Archive for February, 2013

The financial collapse of Brisbane’s airport link tunnel last week adds to a growing list of failed road projects funded as a Private Public Partnership (PPP). Brisbane’s Clem7 tunnel, as well as Sydney’s Cross City and Lane Cove Tunnels have all gone into receivership following lower than predicted patronage and thus an inability for the operator to pay its debts. In each case the road continued to operate and was sold to a new owner at a discount price.

NOTE: You can also add Sydney’s Airport Line to that list of failed PPPs, but it’s a different case and so won’t be discussed here

As this is a private company, the financial burden is on the investors and lenders, rather than the government. Queensland Premier Campbell Newman commented that “one thing I’ll say about Airport Link is this is one decision…that the taxpayers won’t end up paying for, thankfully”.

There is nothing stopping the Queensland Government from being the one to buy the road, something which previous Queensland or NSW Governments have not taken up the option of doing despite the . Even if it doesn’t, the infrastructure is now built, and will eventually revert back into public ownership.

The problem appears to be with the traffic forecasting. This is roughly how it works.

First, the private company works out how much the project will cost to build. Although there can be cost blow-outs, this is still the easiest bit to estimate so let’s assume it’s 100% accurate every time. They then multiply this amount by a percentage deemed to be a sufficient return on investment (usually the going rate of interest plus a few percent) to gives the necessary revenue.

The government then provides the company with traffic forecasts, their estimate of roughly how many cars are expected to use it. The company then divides the revenue figure by the traffic forecast to get a toll amount.

EDIT: It’s been pointed out in the comments section that these estimates can sometimes come from a private consultant, rather than the government. Where it comes from isn’t really all that important, and there was a reason that the description above was “this is roughly how it works”. But it’s a valid point and is noted.

Where it all went wrong is that, for a number of reasons including poor forward planning, freeway construction in Australia has been requiring more use of tunnels (all 4 examples above were tunnels). These cost about 4 times as much as surface construction. So, 4 times the cost means 4 times the necessary revenue, and 4 times the necessary toll. This assumes static traffic forecasts.

In reality, they are not – they are sensitive to price. Higher tolls mean lower traffic. And what happens if you revise that traffic forecast downward? The required toll goes up even further! A vicious cycle if there ever was one.

There are only really 2 solutions to this.

One is to go back to having the government build roads. Michael Pascoe points out that this also has the added benefit of allowing access to cheap debt, as government loans have a lower interest rate than loans to private businesses. However, government debt is now a dirty word, and so this is an unlikely option to be chosen.

The other is to reduce the cost of construction. Good planning can do this in the long term. The M2, which is entirely on the surface thanks to the maintenance of a dedicated reservation for it decades in advance, made an after tax profit of $38 million last year (Source: page 74, Transurban Annual Report). Whereas plans for the M4 East portion of WestConnex have it slotted like the Eastern Distributor rather than tunnelled in order to cut down on costs. While there remain a number of concerns over WestConnex itself, the aim of finding ways to bring down its cost of construction is definitely a welcome one.

Sydney has been compared unfavourably to other countries through poor use of statistics. But that does raise the question: is it even fair to compare Sydney (or any Australian city) to the rest of the world? That question recently led to this exchange on Twitter:

Twitter conversation

There are many examples of Australians returning from abroad and singing the praises of the transport systems of European or Asian cities. “If only we had a system like that over here” they say. But that is like putting a square peg through a round hole. Those cities have a different history to Australian cities, and have developed differently as a result. You cannot then just overlay their transport system on a different city anymore than you can wear your cousin’s suit to a wedding when that cousin happens to be taller and skinnier than you.

European cities were built over hundreds, sometimes thousands of years. In many cases, most of their footprint was set before the time of the motor vehicle. In fact, their urban layout was more likely to be shaped by long distance commuter trains and short distance trams, leading to a city and transport system that is well suited to public transport.

Asian cities, and Middle Eastern ones to a lesser extent, are much newer, and were built during a time when most Australian households had a car in every garage. But 2 things were different. First, most Asian households were poor, and could not afford cars. They relied instead on public transport or bicycles. Secondly, Asian governments tend to have much less of an interest in due process and individual rights than Western governments. If your house is in the way of the new rail line, you are moving out and the bulldozers are moving in. This has allowed them to fast track shiny new underground metros and high speed rail in a way that would never be possible in Australia.

So that leaves the New World (leaving Africa aside): North America and Australasia  plus also South America. Countries made up predominantly of migrants, which also saw big population booms in the aftermath of World War 2. This was a time when the private motor vehicle began to really spread, when increased populations were settled in the city’s fringes leading to the beginning of urban sprawl, and when governments began building highways instead of railways in order to move people around.

When comparing Sydney to cities around the world, it is places like Auckland, Los Angeles, and Vancouver that comparisons should be made with, not London, Paris, or Tokyo. And when looking at how to improve Sydney’s transport system, it is successful New World cities which hold the answer.

Strangely enough, Los Angeles probably provides the best example of how Sydney can improve its transport system. (The link is to a piece by Jarrett Walker at Human Transit, which is well worth a read for more details). Despite being well known as a car dominated city, LA has been making an effort to improve its public transport system, and shares many similarities with Sydney. Both cities have large footprints and built a network of highways after the rise of the car following World War 2. Each is also made up of multiple urban centres with a dense core, Sydney has the Sydney CBD, Parramatta, Liverpool, Penrith, Chatswood, etc while Los Angeles has downtown Los Angeles, Santa Monica, Westwood, Burbank, etc. However they also have their differences – Sydney has a strong heavy rail network, while Los Angeles has an well laid out grid street system.

