8 lessons about great transit I learned riding the Paris Métro
Paris has one of the world’s great subway systems. Beyond its truly impressive coverage and service quality, here are eight wonderful details about how it operates that US systems would do well to mimic.
Door knobs on a Paris metro train.
1. Door knobs speed trains
In DC and in many US subway systems, when trains pull into stations passengers wait for the train operator to open the doors. That adds a few seconds to every stop while the train idles on the platform, doors shut. Waiting passengers tap their feet and cross their arms.
All those seconds, at every station, every trip, all day, add up. The result is not only less happy riders, but also slower trains that come less frequently and carry fewer people than the system’s theoretical maximum.
In Paris, those delays don’t happen. Each door has a manual knob or button that passengers can push to enter or exit at their own pace. For safety, the doors are all locked while the train is moving quickly. But as it comes to a halt the doors unlock, and passengers can immediately open the doors to exit trains.
Here’s a video, showing how the whole operation makes exiting a train noticeably faster than on WMATA:
WMATA did have automatic doors up until 2008, which were faster than the operator-controlled doors of today. But that was eight years ago, and there’s no indication they’ll be fixed any time soon.
Why do WMATA station platforms have so few seats? Especially at side platform stations, why not just line the entire platform with one long bench?
Check out Paris’ Chatelet station, where that’s exactly the layout:
Most Paris stations aren’t like Chatelet. Frankly, with sub-five-minute headways most of the time, a lot of seating isn’t as crucial there as it is in DC. But there’s been many a day I’ve stood for 15 minutes in a WMATA station wishing it had this feature.
3. Flip-up seats add capacity
The first row of seats inside Paris’ train doors flip up. On sparsely-populated trains, riders can sit in the seats comfortably. On especially crowded ones, riders can stand, creating more space on the train.
Yes, riders in Paris sitting on these seats do seem to usually get up and create more space when the train gets crowded. It seems to be part of Paris transit etiquette, like standing on the left on DC escalators. Not everyone does it, but enough do to make a difference.
This arrangement also makes it easier for people in wheelchairs to ride without blocking the aisle.
4. Open gangways really do work
US transit systems are slowly beginning to catch on to the benefits of longer open-gangway trains. If passengers can move from front to back of trains without getting off, that makes trains less crowded and boosts capacity.
All new or recently refurbished lines in Paris have open gangways. And they’re wonderful.
5. Great late night service is possible with only two tracks
Paris’ metro lacks express tracks just like DC’s, and it runs basically comparable hours to WMATA. It’s also decades older than Metrorail. It must have at least similar maintenance needs, and no more time in the day to accomplish them.
Yet somehow Paris manages to run frequent trains late into the night.
A train every 4 minutes at 10:21 pm.
I have no idea how they do it. When do maintenance workers do their work? How do they keep up tracks with trains coming every four minutes?
I wish I knew. If you know, send Mr. Wiedefeld an explanatory note.
And though a bridge over the Seine is a special place, Paris’ els have nice aesthetic touches elsewhere too.
7. Wayfinding can be beautiful
“If you can make something pretty, why not make it pretty?” My wife and I kept coming back to that thought as we explored Paris. These signs, telling riders which direction their metro train is headed are one example of why.
In DC we already put location-specific bus maps and neighborhood maps inside every Metro station. Why not unique maps for destinations to which infrequent riders often travel, like airports and stadiums?
What details like these have you noticed on other countries’ transit systems, that you’d like to see imported to the US?
These storefront maps show which parts of US cities are most lively
These maps show nearly every retail storefront in central DC compared to those in New York, Detroit, and other cities. Since retail streets are usually the most lively streets in a city, the maps offer a nice proxy illustration of urban vitality.
In general, the more red dots you see in a small area, the more lively that part of town will be. More stores, after all, mean more destinations for people to visit.
Here’s the DC map in greater detail:
Image by City Observatory.
You can easily see retail streets like U Street and H Street, and bigger clusters like Georgetown and Dupont Circle. On the other hand, primarily residential neighborhoods are mostly blank.
