Thursday 7 April 2011

State of the Art Bikeway Design, or is it?

Please note that this blog post has caused considerable confusion. It does not recommend a design, and the sketch presented is not supposed to be a guide for other countries to use to design their traffic lights. The design shown here has in some circles become known as "a standard Dutch junction" but was not the intention of this blog post and is a myth. Other blog posts show the multitude of different designs of traffic light junctions which are really used in an average Dutch city.


An association of transportation experts of 15 major US cities (NACTO) recently published new guidelines for bicycle infrastructure. They claim they are ‘innovative’ and ‘state of the art’ and based on ‘an extensive survey of expert knowledge, [and] existing guidelines from countries and cities around the world’. Some US planners do indeed visit the Netherlands to look at Dutch cycling infrastructure as David reported in yesterday’s post. But looking at the new NACTO guidelines we doubt they have too.


Just focusing at track widths we read on the NACTO website: "desired width for a cycle track should be 5 feet. In areas with high bicyclist volumes or uphill sections, the desired width should be 7 feet". This is actually very narrow, 5' = 1.5 m and 7' = 2.1 m. The standard width for one way cycle paths in the Netherlands is a minimum of 2.5 m ( 8'). Wider ones are not uncommon. For bidirectional use the minimum is 3.5 m (11 '), but most modern cycle paths are 4 m (13 ') or more. Although Dutch sources like CROW are quoted as references the Dutch standards were certainly not adopted.

The biggest problems with these guidelines lie in the intersection designs. For instance, NACTO states "typical international best practice is a two-stage turn". We couldn’t disagree more! The shown queuing boxes are a terrible solution. They not only slow cyclists down but put them in a very dangerous position in the middle of the junction where cyclists have to wait while motorised traffic passes on all sides. This is something that you will never see implemented in the Netherlands!

NACTO bike lane / turn lane design

The advised construction of ‘bike lane / turn lane’ is a way to maximize conflict between cyclists going straight on and drivers turning right. Again, this is something you very rarely see in the Netherlands. This type of design was tried, tested and deemed undesirable. The Dutch stopped building lanes like that a long time ago. A few do still exist (I know just one remaining junction approach like that in Utrecht) but they are phased out as soon as possible. Junctions like that seem more usual in Denmark and David has already criticised one such junction in an earlier post.

So what then is the Dutch solution for the junction approach? Where is a Dutch cyclist positioned on a junction and how do the Dutch create a safe left turn? The Dutch standard junction design solves all those issues at once. So you can ask: would this solution at all be possible in other countries? We believe it is and with the help of the NACTO drawings including their advised widths of car turn lanes we were able to create an animation of a Dutch style junction in the US situation.


If anything, this animation makes clear the space is there! But what's far more important: this type of junction eliminates conflict in turning and crossing movements far better than the advised solutions. So we question where NACTO looked for this "European best practice" which is actually nothing at all like what is implemented in any city in the Netherlands.

Standard Dutch turning lane / bike lane design

However, of course "Europe" is not one place, and to talk of copying "Europe" is rather meaningless. No other country has the same standards as the Netherlands does, nor does any other country have the same participation in cycling that the Netherlands does. As David would tell you: "copying 'best practice' from the UK, for instance, would get you no-where at all".

---

7th April update
It has become clear that because details of the timing of traffic lights were omitted in the above post, some aspects of this design are causing confusion to some readers. With this design:
  • Cyclists can always turn right on a red traffic light, and are protected from any interference from motorists as they do so. Motorists cannot make a right turn on red.  Each cycle path is a minimum of 2.5 m wide, and conventionally they will expand in width at busy junctions, so there is space for cyclists to pass each other to make the maneuver.
  • With or without cycling infrastructure, Dutch traffic lights avoid conflict in a way that those in other countries do not. Many traffic lights at a cross-roads in the UK and USA simply have two states. i.e. N->S and S->N are green simultaneously while W->E and E->W are red and vice-versa. Drivers can go straight on, left or right and those approaching in opposite directions will have to cross each others' paths. However, in the Netherlands it is normal for the turns to have their own traffic lights which have different timings so that conflict is avoided.
  • Synchronization with cycle path traffic lights works in the same way, maximumizing throughput while keeping danger at bay. When motorists have a green light for going straight ahead, cyclists also can ride straight ahead without right or left turning motorists having permission to cross their paths. However, when motorists are given a green light for a right turn this is separated in time from the cyclists' straight on green so that conflict is avoided.
  • You may sometimes have to wait twice to make a left turn. However, you don't have to wait at all to make a right turn. On average, this cancels out and cyclists are not disadvantaged.
  • At the other popular design of crossing, with simultaneous greens for cyclists, who go in all directions at once, you still can make a right turn at any time, and only ever have to wait once to make a left turn. Cyclists then have an advantage over drivers.
In general, the timing of traffic lights does not disadvantage cyclists on the cycle path. In fact, in some instances, cyclists get a green light twice as frequently as drivers do. This is only possible to do if the modes are separated and have their own traffic lights.
---

A note from David
Please note that while this is referred to as a "standard" junction design in Mark's text above, the word was used to mean "unexceptional", not to indicate that this is a standard which is worked towards. Junctions which exactly resemble this design are not actually common in the Netherlands. Only one junction in Assen closely follows this design and that's already planned to be replaced with a more modern and more convenient (for cyclists) design as used in the rest of the city.

A footnote about experts
It appears that several members of the consulting team responsible for the NACTO design standards work for the same company as was responsible for both the lacklustre design standards for Ontario and the inadequate plan for Los Angeles. Better "experts" than this need to be consulted if progress is to be made.

Update 2014: NACTO adopt a similar design
It appears that NACTO took these criticisms on the chin and are adopting a design of junction very similar to Mark's sketch above. However, you must be very careful that it is not over-applied. This design needs to be one of a range of solutions, amongst better solutions, not the solution or the best solution that you have alongside lesser solutions.

28 comments:

Anonymous said...

This is a very important and informative entry, though it leaves me feeling incredibly demoralized.

I've written to NACTO about my dismay over their inadequate guide, and I wish I could do more to broaden their awareness.

As an American, I'm very worried for the future of transport in our nation, especially when our "innovation" is still abysmally regressive.

dr2chase said...

Much like the gorilla walking through the teams-passing-balls video, if you're not expecting it, you won't see it.

Em said...

An excellent explanation and the video makes everything 100% clear. Thank you very much!

Paul Martin said...

Excellent post, Mark, and a fabulous animation showing how it can so easily be done... with the right people in control.

Eli Goldberg said...

I left them this message and am curious what they'll say.

Hi! I'm a big fan of your work since Day 1 (and have been following you on Facebook since your establishment).

As someone who has lived and biked in Holland extensively, I wanted to call out a question raised by David Hembrow on his blog that you say:

"typical international best practice is a two-stage turn".

I have lived in several Dutch cities with 40+% bike mode share, and I have never seen a two-stage intersection. From the pictures on your website, they don't like very inviting (perceived/subjective safety).

Instead, I (and countless other families, children, senior citizens) have used countless intersections of the ones that are beautifully depicted on the website I've provided above.

I realize you are promoting the design patterns that have shown success in American cities. But surely you are aware that even the most bike-friendly American cities have mode shares (~8-14%) that rival the worst Dutch cities that I know of.

I would encourage your organization to revisit more focus on prevailing American patterns that have raised cities from 1% to 5-8% bicycling and instead to look at the design patterns that have enabled 30-50% mode shares.

Thanks again for your great work,

Eli

Frankas said...

jep,
really there are two different systems of left turns. The first you mention - and which NACTO copied - was the German one, and meanwhile it's not applied anymore.
As a German living abroad I realise that in many countries they tend to copy ideas from Germany, specially because it's so well formalised...
And finally the Dutch started very late to offer information in English (in fact they probably do things for themselves, not for others :)

Unknown said...

Excellent video animation, so clear and simply expressed. This should be compulsory viewing for all town planners.

We have these queueing boxes here in Bordeaux, along the inner ring road amongst other places and they are dangerous. Not only do the motorists cut across the cycle path without seeing you, they start to pull into the cycle path before the turn, especially when there is traffic queueing at the lights.

Having ridden in Holland, the system there feels so much safer, in large part because you know that the drivers can see you, as you say.

Edward said...

Thank you so much for this video. It really does show how easily it can be done well. The tragic thing is that authorities in the English-speaking world will still congratulate themselves on NACTO style infrastructure.

Kevjs1982 said...

Following British BEST practice might not be too bad, of course being the UK we can't just have a cross roads it needs to be a roundabout - but I use this one on a regular basis Google Streets which works reasonably well (the 20mph speed limit on the main road to the right also helps)

Mark W. said...

@kevjs1982 that is not bad at all, in fact the Dutch have started building roundabouts everywhere recently, but with a cycle path all around it on which cyclists have priority over all motorised traffic. Believed to be even safer than this junction. It is just as amazing to see a junction being transformed into a roundabout on a space that you would never believe to be big enough.

As for this video. I am glad you all appreciate it. It is of course only a representation of a model junction. No two junctions are the same so in real life there are always slight variations. Goal was to show there is space even if you don't take some from pedestrians and cars. And I don't think there was ever a building removed for cycling infrastructure either, not even in the Netherlands!

Unknown said...

Mark, what a great video. I love the way it relates the Dutch intersection design to the American. I'm a professor in transportation engineering and will recommend that all my students watch it.

I'm sure that the NACTO guide authors did not mean to recommended that cycletracks turn into bike lanes on intersection approaches, because New York City, one of the members contributing to the guide, keeps most of its cycletracks separate. But the way this page of the guide is written, it's easy for a reader to get this understanding. I hope the NACTO guide can correct this soon, because I don't think the intention is to prefer that solution.

The issue about "typical international best practice is a two-stage turn" is a simple misunderstanding. That statement has nothing to do with the bike lane situation described in this post; it refers to left turns. A 2-stage left turn is exactly what you show in your video for a Dutch intersection: first cross one street, then the other. That is in contrast to a 1-stage left turn, in which a bike turns left like a car, from the a left lane. I'm sure you will agree that two-stage left turns are a best practice.

Regarding cycletrack widths, recommended widths are a 2-edged sword. Wide cycletracks are wonderful and should be encouraged. But American engineers tend to take minimum values rididly. If the minimum recommended width is, say, 2.5 m or 8 ft and the space available is less, many cities will simply say, "Then we can't have a cycletrack; there must be only a bike lane." As long as American cities don't consider cycletracks necessary -- as long as they are just an option to be compared with bike lanes -- I believe it's better not to put stringent space requirements on cycletracks, or we'll just end up with a lot more bike lanes.

You can find this same reasoning in the CROW Guide. While the recommended minimum cycletrack width on sheet V19 is 2 m (and 2.5 m if there are 150 bikes/hr) and the minimum verge width on sheet V21 is 0.35 m, the CROW guide also has a section (p. 118 of the English edition) where it tries to show that it's possible to fit a cycletrack into the same space used for a bike lane. There, in trying to make the cycletrack option more competitive with a bike lane, it suggests having a 1.8 m cycle track and a 0.30 m verge.

Anyway, keep up the good work of helping us foreigners understand the wonders of Dutch bikeway design.

David Hembrow said...

Peter: the type of two stage left turn which we both criticize requires cyclists to position themselves in front of other traffic which is waiting to cross their path.

We also criticize the very idea of placing cyclists on the right hand side of right turning vehicles. This doesn't happen with the Dutch example because in nearly all cases the traffic lights are synchronized in such a way that right turning drivers are not given a green light until after the cyclists have cleared the junction and the cyclists light has gone red.

I think it's also worth pointing out that this design allows cyclists always to make a right turn on red in complete safety, with no conflict at all and no waiting for traffic lights. Therefore, even though you sometimes have to wait twice to make a left turn, the average number of waits is still one. If cyclists have to merge with the main stream of traffic, this advantage disappears.

You may also be interested to know that there are increasing numbers of junctions here which actually allow cyclists to make single stage left turns also without any conflict with drivers. This is combined with right turn on red for cyclists, and often also with giving cyclists a green light twice as often as drivers. Overall, this reduces the average waiting times for cyclists to about a quarter as long as they are for drivers.

Anonymous said...

Mark: what is this then?

http://maps.google.co.uk/?ie=UTF8&ll=52.372255,4.89295&spn=0.001094,0.003549&z=19&layer=c&cbll=52.372332,4.892957&panoid=DOFYTcYJbZJLSMsci7KjcQ&cbp=12,14.62,,0,20.48

No cycle lanes at all, cyclists turning left must wait in between vehicle lanes - between the bus and the taxi on the Google image. This is Dam Square in Amsterdam.

Anonymous said...

What about bike boxes? Do you find these anywhere in the Netherlands? I ask because they are also promoted in the US as state of the art Dutch bicycle infrastructure. (perhaps they got this idea from the Crow design book as well)

http://bikeportland.org/cats/infrastructure/bike-boxes

http://www.streetfilms.org/how-to-use-a-bike-box/

David Hembrow said...

Anonymous: What that is is another example of what is not "state of the art bikeway design". The Netherlands is not perfect, and the centre of Amsterdam really isn't the best place to look to find the best infrastructure.

However, though I rarely cycle in Amsterdam and don't know this area, I suspect this also isn't quite what it looks. To me, it has the appearance of a road which is a main route for motor vehicles, but which is not the main route for cyclists. I suspect you'll find North-South cyclists mainly use a route which is a couple of hundred metres west or east of this point.


And bike boxes ? Yes, you do see them in the Netherlands, but not that commonly at least in this area. I can't think of any in Assen, though I do know a few in Groningen. This is also some way away from "state of the art bikeway design". I suspect it was an idea from the 1980s.

Bike boxes are a way of getting cyclists ahead of cars, but not a very effective one. I used them a bit in Britain, where they created conflict, especially in feeder lanes to the bike boxes, which were not a very good place to be should the light turn green and other vehicles start turning across your path, but also when you reach the bike box and find there's an SUV where your bike ought to be.

Clark said...

I'd like to see something like this done in a few intersections here in Vancouver. We're starting to get some very good infrastructure now but a few of the intersections need some work and rethinking.
We have some of these same kind of intersections where the painted bike lane moves over left and cars move right to turn. It's kind of scary.

On the one separated lane with advance turn signals, it's the cars that go first, then the bikes and pedestrians. It seems to work very well.

Re. the NACTO guidelines, one wonders if maybe there's some conspiracy going on to make bike infrastructure so bad that it's doomed to fail. This is less likely than just them not knowing the best way. I'd like to think that they mean well but still need to learn more.

I agree with what has been said about bike boxes. They're too awkward to use some of the time and often have a car sitting in them, either through ignorance on what they're for or intentional anti-cycling hostility that we're now seeing being fomented by trashy "news" media. They're maybe useful in a few instances but not something to use for turning left.

Mark W. said...

@anonymous. That is a picture of Rokin in Amsterdam. That is indeed very bad design. This part of the street has been like that for at least 40 years. No cycle lane at all. So that has little to do with the discussion here, in which we discuss that it is a bad idea to have a cycle lane in the middle of car lanes and sell it as good modern design.

Rokin has been a working area for quite a number of years now. They are building a metro tunnel right under the street. So everything is a bit out of the ordinary. As you can see from looking at the same spot from 50 meters back: it is not so grimm as it appears. When the metro is finished this part of the street will be redesigned too and get cycle lanes or tracks like the rest of the street already has.

Mark W. said...

Bike boxes are not very often seen because bike boxes only work in combination with bike lanes, not with separated bike tracks. The Netherlands has far more bike tracks than bike lanes. They only appear in minor streets and when there is no more than one car lane, with an extra turning lane at most. Bike boxes do have an 1980s feel to them. You will never see them on junctions in major roads because those would have separated cycle paths to begin with and as said, you don't have bike boxes when there are cycle tracks. You never see bike boxes in the middle of a junction as part of a two part turn. And four lanes with a bike box as proposed here... That was certainly not designed after any Dutch example!

Kevin Love said...

What a horrible standard!!

Among other things, it endorses door zone bike lanes. Yes, road design where the most dangerous place on the entire road to ride a bicycle is in the bike lane.

“When placed adjacent to parking, a solid white line marking of four inch width should be used between the parking lane and the bike lane to minimise encroachment of parked cars into the bike lane.”

Note that “parking,” of course, refers to car parking. Bike parking? What’s that?

Or how about this little gem of a bike lane standard:

“The desirable ridable surface adjacent to a street edge or longitudinal joint is 4 feet, with a minimum width of 3 feet. In cities where illegal parking in bike lanes is an concern, 5 foot wide bike lanes may be preferred.”

A three foot wide bike lane? Elbow to elbow, I take three feet just riding. Guess what? There’s also got to be at least 1 1/2 feet of “swerve room” to safely avoid debris or obstacles. A three foot bike lane is a dangerous joke.

And just how is a five foot wide bike lane going to deter illegal car parking in bike lanes?

Instead of wasting large amounts of time, money and resources re-inventing the wheel they could have just used the Dutch CROW bicycle standards. Which, of course, do not allow for lethally dangerous stupidities such as door zone bike lanes.

Neil said...

Though of course as we saw in the recent video of the secondary route, they can have quite a lot of door zone bike lanes.

Anonymous said...

The centre of Amsterdam is indeed not the best place to look for good cycle infrastructure. In fact it is seldom a pleasant place to cycle, because of the volume of traffic relative to the width of the streets. There are no alternative cycle routes of the kind David suggests, because there is no room.

But that raises the question: if the high-quality Dutch cycle infrastructure is not in central Amsterdam, then where is it? Is it specific to certain places? And is that not a relevant factor when exporting 'best practice'?

David Hembrow said...

Neil: I've been "doored" in Britain after I was squeezed by a car on one side and then someone else opened their door. I know exactly what it's like.

However, what you see here is not the same. "Dooring" is not a common concern of Dutch cyclists. The combination of facilities which generally keep you clear of car doors, and drivers looking before they open their doors results in it being so rare that a few months back I actually had to explain to Mark what it was, and why people kept asking about it.

Anonymous: There are plenty of separated cycle paths in Amsterdam, and also plenty of roads which are prioritized for cycling over driving. In the centre you'll find more of the latter than the former, but there are also still several examples of good cycle paths even within 200 m of Dam square.

There are also exceptional things, such as a cycle route from the seaside town of Zandvoort on the west coast all the way to the central station in Amsterdam, coming in through Westerpark, on which you hardly interact with cars at all.

The simple answer to your question is that you will find good quality infrastructure everywhere, except in a few exceptional places. If you cycle only around Dam square you will perhaps be disappointed, but this is not something you need to make much effort to find.

John in NH said...

David and Mark, that is wonderful! I really like the way you all put that design together showing what we could be doing. The big deal is turning lorries (trucks) those making a right turn at the intersection. The current standards in intersection design are set for the largest expected vehicle (I am pretty sure…), banning them in certain areas can be effective though.
If you made the intersection curb separation mountable there would be a severe risk of having a truck/bus smack right into the cyclist, but it would allow trucks to make the turn, the risk might be mitigated by moving the stop line of the other street farther back, like the example of placing the ped crossing between the cycle lane and stop line. It is very doable but I am very much concerned with truck traffic interactions, however I think that allowing the truck to swing further into the intersection and have room to swing back in to the proper lane by moving the car stop line back might be effective... do the Dutch have large 18-wheelers that commonly go through many downtowns here in the US?
Cheers
John

Unknown said...

@ Kevin and others: Rather than trash the new NACTO guide, try a positive attitude for a minute -- be glad that Americans have produced a guidebook recommending European designs that until now had been frowned upon or outlawed in other American guides. If you take the attitude that this might be written by your friends, people who have the same objective as you, you might see more agreement with Dutch guidelines than you think.

Example 1: the "3 foot bike lane" quote is a misreading of the text; that's the minimum uninterrupted surface at the tire level. The Dutch minimum for that dimension (CROW manual) is 1.0 m, essentially the same. NACTO's recommended width of a bike lane is 6 ft, or 1.8 m.

Example 2: The 2-stage left turn. A misreading I've already explained led to a flurry of criticism, when the NACTO guide is actually encouraging a 2-stage left turn as done in NL. Suggestions on improving design details are always welcome, but don't miss the big picture: this is the first American guide recommending that bikes be guided to make 2-stage left turns, rather than being expected to turn left like a car, from the left turn lane.

Example 3, with regard to door zone bike lanes: Are you aware that the NACTO guide is the first American guide that recommends that bike lanes next to parking lanes extend 14.5 ft (4.4 m) from the curb, so that bikes can ride without fear of dooring? That reach is actually greater than what CROW recommends (as it should be, since American cars are so wide!!). And while you may not like it that NACTO also allows narrower bike lanes in a constrained situation, read my earlier post about lane widths: often the choice is a narrow bike lane or no bike lane at all, and if the minimum is set too high, the city engineer will say "No bike lane." Also, to be fair, compare NACTO with CROW. CROW recommends having wide bike lanes and marked buffers, as does NACTO; but CROW doesn't require those generous dimensions. It also allows a 1.5 m bike lane next to a 1.8 m parking bay, for a total reach of only 3.3 m, or 11 ft; NACTO's minimum total reach is 12 ft. Both countries face the same issue: we want wide, safe dimensions, but in older cities where space is constrained, it's sometimes better to have non-ideal dimensions than to have no bike lane at all.

We Americans always have a lot to learn from Dutch bikeway design; but if your hope is that we come to embrace Dutch practice, then I hope you'll see the NACTO guide as a big, though not final, step in the right direction.

John in NH said...

I agree with you Peter, overall I do, I understand your corrections on CROW vs. NACTO, but in a way that is not the point. The point is, why are we not taking what has been proven to work time and time again, taking that and adapting the best we possibly can. We are we taking 20 year old designs and calling them good, why? Are we afraid to work for the best? Are we afraid to show what is the best and allow advocates to push engineers to enact the best, instead of just saying, "nope can’t fit". Why are we reinventing the wheel once again? Yes it is better, but not what should be happening.

Anonymous said...

A research report, on which German "NACTO" is based(Empfehlungen für Radverkehrsnanlagen, from 1995, new in 2010), is: Schnuell, R.; Alrutz, D. et al.: Sicherung des Radverkehrs an staedtischen Knotenpunkten; Forschungsberichte der Bundesanstalt für Straßenwesen, Heft 262; Bergisch Gladbach 1992; German short version in http://www.nationaler-radverkehrsplan.de/eu-bund-laender/eu/brr/brr-037-1993-de.pdf (Bicycle research report No 37, 1993). I haven´t found an english short version in the net. It must be there in some transportation-related libraries.

DG, Hamburg

David Hembrow said...

Anonymous: I've seen this sort of stuff before. It really depends what you measure. German cycle paths are not remotely the same as Dutch cycle paths.

Manouchk said...

Your blog is very interesting. I've seen your video can high interest. I participate of a group in Brazil that is discussing and making divulgation of bicycle for transportation, focussing on developpment of cycle lanes...

Emmanuel M. Favre-Nicolin
Blog Vitória Sustentável
http://vitoria-sustentavel.blogspot.com