Ask Gladys

Posted: February 23, 2013 in Transport
Tags: , , , ,

MX ran a feature yesterday called #AskGladys in which readers could tweet in questions to the Transport Minister Gladys Berejiklian. Those that were forwarded to her were all answered an then posted to Twitter.

The responses give October as the date that the new 2013 timetable will be introduced, which means the date it is made public will be some time before that. They describe the new timetable as having more consistent patterns, which appears to mean a more harmonized stopping pattern. They also state that frequencies on the North Shore Line will be increased to the maximum 20 trains an hour currently possible, up from the current 18 during the morning peak.

Some highlights are included below, with the rest available by searching the hash tag #AskGladys on Twitter.

Note: It might be worth reading part 1 and part 2, which provide some context and outline the problems with the current timetable, if you haven’t yet done so.

Cityrail has been simplifying its network ever since the Clearways project was announced in 2005 around the same time as the major timetable changes were introduced that year. The idea behind Clearways was to increase capacity (via additional “turnback” platforms and/or track amplifications) around the network where pinch points caused bottlenecks and to separate the network into 5 separate sectors (which would then converge into 3 sectors in the CBD). This is known as sectorisation, and involves creating sectors that run as independently from each other as possible. As a result, delays in one sector do not spill over into other sectors.

The Cityrail network currently has 3 sectors:

  • Sector 1 – made up of the Eastern Suburbs and Illawarra Lines

2013-01-20 Cityrail Map Sector 1

  • Sector 2 – made up of all the lines that use the City Circle, plus the Cumberland Line (Note: the Inner West Line between Strathfield and Homebush shows up faded in error.)

2013-01-20 Cityrail Map Sector 2

  • Sector 3 – made up of all the lines that use the Harbour Bridge

2013-01-20 Cityrail Map Sector 3

(In practice, sectors 2 and 3 are not entirely separate, with trains on the Western Line and South Line sharing some track between Granville and Homebush, as well as the Western Line and Cumberland Line between Blacktown and Harris Park.)

A Herald report from 2012 revealed that one plan would involve fully separating Sectors 2 and 3. Currently the 2 track pairs between Blacktown and Homebush are used to separate local (all stops) services from express services. This allows express trains to overtake slower local ones. Separating trains on these tracks by sector rather than by stopping pattern then means that an express service could get stuck behind a slower local service. The solution to this would be to also harmonise stopping patterns – with sector 3 running only express services and sector 2 running only local services.

If implemented to the fullest extent, the Richmond and Northern Lines would be separated from the Western Line. Richmond Line trains would become part of the Cumberland Line, running all stations to Campbelltown. This would eliminate a conflict that currently exists at Granville where a flat junction is used by Western Line and South Line trains (by sending Richmond Line trains on the Cumberland Line’s flyover at Merrylands and sending Western Line trains on the Northern track pair not used by the South Line, thus avoiding the flat junction). Northern Line trains would use a third track pair that begins just before Strathfield at Homebush Station and then ends at Sydney Terminal at Central Station, effectively creating a fourth sector. Inner West Line trains would be truncated to Homebush, which relieves some pressure on the heavily used Lidcombe to Homebush portion of the network, allowing South and Western Line trains to pass through there more easily.

This would allow Western Line trains to run faster (by permanently skipping many stations)  and more frequently (as they are not sharing any track with Richmond, Northern, or South Lines as is currently the case). Passengers at stations like Toongabbie, Pendle Hill, Wentworthville, and Harris Park would need to catch a Cumberland Line train and change to a Western Line train if they are going into the city. While passengers at stations like Clyde, Auburn, Lidcombe, or Flemington could change to a Western Line train for a faster journey, or stay on a slower all stations South Line train for a direct one. On the network map, this is what it could look like (again, this is purely speculation based on rumour at this point).

What the Cityrail network might look like after the 2013 timetable is implemented. Click on image for higher resolution. (Souce: Cityrail.)

What the Cityrail network might look like after the 2013 timetable is implemented. Due to an error, Auburn should be the blue South Line only, not the yellow Western Line. Click on image for higher resolution. (Souce: User created from Cityrail.)

Creating these truly independent sectors would also allow for harmonisation of stopping patterns and rolling stock. With high enough frequencies, this will also mostly do away with the need to worry about delays. After all, if a peak hour train comes every 3 minutes and all the trains on that line have the same stopping patterns, then a 3 minute delay effectively puts everything back to normal.

It also makes many commutes easier – with commuters just taking the next train rather than waiting for their train, which will help to reduce station overcrowding on congested CBD stations (by requiring commuters to transfer to another train once they are out of the CBD). Frequencies will also improve, ensuring that commuter wait times are kept to a minimum and allowing many commuters to travel without having to worry about consulting the timetable first.

Higher off-peak frequencies could also mean shorter trips by way of reduced wait times. Parramatta currently has 5 trains an hour into the CBD during the off-peak, meaning a maximum wait of 15 minutes. Increasing this to 8 trains an hour would mean a maximum 8 minute wait, or 4 minutes on average. Similarly, someone taking the train from Pendle Hill currently has to wait 30 minutes for the next train during the off peak, which often means either arriving much earlier than necessary or taking the risk of missing the train and waiting half an hour for the next one. Either way, this means a longer overall journey time. But having 10-15 minute frequencies, and then transfering to a frequent (and express) Western Line train into the CBD, could result in a faster and more reliable journey, despite the removal of direct services. Someone wishing to make a North/South trip, say from Quakers Hill to Merrylands, will now have easy all day access by rail.

The main downside is that it will force many people to transfer to another train. Many commuters on the Richmond and Northern Lines will need to transfer to another train if travelling into the CBD, as they will no longer have direct access.

