Monday, January 16, 2012

Looking at SEPTA

Lately, Alon Levy of Pedestrian Observations has been taking a long, hard look at the MBTA commuter rail operation and came to the conclusion that most of the MBTA's plans to expand the physical scope of its operations were misguided. Instead, his suggestions were to maximize the efficiency of the existing operations, primarily with the use of electrification and infill stations, likely with a long look toward uniting the disparate networks operating out of North and South Stations (although the long-term aftereffects of the Big Dig boondoggle will make that unlikely for a generation or more). SEPTA took the same approach with its Regional Rail network thirty years ago, as I pointed out in the comments, and there is no clear evidence it worked better or worse. In any event, however, I want to undertake the same kind of analysis as Alon's, looking mostly at PA (NJ is another story altogether).

This analysis will start along the lower Delaware and work more-or-less clockwise.

1. The Wilmington-Newark Line happens to be one of the three commuter rail lines that also utilizes high-speed designated lines. Since SEPTA has not relocated its stations to park-and-rides on the urban fringe, and since the Marcus Hook local works together with the local and express trains to Wilmington and Newark as well as the express Northeast Regional and Acela trains (which stop at Wilmington), service on this line is adequate--1 tph at the worst; 4 tph at rush hour. Wilmington is intercity, however; a little dinky service between Marcus Hook and Newark may be called for to handle local service there. East of Island Avenue this line operates exactly halfway between the two closest-together branches of the Subway-Surface trolley network.

2. The Airport Line, SEPTA's newest line, suffers from underridership (as do most airport lines that only serve airports). Not helping this one bit is the Airport approach this line takes, a massive fishhook. A new Airport approach would be useful, shortening and speeding up the line; commuter service would return to the line at Eastwick (and if high-speed rail is desired along an Airport routing, a little further to the east as well). Service would be returned to the original routing (IIRC the Reading Chester Branch) via a viaduct over PA 291 (Governor Printz Bvld./Industrial Hwy.) with a stop in Essington before interlocking with the NEC just west of Eddystone.

This also offers a new service paradigm for expresses, and also creates a problem with two local services both heading toward a natural terminus in Chester; one of these (the Airport one, it would have fewer stops and thus reach Chester faster) can be valved up the Chester Creek Branch--which has just begun a rail-trail conversion--to about Aston; any further would overlap with the West Chester Line (etc.: see below) complex. Bringing heavy rail to the Airport would be a wise idea--it's an employment hub. If this happens, heavy rail can use the existing Airport stations and the current I-95 viaduct for end-of-line storage (if desired) while commuter rail--and possibly high-speed rail--would have a concentrated stop at one station as close to the middle of the concourse as possible.

Finally, the current Airport Line junction with the NEC is inferior, and particularly so if a high-speed routing is desired; the technically superior routing would go straight up the Chester Branch past the interlocking with the CSX main and (under, most probably) the CSX cutoff between its main and the High Line. Infill stations would be desirable--at the Post Office, 63rd St., and Bartram's Gardens would offer the greatest balance between distance, jobs access, and residential access.

3. The Media/Elwyn Line displays the core problem with SEPTA Regional Rail: it has lost service since 1980, when the diesel lines were all shed, with one exception. Current service terminates at Elwyn, which is in the middle of nowhere (technically on the outskirts of Media) and has no park-and-ride. As you can expect, the ridership there is low.

This is not, however, the historical state of things. This line is known as the West Chester Branch, and for good reason: it goes to West Chester. Rails still exist all the way to West Chester. The West Chester Railroad serves freight customers along this line. It was electrified all the way to West Chester; the catenary poles still exist.

Of course, at the end, service was highly infrequent, and the lack of maintenance along the line made for a bumpy ride. SEPTA terminated commuter service to West Chester and beyond Paoli in the late '80s, and then offered a choice to Chester County, so the story goes, which to reopen.

However, West Chester is a natural terminus, and would draw reasonable ridership. The line passes by two schools (West Chester and Cheyney Universities), and the area this extension covers is still a core Philadelphia commuter market. Service on such a line should be hourly at the least (although service on all existing SEPTA lines is hourly at minimum already).

4. The Octoraro Branch refers to a disused route splitting off from the West Chester Branch around Lima and heading west to Kennett Square and eventually Nottingham via Oxford. The line east of the Brandywine Creek was apparently washed out in 1977 and has lain abandoned since; to the west, it is operated by the East Penn Rwy., a short line running several spur operations within the Philadelphia area and Lehigh Valley. Freight traffic runs all the way to the Herr's snack food plant in Nottingham, close to the Maryland border; at one point, this route connected into the Northeast Corridor via a Susquehanna River sub, and would have functioned as an alternative freight gateway for the PRR.

