New York Regional Rail (not S-Bahn)

The discussion of regional rail in New York usually focuses on through-running, with neat S-Bahn-/RERstyle maps showing how lines run. But it’s also instructive to look at longer-range lines, under the rubric of RegionalBahn in Germany or Transilien in Paris. I’ve argued against segregating long- and short-range commuter trains in New York, on the grounds that its infrastructure layout is different from that of Berlin or Paris.

However, it is still necessary to conceptually plan longer-range regional rail in the New York region – that is, how to serve destinations that are too far to be really considered suburbs. I think that those lines should through-run, which makes the planning somewhat different from a standard intercity integrated timed transfer network, but the choice of where to go to, what frequency to push for, and so on is still important. This post should be seen as a pre-map version of what I drew for Upstate New York and New England, but for the Tri-State Region and satellites in Pennsylvania. It should also be seen as a companion to any high-speed rail proposal, albeit unmapped because I am still uncertain about some visible aspects.


The scope of this post is anywhere one should be able to get to from New York without resorting to high-speed rail. This covers the combined statistical area and its penumbra. In practice, this post will focus on areas that are off the Northeast Corridor than on areas that are on it. On the Northeast Corridor, I’ve talked about low-speed solutions toward New Haven putting it slightly more than an hour away from New York; instead of repeating myself, it’s better to discuss other destinations.

So what are the satellite regions around New York, excluding the city’s own suburbs? Let’s make a list:

  • Eastern Long Island far enough to be outside the commute zone, like the Hamptons
  • The Jersey Shore, likewise focusing on what’s too far for commuting, like Toms River
  • Trenton
  • Allentown and the Lehigh Valley
  • The Delaware Water Gap Region and possibly Scranton
  • The Mid-Hudson Valley on both sides of the river, i.e. Newburgh and Poughkeepsie
  • Historic city centers in Connecticut: Danbury, Waterbury, Bridgeport, New Haven

For the most part, they already have commuter rail service. But travel demand is usually not very commuter-oriented. Some of those lines have service that accommodates this fact, like express LIRR service to the Hamptons at popular weekend getaway times. Others don’t. Newburgh, Allentown, Toms River and Delaware Water Gap have no service at all, though Delaware Water Gap is on the under-construction Lackawanna Cutoff.

The need for electrification

All trains touching New York must be fully electric. This means spending not a lot of money on completing wiring the LIRR, Metro-North, and New Jersey Transit, and ensuring further extensions are electrified as well. Diesel trains are slow and unreliable: the LIRR’s mean distance between failures is around 20,000-30,000 km on the diesel and dual-mode locomotives and well into the 6 figures on the EMUs. New Jersey Transit’s diesels also tend to only serve Hoboken, which forces an additional transfer; NJT’s new dual-mode locomotives are extremely costly and low-performance.

This kind of completionism is especially valuable because of fleet uniformity. Boston is reticent about electrification because it likes having a fleet it can maintain all at one place, and it requires some additional resources to expand a railyard that can accommodate future electrification. In New York this works in reverse: a large majority of the network is electrified, and getting rid of the diesel tails increases efficiency through scale.

The issue of express service

All of the tails in question are far from New York, generally 100 or more km, and close to 200 km for Montauk. This introduces tension between the need to run intense local service to areas 15 km from Manhattan and the need to maintain adequate speed at longer range. The solution is always to prioritize shorter-range service and make regional rail the most express pattern that can fit within the through-running paradigm. This works well where there are four tracks allowing long-range express service, as on the Northeast Corridor and the Empire Corridor, including tie-ins like Danbury and Waterbury.

Elsewhere, this is compromised. EMUs can still beat present-day diesel trip times, but the average speeds of the 30-30-30 plan for Connecticut are not realistic. This is a tradeoff; it is possible to run express trains to the Hamptons on the Babylon Branch, but it imposes a real cost on frequency to dense suburbs and should therefore be avoided. If there’s room for timed overtakes then they are welcome, but if there’s not, then these regional trains should really run as S-Bahn trains that just keep going farther out.

