Wednesday, 2 December 2015

Edmonton standard intersection vs roundabouts

In Edmonton, our roads, have a bit of an issue. A list that spans the distance from here to Mars, no wait, here to Betelgeuse.

Ok, hyperbole over, it still is a problem. I am focusing on how intersections are designed.

It is a pretty standard type of intersection, you can find it all over the city at arterial to arterial road and sometimes smaller intersections. The details vary, but the usual plan is having a left turn lane, fully dedicated to left turns, which can be increased to two left turn lanes at some times. Two lanes per direction separated by a median, with a 3 metre wide multi use pathway on one side of the road and a 1.5 metre wide sidewalk on the other, with the two curbside lanes being extra wide to permit vehicular cycling and a car to overtake within the same lane. There are right turn slip lanes, with zebra crossings for pedestrians. Usually there is during peak hours an 8-15 second protected left turn stage in both directions on a road, followed by through movement and pedestrians along with left turns being allowed but being required to yield to oncoming traffic and pedestrians.

Some of the problems include left turns being allowed to yield and then go, which relies too much on human judgement for a tight situation, and has the risks associated with the yellow light, if you are trying to desperately get through the yellow, both the left turner and an oncoming vehicle would speed.

The slip lanes permit far too high speeds to yield safely at that sort of volume.

The lanes are too wide, and encourages speeding, and having wide curb lanes creates dual networking.

The sidewalk and path are too narrow, they both need to increase by at least .5 metres, with the path possibly having 4 metres, and it lacks a separate sidewalk. Plus you should be allowed to overtake on the paths, prohibited or at least discouraged by the solid yellow line.

Pedestrians lack any place to take refuge in the middle of the street if they cannot go fast enough to cross in one stage.

And bicycles have to dismount in order to cross using the pathways. Also the corner radius is very sharp. You have to slow to a crawl in order to just cross the slip lane.

There are some ways to improve it, like using a new design for the slip lane which increases the sightlines and lowers the speed, Edmonton is using this design on new implementations. You can have protected prohibited signal staging for the left turns, bicycles can have elephants feet markings and bicycle specific signals, the corner radius can be widened for the path, and median refuges can also be added for the pedestrian/bicycle crossings.

But this does not solve the fundamental flaw in the intersection design. Traffic lights should be used sparingly. And divided roads also should be used much more sparingly. Here is what I propose to be the new standard intersection for arterial to arterial (to be renamed distributors for the most part) intersections.

It can be smaller too, with this sort of design:

It has a 3.5-4 metre wide bicycle path, bidirectional, or we could have a minimum of 2 metres, 2.5 metres if at all possible, for a pair of one way, sidewalks, at least 2 metres wide, at least 8 metres of distance from the roundabout for cars, and preferably 10, maybe up to 20, very clear priorities, and best of all, exactly 0 traffic lights. It is also a very compact intersection, and cyclists have a good turning radius. The space is used very efficiently, no more space is used than traditional intersections in Edmonton, and probably you could fit it within a traditional intersection. I do not have a copy of the Dutch design guidelines for roundabouts, as they cost quite a lot, but lets try to make a plan for how big this actually is. 

Let's assume we have a design speed of 25 km/h for the roundabout, 2.9 metre wide lanes for motor vehicles and 2.5 metres of space between the two directions at the bicycle crossing, an 8 metre wide space between roundabout and bicycle crossing, a 3.5 metre wide bicycle path, a 50 cm wide buffer between cycle path and sidewalk, and a 2 metre wide sidewalk. I find that this is about 50 metres from one edge of the roundabout to the opposite side of the sidewalk. This seems large, but let's compare with regular intersection. 95 metres I calculate. So yeh. That's something to say about how much space we actually have. And we could have 10 metres of space, and bidirectional crossings on all arms of the roundabout if that would be useful, like to link with another neighbourhood through a shortcut. 

It has a relatively small circle in my diagram, but it would be much larger, to account for trucks, who are able to navigate Dutch roundabouts, with good visibility of pedestrians and cyclists. You encounter the traffic as a cyclist and pedestrian at a 90 degree angle. Cars are at an angle so that any collision with another car is likely to be a low angle sideswipe or a fender bender, and at ~30 km/h, very unlikely to hurt anyone in a car or large vehicle, let alone kill them. Most roundabouts in Assen look pretty similar to this, cycle crossings at 90 degrees to the cars. I do assign the right of way to cyclists, but using the physical design like Assen, I hopefully make them about as safe but with slightly quicker journeys. And in a city like Edmonton where cycle priority a new thing, this could be quite useful. And nothing would get a politician in the Netherlands thrown out quicker than making cyclists yield at literally every single cross street, stopping as they do so (like Sherwood Park). The cyclists have larger corner radii though.

