Wellington Scoop

Barriers to light rail – they’re imaginary or mythical

by Kerry Wood
My two earlier articles showed that light rail is potentially the best value for public transport in Wellington, and that it must be well-integrated with bus services. This article shows why the barriers to light rail are much lower than many people believe (Global Research survey, for LGWM, 2018).

Perhaps the most important perceived barrier is damage in a large earthquake. However, the seismic resilience of steel rails set in reinforced concrete slabs is hardly worse than roads, and perhaps better. Christchurch’s heritage tramway — with modern tracks similar to light rail — was hardly damaged in 2011.

After a large earthquake, the only immediately available surface transport modes are walking and cycling; bikes can be carried over slips and rubble. The real public transport question is recovery time. Many light rail systems can be found around the Pacific ‘ring of fire,’ in Japan, Korea and the United States; design codes are available. The Christchurch tramway was quickly restored when tourism became a priority, and rapid-transit could expect to be much further up the priority list.

Other disruptions are breakdowns, crashes, derailments or power-failures, all needing detailed recovery plans:

. Planning for breakdowns (including power cuts) includes using two substations on the same site, each supplied from either of two high-voltage circuits. A crippled vehicle can be towed by a converted road-vehicle, or pushed by a following light rail vehicle (there is a safe procedure). Daily computer-diagnostic checks boost reliability, and can also be useful for a breakdown. Derailments need special equipment, quickly available.

. Most light rail crashes are caused by motor vehicles obstructing the tracks: 95% of crashes in one UK survey. Total casualties on six UK systems in 2002–3 were three deaths (possibly including suicides) and seven minor injuries.

. Obstructions can be managed by single-line working, using permanent cross-over tracks designed to handle unusual situations. Temporary light rail tracks laid on the road surface are a long-term option.

In many cases the best first-response to a serious incident will be a prearranged plan to pull buses off other routes, maintaining a reduced service everywhere.

Another disruption needing management is the light rail construction phase, especially in the central area. In Wellington this is relatively easy. Parts of the golden mile are two-lane and — short of widening — cannot carry both buses and light rail. The simplest solution is light rail on the waterfront and buses on the golden mile.

Most other light rail problems are mythical:


Cities using modern light rail and having a smaller population than Wellington (WCC area, 211,000) include Besançon (Fr, 135,000), Frankfurt am Oder (D, 58,000) and Valenciennes (Fr, 44,000).

Population density

Light rail viability depends not on population density but on ridership, and costs break even with buses at about 3000 peak-hour passengers an hour. At 6000 passengers an hour, central Wellington can already justify light rail. Light rail is competitive because of low operating costs, dominated by the cost of employing drivers. A light rail vehicle 66 m long (the same as proposed for Auckland) can carry as many passengers as six or eight buses.

In Wellington there are multiple sources of all-day traffic, as well as commuters. All-day feeder bus routes would deliver passengers to light rail at Miramar, Kilbirnie, Wellington Hospital and the Railway Station. Most present-day bus passengers would save time by connecting to or from light rail at these hubs. Present-day peak-only services would become feeder routes. More all-day ridership would come from passengers, visitors and shift-workers going to and from the airport and hospital, as well as shop customers and staff in Miramar, Kilbirnie, Newtown and on the golden mile.

Inflexible on fixed routes

Inflexible public transport is an advantage:
. A commitment to service. A light rail stop will obviously be around for a long time, boosting land value and facilitating medium density development.
. Costly hubs, a necessary part of rapid transit, have been properly thought through, located and modelled to maximise ridership. The best places for hubs are the busiest destinations.

Buses are too flexible, allowing planners and engineers to close routes for a parade, or design compromised bus lanes. A common sight in Wellington is buses running with two wheels in a bus lane and two in a traffic lane, or entirely in the traffic lane because the bus lane is blocked.

Poor resilience

The mythical part of the resilience problem is disruptions caused by parades, street parties and demonstrations. All were raised by the PTSS Option Hearings Subcommittee and all are easily managed using three measures:

. Keep light rail off the parade route, or relocate the parade route.
. Relocate regular events such as street parties (or perhaps the light rail route).
. Schedule other events at times when ridership is low enough for buses to substitute for light rail.

