Sunday 29 January 2017

Kerb Your Enthusiasm: Stepped Cycle Tracks

I blogged about kerbs over three years ago and this week, I return to the subject (albeit briefly).

In truth, this is more of an update on kerbs and their use with stepped cycle tracks (from a UK perspective). Believe it or not, there have been some UK kerb developments, but I think we are still missing tools which would really make things easy (which the Dutch already have been using for years).

You might want to get up to speed by reading my previous post, but the key reasons we use kerbs are as follows;
  • Retain the edge of the top layers of a pavement (I use this in the structural sense of a carriageway, footway, cycle track etc, rather than the often used substitution of footway),
  • A demarcation between different areas or uses of a highway - the obvious here is a kerb between a carriageway and footway,
  • To provide a check or channel for surface water management,
  • To provide restraint to prevent vehicles leaving the carriageway
In designing cycle tracks, these principles hold, because we are ultimately building little roads for cycles and so we need to use kerbs as part of that process.

Using kerbs as restraint to prevent vehicles leaving the carriageway is clearly very important as is providing demarcation. The way in this can be achieved varies considerably, but in general, it leads to the position that kerbs next to motor traffic are there to stop (or at least discourage) encroachment into cycling space and that kerbs use to demarcate space from pedestrians should be forgiving.

Stopping encroachment by traffic is a matter of degree, and depends on the height of the kerb face presented to the carriageway (as well as the amount of lateral space between the carriageway and the track). The usual type of kerb found edging a carriageway in the UK will be half-battered (HB2 type) with a nominal "face" (or height) of 100 to 125mm;

As common with UK kerbs, the unit is 915mm long (3 feet). The overall height of the kerb is 255mm, it's 125mm wide and on the face, the top section is battered back from the vertical by around 12.5 degrees. If you used this type of kerb by the carriageway to support a stepped cycle track, there is no particular issue until we get to driveways, where we would drop the kerb height down to 25mm (using a transition kerb and a bull-nosed kerb; a bull-nosed kerb is square rather than battered with a rounded "nose" at the traffic side).

The traditional way of doing this in the UK leads to the path dipping to meet the low kerb. I've explained ways of dealing with this issue for footways before and the same techniques could be employed with cycle tracks;

The image above shows the typical way we have dealt with transitions to driveways on the left. If this is a cycle track, it leads to a very uncomfortable experience, especially the further right we get. For people using tricycles, the approach is highly likely to tip them over. The image on the right shows a buffer strip within which the transition is made, but in which the track is kept with a nice and constant shallow crossfall (1 in 50 works well for drainage); this pitfall can be seen on this example from Leicester below;

The buffer zone can be as wide as we like - indeed, if it is a couple of metres, then it can be grassed and planted with trees which gives lots of protection from the traffic and in the event anyone falls off their cycle, it's into a verge and not live traffic. Such a wide buffer might not be possible and really, anything getting down to half a metre would need a different approach using quadrant kerbs.

This is all well and good, but relies on being a bit clever with UK kerbs. What we really need is a UK version of the Dutch "inritbanden" kerb to provide a transition from carriageway to stepped track;

The photo above is not of a stepped track, but the kerbs I'm interested in is the row by the carriageway edge which ramp up to footway level. It would be very simple to make these to UK dimensions, but they do of course still rely on using a narrow buffer within which to lay them.

In those tight areas, we might feel that the buffer is a luxury we can do without in terms of maximising track width. In these cases, if we are trying to keep the track nice and level, we have two options. First, we can keep the track at one level and use ramped kerbs from the carriageway to the track as used here in Cambridge;

The kerb next to the carriageway is called the "Cambridge Kerb" and is made by Aggregate Industries. The kerb is about 35mm high and is sloped. From a cycling point of view, it's easy to ride up and down at a shallow approach and so it's forgiving. It's also easy for drivers to do the same thing and so it is less useful in preventing incursion.

The other way to do it is with standard kerbs, but keeping the kerb face lower than the standard 100-125mm and dealing with transitions over a longer distance than the usual single (915mm) kerb. A bull-nosed kerb with a nominal 65mm face gives a reasonable amount of protection from moving traffic and if we transition up and down to the 25mm kerb at driveways over two kerbs (1.8m), we get very gentle level changes. Even close to the kerb, the ride is good and the change in level negligible, even on a tricycle;

It's hard to make out, but above is a track with a constant level at the rear and a low (65mm) bull-nosed kerb at the front. Transitions from the higher bull-nosed kerb to the 25mm at driveways occurs over 1.8m. It's a compromise, but works well. It's also rather complicated to design and so needs a bit of effort to get it to work!

With stepped tracks, there is always a possibility of drivers getting onto them. The layout using bull-nosed kerbs is less likely to be encroached on under normal traffic speed, but at low speeds, drivers can mount the track reasonably easily, especially at the driveways. This can be useful in the event they need to move out of the way to let an emergency vehicle through, but it is very tempting for people delivering to park there.

So much for the carriageway side, what about the other side of a stepped track? We are starting to see some fairly decent cycle tracks being built in the UK, but we often struggle where they are next to a footway. We cannot treat cycles like motor vehicles here and provide a vertical kerb to prevent incursion onto the footway. Well, we could, but high kerbs (over 60mm) will catch pedals (as on the photo below) and low vertical kerbs can grab wheels.

We've often deal with the demarcation between cycle tracks and footways with a special 200mm by 200mm block with a raised centre section of up to 20mm (with sloping sides) to act as a tactile warning to visually impaired people (see the photo below). Unfortunately, it's use often goes hand in hand with acres of tactile paving which is a pain for people walking and cycling;

The advantage of this demarcation block is that people using cycles for mobility purposes (such as a hand-cycle) can easily bump over the block to leave the cycle track (at a shop for example). Personally, I prefer a step down to the cycle track as it helps define clear space.

The photo above shows a stepped track, but the kerb at the rear is vertical and so can grab wheels and can be difficult for people to bump over, so we need something better.

The photo above shows a Dutch cycle track and the demarcation kerb has a slope of about 30 degrees and so meets the objective of being forgiving and can be bumped over if needed.

In the UK, we don't have this profile, we have a full-battered or splay kerb with a 45 degree slope;

It's clearly more forgiving than a vertical kerb, but with a 75mm face, it's still a pedal catcher. I don't recommend bouncing up and down such a kerb too much, it's still too steep, but it's fairly forgiving. Some have used this type of kerb, but buried it deeper to get a lower face;

The trouble is, this is still a bit of a wheel grabber because the angled area is just too narrow. The Cambridge kerb would provide a perfect solution here, but it's only available as a full-height (255mm) kerb which is over the top for cycle traffic and (I understand) requires a minimum order of 1000 metres! A half-height Cambridge kerb (150mm) would be the idea solution.

However, we do have another kerb we can play with and it's off the shelf. Going back to the half-battered kerb I mentioned earlier, it is available in different sizes, including a 150mm high variant (HB3);

It has the same profile as the full height kerb, but it has been shrunk to give less on the vertical part on the bottom. This profile is more usually used on bridges where the half-battered profile is wanted, but where there is no need for a deep kerb (as the surfacing goes onto the bridge deck rather than having a thick road construction). We don't use it like this for cycle tracks, we need to apply some lateral thinking;

By turning the kerb through 90 degrees and laying on its back, we get a gentle and very forgiving slope with an overall upstand (or step down) of about 35mm;

One disadvantage with using this kerb in this way is that because the elements are straight, we cannot lay it to tight radii. Actually, this is a distinct advantage as it stops designers proposing stupidly tight corners! You can just about get away with an internal radius of 8 metres and so it keeps the track flowing nicely.

When we put a decent and forgiving kerb at the rear of a stepped track and the best traffic deterring kerb we can at the front (with a buffer if possible) then I think we can provide some pretty decent infrastructure. The other advantage of stepped tracks is that we can take space from the carriageway by building up layers, rather than digging down and this is great because it generally avoids the need to get involved in the costs and effort of dealing with buried utilities. Therefore, they represent a pretty good way to retrofit our streets. But, details are everything, including how we use the humble kerb.


  1. Why would you even want people to be bumping up and down the kerbs in the first place? Good design is providing a decent infrastructure to use and shouldn't be promoting a choice. They certainly don't try to promote choice in the Netherlands. Its the same reason there should not be an ASL on a road next to a cycle track.

    1. Exactly, we should just build stuff that works; the stepped nature of the track to footway gives some cues to visually impaired people (and everyone), but we need a gentle kerb to make the track work for all.

  2. My only concerns with these low heights comes from reading this document; Effective Kerb Heights for Blind and Partially Sighted People - Research Commissioned by The Guide Dogs for the Blind Association (Guide Dogs).
    The results say "All the participants detected the kerbs of 60mm, 80mm and 120mm when stepping up or stepping down from the kerb as well as when approached straight on and at an oblique angle. One participant consistently failed to detect the 50mm kerb when stepping down. With heights of 40mm and lower, some participants failed to detect the kerb when stepping up or down."

    With the conclusion being "This research has identified the effectiveness of each kerb height in an internal controlled environment. For confidence that a kerb is detectable by blind and partially sighted people, it is recommended to install a kerb of 60mm or greater. This applies to kerb profiles approaching vertical and any profile that is significantly different from this would need to be tested. The effects of lighting, weather, additional cognitive loading, and the practicality of using the kerb to assist navigation were not considered in this trial."

