A year ago yesterday, and in the middle of Auckland’s second lockdown, speeds on the entire Auckland rail network were slowed to just 40km/h as a safety precaution after Kiwirail revealed wear on the tracks, known as Rolling Contact Fatigue (RCF), was more widespread and severe than they previously realised. At the time they said they urgently needed to replace 100km of track across the network.

A week later they shut the entire Eastern Line between Quay Park and Westfield down for two weeks in order to accelerate that work. That two weeks extended to four weeks and over the subsequent six months or so the rest of the rail network experienced various closures as Kiwirail worked to fix the issues. As of early February, with just the Pukekohe line to go, Kiwirail said:

“So far we have replaced more than 112km of damaged rail which is 84 per cent of the required re-railing work overall. At the same time, we’ve replaced close to 20,500 sleepers and destressed 95km of rail track.

They now say they’ve replaced 130km of rail which is significant when you consider there’s just over 180km of track passenger services run on.

After thinking the work was all done, more issues surfaced in mid-June which saw Southern Line slowed down again and services reduced as a result which lasted till late July. This and issues like left over rail causing $250,000 of damage to Auckland Transport’s trains and Kiwirail prioritising of freight prompted letters from the Mayor and AT and included calling on the government to put people with asset management and public transport experience on the Kiwirail board.

The impact off all the disruption and unreliability on train ridership has been significant. Following the first COVID lockdown usage on all modes of public transport was recovering compared to the same time pre-covid at a similar rate – rail was only about 5 percentage points lower than buses. But as we came out of the second lockdown we saw a notable change with rail usage lagging significantly with rail about 17-20 percentage points lower than buses. Notably ferries also saw extra benefit in September following the damage to the Harbour Bridge and have recovered much better this year.

Now an independent report has been released providing a route cause assessment for how the network (AMRN – Auckland Metro Rail Network) got into this position.


There were three key areas studied, the track, the vehicles and the wheel rail interface (WRI). Here are the findings for each of them.


The report highlights that while there has been investment in network improvements, such as double tracking, new stations, new signalling, electrification and new trains, the “existing track and civil infrastructure, including historic formation was not upgraded under any of those programmes“.

Under the section on the track there are four key areas.

  1. Historic under investment in the track asset prior to 2014 up to September 2020
  2. Insufficient rail grinding from 2015 to 2020
  3. The existence of multiple sites where the track condition is sub-optimal in engineering factors known to accelerate the growth of RCF:
    1. Track geometry and gauge exceedances including at welds and bolted joints
    2. Aged timber sleepers unable to hold rail in place adequately
    3. Historic wheel burns/squats causing sudden dynamic loads
    4. Sub-optimal application of cant, mainly from uncorrected past practices
    5. Significant sections of the network have low track modulus (low combined stiffness of rail, sleeper, ballast, and formation), at times aggravated by poor drainage.
  4. The speed of RCF propagation varies between very dry and very wet climates. It is likely that Auckland’s climate has been a partial contributor to the accelerated growth.

Interestingly the report sates:

In 2014, prior to the commencement of the new electrified service, AT engaged Network Rail Consultants to evaluate the overall state of the infrastructure. The evaluation effectively concluded that substantial investment (~$100m) in the AMRN [Auckland Metro Rail Network] track assets was needed to ensure it would be fit for purpose for the proposed EMU operation. This investment was not approved and the parties instead relied on increased inspections for safety, track speed restrictions, and accepted the infrastructure would provide lower levels of service.

I wonder who declined this investment. We’re now paying $200 million to fix the network so I wonder how much of this is a case of penny pinching coming back to bite us, especially as they also note there has been rapid and non-linear, accelerated growth in RCF. A separate part of the report notes rail wears in a linear pattern to a point but after that happens exponentially. To avoid that Kiwirail need to do the maintenance before reaching that tipping point.

The Vehicles

Back when this issue first emerged there was a lot of finger pointing at Auckland’s electric trains, though given this issue also exists outside of the electrified network that was always a long bow to draw. However, the report does note some issues with the trains as a result of trying to make the trains more comfortable for passengers.

  1. The AM class EMUs were designed with a high primary yaw stiffness to improve passenger ride comfort. However, this increases a vehicle’s propensity to cause RCF. The modelling commissioned did not fully assess the “RCF damage index” relating to this class of vehicle.
  2. The AM class EMU wheel profile was modified by the manufacturer from the KR standard profile to avoid anticipated high wheel flange wear, itself related to the high vehicle stiffness. During running rights approval KiwiRail acknowledged that modelling by the supplier of the proposed alteration, showed it to be safe, but had concerns that the change would impact on the rail maintenance requirements. Modelling in this study showed the profile change increases the formation of RCF over the KR standard profile.

