Sunday 30 May 2021

What Is A Dutch Style Roundabout?

Look, I don't want to be that guy, but when I see a UK cycling scheme trumpeting a "Dutch-style" roundabout, it's going to get my attention and I'm going to give it a closer look.

North Tyneside Council is currently consulting on a raft of cycling schemes which are being delivered over the next couple of years using funding from the UK Government's Active Travel Fund (ATF) and the Transforming Cities Fund (TCF). The former is essentially part of the response to Covid where 2020/21's fund was more about pop-up and interim active travel works with 2021/22 being about making things permanent or extending them. The latter closed to applications in 2018 and is about investing in public and sustainable transport in city regions. Both funds are for England only.

The principal problems with the funding are that they are part of the long stop-start tradition with bidding that has made delivering cycling (and walking) infrastructure difficult to plan for. Inevitably, local authorities don;t have certainly in building their own capacity and so lots of projects are delivered using consultants' help (and there I declare a general interest), rather than being able to deliver the core themselves.

I thought it important to set some context here because the fact is that North Tyneside is actually trying to create a better cycling network which is head and shoulders above many English authorities. The council says that it has £1.6m ATF funding and £7m TCF funding, although the ATF money is time-limited to March 2022 (another Government-imposed deadline). Have a look at the plans yourselves and I would encourage locals to respond. For this post, I am going to concentrate on the aspect that caught my eye.

The roundabout in question is on the A191 Rake Lane and forms a junction with Billy Mill Lane and Brookland Terrace (In the consultation it is called the Rake Lane Roundabout). The A191 and Billy Mill Lane are large single carriageways with little frontage development (there are parallel streets operating like service roads). This roundabout is part way along wider plans for the A191 corridor between the New York Road roundabout in the west and the Preston Road North roundabout to the east. 

The general treatment for the A191 is to reduce the speed limit to 40mph west of the Rake Lane Roundabout (New York Way) and then provide cycle lanes on the carriageway with wand or orca protection. To the east of the roundabout the speed limit is already 30mph, but the same lanes with wands/ orcas are proposed (there is already an eastbound advisory cycle lane). The cycle lanes pass in front of bus stops and there are also long gaps in the protection at junctions.

There are also parallel zebra crossings proposed (generally where refuges are removed) and a couple provide access to the North Tyneside General Hospital which is a good place to serve. Unfortunately, the westbound cycle lane gives up at the hospital entrance (below).

From the plans at least, the cycle lanes appear to be advisory which is immediately an issue because drivers are allowed in them from time to time - this is in direct conflict with putting in the light protection. These should be mandatory lanes. But it's academic because the A191 looks pretty much terrifying for most people and I really can't see how paint and a few wands or orcas will make people fee safe. This is especially true of New York Way which will have a 40mph speed limit (below).

New York Way is about 10m (maybe a touch narrower). If the minimum cycle lane width of 2m is provided (6.4.2 in LTN1/20), that leaves traffic lanes of 3m which are not appropriate on a 40mph road like this. I suspect these lanes will be 1.5m which will feel very exposed. The road also has little frontage access and so I wonder who will be using it, especially as the light protection ends at the west end.

Figure 4.1 from LTN1/20 (above) is useful in this context. For a 40mph limit with cycle lanes, we get a pink box which states;

"Provision suitable for few people and will exclude most potential users and/ or have safety concerns"

Light segregation gets us into the amber area which states;

"Provision not suitable for all people and will exclude some potential users and/ or have safety concerns"

For my mind, the distance between wands/ orcas on the plans isn't particularly dense and I simply cannot see how most people would be reassured. I wouldn't cycle on New York Way full stop.

For the eastern side of the corridor, we can use the same figure to look at what happens with a 30mph limit. For light segregation at all traffic flows we are into the green, but again, I would still have concerns about the level of protection provided. This is an A-road corridor with a character that is not built up and urban; it is suburban fringe and this shouts cycle tracks to me. Some sections have the service road which could be properly filtered as a kind of cycle street too.

OK, that's a fair bit to talk about before getting to the roundabout design, but the wider network context needed to be considered. The roundabout itself is is very large at around 75 metres diameter (it's a bit oval) to the carriageway edge of the circulatory area. There are two wide traffic lanes around it and each of the 4 entry roads are single carriageways which splay out to 2-lane entries. The exits are single lane, but very wide and the arrangement means that drivers are able to drive quickly into and out of the roundabout. 

The roundabout dates back from the end of the 1960's/ early 1970's. Looking at Old Maps you can see Billy Mill Lane being widened and straightened on the 1966/69 OS map and by the 1971/77 OS map, the roundabout exists. You can look around to see housing development in the general area from the early 1950s through to the 1970s - a growing suburbia with road building - it seems nothing changes.

For the cycling scheme, there are two proposals for the roundabout. I'll look at Option 2 first;

This option has mandatory cycle lanes on the two Rake Lane approaches and the Billy Mill Lane approach (which of course makes more sense that the advisory lanes on the wider plans). Brookland Terrace has advisory lanes. In all cases, the cycle lanes end at the roundabout entrances.

Each arm receives parallel zebra crossings. Brookland Terrance is single stage and the others are 2-stage given the wide islands which makes sense. The centre of the circulatory area is hatched, although the single traffic lane is still wide. The existing kerb lines appear to be retained, but the cycle lanes entering the roundabout and some other hatching means there are single traffic lane entrances and exits.

For people cycling through the junction, they can either stay on the road to mix with traffic or somehow get to the shared-use areas at the zebra crossings (I assume by a dropped kerb which is accessed sideways) and navigate via 2-way cycle tracks. Because of the shared-use areas by the crossings, there is lots of tactile paving to contend with.

2-way cycle tracks in the context of a very large roundabout actually makes some sense. For example, turning right from Billy Mill Lane means you can cycle anti-clockwise around the junction via 4 parallel crossings at a distance of about 120 metres. If you did the same in a clockwise fashion, this would be much twice the distance using 3 crossings. 

Essentially the design adds cycle tracks to the existing footways around the junction and adds parallel zebra crossings. But, it is classic dual provision for the "confident" and "cautious" cyclists when we should be designing for all (or at least 95% of people). The fact that people have to use cycle lanes (even with light protection) is probably still the reason that most people won't use the new infrastructure.

Option 1 is the headline grabbing "Dutch Style" roundabout which I'll look at in a minute, but first the arrangement;

With this option, the main cycling principle is a one-way orbital cycle track (clockwise) which is created from the outer part of the circulatory area of the roundabout, separated by "flexible surface islands" according to the plan. I don't know what this means, but I assume some sort of bolt-down product.

All of the arms have 2-stage parallel zebra crossings and single traffic lane entries and exits. This is good because people cycling (and walking) only have to deal with one traffic lane at a time and drivers don;t have to worry about what is happening on the other crossing as much as they would with a single stage crossing. This is a good design principle.

The triangular "splitter" islands for each arm have been narrrowed a little and they will extend into what was the outer part of the circulatory area to create a narrower traffic lane on the roundabout and protection for people cycling. Pushing general traffic towards the centre is better than the other design of pushing it to the outer edge simply because it tightens the turn drivers have to make, but it is still about a 70m diameter give or take which really isn't going to slow drivers that much.

Looking closer at the layout (Brookland Terrace and the eastern arm of Rake Lane in this case), we can see how new splitter islands have been added which provide buffer for cycle traffic from general traffic on the entry and exit points as well as giving decent pedestrian islands between the cycle tracks and traffic lane crossings. One point to mention here is that parallel zebra crossings should have the zebra stripes and elephant feet (for the cycle size) 400mm apart. This is a stupid rule in the Traffic Signs Regulations & General Directions 2016 which technically stops us curving the cycle part of the crossing. The zebra strips are also wrong as they need to be complete rectangular blocks, but I'll put that down to a drafting error.