Unfortunately the data clearly isn’t perfect: The retail complex in Columbia Heights seems to be missing, as are the giant gift shops in the Smithsonian museums, and some neighborhood corner stores.
Still, the maps are an instructive illustration of urban vitality in general. You can see patterns here, and those patterns are real.
Zooming out to the regional scale, downtown areas outside the District like Bethesda, Silver Spring, and Alexandria become prominent.
Bethesda and Silver Spring are the clusters at the top. Alexandria is at the bottom. Image by City Observatory.
Compared to other US cities, DC looks decently lively. The country’s dense, transit-oriented cities like San Franicsco and Boston fare well (New York is a crazy outlier), while economically disadvantaged cities like Detroit and sparser more suburban-style ones like Raleigh show fewer stores, indicating less urban liveliness.
Of course, retail storefronts are a simplistic way to look at this. New York’s streets have a lot of stores because New York is tremendously dense, so there are lots of customers to support them. On the other hand Tysons Corner has a lot of stores because it’s a big suburban mall that people drive to from miles around.
Even suburban malls offer a sort of liveliness, however. So while these maps may say little about walkability, they are a good proxy for liveliness.
Why widening highways doesn’t work, in one simple gif
Decade after decade, American metropolitan areas continue to widen their highways in order to reduce congestion. And decade after decade, congestion just keeps getting worse. That may be counterintuitive, but it’s because of a phenomenon called induced demand. This simple gif illustrates how it works:
Surely one more lane will finally solve our congestion problem, right? (Slightly better GIFF. Feel free to copy) pic.twitter.com/uDJwqVT3WI
Of course, it’s a little more complicated than this gif. Congestion keeps increasing not only because more people drive, but also because more people drive farther. And because the more highways we build, the less walkable and transit-accessible our cities usually become. And because the more desperate our congestion situation becomes, the more some groups attack using money for anything other than more highway widenings.
Highway congestion is a negative feedback loop. The only way to really solve it, besides economic calamity, is to break out of the loop by attacking its root causes. Rather than applying highway-widening band-aids that only work for a few years, build urban communities with multimodal infrastructure, in which it’s just as convenient (or more so!) for most residents to get around without a car than with one.
That doesn’t mean no new roads are ever needed. New communities and densifying ones need streets, after all. But it does mean we should be skeptical of plans to make highways bigger. In the long term, that money is usually better spent elsewhere.
Metroway runs between Pentagon City and Braddock Road Metro stations. For much of its route, between Crystal City and Potomac Yard, it runs in dedicated bus lanes, making it the Washington region’s first real foray into BRT.
Through Potomac Yard, the transitway runs in a totally exclusive busway—a completely separate road from the regular lanes.
27th & Crystal station.
Stations in the busway have substantial arched roofs and attractive wall panels.
South Glebe station.
Through Crystal City, bus lanes and bus stations hug the curb.
18th & Crystal station.
Since northbound buses run a block away from southbound buses, bus stations are smaller through this section. More like large bus stops.
23rd & Clark station.
Crystal City is pretty quiet on Sundays, so there weren’t many opening day riders and buses only came every 20 minutes. During the week there’ll be a lot more riders, and buses will run every 6-12 minutes depending on the time of day.
Head over to Crystal City and check it out! Or see more pictures of both the Arlington and Alexandria transitway sections via Flickr.
90 new rowhouses at a Michigan Park seminary could help address the housing shortage
The St. Joseph’s Seminary in Northeast DCs’s Michigan Park neighborhood has a large eight-acre property, but the seminary only uses two acres. Rather than let the rest sit empty, they plan to add 90 new rowhouses on four acres, and turn the rest into a park.
The historic seminary building, as seen from 13th Street NE. Photo by Jonathan Neeley.
The Josephites, as the seminarians call themselves, have been working with developer EYA to build on the site. EYA’s proposal is called 12th and Allison. It focuses on the northern part of the site, preserving the southern part and its historic seminary building.
Location of St. Joseph’s Seminary. Image from EYA.
The Josephites would retain ownership of the southern part of the property, which includes both the building itself as well as a little over two acres of open space.