On that first point, in the more detailed part of the report it suggests high stiffness is only an issue where the track conditions aren’t good, which they weren’t

The vehicle stiffness has been shown to have little impact on RCF for perfect track. However, logic, literature, and the Matangi design experience indicate that a vehicle of high stiffness, on low modulus track with maintenance tolerances that are more suited to a freight railway is likely to accelerate RCF.

That last line highlights the real issue, Kiwirail were maintaining the Auckland network to the standard of a little used freight railway.

Wheel Rail Interface

The points for the wheel rail interface are all covered above however they’re listed below too..

  1. Lack of comprehensive grinding since 2015
  2. Lack of artificial rail inclination on track structures
  3. Lack of optimised wheel and rail profiles


There are 8 recommendations in the report. These can generally be summarised as

  • Fix the tracks
  • Set up a 30-year maintenance programme to keep the network up to spec
  • Optimise the rail profile and the wheel profile of the electric trains to reduce issues.

For both those last two there is emphasis on the need for the various organisations to work together in order to achieve the lowest total cost of ownership

Finally, one other thing I noticed in the report. There’s a section which gives a development timeline for the Auckland network. It includes this comment perhaps suggesting why, despite buying enough trains for all services (except Onehunga) to be 6-car trains, it’s not uncommon to get a 3-car train at peak times.

Share this


  1. Was it Gerry Brownlee, or Simon Bridges in 2014 when “This investment was not approved”? They changed on 6 October 2014.

    1. Either of them: not the brightest of cookies, and knowledge of rail of both probably sitting at an abysmally low level.

      1. Bridges was actually fairly good don’t you think? More seemed to get done back then, everything seems stalled now.

        1. I have no idea how you reach that conclusion, Bridges suppressed the report on the third main, delaying it for years. The Eastern Busway was stalled and going nowhere for years under his watch. Lots of things are underway and happening now, maybe you should get out more.

        2. “Getting things done” is, to me, no substitute for “knowing what needs to be done”. Best to know first, before doing.
          Both Brownlee and Bridges were / still are hopeless at their portfolios. I think the only time we’ve had anything like a Minister that knows what they are doing in the Transport field was Julie-Anne Genter.

          But to be fair, most of the time in politics the Minister in charge will know bugger all about their actual subject. For some reason, construction and building always gets the lowest clued-up minister possible. The last several ministers – both Nat and Labour – of housing / construction have been complete idiots with Zero knowledge of what they are talking about, with speeches written by flunkeys also with no clue. It’s painful – and the end results show up in the crap getting built in our cities.

  2. Later in the report, it says the $100 million track upgrade identified as required in 2015, was rejected by the NZTA. After which nothing much happened, almost zero grinding resulting in the 40 km/h debacle last August.

  3. Hopefully the root organisational cause has been solved with kiwirail moving to the different funding model. I presume this means that they wont have to go to ministers every time they have to conduct a large-ish maintenance program.
    In 3 year terms these politicians are always going to prefer having big new capex projects done under their name, not boring maintenance that is seen as a cost to minimise.

    Does anyone know how expensive grinding is? Of course its going to be way cheaper than buying and installing new steel. Perhaps we need to purchase better equipment https://youtu.be/x0k8HX0GHyI

      1. They do , so why don’t KR buy say 2 ,1 for the North Island and 1 for the South Island . Or when Hillside is finally up and running build our own .

        1. Given probably 60 % of the work is on the Auckland and Wellington metro networks I’m not sure one in each island would be much use.

  4. The issues that arise between the contact surfaces of steel wheels and steel tracks are probably way more complex than most people realize. Present GA company excluded of course. But for those that don’t know, there was a fascinating series of articles in TAUT magazine last year, (Trains and Urban Tramways, the magazine of the Light Rail association in the UK) about the difficulties of getting the right surface angle on the tracks, the characteristics of the steel in the wheel, the differential wear on wheels in corners etc. Incredible. A lot more tricky than I ever thought – and evidently even the Brits (who have loads more trains than us) find it difficult, so it’s no wonder that we are struggling with probably lower quality maintenance and less experienced staff.

  5. Wow, thanks Matt. Great post. That we’ve ended up with two major parties that have both sunk extreme amounts of money into road building, and both failed to set up a proper rail maintenance and upgrade programme is an outrage.

    “This investment was not approved and the parties instead relied on increased inspections for safety, track speed restrictions, and accepted the infrastructure would provide lower levels of service.”

    So even if they’d been lucky, and the decision not to fund basic maintenance hadn’t led to this egregious failure, they still made a decision knowing they were compromising the network and lowering the value for money the public got from the investment in electrification.

    That last item – does that mean that basically, we can fork out for new trains but they can never be addiditive? KR will ensure we only get the services that fit into the existing running rights?

    I imagine this is one thing AT’s CEO has been trying to change. Maybe we need to try to get copies of his attempts.

