Wellington Scoop

Light rail – on the wrong track?

by Brent Efford
Whether 2019 brings any real progress towards light rail and the essential ultimate achievement of a complete rail transit spine serving all of greater Wellington remains very much up in the air, with signs that the Let’s Get Wellington Moving project may be headed down the wrong rail transit track.

Its study outcome is now expected to report in April. What is of concern is that LGWM has apparently fallen for the blandishments of anti-tram-train advocates FIT and may be recommending only a local streetcar version of light rail, operating on a different track gauge and only south of Wellington Railway Station. Not even to Johnsonville.

If true, that’s goodbye to dreams of a regional system which would attract commuters off the state highways, and goodbye to future generations having the choice of an automated rail ‘direct through service’ to central Lower Hutt, Whitby, Wainuiomata or wherever.

This result was indicated in the Stuff/Dominion Post article What is light rail, and how would it work in Wellington? of 12 January. In response, Demetrius Christoforou of Trams-Action summed up the problems with the ‘local tram’ version of light rail in this letter to the Dominion Post (published in abridged form):

While it is encouraging that light rail is at last starting to gain traction among decision makers, I fear that the proposal put forward by FIT is defective for the following reasons:
1. It takes no account of the existing suburban rail system. Having two stub terminals at the busiest point is the worst possible configuration and will do nothing to compete with Transmission Gully and encourage those in the wider region to shift to public transport.
2. Insistence on two separate systems means they have to find an expensive new depot instead of using the existing Matangi facility.
3. Light rail benefits businesses along its route, but the FIT route misses the Golden Mile and Courtenay Place, the areas of highest demand.
4. Their route is highly convoluted and includes a totally unnecessary tunnel from Taranaki St to Adelaide Road: higher cost, slower service.
5. Going to Miramar first and then the airport is the height of folly. The natural route is along Coutts St and under the airport runway. Airport passengers will not appreciate being taken on a tour of Miramar. Such a counter-productive route will doom light rail to failure.

Another correspondent also mentioned the depot problem, and I followed up with this:

Demetrius Christoforou and Peter D Graham have hit the nail on the head regarding the crippling defects in the FIT version of light rail, described to me by one regional councillor as “loopy”.
The alarming thing is that the gauge-incompatible, limited, inconvenient and more expensive version of ‘light rail’ advanced by FIT has, I am told, been adopted by the Let’s Get Wellington Moving study as its proposal. This defies the 140 year-old legacy of local tram-train studies, copious examples of world best practice and expert overseas advice which is readily available, and which I have already supplied to the study as PDF files and web links. Not to mention plain commonsense.
Is LGWM trying to set ‘light rail’ as a concept up to fail?

I am currently preparing a presentation to the NZ Chapter of the Railway Technical Society of Australasia about light rail and tram-train, entitled ‘Direct through service – tram-train for a complete rail system’. ‘Direct through service’ was the translation of the description of the Japanese tram-train Echizen-Fukui railway where average weekday ridership rose from 3,800 to 10,900 daily. Wouldn’t that be nice for Wellington.

Or we could look at the Karlsruhe model, where a region almost identical in population (440,000) to greater Wellington has 12 times the rail ridership. Demetrius Christoforou reported last year:

I have recently visited Karlsruhe, a city of similar size to Wellington, where they are celebrating 25 years of tram-train service. I have seen it in action first hand and can attest to the convenience of riding the rails from the countryside all the way into the centre of town. Visible results are the absence of a mad peak hour rush at their main railway station and a lack of traffic congestion.

Many Wellington observers mistake our “mad peak hour rush” where 85% of the region’s total rail passengers get on or off their trains as evidence of huge usage that mere light rail couldn’t possibly cope with. The numbers tell a different story: Wellington’s total rail ridership (all four lines) of about 45,000/weekday is matched or exceeded by each of three Melbourne tram lines, and by many street-based lines in cities like Prague, Budapest, Vienna, etc.

Two of those Melbourne lines, 86 and 96, are routed via Bourke Street, which has shorter trams and more impediments to tram movement (mainly many no-tram-priority traffic signals) than a properly-engineered Golden Mile light rail line would have. The busiest stop is in the Bourke Street Mall. No “mad peak hour rush” even there!

Meantime, our (diesel) bus system continues to breed dissatisfaction and complaint, with the promised ‘100% electric’ system looking a very long way off indeed.

What with the bustastrophe forgetting its earlier aim of ‘rail penetration of the CBD’ and a regional light rail (tram-train) system, the Regional Council has destroyed any claim to be a competent planner and manager of public transport. Desire for an independent regional transit authority is now often heard in discussions.

The local body elections due in October will be interesting – will the few progressive Wellington City regional councillors who actually tried to save the trolleybuses and prevent the bustastrophe nevertheless get the electoral blame? Will the majority of councillors from the outlying cities, focussed only on ‘four lanes to the planes’ and who really deserve the blame, still romp in just on ’name recognition’?

Brent Efford is the New Zealand agent of the Light Rail Transit Association.


  1. Michael Gibson, 24. January 2019, 14:52

    Can Brent stand in October please?
    I agree entirely that “the majority of councillors from the outlying cities” deserve blame, especially since their main aim has been to reduce their own rates.
    However most of Wellington City’s representatives have been pathetic in not standing up to these out-of-towners. At the very least they should now begin announcing their intention not to stand again and encourage qualified candidates to replace them as our representatives.

  2. Glen Smith, 24. January 2019, 19:01

    It is good to see discussion around the specifics of across town transport design. This is something that should be led by our planners presenting a range of costed options for the public to consider (rather that in secret as at present). Brent Efford’s and FIT’s proposals are only 2 of a range of design/ route options. The best approach in my view is to break possible options into components. Briefly.