This has precedent on busy lines. Trains in the exurbs of Tokyo tend to run at the same speed as an ordinary rapid train, for example on the Chuo Line; there is the occasional higher-speed liner, but usually the trains to Otsuki, Takasaki, Odawara, etc. are just ordinary rapids, averaging maybe 50 km/h. In New York the average speed would be higher because there are still fewer stops even with the infill I’m proposing, which fits since there is more sprawl in New York.

Onward connections

Some of the outer ends in question should also get service that doesn’t go to New York. There is an existing line between Danbury and Brewster that can be used for revenue moves. Allentown lies on a decent SEPTA Regional Rail extension, albeit not on a good one, as the route is curvy. If there are internal bus systems, for example in Waterbury, then whenever possible they should pulse with the train, and it goes without saying trains that do not serve New York should be timed with trains that do.

This for the most part should run on a half-hourly clockface schedule. This means that on an S-Bahn network where even individual branches run every 10-15 minutes, there should be a rule saying every train in 2 or 3, depending on base frequency, continues onward to a distant destination. This is a combination of Northern European planning (timed connections) and Japanese planning (treating long-range regional rail in a megacity as a commuter train that goes further than normal).


  1. Patrick

    The East End, Jersey Shore, and Mid-Hudson all have a nontrivial volume of commuters to NY. But more importantly, they have a high-volume of commutation to intra-Island/intermediate places that better fit the definition of “regional rail” in the textbooks. As usual, it’s not the extreme ends that introduce any uncertainty with how to best handle them, it’s all of the gray, overlapping areas in the middle that cause problems.

    Practically everything within a 50-60 mile radius is high trafficked enough to support full regional rail services through to New York (that would cover as far as Trenton, Toms River, Poughkeepsie, and New Haven), and should be part of any regional rail system.

    Some of the more rural fringes, like the East End, Waterbury Branch/Hartford Line/Shore Line East certainly see less volume and do not have much through traffic to New York, but still retain strong ties to intermediate points. Running these as shorter EMU/DMU shuttle trains every 30-60 minutes with timed connections to core regional rail services at Patchogue, Ronkonkoma, New Haven, Stamford, etc. Most of these segments (at least Montauk, Springfield, and SLE) should have at least some limited direct/express service through to New York during rush hours.

    North and west of the city, the demand seems to be a lot more focused on hub-to-hub trips instead of a lot more local/intermediate travel seen on LI and in CT. For destinations like Albany, Scranton, Allentown, Philadelphia/Harrisburg, running Clocker-style services that are more of a hybrid between inter-city and regional/commuter trains would probably be a better fit. Running these on an hourly interval (but with unreserved seating/flat fares) would likely be sufficient for these destinations. These should have timed connections to local regional rail services (e.g. at Poughkeepsie, Washington, etc.) for local travelers but otherwise run express to New York. I am not sure I see the value in through-running these trains to anywhere other than Sunnyside Yard.

    Whether electrification is extended to the far reaches of the network or these run with dual modes, I don’t think is that big of a difference—the number of stops they would make in the regional territory would be limited to 2-3 each—but universal electrification would of course be ideal. At the end of the day, we are only talking about a small handful of trains per hour (MTK, SPK, SPG, NLC/WLY, ALB, SCR, AWN, PHL/HAR).

    • adirondacker12800

      Torresdale, in the far northeastern corner of Philadelphia is as far away from Penn Station New York as New Haven, Poughkeepsie or Speonk. The railfan boyz drool at the thought of running NJTransit trains THROUGH!! !! !! Trenton to Philadelphia. 10 and 12 car NJTransit trains with multilevel cars can be standing room only and SEPTA runs four to six car single levels. You could run 10 car trains to Philadelphia but they would be mostly empty between Trenton and Philadelphia.

      Albany has once an hour-ish service now. Pick a train that does not come from the west or comes from Montreal. Also know as the Late for Sure Limited, the Maple Late or the Adi-runs-late. The ones from Buffalo are better but a bit iffy. They can sell out but that is a function of Amtrak’s equipment shortages.

