Saturday, 21 August 2021

Cycle Tracks In Tight Spaces

One of the issues we sometimes come up against is where we have a road with traffic volumes requiring cycle tracks to enable people to cycle, but where there isn't enough space - what are the options?

Of course, there is always enough space if we are willing to be radical, but for lots of reasons, a road may still be needed to move motor traffic and making wider network changes may be way beyond a project scope or funding. It would be better to be working to detailed plans, but we have to be pragmatic. 

Every place is different and so I can only generalise, but there are some basic dimensions for us to consider. For walking, the minimum footway width given in Inclusive Mobility is 2 metres for most situations, although 1.5 metres is considered an acceptable minimum (1 metre could be used over short sections, but that's all). More space is needed where pedestrian flows are higher, such as outside shops.

For cycle tracks, LTN 1/20 (Table 5.2) has 2 metres as the minimum for with-flow (1-way) cycle tracks for fewer than 200 users an hour at peak, with 1.5 metres being the absolute minimum at constraints. Obviously as flows increase, the space needed increases. For 2-way cycle tracks, 3 metres is the minimum with 2 metres at constraints for fewer than 300 users an hour at peak. 

In the case of walking and cycling, we also need to think about effective (usable) width. People take up physical space and so walking by a vertical highway boundary will have people moving out a bit as they cannot walk right at the rear - there is no guidance, but you'll easily lose 200 or 300 mm. 

Cycling loses up to 500mm where there's a vertical element over 600mm high (LTN 1/20 Table 5.3). There's also the case of having a buffer between people cycling and traffic. LTN 1/20 suggests 0.5 metres (no buffer as the absolute minimum) at 30mph (Table 6.1) and 1 metre (0.5 metre absolute minimum) at 40mph. I'm also assuming that a forgiving kerb is used between the footway and the cycle track to absolutely make best use of the space. Having no buffer also allows stepped tracks to be used which again maximises space and allows overtaking.

We're probably looking at something like the layout above. At this point, the footway and cycle track are both 1.8 metres wide.

For driving, Manual for Streets, gives some suggestions (Figure 7.1). Probably something around 5.5 metres is OK for 2-way general traffic, although on busy HGV and bus routes, 6 metres is probably the operational minimum and perhaps 6.5 metres on bends. Certainly in my experience, different bus operators will have different views and even 6.5 metres gets push-back from some operators.

So, what does the minimum look like? Well, if you squeeze things down, you can fit everything into 11.5 metres, but you'll end up with pretty stingy footways and cycle tracks. Active modes are clearly the most space efficient at 6 metres of the width with driving modes taking 5.5 metres. Of course, if we had 6 metres for general traffic, that's 12 metres.

An operational minimum looks a bit better for walking and cycling which each get 2 metres on each side of a street where we have with-flow cycle tracks (8 metres for active modes) and general traffic gets 6 metres - a total of 14 metres.

This starts to show how much space we need for moving motor vehicles. If we wanted to add a car parking bay, then we'd need another 2 metres. maybe 2.5 metres for a loading bay. This immediately throws up a point that on-street car parking is a luxury we cannot afford in terms of space. What I mean by that is if we are retaining a road as a movement corridor (so we need to protect people cycling from general traffic), then we cannot afford the space for parking. Loading can easily be accommodated by having off-peak loading from the carriageway with or without marked bays - drivers will just have to go around loading vehicles.

There is one more iteration of space efficiency and that's going for 2-way cycle tracks. A 2-way cycle track at 3 metres width isn't bad and would give more space for overtaking. They can be easier to thread through signal-controlled junctions, but they are not easily accessible from the side of the street without a cycle track.

We can go tighter where there are constraints, although I'd bet a small sum of money that it would be the walking and cycling infrastructure that gets squeezed by designers because once we squeeze the space for general traffic, we'll lose two-way running. In fact, this sort of squeezing could actually help create more space for active modes. Let's take our stingy 11.5m corridor and make general traffic one-way.

The one-way traffic lane is 3.5 metres now because we need a bit of space within which to manoeuvre between the kerbs. In this situation, loading is not going to be possible as the road would be blocked, but buses could stop as it would only be for a few seconds. We can grab a little bit of buffer space between traffic and the cycle track, although when used by pedestrians as a refuge space, it is still very tight. Alternatively, we could run with-flow cycle tracks in the same width.

The problem with converting a road to one-way will be where there is a bus route. The usual UK approach is to have bus stops in each direction as it's much more legible for users. In theory, in-bound and out-bound services could be on different streets, but it could mean people walking longer distances to/ from stops. It does depend on how the streets are laid out because some people might be walking a longish distance to/ from a pair of stops and splitting in-bound and out-bound services might actually mean one of the pair of stops is much closer to use and we shouldn't dismiss the idea.

As ever, we are getting into network planning issues because there may not be a road network which supports splitting in-bound and out-bound bus services. Where we have this, then the answer to protecting people cycling is perhaps more of a hybrid solution. If we developed a circulation plan, we could have general traffic running into and out of a town or city centre on different roads, while still running 2-way bus services on both. If we only have bus traffic running in one-direction, we could arrange the cross-section to have people cycling sharing with buses.

This is a 12 metre solution, although mixing with buses in one-direction is a compromise and if the bus direction serves multiple routes, it could quickly get to a point where it feels both really uncomfortable to cycle along and bus passengers get held up too much.

The natural progression from these various solution is to change the street to one where most private motor traffic is removed, but that is maybe a future decision, especially where we are still trying to move freight. Bus gates or sections of bus/ cycle streets can be a solution, but again, these could end up being unsatisfactory for cycling and bus passengers. There is also a consideration on where exactly the space needs to be divided up because within a town centre or city centre, we really should be removing all but essential motor traffic - again, not something a single scheme can usually deliver.

The other issue with what I have been showing is they have a lack of space available for landscaping and planting, so we end up with wall to wall hard surface and as cycling is often being retrofitted, there will be situations where the only space available are existing verges with mature trees which immediately creates a situation where we're ripping out trees for cyclists rather than having the correct discussion about how we take away space for (private) motors.

One cross section to think about is a pretty common arrangement of a 7.3 metre wide carriageway with 1.8 metre wide footways to give 10.9 metres to play with.

Again, having general traffic running one-way opens up space to add cycle tracks and at 1.9 metres wide, it's a compromise, but not the end of the world. 

Even if we started with a 6.5 metre wide carriageway, there is just about an option which can be squeezed in if we use mandatory cycle lanes. We could take the general traffic lane down to 3.1 metres and then have a pair of cycle lanes 1.7 metres wide. Coupled with a 20mph speed limit I think this is getting us to the limits of what we can do, remembering that the traffic lane is still going to be reasonably busy.

With these narrow sections, bus stops are always going to be a challenge because there isn't space to float the passenger waiting areas and we are probably left with boarder style arrangements (below) which I realise are not optimal for everyone and it goes back to network decisions and bus frequency.

Pedestrian crossings are probably a little easier as zebra and signalised crossings can include cycle tracks and cycle lanes in the arrangements (below), although there can be issues with cyclists either not obeying the crossing rules or at least behaving in ways that some pedestrians are unhappy with - that's a wider discussion around how we often treat people cycling as little vehicles, rather than remembering that they are more nimble than they would be driving which opens up different considerations.

Alternatively, we can create shared spaces at crossings which maintain continuity for cycling, but which can create conflict and have signals which present a collision risk (below). Conversely, this type of arrangement might be a little easier to use with a non-standard cycle if the road needs to be crossed so access a cycle track going in the other direction or premises.

Junctions have similar considerations around space. Side roads and private accesses need careful though in tight situations as there isn't space for continuous treatments that could use entrance kerbs or similar if we are going for a stepped track rather than with a buffer. If we have a 2 metre cycle track, we could gently narrow it to 1.5 metres to incorporate entrance kerbs, but detailing becomes a bit more complicated. 

The pragmatic solution is probably to either go for a lower general kerb face and then ramp down to accesses over two kerbs with a low upstand or perhaps use splayed kerbs (below) and require people accessing to have to drive even more slowly. Side streets would need to be flush and so maybe using gentle raised tables for general traffic could be a solution or dropped cycle traffic to carriageway level for a very short distance.

30 splay kerbs used at a private access between the
carriageway and cycle track (vertical elsewhere)
and between the cycle track and the footway.

Signalised junctions are probably a little easier in tight situations and the general road space is more generous. Again, though, decisions are required on the space motor traffic actually needs as we often add lanes for capacity on the approaches. Sticking to one traffic lane gives more space to play with in terms of giving people protection and if the road we are looking at is more about traffic flow in and out of a town or city centre, then banning turns can help, especially if we think about the routing of general traffic at a network level.

It's an interesting challenge to see what we can squeeze into a space, but we really need to be able to think bigger in many cases because the thing that often constrains us is having to maintain two-way operation for motor traffic. Once we can see beyond that, we have all the space we need to enable cycling in many cases. Easy for me to write of course, but the solutions are rarely technical, but often political.


  1. Timid, motor-pandering, effort. Can be improved in several obvious ways:

    0. Cycling is traffic. Instead of saying ‘people cycling and traffic’, say something like ‘modes’.

    1. Stop comparing the width of one mode against two modes artificially inflated by adding them together. Replace with a direct 3-way comparison or 4-way if also includes horsing provision.

    2. Do not have [presumably bi-directional] walkways on both sides. Far too much space wasted on this under-performing and energy-inefficient mode.

    3. With cycleway on one side and walkway on the other, you don't need to be too obsessed with mutual isolation. Preferably surface red and yellow respectively for clarity and declare non-primary mode to be least vulnerable for Highway Code purposes; tactiles on raised walkway only and centre line in semi-lowered cycleway.

    4. Do have compulsory verge/ buffer of minimum 0.3 m at highway boundary and minimum 0.7 m between modes at lowest speed differential. Put practically all of your street furniture in centre of the latter and almost none in the former, limited to paired signage pole with headroom clearance.

    5. Can do at least as well as your ‘limits of what we can do’. 0.3 m + 3 m cycleway + 0.7 m + 3.1 m (one lane) carriageway + 0.7 m + 2 m walkway + 0.3 m = 10.1 m 🤗.

    6. Introduce a distinction between ‘desirable minimum’ and ‘absolute minimum’ for carriageway widths. As arbitrary 0.5 m reductions are foisted on walking and cycling then it's only fair to do the same for motoring.

    7. Greater vertical separation is more effective than horizontal and ought to be maximised where space is limited. Your half-stepped tracks are atrocious in this regard; effectively no separation at all.

    8. A 30° straight access ramp on 0.45 m Weetman Wiggle gives 0.26 m rise—worst for manual wheelchair, but tolerable—enough height to deter some errant motoring. Would comprise ramp kerbs between quadrant kerbs projecting into carriageway and don't care whether interstitial transition kerbs; makes assembly stronger.

    9. Carriageway for motoring should always be ramped up/ down in preference to active mode ways, regardless of escalated scale of engineering necessitated. Applies to grade separation (large deflection) in addition to side roads, always ∼sinusoidal.

    1. Thanks for your comments, but starting a reply being rude doesn't exactly encourage me to bother with a detailed reply. However, the post was about general dimensions rather than analysis of the individual components and how they may influence width.

      On the footway issue, I fundamentally disagree. General traffic lanes should be the thing to go rather than a complete footway. Where you have this (such as Switzerland and Belgium to name two places) it's because nothing has been done to the general traffic arrangements and people on one side of the street have to cross to get to a footway.

    2. Construction activities are directed. To anyone with an engineering mindset, an ‘analysis of the individual components’ would be a pre-requisite for discussing ‘general dimensions’—e.g. a 1.8 m walkway ceases to be so when some highway bodger later bungs a load of obstructions in the middle of it 🤷. Such backwards workflow leads to costly production of a design, consulting on that, then a necessarily [wildly] different detailed design which gets built without further feedback only to require crazily remedial snagging or endless buck-passing and resentment. Alarm bells are ringing 🔔…

      I was not seeking your agreement, having (apparently) misunderstood the purpose of this 'blog to be the polar opposite! Unconvinced you've the position, influence or inclination either way. If you found my comment rude, perhaps you're a bit too egotistic to be working in highway planning or boasting yourself up under the name ‘Ranty Highwayman’? Complaining in the comments that not enough general traffic lanes removed under your own post decrying the further removal of motoring space 🤦.

    3. God I love South Park.

      Thanks for the blog post. I'm part of a small group of Bham based campaigners for more cycle lanes. The information about different widths and configurations is really helpful.

  2. This is a very helpful summary. Lack of road width is the principal reason we don't have more cycle lanes being built.

  3. Great site. Sorry to post anonymously but I am still working as a civil engineer in the transport planning world.
    You might be interested to have a look at the new development around Barking Riverside. They have added generous space for walking and generous space for cycling. But the trouble is that the building line to building line distances then end up being around 30 metres. So the overall impression is one of Los Angeles rather than Amsterdam.

    1. Barking Riverside is quite odd and they certainly could have gone denser