This post is part of this series on learnings from the Dutch CROW design manual for bicycle traffic.

In this part of the series looking a road junction design as perfected by the Dutch, we’re going to look at what to do when a right of way junction or a roundabout won’t cut it and we need to control the traffic with lights.

Our table of provision shows us that once a roundabout is no longer suitable due to traffic volume, our remaining options are the delightfully named traffic control system (TCS) aka traffic lights, or grade separation aka bridges and tunnels.

First we must set the scene for the UK, we have different rules regarding light phasing than the Netherlands and these will have to be taken into account.

• In the UK, pedestrian light phases must be separate from vehicle movements, so a pedestrian green can not be shown if vehicles have a green to turn across the pedestrians unless another set of red lights are used around the turn in effect creating a separate tucan crossing. In Holland, pedestrians can be given a green phase along with turning vehicles which must give way (zebra markings are used), cycleways are never treated this way (always given a red if motor vehicles have a green) unless they cross the turning away from the main junction.
• In the UK we have repeater lights on the opposite side of the junction or otherwise beyond the stop line. These lead to encroachment by vehicles across the stop line. The Dutch don’t have these, instead the stop line is set back from the lights to keep the lights in vision.
• In the UK, in contrast to the pedestrian rule above, a green light does not give you priority over motor traffic coming from another direction, it simply gives you the right to proceed, when turning you are often required to give way to oncoming traffic. The Dutch system of Sustainable Safety does not allow such sub-conflicts, and so a green light is tied more closely to a specific traffic movement (going straight on or turning), meaning a green light to go in a specific direction means that there will be no other traffic to come into conflict with.

If we want safe light controlled junctions for everyone, we should be looking seriously at this final point. So with this in mind, let’s look at two typical “blueprint” junctions and how the Dutch would design them.

T-junction

A cross road junction is a very common type of intersection, but the CROW manual makes an interesting observation about them; each entrance to the junction has three directions the traffic can travel. If instead, the cross roads can be turning into two T-junctions, then we reduce the number of potential movements and thus the number of potential conflicts.

There’s obviously not always the space to spread out a cross roads in to a pair of t-junctions, but there is sometimes scope to remove an arm of the cross roads redirecting traffic a different way or limiting permitted turns.

So let’s look at a typical t-junction and how traffic movements of different modes can be isolated in space and time.

Each mode (pedestrian, cyclist, motorist) has it’s own dedicated space. On larger distributor or through roads, the cycleways might be bi-directional and only on one side of the road and the pavement may only exist on one side also.

Possible conflicts within motor traffic is removed by having dedicated lanes for each direction and by not having stacking lanes which vanish straight after the junction.

Dedicated lanes per direction allow for the separation of turning movement in time, meaning that each phase of the lights caters to a single direction of travel for all modes.

When the right turn phase is happening, motor vehicles, cycles and pedestrians from the main road can turn right, while motor vehicles, cycles and pedestrians on the side road can turn left.

So we have three phases, each of which incorporate all modes of travel and so provide an equal share of time to each mode as well as keep each mode physically and temporally separated.

Separate cycle tracks cause concern from many groups of people mainly due to the perceived conflicts between modes that they can create. By looking at how the Dutch design such cycle tracks to travel throughout points of conflict, we can see that actually rather than create conflict, they alleviate it as long as they are treated as part of the overall traffic flow with equal priority as motor traffic.

The only remaining conflict which does cause a worry for people who have not seen such schemes working in the flesh is the conflict between cyclists and pedestrians at the junction crossings. As always, the devil is in the details of how the pavement and cycleway are delineated, something of which a drawing cannot show.

Other than going to Holland to see for yourself, a quick peek at Google street view is enough to see how the cycleway is treated exactly like a mini-roadway with small curb and a distinctive tarmac colouring to visually and tactically separate it from the pavement.

People usually quote the northern provinces of the Netherlands for their cycling infrastructure, but the quality is the same across the whole country. Above is a t-junction on the small and narrow inner ring-road in Amsterdam, somewhere I visited for the first time last year.

Cross road junction

So when a light controlled crossroads can not be avoided, the same principles as for the t-junction apply but we need a little more space and add a fourth phase to the light sequence.

Again, every mode and every direction gets it’s own space. If we look at the phases we can see how all movements can be catered for in four phases without disadvantaging any mode.

The second phase is a reverse of the first, allowing east/west movement rather than north/south.

The third phase is the turning phase. Motor traffic from the north/south turn right while traffic from east/west turns left.

The fourth and final phase is the opposite of the third phase for turns in the other directions. Detector loops can of course be used to detect the presence of motor vehicles and cycles to skip turning phases.

Unfortunately cycles and pedestrians are prohibited from turning right in a single phase and must use the two straight on phases one after the other to form a right turn. The order of the phases can be optimised to reduce waiting time for the most popular right turn direction.

One draw back is that this junction scheme requires three motor traffic lanes per entrance arm so as to separate movements in space and time which is not always possible (although with only one exit lane per arm, this is not actually as much of a problem as you’d think). An alternative is to restrict movement in a particular direction to save on required waiting space, or to combine the straight on and right turn lanes and the corresponding phases.

The overall throughput of the junction is reduced and is not suitable for routes that have a busy single straight ahead direction.

With the arrangement of standard TCS junctions in place, the final junction style we have to look at is the TCS turbo roundabout which is the grand-daddy of them all and which we’ll look at in the next post.