    1. Rubber vs rail. There is in my view compelling evidence to add across town rail including capacity, long term cost and attractiveness (with empirical data showing rail outperforms bus in essentially all outcomes). However the most compelling is that we have an existing extensive rail network and the aim should be seamless across town lines for commuters from the north.
    2. Transfer at the Station (ie separate or seamless networks). I absolutely agree with Brent we should be aiming for seamless corridors rather than FIT’s separate light rail proposal. There are two possible options A. Running heavy rail (eg Matangis)- this would require expensive over-engineered corridors as per Neil Douglas’s article and would limit future rail expansion B. Track sharing of ‘heavier’ and ‘lighter’ rail units (effectively tram-trains) as has been successfully introduced overseas but hasn’t been seriously investigated here despite its logic.
    3. Corridor across the main CBD. There are only two real options – the Golden Mile and the Quays. There is, in my view, compelling evidence for FIT’s proposal to add a Quays corridor including long term capacity (a single corridor will be inadequate), difficulty of construction (and hence cost) and disruption during construction. However the two most compelling are A. quality of corridor -creating a high quality rapid ‘bypass’ of the CBD for across town commuters and B. Maintaining a high capacity bus corridor for across town bus lines – a Golden Mile route displaces bus lines forcing unnecessary transfer penalties onto a high number of bus commuters.
    3. Route south of Courtenay Place. There are two real options – Taranaki St and Kent/ Cambridge Terraces (whichever is chosen for rail the other can be serviced by buses). This depends on options for heading East from Taranaki St. A Taranaki St route (as per FIT) runs more centrally but requires an expensive tunnel at the southern end (before Arras Tunnel construction I sent the NZTA concept plans for a parallel rail tunnel to minimise costs but this was ignored). A Courtenay Pl route (as per Brent) displaces Golden Mile buses again. The other option (in my view the best) is Wakefield St onto Kent Tce. This is technically easier and picks up some key destinations including Conference Centre/ Te Papa / eastern Courtenay Pl and Embassy/ Basin Reserve.
    4. Route across Mt Victoria. Both Brent and FIT propose a route via Newtown and then a separate rail tunnel. The logic of this defies me compared to a direct route integrated into the NZTA plan for a second Mt Victoria tunnel/ Ruahine St/ Wellington Rd changes (that will proceed anyway) which picks up the Colleges/ Hataitai and Hataitai park/ Kilbirnie Park and northern Kilbirnie. A Newtown route is longer, slower, technically difficult, almost certainly more expensive, forces all commuters from the East to unnecessarily go via congested Newtown and displaces car capacity and parking (making it politically difficult). The justification is it picks up Newtown commuters including the Hospital and Zoo. However passenger data shows few originate in Newtown, almost all being ‘through’ commuters from the East (which would be serviced by a direct route) and South (which will continue to be serviced by bus including a frequent Island Bay service). Hospital patients are most likely from Wellington suburbs and hence bus commuters. The only reason for rail via Newtown is if the plan is to unnecessarily force a transfer penalty on large numbers of bus commuters. Newtown should be serviced by north-south and east-west bus routes.
    5. Route East from Kilbirne. The only two real options are Coutts St and Rongotai Road. A dedicated Coutts St rail corridor would require removal of parking from both sides of the road, a new tunnel under the airport and ends up a long way from Miramar. Rongotai Road has ample width, starts at the main Kilbirnie bus hub, picks up the Indoor Stadium and leads to the northern end of the airport nearer Miramar.
    6. Route past the airport. This either has to go via a new tunnel under the airport or around the northern end of the runway. A route around the northern end either has to go via a tunnel under the high ground at the northern end of Calabar Road or take a difficult circuitous route via Miramar (as per FIT) which I agree with Brent will inhibit airport commuters. The best option, in my view, is a combined rail, bus, cycle, pedestrian tunnel linking Caledonia Road to Rongotai Road via Cairns St. This solves the across airport difficulties of 4 transport modes in one structure while enabling a future full Miramar rail line to pair with either Lower Hutt or Johnsonville.

    As can be seen there is a wide range of options. The NZTA, GWRC, WCC and LGWM should be exploring them and presenting them to the public. Don’t hold your breath on that one.

  3. luke, 24. January 2019, 20:49

    my route thru the cbd would be along the quays, chaffers st (take out some of the New World carpark), blair st, cambridge/kent, cutncover under the basin, adelaide rd.

  4. John Q., 24. January 2019, 22:52

    same as me luke.

  5. Mike Mellor, 24. January 2019, 23:04

    That LGWM may be heading down the rail track is excellent news, cause for much celebration – assuming, of course, that it’s confirmed. The key point is that the principle of rail through the city may well be accepted. What type of rail or the precise route to be taken then become a details – albeit very important ones – to be worked through when designing the system.

    The two main options are the comprehensive tram-train model, with the potential for through service to and from the existing KiwiRail network; and simpler conventional light rail running just within the city. Both have their advantages and disadvantages, but either, following whichever route it is that evolves, would be a significant improvement on what we have now, and well worth having.

    Blaming regional councillors will get us nowhere – they are there to represent their constituents’ interests. It’s a fact of life that councillors from Wellington city are in a minority on GWRC, so the other councillors needed to be persuaded of the benefits, and that will be difficult unless advocates achieve common ground and make compromises. The positives of rail need to be explained, emphasised and developed, rather than focusing on the perceived negatives of the opinions of others – such public bickering will surely doom light rail to failure.

    But if there is good news and light rail does come to Wellington, GWRC may be no more than one partner in the enterprise. In Auckland NZTA is leading the light rail programme, so that could well happen here, too. An alternative way could be to follow the Waikato model, where a joint committee of all the regional, city and district councils, with their votes appropriately weighted, governs public transport in a spirit of co-operation.

    Councils working together and rail advocates working together – wouldn’t that be good?

  6. Geoff Cameron, 24. January 2019, 23:25

    Lite rail is only ever going to be city only. Need heavy rail imo to encompass greater wgtn. Until then white elephant.

  7. Ralf, 25. January 2019, 8:09

    My expectations are pretty low. One of the reasons I think the LGWM has been delayed is that the main intention is to unlock government money for road tunnels. LR is in there only as a sweetener for the government which seems keen to provide money for sexy public transport. Once the government money is unlocked, road building will proceed and public transport improvements will be delayed until they will be cancelled (as being too expensive, plus there will be a government change at some point, which will reduce the contribution for rail). So, I agree with the article’s headline.
    What I would like to see would be a robust discussion around options, cost and long term planning, as Glen Smith is proposing. If that would be open, my confidence that there is a serious interest in rail in the city would increase.
    Brent’s suggestion to use Lambton for rail is in my opinion not a good idea. We should have separate corridors for rail and bus (with connection points obviously) and rail should have a much bigger spacing between stations than buses. So to me it seems that it makes sense to keep buses on Lambton and find another corridor for rail.
    Implementing something like the https://en.wikipedia.org/wiki/Karlsruhe_model could increase support from the outer suburbs.

  8. Kerry, 25. January 2019, 8:34

    FIT is not anti-tram-train and would support a good case, but has never seen one.
    FIT sees a light rail extension to Johnsonville as good sense, but with difficult barriers:
    — Standard light rail vehicles, with the dimensions proposed by Brent, AT (Auckland) and FIT, will not fit through the Johnsonville (or Kapiti) railway tunnels.
    — Tram-trains cannot be longer than the 75 metre limit set by German regulations based on experience. Running short trains on the Johnsonville Line would halve the capacity, but the Matangis could double it, with 8-car trains.
    — The Johnsonville Line capacity is four trains an hour, each way. Any increase would require double-tracking, at a guesstimate $1bn.
    Extending tram-train operations beyond the Johnsonville Line would be far more difficult: other lines have freight trains. There are light rail studies for Wellington, but not tram-train studies. There has never been a case for tram-trains.

  9. John Rankin, 25. January 2019, 12:09

    Noting the comments from @GlenSmith, @MikeMellor and @Kerry, I suggest that Wellington ought to evaluate at least 2 light rail rapid transit options:

    1. an on-street mostly at-grade option using low-floor vehicles, Auckland standards and a quick and easy interchange at the railway station, with a provision to upgrade the Johnsonville line in future, to allow for low floor vehicles and a more frequent service (FIT’s proposal, with or without modifications)

    2. a metro-style, dedicated corridor off-street option using high-floor vehicles and Kiwirail gauge, grade-separated as necessary, plus a proposal for how through-running with main line services would be accomplished (some have suggested this option, but it has not received the attention or detailed study it deserves)

    Option 1 would be cheaper, but closes off the option of future integration with existing Hutt Valley and Kapiti Coast services. Option 2 would initially be more expensive, but in the long term may be better value. When I was in Vancouver before Christmas, the municipality of Surrey was having exactly this conversation. They had originally chosen their version of option 1, but now many are arguing for option 2.

    As I understand it, @GlenSmith’s proposal is a more developed version of option 2 and I would like to see a detailed proposal, with maps.

    For a tram-train option to be considered, it would need to demonstrate how it proposes to deal with the low floor problem, the maximum length problem, the maximum frequency problem, and the safety-case problem. As @Kerry points out above, these appear to be show-stoppers.

    The best approach may be for LGWM to adopt a technology-neutral procurement strategy, setting out the requirements for a rapid transit service, including the long term vision of a regional rail system that includes Wellington City. I would expect to see proposals that include variations of options 1 and 2. I may of course be wrong; a well-crafted technology-neutral request for proposals or registration of interest would let the market decide.

    As @MikeMellor almost says, whatever option is chosen will have weaknesses as well as strengths and involve trade-offs to fit the constraints of the real world. Let’s not make the perfect the enemy of the good.

  10. Kerry, 25. January 2019, 15:11

    I wonder if there is a compromise here?
    Both of John’s route descriptions, both studied for both light rail and tram-train.
    I have been concerned that, in the era of very rapid change we are entering, light rail capacity might not be enough. So, for all four options, a good question would be: what is the capacity?

  11. John Rankin, 25. January 2019, 18:39

    @Kerry’s question about capacity is a good one. Option 1 described above has a maximum capacity of about 10,000 passengers per hour. A world-class system operator might squeeze an extra 10% or so out, with a lot of luck and a following wind, but you wouldn’t bet the farm on it.

    So a model of long-term demand on the corridor is essential. If we conclude that peak demand is likely to exceed 10,000 passengers per hour in the life of the system, option 1 would be eliminated. FIT has estimated a demand of about 7000 passengers per hour after 2050, so plenty of headroom, but it depends on the assumptions you make.

    Option 2 can support at least 20,000 passengers per hour, as it can run longer trains at higher frequency.

  12. steve doole, 26. January 2019, 10:34

    7,000 passengers per hour capacity won’t be enough. To attract more people from north of the city, particularly if the first part of their journeys are by car anyway, a similar solution to central Auckland or Sydney rail is needed, delivering large numbers of people close to their destination, without interchange. Light Metro, as in say Naples or Paris, could be a good fit.

  13. luke, 26. January 2019, 13:05

    tram trains offer less capacity than the heavy rail network we already have and transfers are how networks work. As long as they are frequent and not fiscally penalizing changing between light & heavy rail should not be an issue.

  14. Glen Smith, 26. January 2019, 18:35

    Mike. It is good to see rail being seriously considered but final acceptance depends absolutely on details such as cost, convenience, impact on drivers/ bus users etc. As Ralf says a robust discussion around available options is required.
    Kerry. Just because there are no good case studies for tracksharing in Wellington doesn’t mean one isn’t possible. I haven’t seen any detailed studies for Quays based rail (which was arbitrarily dismissed in the flawed Spine Study) but that doesn’t mean it isn’t the best option.
    John and Kerry. Tracksharing would involve overcoming a range of technical issues but has been achieved overseas and the prize, a seamless rather than fractured rail spine…forever…is a big one and the price of failure – the potent inhibitory effect of a Station transfer penalty…forever- a severe one. We should aim for the ideal and only accept less if the obstacles are insurmountable or overwhelming (which could be the case but unless investigated we won’t know. This is of course is our planners job but they seem incapable of doing this).

    Regarding some specific technical issues
    1. Platform height can be solved by have specific ‘stations’ with raised platforms at network height
    2. Dual voltage units solve problems of high voltages across town.
    3. Compatibility problems with tunnels etc should be overcome by specifically designed rather than ‘off the shelf’ rail unit
    4. Logistics (frequency, capacity, interaction with freight etc) are not my area of expertise but some observations.
    If across town station platform and unit length is a limiting factor then running all units as ‘through units’ is not likely sensible since a large number of people would exit at the station resulting in long, largely empty units running across town. The aim should be to ensure all units across town are full.
    Instead longer trains- up to 8 car units (currently Matangis) – could continue to run to the station to service the current demand to the northern CBD, while shorter (? up to 75m) units run across town (this would have to occur anyway until the Matangis see out their lifespan). The question is how you ‘sort’ people- commuters going to the station will just take the first unit that arrives. My suggestion would be to have across town units ‘shadowing’ Station units by ?2-3 minutes (whatever is the safe minimum) at peak times so Station commuters are depleted leaving only across town commuters for the across town units. While unorthodox, this concept of sorting isn’t unprecedented – ‘express’ buses from Island Bay leave 4 minutes after standard buses, minimising commuters going a short distance. The net effect would be the same as having an up to 12 car train with an 8 car front ‘Station’ section and a 4 car rear ‘across town’ section except the two sections would be physically separated by several minutes. From a timetable perspective, including interaction with freight trains and on the Johnsonville line, units several minutes apart would presumably behave as a single train over most of the network. The exception may be the ‘pinchpoint’ just north of the Station where line’s all meet and units can be up to every minute, a fact which Chris Calvin-Freeman previously gave for not considering seamless through lines. This area, in my view, needs to be rethought. This is only one possible solution to logistical problems. I am sure there are many others.

    5. Safety. These problems have been solved overseas. A review of tracksharing in Germany notes that “a substantial difference between the United States and Germany (and the EU to the extent they are standardizing vehicle safety regulations) is in the philosophy toward train collisions. U.S. FRA regulations presume a crash is likely to occur and demand vehicle strength to maintain car body integrity and maximize occupant survivability. The German
    philosophy presumes that accidents can be avoided and sets about to devise means of supporting that presumption”. Ie the underlying principle should be to put enough safety procedures in place to ensure trains never collide rather than over-engineering heavy units to survive collisions. In Wellington this would presumably require an updated signalling/ train control system but given the recent reports of multiple near misses this looks overdue anyway. Interestingly the same report notes that braking capability is a key factor (how fast a train can stop once the driver sees a collision is inevitable) and lighter units perform better. Light units with high braking capability and adequate ‘crumple zones front and rear seems to be the standard (not my area of expertise
    – a comment from the tram-train group about safety would be interesting).

  15. Mike Mellor, 27. January 2019, 10:06

    Glen, I’m not saying that we don’t need debate – of course we do. But that debate has to be positive and respectful, recognising that in the real world some compromise is inevitable. The “my way or the highway” approach, taking no prisoners, means that the highway, regrettably, becomes much more likely, and plays into the hands of those who do not see rail as part of the solution. Savvy campaigners and lobbyists are never slow to take advantage of a division in others’ ranks!

    We also need to be aware of the risks of over-specification at an early stage, focussing too much on details. There will be lots that will need to come out in the wash after a decision is made, so what I’m suggesting is that now the focus needs to be on making sure that rail (of whatever variety) is on the agenda – that is the absolute essential, without which all the rest is just daydreams. When that has happened, then we get down into the all-important nitty gritty. At least until that time comes, the old adage that unity is strength holds – and its converse is equally true.

    Your points 1-3 and 5 are all valid, and point 4 is interesting, but looks as if it will add quite a lot of operational complexity, something not to be considered lightly when reliability is essential. The thing to remember is that all of these will add risk and cost money (in terms of both capital and operating costs), probably a lot more in both cases than standard plain vanilla light rail. The question is, will the extra benefit (as yet unquantified) of the extras be worth it, or will the extra risk and cost mean that rail falls into the “unaffordable” trap?

    Another adage is that the best is the enemy of the good, and we ignore that at our peril.

  16. Neil Douglas, 27. January 2019, 10:34

    Perhaps our planners could take a trip to Sheffield or invite somebody over – as an ‘innovative first in the UK’ tram-train service between Rotherham and Sheffield has started (began October 2018).
    It’s not a high frequency service at 3 trains per hour. There is also a FAQ for people who want to know more.

  17. John Rankin, 27. January 2019, 12:21

    @GlenSmith and @MikeMellor: Yes, “we should aim for the ideal and only accept less if the obstacles are insurmountable or overwhelming.” While FIT found the obstacles to single-seat journeys for Hutt and Kapiti riders insurmountable and overwhelming, others may find a way. My intuition is it will need a segregated light rail line, rather than the on-street option that FIT and others before it have proposed. Meanwhile, FIT’s proposal is a bird in the hand, worth 2 in the bush, and a stage 2 can convert the Johnsonville line to light rail, when funding is available. Not the best, Mike, but arguably good enough, the Minimum Viable Product.

  18. Glen Smith, 27. January 2019, 13:05

    Mike. Absolutely agree. I said technical issues may be overwhelming (although I have yet to be presented with one) and I am perfectly prepared to be proven wrong in this or route issues. We won’t know unless we investigate options. A few comments.
    If we are expanding rail-based PT from the north, then we have to add more peak-time units irrespective of whether these terminate at the Station or run through, so logistical problems on the general network and the ‘pinchpoint’ area just north of the Station need to be solved either way.
    The current signalling/ train control situation, with a number of near misses, should be viewed as unacceptable and will get worse whether we add ‘Station’ units or ‘through’ units. A high-quality train control system should be seen as essential either way. Given a high-quality control system and fully dedicated corridors with full right of way, I can’t see why the ‘shadowing’ units, while an unconventional solution designed to address Wellington’s specific scenario, would present any major operational complexity compared to running units at other times and, since they would largely behave as a single unit in terms of overall timetabling, could make logistics easier.
    In terms of cost, the evidence of the potent inhibitory effect of Station transfer is compelling with only 15 percent of rail commuters transferring to buses despite having this option available – compared to over 50% of cars being ‘through’ traffic that continues on through the Terrace Tunnel. Will this change dramatically if it is light rail people transfer to rather than buses or is it the transfer process that is the potent inhibitor? If we fail to attract people out of their cars, what will be the additional congestion and other societal costs over the next 20 years? 50 years? 100 years?