      Nice well rounded numbers Springfield is as far from New York as with twice as many people if you lump it with Hartford. Up until recently the Hartford line was slower than a bus. New York has spent money on the Hudson Line over the years and the trip is faster today than it was in the heyday. Rumor has it could even be faster but Metro North would have to spend more money to maintain the track to a higher class. They don’t. Scranton is also just as far with half as many people as Albany. They pulled up the rails between Port Morris NJ and the Delaware River many years ago.

      There aren’t a whole lot of people on the East End.
      Not much traffic on it except for summer weekends.

      Toms River hasn’t had rail service in a very long time

  2. df1982

    What are your thoughts on through-running HSR to Long Island? Your plans for regional rail don’t seem to segregate HSR services from commuter trains on the NEC, so it wouldn’t necessary need its own tracks (just upgrades to one of the existing lines), but it always seems to get left off your proposals (and those of many other NEC plans) despite being able to serve a pretty large population that otherwise has slow transport connections to the rest of the north-east. A fast train to the Hamptons could probably attract a lot of patronage from things like the Hamptons Jitneys and charter planes to East Hampton airport.

    The other benefit is that it could make up for the demand imbalance between NY-Washington and NY-Boston, so you could have for instance 8tph running Washington-NY that then splits to 6tph for Boston and 2tph to Long Island (or whatever). Or you could have something like a Phildelphia-East Hampton Clocker-style service.

    But I admit I’m in two minds about it myself so I would be keen to hear your thinking.

    • adirondacker12800

      There is a deep fascination with Stamford for some reason. Fairfield County and southeastern Westchester is bit over a million people? They get all the trains to Boston. Long Island gets a ferry to New Haven or slogging on the LIRR to Manhattan to backtrack. The electronic legerdemain of modern trains makes it possible to send trains all the way to Washington D.C. now. There could be a train once or twice an hour that makes all the major stops between Springfield Mass and Philadelphia. The trains all the way to Boston could be going through Jamaica on their way to Yaphank where they make a left to get to New Haven. They could sneak in a train from Saratoga Springs to Farmingdale but nah, Stamford.

      • Nilo

        A long Island Sound tunnel is very very expensive compared with fixing the segment between New Rochelle and Stamford.

        • adirondacker12800

          There are no naval bases west of New London that the Navy can object about. It doesn’t have to be in a tunnel. It can be on a causeway like the Chesapeake Bay Bridge Tunnel without the artificial island and tunnels

          • Eric2

            Still expensive, and the environmentalists wouldn’t allow it (with good cause)

          • Onux

            Naval bases or no, Long Island Sound is an active seaway, which means the Coast Guard, shipping, fishermen, recreational boaters, and yes the Navy, will never allow a causeway that does not include tunnels for ships to pass (or bridges with the same clearance as the East River bridges). Never as in under no circumstances whatsoever, with no exceptions of any kind. Ever.

            Note this means multiple tunnels/bridges as requiring boats at one shore to make a dozens of miles detour to reach a single channel is a non-starter (inshore bridges could be smaller for fishing/recreational boats).

            China completed the Hong Kong-Macau causeway, with an over water span of similar length (including bridges and a tunnel) for about $20B. Alon thinks DC-Bos in 4 hr can be done for $15B. It makes no sense to spend a sizable fraction of that just to build a bridge that has no ROW or tracks at one end.

            On top of all that, Penn station to Yaphank is 58.6 mi, your route adds at least another 31+, or 90+ mi to New Haven. New Haven today is only 75 mi from Penn. Why spend so much money for a longer route?

            Even with 60% of LI’s population being in NYC, it is still large (Nassau+Suffolk is larger than Rhode Island+Delaware) and should absolutely get intercity service. Outside of NYC the majority of the population is along the south coast. The best bet is to build a tunnel from Hoboken to Atlantic Terminal (also so commuter trains can directly access downtown, the 3rd or 4th largest job center in the country), with intercity trains serving DT Brooklyn (basically the densest part of LI) and then out to Babylon with a few stops past Jamaica.