There do exist alternatives, Simon blogs at Fixing Sydney Transport about how Parramatta can be made the terminus of the Cumberland Line, thus maintaining the second track pair West of Parramatta free for Richmond Line trains. Doing this would allow Richmond and Epping Line trains to keep their CBD access, while still eliminating a conflicting move (by Western and South Line trains) on the flat junction at Granville that currently exists. It would not allow a complete harmonisation of stopping patterns, but does deliver some benefits of the complete sectorisation without most of the disadvantages it would bring. There are merits to this option, and would be an improvement on the status quo.

Ultimately, the government’s decision to run the Northwest Rail Link (NWRL) as a completely independent line (which could become the fifth sector), means that the existing Harbour crossing will need to be run at maximum efficiency during the decade between then NWRL’s completion and when a second Harbour Crossing is built, as this will become one of the biggest bottlenecks on the network. The easiest way to achieve that is to implement the sectorisation outlined above. So if it doesn’t happen in this year’s 2013 timetable, then expect it to happen when the NWRL opens at the end of the decade.

The Herald reports that changes to the STA will allow Sydney Buses to maintain existing services with fewer buses and drivers via a more efficient allocation of buses between depots. Currently, many buses must travel long distances to the starting point of the bus route and then do the same to return at the end of their run, a process known as dead running because the bus is unable to carry any passengers.

Dead running bus

The changes will see buses spend less time showing “not in service” displays and more time transporting passengers around. (Source: Author.)

While commuters will not notice any differences with service levels, the reduced number of buses needed to maintain the same service levels means the STA will be able to retire some of its older buses, particularly non-accessible and non-air conditioned buses. No bus drivers will lose their jobs directly, though the STA plans to reduce staffing levels via natural attrition.

These changes were made possible by the reorganisation of the Transport Department into Transport for NSW, which also involved putting all government transport agencies directly underneath it and transferred responsibility for planning to Transport for NSW. This meant that rather than assigning buses to routes at a depot level, it is now done at a Sydney-wide level, allowing for efficiencies like these to emerge.

What is disappointing is that the savings that arise from these changes will be going into general government revenue rather than back into public transport. The same is true of the government’s share of station access fees on the Airport Line.

Finding savings in public transport (or raising extra revenue) should be encouraged as a way of expanding services without having to increase its budget. After all, public transport costs the state government billions each year in operational subsidies. But this cannot be done if these same savings are then used for other purposes.

With reliability on the Cityrail network sagging and extra capacity needed to handle the growing transport demand, the solution that the NSW government is seeking has been to reduce network flexibility (see the previous post on the 2013 timetable for an explanation of how reliability, capacity, and flexibility are interrelated).

Greater flexibility, the ability to get a direct train to the CBD from any station on the network, can reduce capacity and hurt reliability for a number of reasons. These include the need to have flexible (i.e. different) stopping patterns on the same track, the need to run separate lines on the same track, and the need to use flat junctions (see below). All of these things mean that increased flexibility also adds complexity to the network. Meanwhile, a simple system would require many passengers to change to a second or even third train in order to reach the CBD. This is a potentially unpopular option and is why Cityrail has opted for flexibility up until now.

Flexible stopping patterns

Some trains are express services, while others are all stations or limited stops services (which will collectively be called local services).

Often there are two pairs of tracks running side by side (such as from Central through to Hurstville or Parramatta), then one is generally designated for express trains and the other for all stops trains. This allows the express services to overtake the slower local services.

In other cases there is only a single track pair (such as much of the Northern Line between Hornsby and Strathfield), meaning that express services can get stuck behind local services. Here you must make a choice between flexibility (i.e. running both express and local services) or capacity (i.e. running just one type of service, generally local, at higher frequencies).

Independent lines

The Cityrail network operates 3 different “sectors”, effectively independent lines. However in some parts of the network (the section between Granville and Homebush in particular) it is necessary for trains from different sectors to mix in order to maintain flexible stopping patterns. This means that delays on one line can often spillover onto another line.

Take the huge delays that occurred 2 weeks ago that originated on the North Shore Line due to overhead wiring issues. This had immediate impact on the North Shore Line, and soon spread to the Northern and Western Lines whose trains feed into the North Shore Line. However, as Western Line trains also share track with the Inner West, South, and Bankstown Lines, these too soon began to suffer delays, which later flowed on further onto the Airport and East Hills Line. Only the Eastern Suburbs and Illawarra Lines, which is the only one of the 3 sectors to be completely separate, did not suffer delays.

Independent lines would not have prevented the delays, but it would have quarantined them and stopped them from flowing onto others.

Flat Junctions

Where two roads intersect at traffic lights the signals alternate so that only traffic on one road can move. For every minute that one road has a green signal, the other road loses a minute of green. This is effectively what happens in the world of railroads when a train crosses from one track to another that are separated by a third track in between them going in the opposite direction. These are referred to as “flat junctions” (see diagram below for an example), and every train that crosses one prevents another train from using the other track at that moment. You can avoid the use of flat junctions through the use of dives and flyovers (the best example of which can be found between Redfern and Central Stations), but these are not always available nor feasible to build.

Flat junctions cause conflicting moves, preventing another train from using the track that is being crossed over. Click on image for higher resolution. (Source: Sydney Rail Future Report, page 7.)

Flat junctions cause conflicting moves, preventing another train from using the track that is being crossed over. Click on image for higher resolution. (Source: Sydney Rail Future Report, page 7.)

The diagram above shows a section of rail near Macdonaldtown, right before the Western and Illawarra Lines merge. The maximum capacity of each track is 20 trains per hour, and each train that crosses a track via a flat junction effectively uses up one of those 20 hourly “slots” on the track that it crosses. So the 7 trains that move from the “Up Main” track to the “Up Suburban” track use up 7 slots on the “Down Main” track, restricting that third track to 13 trains per hour rather than the maximum of 20.

More importantly, conflicts due to flat junctions hinder reliability. When a train runs late across a flat junction, it delays not just the trains on its line, but also the line that uses the track that it crosses. The 7 trains that used the flat junction above can be timetabled in,  but if one of them runs late and stops following the timetable then the delays flow on to the trains on the “Down Main” track, even though. These delays further constrain capacity, because for every 3 minutes of delay you effectively reduce the maximum capacity on a line by 1 train per hour. The line across the Harbour Bridge is a great example – here you often only see 17 trains crossing the Bridge during the busiest hour of the day even though 19 trains are timetabled, so actual capacity is 10% lower than timetabled.

A second airport in Sydney’s West should be built at Badgerys Creek, according to a panel of experts who addressed a forum at Parramatta earlier today. The forum, organised by the Western Sydney Community Forum and Western Sydney Business Chamber, included academics, social workers, and consultants as speakers who unanimous backed the proposed airport, arguing it would bring significant economic benefits with only minor noise pollution. The panel also rejected Wilton as an alternative, with a co-author of the recent Joint Study on Aviation Capacity saying that the study did not conclude that the options were “Badgerys or Wilton – the options were Badgerys, Badgerys, Badgerys, Badgerys, Badgerys, or Wilton”.

Current and proposed Sydney airports. Click on image for higher resolution. (Source: Google Maps)

Current and proposed Sydney airports. Click on image for higher resolution. (Source: Google Maps)

This is despite the there still technically being bipartisan opposition to an airport at the Badgerys Creek site. This opposition can be traced back to the construction of the third runway at Mascot’s Kingsford-Smith Airport (KSA) in 1994. Chris Brown, a member of the steering committee of the Joint Study on Aviation Capacity, described the campaign against a third runway as a grassroots movement that was responding to huge noise impacts on a large population. Meanwhile, the opposition to Badgerys Creek was led by those at the top, according to Mr Brown. It was argued at the time that Western Sydney should not be the dumping ground for Inner Sydney’s aircraft noise. In addition, the extra capacity provided by this third runway at KSA made politicians complacent on the issue of an airport at Badgerys Creek, according to the former Federal Airports Corporation and Australian Rail and Track Corporation head, Barry Murphy.

The opposition to Badgerys Creek has begun to wane recently, particularly when the issue of the economic benefits that an airport could bring to Western Sydney was raised. Mr Murphy points to cities like Dallas-Fort Worth and Atlanta, which leveraged a large airport to huge economic benefit. Director of Planning with Cox Richardson, Bob Meyer, argues that Western Sydney’s jobs shortfall of 163,000 jobs is putting a huge strain on the transport network, and that this will only get worse by mid century when that jobs shortfall is projected to rise to 406,000 jobs. He believes Western Sydney has abundant supplies of industrial zoned land that could be used for employment, but that it needs a catalyst such as a new airport in order for this to happen. This would then stem the jobs shortfall afflicting Western Sydney.

The panel also dismissed the impact of aircraft noise on Western Sydney. An A380 makes only half the noise of a 747 jumbo jet on take off, according to Mr Murphy, while the smaller A320 makes 75% less noise than a 747. Mr Meyer showed attendees a map of the hypothetical flight paths from Badgerys Creek, which showed that most of the noise would occur over industrial lands rather than residential. This, along with lower urban densities, means that slightly over 300 homes would be affected by high levels of aircraft noise from Badgerys Creek, compared to almost 30,000 from KSA.

Note: This post was starting to get a bit long, so I’ve split it up into 3 parts. Here is part 1.

A new Cityrail timetable is released each year, and while usually this means adding a couple of extra services or moving some trains a few minutes either way, this year’s will be the most dramatic change since the 2005 timetable.

“a brand-new rail timetable is being written from scratch, and will be released this year to provide more express trains, quicker travel times and ultimately improve the customer journey”Gladys Berejiklian, Transport Minister (15 January 2013)

Writing a timetable is a balancing act involving trade offs. One way of looking at it is by breaking it down to 3 different variables, and a decision that needs to be made on which of these 3 to prioritise. Given the problem of limited resources, you need to pick two of those three and forget about the third. If you try to get all three, then you might end up with only one. Those variables are:

  1. Capacity – the number of trains that can be run on a particular track of rail line per hour
  2. Flexibility – the ability to get from A to B without having to transfer to another train (generally this means being able to get from anywhere to the CBD, and vice versa, on a single train)
  3. Reliability – how well trains run to timetable (rather than being delayed or cancelled)

Take the 2005 timetable as an example. Cityrail had become notorious for its unreliability. On time running data shows that in 2003, 80%-90% of trains ran within 4 minutes of the timetabled time. But in 2004, this figure had dropped to 50%-60%, and remained that way until mid 2005 when 2 things happened. First, on 1 July 2005, on time running was redefined from being within 4 minutes of the timetable, to within 5 minutes of the timetable. This resulted in an immediate statistical, though not actual, improvement to reliability (from 65% in June, to 77% in July). Then in September a new timetable was introduced, cutting 1,350 weekly services and slowing down the remaining services. The government had decided to cut capacity in order to improve reliability, while maintaining flexibility. And it worked, with on time running improving right away (from 78% in August, to 94% in September). It then stabilised in the low 90% range. While many complained about the longer journeys and lower frequency of trains, the 2005 timetable did finally return Cityrail to a reliable service.

