The RPA’s Fourth Regional Plan

The RPA has just put up its Fourth Regional Plan, recommending many new subway and commuter rail lines in New York, ranging from good (125th Street subway, Brooklyn-Lower Manhattan regional rail) to terrible (Astoria Line extension to the west rather than to LaGuardia, which gets a people mover heading away from Manhattan). I have a poll for Patreon supporters for which aspects I should blog about; I expect to also pitch some other aspects – almost certainly not what I said in my poll – to media outlets. If you support me now you can participate in the poll (and if you give $5 or more you can see some good writings that ended up not getting published). If you want to be sneaky you can wait a day and then you’ll only be charged in January. But you shouldn’t be sneaky and you should pledge today and get charged tomorrow, in December.

It’s hard to really analyze the plan in one piece. It’s a long plan with many components, and the problems with it don’t really tell a coherent story. One coherent story is that the RPA seems to love incorporating existing political priorities into its plan, even if those priorities are bad: thus, it has the AirTrain LaGuardia, favored by Cuomo, and the Brooklyn-Queens Connector (BQX), favored by de Blasio, and even has tie-ins to these plans that don’t make sense otherwise. Some of the regional rail money wasters, such as Penn Station South and the new East River tunnels from Penn Station to the LIRR, come from this story (the LIRR is opposed to any Metro-North trains going to Penn Station under the belief that all slots from points east to Penn Station belong to Long Island by right). However, there remain so many big question marks in the plan that are not about this particular story that it’s hard to make one criticism. I could probably write 20,000 words about my reaction to the plan, which is about 15 published articles, and there are, charitably, 5 editors who will buy it, and I’m unlikely to write 10 posts.

I’ll wait to see how the poll on Patreon goes, and what editors may be interested in. There are interesting things to say about the plan – not all negative – in areas including rail extensions, transit-oriented development, and livable streets. But for now, I just want to zoom in on the crayon aspects. I previously put up my 5-line map (4 MB version, 44 MB version). The RPA proposal includes more tunnels, for future-proofing, and is perhaps comparable to a 7-line map I’ve been working on (4 MB version, 44 MB version):

I was mildly embarrassed by how much crayon I was proposing, which is why what I put in my NYU presentation 3 weeks ago was the 5-line system, where Line 1 (red) is the Northeast Corridor and the Port Washington Branch, Line 2 (green) is much the same but through the new Hudson tunnels, Line 3 (orange) is the Empire Connection and the Hempstead Branch, Line 4 (blue) connects the Harlem Line and Staten Island, Line 5 (dark yellow) connects the Erie Lines with the Atlantic Branch and Babylon Branch, and Line 6 (purple) is just East Side Access. In the 7-line system, Line 6 gets extended to Hoboken and takes over the Morris and Essex Lines, and Line 7 (turquoise) connects the Montauk Line with the Northern Branch and West Shore Line via 43rd Street, to prune some of the Line 5 branches.

With all this extra tunneling, the map has 46 new double-track-km of tunnel. With just Lines 1-5, it has 30; these figures include Gateway and the other tunnels highlighted in yellow (but not the highlighted at-grade lines, like Lower Montauk), but exclude East Side Access. In contrast, here’s what the RPA is proposing:

Counting the Triboro-Staten Island tunnel and Gateway starting from the portal (not at Secaucus as the map portrays), this is 58 route-km, and about 62 double-track-km of tunnel (the Third Avenue trunk line needs four tracks between 57th and Houston at a minimum), for substantially the same capacity. The difference is that the RPA thinks Metro-North needs two more tracks’ worth of capacity between Grand Central and 125th, plus another two-track tunnel in the Bronx; from Grand Central to Woodlawn, the Fourth Regional Plan has 19 km, slightly more than 100% of the difference between its tunnel length and mine. My plan has more underwater tunnel, courtesy of the tunnel to Staten Island, but conversely less complex junctions in Manhattan, and much more austere stations (i.e. no Penn Station South).

As I said, I don’t want to go into too much detail about what the RPA is doing, because that’s going to be a series of blog posts, most likely a series of Streetsblog posts, and possibly some pieces elsewhere. But I do want to draw a contrast between what the RPA wants for regional rail and what I want, because there are a lot of similarities (e.g. look at the infill on the Port Washington Branch in both plans), but some subtle differences.

What I look for when I think of regional rail map is an express subway. I’ve been involved in a volunteer effort to produce a regional rail plan for Boston, with TransitMatters, in which we start by saying that our plan could be a second subway for Boston. In New York, what’s needed is the same, just scaled up for the city’s greater size and complexity. This means that it’s critical to ensure that the decision of which lines go where is, for lack of a better word, coherent. There should be a north-south line, such as the Third Avenue trunk in the Fourth Regional Plan or my Line 4; there should be an east-west line, such as the lines inherited from the legacy Northeast Corridor and LIRR; and so on.