  19. Mike Mellor, 27. January 2019, 17:09

    Glen: the complexity I’m referring to is linking street running in the city with the mainline rail network: each has its own characteristics and limitations, and this issue has never been looked at in any depth – for example, I don’t think it’s ever been discussed seriously with KiwiRail. For an extreme example of the potential difficulties, the Sheffield example quoted above by Neil Douglas above was three years late and five times over budget.

    It needs to be, and linking the two types of operation will create a pinchpoint at the junction (and there’s no way that units could be “every minute” there – no conventional railway anywhere in the world operates at that frequency), since not all mainline services will be able to continue into the city.

    And sorry, I don’t understand what you mean by “shadowing”, nor how two separate trains/trams can be treated as if they’re one: as far as I can see, they’re either one train/tram (if coupled together) or two (if they’re not).

    It’s not necessarily that technical difficulties will be overwhelming: it’s whether the costs and associated risks of overcoming those difficulties will exceed the benefits, and potentially undermine the whole project.

  20. John Rankin, 28. January 2019, 10:42

    @GlenSmith re point 4 above.

    I fear I am misunderstanding this point. Suppose a current Matangi service is every 15 minutes. As it reads to me, point 4 proposes replacing this with a 30 minute service on a longer train plus a shorter, through-running train 2-3 minutes later. This would be a roughly cost-neutral change.

    So everyone will see their average wait time double, in order that the few who are travelling beyond the railway station can enjoy a one-seat journey. In this example, what proportion of current rail users will switch to driving, because we have added 15 minutes on average to their daily commute (7.5 minutes each way)? Why is this better than changing at the railway station, which will require on average a 2.5 minute wait (for a 5 minute service)?

    Is my interpretation correct. If not, could we have a clarification, please?

    There is an adage which PT service designers do well to remember: one way or another, the Piper always gets paid.

  21. Glen Smith, 28. January 2019, 14:51

    John. No, I am anticipating significant PT growth so demand to the Station would be the same requiring the same size and frequency of Matangis as at present and the across town units would be ADDED to cater for the new across town demand. If we can achieve the same PT share as is achieved to the northern CBD then Terrace Tunnel traffic could theoretically (by my calculations) drop by a third making duplication unnecessary at present ( likely needed eventually but could be several decades away)

  22. John Rankin, 28. January 2019, 18:33

    Glen. Got it; thanks. So doubling the number of trains? One high capacity Matangi, followed by one low capacity cross-town unit. Is that technically feasible? Operationally, it’s clearly easier and cheaper just to run longer Matangis.

    We must hope that all options will be properly considered and evaluated, once Wellington gets Cabinet approval and funding to do light rail.

  23. Mike Mellor, 28. January 2019, 23:20

    Glen, further to John R’s points above, three questions:
    1. How will track capacity for the large number of new movements through the new junction that will be needed north of Wellington Station and then beyond be created?
    2. How will the very high level of reliability and punctuality on both the new and existing lines that this intensive operation will require be achieved?
    3. Are there any operating examples of the type and scale that you are proposing?

  24. Jonny Utzone, 29. January 2019, 10:47

    And here is a fourth question: How do you turn the trains/trams around at the southern end terminal wherever that is? Wellington Rail Station has 9 platforms and most are occupied during the AM and PM peak. How many platforms will there be at the southern end terminal?

  25. Roy Kutel, 29. January 2019, 10:51

    @JR or a $5 million per kilometre Trackless Tram (elongated trolley bus) that hopefully NZTA is seriously considering given the cost and disruption savings. If an elongated bus is good enough for Brisbane and Perth (and nearly good enough for Auckland before free-spending Phil Twyford arrived) it is certainly good enough for Wellington.

  26. Alf the Aspirational Apteryx, 29. January 2019, 13:43

    Will tooting be outlawed in tram tunnels? I certainly hope so, perhaps Cr Calvi-Freeman might be invited to comment?

  27. John Rankin, 29. January 2019, 14:16

    @RoyKutel: you may wish to look that particular gift horse carefully in the mouth. For example:

    Proponents make much of the fact that trackless trams do not need rails, a specially-prepared road bed, or underground services relocation. This reduces construction costs and the associated disruption. However, a key benefit of an on-street light rail bed is that it is essentially maintenance-free when built properly. If the trackless tram runs on regular roads, with regular underground utilities, whenever you do road maintenance anywhere along the line, or need to get at any underground utilities, the entire service comes to a stop for the duration of the works. What is the contingency plan to deliver service continuity during such works, which will inevitably be needed? You might say that the driver can simply drive around any obstructions. So, as well as providing 2 dedicated lanes for the trackless-trams, we also need to provide for 2 shadow lanes along the entire line as a contingency. This suggests that an autonomous trackless tram will always have to provide for a human driver. However, on-street autonomous light rail vehicles will not need a back-up human driver, eliminating the biggest single operating cost item.

    Whether we use conventional light rail or trackless trams, the need for a dedicated 2-lane corridor doesn’t change. If for example FIT’s proposal comes to fruition, we would still need a tunnel under Mt Albert and another under Mt Cook. We are still going to need grade separation where the line crosses SH1 in Te Aro, and perhaps in other places. So the idea that one could build a trackless tram system in Wellington for $5 million / km is simply not realistic.