2013-01-17 On time running 2005

Fast forward to today, and on time running is again on the decline. This time it has been caused by overcrowding on trains, leading to higher dwell times on major stations. This overcrowding is in turn due to rising demand for public transport, and also a lack of capacity to deal with the higher demand. Ironically, one of the biggest constraints on capacity was the timetable changes from 2005. Clearly, the 2005 timetable has run its course and needs to be revised again. Whereas last time the government chose flexibility and reliability over capacity, this time it looks to be opting for more capacity at the expense of flexibility in the hope of reverting to improved reliability.

These problems are not exclusive to Sydney. Melbourne has seen similar problems, and has also been tackling it via a more simplified, albeit less flexible, rail network. The video below outlines these problems and the solutions to these problems from Melbourne’s perspective.

Note: You can also get additional capacity by building new infrastructure. But the problem of limited resources and the trade off mentioned earlier still exists. The difference is you now have more resources to allocate between those 3 priorities, but a choice will still need to be made on how to allocate them.

Friday’s post on whether the Northwest Rail Link (NWRL) should be a metro generated a large volume of comments, sufficient enough to warrant a new post to present some highlights and respond to the points raised in them. It’s worth reading the original post for some context if you have not already done so.

The NWRL alignment. Click on image for higher resolution. (Source: Transport for NSW.)

The NWRL alignment. Click on image for higher resolution. (Source: Transport for NSW.)


“People have commented that the dwell time at Town Hall, the most densely used station is rarely more than one minute, loading and unloading takes less and trains wait for their timetable departure – this means that 2 minute headways through the existing City Tunnels are feasible. Halving the existing signal spacing on the North Shore Line and the Bridge would enable trains to run closer together.” – Dudley Horscroft

“If headways could be reduced to 2 minutes with improved signalling I believe that double decker stock will move more people per hour than single deck.” – moonetau

The argument here is that signalling will allow double deck trains to achieve the same frequency that moving to single deck would provide. The problem with that argument is that the same signalling would also improve the frequency of single deck trains, so single deck would still achieve greater frequency and lower headways than double deck. That is why the single deck Paris metro can run 40 trains per hour, while its double deck RER system runs 30 trains per hour.

Smaller tunnels

“Certainly building the tunnels so that double deck trains will not be able to use the NWRL is crazy. It means that in the future someone will have to design future double deck trains to fit into the smaller diameter tunnels! Reports indicate that the reduction will be only 400 mm.” – Dudley Horscroft

Smaller tunnels remains the least defensible part of the NWRL as proposed. Concerns like this are justified.

NWRL decisions are political

“It is easy to criticise the NWRL given many of the decisions have been made for political rather than operational reasons. The desire to separate it from Cityrail is to weaken the role transport unions will have in running it. Rightly or wrongly the current government believes all the problems with Cityrail are down to the staff.” – Jim

“I would include the entire decision to build the line as one being made for political rather than operational reasons.” – Simon

The decision to build the NWRL as an independent line, to be operated privately, does appear to be an ideological one based on the belief that the transport union is at least partly to blame for Cityrail’s high operating costs. Jeff Kennett’s Liberal government in Victoria took such an approach when it eliminated guards from its trains in the 1990s. However, while these decisions were political, they most definitely do have operational impacts, ones which in the Victorian case resulted in an improvement rather than deterioration to the transport system. This is not necessarily a bad thing.

Widening the stairs and/or doors

“There is no reason that stairs should restrict the access to and from the decks – the trains are about 3 metres wide, and half this could be stairs from top deck with the other half stairs from the lower deck.” – Dudley Horscroft

The problem with widening the stairs, is that it doesn’t eliminate the bottleneck, it only shifts it deeper into the train carriage. The space between seats in the saloons upstairs and downstairs is only wide enough for 1 passenger at a time to approach the stairs, so even if the stairs are wide enough to handle 2 passengers, only 1 will reach it at a time. Removing seats can allow more passengers to reach the stairs, but by getting rid of seats you are eliminating the primary advantage of double deck trains over single deck ones – higher seated capacity.

“Introduction of carriages with 3 wide doors per carriage like the MI09 on the Paris RER line A will significantly reduce dwell times. – moonetau

Double deck trains on the Parisian RER system do achieve lower dwell times, allowing them to have headways as low as 2 minutes (compared to Cityrail’s 3 minutes). However, it remains a fact that Paris’ single deck metro trains still have lower headways at 85 seconds.

“Moonetau is right re “Current two door per side double deck rolling stock is half of the problem.” Three doors would be better, but inserting an extra door in the middle of a car would be rather difficult to say the least. Better to add an extra wide door at each end of the car where there would be less of a problem, the floor is already at platform height. This would give 32 doors per 4 car set, 30 if it is not possible to fit in an extra door adjacent to the driver’s compartment. Together with improved internal stairs this should markedly reduce dwell time.” – Dudley Horscroft

Similarly as with stairs, widening doors would not decrease dwell times as they are not the bottleneck. Wider doors with single deck trains, however, would see a drop in dwell times.

Global economic arc

“What bothers me about running this line as a metro is that is it connecting the least densely populated suburbs of Sydney with the ‘global arc’.

It is servicing semi-rural suburban areas and ignoring the most densely populated suburbs. That’s what metros should be doing; servicing high density areas and doing so quickly.” – Thought

“The fact that that there may be significant numbers alighting /boarding at global arc stations does not really matter as they will never reach the number of transfers TH and Wynyard (must be nearly 40,000 per hour in the am peak) and will not influence dwell times.” – moonetau

There is much uncertainty about this. Take, for example, the fact that more resident of North West Sydney work in the Global Economic Arc (7.7%) than in the CBD (7.3%), or that places like Macquarie Park are forecast to see a high rate of jobs growth in coming decades. While Northwest Sydney is mostly low density suburbia, the areas around the corridor of the NWRL are not, and will only get more dense once the line is up and running.