The one big incoherence in my plan is the lack of a transfer station between Line 4/6 and Line 1/3 at Madison and 33rd. This is on purpose. Line 2 connects Penn Station and Grand Central, Madison/33rd is well to the south of Midtown’s peak job density, and Lines 4 and 6 shouldn’t be making more stops than the 4 and 5 subway lines, which go nonstop between Grand Central and Union Square.

The other weirdness is that in the 7-line system, unlike the 5-line system, there is no way to get between the Northern Branch or the West Shore Line and the rest of New Jersey without going through Manhattan. In the first map of this system that I made on my computer, Line 7 has an awkward dip to serve the same Bergenline Avenue station as Line 2. But I think what I posted here, with two separate stations, is correct: Lines 6 and 7 are lower priorities than a subway under Bergenline Avenue, which would make intra-state connections much easier. It’s difficult to depict rail extensions at different scales on one geographically accurate map, and doing a schematic map like the London Underground isn’t useful for depicting new lines, which should make it clear to readers where they go. But the 7-line system must be accompanied by subway extensions, some covered by the RPA (Utica, Nostrand) and some not (Bergenline, again).

I recently had to give a short description of my program for good transit, and explained it as, all aspects of planning should be integrated: operations and capital planning, buses and light rail and subways and regional rail, infrastructure and rolling stock and scheduling, transit provision and development. When I make proposals for regional rail, they may look out there, but the assumption is always that there’s a single list of priorities; the reason I depict a 7-line map, or even a 9-line map (in progress!), is to be able to plan lines 1-3 optimally. Everything should work together, and if agencies refuse to do so, the best investment is to make sure those agencies make peace and cooperate. The RPA plan sometimes does that (it does propose some regional rail integration), but sometimes it’s a smörgåsbord of different politically-supported proposals, not all of which work together well.


  1. newtonmarunner

    Thanks so much for this, Alon. Lots of stuff to criticize.

    I’m more focused on the subway element. It’s good that they’ve included Utica, Nostrand, 125th, Triboro, and Northern. The first three, you’ve remarked, are gems. I think the big problem with SAS/Northern is that it continues the reverse branching. This hurts frequency, speed, and reliability.

    I think UES should have 8 tracks — 4 on Lexington, 2 on 3rd, and 2 on 2nd. Lex should remain as is, 3rd should hook into the Broadway Express (and eventually continue to the Bronx on 3rd, pruning Dyre Ave), and 2nd should go to 125th (possibly 125th) and Utica vía Williamsburg/Bed-Stuy. Grand Concourse would get CPW local and Wash. Heights would get CPW express. 2 and 4 in the Bronx would be interchanged.

    For Queens, I’d have a new 50th St. subway, and eliminate interlining. Here’s where I’d go:

    Northern to 63rd to 6th Ave. Local
    Flushing to 59th to Broadway Local
    QB Express to 53rd to 8th Ave Local
    QB Local to 50th (to Jerz or Hudson Yards)
    Astoria to 42nd (to Jerz)

    G and new Astoria would transfer to at Court Sq. to 50th St tunnel rather than 53rd; Northern would transfer to QB Local/50th at Northern. This would relieve the 53rd St. tunnel, and allow SAS an easier to board connection at the 55th/2nd St. stop.

    Add a 42nd and 2nd stop for SAS connections to 42nd.

    In Brooklyn, have 4th Ave Line go to 6 Ave Express, and Brighton go to Broadway Express. Nostrand should get Broadway/7th Express, and Liviona should get Lexington Ave Express.

    Well, that’s about all I have to say about that.

    • Alon Levy

      That Jersey loop… they’re spending RER A money (two dedicated tunnels across the Hudson!) to build something that looks even uglier than the RER C.

    • Alon Levy

      Brooklyn, certainly. That creates a four-track line with stations at Fulton Street and Borough Hall, and maybe also Flatbush Avenue if the tunnel to Staten Island runs near 4th Avenue (which it shouldn’t, it’s too circuitous). Advantages include less underwater tunneling, service to Red Hook, and a one-seat ride from points north to Downtown Brooklyn; disadvantages include a less direct ride from Staten Island to Manhattan and possibly higher construction costs coming from having to double-O-tunnel underwater.

      Bayonne… the problem there is that it’s too circuitous, except from the outer part of the North Shore. The one advantage of the option is that it encourages an east-west alignment in Lower Manhattan, which may be easier to find a station box for than a north-south alignment.

    • Ralfff

      Bayonne: as a political reality there was a golden opportunity to extend HBLR to Staten Island over the Bayonne Bridge when the bridge raising was being planned and as usual it was completely wasted for political reasons. It occurred to me, though, that this would be a more sellable proposal to the state of New Jersey if the line continued through the middle of Staten Island and some sort of tunnel or bridge to Perth Amboy, allowing that line to begin and end in New Jersey and connecting two disconnected parts of that state. Of course, the existing Tottenville station is right there so it’s hard to justify a new right of way for light rail through the middle of Staten Island where political resistance will likely be extreme anyway.

