Organization and Electronics vs. Concrete in Washington

There’s a discussion going on at Greater Greater Washington about future expansions of the Washington Metro, adding more coverage and capacity; read both the posts and the comments, because there are great debates about just how much concrete really is needed. The post itself mentions various possibilities Metro has been looking into, a few good and many really awful.

Part of it is that the nature of such discussions favors concrete – it’s much easier to discuss a fantasy map than schedules and organization. Indeed, on my three regional rail posts on The Transport Politic, most comments concerned the proposed through-routing map and infrastructure to be built rather than schedule integration. The reason the comments on the GGW post are so good is that many eschew this and instead talk about other things. Even the idea of separating the Blue Line from the Orange Line in the city, which looks sound to me and is not yet another outbound extension to the exurbs, is suspect and there’s a serious suggestion to build light rail to relieve the capacity problem instead.

Discussed in the comments but not by Metro is the possibility of converting the commuter lines to rapid transit. Only one, the Penn Line running along the Northeast Corridor to and beyond Baltimore, is even electrified, and the rest are owned by CSX and Norfolk Southern. This would be far superior to adding more outbound Metro extensions, which have very high costs: the Dulles extension is $180 million per km despite being predominantly above ground.

The Washington Metro, and even more so BART, is more an S-Bahn or RER system than a subway. The stop spacing is very wide, and the lines branch out and go deep into suburbia. Unlike BART, Metro sometimes gets it right and has good transit-oriented development, though it too has its share of parking lot stations. The main difference is that due to poor organization (FRA regulations, pure agency inertia), the Washington Metro exclusively uses greenfield alignments, whereas S-Bahn and RER systems use predominantly existing commuter lines, with strategic tunnels built to provide service to the urban core.

There are two potential problems with relying on legacy commuter lines, aside from organizational difficulties that should be the first to be tackled. First, those lines may have capacity problems; this is not true for the Penn Line, but may be true for the other, lesser-used lines, because of freight conflicts. Second, the lines may not run to the desired destinations. Both concerns can be mitigated at much lower costs than pouring concrete on new lines.

First, the Penn Line is 64 km long from Baltimore to Washington, has 8 stations, and has no sharp curves except at the Baltimore and Potomac Tunnels, a $750 million replacement for which has already been studied. A 160 km/h train with the acceleration profile of the FLIRT and 30-second dwell times at intermediate stations (easy with level boarding) could do it in 36 minutes. Add Swiss-standard 7% schedule padding and about 30 seconds for slowdown through the B & P replacement and this is a total of 39 minutes end-to-end. A speed boost to 200 km/h would save maybe 2.5 minutes, since regional trains accelerate slowly at high speed. Local trains currently do the trip in an hour.

The Acela, a substandard train, is currently scheduled at 38 minutes, so all trains would travel at about the same speed, eliminating all capacity problems. (Current peak throughput is 7 tph, so the slight unevenness in the travel speed is a non-issue.) A mild Acela speedup, involving trains running at 200 km/h with no slowdowns, would speed the train to 22 minutes, and require just one mid-line overtake if peak traffic is to be 4 tph each for Acela and commuter trains. 6 tph each would require two overtakes and a lot of discipline, but would be doable given the capabilities of ERTMS. Full-fat high-speed rail in the Northeast would do the same trip in about 16.5 minutes, and require the line to be fully four-tracked; this is already part of MARC’s long-term plan. There is in other words no real problem with capacity as far as conflicts with intercity trains go.

Second, one often-overlooked point about S-Bahn/RER networks is that they have a fair amount of greenfield track, often in tunnel, constructed strategically to connect to important destinations off the existing rail network. For example, tunneled alignments bring regional trains to Charles de Gaulle and Zurich Airports. If there’s an important suburban destination not reached by Metro or a rapid transit system based on the five existing commuter lines, it should be fine to construct a spur – for example, the extension of Metro to Tysons Corner is a great idea. However, such spurs should be kept as short as possible, especially airport spurs, since airport connectors tend to underperform.

Look again now at the suburban lines proposed by Metro in the link. The Brown Line is a duplicate of the Red Line, which has no serious capacity issues except at its center. The Beltway Line skips the major centers that a circumferential should hit (for example, Arlington/Alexandria), defeating the entire purpose of a greenfield alignment. The outbound extensions would just create more transit-oriented sprawl, with people driving to stations and taking trains only at the peak. And it would all cost much more than electrifying track, purchasing good rolling stock, and running it with high schedule discipline.


  1. jim

    South of Washington, the CSX RoW as far as Alexandria used to be four tracked (it was still four-tracked when I first moved to Alexandria in 1980). The NS RoW between Alexandria and Manassas has very light freight traffic — a work train between Van Dorn and Gainesville and a unit ethanol train to Van Dorn for transhipment — and is grade separated north of Clifton. So restore the tracks north of Alexandria, electrify say twenty miles south of Washington (to Burke), buy ten to a dozen trainsets and run Baltimore-Burke on fifteen minute headways. There’d need to be some rebuild of stations to support the additional pax flow (particularly at L’Enfant Plaza and Alexandria which would be natural Metrorail transfer points) . Washington Union Station has four (of eight) through tracks completely unused: tracks 27 and 28 can’t even be accessed from the station. They’d have to be re-enabled and concourse access built. But this is small change compared to building new heavy rail.

