Don’t know what you got, ’til it’s gone

As I believe I’ve mentioned, I moved this summer, and I had a six week period in between when my old lease in Point Breeze ran out and when my new place in Francisville was ready for move-in. In the meantime, I crashed in the spare bedroom of a friend and former flatmate in Swarthmore Borough. And, while I’m immensely grateful for the hospitality, it was a soul-crushing experience to be constantly reminded, for the entire time I was out there, how terrible it is to be transit-reliant in the suburbs.

The sad part is, it doesn’t have to be this way. Swarthmore has comparatively excellent transit service for a suburb. It’s traditionally the top station on the Media/Elwyn Line by ridership. The 109 bus, which runs through Swarthmore on PA-320 on its way from 69th Street to Chester, is one of the best in the Victory Division for frequency (and, relatedly, ridership). And yet service, by absolute standards, is just not that good. The Media/Elwyn Line runs once-hourly outside of the rush hour peaks, which is fine for a pre-planned trip to a scheduled event in Center City, but no good for a more spontaneous walk-up trip. The 109 has the speed and comfort drawbacks of buses, and only goes to the asphalt wasteland of the Baltimore Pike STROAD corridor in one direction, or Chester in the other. Chester is either the region’s most distressed or most undervalued asset; one can connect to almost anywhere in southeastern Delaware County, as well as the Airport, Wilmington, and Newark, but the density of lines on the map belies their inconvenience. Most of the bus routes at Chester TC run once an hour; the Wilmington/Newark line runs once an hour to Pennsylvania points, and less often than that to Delaware. Meanwhile, the rider experience of actually making a transfer at Chester TC is marred by the obvious signs of severe and prolonged economic distress that confront you in literally every direction in Pennsylvania’s oldest city.

Contrast this to my new flat, still piled high with boxes and disassembled shelves. I am literally around the corner from the Girard Avenue stop on the Broad Street Line, where locals run 5-8 tph through much of the service day, plus express and Ridge Spur service, often good for another 5 tph and 4 tph, respectively. The 4 and 16 buses provide additional service on Broad Street, and the 2 runs every 20 minutes off-peak on 16th and 17th Streets. The 15 trolley (currently bustituted for track and platform work, for the balance of August) is plainly visible, but barely audible, from my front stoop. It runs every 15 minutes. These are the frequencies at which I no longer even check to see when the next trip is, I just put my sandals on and start walking, unless I’m connecting to an infrequent Regional Rail line. The very act of checking schedules is as likely to prolong my wait time by causing me to miss a train, trolley, or bus, as it is to shorten my wait time at all.

Now, granted, this is a location that is fantastically well-served, even by city standards. That’s not an accident, given my search criteria. And maybe I’m just spoiled. But go back to that point I made about the service being so frequent that I don’t check schedules for ordinary, nonconnecting rides. This is the psychological hump that most people need to satisfy before they will consider living without access to a car. I recognize that I’m personally unusual in my willingness to choose transit over driving, even when I have the unrestricted choice (I live in a household with more than one working adult, but only one car, a slowly shrinking demographic, according to Jon Geeting’s crunching of Census Bureau data). But I think that the basic concept of transportation you don’t have to think about is the critical one, and that the personal details are going to be mostly trivial. In the city, in addition to walking or biking around neighborhoods, transit can fill that role. In the suburbs, even in dense, walkable/bikable suburbs like Swarthmore, it can’t. Or, more precisely, as of now, it doesn’t.

So, what frequency qualifies as “frequent enough”? Obviously, this is not going to be the same number for everyone, nor is it even going to hold equal across modes, nor should it. The average person is prepared to wait longer for a faster ride, a more comfortable ride, a more predictable and reliable schedule, or waiting in a place that offers more protection from the elements. Of course, on all of those points, it provides an edge to grade-separated rail over mixed-traffic bus routes, with intermediate-order transit modes occupying intermediate positions. But for guidance, we may do well to look to the great bus transit capital of America, Los Angeles. In 2006, LACMTA published a map of all routes that ran every 12 minutes or fewer at midday, on the stated assumption that it was the service frequency that allowed riders to dispense with carrying timetables. Later versions of the map, including this one from August 2012 [PDF], relaxed that condition to 15 minutes. The original 12 minute criterion is probably the best for local buses in city traffic, but the 15 minute criterion takes in all of LA’s skeletal rail transit system at midday, which would explain the change. So, as a first approximation, I would say that “frequent enough” headways here in SEPTAland are 12-15 minutes on local buses, 15-20 minutes on trolleys and light rail, and 20 minutes on short Regional Rail lines, and 30 minutes on long Regional Rail lines. Clearly, that’s a long way away for most of the system, outside of rush hour, but it’s a good set of goals for the highest-priority lines.

