Sunday, January 20, 2008

Copenhagen Cycling: Not what it seems...

(Image of Copenhagen vehicle use from BlogNetNews 17May2007)

Patrick McManus says in “The Deer on a Bicycle” that he realized writing humor was easier and more remunerative than doing journalism. He dashed off a bit of comedy in an hour and it paid the same as a well-researched serious piece. Now, here on CycleDog, remuneration or writing for money is a non-issue. But I have to agree that serious writing takes far more time and care. Comedy is easy and I'm fundamentally lazy, so you can understand the undeniable attraction of trying to write humor.

With that said, please understand that the following is decidedly not humor and it took more time than I care to admit. Still, it's good information and I'm happy to share it with all of you...Ed

Many people equate 'bicycle friendly' with a network of bike lanes, paths, or cycle tracks. Sadly, even our own League of American Bicyclists seems to view this as a necessary criteria for conferring BFC status. Cyclists and cycling organizations are loathe to re-examine their underlying assumptions regarding such facilities, and when a study appears which contradicts their cherished assumptions, that study is ignored. Many facilities advocates point to Copenhagen as a Mecca for cyclists and a model for their idea of a modern city that embraces the concepts of multi-modal transportation. Yet this study, a longitudinal examination of crashes both before and after the installation of cycle tracks (Cycle tracks are separated from the roadway and raised slightly above it. Pedestrian sidewalks are then slightly raised above the cycle track.) indicates that such facilities increase the risks to cyclists, pedestrians, and motorists.

Also, I believe that it's desirable to understand the history of the conflicts over bikeways ( I use bikeways in its generic sense that encompasses all types of bicycle facililties) if only to avoid the mistakes made in the past. To that end, I've appended Wikipedia material that outlines some history and discusses the issues in a balanced format. But make no mistake, even the Wikipedia entry says that adding bicycle facilities to roadways increases the number of crashes at intersections. This is well established and references to it go back nearly 60 years.

(From Wikipedia on segregated cycling facilities)

The issue of the safety of segregated cycling facilities has been one of extreme controversy since the 1930s. Since then, the established cycling lobby in the UK and Ireland has taken a critical and measured view of their utility and value[14]. In 1947, in response to official suggestions that cyclists should use cycle-tracks, the CTC adopted a motion expressing determined opposition to cycle paths alongside public roads.[5] In 2007, official claims of safety for cycle tracks provoked a position paper from the umbrella body for UK cyclists' groups stating "Cycle Campaign Network knows of no evidence that cycle facilities and in particular cycle lanes, generally lead to safer conditions for cycling" [15] On the other hand, the proponents of segregated cycling facilities frequently proclaim them as being necessary to the provision of a "safe" cycling environment. However, in many cases their proponents have no established interest or expertise in cycling promotion and may include representatives of various interests such as the motoring/roads lobbies (including automobile clubs), traffic engineers, the environmental lobby, political parties, the health sector etc. In contrast to such claims, reviews of the international literature suggest a predominant finding associating the use of roadside urban segregated cycle paths with increases, some significant, in the rate and severity of car/bicycle collisions [16]. The argument has two sides involving both direct and indirect safety.

Marin county in California did a study that found a 66 percent increase in cycling between 1999 and 2007. The method employed was to simply count the number cyclists riding through a given point over a 2 hour period. (SEE NTTP COUNTY SURVEY REPORT ) I have to wonder if all conditions were controlled, that is, were the surveys conducted in similar wind and weather conditions both in 1999 and 2007. The study does not address this. As an exercise, I loaded the data from Table B-2 into a spreadsheet and removed suspect data that indicated very low counts. The outcome did not substantially change.

66 percent increase in bicycling during commute hour

Mark Prado and Brad Breithaupt

Article Launched: 01/05/2008 11:25:31 PM PST

A county study shows more people are rolling around on bikes these days.

As part of Marin's $25 million federal pilot bike plan, the county took a count of local bicycle and pedestrian trips to create a baseline from which to judge the program.

Overall, there has been a 66 percent increase in weekday commute-hour bicycle traffic and a 33 percent rise in weekend riding, when compared with a similar count conducted in 1999.