    The specification for trackless trams states that the batteries require at least 10 minutes for recharging at the end of each trip. If the service provides for frequencies up to every 2.5 minutes, this means making room for 4 trains recharging at the same time at the end of the line. This is true whether the trains are on steel rails or rubber wheels. However, rubber tyres on asphalt have about 10 times the rolling resistance of steel wheels on rails. So the trackless trams will need bigger batteries and take longer to recharge than the equivalent light rail vehicle. My own view is that hydrogen fuel cells are the future for urban rapid transit vehicles, rather than batteries.

    Most of the objections I have raised are much easier to overcome in Perth and Brisbane. As I noted earlier, a good way to make sure trackless trams and other technologies are properly considered and evaluated is through a technology-neutral procurement process. In this way, the supplier carries the risk for delivering a technology that is fit for purpose.

  28. Roy Kutel, 29. January 2019, 15:54

    JR good point re batteries and I support a tech-neutral process. Me? Well I reckon an elongated trolley bus with re-chargable batteries to enable off-grid running would be eco-fantastic for Wellington, affordable and have a BCR way higher than a tram. If long term rail enthusiast Peter Newman can change his fixation from steel to rubber wheels, then surely Wellingtonians can and we can look forward to a system that does all the main corridors not just the rail station to the airport. One small problem. We would have to get rid of the GWRC first.

  29. Kerry, 29. January 2019, 16:09

    Jonny – Turning light rail is just like turning a train: the driver walks to the other end.
    The airport terminus will need three tracks: two for coming and going, and one spare. The spare is space to get a damaged or broken-down tram out of the way, until there is an opportunity to tow it to the depot.

  30. Dave B, 29. January 2019, 18:09

    The most efficient turnaround arrangement for high-frequency services – whether train or tram – is a turnaround-loop (sometimes called a ‘balloon loop’) which means drivers do not have to stop, change-ends, then reverse. They just trundle on round.

    The need for 9 platforms at the present Wellington Station is largely dictated by the way it is operated. Many trains are timetabled to lay-over for significant periods of time at the platforms, thereby tying them up. This lay-over time is good for preventing lateness-on-arrival translating to lateness-on-departure, but it is not good for efficient usage of platforms.

    Britomart Station in Auckland handles a similar level of service with only 5 platforms (one normally kept as spare), with much shorter turnaround-times. Once the City Rail Link opens and Britomart becomes a through-station, turnarounds in the CBD will become a thing of the past and all layover-time will be scheduled at the suburban-ends of the routes where it should be.

    The other problem at Wellington Station is that it has an adjacent train-storage yard and trains entering and leaving this currently have to jostle for paths with trains in-service.

    An extension of Matangi services to an alternative turnaround at *some point* south of the present station would greatly de-pressurize this current pinch-point and solve a number of problems which are going to have to be faced sooner or later anyway.

    Extension of heavy rail needs to be considered in the LGWM process. There are affordable ways in which it could be done.

  31. Glen Smith, 29. January 2019, 18:34

    John. A couple of points. You say that transfer time is likely to be 2.5 minutes (optimistic to get everyone off a matangi train with their airport baggage, across to a new platform and onto a new train even if the light rail is on time) but this is the ‘walk and wait’ time separate to the ‘disincentive’ time which from Neil Douglas’s research was an average extra 7 minutes for an across platform train to train transfer. This makes around a 10 minute penalty (ie when people are deciding whether to go by train or car they would view a 30 minute rail trip with transfer as a 40 minute trip). At present people can get off a train and onto a bus within a few minutes to get to the other end of the Golden Mile or beyond but the empirical evidence is they wont do this (recall the Station final destination egress data – virtually nobody transfers to bus and virtually nobody whose destination is beyond the Library are prepared to travel by train). FIT seems to assume that people who won’t transfer from Matangis to units with rubber wheels will suddenly change and transfer to a unit with steel wheels without providing any evidence or logical rationale as to why this would be the case. Like it or not, people are very reluctant to use rail if it involves a transfer – we need to get rid of it (if possible).
    Running just long Matangis would certainly be easier but won’t remove this potent disincentive (unless you can find someone prepared to pay for a heavy rail engineered corridor and 180m long across town station platforms)

  32. Glen Smith, 29. January 2019, 19:30

    Mike. Linking across town rail with network rail by ‘track sharing’ certainly adds complexity …. and yet it has been achieved by multiple cities overseas (and the prize of success is a big one). Perhaps, as Neil Douglas suggests, we could consult with them about barriers they faced and how they overcame them rather than reinventing the wheel – then the cost and time over runs might disappear.
    I’m not sure what you mean by ‘street running units’- from that I picture low speed truly ‘light’ rail units threading their way slowly through multipurpose streets with pedestrians and cyclists dodging around them. That isn’t what I’m talking about. Instead high quality dedicated/ segregated corridors that, while at street level, are free from any other vehicles/commuters and all interactions with cars/ cyclists/ pedestrians occurring at specific light controlled intersections with full rail priority, or avoided by underpasses/ overpasses.
    The solution I propose is one to overcome the practical issues associated with making tracksharing work in Wellington. The fixed predetermined features are
    1. A fleet of Matangis with long life expectancy that can be up to eight cars long (180 metres) and so have high capacity but can only service the Station, not across town (unless Dave can convince planners to pay for heavy rail).
    2. Across town corridors that are only likely to be able to accommodate perhaps 90 metre platforms and run units that are ?75 ? 90 metres long.
    3. An existing network that has to accommodate freight, ‘express’ units (that require longer time slots or they ‘catch up’ with other units), and the single Johnsonville line that limits train frequency.
    The solution of a running Matangis first and then an across town unit immediately after (with a modern train control system such as the European Train Control System this can safely be as little as 90-120 seconds later – this solves your question 2 above) is a unique but practical way of solving issues around this. This may only need to be at peak times – at non peak times just a 4 car across town unit may be adequate for demand. And some units could just be from the station with transfer if logistics demand this at peak time.
    The main result of this arrangement would be to limit the number of Station commuters occupying seats on the more limited capacity of across town units. But it would also likely solve problems with freight, express trains and perhaps Johnsonville. For example- non peak trains to/from Kapiti are every 20 minutes presumably giving freight a 20 minute ‘window of opportunity’ to get to the Wellington yards (or a siding). If you space units equally, then this drops to 10 minutes. But if you run 2 units just 2 minutes apart, the ‘window of opportunity’ stays at 18 minutes. Similarly express trains may have a longer window of opportunity as may Johnsonville trains (although I note the 3 passing loops on the Johnsonville line have limited length capacity – it might have to be consecutive siding loops).
    This is only exploring solution possibilities.
    In terms of rail frequency through the ‘pinchpoint,’ between 7.18 am and 8.38 am there are 33 arrivals scheduled at the Station, fitting closely the CCF’s statement on 7 April 2018 that trains arrive up to every 2 minutes (and in fact units are scheduled to arrive at 8.13, 8.14 and 8.15 am which, last time I looked at the clock, was every minute).
    The problem with the pinchpoint is it stupidly forces Hutt and Kapiti lines to merge, limiting capacity. From recollection I e-mailed some practical ways of increasing capacity by separating these lines (and solving your question 1 above) that involved minimal track changes while also facilitating separation of across town units on the east (without having to cross all the other tracks to get to/ from the most western tracks).