In home news, I’m going to still be digging out from boxes for the next while, but I’ll try to get back to a full posting schedule before Labor Day.

Join the Conversation


  1. I think this is why PATCO is so succesful on the Jersey side of the river. I have friends who take it who never look at the schedule before they show up. I only look at the schedule because I hate just missing a train. The headways are ridiculously frequent at rush peaks, like every 4-5 minutes, abs even after, theyre every 15-20 til really late (and trains still run even after *that*; the line operates 24/7).

    I was shocked to hear how infrequently SEPTA trains run even to points in the city. Only every hour to Manayunk? How in the world is this ok. I just don’t get it.

    1. Precisely. PATCO is amazing, because it’s subway headways over commuter rail distances. That’s the promise held out by most postwar/Great Society subways in the US, but the requirement for backward mechanical compatibility with Broad Street Line kept PATCO out of the engineering weeds (looking at you, BART and WMATA).

      I’m going to keep banging on this drum until I break it: once an hour to Manayunk is very much NOT OK. Because it’s not just Manayunk, it’s also northern Strawberry Mansion, lower East Falls, Conshohocken, Norristown, and points beyond. It really ought to be the best branch in the system.

      1. I would not consider Lindenwold to be Regional Rail distance. Not compared to the scope of the system on the Pennsy side of the river. Lindenwold is about 14 miles from City Hall. That’s not quite the edge of the old RRD Zone 3. Depending on how you do the math, it gets you to Chester, Media, Wayne, Conshy, Oreland, Willow Grove, Bethayres, and Cornwells if you really stretch.

        PATCO is great. There’s no question of that. But the post-war systems are apples and oranges with Regional Rail. Both WMATA and BART are supported by regional rail (WMATA covers about the same distance from the center of DC as does PATCO). There’s a good argument to be made for converting the inner portion of SEPTA to this transit-style of operations. But the FRA would probably have something to say about it.

        1. FRA should have no problem with transit style of operations, as long as the rolling stock is FRA-compatible. For relatively-short consists of Silverliners, that’s not a huge penalty, unlike HSR. The real problem is going to be on the labor side; staffing has to come down if frquencies are to go up without breaking the bank. The established ways to do that are gating the system (PATCO’s solution) or POP. SEPTA’s NPT implementation on RRD (partial gating) is an attempt at a third way, and I don’t think it’s going to work.

          My goal would to get Media, Norristown, the Chestnut Hills, and Fox Chase more like PATCO than like Paoli and Lansdale. You can’t quite ramp up the frequencies to PATCO levels, but you can come close.

          1. Silly me, I was thinking full-on transitization, not just increased headways. Yeah, no FRA problems there.

            NPT is going to be a giant fail on the RRD.

            I might add Chester and Bryn Mawr to your list. No reason why you can’t run inner service more frequent than outer service. I still like Vuchic’s original idea of running everything going past Bryn Mawr express with separate Bryn Mawr locals. With a reconfiguration of the interlocking it should be easily doable (I’d love to see high island platforms, but it’ll never happen).

            1. For Chester: the interlockings needed to do Chester short turns don’t exist. I’m not sure there are any left at all between PHIL (Airport Line) and HOOK (Marcus Hook). Amtrak is plain not going to play ball on building them back, either. It’s not as though there are load balancing issues on the WIlmington Line outside of rush hour. Quite the opposite. Really you want to send all of the off-peak trains to Wilmington to pick up that traffic for cheap. Chester just doesn’t get you enough by itself; there’s significant westbound traffic from Chester in the morning.

              For Bryn Mawr: Again, I’m extremely leery of relying on Amtrak being co-operative here. Otherwise, I agree, nothing stopping it. There’s been some rumbling that PRIIA-related maneuverings have Amtrak’s long-term hold on the Keystones in jeopardy, which if true, might make for some very…creative stopping patterns on a unified SEPTA Main Line service. Also in play is a plan to move the local/express changeover point from Bryn Mawr to Villanova, with associated interlocking changes.

              Do not get me started about island platforms at any Main Line station, but especially Bryn Mawr, Ardmore, or Overbrook. They make way too much sense, but even the abortive plan for redoing Ardmore from a few years back didn’t include them.

