With the exception of how people access Britomart, the construction works for the City Rail Link so far have had little impact on rail users. But as we wrote last week, that will change from the middle of next year when the Mt Eden station closes for 3½ years to enable the western line to be dropped by as much as 4m as part of grade separating Normanby Rd and to accommodate the construction of the CRL tunnels.

There has understandably been some concern about the impact of this including questions about temporary stations. I’m not an engineer but I can appreciate it’s going to be difficult to do this work while maintaining a single track through the site and trains running at 10 minute frequencies, especially when you consider all of the health and safety requirements and that compared to previous examples of track realignment, this time there is also all the issue of overhead wiring to contend with.

The CRL works will see the removal of the Normanby Rd and Potters Ave road crossings along with the Ngahura St pedestrian crossing. The latter two will be replaced with pedestrian and cycle bridges. But that still leaves 15 road level crossings and an additional seven pedestrian crossings on the Western Line and as train frequencies increase post-CRL these will become increasingly important to deal with.

Some crossings will just be able to be closed but for that that can’t, in order to grade separate them, some may require tracks to be lowered or realigned. There is the additional challenge that many of the level crossings are also next to stations and so if tracks need to be lowered then that will also impact those stations too.

The current Regional Land Transport Plan, the 10-year investment programme that is refreshed every three years, currently includes about $239 million for closing roads and grade separation but most of that is not due to be spent till after 2024. That itself is just over half of the $424.3 million that was estimated to be needed for the network.

Given all of the above, one of the risks I can see is that just after we complete the CRL and convince a lot more people to use trains, that we then start going through a similar process as with Mt Eden resulting in lengthy periods of disruption and possibly closed stations while we roll through a grade separation programme. This got me thinking,

  1. Assuming the funding was available – which would require the government stepping in to provide it, could we get all of the crossings that require major disruptive works done before the CRL opens?
  2. Given how disruptive it would be, would it be better (and even possible) to ‘rip the plaster off’ and do it all in one go, even if that meant shutting the line down for a number of months? Another way of putting this is, would it be better to have one short but highly disruptive period or less overall disruption but have it drag on for a lot longer?

Of course even if there was the money to do this there would first need to be a considerable amount of planning and we can get an appreciation for some of the challenges from a study completed in 2014 and covered in this post. The study didn’t recommend specific outcomes but looked at the options for grade separation of the crossings AT didn’t think could be closed. Challenges include at Glenview Rd in Glen Eden where options such as putting the rail line under the road would require regrading 1.2km of tracks

Some of the geometric challenges will be even more pronounced now that there will be a future requirement to accommodate 9-car trains but perhaps such works could enable those longer platforms to be built sooner.

If we were to go for a highly disruptive option it would be interesting to understand just how long that would need to be for. Some of the big Christmas/New Year closures we’ve seen have included

  • A major regrading of the rail line between the Parnell Tunnel and The Strand along with rebuilding the underpass though to the domain closed the line for 4½ weeks and was needed to enable the future Parnell Station.
  • Similar closures occurred to lower the tracks around the Avondale station as part of double tracking works.

So Western Line users, would you be prepared to put up having no train for a few months if it meant there was no more disruption after the CRL opens? I think it’s something worth considering.

Share this


  1. Funny you should mention the Avondale Station redevelopment. If ever there was a case of ill conceived, badly thought out development then the relocation of Avondale Station was it. Why Ontrack—as KiwiRail was then designated— thought it acceptable to lower the tracks without addressing the St Jude Street crossing beats me. And then to add two circuitous pedestrian at grade crossings rather than the planned pedestrian over bridge. Of course, it’s all a matter of funding and the station was re-located as a part of the funding-constrained Project Dart, not helped by the newly-appointed Minister of Transport cutting all rail-related Capex to the bone. Even so, there were other funding sources that Ontrack could have accessed like the level-crossing elimination funds set aside by ARTA, etc. We really do need better transport planning. And we also need politicians to push, shove and demand nothing but the best, not to acquiesce in the face of institutional pig-headedness.

    1. The ideal solution would be to trench the line (plus a km or two on either side) so it goes under the street and rebuild the station several metres down. That isn’t a 21 day job even with all the planning and resources in the world!

      The only real solution to St Jude Street is to permanently close it to traffic. An underpass for pedestrians and cyclists could work.

      1. Closing it is also not practical as it is the direct extension of New North Rd to Great North Rd and New Lynn and it takes a lot of vehicles. Alternatives would either direct traffic through the town centre, suburban roads or be a long way out of the way. Lowering the tracks would require reconstructing the Blockhouse Bay Rd bridge. Its foundations were exposed and reinforced with the DART works. Part of those works was the shifting of the station from the bend on the west side of Blockhouse Bay Rd temporarily to the east. I thought this should have been where it stayed, at least until St Jude St is sorted out.

        1. If you close it it won’t be a a direct extension of New North Rd, and it won’t take a lot of vehicles.

          Just because we have a shitty trafficked road now doesn’t make it a good idea. It already directs traffic through the town centre into a shitty, dangerous roundabout. Get rid of the road and the roundabout and you’ll do Avondale two favours.

        2. Closing St Jude St would deprive AT of the income from one of its more profitable red light cameras although it would improve the quality of space in the vicinity of the station immeasurably. Could be the start of a process to remove cars from Auckland streets.

        3. The red light camera on the corner of Blockhouse Bay Rd/New North Road/Crayford Street East/St Jude Street still gets its fair share of red light runners going straight on Blockhouse Bay Rd alone.

        4. Closing it isn’t going to stop the traffic. It will divert via Rosebank Rd and through Avondale town centre, or through residential Chalmers St, or further out via Wolverton St or similar. None of these would be doing the locals a favour.

        5. Anthony, do you also dispute the concept of “induced traffic”? Because if you do, you are out on a limb. Closing a road has the exact opposite effect of induced traffic – it makes some traffic disappear. That isn’t a fringe concept and is reasonably commonly known. So I’m surprised you saying that “closing it isn’t going to stop the traffic”.

  2. “…Why Ontrack—as KiwiRail was then designated— thought it acceptable…”

    Well Glen Eden is an absolute stunner of a mess as well. That, combined with the testimony of everyone who has to drive around west Auckland, is the clue. Nothing is done properly out west because no one rich or important lives there. The cheapest solution delivered after every possible cost-saving band aid and do nothing justification is how the West is run.

  3. If it takes 3 1/2 years to build the CRL at Mt Eden station then you can reckon on a closure of about 9 1/2 years to do the grade separation of the Western line.

  4. In addition to the CRL there is also the possible port move which also would increase volumes particularly in the more western areas.

    In addition to level crossings in terms of planning there are also
    – existing bridges “upgrades” needed like the one on Titirangi Road by Bunnings in New Lynn that should ideally be widened.
    – ensuring any trenches have space for bypassing train lines for freight and ability to provide express services in these further out areas.

    While I understand budgets come into all projects, such core infrastructure should have a well documented and thought out target strategy based on best practice and design (not political whim).
    This also would help highlight short sighted decisions as such and make it clear rework will be needed later at some point, and so be costed. It would also force project stakeholders justified and non alignment.

    1. You are right about thinking about freight trains if they are going to the seemingly mad thing of moving the port to Whangarei. We had a property next to a rail line and the noise made by the container trains was beyond belief. The prospect of these deafening things running all night is something the residents will cotton on to soon I guess.

      Ironic I see that Helen Clark is now championing it – she who didn’t like the happy noise from charity concerts in Eden Park.

  5. Note that in Victoria the “Level Crossing Removal Project” (https://levelcrossings.vic.gov.au) are very politically popular – they are seen to speed up both car and train journeys! They’ve done 30 out of the planned 75 so far.
    It’s being done via a mixture of road closures, over/underpasses and in a few places “Skyrail” elevated rail over a longer section, often with rebuilt stations.

  6. Mt Eden station is on a lava flow isn’t it? So the entire lowering is through rock, making it a slow process. It took them months to get through lava at Grafton.

    But is the NAL being lowered 4 metres? I thought it was staying more or less where it is, with the CRL tracks being the ones on the lower level through the new Mt Eden station?

    1. Yeah there will be lava involved and as you say, that makes things harder.

      I initially thought it was staying the same but have been told that it’s being lowered from just after Grafton to about Potters Ave with it up to 3-4m lower in the middle which is for Normanby Rd. CRL tracks will obviously continue even lower.

  7. If the port does get moved, and assuming that means they need to build the Avondale Southdown line, I wonder if that would make heavy rail the best option for the airport again. And in that case I wonder if it makes sense for Airport trains to go via Mt Albert rather than Penrose to give the network better balance. And in that case the frequency between Avondale and city would be a lot more and any level crossings would have to go.

  8. A real “rip off the band-aid” outcome would be an entirely new Western Line, bulldozed through roughly parallel to the existing line with a 21st century alignment allowing trains to use their full performance envelope. Would probably require extensive tunneling and viaducts.

  9. I see this morning the Western Line was closed for a few hours after a truck hit the bridge at Titirangi Rd, which was previously a level crossing.

    It would be interesting to see the statistics of line closures at grade separated locations vs level crossings. Here in Taumarunui the Short St grade separation sees the NIMT close several times a year, as every bridge strike by a truck means the trains have to stop until it’s inspected.

    I suspect grade separation may have higher closure rates than at-grade crossings.

    In terms of trains, 90kph through a level crossing or 90kph under a bridge, it makes no difference. The primary benefit to grade separation is faster journey times for road vehicles. The secondary benefit is less equipment to maintain. But it doesn’t do anything to make the trains faster, so I question the importance being placed on removing level crossings. It’s a lot of $$$ out of the rail budget to make car journeys faster!

    1. That’s not correct, on stations where there is a level crossing next to it, trains approach the station slower and depart slower. So removing level crossings have benefits for trains too. And it doesn’t need to come out of the rail budget, in fact it shouldn’t, it should come from the roads budget.

      As for being closed more, that likely depends on the type of grade separation, rail under road would be a lot less disrupted

      1. “That’s not correct, on stations where there is a level crossing next to it, trains approach the station slower and depart slower.”
        Where’s your evidence for that?

        I’m not saying you’re wrong, I just want some evidence for it. Because I can’t see why it would make any difference. And I’ve never noticed any difference in approach speeds on the Wellington network.

        1. The question is: after closing the doors, etc., does the train have to wait for the crossing to close? That would be trivial to observe even for a casual observer. (it is a bit harder if you don’t live in Auckland of course).

        2. Daniel, you’re correct in thinking that there doesn’t need to be be a difference, but in Auckland the trains run under the ECTS, which takes control away from the train engineers and has automated safety features, which include very slow running of the train when it is near a level crossing. I don’t think it’s something the engineer can override unfortunately. It does make it “safe”, but also VERY slow sadly.

        3. @Roeland:
          Surely the bells begin ringing and the barrier arms come down before the train even stops at the platform?

        4. @Bevan.
          Isn’t ETCS normally used on high-speed rail lines or freight lines? Why did Auckland spend money on it?
          Surely this should be tweaked (or abandoned altogether)? If anything; so cars don’t have to wait so long at the level crossings?

        5. Daniel – the barriers don’t start going down until close to when the train is about to leave. I think the TM presses a button but someone may correct me on that.

          The reason the train approaches the platform slowly is because there is a red signal until the barrier arms go down and ETCS won’t let a train approach a red signal at speed.

          There are some partial solutions to this but I won’t try and explain them as they are a bit beyond my level of expertise.

        6. Maybe I’m wrong but I’m not noticing any slowdown in train speed over level crossings in the following video from about a month ago:

        7. Ah I can see what Jezza’s saying:
          So the arms come down at the crossing when the train stops.
          But the train’s acceleration from the station across that crossing really doesn’t look to be any different.

          So I’m still questioning this idea that level crossing’s slow the trains down.

        8. The issue with stations next to level crossings is to do with the ETCS signalling system and the Automatic Train Protection – something Wellington doesn’t have – I recall reading about it in AT board reports. There are two issues at play, they have been reduced from what they were but still exist.

          ETCS level 1 (which we have) communicates the signals to the train via bailise on the tracks (yellow boxes you can sometimes see on the sleepers). That tells the train what the signal settings are on the upcoming signals and displays it for the driver in the cab. There are other levels of ETCS and which we may need to upgrade to in the future


          Auckland’s trains also have ATP which prevents trains from running red signals and will automatically slow trains down that are approaching a red light too fast.

          When a train is train is approaching a station it is stopping at with a level crossing on the other side the barriers don’t go down till after the train has stopped to reduce delays to drivers. There will be a signal between the platform and level crossing to make sure the train doesn’t go through it without the barriers down. As such, the train has to approach the platform slower so the ATP doesn’t kick in and apply emergency brakes etc.

          When the train goes to leave the signal has changed but the train doesn’t know that till it passes the next bailise so it has to crawl forward towards the signal till it does that. Once it’s picked up the signal has changed it can accelerate faster. That’s why in that Avondale video the acceleration noticeably picks up a few seconds after it departs.
          Things have been improved from what they were as additional bailise have been added to reduce that second issue and settings have been tweaked to reduce the impact but it is still there.

        9. From where I’m sitting; the impact of the safety control systems on the speed that trains travel across level crossings looks rather negligible.
          And nowhere near enough to even be sensibly considered as a justification for spending hundreds of millions in completely grade separating the western line.
          There is a list of infrastructure projects which are awaiting funding such as:
          * the third main between Westfield and Papakura
          * curve easements along the main trunk line between Papakura and Hamilton (also essential if this regional rapid rail is to be any success), * A new station in Drury
          * Southdown-Avonside link
          * a new harbour crossing (and I know you pretend that won’t be needed)
          * replacing the port of Auckland (which is likely to also mean improving the NAL and will probably also allow a feasible extension of services to Helensville)
          * connecting Auckland airport to rail.

          Yet you’re advocating sinking money into something with such a weak justification and which would gradually occur on an need-to basis anyway?

    2. Geoff. The other main aspect of grade separation you have not covered is safety. Isolation of trains from pedestrians and road users tends to reduce the occurrences of trains being out of action due to colliding with people who are, for whatever reason, on the tracks.

      Collisions with trains are often fatal. There is a social cost incurred for fatalities and injures occurring on public road which is likely to be a similar cost for incidents that occur on the rail corridor.

      Overhead bridge strikes can cause disruptions just as level crossing collisions, but occur much less frequently and with a lower probability of any loss of life.

      1. “Overhead bridge strikes can cause disruptions just as level crossing collisions, but occur much less frequently”

        But how do you know they happen less frequently? That’s why I think it would be interesting to see the statistics.

        The Short Street bridge near me gets hit about six times a year. How many level crossings have collisions six times a year?

        Road-over-rail bridges are less disruptive to trains as Matt points out, but are they better? Vehicle collisions happen on them more frequently than an average stretch of road, because the space is usually more confined, and vehicles colliding with the ends of the bridge happens a lot too.

        So, it would be interesting to see statistic for all types of occurances on road-over-rail and rail-over-road bridges, and compare the injury/fatality rates with level crossings. I’m not convinced they offer an improvement, or at least not as much as one may assume.

      2. How often does anything collide with a train?
        While the results are catastrophic; does it really a frequent enough occurrence to justify the hundreds of millions in cost that complete grade separation would require?

        1. Train collisions happen more often than you think. You only have to see the countless footage of near misses from people and vehicles who are obviously to or blatantly ignore barriers and signals.

          While many people value their lives, some people also value their time. There is economic value to grade separate the rail corridors.

        2. But are these collisions frequent enough to justify spending hundreds of millions?
          Look at the infrastructure improvements needed across the nation and then ask yourself if this is really any priority?

        3. Yes it does justify it, and yes it is a priority.

          Should have been done 100 years ago of course – no time like the present. Do it once, do it right.