  33. Roy Kutel, 29. January 2019, 21:14

    So Dave and Kerry – would you reduce Wellington rail station to 3 or 5 platforms?

  34. Kerry, 30. January 2019, 8:17

    Roy. By all means close platforms when they are no longer needed, but I cannot see it happening soon. Tram-trains are neither cheap nor easy, and the first step is a preliminary assessment to see if a Wellington study would be worthwhile. This paper gives useful guidance (Hint: start with the definitions)

  35. Roy Kutel, 30. January 2019, 9:05

    Kerry your guidance paper is not easy to download (facebook/google – possible payment) so I’ll wait for a ‘preliminary assessment’. I would have thought that tram-train would have been one of the options sifted through by the $1 million Spine Study and put in the e_bin by the clever people at GWRC before their detailed study found el-cheapo diesel bus priority had the highest BCR for Wellington Railway station to the Airport.

  36. greenwelly, 30. January 2019, 10:10
  37. Roy Kutel, 30. January 2019, 10:58

    Thanks GW. I note this sentence: “Successfully introducing a Tram-Train system requires extraordinary good cooperation between many stakeholders. This requires time and good institutional cooperation. Tram-Train-planning must be integrated in the city and regional planning”. Doesn’t sound like Wellington does it?

  38. Kerry, 30. January 2019, 12:49

    Greenwelly, thanks for the revised link.
    Brent, perhaps you could put the new link in your RTSA paper

  39. John Rankin, 30. January 2019, 16:30

    @Glen, some comments.

    1. My previous comment specifically said “wait time” not “transfer time” allowing comparison with, for example, a 7.5 minute average wait for a 15 minute service at say Johnsonville.

    2. Why wouldn’t light rail be on time? The operating characteristics are well-known and widely implemented. It would take heroic levels of incompetence to be unable to operate an on-time light rail service.

    3. Most of the airport demand comes from people working there, followed by visitors going to and from the city centre. Everyone else is a distant third. 1000 people working at the airport generate more demand than 200,000 residents making one trip a year. If the business case depends on people from the north of the region going to the airport to catch a plane, I suspect it will not stack up. FIT’s proposal still makes sense if the airport closes; the line just ends one stop earlier.

    4. Among other data, FIT noted research by one Neil Douglas that perceived transfer penalties for rail-rail < rail-bus < bus-bus and are a function of the badness of the transfer experience. So a high quality rail-rail transfer is the least worst option, which FIT's proposal provides. Neil may have a comment. The current rail-bus transfer is a case study in how not to do it.

    5. A congestion charge on drivers may be a better value way to offset the residual transfer penalty than trying to design it away.

    6. FIT's proposal is our take on a Minimal Viable Product: we can spend more and get more benefits, but not less. A Prius is better than a Corolla, but you have to decide how much you are prepared to spend. I fully support your efforts to push the envelope, but let's avoid buying nothing because we can't afford everything.

    7. Having lived and worked in cities with great public transport (and much higher ridership that Wellington), where transfers Just Work, I don't have the "avoid transfers at all costs" attitude so common here. But people have to trust that connections will work. This trust is rightly missing here, because connections generally don't work. So I'm interested in the research on what you have to do to deliver trustworthy transfers.

  40. Dave B, 30. January 2019, 19:10

    With regard to forced transfers between different public-transport services, yes, these form an integral and necessary part of overall network function.

    However there is a simple and obvious criterion which determines where they should be imposed and where they should not be imposed.

    If a passenger-flow is minor, and if by offering a transfer to a connecting-service you can provide additional benefit for this minority-flow without inconveniencing the majority-flow, then win-win. You provide it.

    If a passenger-flow is a major arterial one (such as 15,000+ people arriving in Wellington by rail every day), the last thing you should do is impose a transfer on all of them. The De Leuw Cather study of 1963 estimated that 75% of rail passengers would wish to continue on a rail-extension if it were provided. Forcing 12,000 people to transfer between modes is a guaranteed way to impede the success of the network. You don’t do this if you can possibly avoid it.

  41. Margaret H., 30. January 2019, 22:08

    DB but 85% take a 5-10 minute walk from the rail station to their destination in central Wellington. What’s wrong with that. Nice and healthy.

  42. Ross Clark, 30. January 2019, 23:38

    How much difference, in actual passenger numbers, would be made by:

    * A Wellington city tram network, terminating at the railway station

    * Tram-train of a sort which would remove the need for transfers for journeys between the railway hinterland and (a) the CBD and (b) including the wider Wellington city as a whole.

    I am working from the POV that two-thirds of the current CBD workforce are within a comfortable ten- or so minute walk from the railway station, and that people who work in Te Aro are more likely to drive – even if a direct bus is there – because of the availability of parking.