To the north, EYA would extend Webster Street through the block, connecting to 12th Street (Webster currently ends when it hits 13th, on the east side of the property). Surrounding the new Webster Street would be 90 rowhouses, most of them north of the new street.
Triplexes everywhere versus rowhouses and a park
Today the seminary grounds look sort of like a park. But they’re not. The seminary is private property, and if the Josephites sell part or all of it, that part can be developed according to however it’s zoned.
Almost all of Michigan Park is zoned R-2, which allows “semi-detached” housing like duplexes and triplexes. Many of the blocks surrounding the seminary are lined with the latter.
But developers rarely build triplexes these days. They require so many setbacks that it doesn’t pencil out to add in any communal open spaces like parks. But the setbacks are rarely large enough to be very good private yards. For new construction in the city, regular rowhouses are more popular with both sellers and buyers.
Thus, EYA hopes to rezone the property to R-5A, the same as Providence Hospital across the street. R-5A zoning would allow for normal rowhouses, which in turn could be clustered together, allowing for better community open spaces.
But rezoning requires city action, and that opens the door to controversy.
At a community meeting in October of last year, a number of residents made it clear that they were opposed to development of any kind. At least twice throughout the winter, opponents spread these flyers throughout the neighborhood, forewarning against the evils of building:
Opponent flyer. Photo by Jonathan Neeley.
Michigan Park is a cozy, moderate-density neighborhood. It’s fair for residents to wonder about the impact of a new development, and hope to influence it. But hyperbole like that isn’t helpful and isn’t true. Saying this project would “irrevocably damage our community” is a stretch.
90 new units on two big blocks won’t turn the place on its head. That’s only a little denser in total than the surrounding blocks of duplexes and triplexes.
By clustering the development mostly north of Webster Street and preserving more open space south of it, the northern block will be noticeably denser than triplexes, but in return the historic seminary building and much of the open space on the south will be permanently preserved, designated historic, and off-limits to future development.
That’s a good trade. Right now, if the Josephites wanted, they could sell their entire property and develop 100% of it as duplexes and triplexes by-right, whether anybody objected or not. Rather, in exchange for rezoning to allow rowhouses, the seminary and considerable open space will be saved.
In April, EYA will present its latest plans at local ANC meetings. They’ve reduced the density of the proposal from 180 houses to 90, and promised to design the buildings in a high-quality, contextual way.
After that, they’ll submit for zoning approval, and apply to the District’s Historic Preservation Review Board to designate the seminary building as a landmark. Expect hearings on it this fall.
The Purple Line will have America’s longest railcars
According to the latest plans for Maryland’s Purple Line, it will have the longest transit railcars in America. Each train will have a single 136-foot-long five-segment railcar. They’ll practically be open-gangway trains.
A Purple Line railcar compared to Metro and DC Streetcar. Image by the author.
Purple Line trains will be Urbos model trams, built by Spanish company CAF. Urbos trams are modular; you can make them as long or as short as you want. These will be unusually long ones.
At 136 feet long, they’ll be 2 feet longer than the closest US competitor: Austin Metrorail’s 134 foot cars. But Austin’s cars are DMUs, a sort of commuter rail / light rail hybrid, built for longer distance and fewer stops compared to the Purple Line.
The next biggest US light rail cars are Dallas’ 124 foot cars.
Dallas light rail car. 12 feet shorter than the Purple Line’s cars. Photo by Matt’ Johnson on Flickr.
Longer is better
Having one long railcar rather than multiple short ones has a lot of advantages. There’s less wasted space between cars, less expense per rider, and passengers can move back and forth inside the train to find the least crowded spot. Overall, having one long open interior increases the capacity of a train by about 10%, and it costs less.
The downside is you can’t pull individual cars out of service if something goes wrong. It’s all or nothing. But as long as everything works, long railcars are great.
Since the Purple Line will be operated by a private company that faces penalties if it doesn’t meet service requirements, the onus is on them to keep trains in service.
An open interior train on the Paris Metro. Photo by BeyondDC on Flickr.
In transit jargon, these open interior trains are called “open gangway,” and almost everyone else in the world uses them, except the United States. For the Purple Line to move in that direction makes it a national model.