          Following order of importance:
          Porters Ave first
          George St second
          Morningside Drive
          Asquith / Rosgrove
          Saint Jude Rd

        4. And what sort of monetary cost are you estimating there Guy?
          And what are you planning to sacrifice/delay for this luxury project?

          Should’ve been done 100 years ago? Given that it was a single track until about 10-15 years ago; that would’ve made it too expensive to double-track.
          Hmm. Maybe you need to think your opinions through more?

  10. And this from AT this morning ;-

    “Due to a truck hitting a rail over bridge near New Lynn, all trains are currently cancelled or delayed on the Western Line. Trains can not run over the bridge until it is inspected. Scheduled buses are accepting all train tickets and AT HOP cards.”

    As far I can se1] put bridged over the lines or 2] fit steel steel barrier arms on all crossing and make them double on both sides of the track and have chains hanging off them to put off the idiots that try to jump through the barriers as they come down . That way they then wiil have trouble trying to explain to their insurance companies how the paint work got damaged ? .
    In the old days when I got my licence it was slow down , change down and be prepared to stop . But now it seems to be change down , speed up and go like hell to miss the barriers as they come down . And don’t say they are time poor as if you get hit by a train , you will have a long time either in hospital or the local grave yard .

  11. On a suburban line with closely spaced stops the differences in journey times between having the availability of a “full performance” envelope and the current fleet, (slow door operation excepted) is little. High top speed is unobattainable between closely placed stops anyway. The time at stations is the same. The current alignment does have speed constraints at grade crossings and perhaps on some curves.

  12. With the reduced running on the Western while CRL is being built they should look at doing as much as they can now. With the reduced running it means they could run single track past these crossings while lowering/raising the other line, then switch over.

    As for trucks hitting bridges, they need to install overhead barriers before the bridges so that the bridges themselves never get hit (same goes for that motorway one that always gets hit around Penrose).

    1. What reduced running? As far as I know, the frequencies on the Western Line aren’t being reduced during the CRL construction period. The only difference is that trains won’t be stopping at Mount Eden station for approximately four years.

    2. Penrose Rd Bridge does have an early warning system. Flashing you are over IRRC but I guess now and then it is ignored. Yes an ugly light structure before a bridge may work but would fling debris around the place when struck, enter than a rolled truck perhaps.

  13. The Melbourne Level Crossing Removal Authority is fully paid for by the Victoria State government, with $2 billion initial funding about five years ago to remove about 30% of all 177 crossings. This was later topped up with another billion or so to allow removal of a further 25 crossings. Rather than treat each crossing as a project in its own right, the plan has been to deal with multiple crossings on the same line as a “box set” – so in some cases the line has been elevated for some kilometres, allowing grade separation of a series of level crossings. Here in Auckland some thought has already been given to a set of 4 crossings between Takanini and Papakura, and of course the 8 crossings on the Onehunga Line. Even with their generous level of funding, some Melbourne level crossings will simply be closed, but there is a limit to how many you can close as severance issues arise for communities on either side. In addition, although closing a level crossing may appear to be a cheap solution (i.e. free), traffic denied one crossing point will quickly seek out nearby alternatives and potentially overload them and intersections on the roads that feed them – fixing that mess could be quite expensive and potentially comparable to fixing the level crossing being considered for closure.

    1. Unfortunately the cost is actually far, far greater. The initial 50 level crossings cost $8.3 billion AUD (https://www.theage.com.au/national/victoria/level-crossing-removal-program-poor-value-for-money-auditorgeneral-andrew-greaves-20171214-h04ed0.html). That is about in-line with a rough estimate of $100-200m per crossing.

      The additional 25 crossings are being removed at an estimated cost of $6.6 billion AUD (https://www.danandrews.com.au/removing-75-level-crossings).

      Based on those two figures, the total estimated cost for the 75 crossings is $15 billion AUD. That’s well over 3x the cost of CRL.

      $2 billion was the approximate cost of the Caulfield-Dandenong Skyrail project, which removed 9 crossings.

      There are definitely huge benefits, but it is crucially important to recognise that the cost is actually extraordinarily high.

  14. As a mainly Eastern & Southern line user. Yes do it all at once!
    But seriously, I think doing as much as possible in strategic multiple areas would be ideal. Post CRL things are only going to (or really a latent demand “want to”) get busier with those great time savings into the city centre for western users, population increases & “hopefully” some modal shift.

    Some of these level crossing are going to cause a bit of disruption & death or serious injury if you ask me. Maybe some island platform upgrades as well. Was actually at the Kingsland Station last week & some kids jumped down & crossed over the rails to get to the other side during a peak morning period.

  15. It’s probably not worth the lots of money and time needed.
    It should only be done on a need-to basis; if the level crossing holds up too much road traffic. It’s not like level crossings hold-up the train.

  16. All depends how seamless they can make the rail bus rail Connections. If it an be done right it might not be to bad. I never use the western line so can’t help.

  17. The sooner these are fixed the lower the personal and economic impact. Post CRL the whole rail network will be ever more vital to the city’s economy and particularly to relieving the dominant road networks, enabling them to function better more often.

    There is surely a case for making a substantial move on the western line ones over the coming summers. Though some are very hard (ie expensive), and seriously need considering for simple closure; yes you Morningside!

    The counterfactual is there are perhaps two big uncertainties in 9-car sets and possible new huge freight use to NorthPort.

    So perhaps the key work should be on the southern group (aligned with 3rd main, new bus interchange stations on bridges above), as post CRL the South/Eastern pair would then be entirely grade separate. Though that would very much leave the Western Line as the poor cousin, as it already sort of is.

  18. Some level crossings should be easy to remove for a council and government into so-called Zero tolerance to death and injury on the roads.

    AT could start with removing the lethal pedestrian level crossings they have installed on many stations after their modernisation for convenience sake. They even installed them where there weren’t any before, Glen Innes and Mt Eden spring to mind, all to save pennies. They are routinely ignored by pedestrians and they are utterly lethal but ever so cheap.

    In the fast stretch of line between Te Mahia Platform and Tirinui Rd there are 5 level crossings especially for freight or empty passenger trains (they are the most high risk because people wrongly think they will stop at platforms) and between Takanini Platform and Tirinui Rd pedestrian level crossing it is close to the fastest on the network and yet at 110 km/hr for EMU’s and 80 for Freight, little ol’ humans are allowed to wander out past the bells and lights to their demise.

    Undergrounding exactly as per Kingsland and Parnell is the only way to go. But it depends on how serious they are about their newly found desire to reduce death and injury on our commutes!

    1. The best time for replacing at grade rail crossings on the Auckland suburban system has already past. The next best time is before the new rolling stock and the CRL make increased rail frequency possible. The more trains running the more potential conflicts, deaths and disruption between rail, road and pedestrian traffic. It needs fixing. We know fixing will be disruptive, we know the disruption is more or less in proportion to the patronage on the rail line.
      We know patronage is roughly proportional to rail frequency.
      All this points to is the next best time to provide the required grade separation is as soon as engineeringly possible. It will never be cheaper.
      But nothing can be started until the polititians provide the seed funding.
      The ball is in your court, Phil, all both of you.
      And for those polititians on the other side, stop trying to thwart long term solutions to our inefficient transport by championing the causes of the disrupted.

  19. You’re trying to make a de facto metro out of a general purpose railway, not a cheap proposition. Maybe best to get the low-hanging fruit and upgrade the rest.

    1. I was wondering if that was the ulterior motive for this very unwise investment.
      Honestly: The network has so much potential to be an S-Bahn and provide regional services, something better than some shoehorned pseudo-metro, that would be lost. And do the suburbs of detached housing in Auckland have enough population density to support a metro anyway?

  20. Yep, level crossing removal should have been part and parcel of CRL considering the changes to train frequencies enabled by it. Although that would have blown out the BC ratio…

Leave a Reply

Your email address will not be published. Required fields are marked *