        • adirondacker12800

          No and the west side of Staten Island, that isn’t already developed, is mostly park that floods during high tide. There is never going to be much of anything out there. Sending the North Jersey Coast trains express through Staten Island and Brooklyn, on the SIR ROW, might be interesting. Pair it up with the Harlem line or the Port Washington branch. Or local service to the inner suburbs. Or ….

  2. newtonmarunner

    I keep looking at the RPA map, scratching my head. Why Columbus Circle/59th St. and not 42nd St.? Why get rid of above ground Harlem Line for 3rd Ave. tunnel in the Bronx rather than a 3rd Ave. subway (in addition to the Harlem Line and using the Q), which combined with the Harlem Line, provides better coverage for places like Morrisania and places north of Crotana Park than just putting the current Harlem Line underground? Lots of work still necessary on this.

    • Lawrence Velázquez

      I assume they want to use the Weehawken Tunnel ROW under Bergen Hill. From there, a direct Hudson crossing lands at about 55th Street.

      As far as I can tell from a quick skim, the Bronx tunnel is a supplement to the Harlem Line, not a replacement. The existing line is already four-tracked, and without fare integration I don’t think that corridor has enough demand to justify concrete.

      (As a rule, I’m skeptical of central Bronx rail proposals. The Plan’s claim that “the Third Avenue corridor…has very poor transit access” is misleading. Most of 3rd Avenue is 2–3 blocks from the Harlem Line; the poor access comes from thin scheduling and lack of fare integration.)

      • newtonmarunner

        I’m generally more supportive of Bronx subway proposals. The Bx15 (3rd) and Bx41 (Webster) have over 50K riders combined, and Bx19 has nearly 30K riders — much of the service being unidirectional. As I wrote before, some parts of Morrisania and East Tremont are a pretty far walk to the Harlem Line or the 2/5. Even without skip-stop and with fare integration, these people will lack rapid transit coverage for such a densely populated area (66K/sq. mi.)

        I do very much like a 3rd Ave. Subway as even if it doesn’t leave UES, it allows SAS it’s own C division and UES connections to Midtown West without reverse branching. Obviously, the further you go north with 3rd Ave., the more things you accomplish. The 6 could change for the Midtown West job cluster at 138th if the mythical 3rd Ave/7th Ave/Broadway Subway is brought there. It could provide another path to the Bronx Hub if brought up another stop. Another stop would fill the transit desert in Morrisania. The Line could go up just north of Crotona Park, to the hospital (filling a transit desert in East Tremont), and then prune the Dyre Ave Line. This would increase capacity, and would obviate the tramways Alon wants on Webster, 3rd, and Southern.

        • Eric

          A 3rd Avenue Subway in Manhattan would basically serve the same purpose as four-tracking the 2nd Avenue Subway. As such the “local” tracks should continue west on 125th, and the “express” tracks should continue to the Bronx (I would have them take over the Dyre Avenue branch).

          • newtonmarunner

            Exactly. Two tracks are two tracks are two tracks — be it Regional Rail on 3rd as the RPA proposes, Subway on 3rd as I propose, or 4-tracking on Second as you propose.

            I’m not as much a fan of express trains on 2nd as I am local trains. Inwood, Washington Heights, etc. should have a 2-seat ride to UES Local stops, so the Bronx should definitely get the express if there is one. But UES locals should have a 1-seat ride to the Midtown West job cluster, the West Side lines at Times Sq., and the 6th Ave. lines at Herald, Sq., so I prefer all local on UES with similar stop spacing as the current SAS (wider than Lexington Ave. Local but shorter than Lexington Ave. Express). Hence, I come to bringing the Q to 3rd Ave. with similar stop spacing as SAS.

          • newtonmarunner

            Also, the stop spacing on SAS is wide enough such that someone on UES living on east of 2nd Ave. can see the SAS line as a reasonable substitute to the Lexington Ave. Express Lines (which are a longer walk for someone living east of 2nd Ave.) to FiDi. This relieves the Lexington Ave. lines.

          • newtonmarunner

            One other thing — Extending the Q to UES to Mott Haven to Morrisania to East Tremont (all along 3rd Ave.) to Dyre Ave. should be done at the same time (in stages) the 125th/SAS is extended to LES’s 1st & Houston and Houston/Ave C to Williamsburg (Metropolitan/Grand/Bushwick) to Malcolm X to Utica.

            Northern should go into 63rd w/ QB Local to 6th Ave. QB Express should go into 53rd. If more room is necessary, or the 53rd St. tunnel needs relief, then build the 50th St. Subway (stops at Court Sq, 3rd, 5th, and 7th) with Northern or QB Local hooked into it.