Kennett Square is an agricultural job center, and Lincoln University, a bit further on, offers a ridership draw. Additionally, the route passes Chadds Ford, a major regional draw (the home of Longwood Gardens and the Brandywine Museum). Finally, there is no direct highway route from the Chadds Ford area to the city. A good case can be made for hourly trains to Kennett Square, with bihourly trains to Oxford (Lincoln University).

5. The Newtown Square Branch is an old PRR branch serviced by diesels; service beyond Grassland (Eagle Road, perhaps? If anybody knows anything more, can they fill me in?) stopped in 1963, and all service terminated in 1981. Commuter traffic on this line was light for much of its history; it competed with a rather more direct trolley line running down West Chester Pike. Reinstating this trolley service (now Rt. 104) should be prioritized. However, this does not mean that this route should be allowed to decay away (as has been happening); it needs to be banked for the long term, and those parts that have already been lost returned to it. A rail-trail would be the most beneficial option here--the banking is a just-in-case if light rail along West Chester Pike ever reached saturation.

6. The Paoli/Thorndale Line is the second of the three Philadelphia-area commuter lines along high-speed designated corridors; the fastest express service is the Keystone Service, a push-pull utilizing AEM7s, Amfleet coaches, and ex-Metroliner EMUs converted into cab cars. Service right now terminates at Thorndale, which, like Elwyn, is in the middle of nowhere; the natural terminus for this line would be Coatesville (Thorn Interlocking allows turning; a Coatesville or Parkesburg terminus would need to turn all the way out at Lancaster. IIRC, there is a project to improve Parke Interlocking; if this allows for turning, this would make Parkesburg the natural operations terminus of this line).

One of the issues of this line, however, is that it's starting to venture into intercity territory. Lancaster is the main commuter market beyond Coatesville; the least frequent of all trains are the ones beyond Malvern (the "Paoli Local" terminus), the only ones on SEPTA's system which have bihourly frequency.

7. The Cynwyd Line is too damn short. Service needs to be extended to at least a park-and-ride off the Schuylkill Expressway, or possibly Ivy Ridge (where it was in 1986). If extended to Ivy Ridge, a connection could be made with the Manayunk/Norristown Line, thereby increasing service flexibility upstream.

8. The Manayunk/Norristown Line has much higher ridership today than in 1986, when it split riders with the Ivy Ridge line. Extensions of this line to Reading have been on and off the drawing boards for the past twenty years; the Schuylkill Valley Metro disaster basically sunk all of SEPTA's expansion plans for the past decade. Such a route to Reading would, however, be intercity; the natural termini beyond Norristown are Phoenixville and Pottstown, to which extensions can come in that order. Add to that that the line has plenty of capacity, while its primary competitor (the Schuylkill Expressway) is famously ever over-capacity, and this route has some of the region's better potential. This is also the passenger line that comes closest to King of Prussia; unfortunately, it does so in the middle of a giant office park. Perhaps shuttle services can help alleviate the problems that come with that.

9. The Perkiomen Branch saw its last service way back in the 1950s but suburban growth (sprawl) since means that service to Collegeville, with the built-in ridership boost its name suggests is there, and eventually the East Greenville-Pennsburg-Red Hill area. The major problem with this is that this line has been broken in numerous places, including in one place for a golf course (!). That area, however, is much more a part of the Allentown commuter market; a Collegeville spur would make for a better rail restart option.

10. The Chestnut Hill East and West Lines are two semi-overlapping commuter rail lines entirely within the Philadelphia commuter market. Their major problems stem from their infrastructural heritage; Chestnut Hill West has a queerly roundabout routing from North Philadelphia into Center City courtesy of its PRR heritage.

The best fix would be maintain one as commuter rail and convert the other into heavy rail. The usual proposal would be to maintain Chestnut Hill East as commuter rail because it enters the SEPTA Main Line at Wayne Jct., while converting Chestnut Hill West into heavy rail, entering the Broad Street Line around the North Broad/North Philadelphia transit node.

However, this has some major disadvantages: (1) while CHW has that roundabout Center City gateway, its routing north of that is straighter and therefore preferable for commuter rail, and its corollary, (2) CHE is much curvier, and (3) CHW passes mainly through residential areas while CHE comes quite close to Germantown--a major commercial artery--several times.

Combining this with my proposal that my Society Hill-Mt. Airy Line conception, we get a different idea, however. With a 19th St. subway to Wayne Jct., we can use Chestnut Hill East as a heavy rail routing and, with a Swampoodle Connection linking Chestnut Hill West into the SEPTA Main Line, give that line much more direct Center City service. Converting CHE to heavy rail furthermore has the advantage of giving us equipment that can better handle, and in the long run, a superior technical framework for eliminating the sharpest of, that line's curves (Wister Woods and Germantown). This routing also gives us a route that is better equipped to be extended--to Chestnut Hill College, another natural ridership generator a mile away--say. Note: Stenton and Ivy Hill are both definitely underserved by the current options. I may have more ink to spill on that later.

11. The Lansdale/Doylestown Line is the core route of the SEPTA Main Line, and one of the busiest on the Reading side. However, Lansdale, while a major North Penn area center and a natural terminus, is not the natural terminus. Service originally extended to Bethlehem and occasionally Allentown; were local service extended up to the North Penn area's other natural termini (Perkasie and Quakertown) and Bethlehem service extended, (service extension to Perkasie is currently being studied; the relatively bare-bones idea is upgrading the tracks to ridable condition and electrifying the line while utilizing existing infrastructure for stations (to be upgraded later), and extending Lansdale locals) Lansdale would be quite a busy rail junction, leading us to...

12. The Stony Creek Branch, a freight cutoff between Lansdale and Norristown. The primary attraction of this line is service to Merck, which lies right on it; with a busier Bethlehem Branch beyond Lansdale, it would also function as an express cutoff avoiding traffic snarls downstream (particularly in the busy two-track section between Jenkintown and Wayne Jct.)

13. The Warminster Line is the line that currently best accesses the part of Philadelphia's northern commuter market with the worst highway access, and hence has high ridership. Beyond Warminster the line continued as the New Hope Branch, today the New Hope & Ivyland Railroad. This route should be extended to a natural terminus at Ivyland; the next one up is New Hope, which, while a regional draw, would require too extensive an electrification to be viable until better routes are dealt with.

14. The Newtown Branch (part of which is the Fox Chase Line) was the last diesel commuter rail line SEPTA operated, ceasing service in 1983, after SEPTA had run its RDCs into the ground. While the Warminster Line passes Willow Grove, the Newtown Branch has, in aggregate, the best rail route in the northern suburbs' commuter rail market. Warminster Line saturation also suggests it's time to invest in it once again. PA-TEC has an ingenious (if rather park-and-ride heavy) plan to extend service to Huntingdon Valley, a key foot in the door for actual Newtown Branch line restoration--the line's natural terminus is Newtown itself. The Newtown Branch crosses the New York Branch at Ayres Interlocking in Beth Ayres.

15. The West Trenton Line follows the New York Branch from Jenkintown east. At Beth Ayres it crosses the Newtown Branch; it terminates at West Trenton in Ewing Township, another of those stations with nothing around it. For whatever reason, the Reading chose to terminate its Philadelphia commuter service here, and the infrastructure needed for it hence still exists. It is an operational terminus: Yardley is the more natural terminus. Since it's already in New Jersey, two possibilities can be considered: extend it to NJ Transit, or extend NJ Transit down to it, with a shared terminus at Yardley. Such a route would be considered part of the Raritan Valley Line (and short of its electrification there is no good reason to extend SEPTA service to Bound Brook).

16. The New York Short Line is a freight cutoff between the Newtown Branch and the New York Branch with some useful passenger rail possibilities. West Trenton trains using the NYSL would eliminate the need to go through Jenkintown, thus saving time; it would also avoid needing to reinstate the level interlocking at Ayres and send potential Newtown Branch trains through Jenkintown. An unfortunate byproduct of this idea, however, is that it would eliminate service to Philmont, Forest Hills, and Somerton stations. However, an NYSL express cutoff is still a useful idea: West Trenton express trains run local beyond Trevose (which is where the NYSL-West Trenton Line interlocking is).

17. The North Penn Sub, as I recall its name, is the route extending from Fern Rock south, principally along American Street. It was originally the North Penn R.R.'s mainline--hence its name. The entirety of the ROW still exists; a smart move may be to bring light rail service up it, especially if new light rail service is to go along Delaware Avenue (although that is not the smartest proposition right now).

18. The Bustleton Branch is a short branch between Holmesburg Jct. and the Bustleton area. While the section north of Roosevelt Bvld. has been obliterated, the portion south is still intact, and a spur heads from it to Northeast Philadelphia Airport. This is minor expansion written all over it, and would make a good terminus area for a local service from the interior of the Trenton Line. It would also be helpful to extend a spur from the NYSL to this branch, as the current setup forces the not-inconsiderable amount of freight traffic generated here onto the NEC.

19. The Trenton Line--at last!--is the NEC's route north from Philadelphia and is hence the third high-speed designated line in the city. It has sharp curves at Zoo Interlocking and Frankford Jct.; as I've noted before, there is no cost-effective way to do something other than put up with Zoo as the junction between the NEC and Keystone Corridor; for true high-speed service, a Frankford Jct. cutoff along Torresdale Ave. can be arranged. (A little concrete is better than a lot.) Trenton trains are reasonably frequent, and timetables for this line also include a section of the connecting NJT Penn Line timetable to New York--a rare manifestation of organizational competence in the passenger experience. So rare is it that the two can't even be bothered to time their trains for cross-platform transfers! Or accept one another's tickets!