  2. The problem is that ridership levels largely justify the frequency of service. It’d be a great experiment to slash headways and see what happens to ridership, but who’s going to pay for that? SEPTA? The City? The State?

    The problem with trying to get TOD going is that the land isn’t available. The regional rail lines have been around for a very long time, and everything near them is largely developed. So some sort of partnership with a developer at a large (by SEPTA standards) surface lot like, say, Jenkintown (boy, wouldn’t the NIMBY’s love that) wouldn’t get you much. There’s about 4 acres there. Not much room for a parking deck plus mixed-use development. Certainly nothing that’s going to generate enough trips to justify adding extra trains.

    What would happen if, say, Lower Merion, Radnor, and Tredyffrin Townships chipped in to pay for additional service to Paoli? Two extra weekday midday trains (15 min headways), an extra train in the evenings (30 min headways), and an extra train on Sundays (30 min headways). Is there anywhere else on the system where such a thing could happen?

    1. For the funding side: that’s its own post. I have a few ideas there, but I would point out that break-even ridership per RRD run is just not that high, at least with just the marginal costs of labor and power. SEPTA doesn’t need to completely replace driving (good luck with that), but the ridership to justify higher service frequencies is achievable largely through induced demand. I would want that to be added incrementally, in case I’m wrong, for the most part. In the first round, for the ex-RDG side, I would add 1 tph to Norristown, 1 tph to West Trenton, and 2 tph to CHE (while kneecapping the 23 and other local buses!). Nothing to Colmar, nothing to Doylestown, Fox Chase and Warminster gets back-burnered to round 2 along with the third Norristown tph.

      Real TOD involves upzoning along Broad Street, not blackfield development around RRD stations. Maybe a different story on PATCO, NJT, or somewhere else with oceanic parking lots, but SEPTA is pretty good on not having done that really anywhere other than Cornwells Heights, where it makes sense.

      Putting together a three-township (had to double-check that Haverford was being omitted deliberately) coalition for increased Paoli line service is a decent thought, especially if Lancaster or Montgomery Ave gets ripped up for anything. That said, the Paoli Line’s priority is probably high platforms, which is going to need the townships pitching in on planning and historic issues, and assuming current conditions continue, capital funding. They’ll get significantly faster service on locals out of it, though. Also, I don’t see that the jump from 3 tph to 4 tph (20 to 15 minute headways) is really good enough to warrant the money. I could be wrong there. Still, if the townships pay for only 1 tph to SEPTA’s 2, you’re still going to see fairly even 20 minute headways.

      1. I don’t think kneecapping the 23 will work. Look at the area bounded by Chelten Ave, Germantown, Sedgwick, and Allen Lane. Most of Germantown Ave is 1/2 mile away from a stop. It’s worse if you only pump up CHE. The rail lines and the avenue are only close enough together to be interchangeable on top of the hill.

        Otherwise, I think you’re on the right track on the RDG, though I’d be tempted to do Fox Chase instead of West Trenton.

        TOD in the suburbs is a no go. But that’s where most of the RRD stops are. I agree the subways ought to be the focus there.

        I didn’t really think about Haverford Twp, but it’s better served by the Rt. 100, the 103, and the 104. And I suspect that’s a separate post too.

  3. Indeed, this is one of the reasons I moved from Philadelphia to Boston. As a transit-reliant person who needed to get around the suburbs a lot, it was generally quite annoying to get anywhere in a reasonable length of time. Boston, of course, has dramatically worse headways for commuter rail, but is also configured such that it is much less necessary to use it.

    As a side note, given all this talk of which stations deserve high level platforms… what the heck did Primos do to deserve one? As a former user of that station, I was somewhat flabbergasted the first time I came back and saw the upgrade, while so many stations with higher ridership are ignored. About the only think I can think of is that there wasn’t really any existing construction to get in the way, unlike other stops with buildings more substantial than a pre-fab trailer, but even so…

    1. As far as I know, the primary reasons Primos was first were need and cost: busier stations like Swarthmore and Media already use mini-highs to comply with ADA, and it was dead simple to drop the prefab sections of platform into place, with the same one-size-fits-all components that SEPTA used at Ft Washington, Ambler, Wayne, and Croydon. At the schedule change this month, all the peak hour expresses to Secane became expresses to Primos, which is a big tell that the new platforms have been very good for ridership. Rumor has it that Secane is next, when funding can be scraped together.

Leave a comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: