Knowledge is power, especially on bad days

Quick thoughts on SEPTA’s response this morning to the fire in Kensington across from York-Dauphin Station that shut down the El:

Obviously, the root cause of the mess was an enormous fire on someone else’s property that SEPTA could not have prevented, but since “Large Fires in Kensington” seems to be the new normal, at least until someone makes L&I get its shit together, SEPTA might as well have some good plans in the can for dealing with it.

Basically, there’s no way that SEPTA can really have enough spare buses on hand to deal with a disruption this large, on this important a piece of its rail system, at rush hour.  That goes for the El, that goes for the Subway, and that goes for the core of the RRD system.  That being said, they did about as well as could be expected today, pulling buses from multiple depots and off of other busy routes to run the bus bridge between Huntington and Berks.  SEPTA probably could have improved by doing a better job of telling bus riders across the city that their bus service had just taken a minor cut on a cold day, but that’s a relatively minor strike to have as your worst sin.

One thing that I think would be worthwhile in future disruptions, but the technology is a few years away, is push notifications to riders that they should seek alternate routes.  Many El riders in Kensington and the Lower Northeast connect to the El from crosstown bus routes, and would have been best served if they could be instructed to head west to a Broad Street Subway station, which would crowd those buses in a more distributed way, and take some of the load off of the shuttle operation.  That wouldn’t even require pushing notifications to individual riders, although I’m sure that that is coming in the smartphone era, but for major disruptions like this one, having announcements on board buses and scrolling on information displays on buses and at bus shelters would be a major help.

That makes it all the more terrible that the City of Philadelphia didn’t specify any kind of realtime schedule information displays in the new bus shelters it just contracted for with Titan Outdoor — they’re a great improvement in quality-of-life on a daily basis, but in a major disruption like today’s, information is critical.  Why that bad contract isn’t an electoral issue in the upcoming Mayoral and Council contests, I have no idea.  Unlike schools and SEPTA Regional Rail, it’s something where the Mayor and Council have actual authority over, and doesn’t even require City money, but they still muffed it badly.

EDIT: Actually, there is one ongoing SEPTA screwup that exacerbates these problems: the $1 transfer charge.  Consider someone in walking distance of K&A, commuting back and forth to Center City near Market Street.  They use tokens because it doesn’t make financial sense to buy a Transpass unless you transfer or make extra trips on weekends.  Even if they are told what is happening, they are going to be charged $1 for the privilege of taking the 60 to Broad Street and going around the fire zone in reasonable comfort, instead of throwing themselves into the teeth of a chaotic bus bridge operation.  How many Kensingtonians are going to be doing that voluntarily?  A lot fewer than if a transfer was included in the base fare, that’s for certain.

Platform Screen Doors NOW

The new hotness today comes from the unlikely place of the office of City Councillor David Oh (R-At Large), who told free-tabloid Metro yesterday that South Korean company TIS Inc. wants to install platform screen doors on the Market-Frankford and Broad Street Lines.

Platform screen doors (PSDs) are an old and mature safety technology in Asia and Europe, but they have yet to see a major deployment in the United States, apart from a handful of airport circulators and the Las Vegas Monorail.  (The Honolulu Metro, under construction, will open with them in 2017.)  They often go hand-in-hand with driverless systems, since PSDs require a signaling system that can stop the train in exactly the same spot (to within an inch or two) every single time, and also remove the possibility of a person falling, jumping, or being pushed onto the tracks, which removes the necessity of having a person watching out to apply a (probably futile) emergency brake.  In addition to the obvious and well-promoted safety features, PSDs also keep trash and detritus off the tracks, which SEPTA would otherwise have to remove during the off-hours, or risk having it become fuel for a third-rail-sparked track fire.  In addition to the increases in reliability, PSDs also allow for more aggressive train-handling, which shaves time off of the schedule and makes everyone’s trip faster.

TIS’s pitch to SEPTA is that they will install PSDs on SEPTA’s platforms gratis, in exchange for the right to sell advertising on the doors to recoup their investment.  If this is true, even in broad outlines, then SEPTA should be looking to sign on the dotted line as fast as it can find a pen.  Because if subway frequencies are uncoupled from labor costs, then SEPTA can run trains as frequently as the signal system permits, throughout the service day.  Other writers, including Jarrett Walker and Emily Washington, have delved into the link between driverless operation, high frequency (especially high off-peak frequency), and higher ridership, but suffice it to say that having a train that shows up every 5 minutes or less is pretty much the ideal scenario for transit.  And if TIS weren’t offering to foot the bill, and there weren’t a slew of other things SEPTA needed to take care of before they break (trolleys, trains, bridges and viaducts, substations), I’d rank PSD installation as a strong candidate for #1 Capital Budget Priority.

The devil, of course, will be in the details.  The signal systems on the Subway and El might need upgrades to be compatible with automated operation, as might the control systems on board the M-4 and B-IV fleets.  There is also the slim chance that the system turns out to be buggy or just a lemon, like the CBTC system installed by AdTranz in the Subway-Surface tunnel years ago.  If that happens, which is unlikely, but possible, SEPTA would need an exit clause to tear out or replace the PSDs, without getting sued for TIS’s expected revenue from future advertisements.  Speaking of advertisements, some people will hate them, but they’re for a good cause, and are hardly as disruptive as renaming an entire station.

TWU 234 will also be a stumbling block, insofar as they can be expected to object to anything that is potentially the first step to automating any of their members out of a job.  However, automation, even driverless operation in daily service, does not mean the elimination of all train driver positions.  For example: trips in, out, and within train yards are typically always handled by humans.  Also, TWU leadership might want to talk to veteran operators among their rank-and-file about the psychological toll of being at the controls of a train involved in a passenger fatality incident.  Mostly, it will be worth whatever fight SEPTA and its riders have to wage with the TWU, to get to driverless operation. I hope it won’t require much of a fight, but if it should even precipitate a strike, so be it.  PSDs are the best improvement SEPTA can make for itself, its workers, and (most importantly), the public it serves.

Home news

It’s been a very busy few weeks in Greater Philadelphia transportation news. The 24 hour Regional Rail strike. The continuing saga of Delaware’s I-495 bridge. The reintroduction of overnight weekend service on the El and Subway, which despite the lack of staffing at most stations, has so far been a smashing success. PATCO’s travails with the Ben Franklin Bridge reconstruction. Proposals for renaming 30th Street Station, and selling naming rights at Suburban Station.

I’ve barely written about any of it.

I’m sorry. I’ve been busy.

(Warning: pictures and self-indulgence after the jump)
Continue reading Home news

Philadelphia needs driverless subways

Tying a few recent threads together, there have been a lot of jokes on the internet in the wake of the recent derailment of a CTA Blue Line train at O’Hare Airport station in Chicago. But of course, as Eric Jaffe pointed out in The Atlantic Cities, we have the luxury of laughing because the train went off the end of the trackway in the dead of night, with a relative handful of people on board, resulting in none of the 30 injuries being life-threatening. We can also avoid the sorts of knee-jerk responses we see after accidents with nonzero death tolls, like last year’s derailment at Spuyten Duyvil in New York.

Both the O’Hare crash and the Spuyten Duyvil derailment have been linked, in early media reports, to operator fatigue. While there are many technical safeguards in place, and more coming online in the next few years, to prevent operator errors from causing accidents, it’s clear that those measures are not yet 100% effective. As SEPTA prepares to join CTA in the elite club of 24-hour subways, it’s worth noting that people are more fatigued overnight than during the day, and that there is a straight line between fatigue and error rates. (I have plenty of experience with this phenomenon, being someone who sees quite a lot of the other side of midnight.) SEPTA is going to see an increase in operator error rates if it runs overnight; that is an unavoidable fact.

Except, of course, that it is entirely avoidable, by removing the operator from the circuit entirely. This is a drum that Stephen J. Smith has been banging on for quite some time, but he’s right that the time is long overdue for automated trains to come to America. Automating trains, unlike automating cars, is a mature and well-understood technology, and has not only been installed on many new-built transit lines, but has been retrofitted to older lines. One of the better-known examples of a retrofitted line, at least among transit nerds, is the Paris Métro’s Line 1, which first opened in 1900, seven years before the Market Street Elevated, and completed the changeover to automated operation in 2012.

Like many automated lines, new and old, Line 1 paired automation with the installation of platform edge doors (PEDs), which is not a technical requirement of automation. Still, PEDs are a big safety and liability-reduction upgrade that comes relatively cheap bundled with automation, especially at the claustrophobic deathtrap that is City Hall Station BSL, as well as other platform pinchpoints throughout the system, mostly caused by poorly-placed stairways and elevators down to track level. We may have avoided the New York City Subway’s critical design error of placing ubiquitous support columns in close proximity to the platform edge in most places, but not quite everywhere. If PEDs can make the passenger experience at the Broad Street Line’s busiest station better, without bringing the eponymous building down, then I’m all for installing them at the first opportunity. About a dozen or so people per year die after being struck by SEPTA trains, and cutting into that number is also a worthwhile enterprise, although it’s noteworthy how much smaller that number is compared to the traffic violence on the streets and highways.