Infill stations should be considered along the Trenton Line, particularly at Brewerytown, Fairhill, and Juniata Park.

Commuter rail in Philadelphia is a different animal than in Boston. Where in Boston concentrating on optimizing existing service would be more useful than less and less tenuous greenfield extensions, in Philadelphia, development and larger towns tend to exist beyond the edges of existing lines. I hope to get a map of these ideas going; additionally, I would like to see what kinds of service patterns would optimize our resources.


  1. I'll have more to say later, but for now, just one question: how well does SEPTA serve King of Prussia?

    1. Well, there's no good rail access to the area. The closest commuter train currently terminates at Norristown. SEPTA has on the drawing boards a plan to extend a Norristown High-Speed Line (NHSL) branch to King of Prussia via a routing parallel to a Norfolk Southern freight cutoff. Unfortunately, commuter rail along this cutoff is unlikely: it doesn't pass through any substantial towns and has poor connectivity with the SEPTA network.

      King of Prussia is a major bus hub, however. Two buses from Center City serve it, as do a number of suburban buses which extend to larger towns like Norristown and West Chester.

      However, that said, the bus network within the Norristown-King of Prussia-Plymouth Meeting-Conshohocken area is very confusing and almost disorienting. I do not think it is possible, for example, to catch a bus directly from downtown Conshohocken to King of Prussia. You'd have to transfer at Norristown or Gulph Mills.

      Compounding these problems is that, outside of the (generally low-wage) jobs concentrated at the malls, the King of Prussia and Plymouth Meeting areas--they run into one another--are diffused out in the worst kind of office park hell you can think of. Worse still are the facts that this extends along Rt. 202 to Chesterbrook and Rt. 422 to Oaks. This is an area that is hostile to public transportation.

    2. The primary problem in regards to all of the low performing Frontier Division routes is that they all have to meet certain performance standards or else they must be adjusted through schedule and/or routing changes to bring their performance up or else it can be eliminated. I do commend 1234 Market on trying to maximize every vehicle they can but in the case of the Lower Bucks County Routes (127-130, 133) and the Norristown secondary routes (92, 95, 97, 98, 99 past KOP, 131, 139) they are all habitually low performing routes that gets adjusted almost every three years so they can never get a stable ridership base as compared to the 93 or the 96. As it stands now, there is about another year before those routes I have mentioned will be reviewed again so I do expect more changes come 2013.

  2. Five buses: 92 (westbound service ~90 minute interval)
    99 (to Pheonixville and Norristown) 30 min interval
    124 (to 69th St) 30 min interval
    125 (to Center City) hourly service
    139 (to Limerick and Pheonixville) hourly service
    It's also 2 miles to the nearest stop on a high-speed trolley to 69th St.

    1. Matt, just to correct one thing, the NHSL is not a ”High speed Trolley” it has its own Railroad with no grade crossings. Also runs on third RAIL power as our market frankford Line.

  3. Matt: Quick fix.

    92 runs straight south, to Exton. Bihourly. IIRC SEPTA floated a plan to extend it to West Chester last summer.
    99 runs northwest, to Phoenixville, and east, to Norristown. Hourly.
    123 runs south to 69th Street, Market-Frankford transfer, 30 mins.
    124 runs west, to Chesterbrook, and southeast, to Center City, hourly.
    125 runs north, to Valley Forge, and southeast, to Center City, hourly.
    139 runs to Limerick via Phoenixville, hourly.

    Nearest rail stops are on the NHSL. DeKalb St. on the 99, and Gulph Mills on the 124 and 125.

    Hope that helps.

    1. The 92 is basically one massive loop starting in Exton, looping through West Chester (a year and a half ago the 92 was merged into contract route 314 and that was the result) going back to Exton before running out to King of Prussia via Malvern, Paoli, and Chesterbrook. It is a LONG ride......

      The 124 and 125 are two of SEPTA's "Premium Routes" with the others being the 150 (Plymouth Meeting to PARX Casino via PA Turnpike) and the 78 (Cornwells Heights PR to Center City/30th Street IRS via I-95). The fare for all three lines is Base Fare plus three zones. If I am going to pay that much, then give us a few more perks like anything other than that fabric covered aluminum that is the norm and some warning that if the Schuylkill is backed up, the drivers can go local route instead of sitting in traffic.

  4. Speaking as someone who has to travel along 422--I would much rather drive to a station like Pottstown, Stowe, or even Exeter Twp and take a train in all the way to 30th St or even just to Norristown and transfer onto the NHSL & then the MFL, than put up with my current hassle with that abominable highway.