This design can also be used in more rural areas. Design the roundabout for slightly higher speeds, 35 km/h or so, have at least 10 metres between the cycle path crossing and the roundabout, preferably 15, make the cycle corner radii a bit smaller, remove the sidewalk and design the approaches for 90 km/h or 60 km/h in some cases, and you have a well functioning roundabout. 

Larger or busier roundabouts may need more than a single lane. Most roundabouts can do well with one lane on each approach, at least each approach that a cyclist might need to cross. If cyclists do need to cross near one of these roundabouts, it should be where there is only a single approach and exit lane for the arms of the roundabout, and preferably with some distance from where it widens out for more lanes. If this is not possible, there should be A a spiral lane marking and something to enforce that, plus a median refuge at least 2.5 metres wide between each lane. But more preferably, an underpass or sometimes an overpass should be used. These roundabouts should also be used more sparingly, like connecting arterial roads to freeways and divided highways and the largest intersections, 91 St and 23 Ave for example. 

The Dutch have an invention for the world in relation to multi lane roundabouts. It is called the turbo roundabout. Do not try to mix cyclists and motor vehicles whatever you are trying to do with a turbo roundabout. Actually on any roundabout. I know cities and everyone else in the government of North America, and often the UK and Australia/New Zealand, say that because of the lowered speed of roundabouts, about 30 km/h, and the reduced number of conflict points, roundabouts are safe enough for cyclists and motor vehicles to mix well. The key element to why I object to this is volume and visibility. Roundabouts are built to handle (single lane) about 20-25 thousand vehicles per day. Way too high for mixing. Would anyone want to mix in a congested downtown, where the speeds may indeed be about 30 km/h but heavy volume? And almost always they exceed the 2000 vpd threshold I and Dutch road design accounts for. Cyclists also are usually on the side of roundabouts, meaning that at that distance, the sightlines become very poor. This is why cycle lanes are actually prohibited from being used on roundabouts in the United States by their MUTCD (manual on uniform traffic control devices). 

Back to turbos. They have both a spiral lane pattern, but also they go one step further and have actual raised ridges made of concrete, plastic, rubber or something similar, and sometimes even a full curb. Here is a picture of what I mean: Also, it correctly shows how cycling is to be done. Away from motor traffic. This also improves subjective safety, and of course, objective safety. If well designed, the underpass will not become a social safety problem. 

The volumes and speeds are just too high for cycling/walking to have at grade crossings except perhaps over a hundred metres away. If you have no other choice, a HAWK beacon could be used, but should be extremely rare. 

And this also works with public transport. A roundabout in Amsterdam has the cycle path, with priority although in a perfect circle, going around the roundabout, and motor traffic has a pretty normal looking roundabout except for the tram/streetcar tracks running right in the middle. This can also work if the tracks were crossing a particular arm of the roundabout not the middle, but it basically works like a railway crossing. The city does not want new LRT to have massive gate arms and bells and large signals, so because we have no direct equal for the signals they use to stop traffic when the tram approaches, a new one will have to be imagined. But we already have an option for this. Flash an amber signal to cars and no signal to pedestrians, but when the LRT approaches, turn it to solid amber, then to solid red, and to don't walk to pedestrians. And you can also replace the tram tracks with bus lanes too in the same way.

A mini roundabout is a concept from several countries, those from the UK probably know the idea, but I make sure that there is about a 5-7.5 cm area that is like a truck apron, but takes the place of the regular non traversable central island. It actually calms the traffic. And because the speeds are while less than 30 km/h, there is usually too much volume and is usually involving a distributor or arterial road, cycle paths are still required. This is different from neighbourhood traffic circles, where cyclists can share the lane with motor vehicle because of the low volumes and the access road nature of all the approaches. Still the same concept as before. Where cyclists are not on a ring, but with the physical design of Assen with the priority rules of Edmonton, and making sure sightlines are good with setback crossings for cyclists and pedestrians.

So why build roundabouts? 

Well, they help to implement Sustainable Safety. They keep directions down to an absolute minimum, with 8 conflicts on a traditional single lane roundabout, compared with 32 or more for a regular intersections. It combined with cycle paths keeps everything as homogeneous as possible. Similar speed, same direction as much as possible, same/similar masses. It makes it very clear what kind of intersection to expect. In urban areas, you expect to let cyclists and pedestrians go first. In rural areas you expect to have cyclists let you go first but if pedestrians for some reason are present, they go first. You also expect where to find cyclists and pedestrians. This makes it predictable. Roads by using roundabouts make the functionality of the road more obvious. Distributor roads focus on exchange at intersections but flow along most of the road. Like our arterial roads. Roundabouts are a great way to make it clear that traffic can turn, go straight on, cross and go around at these places without causing too much trouble. Even U turns are possible at most roundabouts. 

They also forgive mistakes very well, and reduces the likelihood that they occur in the first place. Crossings have good sightlines, including around trucks and buses. This means that if you spot someone making a mistake, you have enough time to stop. The low speed also helps with this. And if the collision does occur, that the impacts will be as minor as possible. The angles and speeds make it so that collisions between cars are likely to be property damage only, and even if worse things happen, a minor injury is likely the worst you can expect. It is self enforcing too. The curve of the roundabout, and you can add speed tables to the cycle/pedestrian crossings, make it very hard to go faster than the design speed, The island in the middle prevents people from making left turns in the way they normally do, making T bone collisions almost impossible. At multi lane roundabouts, the spiral lane markings and the ridges make it hard to go the wrong way. At Assen roundabouts, a total of 2 injuries among cyclists was reported between 2007 and 2012. The safe physical design made that low injury rate among 21 roundabouts possible. 

Even awareness by the road users is aided. If someone is driving drunk, or distracted for that matter (come on, how many people do it, you know you almost certainly do if you drive. Admit it), their ability to understand the road is reduced a lot. If they fail to do something correctly, the design makes that their mistake is less likely to cause a serious problem. If you are a child of say 8, you are not as aware as older people. But mistakes they cause while cycling or walking would be less impactful, like hitting the curb, which if you have a forgiving curb, there is not a problem. 

So how often can roundabouts be used? Much more often than you think. 

Here is a picture of a map I made (the base is Google, I added the dots and lines) 

Each red dot here is a regular single lane roundabout with a on traversable central island. Each brown dot is a mini roundabout, complete with a raised but traversable central island, and separate cycle paths, and each black dot is a turbo roundabout. Each blue diamond is a bicycle/pedestrian underpass or overpass. They all have one thing in common. Bicycles and pedestrians have their own dedicated space, well away from the actual circle. To give you an idea of how many signalized intersections (either existing or is planned to exist, plans made by the city of course) I replace, just count the red and black roundabouts. I count 21. I am hoping to have just ~2 cyclist injuries at all of them combined. I also forgot to mention on that diagram that there is planned to be accesses to Allard directly from 41 Ave. I remove the plans for that access. I also replace a number of stop signs, existing or planned. They would be where the mini roundabouts are. 14 in total. So quite a lot. 

Roundabouts also improve the environment, by not using electricity for one, and also having space for greenery. By not stopping as often and just slowing down, it produces less exhaust. By removing the element of danger, and making it possible to keep on cycling without stopping or slowing as much, and taking mostly straight lines, it attracts people to cycle, same with walking. This reduces the need for and use of cars and trucks. Transit also benefits, fewer collisions on the intersection itself, and the ability for it to have absolute priority over other traffic (on a transitway on separate lanes), making it faster and more attractive. Again, reducing the number of cars and trucks.

These things also cost less, especially over the long run. The signal itself, the installation, is often 150 thousand dollars or more. This is not even including the costs of the asphalt, curbs and paths and utilities, which can add well over 100 thousand dollars more. Maintenance over the long run is also high. Just ask the city of Carmel Indiana, where from pretty much any location in the city you are never more than a kilometre from a roundabout somewhere. Their investment dramatically reduced their expenses, and now they have more money for other things, like pothole maintenance. They have many more living people and non injured people than they would have otherwise. Much less pollution. And if you think the slow speed of roundabouts is bad, the Dutch have a saying: "A living person is faster than a dead one". 

Video about a roundabout location here:

No comments:

Post a Comment

Thanks for commenting