Driverless trains also provide financial encouragement to transit agencies to operate far more frequent service in off-peak hours. This effect is strongest in systems where train length can be adjusted based on hourly demand, as SEPTA theoretically can, but currently does not. With driver compensation out of the marginal cost equation, it’s approximately the same amount of money to run three two-car trains as one six-car train, but three times as many train departures is obviously much more attractive to riders, even if they have to walk down platforms a bit more than they already do.

But this week, there’s another issue that insistently argues in favor of automating Philadelphia’s rapid transit lines: industrial action. Labor negotiations between SEPTA and TWU 234 are continuing, and the rumor mill surrounding the process (TWU head Willie Brown is talking to reporters; SEPTA is not) suggests that, despite there not yet having been a strike authorization vote presented to the TWU 234 rank and file, the two sides are far enough apart that a strike sometime in the month of April is likely. In response to the imminent threat of the third SEPTA strike in eight years, State Representative Kate Harper (R-Montgomery) has introduced legislation that would completely remove TWU’s right to strike. Nevermind that TWU is reaching Italian levels of triggerhappiness when it comes to industrial action, and thus has a public approval ranging somewhere between Jack the Ripper and the United States Congress. I strongly believe that completely revoking TWU’s right to strike is an overreaction to its abuse of that right. It also might endanger SEPTA’s federal funding, to the detriment of labor and riders alike, although that’s more ambiguous. But driverless subway trains, while not actually laborless, can be run with skeleton crews of management replacements, which is an infamously bad idea for human-operated trains. Driving a train is highly specialized skilled labor, but cleaning a train at its terminal is not. If the subways kept running, even perhaps at reduced capacity, during a strike, alongside (mostly) strike-proof Regional Rail and a bikeshare system coming online a year too late to pick up slack during this round of troubles, then Philadelphia would be functional in a way that it is not during a transit strike today. That reduces the power of TWU’s bargaining position, but not nearly as much as it would be reduced by Rep. Harper’s bill, or by a clone of New York State’s Taylor Law, which may be the alternatives facing TWU. This is not a surrender by labor so much as it is nuclear disarmament; if conflict is inevitable, as Willie Brown seems to think (and this is his second act at the helm of TWU 234), then conflict should be thinkable. The other route, the nuclear deterrence option of Mutual Assured Destruction, doesn’t seem to be enough to keep TWU off the picket lines.

So yes, let’s add automating the subways to the queue of capital improvements in the Catching Up plan. It’s a nontrivial investment, but it will be paying dividends in blood, fortune, and labor relations for decades to come.

We’re getting late night subways, maybe! Is it really all that exciting?

(spoiler alert: maybe)

Never let it be said that SEPTA doesn’t listen to feedback. Local found-art blogger Conrad Benner (a/k/a Streets Dept.) posted a change.org petition to restore 24/7 rail service to the Subway and the El on Tuesday. On Wednesday, that petition had 1,000 signatures. And today (Thursday), SEPTA’s GM Joe Casey and CFO Rich Burnfield are on the record with Paul Nussbaum in the Inquirer about extending Subway and El service to 3:00a on Friday and Saturday nights. So much for change.org being the McDonalds of slacktivism! That’s got to be a new record for progress.

But it’s not really all that surprising. As Sandy Smith pointed out in the comments section of Phillymag, this is a Back to the Future move, since the BSL and MFL ran 24/7 until 1991. SEPTA bustituting its main rapid transit lines every night was not originally sold as a budget-trimming measure. It was because, in the recession of 1990-91, the homeless flooded into stations and concourses for shelter, and because train conductors handled fare collection overnight and station gates were left open (the practice still survives today at Chinatown Station in off-hours), there was no legal recourse to clear squatters off the platforms. This was a nuisance to riders and a perceived safety and security threat, so SEPTA ran up the white flag and switched to buses. The fact that Benner and so many of the other petitioners are citing the perceived security of station platforms being higher than the surrounding streets, would be completely fantastical to the proverbial Rip Van Winkle who spent the last 20 years in a coma.

It’s actually somewhat obscure what the bottom line impact is, for buses vs. rail. In 1991, the El and Subway still ran two-person train operations, and the massive bank of supercapacitors reclaiming electricity from braking trains was still science fiction, so the operations cost of running trains was much higher then as compared to now. (The sluggish and worn-out M-3 Almond Joys were still running on the El, too.) And the demand for overnight transit was far lower in 1991: Center City and the neighborhoods around it weren’t yet in the midst of a national urban renaissance as they are today. Nor did 1991 have the advanced policing techniques and cheap and ubiquitous CCTV cameras that we have today, which provide an alternative to having a SEPTA policeman on every train and every platform.

What I can testify to from personal experience, is that today the Owl buses do a very brisk business, all night, despite their unattractiveness to choice riders. My fiancée’s morning commute often involves taking the last Broad Street Owl bus of the morning to Suburban Station, and sometimes I’m right there with her. (The 5:00 Subway train from Fern Rock doesn’t get to City Hall in time to make the 5:26 RRD train to Delaware.) That bus is standing room only at Girard Avenue. While the people staggering out of bars at closing time are a loud constituency for overnight train service, the vast majority of the travel demand during those hours comes from people just trying to get to or from work.

So, full speed ahead for restoring SEPTA’s subways to their rightful former glory? Not so fast. There are a lot of factors that go into a decision like this. One reason SEPTA is in better shape today than it was 23 years ago, is that it’s made good use of all those overnight closures to bring the system back to a state of good repair, without reliance on frequent weekend closures or single-tracking. Compare this to PATCO, which retains 24/7 service but is now paying for that with an inability to conduct vital repairs to the Ben Franklin Bridge without melting down. Also, SEPTA is not going to open up its platforms overnight again; and that necessitates the existence of over 60 booth trolls to handle fare collection duties during the graveyard shift. Even if those were non-union positions, and they’re not, that’s a lot of cash to spend on salaries. Now, it’s true that a lot of those positions might be dispensed with after the full rollout of NPT, but that might negate the security advantage of being off the sidewalks in the middle of the night. There are other positions that SEPTA doesn’t currently need to staff overnight, like dispatchers in the subway control center at 1234 Market Street.

Some of the cost considerations are actually in favor of restoring rail service; still others are of indeterminate valence. With service every 15 minutes, El service can be covered by six train operators, as opposed to ten bus drivers. The wear and tear of operating a full six-car train is definitely more than the cost of running a single 40-foot bus, but is it 67% higher? Probably yes. Is the cost of electric traction higher than the bill for diesel fuel? Probably not, given regenerative power and off-off-peak electricity costs. Can SEPTA offset maintenance costs by running shorter trains? It would require breaking up sets in 69th Street Yard, which would require additional labor there, but you could start doing that at 9:00p, and not wait for midnight. The M-4s have been run in 4-car sets on Sundays in the past, but nobody outside SEPTA knows if 2-car sets can be run on the El. The Subway ought to be able to run the two-car trains that run on the Ridge Spur during the day, assuming that trains don’t gap out and lose contact with the third rail at the crossover north of Pattison Station.

And even with SEPTA Police being awesome and cameras being festooned off of every piece of property SEPTA owns, it’s still true that we don’t know for sure that stations are safer than street corners at 3:00a. They might, but they might not. Perceptions are one thing, statistics are another. And not all perceptions are created equal; as I pointed out in a comment to Phillymag, I feel perfectly safe taking the Owl buses at whatever hour, “because I have the privilege associated with being a 30 year old, 5’10”, 200+lbs. man of visually indeterminate ethnicity”. Not everyone is so fortunate. And it’s cynical, but somebody at SEPTA has to have noticed that they bear no liability for anything that happens to anybody waiting on a lonely stretch of sidewalk, while they can and certainly will be sued for anything that happens on a station platform.

With all this legitimate uncertainty about the pros and cons, it’s no surprise that SEPTA is playing it very safe, and is only talking about a pilot program to extend train service until 3:00a on Friday and Saturday nights. (The speed with which they responded, though, indicates it was something already being tossed around conference rooms at 1234 Market St.) Given the friction involved in the changeover from rail to bus and back, with buses covering routes 40% slower than their steel-wheeled counterparts, and given that the trains start service at 5:00 sharp, I fully expect any extension to 3:00 to actually be an all-night service. There’s no benefit to be gained from closing for two hours, except possibly spite. I also expect a Friday-Saturday pilot to succeed, since SEPTA’s other nightlife-oriented service pilot, the late night trains on the Manayunk-Norristown, Trenton, and Paoli-Thorndale Lines on Friday and Saturday nights, are still running after several years. And those serve a much smaller market, bringing the barhoppers of Manayunk and Center City (only) back to their dorms in Overbrook, Villanova, and New Brunswick. (Yes, I’m stereotyping. But so was whoever cooked up the schedule, and they were right.)

