13 December 2012

Why Island Platforms Rule

Many words have been used to describe the myriad reasons why island platforms are best, and why express train traffic should be on the outside pair of tracks.  To amplify those points, here is a picture of what a Caltrain station could be.

In a true public transit system, trains and buses should connect seamlessly.  Anyone who has transferred between Caltrain and buses knows that the transfer is rarely direct; it usually involves a circuitous path on foot, occasionally requiring an umbrella, followed by a slow bus ride around a bunch of convoluted islands that are often configured opposite to the bus' main direction of travel.  Combine that with irregular timetables that are poorly adhered to, and what you've got is a time-wasting, ridership-killing mess of a connection.

It doesn't have to be this way, and the addition of passing tracks for the blended system is a perfect opportunity to rectify the situation with seamless transfer stations that:
  • provide the shortest route from one vehicle to another
  • allow each vehicle to pick up or drop off passengers without going out of its way
  • provide shelter from the elements
  • provide permeable access to pedestrian, bike, bus, and other road traffic
  • enable coordinated clock-face schedules


  1. You are assuming that Caltrain can run on schedule. My daily experience suggests otherwise.

    Also I thought Caltrain and HSR were planning on different platform heights...? It would be pretty nice if Caltrain can figure out that level boarding platforms will help with ingress/egress.

  2. I belive Caltrain will run ON schedule after electrification and level boarding. EMU will reduce the mechanical failire, level boarding will reduce uncertain time addition of wheel chair lift usage.

    1. If only Caltrain was planning (let alone even thinking about) level boarding ...

      I've yet to hear or read anything from Caltrain about level (roll-on, roll-off) boarding. Seems like the now-or-never-in-our-lifetimes moment is upon Caltrain with electrification and fleet replacement. CEO Scanlon and his team are seriously lacking drive and vision toward what Caltrain can and should be. Sad. Very sad.

      Did I miss it, or is Caltrain as completely and inexplicably "out to lunch" on level boarding as they are on numerous other subjects Clem has been blogging about for years now?

    2. The good news is that Scanlon talked to the board about level boarding a couple of meetings ago. The challenge is that it will require getting rid of an obsolete CPUC rule designed to let freight employees hang off a side of a train standing on a step to look for obstacles. Not clear yet whether eliminating the rule will take administrative or legislative action.

    3. Adina,

      It would have taken one phone call to Feinstein's or Boxer's office any time in the last 10 years.

      I know, because a high-up Boxer aide told me so.

      Scanlon and :"CBOSS Bob" Doty know this because I relayed that information to them in person.

      Everybody at Caltrain, without exception, is criminally incompetent and interested solely in destroying public funds and public facilities. The agency must be destroyed.

  3. It will take more than nicely designed stations and clockface schedules to enable train/bus integration.

    It will take a change in objectives, success metrics, and strategy for the transit system. When MTC looked at transit performance for the transit sustainability project, they only looked at trunk train and bus lines, and measured their success by passenger miles. They wanted to leave local bus service to the agencies, and had no concept or measure for connectivity.

    In the region, trains are commonly seen as transit for the rich, and public buses are seen as transit for the poor. This perception is held by transit agency boards and staff. So you'll see a slow bus running parallel to a faster train route; the train is the high-priced, high quality product, and the bus is the discount low-end product. Never mind the Google buses, their rubber tires are different somehow.

    There isn't a strong concept that different vehicle technologies have different properties that enable them to serve different roles in a single transit system system. Pricing systems are designed per service, with very weak transfer and cross-service pricing.

    Until this mindset changes, "transit centers" will be buildings that happen to serve trains and buses, and are not designed to help passengers change from one to the other.

  4. One downside to island platforms is that they are less efficient for passenger circulation. Imagine a situation (in Clem's diagram) where a train has just offloaded a large number of passengers, and they are all coming down those stairs. Someone trying to go up the stairs is like a salmon swimming upstream.

    1. Mr Engineer,

      You're decribing a bat-shit insane passenger circulation, such as that at Berkeley BART or at the $4 billion(!!!!!!!!) Transbay Terminal, involving single direction escalators and single width stairways. That is a recipe for disaster. (Everything in the Transbay "design" is a disaster, without exception.) While bat-shit insane passenger circulation is certainly what Caltrain and CHSRA and America's Finest Transportation Planners deliver, it turns out that it isn't the only possibility.