    Of course tactiles will help identify crossing places, but lack of adequate upstand may contribute to partially-sighted peds not being able to identify when they are actually in a place of danger.

    Any thoughts?

    Andy R.

    1. It's a quandry. A 60mm kerb has us in pedal strike territory and we don't have a forgiving kerb to give this face together with a gentle slope to be useful to wheelchair users.

      However, given the 12 to 20mm upstand on a delineator block, the 35mm upstand in my example is better. The key here (I think) is that the greatest risk to VI people is when they meet a kerb at right angles.

      It's not a great photo, but ion this example, a VI person walking left to right from the side road would meet an upstand of at least 60mm

      This is possible because the track drops to road level and then back up again as it crosses the junction. Once clear of the side road, the kerb swaps to a forgiving one and the risk to a VI person is similar to if we had used a demarcation block without the step.

      In this stepped track example, we also have a 65mm kerb face on the track edge, although at driveways, we have 25mm, but is that different to situations without cycle tracks (i.e. most situations).

  3. A good useful article. I think a lot of bad designs come about just because designers don't know any better and don't have time to learn or to innovate.

    When laid on it's side, the radiused corner faces the cycle track. Would you lay the surface on the cycle track to the top of this radius or the bottom (looks that way in the photo)? At the top, there is only a 20-25mm level difference between the two surfaces. At the bottom, there is an additional 15mm upstand that aids recognition of the different surfaces, but might catch skinnier bike tyres. (I'm sure you've tested it out!)

    It's a pretty good solution using off-the-shelf components, but when laid on it's side, the kerb is slightly higher than the top layers of most footway/cycleway constructions, but not deep enough to match the whole construction depth. Would that make it a pain to install on site, and how might you avoid that? Or is that something for the contractor to sort out?

    1. The issue is that we were using a kerb not really designed for what we were using for. There will always be a slight chamfer because the concrete would stick to the mould and fret away.

      The radius is less that it would be for a bull-nosed kerb as we are dealing with about 15 degrees of a circle rather than 90. Laying the surfacing tight to the bottom of the radius seems fine - I've not tested with skinny wheels yet, but fine for my 700c hybrid and 20"folder.

      In terms of the installation, the depth (now it's on its side) is 125mm at the back and about 90mm at the front (about a 35mm upstand). For surfacing, it means 60mm binder course with 30mm surface course works a treat (using 10mm AC10 or similar).

      At the rear, 125mm is enough to get some decent backing and still allow the concrete kept low enough for surfacing/ paving behind. It does take some skill by the contractor and the gang who build my example have done a couple of schemes now. Don't forget, most of the kerb is buried, so it seems to be keeping stable.

    2. Actually, I messed up that description, the radius is less than a quarter of a circle compared to a bull nosed!

  4. Thanks for both these articles.

    As I understand it you recommend kerbs are laid flush between cycle lane and a side road (and also flush between the cycle lane and carriageway at side road?).

    At a driveway things are slightly different, with a 25mm step between the carriageway and the cycle lane and preferably a very forgiving kerb on the inside.

    A report of an accident has drawn our attention to this as a cyclist came off and was injured joining a cycle track from the road at the entrance to a small industrial estate. There is a bull-nosed kerb with a step of 25-35mm. The cyclist was on a road (racing) bike with small tyres and approached it at a shallow angle not realising there was a vertical lip, and came off.

    Reading this I understand the balance between keeping vehicles out and making the transitions forgiving.

    I wondering if at wider entrances (more side street than a driveway) a flush or nearly flush transition might be more appropriate?

    I think cyclists may be expecting similar treatment to a side road due to the geometry and are more likely to approach it at a shallower angle too.

    Kind regards,

    Waltham Forest Cycling Campaign

    1. Have you the location - it's easier to understand in context; there are lots of examples of junction type accesses which use a dropped kerb with an upstand and I don't know why. If a cyclist is turning onto their driveway, the of course know about the upstand, but if turning into an access as you describe then it's a trap - definitely needs to be flush. Of course, if there's a cycle track crossing such an access then the problem is avoided!

    2. In this case the tracks had started a short distance before, the cyclist had missed the entrance to the tracks as they weren't very familiar with the road. It's fairly easy to do as the previous section of road doesn't have a cycle lane leading to the track (it's narrow and a bus route). The stepped track starts after a side road and so it's easy to miss as you're checking for vehicles emerging.

      The previous entrance seems to be quite flush and the next one has more of a lip on the entrance.

    3. Actually, added to that there's a short shared section just before, which some cyclists will prefer to avoid if going at speed, meaning they would enter the cycle track at that entrance.

    4. Slightly further on there's a side road with a very slight lip too, although it's generally smaller than for the entrance.