      • newtonmarunner

        I should add that bringing an additional subway line to the Bronx shouldn’t happen until (1) the Harlem Line gets subway-like service, and (2) the reverse branches are untangled.

        The other reason I detest the 3rd Ave Regional Rail Line — and prefer eventually bringing the Q to Bronx on its own two tracks through UES —is the destinations the regional rail line hits. Grand Central, Union Sq., and W. 4th on Alon’s map are an order of magnitude more valuable than 3rd and 42nd, 3rd and 14th, Houston, and Canal. Park Ave. is closer than 3rd to Madison, 5th, and 6th, and has connections to the Lexington Ave. Lines. Union Sq. and W. 4th are secondary job clusters: if the reverse branches are untangled (namely Brighton and 4th Ave. in Brooklyn but also Washington Heights and Grand Concourse), that’s an additional path to Union Sq. and W. 4th, effectively relieving Barclays, Canal, and Lafayette as transfers. Also, it relieves the A from the Bronx. As for Canal St., Lexington Ave. Express has two stops between 42nd and Fulton — Union Sq. and City Hall; with stops at 14th and Canal, the RPA 3rd Ave. RR line is the same. But RR should have wider stop spacing than an express subway line, so this makes no sense to me.

        Just my latest $0.02, though probably not worth that much.

  3. Abraham Nemitz

    Anytime I see a proposal to merge NJTransit Commuter rail with Metro North and LIRR (which I fully support) I have to wonder what about SEPTA?

    The benefits of through running are just as important at the Philadelphia end of NJTransit. I think any realignment of regional rail operation has to include Pennsylvania and Delaware in the analysis.

    • ckrueger99

      I’d like to see NJT go south to Cornwells Heights while SEPTA goes north to Princeton (Junction) on the NEC and to Newark via the West Trenton/Hopewell branch (which would have to be electrified). I see benefits to both SEPTA and Amtrak operational simplicity in building a connection from NEC to SEPTA mainline at North Broad, allowing a more direct connection from Center City to Trenton (and beyond) and simpler operations for Amtrak around Zoo (and giving Amtrak the option of sending Keystone trains through Center City). This map isn’t very clear, but it’s the best I’ve got for the moment:

    • Alon Levy

      The operational benefits of through-running are considerable, yes – the alternatives are to keep reversing direction like today, crossing opposing intercity traffic at-grade, or to build terminal facilities and flying junctions in Trenton. That said, there is one big drawback to NJ Transit-SEPTA through-running. There’s much more demand on NY-Trenton than on Trenton-Philly. This means that if the trains are equally long, there will be much more frequency north of Trenton than south, which means many trains are still going to terminate. Today the trains aren’t even equally long – SEPTA runs short trains (I think 4-5 cars), NJ Transit runs very long trains (8-12 cars).

      • ckrueger99

        Very true, Alon, which is why the Hopewell branch has to come into play. By making Hopewell viable (in 2 directions, even!), the NJ/PA balance becomes more equal. Plus, Central Jersey benefits 2 ways: 1. residents commute to Philly/Jenkintown and 2. PA residents commute to jobs at Princeton, New Brunswick and Metropark. So there’s plenty of incentive for NJ’s new governor to talk to his PA counterpart.

        • Alon Levy

          How is the West Trenton Line going to help people commute to Princeton, New Brunswick, and Metropark, all of which are on the NEC?

          • adirondacker12800

            Right now they get on shuttle buses from the NEC to places along the West Trenton line that someday, maybe perhaps, will have stations. The one that I know about, the big office/hospital complex growing up near the Trenton Airport, is designed with that in mind. There are a lot of people along the West Trenton line, who want to go to New York or Philadelphia and points beyond. Who drive to a NEC station. Or just drive. Enough of them that they could walk or take a much shorter drive to a West Trenton station.

          • ckrueger99

            It doesn’t, but developing this line jointly between NJT & SEPTA makes the benefits to the 2 states more balanced.

          • F-Line to Dudley

            NJT West Trenton commuter rail was also supposed to be augmented by a RiverLINE light rail extension from downtown Trenton to West Trenton, using stretches of the active Conrail Bordentown Secondary (Trenton end), active Conrail Trenton Industrial Track (W. Trenton end), and an abandoned ROW along the canal through downtown spanning the two to hook it together. The commuter rail line comes first in most build scenarios because it’s easier to implement; the existing RiverLINE needs a substantial shot of double-tracking $$$ on its current route to increase service levels before it can finally behave like a light rail circulator truly worth it’s salt. But the overall transit vision for the state capitol region envisions a big multimodal hook-in at West Trenton.

            Wholly separate audiences, wholly separate upside from the NEC re: who/what gets tied in. The most relevance the West Trenton Line has to current NEC services is maybe blunting a little of the extreme overcrowding on Trenton-terminating NJT and SEPTA trains so those services have a little more found flexibility to address other things en route instead of having to always be all things to everyone right to the last stop. But for the most part it’s scratching its own itch.

      • al

        Amtrak could run HSR and NEC Regional, while the commuter rail organizations run a relay type system. The personnel can switch out at territory boundaries, while the equipment continues down the line. Amtrak would run express and super express (HSR), while the commuters run local and short turn service.

      • adirondacker12800

        NJTransit has terminal facilities west/south of Trenton in Morrisville Penna. When they did that they disentangled things. From the looks of the satellite images, with expansion in mind. When NJTransit finds the money to build the loop west/south of the new station in North Brunswick, it is going to take a very long time before SEPTA and Amtrak clog things up in Trenton. If ever.

        • ckrueger99

          I would think it would make sense for SEPTA to arrange w NJT to share Morrisville, as this could service both Trenton and West Trenton lines.

          • F-Line to Dudley

            Morrisville’s a Norfolk Southern-owned freight branch feeding a very busy truck container and bulk-transfer yard. And that branch also spanned an ex-Reading passenger line (West Trenton) and an ex-Pennsy line (NEC), so pre-1984 before the unification of Regional Rail the routings still reflected the historical competitive separation even 10 years into the Conrail era. Those are the only two reasons why that 6 miles of connecting branch hasn’t been used for alt-routing passenger service to Trenton all along, and hasn’t taken up more mindshare as a viable proposal in the ensuing decades.

            Wouldn’t be very hard at all for SEPTA to purchase the West Trenton-NEC connector from NS; as Class I’s go NS isn’t an unreasonable ogre to do business with. They’d just have to make sure electrification clearances under the 4 highway overpasses on the connector are adequate for double-stacked container freights (West Trenton Line between Neshaminy Falls and West Trenton already is), and to buff out the track layout on the branch to avoid traffic conflicts with freights that are crossing from the north side of the ROW to south side in order to reach the yard (plenty of room for that).

  4. newtonmarunner

    The more I look at the plan, the more work I think it needs. I mentioned that the Third Ave. regional rail line doesn’t coordinate well with the subways. Neither do the Northern and Utica. The stop spacing seems more about coverage and making the subways faster than integration with buses. While Northern hits Northern Blvd. (transfer to the QB Local), the only bus route it hits is Junction and the Q72. 80th St. doesn’t have any buses, nor does 110th. Also, key is that Northern misses its Triboro stop connection.

    The Utica Ave. subway also needs work. Intersecting Utica is Church with its B35 Bus, Glenwood with Triboro, Avenue H with its B6 bus, Flatlands Ave. with the B82 bus, and Flatbush with the B41 bus. All of the bus routes have 27K+ riders, with the B6 having over 40K riders/weekday and the B35 nearing 40K riders/weekday. Obviously, Glenwood and Avenue H are too close for both to get stops, and the RPA is correct to hit Triboro rather than the B6 as most people won’t transfer to a bus the way they do a subway. Yet Utica misses Church and its highly popular bus route, and instead opts for Winthrop and Clarendon, which don’t carry any bus routes. This is a pretty glaring flaw in the Utica Ave. Subway.

    At the end of the day, something has to give. Transportation policy is all about tradeoffs. I just think the RPA in too many cases made the wrong tradeoffs.

    • al

      RPA also left out a gimme. A New Haven – Hells Gate – LIC – Jamaica/JFK line. It would create a 1 seat ride between Queens and points northeast. A flying junction between the Freemont Secondary and Main Line at Woodside, or at Maspeth with the Lower Montauk would be a long term upgrade.

      • newtonmarunner

        I think the biggest gimme the RPA left out is what everyone else is talking about — not connecting Penn and Grand Central and hooking the new Hudson Tunnel Lines into ESA. Really nothing compares to that miss.

        As I’ve written above, I think on the subways, given the reliability problems NYCT has, should eliminate reverse branching and inner-lining. You can use some of the subway extensions, e.g., SAS to Yankee Stadium (should go to FiDi — not — take the Q, though), as political sweeteners. [Granted, so many of the extensions don’t belong, e.g., BQX and Willets Pt AirTrain, or are backwards, e.g., Astoria and Hudson Yards extensions, hooking Northern into SAS.]

        The basic gist of the RPA plan is that they’re more concerned about access, coverage, economic development, and avoiding political headaches than they are about capacity, operating expenses, and good transit. The result is far more money than what Alon has doing far less things than Alon does.

        • adirondacker12800

          The 63rd Street tunnels aren’t high enough for pantographs to fit in. Third rail would have to laid all over New Jersey. No thanks.

          • Alon Levy

            Threestationsquare resolves the 63rd Street tunnel size issue by connecting ESA to Staten Island. (I do by just saying they should third-rail parts of the Morris and Essex Lines, or get trains with lower roofs.)

          • adirondacker12800

            Third rail made sense in 1899. It doesn’t anymore.

          • orulz

            Third rail still makes sense today when you have an unelectrified line, a large installed base of third rail to connect to, with existing height-restricted tunnels, a significant amount of third rail-only rolling stock, and no need for high speeds.

            Alon’s approach seems (to me) to be the right one for the Erie lines, which are not currently electrified and must connect to the height-restricted LIRR Atlantic Branch. The ceilings are way to low to accommodate existing M8s with retracted pantographs. Even with a new railcar design, it would still be difficult – the floors have to be 4 feet 3 inches, and the vertical clearance is only 13 feet, there may just not be enough space to fit pantographs while maintaining adequate headroom.

            It’s a little less clear cut on the Morris and Essex lines which are already substantially electrified. Dual-mode rolling stock is probably the appropriate solution here, as long as the 63rd Street tunnels were built tall enough to accommodate at least retracted pantographs (although, I’m not sure if they were).

          • Alon Levy

            Are there height restrictions on the Atlantic Branch? I know there are some in the 63rd Street Tunnel, which ESA hooks to, but I don’t think it’s infeasible on the Atlantic Branch?

          • F-Line to Dudley

            Third rail’s not going to scale well at all for new commuter rail installations because of the need for 750V DC substations every 3-6 miles. With 125+ route miles’ worth of current ex-Erie diesel lines (75+ mi. cumulative on the Main/Begen/Pascack, ~50 mi. Suffern-Port Jervis) plus all the expansion lines in that area, the cost difference in doing any of that up even partially as 750V DC instead of 25 kV AC is far too extreme to work. The NEC, by contrast, does all of New Haven-Boston @ 25 kV AC with only 4 subs. And if NYHSR did up the entire Hudson Line + Empire Connection to Albany as 25 kV AC, junked the third rail entirely north of Spuyten Duyvil, and had Metro-North buy M8 replicas to power-switch from third rail to overhead at Spuyten Duyvil they could do up the entire 140+ mile NYP-ALB corridor on 3 subs…4 tops. And end up retiring twice as many existing DC subs between SD and Croton-Harmon in the process. Just to extend existing third rail 40 miles from Croton-Harmon to Poughkeepsie would be something like 9 or 10 more DC subs on top of retaining all existing ones, so that’s illustrative of the challenge of planting the DC electrification flag on a second system.

            Long Island is a slightly different story because DC trunk feeders already crisscross the whole western half of the island with extreme density as N-S-E-W interconnections between lines, such that the costs of extending LIRR electrification are offset by all those interconnections making placement of new subs (and land acquisition therein) much more flexible. It’s no undue burden to finish up the electrification job to Port Jeff, Oyster Bay, Babylon-Patchogue, and on the Central or Lower Montauk branches when they’re all in easy reach of that cross-island DC feeder thicket. Only when you get way out on the east end on those hundred-plus linear miles of past- Zone 12 trackage to Montauk and Greenport does new construction start scaling too poorly to bother with and become a cost analogue to the horrendous scalability of DC into Jersey, etc.

            You can use the M8 design to switch to third rail in the tunnels, but you have to have that pantograph on the roof for running out in the field so clearances will need to square the height difference between ESA 63rd St. (i.e. not one inch taller than an M1/3/7/9 car) and an M8 with its pantographs and roof-mount electronics. Put on your thinking caps, because troubleshooting the tunnels and/or vehicle designs for a few inches of height are going to end up the far better use of resources than trying to fling new DC electrification several counties out. Tunnel mods might be a hard sell the way the 63rd St. bore was originally constructed, but at least the engineering difficulty level is self-contained to a single-point fix in a very short-length project area instead of negatively warping scalability all across that second system.

          • adirondacker12800

            The DC substations already have an AC transformer at them.There wouldn’t be a need for much land acquisition, if at all. Changeovers happen outside of tunnels, away from places with lots of traffic. Or where the traffic can get around it when it doesn’t go well.

          • orulz

            I would agree that if some slight modification to the tunnels can magically improve clearances, or if there is some off-the-shelf railcar somewhere in the world that happens to meet these rather stringent clearance requirements, then that’s obviously the way to go. However, if neither would work, building the Erie out with 750vdc third rail might be more expensive than using 25kv AC, but I doubt writing a blank check to substantially rebuild or replace the Atlantic Avenue and 63rd street tunnels could possibly be cheaper.

            As a result of Sandy, LIRR replaced three substations on the Long Beach Branch, for a total cost of 56.5 million. From that description it sounds like those substations were a little on the ‘special’ side to harden them against sea level rise and future storm surge events. But even assuming that’s a typical cost, it’s about 18 million each. So if you figure on two dozen substations to cover the Erie lines, that’s $432 million. A lot of money? Yes. But way, way, way less than anything involving tunneling.

