Junction malfunction on Hills Road

This article was published in 2018, in Newsletter 137.

Hills Road approaching Queen Edith’s Way.
Image as described adjacent

The Hills Road junction with Queen Edith’s Way (QEW) and Long Road was completed in the autumn of 2017, so we’ve had a few months to assess how it works in practice. I’m sorry to report the experience is poor for cyclists in the evening rush hour, and for much of the day.

On the southbound approach to the junction the Cambridge kerb disappears, the cycle lane forks, a new vehicle lane starts on the left, and the lead-in cycle lane continues between the two car lanes (see diagram). The kerb-side lane is for cars going left and straight ahead, and the right-hand lane is for right-turning vehicles only. Drivers wanting to go straight on (the majority) must cross the lead-in cycle lane to reach the kerb-side lane.

The problem is that the lead-in cycle lane is usually blocked by queuing cars in the evening, halting the progress of cyclists. Worse still, when the lights turn green, traffic sets off as cyclists are approaching the lane split, and drivers cut them up. This is quite dangerous.

Queuing cars block the cycle lane.
Image as described adjacent

Other places in Cambridge also have a similar layout where a lead-in cycle lane is between two car lanes: the top of Hills Road bridge and the Trumpington Road approach to Brooklands Avenue. All are relatively new layouts, but the Hills Road approach to the QEW junction is very different from the other two junctions because straight-on car traffic uses the kerb-side lane, while the other two cases follow the much more typical pattern of reserving the kerb-side lane only for left-turning traffic. In these latter two cases, we rarely have the problem of blocked cycle paths or cars setting off without giving way to cyclists. Why is this?

Having looked at all the junctions, the problem at the QEW junction appears to be caused by the shallow angle of the path of moving cars as they cross over the lead-in cycle lane to queue up in the kerb-side lane. This means each car must take a longer time passing over the cycle lane and is more likely to queue on top of it rather than leaving a gap for cyclists to use. Since the kerb-side lane is used for straight-on traffic, it handles a large volume of traffic and tends to build up queues on every red signal. Compare that to the sharper angle of entry into the kerb-side lane on top of the Hills Road bridge: drivers there have to make a conscious choice to cross the cycle lane and hence do so more considerately and safely.

Trumpington Road: unobstructed.
Image as described adjacent

The Hills Road approach to QEW has a very gradual start to the kerb-side lane; the distance from the end of the Cambridge kerb to the point where the kerb-side lane reaches its full width is at least 37m. Compare this with the Trumpington Road approach to Brooklands Avenue: the same measurement made there is only 13m (see photo above). On the Hills Road Bridge, the kerb-side lane opens up to full width immediately without any transitional approach.

I also find that the use of raised traffic islands instead of painted road markings plays a role. On the Hills Road approach to QEW, the splitting-off point for the kerb-side lane starts off as a cycle lane: the road surface is red for 9m with a cycle symbol on the road (see photo below). These painted cycle symbols are routinely ignored by queuing drivers. This means that cars enter the kerb-side lane before the lane is wide enough to fully accommodate them, and the lead-in cycle lane is blocked. At times of day when there is no congestion, drivers shoot across the cycle lane at speed, which is dangerous and intimidating to people trying to cycle there. By contrast, on the Hills Road bridge a traffic island with bollards prevents cars from queuing in the kerb-side lane too early (see photo above). Similarly, at the Trumpington Road approach to Brooklands Avenue, the short lead-in distance to the kerb-side lane indicates clearly to car drivers that they should wait until there is room to enter the lane. In both cases, drivers must make an explicit lane-change manoeuvre, which seems to be done more carefully.

Nothing prevents drivers from entering the kerb-side lane too soon.
Image as described adjacent
We should learn from this experience, share the knowledge for future designs, and hopefully fix this specific problem

We can test these observations. Set up some cones to eliminate the taper on the approach to QEW, so that the kerb-side lane can only be entered where the lane is already at full car width. The area I propose to cordon off is currently a cycle lane complete with red tarmac and a cycle symbol painted on it, where people shouldn’t be driving anyway. If this pilot programme is a success, then we could consider how to make it permanent.

What are the knock-on effects of this proposal? Effectively it reduces the left-hand queuing space by one car-length (i.e. from 11 cars to 10 cars) but the junction capacity will remain unchanged. At evening rush hour, between 5 and 14 vehicles can use the kerb-side lane in one traffic light cycle. However, owing to queuing on the approach to Addenbrooke’s roundabout, the cars exit the QEW junction slowly and in clumps. The second half of the queuing cars has to wait for those ahead of them to shuffle up and pack more closely before they can exit the junction. There is no rush for queuing cars to get into the kerb-side lane and plenty of time to give way to cyclists.

Top: Hills Road bridge. Below: Proposed experiment using cones to improve safety and protect the cycle lane by eliminating the taper on the approach to Queen Edith’s Way.
Image as described adjacent
Image as described adjacent

The right-hand lane only has 3 or 4 cars queuing at evening rush hour so there is always some space for them. When lights are green for the right-hand lane they are also green for the kerb-side lane, therefore right-turning traffic is unaffected by this proposal. In the morning rush hour this junction approach is opposite the peak flow of inbound traffic. The right-hand lane is much busier, and the kerb-side lane fills up less often, thus queuing behaviour is different and the lead-in cycle lane is blocked less often. Again, car junction capacity would be unaffected by this proposal.

This proposal mostly helps people cycling during the evening peak hour, and probably has little effect in the morning. The scheme delivery team has been responsive and are in discussion about the best solution. The current thinking appears to involve posting a road sign ahead of the junction to explain the layout of the cycle lane between two lanes of motor traffic. Whether this helps we will have to wait and see.

The QEW junction could provide a good case study on how to design layouts with a lead-in cycle lane between two other lanes, especially as it is physically easy to test the theory that a tapered approach encourages cars to ignore the cycle lane. We should learn from this experience, share the knowledge for future designs, and hopefully fix this specific problem. Check out the discussion in our Cyclescape thread 1298.

From the Editor:
The design of this junction with advanced stop lines and lead-in cycle lanes is fairly standard across the UK but we do not regard it as best practice design. The problem is that it is not inclusive design: asking people to cycle between two lanes of motor traffic might be reasonable for confident cyclists but does not enable people who are new to cycling and feeling more vulnerable on the road. That’s why we campaign for better, safer and protected junction designs based on best practices from the Netherlands. With separate and protected space, the problems described in this article could have been completely avoided and the overall experience here would be easier for everybody regardless of their ability.

Tess Jones