      In contrast, nobody with even a single functioning brain cell in their heads would design or construct a new station with such uni-direction vertical circulation. You'll note that Clem's sketch shows pairs of full-width adjacent stairs and/or escalators.

      For really high-throughput platforms (eg any major city central station new construction, any modern metro) one can go with pairs of escalators flanked by or flanking pairs of stairs. But we're not talking Caltrain or the SF Peninsula then.

      In summary: island platforms work just fine if you don't employ certified morons as your "architects" and "planners". Just cut-and-paste Clem's diagram or cut-and-paste any contemporary design from the first world and you'll be fine.

    2. Yeah, as someone who has recently started to commute to work by bike and BART, I"m constantly amazed at how poorly BART stations are designed. Like Fremont BART, one elevator (like all BART elevators small and slow and nasty), one set of narrow stairs, and one up and one down escalator. I'd think concrete ramps, like you see in sports stadiums, would be a massive improvement for getting people (and bikes, strollers, etc) up to BART's elevated stations (which is like all of the above ground stations).

    3. Not to criticize Clem's artistic skills, but the stair/escalator in the diagram does not seem nearly wide enough. And yeah, there could be multiple banks of stairs and escalators -- but for what purpose? There is nothing inherently advantageous about a center island platform for bus-train xfer.

      For example: your standard Dutch commuter rail station. BTW, note the bike ramp leading straight up to the platform.

    4. What to do if F-S-S-F thinking does NOT prevail? Two islands between each F-S pair to allow cross platform transfers ( the zero-MPH passing track option)? Or possibly our worst fear: two outside platforms?

    5. @Drunk: maybe the island platform isn't inherently better for bus-train transfers, but it is better for other reasons.

      1) Operational flexibility (wrong way running without platform change)
      2) Easy turnback sidings beyond the island, with no interference to express
      3) No confusion among users about which platform to use, easy way finding
      4) Wider comfortable platforms with more room away from moving trains
      5) 1x Amenities (TVMs, information signs, lighting, shelters, benches, etc.) instead of 2x
      6) Fewer opportunities to be isolated (crime)

      Need I go on?

    6. Mr Engineer,

      Reading off the pixels in Clem's diagram, the stair/escalators depicted appear to be about 3m wide total. Given that a single 1m tread escalator is good for 6000 people/hour (at 0.5m/s) or 8200 (at 0.75m/s), a pair of those fitting inside 3m is hardly unrealistic for the suburbs or Caltrain (or sub-suburbs like San Jose.)

      For a typical low- to medium-traffic suburban (which is what all Caltrain's stations are, including SF, will be), a stairway width of 2.4m is typical European practice. So again ~3.0m is not unrealistic, if not mega-over-generous.

      Just FYI, the under-construction Stuttgart 21 underground main station, whose platforms will see many many times the traffic of any Caltrain stations', ever, features multiple banks of 1.65m(escalator)+2.6m(stair)+1.65m(escalator), with 2m of clear space on either side, all fitting into 10.0m platforms. (Those dimensions are outside-to-outside, so the escalator step width is more like ~1.3m and the stairs ~2.4m) (Some photos of contemporary German station platforms and stairs here if you care.)

      As an aside, a huge advantage of elevated stations is that the cutout "holes" in the platforms needed for escalators/elevators going down are quite small -- on the order of 7m in length -- whereas the obstruction presented by stairs/elevators ascending up from a platform is over twice that! Watch your head!

      And re your Dutch Almere Muziekwijk link: those ramps are for wheelchairs, not bikes. Also: 1. No riding allowed!; 2. Able-bodied cyclists carry up stairs rather than waste several minutes walking on long detour ramps; 3. Bikes get locked up and left outside, not carried or ridden onto the train anyway.

    7. Richard,
      It isn't clear if a stair or escalator was intended in the diagram. For 2 escalators, then I agree there is no issue. But for just stairs, the crowd coming off the train will fill the available width. I find that hugely annoying when "salmoning" up the stairs to catch a train about to leave.

      And yes, should have put "bike ramp" in quotes -- in my experience, those ramps are for bikes (regardless of what the signs say).

    8. The diagram can be tweaked in about 5 minutes. The escalator is from an older diagram that I hacked up. If I had to redo it, I'd use a T-shaped stair with the bottom section splitting in each direction of the sidewalk. Just the elevator, and no escalator. The stair needs to be split into sections anyway, 12 ft max rise for each.

    9. At major transfer stations, people would be going up to empty platforms anyway. Explanation: if you have a timed connection at :00 between regional trains in both direction and a local bus, then because the trains need to keep moving and can't dwell too long, the trains in both directions will arrive around :00 and depart at :00 also, plus or minus 1 minute. The buses will arrive a few minutes before, say :55, and depart a few minutes after, say :05. So in reality, the sequence of events is as follows: bus arrives, you get up to a platform without any disembarking train riders on it, trains arrive, people get off, you get on the train, people get down from the platform to the bus bay.

    10. Adirondacker1280019 December, 2012 23:01

      At major stations the people going up ( or down ) to the platforms can use the escalators running in "their" direction. A station with so little passenger traffic that it doesn't have escalators won't have a problem with hoards of people going in any direction.

  5. The ideal center-platform rail transit stations for most stops should have the following properties:
    1) The station’s center-boarding-platform should be narrow enough to tolerate a constant spacing between SS tracks within a FSSF scheme for most of the corridor length in order to permit straight express tracks and allow for future platform lengthening and infill-stations without repositioning main-line tracks but wide enough to allow for single track gap-train storage between stations. Driverless train operation combined with widely distributed gap-train storage throughout the corridor would permit immediate gap train insertion the second the case where an added train would improve service becomes apparent. Platform edge barriers, necessary for driverless train operation, would permit all but 6 inches of the entire platform width to be usable for passenger occupation even when trains are rolling past the platform edge at 200 km/hr. Arriving train air-suspension pressure measurements could be used to estimate that train’s load distribution; information that could be sent ahead to the train’s next-stop in order to inform waiting passengers where to arrange themselves along the platform to be near empty train seats. Well optimized waiting passenger distribution along platforms becomes increasingly important as train lengths are extended; a cost-effective approach as traffic increases. (Doubling train lengths in a situation where close-up time and dwell times are near to equal, will increase station capacity by 85%.) Widely distributed easily read and interpreted station flat-panel displays could shorten dwell times; a low-cost way to increase station capacity compared to adding parallel tracks and platforms. Also convenient cross-platform or same platform transfers are more likely possible with a compact high traffic density per track station design.
    2) Train-access-platforms should be below grade but no more than necessary below over-crossing bridges in order ease walk-in access and minimize grade-separation construction cost. A narrow train access platform width design could still allow plenty of access ramp width available at both platform ends to accommodate ticket vending machines and fare payment confirmation barriers. The station center should be reserved for a bank of elevators connected to every floor of two high rises adjacent to both sides of the railroad right-of-way. For close to the railway tracks high density transit oriented development cases a plaza directly above the tracks would be an efficient noise attenuation approach. Note: An elevated track-way station design would likely discourage private sector investment in a tightly integrated transit-oriented-development described above.
    3) A station design that most effectively discourages crime is when the entire length of all passenger walk-ways can be seen from most surrounding at-grade overlook positions. A shallow-depth open-cut center-platform station with transparent ramp and platform edge barriers station design appears to go a long way toward meeting late-night mutual visibility security concerns.

  6. I probably am missing something basic here - but is this meant to be a station with no high speed rail service? It does not look to me like there is any way for anyone to board or exit the HSR trains on the outside tracks. Is the concept that either Caltrain or HSR trains would board using the platform in the middle - and the outside tracks are simply overtake lanes for trains (EITHER Caltrain or HSR) that don't stop at this station (which sounds like the kind of thing Clem would like - and that Caltrain and HSR planners seem to consistently reject)?

    1. This would be for a local station. For an express/HSR station, you would probably have two island platforms. Current plans show 4 side platforms ... somehow.

    2. Don, see http://caltrain-hsr.blogspot.com/2010/10/station-design-101.html

      The majority of Caltrain stations will be (won't be, but ought to be if there were anybody with a non-functional brain involved) either two tracks with a central island platform or four track with a central island between the inner tracks. (express tracks outside, express trains do not stop.)
      =*= or ==*==

      A smaller number of stations would be four tracks, two island platforms cross-platform transfers between local and express trains running in the same direction.

      (And perhaps there might be stations that look like *=*=* or =*== etc for various compromise reasons.)

      Which stations are "major" (more than two platform tracks) and which are "minor" (two platform tracks), which which are turnback locations, and which which provide train-train transfers are (or would be, ought to be, aren't being ...) decided by a combination of (first and foremost), a service plan that prioritizes passengers and their needs; by existing and potential station ridership; by availability of right of way, and by the logistics and economics of matching extra tracks to a desirable service plan.

    3. Thanks for the clarification. I see now that I confused myself by seeing all of the discussion of transfers between vehicles and thinking it referred not just to transfers between trains and buses (which was Clem's main point), but also to transfers between trains (which I think Clem still holds as good principles to follow, but which are not addressed in this kind of minor station design).

  7. Caltrain has posted the November JPB meeting CBOSS status presentation slides:
    Communication Based Overlay Signal System Project Status

  8. Local Policy Makers to Discuss Caltrain Upgrades Thursday
    The meeting, which is open to the public, will include an update on Caltrain electrification and high speed rail blended systems.

    See also: Caltrain Modernization Program

  9. Great graphic but you miss the opportunity to illustrate how lightweight the upper shade/rain structure could be. No need for complete coverage of the HSR tracks... an airy almost sail like structure would be sufficient.

  10. I think that is the gantry from which the catenary lines are suspended, not a shade/rain structure.

    1. Correct. It's probably not the most graceful gantry, but outside supports keep the configuration as narrow as practical and the platforms free of clutter. The real thing is probably not going to win any style points, if the relatively recent New Haven - Boston electrification project is any indication

    2. Clem, it turns out (as any field trip to recent-ish construction Central Europe will reveal) that if the overhead system is designed in conjunction with the passenger amenities, then there is no clutter.

      If, on the other hand, you have America's Finest Transportation Planning Professionals, then you end up with Caltrain's electrification 'design", in which the placement of stanchions makes absolutely no acknowledgement that stations or platforms or passengers even exist. (Check out Palo Alto, California Avenue and Santa Clara stations and observe how fucking wonderfully they place the concrete cutouts for future poles right where they are most in the way. Because that's what our own globally unique "design standards" that we happened pulled out of our own rears say we "must" do.

      Using standard German (Re200) catenary design and standard German platform canopies (what? PASSENGER AMENITIES? protection from the elements? NEVER in California) and standard HSR and S-Bahn compatible platform height, here's a cross-section of what a standard FSSF single island platform including simple overhead looks like. Not much clutter. Very little or no passenger obstruction beyond that needed for other platform furniture or canopy support.

      And here's what that sort of thing looks like in real life (Dresden Mitte). The clutter is underwhelming. The utility of obvious. The lack of US involvement is manifest.

      It can be done. All you need to do is have a designer involved and give a damn about passengers. If on the other hand you just have US-quality transportation "engineers" crap out stuff from their CAD system, all you get is maximum cost and a huge "F YOU!" to train riders and to transportation system efficiency and effectiveness.

    3. Adirondacker1280022 December, 2012 22:50

      Is it going to matter much where the concrete cutouts are if they are putting in level boarding?
      They don't need to travel outside of the US to see reasonably well designed stations with level boarding and OCS. The Northeast is fairly lousy with them. So is the Southside of Chicago.

    4. Adirondacker1280023 December, 2012 21:02

      what's wrong with it? IIRC it's in a floodplain, Not an ideal place to be building underpasses The stationhouse is a historic building so they had to work around that too. You'll notice that the light stanchion is level boarding ready. Why they didn't put in level boarding while they were busy building the overpass is a good question. Probably budget constraints. They didn't put in accomadations for eventual electrification because they were building the overpass because electrification was coming.
      They had to build an underpass or an overpass because it wouldn't have been a good idea to have passengers walking across the tracks when the tracks host trains that go 150 MPH through the station. The reason for the electrification.

    5. "what's wrong with it?"

      Dear God.

    6. Adirondacker1280023 December, 2012 22:08

      yes I want to know what's wrong with it other than they didn't spend the money to put in level boarding.

    7. Well I'm not entirely sure what the critical flaw is, but I can see a few issues:

      - Very little space between the tower/catenary poles and the edge of the platform. Now, presumably this is a low volume station so platform crowding won't be an issue (presumably...), but it sure looks uncomfortable to be there when a train is speeding through the station, and at the very least it looks like no one can move along the platform while a train is passing through.
      - The overpass is higher than it needs to be. It probably should be an underpass, but even ignoring that, is there any reason to have it more than a few feet above the wires?

    8. @ Joey

      It does look like the overpass is in fact only a couple feet above the wires. I think parallax is making it hard to see how high it is.

      @ adirondacker12800

      Another problem is that the platform is only a few feet wide. Looks kind of terrifying. How much would it have cost to make the platform wider (and safer)?

    9. The rest of the platform also looks dangerously narrow, i.e. not around that wooden tower thing. Also, why do people who design these things not put in cover for people? Sun and rain folks, people don't like them while waiting for the damn train (Caltrain is notorious in this regard). That wooden thing is a fucking historical building? I'm a railfan, but Jesus, that's a stretch. I hope they get sued when someone is killed on that narrow platform.

  11. This comment has been removed by the author.

  12. By far the most important design aspects for the future viability of a rebuilt Caltrain/CHSR Peninsula Corridor are station designs with the following properties in addition to level boarding:

    1) Their location should be as close as possible to most passengers’ destinations.
    2) Cross-platform-timed-transfers in both directions should be provided at all rail transit intersections.
    3) They should be in shallow open-cut positions in order to encourage closely integrated with those stations private sector financed transit-oriented-development.

    In order to accomplish all this at an affordable price the present electrified Caltrain/CHSR 23’6” overhead clearance requirement should be reduced by at least 10 feet.

  13. Why would a person travel between San Francisco and Los Angeles in 2:40 on a CHSR train when he can make the same trip gate-to-gate in 1:15 by airplane with convenient moderate cost rental cars at both air terminals?
    By far the most important competitive property of any rail transit destination stations is whether they are within a short walking distance to locations of interest to a large number of people. Examples for Caltrain/CHSR stations: A Downtown San Jose Station located a short walk from San Jose State University, a BART branch dividing from the line under construction at Milpitas calling on many stops within the Golden Triangle on the way to a Caltrain/CHSR cross-platform-transfer-station directly below the San Jose Airport Passenger Terminal, an elevated station in San Francisco at Third and King with ramps directly to the adjacent AT&T Baseball Park with tracks connected to the 22nd Street Station by a nearly straight rail line with a midway station adjacent the San Francisco State University Campus in the Mission Bay Development, and the Montgomery and Embarcadero stops within the established Market Street Subway. Timed-transfer-cross-platform connections to all intersecting rail services could be developed.
    In order to minimize new subway tunnel construction distances and to fit yet to be designed and constructed Caltrain/CHSR trains onto already established BART Downtown San Francisco and Bay Crossing Tunnels dual gauge track-way and third-rail electrification capable Stephenson gauge (1435 mm) Caltrain/CHSR rolling stock should be built. {Note: The 34.5 mile Seikan Tunnel Hokkaidō Shinkansen project (which started construction in 2005) will include laying dual-gauge track linking the tunnel into the Shinkansen network, so Shinkansen trains can traverse the tunnel to Hakodate (scheduled for 2015) and eventually Sapporo.}
    Any affordable track-way design along this proposed Caltrain/CHSR route would have a minimum vertical clearance requirement within tunnels and overcrossing bridges considerably below the California High Speed Rail Authority’s present minimum vertical clearance standard for overcrossing structures of 23’ 6”. Any rail CHSR car must have sufficient interior height to accommodate a level walkway from car to car over each boggy that accommodates large traction motors, a compliant suspension, and enough sound insulation for comfortable CHSR trains. The TGV Paris SUD EST car structure meets these internal space requirements and yet retains an external vertical car height of 3.420 m or 11’ 3”.
    Would Public Utilities Commission of the State of California allow a significant height clearance reduction? Consider the CPUC Resolution No. S-1040 amendment to General Order No. 26-D signed November 12, 1963.
    A. Clearance in Tunnels, Subways, and Tubes. … 3. The minimum vertical clearance above the top of the lowest rail shall be not less than 12’ 6”.
    D. Herein shall apply to the operation of passenger car equipment of 10’6” in width, 11’3” in height above top of rail…

    A ten foot reduction in the grade-separation overhead crossing clearance requirement and allowing for a 6.5 foot constant additional clearance in order to allow for track-way structure and overhead bridge thickness that will remain constant no matter what rail-top to overcrossing-bridge-bottom standard is finally adopted if open-cut-grade-separation construction costs are proportional to the square of the open-cut track-way depth a 10 foot reduction in that depth will have the following impact on grade separation costs:
    {1 − [(13.5 + 6.5)/(23.5 + 6.5)]^2}*(100%) = a 56% reduction in initial construction costs