Using these long trains was one of the changes project officials made in response to Maryland Governor Hogan’s demands to reduce the Purple Line’s costs. One long railcar rather than two short ones coupled into a train saves money and keeps train capacity high enough to work.
Hogan’s other changes made the Purple Line a lot worse. They reduced train frequency, eliminated the direct transfer to Metro at Silver Spring, and reduced the electrical power of the line, limiting its capacity. But the move to longer railcars with open interiors may be a silver lining.
Crystal City transitway station. Photo by Arlington.
Georgia Avenue’s bus lanes will run just four blocks, from Florida Avenue to Barry Place. They’ll be curbside lanes, with normal bus stops on the sidewalk.
Location of Georgia Avenue bus lanes. Image from DC and Google.
Four blocks is short, but this location is specifically one of the slowest stretches WMATA’s busy 70-series bus line passes through. Bus lanes here will speed the entire line.
Just as importantly, this will be a test project for DDOT to study, and to learn about bus lane implementation. In May, crews will add red paint to the roadway to make the bus lanes more visually obvious. By adding the red surface later, DDOT will gather data on whether the red really does dissuade car drivers from using the lanes illegally.
Red-painted curbside bus lane in New York. Photo by NACTO.
If Georgia Avenue’s four block bus lanes prove successful, they could provide a model for the citywide transit lane network envisioned in moveDC. They could also one day form the backbone of a future Georgia Avenue streetcar.
The new Crystal City transitway section will run from Crystal City Metro south to Alexandria, where it will join the existing busway. It’ll be a mix of curbside bus lanes and fully exclusive bi-directional busway.
Crystal City transitway. Image by Arlington.
The DC region once had 60 miles of bus-only lanes. With these projects finally happening, and others like 16th Street on the horizon, it’s exciting to see a reborn network begin to take shape.
Closing Metro lines for months could work, but only if the region provides transit alternatives
WMATA may shut down entire rail lines for months in order to catch up on maintenance more quickly (though, officials noted, no decisions have been made yet). If a shutdown does happen, Metro must thoroughly prepare, communicate, and provide riders who rely on Metro with reasonable alternatives.
Special event bus shuttles. Photo by DDOT DC on Flickr.
A months-long shutdown may make sense
For years, WMATA has been struggling to perform maintenance at nights and on weekends. There’s so much work to do that they can’t complete it all.
A months-long shutdown would theoretically put an end to that, or at least significantly reduce the need for weekend track work. It would let WMATA catch up on all its maintenance needs for a line in one fell swoop. We’d be trading a few months of pain for years of happiness.
WMATA GM Paul Wiedefeld says he is indeed considering the idea. He adds, “In the last few years, we’ve been trying to do this [maintenance] in a sort of piecemeal way, and basically we’ve alienated everyone.”
He’s right. Working at night and on weekends is fine when you’re just doing preventative maintenance. But after years of increasingly terrible weekend service, it’s become clear that model won’t work with WMATA’s need for major rebuilding. With no end to the rebuilding in sight, it’s time to try something new.
This is getting old. Photo by the author.
Closing an entire line end to end may not prove necessary. You can rebuild the Virginia section of a line without closing the Maryland section, for example. And closing the core is much harder than closing outer sections. But closing long segments of a line, say four or five stations long, may well make a lot of sense. On the other hand, if a whole line needs work, maybe shutting it down completely is the way to go.
Whether or not WMATA has the equipment and work crew capacity to do such a big job is an open question. But if so, or if it can expand as needed to do so, it may not be a terrible idea.
Leaving riders without options wouldn’t be acceptable
So yes, it’s very possible that closing major segments of Metro lines for months would be the best way to get this painful decade of rebuilding behind us.
But we absolutely cannot simply shut down Metro and hope for the best. Metrorail is not an optional service for the Washington region. Hundreds of thousands of people rely on it every day, including many who don’t have access to cars. Telecommuting saved us for one day, but can’t work for months on end.