  43. Richard M., 31. January 2019, 6:59

    Ross Clark – My estimate is a max uplift of 20% in rail patronage for a theoretical tram-train (which will never happen for practical and economic reasons such as Wellington Rail station is a heritage building and the Matangis will last another 25 years).

    GWRC would get consultants back in with the same peer reviewers to predict a zero uplift (even perhaps a reduction) for a tram network because of the unpopularity of suburban transfers. That was the result of the 2013 Spine Study using the Wellington Regional Transport Model. And given the same model would be used again, the same forecast would eventuate.

  44. Kerry, 31. January 2019, 8:31

    Ross – Take a look at Figure 5.7 in the PTSS Modelling Report. The number of rail passengers taking the bus is well below the number walking, as far south as Ghuznee St (1400 m), but is largely bus-based by the time you get to Vivian St (1700 m). Similarly, walking northward dominates as far as Tinakori Rd (900 m).
    Tram-train proponents worship Karlsruhe but forget that the Wellington Railway Station is only about 550m from the centre of the ‘main activity centre’ (say a range of Pipitea St to Ghuznee St, centre about Grey St), but Karlsruhe is a walk of 1900 m (Naegeli paper: link above).
    About 16,000 rail passengers come into Wellington each weekday, but the capacity of tram tracks through the city will inevitably be lower, say 12,000 at best. Growing ridership will quickly push down the number of passengers who can take advantage of tram-train, even if they want to, while the costs remain as high as ever. Tram-trains are competing with far cheaper conventional transfers, and nobody has shown how it will be done.

  45. Dave B, 31. January 2019, 11:26

    @ Margaret H: “…85% [of rail users] take a 5-10 minute walk from the rail station to their destination in central Wellington. What’s wrong with that. Nice and healthy” Superficially a very good point, and one which helped the infamous Public Transport Spine Study to summarily rule out any extension of rail. However the counterfactual consequence is that people from rail-served areas whose destinations lie more than 10min walk from the station tend not to bother using the train. Chances are they will go by road instead, adding to the city’s traffic problems.

    My claim is that a large amount of potential demand for rail is turned-away by the service terminating where it does. That it is as successful as it is indicates how effective the mode is, so we should be going all-out to maximize its effectiveness by extending it along the corridor that has been identified as part of the region’s spine. Building more motorways and road-tunnels will encourage traffic, not reduce it. Extending rail along this congested corridor will take people off the roads, as it already does in the areas it currently serves.

  46. Jonny Utzone, 31. January 2019, 11:59

    @Dave B – Won’t the roads refill with ‘latent’ car demand or is that an argument that applies only to road projects?

  47. Mike Mellor, 31. January 2019, 13:42

    Anyone contemplating tram-train (“track sharing”, the term Glen S uses, appears to be essentially the same thing) should read the article linked by greenwelly above, and note that while the benefits are potentially significant, there are many obstacles in the way.

    The paper makes the point that tram-train is complex, both institutionally (as noted by Roy K) and technically (as they have found out in Sheffield, at a substantial cost in time and money). Complexity always has a high cost, and that would be very much the case in Wellington, since as far as I’m aware there are no full tram-train systems (trams sharing with all types of train, including freight) on our narrow track gauge anywhere in the world.

    So we’re not just talking about complexity, we’re talking about innovation – an excellent thing, provided someone else takes the risk and does it first (remember Wrightspeed?). Anything unconventional or non-standard is essentially taking something of a gamble, with odds that history shows are generally pretty poor.

    So any benefits (which I’m sure would be there) of converting the Wellington terminus into a through station would have to be greater than the costs of overcoming complexity (both sorts) and of being a world leader. Both of these are highly risky and therefore highly expensive – so beware.

  48. Dave B, 31. January 2019, 14:34

    Hi Johnny Utzone. I didn’t say that extending rail would reduce congestion. I said it would “take people off the roads” (a deliberate choice of words in anticipation of a comment such as yours!). More people may indeed choose to flood onto the roads to fill the freed-up space. But at least there would be a ‘congestion-free alternative’ which that corridor lacks at the moment, and potentially thousands could benefit from it – as they do throughout much of the region which already is rail-served!

  49. glen smith, 31. January 2019, 19:46

    Mike. I use the term ‘track sharing’ rather ‘tram-train’ to indicate it is not ‘trams’ that I am intending to share with ‘heavy’ rail but units that are as ‘heavy’ as possible while avoiding the need for expensive over-engineered across town corridors. In discussing this issue TheTrams.co.uk states “Virtually every tram systems can be considered as light rail, but only those light rail systems which feature street running can be called trams.” It clarifies this street running (ie tram) as “a vehicle which runs on fixed rails and is designed to travel on streets, sharing roadspace with other traffic and pedestrians”. The corridors I propose are dedicated/ segregated (similar to our mainlines but with lighter engineering) therefore what I am proposing isn’t ‘tram-trains’.
    In term of sharing with freight the 2002 review of track sharing page 28 notes “Dueren, Kassel, and Saarbruecken have the dominant use of the tracks for the light rail vehicles, with a very limited level of freight traffic also using the lines. Certain Karlsruhe and Cologne lines have their light rail vehicles sharing the tracks with significant levels of freight trains, as well as regional DB Railway trains. On the line between Karlsruhe and Baden Baden, the observed ratio of LRVs to freight and passenger trains was approximately 8 to 1”. It also notes that ‘One study mission participant reported “while waiting for a light rail train in Saarbruecken, a DB iron ore train passed the platform where I was standing, just minutes ahead of my arriving LRV on the same track.” This was 2002. It seems that sharing ‘light’ rail with freight isn’t quite the innovative risk taking adventure you seem to indicate that it is. You qualify your statement by saying ‘on our narrow track gauge’ to make it probably true but exactly how the gauge of rail track has any practical implications is unclear.

  50. Wellington.Scoop, 31. January 2019, 21:35

    Comments on this article are now closed, because we’ve reached the maximum number that our system can cope with. However you may wish to comment on Undeniable – the case for light rail, where there’s still space for more contributions.