Evaluation should always be part of the analysis of the success or failure of any program. There's always the danger of falling into a fallacious argument, one that many facilities advocates make routinely, and that's a post hoc ergo propter hoc argument. The assumption is that if one thing preceded another in time, the first one caused the other. In the case of the Marin study, the increase in bicycle traffic could just as easily be attributed to increases in population, changing economic conditions, or variations in the weather. The linkage is unknown, therefore basing public policy and public spending on dubious information may be a waste of resources.

Allow me to expand on statistics for a moment. If we did a study of the number of cyclists going through a given intersection, it's obvious that time of day, the season of the year, and the weather conditions are going to have a large impact on our observations. Likewise, it's important to have as large a sample as possible because we need to know how how much variability is present. For instance, if we observed an average of 50 cyclists per day over a 30 day period, and that number varied plus or minus 10%, we could be confident of that count. However, if we established that our 50 cyclist average was based on numbers ranging from zero to one hundred, we could not be confident that any one-day count would accurately reflect the average.

Now, before you start thinking I'm splitting hairs over this, consider that it takes a huge amount of money to do an in-depth study over time. A simple traffic count is much simpler. But when we're proposing to use this information for planning, and that planning involves spending large sums of public money, it behooves us to see that the studies are solidly based.


This is a summary of two reports (here and here) on before and after observations of newly installed cycle tracks in Copenhagen. For those unacquainted with the cycle track concept, it's a bicycle lane physically separated from the motor vehicle portion of a roadway. The separation may be a raised curbing or the cycle track itself may be slightly raised above the road.

Some interesting parts of this study include the compilation of crash statistics for the same roadways and intersections both before and after the installation of cycle tracks. This is the reality of cycling safety – cold, hard numbers. But they went further and asked cyclists for their beliefs about their safety, and found that the installation of cycle tracks boosted cyclist's perceived safety, while in reality road safety deteriorated.

There's a risk of overgeneralizing in stating that this is often the case. Local governments and cycling organizations have a vested interest in touting the increase in cycling along newly installed bikeways, while ignoring whether those cyclists were attracted from nearby streets. The Copenhagen study, on the other hand, looked at existing streets and intersections over a long period, and had the advantage of a mature cycling population, one accustomed to daily bicycle use for transportation.

The construction of cycle tracks has resulted in a slight drop in the total number of accidents and injuries on the road sections between junctions of 10% and 4% respectively. At junctions on the other hand, the number of accidents and injuries has risen significantly, by 18%. A decline in road safety at junctions has undoubtedly taken place after the construction of cycle tracks. If the figures for the road sections are combined with those for the junctions, an increase of 9-10% in accidents and injuries has taken place.”

....The increase in injuries due to the construction of cycle tracks arises because there are more injuries to pedestrians, cyclists and moped riders at junctions. There has been an increase of 28%, 22% and 37% respectively for these three road user groups.

The increase in injuries to women was 18%, whereas there was only a small rise in injuries to men, just 1%. The increase in injuries is especially large among females under 20 years of age on foot and bicycle, as well as female pedestrians over the age of 64. On the other hand, there was a considerable fall in injuries among older cyclists and children in cars of both sex.

...From table 1, it can be deduced that the construction of cycle tracks has resulted in three important gains in road safety: fewer accidents in which cars hit or ran over cyclists from the rear, fewer accidents with cyclists turning left and fewer accidents in which cyclists rode into a parked car. These gains were more than outweighed by new safety problems: more accidents in which cyclists rode into other cyclists often when overtaking, more accidents with cars turning right, more accidents in which cars turning left drove into cyclists as well as more accidents between cyclists and pedestrians and exiting or entering bus passengers.

One aspect of the Copenhagen study revolves around questions it doesn't answer. Actually, that's always part of the process. As one question is answered, it leads to another. Why does the installation of cycle paths lead to increased crashes? Could it be that they attract unskilled riders who presumably learn to avoid some of the more critical errors over time? This question is unanswered. As far as I'm aware, there are no long-term studies of crash rates related to cycling facilities over a period of years. I could be wrong about that, so if anyone is aware of such a study, please note it in the comments.

Eliminating parking to install cycle tracks forced motorists to side streets, thereby increasing the numbers of cars turning at any given intersection and increasing the number of collisions between cars, pedestrians, and cyclists as an unintended consequence. Illegally parked cars increased the crash rate further.

The construction of cycle lanes has resulted in an increase in accidents of 5% and 15% more injuries. These increases are not statistically significant.

Yet page 1 says "The amount of data is enormous with more than 8,500 accidents, 1,500 traffic counts and 1,000 interviews investigated and many results are therefore statistically significant." With a sample that large, it's difficult to dismiss the above increases as not statistically significant. Page 2 says "A decline in road safety at junctions has undoubtedly taken place after the construction of cycle tracks. If the figures for the road sections are combined with those for the junctions, an increase of 9-10% in accidents and injuries has taken place." It cannot be both ways. Either crashes declined or they increased. They cannot do both.

(Perceived Risk)

Cyclists feel most secure on roads with cycle tracks and most at risk on roads with mixed traffic. This is true for all cyclists, irrespective of their gender, age, purpose in cycling or familiarity with their route. Figure 3 shows that conditions in mixed traffic create considerably more feelings of risk than conditions on cycle tracks or cycle lanes. Cycle lanes are a middle path so to speak: somewhat less secure than cycle tracks, but considerably more secure and satisfactory than mixed traffic. Increased car traffic leads incidentally to cyclists feeling more at risk.

...increase in entering/exiting bus passengers (where cyclist pass in a lane or cycle track to the right of the bus stop) +1951% for accidents and +1762% for injuries (presumably all types, including both cyclists and pedestrians)

62% of cyclists in the current study on cyclists’ perceived risk answered that in general they feel secure in the traffic of Copenhagen. This is close to the result of the Bicycle Accounts of the Municipality of Copenhagen of 2004, where 58% of cyclists said that they felt safe when cycling in Copenhagen. This figure can be compared to the fact that 87% feel secure on roads with cycle tracks and 86% feel secure at signalised junctions with advanced cycle tracks and blue cycle crossings.

Taken in combination, the cycle tracks and lanes which have been constructed have had positive results as far as traffic volumes and feelings of security go. They have however, had negative effects on road safety. The radical effects on traffic volumes resulting from the construction of cycle tracks will undoubtedly result in gains in health from increased physical activity. These gains are much, much greater than the losses in health resulting from a slight decline in road safety.

While it's probably true that increased physical activity and the accompanying health gains offset the documented decrease in road safety, there is nothing in this report to support such a claim. Making an unsubstantiated claim is highly dubious conclusion in the absence of any supporting documentation. This highlights the fallacy of facilities advocacy, insisting that decreased safety has a positive overall impact. Why not inform people that riding in traffic is both safer and more beneficial to one's health? This report shows that cycle tracks and cycle lanes are merely 'feel good' measures.

How can we use this information? Bike lane apologists will dismiss it out of hand because it does not fit into their carefully defined views. It's hard to argue with a closed mind. But as good citizens and good advocates for cyclists, and firmly believing that we should empower individuals, it's imperative that we insist that these statistics illustrating the fallacy of constructing separated facilities be included in any discussion regarding their construction. Doing so is a service to all bicyclists as well as an exercise in good citizenship. I do not want to see my community building something that reduces road safety and ultimately results in a greater number of injuries.

Labels: ,


Blogger Fritz said...

My family and I spent the day in Marin County yesterday (driving, alas). Marin is at the north end of the Golden Gate Bridge and is the county just north of San Francisco.

Because Marin County is so hilly, steep and criss crossed with valleys and ravines, the roads are narrow and winding. Outside of the central parts of the small cities, I didn't notice much in the way of bike facilities, but I saw literally hundreds of cyclists of all kinds. Most were recreational roadies on their Saturday group rides, but I saw a good number "utility" cyclist as well as tourists and other "regular" people who were Just Riding Around on all kinds of bikes.

One in particular stood out to me -- a teenage girl in blue jeans with a pink backpack going downhill at about 20 mph. The road (California State Highway 1 west of US 101) is way too narrow to share, so she confidently took the absolute center of the lane as a line of tourist traffic kept pace behind her. She pulled over at a wide spot to let traffic pass.

I have no idea what the cyclists who are car apologists would expect of this teen cyclist besides call her "arrogant" and expect her to stay off the road, but I think she did great.

I expect the increase in cycling in Marin County, BTW, is probably more due to an increase in gasoline prices than anything else.

5:12 PM  
Blogger WheelDancer said...

Very interesting read, thanks for the post!

I wonder how much of the increase in accidents can be attributed to less experienced cyclists not yet savvy about riding. If the numbers of cyclists goes up it only stands to follow that the number of inexperienced cyclists is also increasing. I would be interested in a study that included an education piece for chipping away at mistakes of ignorance new cyclists tend to make. I suspect these mistakes might also be a bit higher with the various bike lanes since the comfort level is higher and that higher comfort combined with inexperience isn't likely to help the accident statistics.

8:00 PM  
Blogger tOM said...

1. The accident increase was less than the cycling increase.

2. Accidents decreased between intersections. Now effort can be made to solve intersection problems - or streets can be closed off. Bicycle thruways!

tOM Trottier, VP,

10:44 PM  
Blogger Ed W said...

Here's the problem with simple traffic counts. If a newly installed bike lane, path, cycle track, what-have-you, draws cyclists from adjacent streets, does it really increase cycling in the area or does it merely concentrate the existing cyclists? If we had three parallel streets, A, B, and C, and a new facility were installed along B, we could reasonably assume it would attract cyclists from A and C. To get a better understanding of traffic patterns, we'd have to do a before and after count on all the streets simultaneously. And those counts would have to be smoothed over a period of several days or weeks in order to get a realistic picture unaffected by wind, weather, or other unforeseen factors. Obviously, that's a huge and costly undertaking.

Now, that's not to say that the data is worthless, just that it's incomplete and brings more questions than it answers.

5:46 PM  
Blogger fridrix said...

Your argument is ridiculous. Why don't we just scrap sidewalks too, and have pedestrians walking around in the street, "taking the lane." I want a bike path. I don't want dozens of 4,000-10,000 pound vehicles roaring within a couple of feet of me every minute. I try to reduce my exposure to vehicular traffic.

4:02 PM  
Blogger Ed W said... fear is understandable but misplaced. The crux of this report is that the development of cycle tracks has actually decreased road safety while cyclists believed their safety was enhanced. That's not an argument. It's simply the truth.

If we're to be effective advocates, we must insist on measures that provide real benefits to cyclists rather than imaginary ones.

9:07 PM  
Blogger tOM said...

Increased or decreased safety?

Accidents go up, but less than the increase in cyclists. So is this safer or less safe? I say more safe.


11:00 PM  
Blogger fridrix said...

I'd rather wreck on a bike path than on a busy street.

8:52 AM  
Blogger Lou said...

tOM, you have it wrong. The increase in accidents to cyclists is much higher than the increase in ridership. The 5% and 15% figures are for all types of transportation, automobile, bicycle, moped and pedestrian due to bicycle lanes.

If you look at only the bicycle/moped data, you will notice an increase much more than the increase in cyclists. This is true for both bike lanes and tracks.

11:40 PM  
Blogger tOM Trottier said...

from the pdf:
"The construction of cycle tracks in Copenhagen has resulted in an increase in cycle traffic of 18-20% and a decline in car traffic of 9-10%. The cycle tracks constructed have resulted in increases in accidents and injuries of 9-10% on the reconstructed roads."

But I do think that just constructing tracks or lanes along existing streets in a dense network makes intersections more dangerous. There are several ways to ameliorate this.

The best way is to reduce the number of intersections by closing off some intersecting roads to cars. This mean careful planning and community buyin.

Secondly, bicycle lanes can be routed to the left of right turn lanes.

Thirdly, special intersection treatments. From the pdf:

"It should be noticed that blue cycle crossings, retracted stop lines for cars and pre-green lights for cyclists have been used in only very few places on those streets where cycle tracks have been constructed in Copenhagen in the studied after periods. More extended use of these safety measures would very probably have improved road safety.

What is important is the increase in cycling and reduction in pesky cars.

8:08 AM  

Post a Comment

<< Home