My post yesterday about the hot mess that is the proposed Tamaki-Ngapipi intersection resulted in a lot of discussion, especially around the design and the role consultants play. Reader George who is also an engineer decided he could come up with a better design and posted it on twitter last night. He says that the design fits well within the proposed reclamation.

To me this is a vastly superior solution and one that caters for for all users. Compared to the official version:

  • People in vehicles appear to be no less inconvenienced, there are still exactly the same number of lanes.
  • For those on bikes, whether squeezed into lycra or just out for a cruise with the kids, it appears to be considerably safer and offers more options, such as only having one crossing to contend with to turn right onto Ngapipi
  • For those on foot also benefit, especially on the northern side where the cycleway is more defined and so less sharing is needed.

Perhaps the only thing that is needed is to ensure the cycleway on the northern side is wide enough for bi-directional movement for those who do use the current (sub-standard) shared path.

Tamaki-Ngapipi Concept - George

As a reminder, here is what AT propose

Tamaki-Ngapipi Intersection Consent Plan

I guess the focus of the resource consent AT are currently pursuing is the reclamation and as such I hope the actual design of the intersection can improve within that new wider footprint. Regardless it would be good to have a high quality design right from the get go.

So what do you say AT, how about going back to the drawing board and pursuing an idea like this as the base option.

Share this


  1. If Reader George is an engineer who cycles, I would like him to be taken seriously. Rumours are that AT folk aren’t really cyclists which goes some way to explaining the less than perfect solutions they often propose. And hand drawings are far superior too!

  2. Thanks for sharing my thoughts on this intersection, Matt. It’s been great getting feedback from others to help make iterative improvements since posting this last night, particularly regarding th two way cycling on the northern side which was, in hindsight, an obvious omission!

  3. George’s design actually seems to use less space and is primarily a result of switching the high-speed slip-lanes turning left out of Ngapipi into slower speed 90 degree corners. This surely only has benefit in terms of slowing traffic down at a dangerous intersection when they have the green.

    Great piece of simplifying design.

  4. Why do cyclists going straight ahead have to leave carriageway as they approach intersection? Seems very inconvenient.

    1. Thanks for the question. Cyclists aren’t required to use cycle facilities in New Zealand so would not have to leave the carriageway. The purpose of the deviation away from the carriageway is to make the crossing of the side street as narrow as possible while still accommodating a corner radius for left turning vehicles. On the northern side of the intersection this has the added bonus of allowing cyclists to skip the traffic lights. Eventually the goal would be to have the full route with a buffer area for cyclists rather than just the intersection, in that case the cycle lane will not have to bulb out at the intersection.

  5. I’m missing something here – the number of cyclists I see using the shared path on the city bound side can be counted on one hand in a week. Most are the performance types or more confident commuters who will ride on the road. Wouldn’t it be better to have a decided dedicated two way cycle path on one side and a decent footpath on the other?

    1. I think that we are better to actually keep the cycle lanes on both sides right through to the marina so that cyclists into the cbd can then cross with the cycle traffic heading to the CBD on the Tamaki to Glen Innes cycleway. This intersection would still look the same though, even if you were to move east bound cyclists to the northern side of the road at this intersection.

  6. To simplify and cost reduce why not turn the harbour side road lane east bound into a bidirectional cycle lane, just like in Quay St, and widen the current intersection slightly to allow a right turn lane in Ngapipi and a straight ahead lane toward Okahu Bay? Yes I know it will be single land east bound but mostly that lane is taken up as a car park anyway. You then avoid the need to control Ngapipi Rd with traffic lights that I can guarantee will cause major traffic issues because Auckland can’t do traffic lit intersections that work and do away with the need for a traffic island.

    1. Because west bound cyclists would have to cross the road at some point to get to this lane, and then cross back again when the lane suddenly transforms into a shared path of poor quality. If it were a quality bi-directional cycle path properly separated from pedestrians for the entire length of Tamaki Drive then that might work.

  7. George’s plan looks good to me, especially as a Ngapipi Rd/Tamaki Drive to CBD rider. With the Glen Innes to CBD cycleway being progressively opened that route is only going to get busier.

    The AT proposal looks very unfriendly to cyclists – who is going to be able to get to the right-hand lane to turn right into Ngapipi?!

    1. Bryce hits the nail on the head here: to accommodate a hundred or so semi-trailers a day we simply put up a sign saying ” heavy vehicles use both lanes” with a picture of it. We don’t need to kill cyclists to accommodate truck turning circles.

        1. The AT proposal has the limit lines further back on the slip lanes, so how would the ‘Dutch’ example improve what is already proposed.
          You need to remember it is not only trucks that would struggle with a 90 degree turn but also buses.

        2. Fair point Bryce but they/we manage at thousands of other intersections every day in areas far busier for cycles and pedestrians.

        3. Sure but, as I often say about H&S at work “if we can do better, we should do better”. Removing crossing phases for cyclists removes more partial roadblocks for cycling in general. Which is what the Dutch and Danes have done.