          • adirondacker12800

            It’s 2017, No one is going to allow third rail without grade separation and fencing. That gets pricey.

          • F-Line to Dudley

            The yards at Atlantic Branch used to carry freight trains until that business dried up in the 1960’s. See here photo from 1954 of the Atlantic Terminal yard stuffed full of boxcars, with an MU parked on the left: There used to be several blocks of street-running trackage for threading those loads from the yard to the Brooklyn piers.

            M8’s to Atlantic Terminal shouldn’t be any issue, as you can plainly see in the photo that those freight cars are quite a bit taller than the MU. It’s only the 63rd St. tunnel that enforces a restriction because of the way it was constructed as a double-deck subway tunnel with equal-size bores 40 years before the rest of ESA. The design for the M1’s was predicated on the dimensions of that tunnel: 13 ft. even, which has held for the M3/7/9 car designs. Pantograph-bearing M2/4/6’s are 14’9″ to the top of the roof resistors (figures unavailable for M8’s, but pretty sure it’s identical or no worse than +1-2 inches). So squaring 1’9″ of height difference over a 3140 ft. tunnel segment is the engineering challenge. Forget about sprawling any cost-bloating hacks further out, because they won’t scale well enough for planting the flag on a second system.

            I don’t know what kind of trackbed ESA is going to have through the legacy 63rd St. tunnel segment. If it’s designed for regular ties and rock ballast you might well be able to shave some consequential inches by going to floating-slab track anchors placed straight over the concrete tunnel floor. The North River tunnels, post-Gateway when they’re each taken offline for full top-down rehabilitation, will very likely get changed over to floating slab as a means of installing active floor drainage channels and pump systems that are easily maintenance-accessible under the trackbed slabs. That *may* as a happenstance consequence of the flood protection effort add enough height in those incredibly tight tunnels to accommodate a future change from 12.5 kV to 25 kV electrification clearances over the same max. vehicle heights (i.e. a 14’6″ MultiLevel coach or EMU)…though it’s uncertain without more detailed engineering assessment whether enough inches are going to be claimable. We’ll know definitively one way or the other when the overhauls of those tunnels go into design. That’s a project to watch for ascertaining feasibility in the 63rd St. tunnel. 1’9″ may be a stretch, but if playing similar trackbed games gets you most of the way there the rest of the difference could possibly be made up by roof-notching the inner tunnel lining around the clearance envelope of the M8 roof resistors. It’ll be close, so this should be an active area of inquiry.

          • orulz

            When you say ‘squaring’, do you mean in the literal sense of grinding away some of an arched tunnel’s roof to make it more square, or the figurative sense of just finding a resolution to a problem? Probably the latter, since the Atlantic Branch tunnels have flat roofs.

            That’s an interesting point about freight service on the Atlantic branch, though. I always assumed that Atlantic was the dog and 63rd and the EMUs were the tail, since the tunnels on the Atlantic branch are over 100 years old. But if those are standard boxcars they would be somewhere between 14 and 15 feet tall, no? How can you tell they’re not some type of special low-height boxcar

            Regarding 63rd, at any rate, 1’9″ seems like a lot to make up, given that tunnels are generally built with tight margins to begin with. In a conventional tunnel, it is often possible to dig out the floor. Perhaps this approach could work for the Atlantic Branch if needed – and costs might even be reasonable. But it’s almost certainly not an option for submerged tubes like 63rd. I suspect you’d be lucky to get even 3 inches of extra clearance by tunnel modifications. So that means either a budget-blowing tunnel replacement (which would in all likelyhood cost MORE than 3rd rail electrification), or otherwise significantly reducing the height of the rolling stock, which might well be possible with a touch of creativity – and would certainly be the better approach -as long as it’s possible.

          • F-Line to Dudley

            That pic @ Atlantic is of AAR “Plate B” (15’1″) freight cars, the minimum standard size that is 100% universally interchangeable across the North American freight network. That minimum standardization was ratified in 1946 when the rail network was still semi-nationalized for the war effort in order to correct a hodgepodge of standards inconsistencies, so it took into account LIRR’s clearances at a time when the Brooklyn waterfront was still moving lots of war supplies. You can physically take an LIRR C3 bi-level coach or NJT MultiLevel in there because those meet Plate B standards…just not anything bigger like an MBTA/MARC Kawasaki bi-level or an Amtrak Superliner.

            The primary restriction with Atlantic is a severe weight limit on the Nostrand Ave. viaduct. The only locomotives allowed in there are small yard switchers like LIRR’s work fleet of 1000 HP SW1001 ex-freight engines, which only weigh 115 tons vs. 128 tons for a 3000 HP DM30AC passenger dual-mode. So Atlantic ends up moot for anything with enough power to haul a regular commuter train’s worth of cars, and that’s why there’s never anything but MU’s routed there. It’s no use buying some super-lightweight loco for alt-routing to AT if it’s going to have too few horses to carry more than 3 or 4 unpowered coaches at-speed. There’s also some tight curve restrictions in the tunnel that ban longer cars that have wider-than-normal turning radius on their axles, which is why the Ringling Bros. circus train never attempted a direct visit to Barclays and had to stage the circus there all the way from the yard at Long Island City.

            Fine for an M8, however. 63rd. St. is definitely the only tail wagging that dog. Alternately, in a game of inches you can hope that regenerative braking tech advances enough by the M10/M12 car generations to further shrink the undercarriage space currently taken up by radiators and relocate more resistors off the roof to help lower the profile. Probably not going to get the car roof entirely 100% clean of all non-pantograph junk because those dual AC and DC inputs will always leave the underside pressed for space, but component shrinkage can factor into the 63rd St. equation along with trackbed and roof-notch games. Say, you move enough stuff off the roof to cram any electrical that remains inside the width envelope of the pantograph surface and free up space around the corners of the car roof (i.e. a much narrower roof ‘hump’). Does that make it any easier to notch the tunnel roof a few inches deep at the apex of the tube’s lining with a couple foot wide “channel” of sorts? You’re most likely in the ballpark if trackbed games like floating-slab anchors are in-play, so what combos get closest?

            Nobody said this would be easy, but you can get awfully close to paydirt by working every available angle to try to solve the single-point restriction on behalf of the system instead of warping the whole sprawling system for the sake of the single-point restriction.

          • adirondacker12800

            Add a million people to the city, the city needs four more tracks of subway. 6 or 8 would be better. The Jamaica Avenue El is way past it’s prime. Replace that with four tracks out to where it goes underground again to get to Jamaica. The express trains can substiute for LIRR service to Brooklyn, along the Fulton Street line while Atlantic Avenue is being dug for for 6 or 8 tracks. Four subway and two LIRR/Amtrak… If 63rd Street is unfixable, once the Island is converted to overhead, shuttlize service to Grand Central. 20 an hour means every 6 minutes at Floral Park or Valley Stream.

  5. adirondacker12800

    The legend on is that they didn’t make clearances in the 63rd Street Tunnel bigger because the railroad was forever hobbled by the low clearances on Atlantic Avenue. Someone would have to send one of those new fangled 3-D mappers through both of them to re-establish what the clearances are.

    • F-Line to Dudley

      No, 63rd St.’s dimensions have to do with the construction method, not service choice. It comprised the first prefab immersed tube tunnel done in New York, with 4 sections floated up from Maryland sunk into trenches in the riverbed and joined together by cut-and-cover sections at the shores. Some particular technical aspects of the construction had never been attempted before or since in any immersion jobs, due in large part to the very strong currents in the riverbed that ruled out other immersion techniques. The divider between upper and lower levels had to go right along the diameter of the tube to balance the weight distribution for transport and immersion of the prefab sections. That left the LIRR level vertically constrained and the subway level with a little bit of extra vertical slack, rather than being able to fine-tune the divider to give LIRR more room and the subway less.

      That was the only way the tunnel was going to be built within cost back in 1969, and it ended up winning all sorts of construction industry awards for its novel methods. The design compromise for getting it built at all wasn’t seen as a detriment to future service patterns because the next-gen EMU designs then under consideration for the eventual M1 order would’ve fit it just fine. But past precedent on clearances elsewhere on LIRR had little if anything to do with why they chose to build 63rd St. the way they did. It was all about construction methods and feasibility therein at that one very specific location in the river. I don’t know what the feasible construction alternatives could’ve been at the time vs. what they chose, but it sounds from reading up on the engineering history of that bore that any alternative would’ve cost so much more in ’69 dollars that an addled ’69 NYC could’ve never contemplated doing even this much.

  6. FDW

    Alon, the link to Larry Littlefield’s Room Eight hasn’t been working properly in a while. Wouldn’t be better just to switch it out for his WordPress page? I like Larry Littlefield, he and Johnny Sanphillippo (Of Granola Shotgun) need all the exposure they can get, given how effectively they’re arguing for a Revolution in America.

    More on topic. I’ve been designing an RER network for the Seattle area, I’m designing it alongside the “7-Line” draft of my Seattle Metro network. The Metro part is mostly done (save for blurbs explaining how and why the system is as it is), but I’m only really just starting on the RER. Still, would you be interesting in seeing what I’ve got so far? The only part I’m not sure of is the size of the fleet.

  7. Pingback: The RPA Construction Cost Report | Pedestrian Observations
  8. Pingback: When are Express Trains Warranted? Part 1: Regional Rail | Pedestrian Observations
  9. Pingback: Which Older Lines Should Express Rail Have Transfers to? | Pedestrian Observations
  10. Pingback: Assume Nordic Costs | Pedestrian Observations

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.