The problem with this layout, however, are threefold. First, people cycling and wanting to cross the parallel zebra will need to almost look behind them (more than 90° in any case) for drivers leaving the roundabout which I have shown with green arrows (above) for cyclists and purple arrows for drivers. The crossing of traffic joining the roundabout is not a particular issue because speeds will be lower. 

Second is that drivers will be able to enter the roundabout at speed as shown by the orange arrow. If they can see a gap in traffic, this means the entry speeds are high and with the alignment the exit speeds will be high. As you can see, there is a set back of the cycle crossing from the edge of the roundabout which can fit a car, but it's not enough stopping space at any real speed and so creates shunt risk from traffic on the roundabout.

This is the nub. On the whole, Dutch roundabouts which include people walking and cycling are compact. In the CROW Design Manual for Bicycle Traffic, an urban roundabout would have a diameter of between 25 metres and 40 metres (across the circulatory area) to ensure driver speeds are kept low. With a diameter of maybe 70 metres for the "Dutch Style" roundabout, there is no way that the size is remotely Dutch and in fact, the island at 55m diameter would easily contain most of a Dutch roundabout!

The traffic entries to the roundabout should be perpendicular to the circulatory area and the cycle track should be orbital (and preferably one way) with a Dutch urban roundabout; there isn't a flaring as such, just a radius between the entry lane and circulatory area (above). This means that drivers have to actually use their steering wheels to enter and exit and cyclists cross perpendicularly reducing the need to try and look behind oneself. With the Rake Lane scheme, the cycle track is not orbital, it is more like the rural option which I'll show you in a minute.

The cycle crossing is set back 5 metres which isn't an issue on a compact roundabout as drivers are moving slowly with shunt risk reduced. The crossings are single traffic lanes at a time with a refuge between the two crossings.

The photograph above is the Frederik Hendrikbuurt Roundabout in Amsterdam. It's unusual as it has a tram line running through the centre where everyone gets held by a red traffic signal, but apart from that it operates as an urban Dutch roundabout. you can see the orbital cycle track, the set back to give drivers a stopping place (and space to enter the roundabout - things being broken down by task).

We could also look at the Dutch rural roundabout where people cycling (and walking - it would be a shared-use path) give way to traffic. In this case, there is still tight geometry for traffic, but the crossing points are set much further back to give drivers more space in which to stop if someone crossing misjudges the gap. There is the same perpendicular approach for drivers; no triangular splitter islands making it easier to enter and exit. There is less need to make the cycle track one way and in many cases, there are only cycle tracks around part of the roundabout in any case.

The cycle track is not orbital, it follows the general junction shape which means that cyclists have to slow down and then turn before crossing (above). This does lead to energy losses for the cyclist, but the geometry stops people just continuing across the road as they would do with the parallel zebras of the urban example. 

You'll also see the carriageway drain to the outside which creates an adverse camber (above) which slows drivers down along with an overrun area for large vehicles which keeps space tight for smaller vehicles avoiding the overrun. This will often be the same in urban areas, it's just easier to see in rural examples.

One piece of information I don't have is the traffic flow through the roundabout. The Dutch designs allow for up to 25,000 PCU (passenger car units) a day for single lane entry. The Dutch do use multiple lane entries, but despite there being examples which include cycling, they are much less safe and should be avoided. For UK roundabouts, the geometry which allows higher entry and exit speeds mean we can push more traffic through.

So, does it matter that this scheme is being headlined as a Dutch-style roundabout? In my view, yes. "Style" is doing some very heavy lifting. In fact the only facet of the design which is Dutch is the parallel crossings with single lane entries/ exits configuration, however, without all of the other details it can never be Dutch-style. In fact the only Dutch-style roundabout is in Cambridge which has cycle priority, but also the safety features required to be called Dutch-style.

