely flagging the real scope of the project, with all the resultant benefits, would be better for the government than hiding the details away and having major newspapers writing feature articles highlighting the apparent cost overruns and implying incompetent project management.

The Airport is a major transport node in each city, as well as a large employment centre in it’s own right, and certainly should not be neglected as part of a public transport system. I’ll be focusing on the three main east coast cities of Sydney, Melbourne and Brisbane for this post.

1) In September 2002, 10,300 people were employed at Melbourne Airport in some 9,000 equivalent full time positions, earning an estimated $525 million in wages.
http://www.melbourneairport.com.au/downloads/pdfs/PublicVersionEIS.pdf

The latest job survey by BAC [Brisbane Airport Corporation] indicates the number of people working at the 2700-hectare Airport is expected to top 9000 within the next 6-12 months.
http://www.bne.com.au/content/mediapopup.asp?name=Media_12_04_2005&keepThis=true&TB_iframe=true&width=640&height=480

Rail Links

The first thing that comes to everybody’s mind is airport rail lines. Two have been built in Australia, the first being Sydney’s AirportLink opened in 2000 which has been the subject of much criticism and poor patronage since it opened.

I don’t have an issue with the way this line was built, it is essentially an extension of the East Hills line and connects the Airport to the existing network as well as being a generator of additional paths. It has been poorly managed from the start, with RailCorp having control of the actual line and a private consortium owning the actual intermediate stations.

For passengers using trains at the Domestic and International stations, a surcharge of around $11 on a single ticket over the regular CityRail fare applies, for which no premium service is provided. The double-deck carriages have little space for luggage and services are often overcrowded in peak hours. The only other service to the airport is Sydney Buses route 400 which runs only at 20 minute intervals, but at least has regular fares.

Next up is Brisbane’s Airtrain, opened in 2001. I have less against this line than the Sydney one as the line is entirely privately owned and funded with QR being paid to operate services. Trains ran every 15 minutes at first but were later cut back to 30 minutely, the last one being at 8pm each night. This line is now making an operating profit at least.

I’m not sure what to think of this line but certainly I don’t think the frequency is enough. The private owners are unlikely to pay QR to operate twice as many services until they eliminate the guards which double the crewing costs of operating each train. $13.50 is the price of a single from the Airport to the CBD but there are significant discounts for workers.

There are no Translink services to Brisbane Airport. This line could have been constructed as an extension of the Doomben Branch, currently just hanging on with an awful service level, but was built as a new elevated line from Eagle Junction. This could be seen as a missed opportunity to revitalise a marginal line, but otherwise the chosen route is more direct.

And finally, proposals have often been mounted for connecting Melbourne’s Tullamarine Airport to the CBD by a train service, which was seriously considered by the Bracks Government in 2001.

The original plan was doomed to fail; it set a 20 minute journey time (currently done by road) only achievable with expensive track amplification projects. It mandated special trains fitted out for the run. It would have been an expensive project with a journey time no faster than Skybus that still failed to cater to the large working population.

Map from the original plan

Skybus, the express bus service to the CBD has been upgraded since and now runs every 10 minutes with a 20 minute journey time, for a $16 premium fare that is charged by it’s private operator. I believe the express Airport-CBD is already covered well by this service.

Not all hope of a rail link is lost. Phin has done a post on an affordable airport rail link that suggests a reworking of the Upfield line to serve the airport, also picking up the Broadmeadows area. I could see something like this as being more viable as it would provide benefit to airport workers, who should be targetted just as much by such a link as the passengers.

But wait, we’re getting ahead of ourselves here. Let’s take a look at what regular bus services presently connect the airport to these suburbs:

• Routes 478/479 go between Sunbury, the Airport and Moonee Ponds/Essendon via Airport West. Has an absolutely hopeless timetable, with a four-hour inbound gap in the middle of the day between 10am and 2pm. Most of the route south of Airport West duplicates the route 59 tram.

• Route 500 runs from Sunbury to Broadmeadows via the Airport, Gladstone Park and Westmeadows. Comes about every 90 minutes during the day on weekdays with  nothing on weekends, and the last bus leaves Broady at 4:48pm.