It might be that these stations North of the Harbour do not get the sort of passenger turnover that the CBD does. But they will certainly have a similar or higher level of turnover as other inner-city stations on the network. Given single deck metros are more suited to high turnover style patronage, that is why the NWRL would still work as a metro despite it’s long distance.

Non-CBD connections

“How would one from Parramatta or Burwood or Mascot get to Macquarie Park easily? They can’t.

A similar proportion of residents in the Hills work in Parramatta as they do in Sydney for example. Who has encountered the traffic streaming down Windsor and Old Windsor Roads?” – Thought

“We are still basing transport planning on getting in and out of the city when the majority of journeys are across the suburban area. Many people from the Hills area work in Parramatta to Penrith and many people who work in the Hills come from the western suburbs. They are ignored by the city centric planners.” – Jim

These are valid concerns, but have little or nothing to do with the type of line that the NWRL should be. Whether it is single deck or double deck, it still will not serve the sorts of connections listed above.

However, in the defence of the NWRL, traffic on the M2 is much worse than the Old Windsor or Windsor Rd in my experience. So if tackling congestion is the top criteria, then a Hills to Macquarie Park link is more important than a Hills to Parramatta link (which does currently exist in the form of the Northwest T-Way). Additionally, the NWRL appears to have been criticised simultaneously in parts for both not connecting to the CBD (despite connecting North Shore trains set to arrive every 3 minutes in peak hour) and for being CBD centric (despite linking up the Northwest with the Global Economic Arc, which employs more residents of Northwest Sydney than the CBD does).

EcoTransit recently produced a video attacking the government’s plans for the Northwest Rail Link (NWRL). It’s a well produced video that provides some good background and makes some good points. But it’s also a bit off the mark in some instances, which are discussed below, following the video itself. The video makes two main arguments about the NWRL: that it should be double deck and that it should be publicly operated.

The video gives some background on how metro systems developed around the world, where typically you have a long distance commuter rail system (often but not always double deck) combined with a short distance metro rail system primarily within a roughly 10km radius of the CBD. Commuters from the suburbs would catch a commuter rail train into a central station, where they would change for a metro train to travel within the CBD itself. Residents of the inner city could catch a metro train directly. Commuter rail is designed around peak hour travel, and off peak will often only have hourly services, while metro rail is all day and frequent.

The NWRL alignment. Click on image for higher resolution. (Source: Transport for NSW.)

The NWRL alignment. Click on image for higher resolution. (Source: Transport for NSW.)

The beauty of Sydney’s system, the video correctly points out, is that rather than having 2 separate systems, it combined the two. And so Cityrail trains from the suburbs arrive at Central Station, but rather than terminating there, they continue through an underground CBD subway. This has many advantages, primarily allowing a seamless journey into the CBD, and avoids the need for large amounts of scarce CBD space that would be required for a commuter rail to metro rail interchange. But it also has its disadvantages, such as infrequent trains outside of peak hour and unreliable services during peak hour, which the video does not address.

The video also criticises the previous Labor government’s various metro proposals, the Northwest Metro (which would travel under Victoria Road) and the CBD Metro (a shortened version of the previous proposal). The problem with these metros is that they seemed to be designed as a way of building metros almost for the sake of building metros. They were the wrong solution for Sydney, not because they were metros but because they were the wrong metros. Luckily, they were eventually dumped, but not before the government spent $500m on the project.

Single deck metro systems are designed for short distances to dense city centers with stations spaced about 1km apart. The Northwest Rail Link does none of these, and is inappropriate for a metro system.

A metro would definitely be inappropriate for making a long distance trip to a single employment center, these are trips where passengers get on at suburban stations and then all get off when they reach the CBD. While most of Sydney is low density suburbia, the NWRL alignment is a dense corridor very similar in nature to the CBD. It is full of employment and residential centers, resulting in a constant turnover of passengers both getting on and getting off at many stations along the way. Single deck trains, which lack the bottlenecks that double deck trains’ stairs have, ensure that dwell times will remain low at these stations. In fact, the proposed NWRL’s 47km alignment from Cudegong Rd to the city, which passes through the “global economic arc” of Macquarie Park, Chatswood, St Leonards, North Sydney, and the CBD, is probably the only possible long distance (significantly over 10km-20km) rail line in the Sydney basin that suits metro style operations.

Cityrail doesn’t need to convert to metro to increase the existing frequency from 20 trains per hour to 30 trains per hour. RER has 2 minute headways with double deck trains, so Sydney could get higher frequencies without shifting to single deck.

Comparing the headways of Parisian double deck trains to those of Sydney single deck trains is comparing apples with oranges. Paris has a far more advanced signalling system that allows trains to safely run closer to each other. The point is that single deck trains will, all else equal, always be able to run more frequently than double deck trains. This is due to single deck trains having shorter dwell times from quicker boarding/exits by passengers. You need no more evidence of this than to see that while Paris’ RER system has 120 second headways, the Paris metro’s headways are even shorter at 85 seconds.

NOTE: Dwell times are important, as long dwell times lead to delays. These delays then limit the number of trains that can pass through a given station each hour. Once you limit the number of trains per hour, you are reducing the overall passenger capacity of that line. It is not uncommon for 17 Northbound trains to cross the Harbour Bridge during the busiest hour in the morning when 19 are actually timetabled. So ensuring low dwell times can actually increase passenger capacity.