Assets stranded by new technology

Autonomous electric buses would not affect light rail viability because the same technology suits either mode. The light rail cost advantage would be lost but that is barely relevant in Wellington, because buses need so much more space.

There is no possibility of autonomous cars radically increasing road capacity; cars are inherently inefficient users of road, space, whether moving or parked. Walking and cycling both have up to ten times the person-capacity of human-driven cars, on the same lane-width.

Real-world considerations ensure that autonomous cars cannot close the gap:

. Autonomous cars must usually slow or stop at junctions. Traffic signals are common and many junctions have three or four legs but only one main commuter route.
. Gaps between cars must ensure that a queue can stop safely. Autonomous vehicles should be able work with shorter gaps than the two seconds recommended for human drivers, but only under special and very costly conditions:

. No human drivers in the queue
. Separate routes for pedestrians and cyclists
. Grade-separation at all crossings.

Safety requires scope for closely-grouped autonomous vehicles to make an emergency stop at any time. In an emergency, autonomous control must limit queue deceleration to something manageable by trucks and buses.

Who would want to live on streets meeting these requirements?


  1. Neil Douglas, 11. April 2018, 13:46

    Kerry, informative always but, lest we forget, your UK accident stats fail to mention the Croydon Sandilands 2016 derailment which killed 7 and injured 51. This was the accident in the UK with fatalities since 1959, so the two fatalities you mention for 2002/3 must be something other than tram accidents. (I note the cost graph in your first article originated from Croydon LRT). The Croydon tram came off the rails at 0615 on Nov 9th 2016 at a corner in a cutting (during heavy rain and in the dark) near the Sandilands tram stop. The tram was going at 73kph when the speed limit was 20kph. The accident was the deadliest in the UK since 1917 when a Dover tram killed 11.

    This link gives more info on the Croydon accident:

  2. Roy Kutel, 11. April 2018, 14:11

    Hey Kerry, Melbourne provides some tram accident stats with 15 deaths and over 1,700 injuries that required hospital care between 2000 and 2008. 2013 was a bad year in Melbourne too, with two pedestrians killed after being hit by trams and a car driver killed in a collision. There were also 30 serious injuries.

    And trams are not universally good for cyclists either. Two months ago a Northcote cyclist was seriously injured after getting hit by a tram, with the police warning people about the inability of trams to diverge from tracks.

  3. Kerry, 11. April 2018, 15:40

    Neil. Well yes, there are outliers, such as Croydon: a driver forgot to slow down… The figures I used are from:
    (Steer Davies Gleave, 2005, prepared for the Passenger Transport Executive Group, Table 7.2)

    SDG says ‘such as Croydon’ and Transport for London were interested in light rail for the proposed Uxbridge and Cross-River, so I expect the assumed capital costs given by TfL are reasonable. After all, the 3000 passengers an hour break-even figure is hardly new: it was given in the 1992 Superlink study for Wellington.

    Roy. Thanks for the Melbourne view. I can guess at a range of explanations, but I can only be sure that light rail is safer than buses and far safer than cars. Even in Melbourne, the annual death rate is about two a year, for a very large system using some pretty ancient trams: this isn’t light rail. Older vehicles will tend to be less safe, just like cars. The high proportion of passengers injured in a tram is not a surprise either. Part of the trouble is emergency stops by normally smooth-running vehicles, catching standing passengers unawares.

    Bicycles and light rail are usually only a problem when the bicycle tyres get caught in the grooved rail.

  4. Neil Douglas, 11. April 2018, 17:55

    Kerry, the 2004 TfL cost graph is old and needs updating. It’s also whole of life including construction as well as operating costs. Croydon has 10 street kms out of a total of 28 kms, the rest being old rail line. Croydon’s capital cost works out at NZ$ 50m per km (today’s dollars). Unfortunately, the current Sydney experience is five times higher at $250m per km which would shift the LRT cost curve way skyward. Clearly, Wellington would need to work to a lower engineering standard (maybe Melbourne?).

  5. John Rankin, 11. April 2018, 18:44

    Neil, it would also be useful to know what operating speed the graph uses, since this affects the operating costs. If it uses the UK average of 30 km/hr and Wellington chooses a central city route that only lets us achieve 20 km/hr (say), that will also move the LRT cost curve in a skyward direction.

    Yes, if we pick a waterfront route we’ll lose riders from the outer edges of a 500 metre walking radius, but on the other hand, we may buy so much travel time and frequency that our ridership goes up. It all depends on what we are trying to achieve. Even with a waterfront route, I think it’s unlikely we can get a 30 km/hr average speed for a route through Newtown past the Zoo.

    Do we know what engineering standard Auckland is planning to use for its LRT lines? Is there a reason Wellington would want to use a different standard?

  6. Ross Clark, 11. April 2018, 23:58

    “Gentlemen, we haven’t got any money, so we’re going to have to think” (Ernest Rutherford, attrib.)

    The money issues I and others continue to raise are not a myth. The Edinburgh light rail scheme was initially meant to cost £550m for 11 miles of route. We got 8 miles of route, for about £780m – the city having foolishly signed up to a deal in which they bore all the liability for the cost overruns (well over £200m, in the end).

    Separately, I would like to suggest that connecting the region to the CBD requires a separate solution to connecting the city to the CBD. There is not much in the way of city-to-region traffic, certainly not for work journeys. So a connection between the railway station and about Courtenay Place could well be kept separate from whatever we do up the Golden Mile.

  7. Chris Calvi-Freeman, 12. April 2018, 8:25

    Just finished compiling my PowerPoint presentation on a possible LRT system for Wellington. Presenting to fellow WCC councillors and invited GWRC councillors today. [via twitter]

  8. Helen B, 12. April 2018, 9:01

    Ross you make a very good point but you won’t find a sane private company that will take on all the risk because of the unknowns of the underground utilities.

    Chris it is nice to hear you are spending time developing your ideas but surely it is the job of the paid council officers to develop the plan? I thought this is what LGWM was doing. Hopefully your plan won’t create more confusion.

  9. luke, 12. April 2018, 9:46

    trams cannot swerve to avoid somebody and inadvertantly hit somebody else instead, like a bus can.

  10. Kerry, 12. April 2018, 10:20

    Luke. Light rail tracks are no wider than bus lanes, making them no more difficult to cross, and the ‘swept path’ is marked on the road. Stand, walk, cycle or sit outside the marks and you are safe. Some European cafes have outside seating so close that a customer could reach out and touch a passing vehicle, but I wouldn’t accept that beside a bus lane.

  11. Roy Kutel, 12. April 2018, 10:56

    Luke & Kerry: this Melbourne study shows that increasing the distance between stops reduces accidents since trams don’t have to brake and accelerate so much and the chance of collisions between trams and other road users is lowered.

  12. John Rankin, 12. April 2018, 12:18

    @HelenB: construction cost risk is only part of the risk profile. We also need to consider ridership risk. Should we expect a commercial system operator to assume at least part of the ridership risk? If so, then one consequence is that the operator may reasonably expect to have final say on the route and stop locations, which will affect construction cost.

    LGWM has been much exercised about whether there will be enough demand along the entire light rail route to justify the level of investment needed. Many of the comments on Kerry’s articles have been about how best to deliver the patronage that light rail needs to be economically viable. These comments have not converged on a consensus view.

    I would find it somewhat reassuring if a commercial system operator would take on a share of both the construction and ridership risks. This would create strong incentives and disciplines around system design decisions. Is this too much to expect?

  13. Ross Clark, 12. April 2018, 22:17

    Helen B: You won’t find a sane private company that will take on all the risk because of the unknowns of the underground utilities.

    This was precisely the problem in Edinburgh. The contract was let on the basis that shifting utilities would take a certain length of time and cost a certain amount of money. But there proved to be far more under the asphalt than anyone had realised, so the work took twice as long as expected and cost twice as much. As a result this threw the whole construction programme out of kilter, leading to vastly increased costs on building the part of the network which involves street running. The risk from cost over-runs in projects of this nature is quite substantial, and should makes one very reluctant to take on board estimates of both cost and time, unless very substantial contingencies are built into both. Who bears the risk? is a very important question.

  14. Kerry, 13. April 2018, 10:06

    Ross. Good point but, again, I think it is a manageable problem.

    First, have the route surveyed using ground penetrating radar, using the best gear available, early enough to inform decision-making. Doing it too late in the project is a bigger risk than doing it too early, especially in known difficult areas such as the Basin Reserve and Willis St. At a guess, Sydney would not be going to court if that had been done. Up-to-date radar should be able to spot everything that matters, in three dimensions. Only 20 years ago, cables were sometimes laid just above plastic pipes, in the hope that putting a signal into the cable would allow location of the pipe. Once lost (no drawings) finding unknown drains was very difficult. There was no way of tracing them until somebody came across an unknown manhole, or an unknown connection. I once found a 900 mm drain that way.

    Second, check the grade of drains that have to be lowered, to see how much will have to be relaid. Drains are often the biggest problem, because they have to be laid to a fall (pumping may create more problems than it solves). If a drain is lowered at one point, it has to be relaid downstream, at a flatter gradient,until it is back at the level of the old pipe. It sounds as if Sydney has hit this problem.

    And third, keep the bottom of the track formation as high as reasonably practical. In researching this I came across light rail contract drawings specifying that the top of all services be at least 900 mm below the road surface, and many Wellington drains are higher than that. Two methods are:
    — Keep the concrete slab as thin as reasonably possible.
    — Think about raising the road surface in places, or lay the tracks above road level. This is a common practice anyway (by 150 mm in the UK design code) so that motor vehicles are discouraged but can still cross the tracks at low speed.

  15. Alastair M, 13. April 2018, 13:02

    Does any reader recall the extensive and protracted road works that were put into the stretch of Manners Street when it was returned to bus-only road use (from its previous Manners Mall pedestrian-only use)?
    There was considerable excavation and then a sub-surface reinforced concrete “road” laid below the current road surface. The justification that I recall for all that engineering was to future proof that section of the Golden Mile for light rail.
    I assume that could be seen as a recent working example of the scale of road work that would be required for light rail through the city.
    In the Manners Mall case there was no traffic disruption, as the area was pedestrian only at the time the work commenced.

  16. Patrick Morgan, 16. April 2018, 22:26

    I’m excited about light rail for Wellington, says Phil Twyford at the Transport Summit So, not just more buses. Light rail is on the table. [via twitter]

  17. Jonny Utzone, 17. April 2018, 10:02

    Err Patrick, Light Rail has been ‘on the table’ since 1992 (at least). Show me the money!

  18. Kerry, 17. April 2018, 14:21

    Jonny. This is probably the first Minister who has known about any light rail proposal for Wellington. Certainly Phil Twyford is the first who has expressed interest, at a national ‘summit’ meeting on the new draft Government Policy Statement for Transport.
    This is the beginning of transport policy in New Zealand entering the twenty-first century, recognising the futility of ever-greater road-building, the necessity of managing carbon emissions and the need to design cities for people, not cars.
    Light rail was officially rejected in the notoriously bad but still official PTSS study in 2013. Now it is far more certain than it has ever been.

  19. Neil Douglas, 18. April 2018, 11:18

    @Kerry, Michael Cullen back in 2006 mentioned extending the J’ville line to Courtenay Place in advocating the line’s retention in this letter to WCC Mayor Prendergast and GWRC Buchanan about the idea of spending $70 million to convert it to a Busway.

  20. Kerry, 18. April 2018, 19:14

    Neil. Oops, missed that one, but it doesn’t have as much push behind it as we saw on Monday. And with no plans for the Johnsonville Line for a while, KiwiRail can concentrate more important stuff.


Write a comment: