Auckland Transport is recommending the higher-end median busway solution for a difficult suburban arterial corridor that we studied for them last year. (The corridor extends from Panmure rail station through Pakuranga to Botany via Pakuranga and Ti Rakau Roads.) It would be the main radial rapid transit corridor for Auckland’s far eastern suburbs.
When you’re looking at fitting some sort of “Bus Rapid Transit” (BRT) into a suburban arterial, the question of median vs side lanes is bigger than it sounds. It largely determines the scale, prominence, and apparent permanence of the project.
Putting BRT in the median is the high-risk, high-benefit, high-impact choice. A median BRT solution looks and feels like a separated busway, and its dedicated median station infrastructure makes the service look both prominent and permanent. It obstructs a lot of turning movements that motorists are used to, and is generally harder to compromise.
On the other hand, putting a bus lane on the side, as Seattle’s new Rapid Ride Line A does, delivers a more compromised outcome that has much less impact on car circulation. Stops on the side risk looking like any bus stop, so unless you make a very strong statement with the station architecture, the service may not look like permanent infrastructure. A side lane is also generally shared with cars turning into and out of the street on that side, which is often a source of roughness and can mean significant delay at major intersections.
That’s not to critique the Seattle-area examples, which are designed to a different scale. But I congratuate Auckland Transport for going out to the public with the more ambitious option, and thus triggering a real and enlightening debate.
(Commenters: Note that I’m speaking to both left-hand-drive and right-hand-drive countries with this post, and thus avoiding all references to left and right! I challenge you to maintain similar neutrality.)
Photo: Rendering of median bus lane option, presumably by Auckland Transport, via New Zealand Herald.
I’d be interested to hear some discussion on the access to median stations (BRT or LRT). I see a lot of near misses on Embarcadero in SF as pedestrians scramble across 2 lanes of 35 MPH traffic and then the track itself to get to a Muni train waiting on the platform. It seems like the combination of increased visibility of the transit vehicles plays as much a part as the obstacles between the sidewalk and the platform.
Pedestrianist. To use transit for a round trip, you need access to stops in both directions. So while side platforms spare you from crossing the street in one direction, they require you to cross the entire street in the other. So regardless of whether it's center or side, you'll cross the street once per round trip.
Helsinki has just tightened up regulations on its side bus lanes, and is finding the problem that bus lanes in which turns are permitted are quite easy to abuse – “Not all drivers using a bus lane plan on taking the next right turn. They simply take advantage of the unclear situation”.
As a result, it’s considering central bus lanes instead. It already has trams in the central reservation, so that would mean four transit lanes in the centre, and one lane on each side for mixed traffic. The Helsingin Sanomat provides commentary and a picture of how they envisage it looking – http://www.hs.fi/english/article/YTV+Public+transportation+lanes+on+Mannerheimintie+could+be+in+centre+of+street+/1135235646509
The access challenge mentioned by Jarrett a short time ago rears its head here – “Where do we find enough room for the bus stops, and how safe would it be to cross the street to get to a stop?”
My answer to this is to stop having trams and buses wastefully duplicating each other, and instead have buses run on a rapid basis, stopping at main intersections averaging about 1km apart, thus shrinking the need for providing stops greatly.
Auckland is also investing a large amount of money in upgrading its commuter rail line, including electrification.
Seattle uses the median approach with Sound Transit buses along the freeways.
The other big advantage of the median is that the bus doesn’t have to pull out of its lane to stop and load passengers. This is especially problematic on freeways, because the bus often has to exit on an offramp, load passengers (possibly at a park-and-ride station) and then re-enter the freeway.
If you put the station in the median, though, passengers have a longer walk up and over the freeway to get to the station.
I’m curious what the relevant trade-offs are here, Jarrett, between making people walk a bit farther and having a faster ride overall. Does that factor into the decision at all?
Victoria (BC Transit) recently (2008) shot down a median transitway, largely due to the resistance of the local Chamber of Commerce but recently came back with a transitway running along one side of the street. It will be interesting to see the reaction to the new plans.
Ah, but your picture clear shows vehicles moving on the left. The examples discussed below, all occurring in North America, all assume right-hand drive–substituting right for left in those places where ’tis appropriate is left as an exercise for the reader.