If this is really going to happen, WMATA and the affected jurisdictions would have to work together to provide transit alternatives. We’d need special bus shuttles to replace the shuttered Metro line, temporary bus lanes to make longer-distance bus travel fast enough to be practical, greatly expanded transportation demand management, and more. We’d need a comprehensive transportation management plan.
Such a plan might look something like what Matt Johnson suggested in 2013, when WMATA considered closing part of the Red Line for six weeks. That shutdown hasn’t happened, but many of the same ideas would be necessary anywhere.
It wouldn’t be easy, and it would require sacrifice from everyone, including drivers who don’t use Metro. It’s impossible to stop running a transit line that carries a hundred thousand passengers without making life hard. Substitute buses absolutely would not be as good as Metrorail, not for transit riders and not for car drivers who have to share road space.
But the current situation is hard too.
WMATA will need to study this concept in detail. Then they’ll need to share their detailed findings with the public. What are the real options, what are the trade-offs, how much time and money would this save, and what will the Metro system look like when it’s over? If Metro expects the public to buy this idea, they’ll need to be forthright.
But it’s possible. WMATA could do this and it might be successful, if and only if they take the time and money to plan, prepare, and do it correctly.
WMATA’s PlanItMetro blog has released a trove of data on Metro station use. Here’s one snippet: All 91 stations, ranked by the average number of riders who entered the faregates each weekday in February, 2016.
Rockville misses the forest for the trees with its plan for an 18-lane mega main street
Rockville Pike could one day become a 252-foot-wide mega boulevard with 12 car lanes, 4 bike lanes, 2 bus lanes, and over 50 feet of landscaping. But in designing a street with more than ample room for cars, bikes, and buses, planners abandon any hope the street will be walkable.
The plan for Rockville Pike. Image from Rockville.
Everybody gets a lane!
Rockville Pike is one the most important retail strip highways in the Washington region. Like most 20th Century retail roads, it’s designed for cars, and it carries a lot of them.
Rockville wants to make it a more urban main street, so planners there are drawing up a redevelopment plan. It’s a laudable goal, and it’s not easy on a high-traffic state highway like Rockville Pike.
At first glance, this plan has all the components of a good complete street design: Tree-lined sidwalks, protected bikeways, a center-running dedicated busway. Every mode gets all the street width it could possibly want.
And why not? Why go through the political headache of forcing the community to make the difficult choice between fewer car lanes versus bikes or BRT if you can fit everything in? With a mega boulevard like this, everybody gets what they want, and nobody loses. Right?
Walkability loses, and it’s the most important factor
At 252 feet wide, the new Rockville Pike will be practically impossible for pedestrians to cross. It will take multiple traffic light cycles and multiple minutes for anyone to cross.
Instead of a main street, Rockville will have a barrier. And that is a big problem for the rest of the plan.
Transit oriented development doesn’t work unless it’s walkable. If Rockville Pike is too wide, development on one side of the street will be effectively cut-off from development on the other side. Riders won’t be able to easily access the BRT stations. People will drive for even short trips. The concept of a community where people don’t need to drive everywhere will break down.
If you can’t walk, other multimodal options don’t work. Pedestrians are the linchpin to the whole thing.
To be sure, some level of compromise is always needed. If walkability were the only factor that mattered, all streets would be pedestrian-only. We add in car lanes, bike lanes, and transit because we have to make longer trips possible, and that’s a good thing.
But there’s a balance, and 252 feet veers so far to accommodate long distance travel that it seriously sacrifices short distance walking. In so doing, Rockville undermines the very foundation on which its redevelopment plans rest.
Make pedestrians a priority
The Pike needs to be narrower. Assuming the sidewalks, busway, and three general car lanes each direction are sacrosanct, that still leaves a lot of potential fat to trim.
Are the service roads really necessary if the plan also includes new parallel local streets? Do we really need redundant bi-direction bikeways next to both sidewalks? Could we possibly reduce the 74 feet of various landscaping, buffer, and turn lanes?
These would be difficult trade-offs, to be sure. But there are massive negative consequences to an uncrossable mega boulevard.
If Rockville wants the new Pike to work as multimodal urban place, pedestrians need to become a priority.