Am I in favor of restoring overnight rail service the other five nights of the week? Not really. I can make a solid argument that there’s enough nightlife on Thursday to treat it equally with the traditional weekend, and I would really like it if the trains ran even a half hour later at night and earlier in the morning.

But hats off to Conrad Benner, for moving the Overton Window for this far enough for SEPTA to seize the opportunity to be the good guy. After all, #SEPTA247 makes a great hashtag and great slogan; “SEPTA 20/4 and 24/3” doesn’t get the blood going quite so well. See you all on the late train to Fern Rock.

Mister Gorbachev, open this gate!

Commenter Noah asks the following:

Do you know more about this, from Wikipedia?

“SEPTA’s Market-Frankford Line (also known as the “El”) and all of SEPTA’s Subway-Surface Lines stop at the 30th Street subway station, less than 1/2 block (< 1/10 mile) from the southwest entrance to 30th Street Station. A tunnel connecting the underground subway station and 30th Street Station was closed due to crime and vagrancy concerns."

http://en.wikipedia.org/wiki/30th_Street_Station

Whose decision would it be to re-open this tunnel? I've always been surprised they're not connected underground.

I’m not quite sure if SEPTA or Amtrak owns that tunnel, but the portal on the Amtrak station side now has Bridgewater’s Bar on top of it, so it would be a significant disruption to reopen the tunnel to the public. On the SEPTA side, you can see the stairs down to the passageway underneath the stairs up to street level at the Northwest corner of 30th and Market.

Photo of tunnel portal from SEPTA subway station
The tunnel to the Amtrak station is through the gate on the right — you can see the handrail going down.

Honestly, the decision to close the tunnel was correct at the time. It’s too far from SEPTA-side cashiers and Amtrak-side shopkeepers, and there’s no good sightlines into the tunnel, for it to have eyes-on-the-street security. Today, the option of cheap cameras supplementing occasional foot patrols exists, to possibly provide a middle ground between 24/7 patrolling and closure. I would recommend separate cameras, controlled by Amtrak PD, SEPTA PD, and PPD, for operational clarity and redundancy.

There is one very good side benefit of the closure, although I will be the first to admit it doesn’t look like a benefit when you’re there: because all connecting passengers have to cross 30th Street on the surface, it creates a large flow of foot traffic across that intersection, which helps calm traffic coming in from I-76 and re-acclimates drivers to the city street grid. That’s a hard benefit to quantify, but it’s there. It also tends to create a lot of delays for vehicles coming off of 30th Street, thanks to aggressive mass jaywalking, which I approve of, because jaywalking is the sign of a civilized society. Unfortunately, buses also get caught in those delays. Can’t win ’em all…

Ask not for whom the booth trolls

Rev. Michael Caine, Friend of the Blog and pastor at Old First Reformed UCC in Old City, shares the following story (edited slightly for readability):

Today the SEPTA token booth “non-worker” couldn’t tell when the next southbound Broad Street Spur would come.  I asked, “do you know if there are scheduled times or do they come every so many minutes or just when they come?” She got all loud and responded, “now how would I know that?” It actually made me laugh out loud! I didn’t bother to reply, “because you are [the one] inside the token booth!”

The non-token booth is possibly the most aggravating aspect of being a Subway or El rider; a person is required to accept cash fares and to sell and accept paper transfers, but does not sell tokens or any other fare instruments, nor do they seem to be able to tell the confused rider the most basic information.  This seems to be a setup for frustration and crushed hopes. The most visible of all SEPTA frontline employees, and the most findable, can not help with the most basic of customer service tasks.  No wonder that local message board regulars have dubbed them “booth trolls”, whether that is for their resemblance to mythic creatures or internet pranksters.  Rev. Caine probably has a leg up on the average harried commuter or confused tourist in terms of ability to retain equanimity in such trying circumstances.

This raises the question: whither the booth workers, come the Fall and NPT?  With NPT bringing fare vending machines to every Subway and El station, and the end of paper transfers, the booth troll may be an endangered species; they will have no functions left to do.  Such is the price of progress.  Hopefully, if SEPTA allows the booth workers an opportunity to transfer to other jobs in the agency, they won’t be customer-facing, unless intensive customer service retraining is required.