At any rate, the distinction between median and side BRT isn’t quite as dramatic as you imply. Eugene’s EmX down Franklin Boulevard (seen quite clearly here) runs in the median of Franklin Boulevard (OR99) just north of the University of Oregon. The street, as the name suggests, is an urban boulevard (a rather wide one) with numerous intersections and left turns. When the line was put in, the existing left turns weren’t closed FTMP, but–and this as key–the busway is not shared with vehicles turning left. This did result in a different compromise–in several sections, the busway is reduced to a single lane (used by busses in either direction) as a result of insufficient room to provide both turn lanes and two exclusive bus lanes (and lack of political will to reduce auto lanes).
The former Translink BRT line down No. 3 road in Richmond, BC (which was removed to make way for the Canada Line) operated in the same fashion–busses got two continuous, exclusive lanes (one in each direction) down the median; left-turning vehicles do NOT get in the way of busses. (Unless they get stuck in the turn, I suppose).
Also in the province of BC, along BC99 from about Ladner all the way into Richmond, is a brand new bus lane–on the side (right) lanes of the freeway. Given that this is also the side of the highway on where the offramps and onramps are found, various techniques are used to keep busses from being delayed–there are a few instances of braided ramps, and a few onramps where there is a ramp control signal which blocks merging traffic from entering the freeway (and crossing the bus lane) when a bus is nearby.
But a key point, I think, isn’t so much whether the busses run down the middle or along the side. More important is the degree to which cars are permitted to mix with busses–if cars can stop or stand in the bus lane while making a turn or parking, it lowers the effectiveness. If the bus lane is truly bus-only (perhaps with some physical separation to discourage autos from encroaching on it), does running on the side pose as much of a problem?
Or does separation from the sidewalk give median-running busses an advantage?
Ah excellent, an Auckland post!
This is an interesting proposal because it runs through what is probably Auckland’s most car dependent area and the area which currently has the worst public transport (certainly not a coincidence there!)
The current level of service along the road isn’t particularly high: say 10-15 minute peak frequencies and half-hour off-peak frequencies. There’s also very little money to spend on improving frequencies in the near future – because New Zealand has a Minister of Transport who’s obsessed with building motorways. So this project isn’t assure of success in a way that perhaps doing something similar to the Dominion Road corridor (2 min peak frequencies, 5 min off-peak) would be.
I think what will be the key issue for this project is whether the routes can be changed to terminate at Panmure/Ellerslie (the train stations at the city end of the busway). There’s no real hope of getting a busway standard road any closer to the CBD, and it would just be silly to keep running all these buses into the CBD and out again – when if they terminated at the train stations you could double the frequency with the same resources (shorter runs = increased frequency).
Jarrett, did you model how long a trip from Botany to Panmure might take with the busway built? 15 minutes?
Josh. The number you request is not at my fingertips, but that sounds about right. Re Panmure, yes, there’s an issue about how you get closer in, but ARTA (now AT) chose to break the corridor conceptually at Panmure. It’s one of those moments of “yes, it’s all connected, but if we don’t break it somewhere the problem is just too big to study.”
Scotty. You’re right that it’s possible to create exceptions both ways — well-protected side lanes and disruptable median lanes — but the starting point of discussion is quite different. Median makes it hard to disrupt, side makes it hard not to.
Note to all: Freeways are an entirely different situation. This post is about arterial street BRT, where there are frequent signalised intersections and almost constant local access to the curb lane from side streets and fronting land uses.
One other issue to consider with center-running BRT is that most buses only have doors on the curb side, which makes island platforms impossible. Eugene’s EmX is an exception, with buses that have doors on both sides and island platforms.
Jarett: “So regardless of whether it’s center or side, you’ll cross the street once per round trip.”
Umm… no… with a median lane, you have to cross the street when you start your (round) trip, and again when you end it. with a side lane, you only have to cross on one of those.
Crossing half the street twice is still crossing the street twice – it is not the same as crossing the whole street once!
Also, with median lanes, you always want to cross when you get out teh bus – which is exactly when the traiff will be flowing in the direction parallel to the bus route. So, you always have to fight through traffic.
Median bus lanes do have one large potential advantage: they can be made sufficiently segregated to allow buses to travel at higher speeds the speed limit for adjacent traffic, meaning that transit can *quicker* than driving.
