CORONAVIRUS DISEASE (COVID-19) SERVICES UPDATE: Find out about services open under Alert Level 2. Please note, we are currently experiencing higher than normal volumes of work so please be patient as our teams work at reduced capacity. More information on our services
Roundabouts are the safest form of intersection control for motor vehicle occupants. Numerous studies have shown that, in general, fewer casualty crashes involving only motor vehicles occur at roundabouts than at intersections controlled by traffic signals, stop, or give-way signs. Because roundabouts generally involve slow crash speeds, the forces exerted on people inside motor vehicles involved in crashes at roundabouts are generally below the thresholds at which serious injury is likely to happen, as per the safe systems approach.
However, injury crash rates for cyclists at roundabouts are typically higher than at other intersection types. Furthermore, the safe system threshold impact speed for pedestrians, cyclists and motorcyclists is 30 km/h, which is easily exceeded at most roundabouts.
Multi-lane roundabouts and roundabouts with significant speed differentials between motor vehicles and cyclists pose the greatest risks to cyclists and many people will not feel comfortable cycling in these locations.
The most common cyclist crash at a roundabout involves a circulating cyclist struck by an entering driver who failed to see the cyclist. Motorcyclists are also prone to this, but to a lesser extent. Approaching drivers have a tendency to look for conflicting motor vehicles; ie they look to the location from where another vehicle would arrive at the conflict point at the same time as themselves. Thus drivers may overlook people on bikes who, because they travel more slowly, are already much closer to the conflict point.
Despite these considerations, European research and best practice reveals that roundabouts can be designed to be consistent with safe system principles, so they can be safer and more attractive for pedestrians, cyclists and motorcyclists.
Design factors that contribute to cyclist safety at roundabouts are:
entry speed – the faster the entry speed, the less time drivers have to scan for conflicting cyclists, and the greater the severity of any crash that does occur. Lowering entry (negotiation, and exit) speeds is the focus of designing roundabouts that are safe for all users, including pedestrians and cyclists. See the ‘Motor vehicle entry speed’ under Concept design considerations section below.
visibility – if drivers approaching a roundabout are able to assess gaps in traffic too early, they are less likely to see a cyclist, and less likely to slow down. See the ‘Visibility’ under Concept design considerations section below.
When developing options for cycle routes on corridors with roundabouts, refer to the roundabout information in the Planning section. Where suitable, safe and attractive provision for cyclists, pedestrians and motorcycles at urban roundabouts is not feasible, alternative routes or even alternative forms of control should be considered.
In rural locations, pedestrian and cyclist safety and levels of service are less likely to be of primary concern, but design should aim to provide at least a basic level of safe provision for the expected range of users.
There are a few legal considerations regarding some specific types of roundabouts designed to accommodate cyclists. They include aspects not yet included in rules such as the signs at C-roundabouts necessary to indicate that heavy vehicles should straddle lanes, use of sharrows, and markings that show a transition from a cycle lane to a central position.
Where cycle routes pass through roundabouts there are several options for designing a facility that is appropriate for people on bikes. All options mitigate risks to some extent, however each has disadvantages. The design should be developed in the context, while considering all the types of users expected to pass through the intersection.
A number of aspects should be considered together when providing for cyclists at roundabouts. There is not necessarily a set order in which to consider these, rather an iterative approach may be necessary. These aspects, which are detailed in the following sub-sections, are:
‘Every street is a cycling street’. In other words, people who ride should be able to access all parts of the road network with at least a basic level of safely (except where cycling is specifically prohibited and alternatives are provided). In particular, where roundabouts form part of a cycle route, it is necessary to provide a level of service appropriate to the intended target users.
Methods of cycling through roundabouts
On the roadway
Cycle lanes should not be used at roundabouts, as they put cyclists in a less safe position. Where cyclists use the roadway at roundabouts, the engineering treatments employed should encourage cyclists to ‘claim’, ‘defend’, or ‘take’ the lane. This involves people cycling in a central position within the lane, similar to that of a motor vehicle, so that motor vehicles and cyclists stay in single file in the lane. It is less safe at a roundabout for a person on a bike to keep left and allow motor vehicles to pass them, due to the potential for conflict between different trajectories of users entering, circulating and exiting. The practice of riding in the middle of the lane places people on bikes where they are most visible to motorists and also slows motorists to cycling speed, improving safety. It is not appropriate to expect people to claim the lane on busy or high speed roundabouts. Many cyclists may be willing to take the lane provided traffic speeds are sufficiently slow (preferably 25 km/h) and traffic is not too busy to permit cyclists to merge into gaps. Guidance suggests that this is appropriate for roundabout legs with up to 8,000 vehicles per day (i.e. in both directions, giving 4,000 vehicles per day for the approach direction). For design guidance for single lane roundabouts that are appropriate for cyclists remaining on the roadway see Compact roundabouts.
On paths or separated cycleways
Many people, especially some of those termed interested but concerned (as per Geller’s typology), may not be comfortable remaining on the roadway, except on quiet neighbourhood roundabouts. So at more exposed locations, with higher traffic volumes and speeds, and especially for multi-lane roundabouts peripheral paths and grade separation should be considered in the design – see the Off-road alternatives section for more information on the design considerations. Cyclists using peripheral paths have to cross the exits and entries at facilities shared with pedestrians. While this is much safer for cyclists at busy and faster roundabouts, it is difficult for cyclists and pedestrians to judge driver intentions at exits, and there may be long delays finding a suitable gap. For these reasons at some busy, lower speed urban roundabouts, zebra crossings on platforms have been used.
Grade separation of walking and cycling facilities at multi-lane roundabouts can provide a high level of service for all users. There are examples on highways in Hamilton and Christchurch.
Number of lanes
In general, only single lane roundabouts can be cycling friendly. Roundabouts with single lanes for the entry, circulation and exit are safer and simpler for all road users to negotiate than roundabouts with multiple lanes. They reduce the confusion for people cycling as to where best to cycle and reduce the likelihood of a motorist overlooking someone on a bike. They also minimise the potential points of conflict between people on bikes and motor vehicles, especially if the cyclist ‘takes the lane’ (see Methods of cycling through roundabouts).
Roundabouts with dual lanes may be designed to be safer for confident cyclists, but are never an attractive facility for any type of cyclist. Even for roundabouts with multiple circulating lanes, it may not be necessary to provide multiple lanes on all approaches or exits, and it may be possible to reduce the number of circulating lanes on some sections of the roundabout. As a general rule, a single lane exit should be preceded by a single circulating lane.
Motor vehicle speeds
The importance of managing entry speeds for all types of roundabouts to safe system limits cannot be over-emphasised. While the design guidance in Austroads Guide to Road Design part 4B Roundabouts (Austroads 2015) regarding speed management on approaches appears to focus more on higher speed, rural environments, the principles are equally applicable to lower speed environments. Austroads Guide to Road Design Part 4B advises that urban roundabouts used by pedestrians and cyclists should be designed to achieve entry speeds of 25-30km/h, however Austroads guidance does not yet contain any details about design treatments that help to achieve these speeds.
Lower speeds entering a roundabout increase the time a driver has to process the information they are presented with, and therefore the likelihood that a motorist will see a circulating cyclist and therefore improve safety. Lower entry speeds have been shown to reduce crashes for all roundabout users in many studies, as summarised in Turner et al. (2009). The average entry speeds at the New Zealand roundabout approaches studied by Turner et al. range from about 13 km/h to 40 km/h, with a median speed of about 25 km/h. (Note: roundabout design uses the ‘desired speed’ based on the 85th percentile which is up to 10 km/h higher). Higher entry speed was significantly correlated with higher visibility on the approaches and larger central islands. Previous Swedish research by Brude and Larsson summarised in Turner et al. (2009) showed that the slowest speeds happened when the central island radius was between 10m and 20m. While larger roundabouts allow higher circulating speeds it should be noted that smaller ones may have insufficient deflection and therefore encourage higher entry speeds.
As already mentioned, lower speeds also reduce the crash forces and hence likelihood of serious injury.
Achieving equitable speeds (i.e. a low differential between entering motor vehicle speeds and cyclist speeds) for a roundabout will assist people to be more comfortable to share the lane. By ensuring all approaches are appropriately slowed, both entering and circulating speeds should be controlled and therefore more predictable. Entry speeds are more critical than exits speeds because the main crash type is between entering motor vehicles and circulating cyclists. However exit speeds are important for pedestrians and cyclists crossing the exits.
A method for measuring the negotiation speed of an existing roundabout is given in section 4.4.4 of ‘Improved multi-lane roundabout designs for urban areas’ NZ Transport Agency research report 476. This uses the 85th percentile of speeds of vehicles travelling straight-through (over the 20m distance past the limit line) whilst unimpeded by preceding traffic (ie having a headway of at least 4 seconds).
Entry speeds may be controlled by a combination of effects as detailed in the Austroads Guide, including horizontal geometry, approach widths and visibility. Speed control platforms may also be considered.
The main horizontal geometry features that control motor vehicle speeds are entry path curvature and circulating path radius. Of these, entry curvature is more important and is the measure emphasised in Austroads Guide to Road Design Part 4B. If the entry path radius is much greater than the desirable, the differential between the entry radius and the radius of the circulating path around the central island is more likely to cause vehicles to lose control when travelling around the roundabout.
The geometric elements that result in sufficient horizontal curvature are determined by the size of the roundabout (especially the central island) and the location and angle of the entry. When the central island is too small the straight through movement does not have to deviate much and therefore does not have to slow down much. When the central island is too large, the circulating traffic can travel too quickly. Although every location is different and should be designed to the local conditions, there appears to be a ‘sweet range’ of about 10–15 m for the central island radius of single lane roundabouts.
The type of entry treatment that is most effective in slowing approaching vehicles uses a straight approach that is perpendicular to the roundabout island circle, intersecting like a T-intersection with a tightened left turn (ie a small radius on the entry curve). This is widely referred to as ‘radial’ design. The extent to which a pure radial design can be achieved in practice depends on the size of vehicles to be accommodated.
The width of the entry lane also influences approach speeds. Narrow entries also encourage cyclists to claim the lane, and motorists not to overtake them. However, when combined with tighter kerb radii, narrow entries limit the space available for heavy vehicles to negotiate the roundabout and often some form of widening needs to be considered. Austroads guidance also suggests that sufficient width (5.5m) is desirable to provide room for passing a broken down vehicle. These issues may be dealt with by providing mountable islands or aprons that provide for these movements at low speed, however such areas should also be designed so that they are not an attractive option for drivers of regular vehicles seeking a straighter, faster alignment through the roundabout.
Some effective two-lane roundabout designs use narrow lanes alongside smaller entry radii and may be considered when providing a basic level of service for on-road cycling. See ‘Dual-lane roundabouts’.
Visibility across roundabout entries is an important design parameter. As the intended entry speed for an approach increases, the amount of visibility required also increases. However, visibility also affects entry speed. Approach speeds can be reduced by restricting the distance from the roundabout entry at which visibility to conflicting traffic is achieved.
Austroads Guide to Road Design – part 4B: Roundabouts(external link) gives guidance on three categories of visibility for drivers approaching roundabouts. It should be noted that visibility criterion 1 (view of the roundabout entry from the approaching road) and criterion 2 (visibility from near the limit line of approaching traffic) are considered mandatory because of their impact on safety. However, there is optional advice on criterion 3 visibility. This criterion measures a driver’s view of other traffic from further away on the approach to the roundabout. Satisfying this criterion enables approaching drivers to look for other traffic while still some distance from it, and to judge whether they will need to slow down to give way at the roundabout, or whether they will have unimpeded passage. Although more visibility is often considered to be safer, there is good evidence to suggest that in practice at crossroads and at roundabouts, restricting the visibility until the driver is closer to the roundabout is safer.
Research and experience both show that that limiting the distance from the limit line where visibility is achieved reduces vehicle approach speeds. Also, if an entering driver looks for conflicting traffic when closer to the limit line, then it is more likely that a circulating cyclist or motorcyclist will be seen. Such reduced visibility and approach speeds were shown to reduce crashes for all urban roundabout users. In a study to develop crash prediction models reported in ‘Roundabout crash prediction models’ (NZ Transport Agency research report 386) , Turner et al. (2009) showed that the amount of visibility measured 10 metres back from the limit line was highly correlated with approach speed.
Achieving such a restriction is, however, less straightforward in practice, particularly for larger diameter layouts. Practitioners need to consider the observation heights of all users (car drivers at 1.1m and truck drivers at 2.4m) in assessing this parameter.
It is important that any visibility restriction applies equally to each approach as consistently as practicable to create an even and predictable speed environment on each approach for the benefit of all users.
Inter-visibility between approaching drivers and pedestrian crossing points should always be considered.
In lower speed environments (speed limit 50km/h or below), where entry speeds need to be reduced even further, platforms can be used to create vertical deflection. They should be designed so that they do not introduce a new safety risk. Platforms may be useful for other reasons than just speed control such as providing the base for a raised zebra crossings and cycle crossings. If speeds below 30 km/h are achieved, zebra crossings may even cross two entry lanes with relative safety (see Platform design). Platforms should not be used in higher speed environments.
While the most recent Austroads Guide to Road Design Part 4B: Roundabouts (2015)(external link) does not include marking cycle lanes in roundabouts, however many designers may still be familiar with the guidance given in earlier version of Part 4B which did recommends marking cycle lanes in roundabouts. In New Zealand, cycle lanes should not be marked inside roundabouts. The position held in New Zealand and the most recent Part 4B is in line with the research titled ‘Assessment of the Effectiveness of On-road Bicycle Lanes at Roundabouts in Australia and New Zealand’ Austroads research report (AP-R461-14)(external link) which identified that providing cycle lanes on the approaches to and within roundabouts puts cyclists in a dangerous position.
Forcing cyclists to the left of vehicles inside the roundabout results in more chance for conflict between circulating cyclists and entering (or, to a lesser extent, exiting vehicles), as the cyclists’ intended path of travel is not necessarily obvious to motorists, and motorists are more likely to look for other cars rather than cyclists. This is especially critical for multi-lane roundabouts where ‘looked but failed to see’ crashes can be common.
Instead of marking cycle lanes within roundabouts, cycle lanes on approaches to roundabouts should be terminated at least 30m from the limit lines, to allow cyclists to transition into the general traffic lanes or onto an off-road alternative path where provided. This arrangement is shown in MOTSAM Part 2, Section 3.18 [PDF, 6 MB].
In line with this, no other on-road cycle facilities (eg advanced stop lines, advanced stop boxes or hook turn boxes) should be marked at roundabouts. The only exception is signalised roundabouts, where advanced storage facilities (boxes or lines) for cyclists may be appropriate.
Cycle lanes can be marked on the left hand side of the departure legs of roundabouts. This makes it clear to cyclists where they should be heading when they exit the roundabout as illustrated in MOTSAM Part 2, Section 3.18 [PDF, 6 MB].
Sharrow markings are a useful tool to communicate to both cyclists and motorists that cyclists are expected to share the general traffic lane by taking the same position in the lane as a motor vehicle would (‘vehicular cycling’) to prevent motor vehicles from passing them, including on the approaches to single-lane, low-volume low-speed roundabouts.
There are no specific signage requirements for cyclists at roundabouts.