Thursday, September 10, 2009

Stop Spacing and Arterials

Rob has a good evaluation post up about the slow going Health Line BRT in Cleveland. Ultimately the planners weren't thinking of Rapid Transit and probably had too many people to please when they decided to have so many stops on the dedicated ROW. But this isn't just an issue that keeps BRT below 10 miles per hour, the T Third line in San Francisco has a similar issue. It has its own ROW down the center of Third Street, yet it has so many stops and crossings that the schedule maxes out at 12 miles per hour. Compare this to Rob's example of the Red line in Cleveland that goes 25 mph.

But its a balancing act of serving the most people possible and making the line go fast. For example the FTA sometimes goes to far and the running joke is that the cost effectiveness rating would be a great measure if you didn't have to have stations on the line. But something we might think about when planning these lines is that perhaps we don't need stops so close, and also that signal priority for buses and trains is imporant in making these generally longer haul services competitive with auto trips.


Alon Levy said...

Some of the local subway lines in New York aren't much faster. In Manhattan, the 1 and 6 both range from 12 mph at rush hour to 16 in the middle of the night. The 7 is scheduled at 19 mph off-peak, falling to 14.5 at rush hour.

Jon said...

In Portland for the proposed Lake Oswego streetcar line, unfortunately it looks like the hybrid option will be chosen. The hybrid option calls for the route to deviate from the private right of way and run along a busy arterial in the middle of the route then resume back on the private right of way. And to make the detour off the right of way to the arterial requires several tight turns :(

Portlanders know quite well about slow street running with the existing streetcar and yet this extension looked like a great example of the rapid streetcar concept with streetcar running on private right of way. The line could have been quite fast but this detour a block or two west will kill the speed.

I sure hope there will at least be signal priorty or transit lanes on the street running portion.

I just asked about this on Portland Transport with Chris Smith replying. Though it tied pretty well into this post.

EthanJ said...

If you have a dedicated bus right of way, you ought to be running mixed express-local services with off-line stations.

Justin said...

Wouldn't it be better to serve short-haul trips over long haul trips? That is where the bulk of passenger travel is. People focus so much on speed, and regional travel while ignoring that public transit should primarily serve local travel.

Rob Pitingolo said...

Justin, keeping what you said in mind, what bothers me is when transit isn't able to achieve its maximum potential speed because it has to stop and wait at dozens of red lights. All else equal, if a BRT or an LRT line has the ability to go 15 mph but it only does 10 mph, then that's bad for riders and bad for transit.

Cap'n Transit said...

With regard to the Lake Oswego streetcar, IIRC the street in question is wide enough to give right-of-way to the streetcars and still have room for cars to go both ways. The real question is why no one is willing to consider that option.

Matt Fisher said...

The Transitway, with inline stations, works, but isn't as good as rail. Yes, express buses from the suburbs have the advantage of "one seat rides", but it doesn't show that our Transitway here in Ottawa is an example of how BRT can be a supposed substitute for rail. It wouldn't be the same with the HealthLine in Cleveland either.

Michael Pereckas said...

Wouldn't it be better to serve short-haul trips over long haul trips? That is where the bulk of passenger travel is. People focus so much on speed, and regional travel while ignoring that public transit should primarily serve local travel.

So what is local and what is long haul? Can you put some numbers on that? The shortest bus ride I ever take is around 8km. Sometimes I go downtown from my home within the city limits, about 18km as the bus travels. Is that local? There really aren't any destinations closer than that to my home that are realistically served by bus. In theory there are destinations 8 or 10 km away in the can't-get-there-from-here sort of category as far as the existing bus routes go. Would that be local or not? Work is 14 km, I normally go by bicycle (by bus it takes several times as long with a diversion way out of the way and a long wait for the transfer.) The local drug store is about 2km, I walk there now and then. The bus doesn't go there, even if it did, wouldn't that be too close to take the bus to? You'd spend more time waiting for the bus than the trip takes. Much less aggravating to just walk, and little slower besides.

david vartanoff said...

Local, express, rapid? They all have their markets. The 1 and 6 subways in NYC are for short hops or the closest stop to your origin/destination--which is why in much of Manhattan there are expresses. As to the T Third St LRT line in SF the problem is that even with FEWER stops, it is slower than the predecessor bus, because the buses RAN whereas the streetcars defer to cars.

BruceMcF said...

Variable signal cycles with priority to a waiting bus could in some cases be more important than a dedicated right of way.

But with a dedicated right of way, there is no excuse for not having it ... a transmitter in the bus and a signpost receiver, and a light can receive a signal that a bus has stopped at a stop or passed a stop, and if it was stopped that it is moving again ... and the appropriate priority cycle drops into place.

And, indeed, a BRT line with very few stations is well worth considering. A BRT corridor could just have the Express stations on the dedicated corridor, while locals and semi-expresses in their local segment run mixed with traffic.

Again, signpost signal priority can readily be programmed to recognize a bus route number and know which lane it will be running in and what signal priority it needs.

John said...