@Tom West, why do you need to fight through traffic? If you have a PT stop in the middle of the road, then wouldn’t there either be a law that says while people are boarding traffic needs to stop, or else pedestrian infrustructure?
@Zoltan et al. I think in Melbourne on the rare occasion that there’s separated tram lines and buses along that segment, the buses run along the tramline unless there’s a very particular reason for them not to. However, I can only really think of one or two examples of this happening. An example is along Queens Pde/High St in the vicinity of Clifton Hill, which you can see in Google Street View http://maps.google.com.au/maps?f=q&source=s_q&hl=en&geocode=&q=clifton+hill&ie=UTF8&hq=&hnear=Clifton+Hill+Victoria&ll=-37.785775,144.994276&spn=0.007529,0.009731&t=h&z=17&layer=c&cbll=-37.785707,144.994345&panoid=5zMyRXoUA5WDgnK7tavimw&cbp=12,189.41,,0,9.07 There’s a bus heading north (towards the camera) and a tram heading south. The bus/tram lane ends to go under the rail bridge and over a creek, but comes back a bit further north for a short while. But I must stress this is highly unusual in Melbourne; most roads with trams don’t have enough space for peds and parking and two-way cars and tram lanes; and in any case, buses rarely overlap with trams.
With regard to the question of pedestrians crossing the street to catch the bus: http://deadhorsetimes.blogspot.com/2010/11/flashing-red-lights.html
@Felix the Cassowary
Helsinki does suffer from a silly degree of separation between buses and trams, as if they’re for different people, and therefore in suburban areas, where trams are every 7-10 minutes, I would argue for buses using the median tram lanes.
However, on the Mannerheimintie, and other arterials leaving downtown Helsinki, there are up to about 30 trams and 30 buses per hour, so four transit lanes are necessary.
Is there any literature or studies on locating busways or light railways/tramways in the median versus to *one* side (and I don’t mean side bus or transit lanes – I mean side-aligned with both transit directions on one side of the street in a dedicated RoW with some separation from the street)?
I ask because in the two Canadian cities with LRT, Calgary and Edmonton, we’re seeing a tendency towards side alignments rather than median alignments. Median alignments seem to be used only on heavily-used arterials or expressways and the stations in such corridors are fairly significant structures accessed by overhead pedestrian walkways.
Take Calgary’s West LRT under construction. A few short sections of it will be median aligned in an expressway environment weaving through an interchange (at Crowchild Trail) but for the most part it will be placed on one side of 17th Avenue heading west. Also in Calgary, the recent extension of the Northeast line moved out of the median of 36th Street and to the side of 44th St as it makes its way north. Similarly, the future Southeast line anticipates a side alignment along 52 St SE (see the space on Google Earth provided for it on the west side of 52nd either side of the intersection with Prestwick Gate SE). Only in the Northwest where Calgary is basically stuck in the median of an expressway will they continue to build in a median.
Edmonton currently has some side-aligned LRT along 114 St SW, though they pop into the median of the more heavily-used 118 St SW further south. Most of Edmonton’s existing system runs in an existing or former railway corridor, so not much can be said about it. The future North line will have side-aligned running along 106 St NW and briefly on Kingsway (the rest of it basically takes over streets outright). I suspect this trend will continue for future lines.
Of course the major Canadian counter example is Toronto, where the streetcar system (assuming it survives the current mayor) is gradually being segregated into median alignments that make it virtually into a light rail system.
I can see some distinct advantages to side-aligned running rather than median running. For pedestrians, a side-aligned transit corridor doesn’t quite have the same effect of increasing the effective width of the corridor for crossing purposes – the transit facility and the road are crossed as distinct events rather than all at once when the transit facility is in the median.
For construction, it may be less disruptive and may require fewer utility relocations and the like. Stations can be built larger as pieces of suitable property can be acquired for the purpose, whereas to steer a road around a large median station requires considerably more property. There would also be more opportunities for integrated station developments.