These services only provide an abysmal frequency that nobody would use with a choice, but provide the only direct link – basically, it’s a public transport black hole. As I have covered in They need a train, the main arguments for rail over other modes is speed and capacity. The latter is certainly not presently needed in this instance with so little of a ’service’ being provided.

I’ll cover the issue a bit more in the next section, but for this instance I’ll suggest that a much better bus service is all that money should be spent on for now. I also wouldn’t support the extension of tram route 59 to the Airport for the reason being that a bus could do the job as well.

Who goes to the airport?

Something that needs to be considered in all this is – where are people coming from? I’ve posted the figures above for how many work at Melbourne and Brisbane airports at the top (couldn’t find any specific figures for Sydney), which may be larger than what I have quoted as the figures are a few years old.

There’s a good Railpage post by savethehumans which highlights this issue well. Airline travellers make irregular trips, often with lots of luggage for which going by car or taxi is suited. They come from all directions, as the one airport services all corners of our sprawling suburbs.

Taxi – a popular mode of transport for airline travellers with luggage.

But the airport workers most likely make the same trip every day, and a large number would live in suburbs closer to the airport. If they could be lured to public transport, infrastructure like staff parking could be freed up to benefit airline travelers. Airports should be serviced just as well by public transport as other major commercial centres.

Any kind of airport railway or bus route needs to be a part of the larger picture for it to work. A frequent train service is of little use if there are no useful services to feed it, or if fares are set too high for workers.

Take a look at someone who might live in Arncliffe, and work at Sydney Airport (about 5km) away. They could take the train one stop to Wolli Creek and catch a train to the airport, but there’s only a train once every 30 minutes at Arncliffe, too infrequent to be useful. For the short hop on the train across the creek to the airport terminal they’ll pay a $10 surcharge on the normal distance based fare. Hence it makes more sense to drive.

Depending on the airport’s location, express services to/from the central CBD are unlikely to be of any use for most workers who live more locally. The main potential targets of these services vary; for visitors staying in the inner city, catching the Airtrain or Skybus makes for a handy transfer. If I wanted to connect from the airport to the suburban rail system, these services would also be of use.

Conclusions

I’m not going to say that it’s not worth trying to attract airline passengers to public transport, but the focus on such proposals is often too much on passengers and too little on the workers. There needs to be a balance.

My view on airport railways is that they are best done as a part of the regular suburban network in the respective city, than as special express lines with dedicated rollingstock. In some instances the one service can double up as being ideal for both types of journeys ie Brisbane’s. The benefit of special rollingstock is watered down if passengers need to take their luggage onto regular services to get home, unless their destination is within walking distance of somewhere served by the airport train.

None of the three airports discussed would I consider to be well serviced by public transport, although Brisbane is probably the best except for the poor Airtrain frequency/operating hours and the poor services in the ‘bigger picture’ to take people to where they actually live.

In any case, construction of major transport infraustructure must be backed with sound reasoning and justification, not the claims sometimes used in Melbourne that a line should be built to enhance the cities international reputation, or simply because “other cities have it”.

I’ll cap it off with – how high is transport to the airport a priority, compared to everything else in a large city? Are the existing taxis and cars already working well enough? Is it worth a major capital investment? Put it into perspective.

Looking forward to seeing comments from readers.

In light of Riccardo’s recent post Money for Jam: Contra-peak traffics and encouraging them I have had some thoughts of how you can justify providing a heavy rail service over other forms of public transport.

The main benefit of heavy rail is it’s ability to transport large numbers of people frequently, as well as speed. Both points are essential for any decent suburban commuter rail or metro service as well as a good interurban service, but where does the title shot of Hurstbridge in Victoria fit into all this, a town of 3,400 that gets visited by a 6-car electric set every 40 minutes? Or somewhere like poor old Armidale in the New England region of NSW which has it’s railway kept open for one loss-making CountryLink train per day?

This all boils down to the issue of cost recovery. Heavy rail is the most expensive mode to provide due to it’s high fixed costs yet low variable costs, making it an ideal mode for shifting a load of commuters every 10 minutes but not for a single two-car railcar per week (ala Griffith, NSW).

The justifications for a train service

There are regularly proposals from railfans for new passenger train services that are usually proposing to use the infrastructure purely as it is there (but usually needing upgrading), such as this Railpage thread proposing a service to SA’s Barossa Valley.

Of course if we are to ignore the political reasons that cause many services to exist, such as how Griffith, NSW receives a railcar once per week, running a choo choo purely because it is possible is a poor use of limited resources.

For a suburban commuter railway to be justified, I’d be hoping for a line with a good population catchment and intermediate destinations that would present consistent patronage. Certainly I wouldn’t support building expensive infrastructure (ie third tracks) that would only see use during peak hours.

And of course, I wouldn’t support jumping up from a lackluster bus service to an entirely new railway. This applies for some proposals such as for Doncaster and Rowville in Victoria, both of which should IMO not get railways until a high frequency bus service is not enough.

One pre-feasiblity study (1MB PDF) for the latter by Knox City Council actually went as far as suggesting the railway should be built because it would most likely provide Sunday & evening services at times when existing buses do not. This is a common myth similar to the buses/trams issue; although most suburban railways in Australia run to an 18/7/365 service standard, there is nothing to stop buses from running to a similar service level.

Having said that, there are numerous instances where the construction of a railway would not be warranted (if there was not already one), but providing services along the existing line is viable.

But on the other hand the operation of joke services like what QR provides to Doomben in Brisbane, a line that was electrified to Eagle Farm in 1988, replaced with buses in 1993, and had it’s passenger service re-instated to the present hourly Monday-Saturday service (not including evenings) in 1998 should be re-considered. Either a proper service should be provided or if the capacity of rail is not warranted the service should be ceased, as running such a poor service benefits few people and diverts vital rail funding.

Wrapping this up

I have done little to touch on the issue of country passenger trains in this post, particularily the issue of keeping a line open for only a couple of trains a day or services provided for welfare and/or political reasons. That will be for another day.

The conclusion of all this is similar to my Only Trams Can post, public transport should provide the most appropriate mode for each service. Just because there is already a railway (or even already a passenger service) it does not mean rail is the best mode for that service.

I’m particularily interested in seeing comments from others on this issue as I beleive there is much I haven’t touched on yet.

One common mass transit service model is the ‘commuter’ model. This model consists of usually high speed trains or buses, designed to take suburban white-collar workers to the CBD in the morning and back home in the evening.

The most well known examples are in North America, one such example being Toronto’s GO Transit which operates rakes of bi-level cars hauled by locomotives from seven lines into Union Station each morning, and back in the evening.

The Lakeshore East & West corridor is the only one to operate full-time albeit only at an hourly off-peak frequency, the others being replaced by buses outside of peak periods. In suburbs closer to the CBD, the Toronto Subway provides a full-time rapid transit service.

This service model is particularly wasteful as rollingstock & infraustructure that is needed for the peak service spends much time sitting around idle, and it only caters for trips to the City Centre.

Where have Australian systems headed

Australian cities have a history of well-planned contiguous urban development. Take a look at the suburban centres located along railway corridors as an example. Most lines could be considered closer to urban transit than commuter rail, but service frequencies and patronage patterns are closer to the latter. Very few locations receive a full time turn-up-and-go service.

Much emphasis is put on as peak hour crowding and services for peak CBD workers, and this is reflected in timetables. One example would be QR Citytrain’s Darra Station, which receives nine trains within the space of 60 minutes in the morning peak but only two per hour at other times.

Although most would describe Australian networks as commuter railways, they are by world standards a hybrid between commuter railways and rapid transit/metro systems. In terms of rollingstock, some stock such as CityRail’s DD cars is very much optimised for long-haul journeys (yet still used for short runs!) while others such as the Siemens stock used by Connex in Melbourne was initially designed as underground subway stock.

One strong point of Australian systems is that most services operate 18/7/365, with only few exceptions such as Doomben and Tonsley. This is similar to mass transit systems overseas, but few places can boast having frequencies equivalent to such systems.

In some places such as Brisbane and Adelaide, there is a tendency to run far more bus routes into the CBD than is required, when these vehicles could deliver passengers to the rail system. In the former this is caused by lack of co-ordination between Brisbane City Council and QR, the former planning a bus network that ignores the existance of railway lines in many instances.

Most Australian cities promote the concept of Park-and-ride as a way to get to railway or bus stations, a flawed concept that only serves for peak hour journeys and is only useful for trips to locations that are directly served by the rail/bus system. This is something that might be appropriate for a commuter transit service, but is ‘not good enough’ in a system that aims to cater for more than peak hour CBD journeys.

The divide between tiers

In numerous places there is a tendency for the same “one size fits all” service to be provided to both inner and outer suburban areas. Like how you have the same train servicing Caboolture as serves Albion, or the same train is provided to Pakenham as is provided to Hawksburn.

On lines where an interurban service is provided, this also leads to a sharp divide between the type of service provided on either side of the boundary. For example, Pakenham in Victoria receives an all-stations electric service making 24 stops before the City, but V/Line-served locations immediately east get more comfortable diesel trains, making only four stops in that section.

The needs of a suburban commuter railway are different to that of one serving denser inner areas, paritcular in terms of station positioning. For the latter, what you really need is well described in this quote from Riccardo’s blog, or the list of hobbyhorses to be specific:

42. Stations in the ‘commuter’ belt should not be 1-2 km apart, in fact a good 5 km apart is better as was shown by a study in the leadup to Perth’s new metro rail. Walkup passengers are insignificant to a commuter railway (at a non-city ‘origin’), with buses and car parks the realistic means of getting most patronage. Trains need to reach the optimum speed/catchment passengers tradeoff ie for every new passenger you reach by having a train pull up at their door, you lose passengers who find the journey time too long. In ‘metro areas’ on the other hand, stations should be very close together because walk-up becomes more important

And the issue here is that these corridors are trying to serve both roles at once. Regular proposals are made for new stations on commuter rail lines, such as for Lakeside between Officer and Pakenham in Victoria, that would just worsen the problem. Riccardo has also done a series on the Pakenham problem, much of which relates to poor urban planning than failings of the railways.

How many people are going to the City?

The majority of patronage on Australian rail systems is bound for the central CBD, which often makes proposals for new services to be designed around taking people to the City, as that’s where everybody wants to go, right? Like this one for express buses from Narre Warren and Pakenham to the Melbourne CBD.

An interesting analysis has recently been posted on SkyscraperCity, focused on the metropolitan areas of Melbourne & Sydney. It can be legitimately argued that the CBD and surrounding areas of each city is the largest individual employment centre, but only in the most inner suburbs is it the largest centre.

The amount of CBD workers is also proportional to the distance of the municipality from the CBD. 20.6% of Glen Eira (Vic) residents work in the CBD (slightly outnumbering those working locally), as compared to 5.1% of Frankston residents who do. Yet moves are always made to pander to outer suburbanites with things like express trains, when the focus should be on providing a high frequency service from these inner locations.

This shows that moves such as high speed trains from Penrith to Sydney would do little to cure PT issues
in these areas, as the majority of residents both work closer to home and would go shopping, etc nearby. The header photo of this post is a classic example, a major shopping centre has been constructed next to the Frankston Line in Melbourne, but the nearest station is around 800m away and proposals to build a new one have criticised because it would make trains take longer to reach the City.

This graph from Melbourne shows the modal share of public transport for CBD trips, and for all metropolitan trips, the results not being surprising:

Conclusion of this post

While the central CBD is always a significant employment, business and retail centre, public transport planning needs to focus on more than just delivering people to the City if it wants to acheive significant modal share increases. Services need to be designed to link suburban hubs & promote new developments (ie Dandenong Transit City in Vic).

Things like super-express buses and trains from outer suburbs to the CBD are a waste of resources and encourage sprawl & long distance commuting. Long distance commuting should be discouraged, although this can be done better ways than attempting to sever transport links.