    1. The basic principle of both the Auckland and Wellington Network Access agreements is that neither party ( i.e Kiwirail or AT/GWRC) can make a service change that affects the other party’s access rights, without there being a compensatiry increase in network capacity.

      1. They might have to book that in with Transpower, so they can warn the Gentailers, who need sufficient time to import coal from Indonesia.

  6. I think many of the same problems apply to the Wellington network as shown by the increasing number of service alerts for delays and breakdowns I receive almost every day: I use the Kapiti line.
    I was informed by an official during the week that overall it is the result of 30 years of under-investment in track, signals general engineering and so on.

  7. Interesting comments from the Autech report commissioned by Kiwirail in 2019.

    “Consequently, as a prerequisite, in the KiwiRail-Network of Auckland Metro, the track foundation
    must be intensively improved. This was always related to “pumping” of rails. That means, as long
    as those bad track foundations are not treated, the RCF will re-start even if grinding measures took
    Therefore, the focus to avoid RCF must be set to preventive grinding. Those rails which have
    already deep cracks are lost and need to be replaced. It must be considered that intensified
    lubrication would even increase the risk of RCF since it reduces the “healing” wear.
    Specifications concerning classification of cracks in [KR-011] (which varies from given information
    in [KR-004]!) (5 mm/8 mm-limit) is, according to our point of view, much too broad. Cracks with
    depths of up to 5 mm should not be considered as “light” but as severe! According to our
    experience and common practice at railway and suburban operators rails with crack depths of
    more than 5 mm are to be considered to be dead. “

  8. I have being thinking for some time that maybe we are trying to hard. After all why drag around an extra 100 tonne 3 car unit for services that are only carrying 10 or 15 passengers. At the time of the electrification I was in favour of 2 car units rather than 3 car units for the same reason. Peak services are different and the afternoon peak is starting at 3.30 pm when the schools get out particularly noticeable on the Southern line. These require longer trains. The other thing is frequency off peak 20 or 30 mins seems adequate during peak this can be increased to 10 mins. So less axles less wheels less wear.
    The next thing is use back up bus routes to augment rail frequency already 321 the Hospital bus is a backup to the Southern line my view is it could be extended and streamlined. The airport bus duplicates the Manukau branch. There must be other opportunities out there. These buses also double as feeders and fill in the gaps created by station spacing and now closed stations.
    Lastly space out trains on shared tracks. By this I mean the Southern and Onehunga trains share track between Britomart and Penrose junction. At the moment they run within 5 minutes of each other. Also trains running between Puhinui and Otahuhu run within 5 mins of each other then we have a gap until the next cycle. Presumably this is done to facilitate connections and also to create gaps for freight trains to run. But for a typical passenger from Middlemore travelling to Britomart doesn’t care which way they go. Something that will be addressed with the 3 rd main project I expect.

    1. If we’re really worried about wear on the tracks the other option would be to just stop running trains, the problem would disappear overnight.

      1. That’s called taking an argument to an extreme. But do you wash a single item in your dish washer or do you wait until you have enough to justify the expense.

      1. Ok wait for an Onehunga train (30 min frequency) at Onehunga station. No one is there until five minutes prior to departure then they suddenly appear obviously they have read the timetable. And there are buses that run more frequently which is what I have advocated for above.
        And I recently travelled both way on well patronised buses that runs only hourly off peak between Huntly and the Hamilton transport centre. At one stage we had 26 on board plus the driver. The obvious antiquity of the bus shelters suggests this is a long standing route. Some passengers were picked up and dropped off at location that had no obvious bus stop signs. Probably as good patronage as many of Auckland’s off peak train services. It must be a relief to the drivers that most of the traffic has being diverted onto the express way. Certainly a much more relaxed feeling through Huntly and Ngaruawahia now most of the through traffic has gone.

        1. Ever been on the London Underground, Royce? If so, would you use it if you had to plan each trip based on a ½-hourly timetable? Maybe you would, but I imagine most people wouldn’t. What would happen is that patronage would dwindle to the extent that bean-counters would then claim that a train every ½-hour is all that could be justified.
          I wonder what the effect on off-peak motorway-usage would be if cars were only allowed once every ½-hour.

        2. Yes I have and the Paris metro plus the RER and the Paris Trams. I have mentioned the Onehunga line because it is currently 30 mins freq and I also suggested there are more frequent buses if you need to travel into the city and the new 36 bus will take you to Papatoetoe Station and Manukau which is a faster option than trying to change at Penrose if you are wanting to go south. I see the Onehunga branch as a perfectly adequate commuter railway it isn’t really part of the bus network. Stations on the Southern and Eastern lines are more integrated into the bus network and run with a 20 min frequency off peak. Auckland is not London and we don’t have multiple lines to transfer between. I am questioning whether need better frequency than this off peak. What is the point in running 6 car trains almost empty. Seems to me we are just wearing out the EMU’s and track by running more frequent longer near empty trains. I think people on this site have falling into an orthodoxy that we should be trying to create an Auckland version of the underground but that’s not right. What we have is a mixed freight and commuter railway. Hopefully in the future we will have more regional and long distance trains. Freight is important with trains running to Tauranga, Mission Bush, Northland and Hamilton and mainline shunts to the port, Penrose, Slyvia Park and Wiri. This is more important than near empty passenger trains. Five stations have being removed on the Southern and Eastern line. This has resulted in less destinations for passengers it has also meant buses have had to be deployed to fill in the gaps or in some cases no public transport has being offered. This sites orthodoxy is that this has made the railway more efficient and I will be a heritic and challenge that as well. On top of all this we have lost the international students and now we have working from home. The park and rides are not as full on a Friday or Monday. There isn’t the patronage to justify more frequent longer services off peak.

        3. Royce , are these EMU’s have fixed Boogies or are they on a pivot ? . As I have heard different stories about them and I just want clarification .

    2. I can’t see any suggestion in the report that reducing the size/frequency of trains would assist with resolving the problem. In fact, the summary of the SNC-Lavalin report says “wear is beneficial in that it wears away cracks before they can develop into RCF”, so “less axles less wheels less wear” would be precisely the wrong remedy.

      In addition to that, economics tells us that the sensible thing to do with capital-intensive projects is to use them to the maximum, spreading the cost. Limiting the use of a high-capacity railway as suggested makes no sense, especially when there is no apparent gain.

  9. “No one is there until five minutes prior to departure then they suddenly appear obviously they have read the timetable..”

    Except for those who looked at the timetable at 1:01pm, saw the next train was not for another 29mins and took the car. You know, those people who are unable to fit their life around an O-Line timetable

    I don’t know how many people that would be but I am guessing a lot more than those not in a hurry, don’t have other things to do, and can just wait 29mins for the next train.

    For what its worth, I think the Onehunga should be replaced by the LRT cross-town line here posted a few weeks back and terminate at Ellerslie where people can transfer. Or lean more heavily on buses and buslanes

  10. This from an overseas paper on the subject dated 2/04/2016 — “Rolling Contact Fatigue is a serious hazard to rail traffic and a major problem for railway infrastructure managers across the world”. So as Average Human has eluded to, it is not unique to Auckland or NZ, despite the correspondence from members of this group, and our local “experts” and their political masters. It is also a problem in other industries as well.

  11. Thanks – good post.
    At what point does “deferred track maintenance” become neglect – or criminal negligence.
    The Govt of the day didn’t fund maintenance. $100m saved.
    That the trains still ran was the failure – services should have been terminated – which is the only outcome from not doing maintenance.

    1. I would like to think that KiwiRail won’t be importing any more asbestos-laden diesel locos. That decision was made by a different management team, under a different NZ government. A lot has changed since then.

      1. KiwiRail ordered 10 more DL’s in 2020. The new ones don’t have asbestos, that was only trains in the first batch.

  12. So after taking a low capital investment high maintenance option, the maintenance was not done.
    I’m wondering how extensive the areas of low track modulus were and if most of those areas were just down to aged timber sleepers unable to hold rail in place adequately (given the numbers of sleepers replaced along with the rails).
    Nice to see the recommendation for NZ to have rail grinding capability….

  13. Question , who does the grinding/lathe work to the AT EMU’S , is it KR or do they do it at Wiri in their own workshops . As KR have their own plant at Otahuhu for that purpose could the 2 lathes be slightly different in angle when grinding the wheels ? .

  14. As a railway kid from the 1930/40’s I find all this very interesting. Father did 40 yrs from 1926/66. I guess what transpires in formative years guides sets your future knowledge.

  15. “Kiwirail were maintaining the Auckland network to the standard of a little used freight railway”

    That’s not a correct assessment. Freight trains are heavier and place more stress on tracks than lightweight EMUs. And of course in the case of Otahuhu-Pukekohe far more freight travels those rails than passenger EMU’s.

    The issue is essentially the result of the EMUs having the wrong wheel flange profile for New Zealand.

    1. Well the report states that low modulus freight standards combined with modern stiff EMUs with good ride quality causes RCF. One interpretation is that Auckland’s new EMUs should have been designed like freight trains with shitty handling for shitty track, the other is that kiwirail should have used all the track access charges paid to it by Auckland to maintain the Auckland network to suit Auckland new modern standard trains.

      There are 270 emu’s per day south of Otahuhu. While they might be lightweight there are a hell of a lot more of them than freight.

    2. Kinetic energy = mass x square of velocity / 2

      Going faster has more impact than running heavier.

      You can go online to see plenty of wibbly wobbly lines taking freight trains at walking pace, from the US to China. Smooth rails are for passenger comfort.

Leave a Reply

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