I suspect motorists may be more careful with side-aligned transit corridors (or at least light rail/tramway) as well: a motorist on the “near side” making a turn across the tracks to a cross street (e.g. right turn in a right-side driving country, left turn in a left-side driving country) will first have to face the oncoming track before crossing the track of vehicles heading in the same direction as the motorist was coming from – in effect they can see the first “threat” before even making the turn and they have the width of that first track to see a threat from a transit vehicle approaching from behind.
Motorists from the “far side” carriageway making the opposite turn would have the width of the opposite direction carriageway to see trains moving in either direction. With a median alignment, motorists could turn into the parallel same direction track without trying too hard – and many apparently do in Houston.
Motorists on cross streets, especially those making a turn across facing traffic [aside: this side-neutrality business is causing some awkward terminology…] would not have to worry about sitting on the tracks while waiting to complete their turn as they might in a median-based system. Motorists in such a position in a median-based system may be more inclined to make a mistake in an effort to get out of the perceived danger zone.
In the above, I’ve tended to write about LRT. I can’t decide if buses operating exclusively on one side in what looks like a roadway to everyone else might not introduce its own confusions. As it happens, in bus-mad Ottawa, we do tend to put busways on one side rather than in a median, but our busways are generally grade-separated anyway so the decision to put them on one side doesn’t have to concern itself with how it will be perceived by motorists.
I’d also be interested in a hybrid model, at least for buses. In one direction, put buses in a bus lane at the edge of the roadway. But for the other direction, instead of putting buses in a similar lane on the other side of the road (as is typical), put them in a uni-directional “contraflow” lane on the same side as the other bus lane but separated from it by a median. This allows the median to serve as an island platform for both directions at once. The “normal” direction bus lane could even be separated from regular traffic by a narrow median or curb of its own. The appropriateness of such a design is going to depend on the particular context, but in a suburban environment this uses less space than a median alignment and doesn’t involve nearly as much reconstruction of the roadway while arguably providing better and more visible service than side bus lanes. One might envision the other side of the street having a bikeway as well, thus removing the bike-bus conflicts that are so annoying with side bus lanes.
@Felix the Cassowary: Of course you can force traffic to stop when the bus does. However, I’m unaware of an example where there are median busways with “platforms” for passengers to wait which also require traffic to stop.
(In Toronto, traffic has to stop when the median-lane-running trams/streetcars do, but that’s because there is no platform for disembarking passengers.)
Back to Jarett’s point about speed… yes, bus lanes on the side will get helf up by cars turning out… but if these aren’t busy to warrant a “proper” intersection, I don’t the delay would cause a significant speed disadvantage compared with median lanes.
My observation from Vancouver Canada is that buses (or any large transit vehicle)traveling in the curb lane at high speed is going to adversely affect the pedestrian experience. In particular I am thinking of the diesel 99B line buses on Broadway. It is not so bad with the electric trolleys providing frequent stop service on the same route.
Perhaps curb side transit lanes are more appropriate for frequent stop, lower speed service where there are sidewalks and pedestrians. Imagine sitting in a sidewalk cafe right next to a 60km/hr curb lane express bus or rail line with 2 min headways.
It’s been about 20 years since I lived in Auckland… Pakuranga and the entire eastern suburbs still had lots of sheep pastures when we moved! The BRT project should be a good fit for this area. The bus ride from Howick to Downtown used to take about 1 hour.
What passes for BRT in many places is a joke, no wonder many people mistrust the idea.
Some good “full” BRT inspiration
Not something that many places actually have – 1) the entire network on these systems is fully isolated from regular traffic
2) the investment real stations over fancy buses.
Watch from 0:50
And a Google streetview link from Johannesburg City Hall in South Africa:
BTW ^ these work exceptionally well on major arterial/highways with distantly spaced stations (ie regional service).
Toronto won’t necessarily remain median-oriented forever. The design for the Cherry St Streetcar, if it’s ever built, is side-of-the-road.
Also, the planned narrowing of Queens Quay, if it happens under the new mayor, will make the Queens Quay streetcar median side-of-the-road by eliminating the roadway to the south of it.
@David in Ottawa: yes, there’s some literature. Download the ITDP’s vast bus rapid transit guide. The discussion of locating BRT to one side of the road begins on page 186. The guide says that in most circumstances it’s a bad idea because it leads to more turning conflicts, but if the ROW has an obstacle to one side, such as a park or an ocean, then it may actually reduce conflicts.
Another guide, one of the various US government publications I’ve hunted down (I don’t remember the exact reference), is more negative, saying that putting the lanes on one side could confuse riders as to which direction vehicles will come from, leading to more accidents.
ITDP partner Viva’s recent global BRT trends – great arguments for median stations:
Oh and Built Environment Journal has an entire issue dedicated to BRT this past year (Vol 36 Issue 4).
Infrastructure styles are definitely diverging. “North America is going the way of express bus BRT [aka BRT Lite] as is Europe, while Latin America and other places going the way of train-like BRT”.
Thanks for the tip, but there’s not really that much in there on this topic. They seem to have a pre-determined preference for median alignments without much of a rationale being provided. It mentions turning conflicts from side-aligning, but that occurs with median alignments as well. Indeed, arguably a side alignment has fewer turning conflicts per intersection because a median alignment has *four* turning conflicts (e.g. all left-turns in right-side driving countries) while a side alignment only has *three* (two of them right turns in right-side driving countries) because one of the nominal turning conflicts – from the cross street from the side with the transit facility making a turn across facing traffic – is effectively removed as such traffic crosses the transit facility and the first carriageway well before making the turn, so from a practical point of view it’s no different than vehicles going straight ahead. Besides, all these possible conflicts should be controlled by signals anyway. Rather than actual turning conflicts per se, I would tend to think the bigger issue is motorist confusion during turns over what is a roadway and what isn’t – something a light rail line is unlikely to suffer from. If a road has a lot of property accesses that can’t be closed, then it’s a different scenario as a median busway is protected in a way that a side-aligned busway wouldn’t be – but then we are discussing suburban arterials here and not the kind of corridors typically found in that guide…
I also found that guide to be a tad unrealistic in that it called for island-style median stations, which, if you’re not going to run buses counterflow, requires special buses with driver-side doors. In Ottawa at least, one of the major historic rationales for BRT over LRT was that the existing bus fleet could be employed and that any bus used by the transit agency could be used on the busways (the fact that our buses disintegrate in a few years due to all the salt we dump on our streets in the winter is just one of those things that BRT obsessives don’t like to talk about). For a city like Ottawa to contemplate this, we’d have to institute a 10+ year plan to replace the bus fleet before building the first median busway.
More generally, the guide is clearly aimed more at the developing world than the developed world. The guide’s authors are trying to employ buses as a surface-based “metro-light” system that’s built without much regard to the existing bus transit system, essentially eschewing the flexible phasing-in options that BRT proponents (at least those from Ottawa who proceeded to go soothsaying around the world) love to mention. In other words, it’s the sort of construction philosophy inherent to rail systems and which Ottawa’s BRT lobby has always castigated as a weakness of rail-based systems. Perhaps it’s not too much of a surprise that Ottawa doesn’t appear much in the guide itself… I also note that Ottawa’s quasi-religious obsession with grade-separation everywhere and always is mercifully not present in this guide.
On that note though, the “Qualitative comparisons” chart on p. 769 for Ottawa is so full of errors as to make me wonder about the rest of the cities listed. Of the 13 ‘X’s, 5 are incorrect, and one of the two ‘P’s should be a ‘check’. I’d also say that the ‘check’ for ‘Competitively-bid and transparent contracts and concessions’ is a bit of a stretch considering the cozy relationship between City transit planning staff and their pair of favourite consultants.
To take the conversation further, I’ve read evidence (which I currently cannot find, and I would love it if someone dug it up) that wide boulevards are just about the most dangerous environment there is for pedestrians, as well as creating long, unfriendly walks across the street (separating “side one” and “side two” retail districts.
In this context, which design for bus lanes (or streetcar lanes) better ameliorates the inherent problem with these boulevards? I could see arguments either way. I would tend to guess that the presence of stations in the median at cross-streets would make more of a “continuous area” rather than one massive gap for pedestrians to cross.
Yes, medians provide safe refuges for peds. Good signals provide a crossing of the street for fast peds, while giving slower peds the option of crossing in two cycles. This function is easily integrated with median BRT, but is harder with side BRT.
@David in Ottawa:
Anything else about BRT in Ottawa? (I believe, though, that there is no question that the Transitway is a success.)