Paris is replacing its single deck trains with double deck trains. Having converted all Cityrail trains to double deck, going back to single deck would be a step backwards.

This is true of the RER system, which is their commuter rail network, and something that Sydney did many decades ago. But they are not converting their metro system to double deck. A metro needs to be able to handle high passenger turnover, and this is the achiles heel of the double deck train. They provide lots of seats for a comfortable long distance journey, but they do this at the expense of allowing high numbers of passengers to get on and off quickly.

Double deck trains have 50% more floor space, meaning 50% more capacity.

The former is true, and the latter is also true if the configuration of seating is the same. However, it is not. Single deck trains will have fewer seats, allowing more space for standing passengers. Double deck trains cannot achieve this without a blowout in dwell times as the stairs into the vestibules are only wide enough for one person at a time. Single deck trains have no such constraint and so you are able to remove seats in order to increase capacity without longer dwell times.

A metro’s lack of seating will result in passengers from the Northwest to stand if making a 40 minute journey into the CBD via the NWRL.

Unlike other lines in the Cityrail network, the NWRL is not one where passengers continue to baord the train as it approaches the CBD, then spill out in the city. They will continually board and exit the trains as it passes through job rich areas like Macquarie Park or the North Shore. This constant turnover of passengers means seats will often become available during the journey. Only about half of all commuters predicted to use the NWRL are expected to be travelling to the CBD, with almost half getting off before crossing the Harbour. Additionally, those making the long journeys, say from Rouse Hill to the CBD, will be boarding an almost empty train, thus be almost guaranteed a seat the whole way (a similar mirrored scenario will exist for the return journey in the afternoon, where high passenger turnover will provide many opportunities for a seat if the train is full when it leaves the city).

Large number of passenger will have to change trains at Chatswood, which could lead to many passengers getting stuck on the platforms, particularly if a CBD bound train is cancelled.

This is absolutely true, and one of the biggests risks that the NWRL poses. However, the fault of this is not that the NWRL is being operated as a completely different system to the Cityrail network, but that there is only so much capacity across the Harbour. If NWRL trains were sent directly into the CBD, then it would limit the number of North Shore trains that would be able to do the same. The only real solution here is to build in more capacity. In the short term this means a quadruplication of the track between Chatswood and St Leonards, allowing NWRL trains to continue through to St Leonards, and in the long term it means building a Second Harbour Crossing.

The Transport Minister, Gladys Berejiklian, promised the NWRL would be integrated with the Cityrail network and only made these changes to get the support of Infrastructure NSW Chairman, Nick Greiner, given that Mr Greiner is a big proponent of privatisation.

It’s true that Ms Berejiklian broke her promises on the NWRL. She promised it would be operated with double deck trains and that trains from the Northwest would travel directly through to the CBD. Both of these will not be the case. (Incidentally, the government has also promised that the NWRL trains will not be driverless, which hopefully will be another broken promise given the benefits that driverless trains would bring.)

However, it’s not clear that this was done to appease Mr Greiner. It would appear more likely that Infrastructure NSW was told that the NWRL was government policy and not negotiable, given the government’s desire to not be seen to back away from a transport infrastructure project like the previous government had with its metro proposals. Nor did the government seek Mr Greiner’s approval on other projects (other than WestConnex), as every time the Transport Plan and Infrastructure Plans disagreed, the government opted to take the advice of Transport for NSW’s report.

Instead, this raises the questions over whether privatisation is inherently a bad thing. If privatisation is implemented like the Airport Line was, were a private company owns the stations and charges a station access fee, then it will not work as part of an integrated transport system. However, if it is implemented like the Sydney bus network or Sydney ferries, where the government pays private operators to run the vehicles, but the government sets and collects fares from commuters, then it can be a way of reducing costs while ensuring services are maintained at a contractually set level. All indications are that the latter is true in this case, particularly given that the government has seen Cityrail’s costs spiral out of control. So if introducing private operators is one way of cutting costs, then it can allow for more services with the same transport budget.

Building the NWRL with smaller tunnels will forever shut out the rest of the Cityrail network from using those tracks as well as a future Second Harbour Crossing.

This is unfortunate. It does not even seem that the savings from smaller tunnels will result in a significant cost saving either. In fact, the cost savings will be less than the additional costs that will be incurred in converting the Chatswood to Epping Line to be metro compatible. The separation of NWRL services itself will provide an added benefit of improved reliability through additional sectorisation, the line does not have to be built so as to permanently shut out all double deck trains.

Monday’s breakdown on the North Shore Line disrupted some 50,000 passengers going to work that morning. Trains did not run between North Sydney and Chatswood until 10:30AM, and there were no trains between Wynyard and North Sydney until 8AM. Buses lack the capacity of heavy rail at the best of time, let alone when a section of the rail system shuts down unexpectadly, and buses struggled to deal with the huge numbers of passengers getting off trains.

However, the disruptions were not limited to the North Shore. Trains on the Western and Northern Lines that normally pass through the North Shore Line saw delays throughout the morning. And it was these delays that saw flow-on effects to other parts of the network, to the South, Inner West, Bankstown, Airport/East Hills, and Newcastle/Central Coast Lines. This is because trains on these lines run on the same tracks at various points in the network. The Illawarra/Eastern Suburbs Line, which operates on completely different track to the previously mentioned lines, was not affected.

This separation of lines, operating on separate tracks to prevent delays on one line from spilling over to another line, is known as sectorisation. Cityrail has 3 sectors: Sector One (the Eastern Suburbs/Illawarra Line), Sector Two (all lines using the City Circle), and Sector Three (all lines using the Harbour Bridge). The problem is that trains on sectors two and three share track between Granville and Strathfield. This is despite there being 2 pairs of track along this section of the network, meaning that sector two trains and sector three trains could easily run on completely separate tracks. Word on the street is that the timetable update later this year will do exactly that (more on this next week).