So what would I do? Well, Given the context, I don't think I would be proposing an urban Dutch design, I would go for the rural example simply because this location is not in an urban location. Assuming I wasn't allowed to dig up the central island because of cost or the trees there, I wouldn't be calling my layout Dutch-style, but I would redesign the approaches to provide rectangular splitter islands (maybe 3 or 4 metres wide) to get the perpendicular approaches for drivers. 

The crossing points would be set back at least 10 metres from the circulatory area, but I would keep the idea of pushing traffic to the inside of the circulatory area with as narrow a traffic lanes as I could get away with, but which is still compatible with HGVs. I think I would also make the cycle track 2-way because with traffic priority, it would be much safer than a 2-way track and zebra. Having it 2-way also helps is provide a 2-way cycle track on New York Way, rather than the scary lightly protected lanes.

Above is a quick sketch to the same scale to show what I mean. To the east is New York Way where I've provided a 3.5m shared-use path with a 2 metre buffer from traffic. The crossing point is set 12 metres from the edge of the circulatory area and traffic is pushed to the inside. you'll note the perpendicular approach which helps slows driver speed into and out of the roundabout. There are lots of options for the other arms. If walking and cycling numbers are expected to reach any significant numbers, then we can provide separate space, but the 2-way nature of the layout allows people to choose the quickest option for them.

My sketch is not "Dutch-style", but it does follow sustainable safety principles where we control traffic speeds and where we don't expect people to cross more than one traffic lane at a time. I think the layout better fits the local road hierarchy and while I understand the desire to provide walking and cycling priority, I am not convinced it is the safest option in the local circumstances.


  1. In your proposal, I'd question why you have the shared path hugging the edge of the carriagway as you proceed round the roundabout. Surely it would be safer, as you proceed clockwise round the bottom of the roundabout, for the path to go relatively straight to link to the path alongside New York Way. A longer spur off this path to the crossing would mean that cyclists would be travelling at right angles to the road for longer, which would make the crossing safer in two ways. Firstly the cyclists would have longer to judge the behaviour of cars coming off the roundabout, without having to look over their shoulders, and secondly the drivers would be certain of the cyclists' intentions for longer, as it would be obvious they would be intending to cross.

    1. My proposal requires cyclists to give way and this means slowing down, turning into the crossing point and being prepared to stop. Sure, we could make things more direct, but we shouldn't be aiming cyclists at the crossing point, we need that turn. Good example here

  2. In your analysis of roundabout 'option 1', you have missed the council blurb about cycling having priority. The comments about motor vehicle entry/exit speeds and the excessive, over the shoulder look by cyclists for clear space are not valid.
    With regards to right of way, 'option 1' is very much in alignment with the Cambridge design.

    The real problem will be communicating this new right of way to the drivers. I foresee some serious collisions taking place, then a change of priority as the council run for cover.

    1. I'm sorry, I disagree. Simply placing a zebra crossing doesn't automatically mean drivers will stop and so we need to use the design to make them stopping more likely. It is critical to use the layout to achieve this and that is why I criticise the layout.

      People cycling still need to be aware of what is going on around them and they still need to see what drivers are doing because until they are on the crossing, priority isn't taken. The rules to one side, again, it's the design which helps the user and having to look almost behind you isn't conducive to comfortable cycling. This Option 1 is only like the Cambridge example because it has parallel crossings, it shares nothing else in common.

    2. "Simply placing a zebra crossing...."

      What is in place at the Cambridge and Dutch examples to make the drivers stop, other than signs and road-markings?

    3. A proper urban Dutch roundabout needs to have all of the things I mention to make it safe. The provision of a zebra is used in urban settings because a decision has been taken to prioritise cycling everywhere. The Cambridge example follows the same list of design requirements to ensure things are kept slow and safe. I'm suggesting that this proposal as it stands does not. I mean, the roundabout is twice as large as the maximum set out in Dutch design guidance.

  3. My impression is that most of these (urban one-lane roundabouts) are 50m or more in diameter... Do you know of any more compact ones that could be models for Oxford?

  4. This roundabout has started construction a few weeks ago.