  2. David Alexander

    The Washington Metro, and even more so BART, is more an S-Bahn or RER system than a subway

    My personal comparisons have been to compare WMATA to a not as good Berlin S-Bahn and BART to a not as good Frankfurt S-Bahn. The former is usable for inter-city trips while the latter is much more of a suburb to urban core high frequency network. In theory, with railway reforms, both could have been achieved with UIC-ish style rolling stock and new tunnels connecting the older RR ROW or providing new coverage on their respective greenfield lines.

    though it too has its share of parking lot stations

    How do you trick Americans into riding buses. 🙂

    • dejv

      > How do you trick Americans into riding buses.

      Time connections and integrate fares so the trip requires a single ticket/pass.

  3. ant6n

    I find it strange that they want to untangle the downtown lines running on the same track – surely there are cheaper ways to increase downtown capacity, rather than building redundant tunnels.

    For S-Bahn systems, branching is a feature, not a bug. It means that as you go further and further from the central city, the frequency decreases – and thus follows the potential ridership. In Munich, 7 lines – three at 10min, the rest at 20min intervals – converge downtown to 2min intervals.

    • David Alexander

      surely there are cheaper ways to increase downtown capacity

      Railfan rumour is that WMATA is theoretically capable of 2 million riders per day with full 8 car trains running at the mythical 90 second headway. I suspect current management feels that current service is at “maximum” capacity and new lines are needed for more trains, and I suspect there’s an institutional bias toward building more and having as much area covered as possible with service.

    • Alex B.

      The problem is that there is more demand for Metro on the branched portions of the system than the current headways can support. Metro is already working on different service patterns to better make use of the existing capacity, but these aren’t the greatest long-term solutions.

      The fundamental issue is that while Metro may have been planned as more of a hybrid commuter, S-Bahn type system, that’s not how it is being used within the core – it gets used as a regular ol’ subway.

      Branching can still be a feature, but the branch points need to be further out – hence the desire to untangle the portions of track that are currently interlined through the core.

  4. Steve

    I was doing some research on commuter gaps in the D.C. area and was surprised to discover a ca. 1870 map showing only seven usable commuter lines in D.C.–all of which (three by MARC, two by VRE, and two by Metrorail itself) are in use today.

    There are gaps, of course, but in general D.C.’s periphery is as well served as it was 130 years ago. Any new concrete should be directed to (a) fill in service gaps in the urban core, and (b) fill in gaps in the metropolitan system–primarily along the Potomac–but neither should be done until the current network is optimized, both in terms of organization and in terms of infrastructure modernization (i.e. electrification of ex-B&O lines).

  5. Alon Levy

    The reason I’m more sympathetic to the Blue Line project and the inner parts of the Silver Line is not so much capacity or service to the airport as the need to serve Georgetown and Tysons Corner, which are big gaps in the existing rail infrastructure. It goes without saying that capacity should be resolved by modernizing the operations first; Metro has automatic train operation, but both the system and the workforce have a lot of problems with safety, capacity, and speed. If those two projects are built, there would still be branching on the Orange/Silver Line, which as far as I understand is the system’s most crowded.

    Getting Americans to take connecting buses is mainly an issue of integrating fares and raising bus stop spacing (traditionally 400-500 meters in Europe vs. 200-250 in North America) to the point that buses are faster than walking. But best practice for regional rail is also to count on many walking trips. The Germans locate stations near where people live, decrease stop spacing and use higher-acceleration trains to compensate, and avoid horns and other loud noises that make living near a railroad undesirable.

    • David Alexander

      raising bus stop spacing

      Which runs into the problem where people will complain that the stops are too far, especially for the older riders. One of the reasons that I stopped riding the bus in Queens was that I could beat the bus to the F at 179th Street (or Queens Village) by taking direct back roads over the bus’s indirect routing to the E & J at Jamaica Centre. If it wasn’t for the alternate side of the street parking rules, I suspect that more people would do that. OTOH, there will always be some politician to pander to the people that want a bus stop on every block to minimize their walk, especially for the para-handicapped people who are too able bodied to use para-transit, but suffer from pain-induced condition like arthritis that hinders their mobility. Hell, they’re the reason why the MTA blows money on express buses in Queens and Brooklyn…

      FWIW, I wouldn’t mind taking the bus to my current home station since the parking lot is usually filled by 8 AM, but it needs to run at coordinated schedules with the railroad and it needs to operate until at least midnight or so. The problem is that the peak riders are workers that earn “big money” so they can afford the cost of gas easily and they would demand parking, and politically, they would get it. The planning and political structures here are just as much of a problem as the management at the transit agencies.

      integrating fares

      That’s less a political matter and that requires some degree of cost shaving to reduce the lost costs from the free transfers given the fetish over farebox recovery, but also coordination between disparate agencies. Even at an agency like NJT, the bus and rail operations seem like two different firms operating within the same area.

      avoid horns

      The FRA demands the horn, so they’re going to keep blasting them. Hell, even NYCTA blows a horn at 15 mph when it bypasses a station. There really seems to be some fear that people will get hit by an oncoming train.