Trains from different lines use the same track between Granville and Strathfield, meaning that delays on one line lead to delays on other lines. Click on image for higher resolution. (Source: Cityrail)

Trains from different lines use the same track between Granville and Strathfield, meaning that delays on one line lead to delays on other lines. (Source: Cityrail)

Doing so would not have prevented the faulty wiring that shut down the North Shore Line and led to delays on Monday. But it would have quarantined these delays, preventing them from flowing on to lines that run into the City Circle.

Decades ago, Sydney was a working harbour, but this began to end when Port Botany was opened in 1979, which has led to massive amounts of urban renewal in and around the city center. On the border of the CBD, Darling Harbour was converted into an entertainment, convention, and cultural center, while Barangaroo is seeing office towers being constructed that will soon be followed by a hotel/casino. The neighbouring suburbs of Pyrmont and Ultimo, which were industrial areas until the 1970s based on their proximity to a working port, no longer had reason for being and became filled with up-market apartments , a casino, and a cluster of media businesses. Across the Anzac Bridge is White Bay and Glebe Island, which will soon be converted into an exhibition and entertainment center that will host events while the Convention Center in Darling Harbour is renovated. The result is that huge sections of inner Sydney have been converted from industrial to commercial use, with a large amount of cultural and entertainment areas.

Click on image for higher resolution (Source: Google Maps)

Click on image for higher resolution (Source: Google Maps)

Meanwhile, Sydney continues to see more and more cruise ships arriving into Sydney Harbour, with 36 ships set to arrive this February, and cruise ship traffic will only increase. Originally these cruise ships could dock only at the Overseas Passenger Terminal at Circular Quay, but recently a temporary terminal was setup at Barangaroo.

However, the long term plan is to build a new passenger terminal at White Bay and close down the Barangaroo teminal as part of the site’s redevelopment. This new location has its problems, primarily a lack of good transport and its distance from the CBD, two things that the Barangaroo site has. A new terminal at White Bay should include a ferry wharf at the very least, and potentially a higher capacity transport link such as light rail in order to deal with the high volumes of people moving to and from both the passenger terminal and convention center. Yet the government has made to commitment to either.

One disadvantage of both sites is the height limitation of having to pass under the Harbour Bridge, which would prevent 2 large cruise ships from berthing in Sydney simultaneously. This height restriction led the federal government to allow limited use of the Australian Navy’s Garden Island facility for passenger cruise ships to dock. This location has both a ferry wharf and train station nearby, but is also close enough to the CBD  to reach it on foot. However, the Navy will not be moving, so this location will provide only limited relief, albeit at the most critical of times.

Sydney’s past as an industrial port means it actually has many options for catering to the growth of the cruise ship business, but this also poses the risk that those in charge may make the wrong call, by choosing a poor location and then not providing the essential infrastructure for it to work. Sydney’s future as an international city depends on getting decisions like these right, and at the moment there’s a question mark over how the government will implement the White Bay passenger terminal.

Emergency call points on trains

Posted: February 4, 2013 in Transport
Emergency call point on a Waratah train (Source: Author)

Emergency call point on a Waratah train (Source: Author)

The Telegraph reports that less than half of all trains have emergency call points installed and that retrofitting these on older trains would be prohibitively expensive. This is especially the case with the non air-conditioned S-Sets, which are due to be phased out by Waratah trains in the next 2 years. Here the Telegraph describes the Waratah trains as “troubled”, and tries to paint them as unsafe, pointing out that despite being installed with emergency call points, there have still been:

“189 incidents on trains, including 23 sexual offences and 64 assaults – reported to police in the first three months of 2012 alone”Melissa Matheson, Daily Telegraph (3 February 2013)

This is trying to have it both ways – either emergency call points improve safety and should be rolled out, or they do not prevent incidents from occurring. Ignoring that fact, the “189 incidents” figure is more sensationalism than a serious issue, and something the media is known to do.

These are the numbers. By the end of March of 2012, there were 7 Waratah trains in operation. Some of these came online during the previous 3 months, so there was probably about 5 Waratah trains averaged out over the first 3 months of 2012. Cityrail has roughly 180 trains, so the 5 Waratah trains represent about 2.8% of total rolling stock. There were 74.7 million trips made in the first 3 months of 2012, which means about 2.1 million trips were made on Waratah trains during that period, assuming an even spread of passengers on all trains. If there were 189 incidents on Waratah trains, then the probability of an incident occurring to someone on a trip is 0.0091%. In other words, 99.9909% of all trips had no reported incidents, which suggests that Waratah trains are incredibly safe.

But this raises the problem of the older C-Set and K-Set trains (the old silver models which are air-conditioned) that are likely to be retained even after all the Waratah trains are rolled out. Assuming that retrofitting emergency call points on these is prohibitively expensive (which would seem reasonable), then why can’t stickers be put on the inside of their carriages with the phone number that passengers can call to report incidents? The trains with emergency call points already have similar stickers with information on getting information on delays (“call 131500 and push 2”) placed all over their carriages, and given that most people have mobile phones these days, why can’t this be used as a substitute for emergency call points? It’s true that some people may not have mobile phone access (no phone, no signal, no credit, etc). But the emergency call points can also break down, so it’s not like they are a fail safe measure either.

If ensuring access is the goal, then this is surely a cheap and easy way to ensure it.

Information stickers (Source: Author)

Information stickers on a Waratah train (Source: Author)