This guest post is by Dr Mike Lloyd, a NZ academic who contacted me after I featured his article on the recent MTB bike rage incident that was caught on video and went viral. This post remains the one of the most popular BCC posts. His follow-up article examined ‘the spatial, temporal and interactional order of a rare case of cycle rage’ and looked at the same incident from a videography analysis to uncover the details of a MTB track run ‘gone wrong’. Both are well worth the read! It an absolute pleasure to present Mike’s first guest post – we hope to be hearing more from him – Enjoy! NG.

The Road Ahead: Research on the ‘sharrow’ in cycling infrastructure

(A summary of research by Dr Mike Lloyd, Max Baddeley, and Dr Ben Snyder, School of Social & Cultural Studies, Victoria University of Wellington, New Zealand; the full paper is currently under submission with an academic journal)

‘Sharrows’ first appeared in California in the early 2000s and have now been officially mandated for use on roads in many countries including Australia and New Zealand.   Our research looked at new cycling infrastructure in Wellington, New Zealand, specifically a 450 metre stretch of road where the designers stopped Copenhagen-style cycle lanes and reverted to more standard road space marked with sharrows (short for shared lane arrow).  Here is what a sharrow looks like in this space.

The sharrow is the white cycle with double arrow sign painted in the middle of the red area (indicating a 30kph zone).  For the SUV driver, in this context the sign can mean, ‘be alert for cyclists ahead, and share the road if you come across them’.  This is consistent with aspects of sharrow use in Australia where they are referred to as a ‘Bicycle Awareness Zone’.  This also conforms to the New Zealand ‘best practice’ guidelines which say the sharrow ‘helps reinforce that the carriageway is a valid place for cyclists to travel (reinforcing to other road users to act accordingly’).  The interesting question is ‘where exactly should the cyclist ride?’  Looking at the bottom panel, we can see that there seems to be some guidance in this regard: the widening green bars seem to direct the cyclist to move into the middle of the road, exactly where the sharrows signs are positioned.

A quick googling of ‘sharrow’ would confirm this, as phrases like occupy- claim- or take-the lane will crop up.  However, this is not a hard-and-fast guideline for how a sharrowed area should be ridden.  Traffic experts emphasise that the sharrow is there to help the cyclist occupy the traffic lane when it is safe and appropriate to do so.  So, key questions are: ‘when’ should a cyclist claim the lane, and in doing so ‘where’ exactly should they position themselves in the lane?  Further, any cyclist will also know that it is not only the road ahead they need to be concerned with, but the road behind.  If a cyclist occupes the lane with cars behind, how will those drivers react?  A first way into this issue is to look at the process of attempting to claim the lane in the transition from the Copenhagen-style cycle lanes to the sharrow area.  This is where the widening green bars come into play, but as we see they are no guarantee of success:

Panels 1 to 3 show a cyclist attempting to claim the lane, but in response the driver of the white car speeds up not allowing the cyclist ahead.  In contrast, panels 4 to 6 show success: as the cyclist moves out the driver of the green car backs off, allowing the cyclist to ride ahead claiming the lane.

We do not have a breakdown of the ratio of success to failure in claiming the lane as this was not our goal, however, it is worth noting that our research involved one of the researchers riding to claim the lane; in reality, it is rare to see other cyclists doing so.  Mostly, cyclists revert to a default line to the left of centre.  This is a pity, because as we rode we discovered that sharrows can work to make cycling safer in traffic spaces where cycle lanes are not present.   Here the effect of raised pedestrian crossings and four speed bumps within the sharrow area was signficant. The entrance of both ends of the sharrow area features raised pedestrian crossings, and whereas a cyclist can ride over these with little decrease in pace, vehicles slow to a greater degree, thus giving the cyclist a chance to maintain a lead ahead of vehicles.  When the vehicles get over the raised crossing and increase their speed, catching up with a cyclist, the speed bumps repeatedly give the cyclist a ‘breathing space’.  Of course, this all depends on the speed being travelled: this ‘breathing space’ effect works for a car obeying the 30 kph limit, but not for one travelling significantly over this speed.  Needless to say, not everyone obeys speed limits.

Also, once claiming the lane, there can be a reduction in the temptation to ‘filter’. When there are parked cars to the left, but the central line of traffic is slow or stopped, it is very common for cyclists to ‘filter’ between the cars –  a dangerous area to be cycling in.  Our cyclist’s impression was that once riding to ‘claim the lane’, when the traffic slows, filtering to the left is not so ‘automatic’, rather the cyclist may just slow their pace to match the vehicles ahead, thus reducing the risk of riding in the dangerous space between cars.

This good news has to be tempered though by one of the main findings of the research.  This has to do with how difficult it is to predict in the design of cycling infrastructure how drivers and cyclists will actually interact on the built road.  Small details can be remarkably important, yet hard to plan for.  We were able to realise this because of our dual camera research method, that is, our cyclist had a GoPro camera pointing forward on his bike handlebars, and a rearwards facing camera mounted on his helmet.  The folllowing three visuals capture a near-dooring incident.

In panel 1 the cyclist is riding past three parked cars and a motorbike, and just as he is adjacent to the motorbike, the door on the silver car starts to open.  The video record does not allow us to be definitive, nevertheless, in our view two things can be noted.  First, the opening of the door is a continuous movement (see panels 3 and 4), and second, from a careful scrutiny of panel 4, the car driver is looking forward, not behind or to the right where the cyclist is approaching.  It does not seem either that the driver is looking into a rear-view mirror to check for any vehicle or cyclist behind, nevertheless, we certainly accept that this could be the case.

Interestingly, the inability to be definitive on this point is not of crucial importance, because the more pressing question to ask is, why was the cyclist not aware of the door opening?  This is sensible to ask because, as shown in panel 3 of figure 6, it has opened sufficiently enough for it be visible.  Experienced cyclists develop a strong sense of where they are cycling in relation to parked cars and the potential at any moment for a door to be opened on them, meaning that even a door opening to 10 centimetres is probably detectable. But there is no evidence that the cyclist sees the door opening, as he certainly does not change his line in response to the opening, even though by the time he is directly adjacent to the door it may well have opened even further than seen in panel 4.

In an ‘aha’ moment the answer was provided by consulting the rearwards-facing video record, filmed simultaneously.

In panel 1, the cyclist has entered the sharrow area, claiming the lane with the consequence that the silver MPV behind him slows.  Just after panel 1, the vehicle comes closer but then backs off, continuing to follow at a reasonable distance (panel 2).  Just before panel 1, the cyclist has looked behind and seen the silver MPV, so he is aware of its presence while he rides centrally ahead of it.  Panel 3 provides the answer to the question of why the cyclist was unaware of the car door opening.    At precisely the moment when the car door begins and proceeds to open, the cyclist is looking behind (hence, the tilted screenshot) to see where the silver MPV is in relation to him.  This fully explains the ignorance of the door opening, but, as captured in panels 4 and 5, we now have a much more extensive idea of what happened.  We see that the door was fully opened with the driver emerging onto the road, and we also see how dangerous this situation was.  As indicated by the yellow arrow, the cyclist’s line was directly in the path of the fully opened door.  It was probably only by a matter of micro-seconds that he escaped being doored.

There is more that can be learned from this data, for another pressing question needs answering: if the cyclist was claiming the lane in the sharrow area, why, at this particular point, is he riding a line within the dooring zone?

The answer is available in the subtle change of line prior to the place where the near-dooring occurred.  In panel 1, the cyclist approaches the raised pedestrian crossing riding in the centre of the road, and in panel 2 is seen riding straight over the sharrow sign.  Panels 3 and 4 show though, that just before he gets to the speed bump, there is a subtle alteration in line, taking him leftwards and closer to the line of parked cars. This alteration in line is first due to riding around a manhole cover in the road, which takes the line towards a second cover in the road (at the head of the top arrow), which is also ridden by moving to the left.  These slight alterations in line are continued by riding to the left of the speed bump, the line then maintained towards the circled area ahead (panel 5) where the near-dooring occurs. The cyclist is clearly picking the line of ‘least resistance’ in relationship to the bumpiness of the road, which results in moving him further and further to the left, away from the sharrow line and into the dooring zone.

The subtlety of such alterations in line would be difficult for road designers to predict.  Moreover, other things can happen in the same space that lead the cyclist to a different line.

This second door opening occurs well before the cylist, but is also at a time when the cyclist is riding to the right of the sharrow line, so that he is well clear of the opened door.  This is because the transition from the raised pedestrian crossing to this location has no material objects that encourage alteration in line.   In panel 2 we again see the alteration in line around the cover, but this time when the cyclist gets to the speed bump (panel 3), he rides through the middle.  The reason for this is visible in panel 4: he looks ahead and sees a parked bus taking up significant space in the road, so he anticipates the need to go wider and adjusts his line out more centrally in the road, coincidentally taking him well away from the dooring zone.

So, the exact lines ridden are clearly not solely determined by the material features of the road, rather there is a complex entanglement of the social and material in any particular riding through the sharrow area.  There are patterns in how a sharrow area is ridden, but at the same time these are not sufficient to predict the course of any moment’s riding through this new cycling infrastructure.  The particular line taken in any particular moment is part of a ‘wild phenomena’. To decide on the degree of success of any new cycling infrastructure requires close attention to the detail of how cyclists and drivers actually interact. Thankfully, the availability of cheap and easy-to-use action cameras makes data-gathering relatively simple, leaving the researcher with the difficult task of unpacking the fine detail.  It is an important task that may lead to improved cycling infrastructure design.