      • Alon Levy

        If I can read the ZVV finance and revenue statistics correctly – and that’s a very big if – the farebox recovery there is 69%, and that’s counting capital charges, i.e. depreciation. Nothing in the US is as high; New York City Transit has a farebox recovery of about 40%. Add in the terrible finances of express buses and what you see when agencies cry poverty regarding free transfers is less an interest in cost recovery and more an interest in extracting money from riders, regardless of the consequences to ridership and operating costs. Transit agencies are not the only body that does this; airlines do the same with their charging for everything, and so do tourist traps. At least legacy US airlines are losing market share to Southwest and JetBlue, which don’t destroy the environment any more than the legacy airlines do; transit agencies have nothing to lose mode share to except cars, which do.

    • Alex B.

      Yesterday marked the 2-year anniversary from Metro’s red line crash that killed 9 people. Trains have been running on manual control ever since, as they have not yet been able to isolate and fix the problems with the ATO system. I’m hopeful they can get that system back in working order (and improve upon its design limitations) soon.

      That said, even those kinds of operations and electronic improvements can’t fully address some of the core capacity issues.

  6. ralphgarboushiang

    I’m not much of a transit expert — more of an advocate and a fan — so I defer to others on techincal train issues. But as someone who rides WMATA every day and exits and enters at Metro Center each weekday rush hour, I would point out that capacity on the trains themselves is not the only problem. Indeed, I am not sure that crowding on the trains themselves is a major problem, at least right now. The trains are crowded but not beyond what I have seen in other cities at rush hour and certainly not beyond what one would reasonably expect at rush hour. However, capacity of the core transfer stations is a serious problem. Part of it is poor station design — putting the escalators and stairs in the middle of the station creates crowds in the middle portion of the platform while the ends are relatively uncrowded. The escalators in the middle also make for narrow platofrms that create bottlenecks. As currently designed, those stations simply cannot (safely or comfortably, at least) handle any more human bodies than they currently do. Metro Center and Gallery Place are a nightmare at rush hour. Even on a Saturday afternoon, transferring at Metro Center can try one’s patience. My guess would be that when the core stations were designed back in the “nobody rides transit and downtown DC is dying anyway days of the 60s and 70s”, none of the planners or engineers imagined that those stations would ever serve the volume of people they do today.

    • ant6n

      Tow notes:
      – I think turnstiles can create bottlenecks at stations, and kind of ‘trap’ people on the stations. I think without turnstiles, passengers crowds might more easily flow out of stations if there’s crowding.
      – Some subway systems always have at least two exits. This will mean that people come in at both ends of the stations, and crowd toward the middle – I think (not sure, though) that that can result in better distribution of people on stations/trains. At the same time this can put subway-entries 200m or more apart – which will reduce walking distance to the next station/increase the coverage area of the station/allow for greater station distance.

  7. Mike O

    > I wouldn’t mind taking the bus to my current home station since the parking lot is usually filled by 8 AM, but it needs to run at coordinated schedules with the railroad and it needs to operate until at least midnight or so.

    This is an interesting paradigm, and it alsomshows up with BART. BART has three types of commuters: Finance types are on New York time, and travel between 5 AM to 7 AM, and return starting around 1 PM. Business types commute around 7 or 8 (until the lot fills) and return around 5 to 7 PM. Retail (which includes tourism and hospitality types – 10 percent of SF’s GDP) commute between 9 and 11 AM,, and come home after 7 PM. However, many business types stay in the city for after-hours meetings, etc., and return around 10 PM.

    Parking, especially parking structures, can be expensive to build, but in the area five miles around most Suburban stations, there is lots of space for parking. Smart commuters may park at Walnut Creek’s Lescher center (10 minute walk, or Pleasant Hill’s Embassy Suites, but most won’t figure this out. But if there was a shuttle that took people to these satellite lots, running til midnight (and it would not have to meet every train, it would make BART usable between 8 AM and Noon.

  8. USDOT guy

    Washington Metro does have capacity problems. The “core” parts of the system should have been built with three or four tracks, as was NY’s system — not just to handle the crowds, but also to permit service diversions during track maintenance without totally disrupting service.

    But the biggest problem is that Metro is very badly managed. The operating staff can’t seem to maintain consistent scheduled headways. This in turn causes delays at the points where lines converge. How many times has everyone heard, “There is a train servicing the platform at L’Enfant Plaze. We will be moving momentarily” (insert your favorite transfer station name here).

    And of course, not only are the stations poorly designed to handle large numbers of people, but most of the escalators are broken at any one time.

    So before spending any capital dollars on line extensions, Metro needs to figure out how to run what they have.

Leave a Reply

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

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

Twitter picture

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

Facebook photo

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

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.