I mentioned that I thought there were too many stops at a TRB event for the Euclid Corridor. RTA staff said they got rid of about half the stops the local route used to make, and they thought that was the most that was politically feasible.

I agree with Ethan that they should consider a mix of local and express services if the ridership justifies it.

Jonlin said...

Seattle Streetcar only goes about 8mph... But the Link Light Rail goes an average of 26mph, even though a third of it runs in the middle of the street. Strangely, the completely underground extension to the University of Washington will have about the same speed including two stops.

W. K. Lis said...

From Toronto's Sheppard East LRT EA:
"LRT Stop Locations
LRT stops are selected based on the right balance between good local access and high route speed. The higher the speed of travel, the longer the distance between stops. There were two general scenarios for stop distance for the Sheppard East LRT. Using the typical road network in the area west of McCowan as an example, the two options considered were:
1) LRT stops every 800 metres – design stops like a ‘surface subway’, stopping only at
major intersections (which are 800 m apart on this sample section of Sheppard), with
infrequent parallel bus service (e.g. every 20 minutes) servicing close bus stops in
between. At LRT stops, customers transfer to centre LRT platform from side of road bus stop.
2) LRT stops every 400 metres – stopping at every major intersection, and once in between, with no infrequent parallel local bus service.
A micro-simulation developed for the Sheppard East LRT modelled LRT spacing at 1000 metres
and 440 metres because the study includes sections west of Victoria Park and east of McCowan
where major intersections are more widely spaced. The model resulted in a stop spacing of 800
metres requiring a route speed of 26-27 km and a stop spacing of 400 metres requiring a route
speed of 22-23 km.
The 800-metre scenario was not selected because the full impact of the increased speed of the
Surface Subway applies only to those walking directly to LRT stops. Those boarding local buses at bus stops in between LRT stops have a shorter walk, but a longer wait for service and a transfer to the LRT after a very short bus ride. Further, 800 metre spacing did not achieve as great a speed
advantage as expected – while the LRT stopped less often, the time for customers to board took
twice as long per stop (same number of passengers collecting at half the stops) and the LRT still had delays due to red lights at signalised intersections in between stops (though the model accounted for signal priority to reduce such delays).
The recommendation for Stop Spacing in the order of 400 to 500 metres, depending upon the
pattern of development and cross-streets (with an expected average speed of 22 to 23 km/h), was considered to be the best balance between overall route speed and good local access. For
purposes of comparison, during peak operating conditions, the average speed of the Bloor-Danforth subway line is 30 km/h, the 85 Sheppard East bus service is 17 km/h, and the 510 Spadina
streetcar service is 14 km/h. In assessing the overall customer service provided in each scenario, City and TTC staff recommends LRT stop spacing ever 400-500 metres, with an average stop spacing of 460-to-480 metres, depending upon the additional stops that may be included during detailed design."

Alon Levy said...

Are Torontonians physically incapable of walking 400 meters, from the center of an 800-meter block to either stop at the ends of the block?

W. K. Lis said...

Don't forget that 400m (¼ mile) can be far for seniors or if you you have a cast on your leg. As well, they may not live right next to the route, but away from it some distance, a block or two or three, as well.

Michael Pereckas said...

Sure, but then your actual destination is probably one or two km from the line anyway.

Again, I really think a very expensive motor vehicle system with very expensive infrastructure really should be faster than normal day-to-day human powered movement. If a naked caveman with no technology whatsoever could outrun your streetcar, something is wrong.

Ted King said...

Some points to consider on SF Muni's T-Third LRV :
1) The ROW is mixed - dedicated sections at the top and bottom of the route but a long section of regular street with the tracks in the pavement in the Hunters Point area. There are also lots of left turn lanes and crossings.
2) The run times increased compared to the previous diesel bus (15 Third + Kearny). Part of that was due to the forced transfer of passengers at Arleta + Bayshore (A+B). That's right, those people boarding along Geneva now have to get off at A+B, walk across Bayshore (slow signal), and walk up the ramp onto the platform. It's very easy to miss one's connection at A+B.
3) The Fourth + King intersection is a flaming kludge. Somebody had a brain fart and combined a freeway ramp with a major transit crossing. Combine that with a missing reliever track along the south side of Mission Creek and on into the back end of the N-Judah terminus yard and you have lots of delays.
4) The new yard at 25th+Third has had some nasty signal problems. There have been times when the drivers were told to stop, look around, and then proceed. Other times yard movements / signal problems have added ten to twenty (10 - 20) minutes to the run.

This is not hearsay - I've ridden both lines (15's and T's) and can only score the T-Third as a marginal success. Typical run time from Geneva + Mission to 23rd + Third was forty-five (45) minutes. The 9X + T-Third combination on a good day does it in fifty (50) minutes. It's just as easy to go the other way around (Muni + BART + 48's to 22nd + Third) and takes about the same amount of time.

david vartanoff said...

Ted describes the T well, but misses two items. First, the money was spent for signal preempts but they are not used. Second, Muni drivers are told they are at fault nearly no matter what some idiot car does. Thus they run slow as a job survival tactic.

Matt Fisher said...


You're right. They should have done this.