      1. I would be careful with that approach. Yesterday the experts here told me the designer would be flayed if there was a crash on an intersection they had drawn. I dont believe them but maybe you do. Even if you could convince AT not to worry about an articulated truck cutting across and squashing a car on a regional arterial intersection near a port, you still have the problem that it just looks too tight, even if the truck uses both lanes. That is why designers try not to use minimum radii, the swept path gets very wide.

        1. Truck swept paths are really easy to model using auto-turn or similar software. Either it works with acceptable clearances, or it doesn’t…

      2. As a matter of interest what would your proposed sign show? “Don’t use left lane or a truck might kill you” Or maybe a symbolic picture of a car squashed by a truck cartoon style? Or maybe if you want to reduce risk as far as ‘practicable’ as suggested yesterday you might want to address the bit on page 30 of Austroads part 4 that says “the design vehicle should be able to turn (left or right) from a marked lane without crossing adjacent marked lanes.”
        At the very least you probably need a good reason to tell the investigator why you thought that wasn’t relevant here. Lucky for you the safety auditors would point if out before it ever got built.

        1. Funny that the crescent roundabouts in west Auckland got away with it then, and the Forrest Hill Road/Wairau Road intersection? The other sort of obvious option is to ban semi-trucks if you don’t want to make the compromise. The difference is that I am prioritising measures that will save lives by protecting vulnerable road users rather than ensuring that cars won’t have to wait behind a turning truck.

        2. That is a fair question. The simple answer is they have a different design vehicle. In this case its a Regional Arterial to Regional Arterial which means the left from Ngapipi and right into Ngapipi will be designed for the big truck. I dont know the counts here but you could expect those two turns to be around 5 to 8% heavies and all the other movements almost nothing. The issue of trucks going this way has been a sore point for as long as I remember. It has been studied heaps without much change. The problem is it is quicker to and from much of east Auckland when the motorway is busy and it is always busy. Truck design really hogs space but you dont have much choice. If you get it wrong it will be noticed and show up in the crash numbers, something we all try to avoid. The usual approach is ease the curve so the swept path gets manageable otherwise you end up with incredibly wide lanes. Urban designers love tight kerb radii at intersections but if you have to design for big trucks the whole road ends up wide instead of just the corner. If you are designing for smaller vehicles it is still a good idea to test the big truck beside a car in the next lane. You can’t avoid risk despite what the H&S Act might say, the best you can do is try and provide enough space so people get a fright or a ding but not a serious injury. The other intersections you mention are probably designed for a bus. Local roads will be designed for a small single unit truck using both sides of the road.


          Gotcha, the crescent roundabouts are on Don Buck Road and Fred Taylor Drive. Same classification as Ngapipi.

          I think we need to take a step back though and look at why the two designs are so different. Tamaki Drive (west of here) is the busiest road in Auckland for cycling, and forms a major access route to the CBD. Ideally we would cater for the truck turning radii, and build really good cycling infrastructure, and maintain enough lanes to ensure that traffic can still get to the bottlenecks on either side without additional delay and do it cheaply. Invariably, as traffic engineers, you and I have to make a decision about which of these goals to compromise, and by how much. I have decided that requiring trucks to split two lanes to make the left turn as they rat tun to the port is the better comprimise than building cycling infrastructure that will continue to be a weak point on this route, like Jacobs did. Both of us have made design compromises, the question is then which is the appropriate compromise in this location? To me that is obvious.

        4. Thanks mfwic – that’s a useful reply. I think Sailor Boy’s design improvements could still be made even with the insane radii.

        5. For roundabouts we used to design as a truck and a car when deflection was the main criterion. Designing for two big trucks for a lot of multi-lane roundabouts meant a car could travel straight through a high speed. The the 2009 guide got rid of that deflection test (a mistake in most people’s view) and focused on entry geometry. As a guide it was a dud and was replaced again last year. Most people who have been around a while will still test deflection because the worst thing that can happen is a car not slowing down so the driver clockwise of them might not see them until it is too late. The old way was to reduce approach speed with deflection and match the visibility to that speed. We have had three roundabout guides in less than 10 years so goodness only knows what was used.
          I dont know what was used for the ones you mention, I am just guessing from what I would do. You are right there is a balance or trade off to make. If I was auditing your sketch I would recommend dumping one left lane and just have one wide one painted to narrow it for cars or widen the two because of the truck traffic due to port proximity. Tight turns beside trucks are nasty and have their own crash code.

        6. The thing I meant to say was what you got totally right was to sketch it freehand first. A lot of people go straight to CAD and end up with something that is often bigger than it needed to be. Sketching and then editing means you only add in what is needed to make it work. I still start with butter paper despite all the modern tools. CAD puts your focus onto details too soon.

        7. There are several regular trolls on this blog who post their kneejerk responses to any suggestion that might impede traffic flow, even if it means improving safety, land use, or promoting sustainable, healthy and efficient transport.

          No matter, it’s easy to shrug them off. The tragedy is when stuff like that is posted by someone that appears to be involved in design of infrastructure in Auckland in some sort of professional capacity. That is really worrying. Where are the modern, well educated and informed engineers with vision? That’s who should be building my city.

        8. I don’t always agree with mfwic, and he definitely engages in a bit of trolling here and there, but I think his contributions to the comments section are generally thoughtful and relevant to the topics at hand. Just because someone has a different perspective doesn’t mean they’re an asshole.

          Also, please note that ad hominem attacks are in violation of our user guidelines.

        9. Very true Peter but I think in my case having a different perspective and being an asshole are not mutually exclusive. 🙂

        10. I like to think of mfwic as a useful devil’s advocate. His challenges to my design were not vindictive, and often discussions yield positive results. Being wilfully contrarian isn’t necessarily trolling!

    2. The left turn lane turning into Ngapipi has a much smaller radii. Why does the left hand turning out need a larger one?

  8. Nice. If sometimes there isn’t a ped/cycle crossing demand over Tamaki Drive, you might also be able to have variations on Phases 1 and 3 that allow eastbound Tamaki to also go (Phase 1) or westbound Ngapipi to go (Phase 3). That would probably keep the capacity/queuing/delays people a bit happier.

    The south-east corner (Paritai South Reserve) is probably a physical constraint that requires the whole intersection to go a bit up and left to fit all this in, but it seems do-able. The other key challenge will be getting the vehicle tracking OK for large vehicles using the two left-turn lanes out of Ngapipi (partly mountable corner, perhaps?).

    1. Hi Glen, thanks for the feedback. I agree that your phasing suggestions would be possible, particularly at night, when cycle demand is lower. I think there is enough width in the proposed reclamation that moving the intersection up creates the space needed, but a shift left would also help if it doesn’t interfere with the bridge. I have addressed long vehicle tracking above, using both left turn lanes the turning arc can be full accommodated. There was a little spare width in the proposed reclamation that may make it possible to accommodate the tracking curves exclusively within the central lane on the Ngapipi approach.

  9. Trucks are already banned on this stretch of road between 8pm and 6am, can’t we ban them outright? Why have suboptimal solutions so a few containers can get a shortcut?
    I live closeby and many containers coming this way are empty anyway!

    1. Trucks are not banned between 8pm and 6am they are just discouraged by POA. Trucks going to and from the back of GI and Panmure would go that way anyway.

  10. Good discussion – would be good if the information is captured a feeds into upgrade. May i discuss the role of new safety at work act in all of this since it has been raised already. I’m not aware of anywhere in the new law that states risk must be avoided. Risk should be identified and controlled as far as is reasonably practicable yes, but avoided no. The statement about upstream duties on designers (s). 39 may be correct though.

  11. Firstly, the design of this intersection should not be dictated by a sense of obligation to cater for the desires of rat-running truckers. I’m not aware of any large industrial estates in Orakei or Glendowie that need to use Ngapipi Rd. If they’re going to places further afield for which there are reasonable alternative routes, they should be told to use them.
    Secondly, if long vehicle swept path from Ngapipi left into Tamaki Drive is still a concern, the answer is simple: instead of two left turn lanes as shown in the top-poster’s sketch, have one wider one. Within the basic design that the top-poster suggested there’s plenty of room for one 15m radius left turn lane with the dimensions that mfwic linked to.

      1. When they can but due to the limited hours of POA multicargo and the hours the customers of the transport companies are open, that is not always possible.

        1. it’s mainly empty containers going between Mt wellington and the port. You can tell because they use smaller trucks and trailers that can’t haul much weight. I say take the motorway.

        2. The small trucks and trailers (if you can really tell the difference) like you say are for hauling empties and as there is an empty container park in Morrin Road (normal hours 0600-1600) and various other place nearby that import and export using POA so it makes perfect sense to use the most direct route.

        3. no it doesn’t make sense at all. It takes only 4 minutes more to take the motorway and we shouldn’t prioritize the movement of empty containers in this piece of waterfront.

        4. Yip. Stanley St upgrade and a brand new road to connect to Mt Wellington Hwy were supposed to cater for this.

        5. Its not about prioritising empties it is about moving whatever freight when you can both pick it up and then deliver it. It may only take 4 extra minutes on the motorway in the middle of the night when you have nowhere you can deliver it, any other time it would only be a fraction of the time going that way.

        6. you clearly have never been to ngapipi road. it’s always congested at peak hour so the motorway would be a good choice anytime

    1. Roundabouts are good when there is equal traffic coming from each direction and limited pedestrian and cycle traffic, all of which are not the case here or the current design would not be the problem it is.

  12. One option which could eliminate problem with trucks turning left off Ngapipi, is to have the east-bound Tamaki Drive traffic stop a little further back. Allows the turning traffic to swing wide (effectively sweeping outwards, rather than inwards), but only when it needs to. Simple enough to work.

Leave a Reply

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