            LIRR ridership data is opaque. Although the Babylon Branch has the highest ridership, it is unclear if the Ronkonkoma/Port Jefferson branches combine for higher ridership along that route. It may be that intercity trains from Jamaica out the Main Line to Ronkonkoma would be best. Either way, geography dictates that the best route from NY to Bos does not go through LI.

          • adirondacker12800

            Alon thinks DC-Bos in 4 hr can be done for $15B.
            Gateway is somewhat beyond 20 billion, there goes his 15.
            Alon also thinks he’s going to be able to straighten a curve or two by putting elevated railway over Ye Historick Olde Boston Post Road in Fairfield County. I want his dealer’s number, his dealer has good stuff.

          • adirondacker12800

            They weren’t proposing building a new railroad over an old road in Atherton. They were proposing upgrading an existing railroad. People have wanted to widen Post Road since those new fangled automobiles started to clog, it a century ago. It hasn’t happened. Putting a tunnel under it perhaps but not an elevated. Tunnels get very very expensive.

      • Nilo

        A Long Island Sound tunnel is a very very expensive proposition compared to fixing the span between New Rochelle and New Haven.

  3. Matthew A da Silva

    Is triple track sufficient for overtakes? Thinking about the practicalities of intermingling express trains and frequent local service on the M&E, Harlem Line, LIRR Main Line with existing infrastructure…

    As for Port Jervis/Newburgh and Lehigh Valley, the RVL and Erie Main Line both have quad track ROWS – it’d be technically easy to add the infrastructure needed for timed overtakes for expresses. In fact, in the 1960s the service vision for the RVL was for PATH to take over the local tracks from Elizabeth as far as Plainfield, and trains going further west would run on dedicated express tracks between Aldene and Plainfield, only stopping at Cranford, Westfield, and Plainfield.

    • Richard Mlynarik

      In general, triple track “for overtakes” where the limited-stop trains in both directions use a single (central) passing track, is not desirable because in general symmetrical (“Takt und Symmetrie”) timetables are desirable, and this generally (and desirably!) puts the overtakes (running, or cross-platform) in the same position, hence four tracks.

      America’s Finest Transporation Planning Professionals like triple tracking because they only want or can understand tidal traffic flows (whether full-vs-empty “commuter” or full-vs-empty coal trains) and don’t give a damn about timetable construction or service quality.

      Triple track where local service runs bidirectionally on a single track (with limited sections of quadrupling for meets) while limited service runs semi-independently on the two other tracks can work pretty well. My go-to example is on the Swiss main line along Lac Leman: three tracks, ~28tph, 4tph all-stop locals (some quadrupling around some local stations, and 4tph terminating, reliably, at a single platform.)

      Click to access G101.pdf,+Gare/@46.3173354,6.185701,17z

      That said, SBB is nearing completion of four-tracking the triple-track capacity pinch-point between Lausanne and Renens further north:

      Click to access G113.pdf

      • Alon Levy

        The TM Worcester Line regional rail timetable works like that, except that the bidirectional track is the express track, not the local track. The overtakes are on the fly and don’t have to be at the symmetry point; triple track is really annoying to deal with but it is possible.

    • Henry Miller

      Triple tracks can work with careful planning. As Alon said, it is really annoying, but you can plan it out up front and make it work.

      The problem is your upfront plan forever locks you into that service. You will run your trains at the same speed, at the same frequency, on the same schedule, at the same length (you can go shorter), and to the same stations. You have to account for all of those in order to get your switches in the right place so that the overtakes happen, and any change to one means your switches are in the wrong place. No adding more trains to a popular line. No adding extra service for some special event on the line. No adding infill stations. To get the most out of this system you also need to choose where you put your stations, which may not be the right place for a station by any other metric. You can take away service if you need to save even more money, but you can’t adjust the other trains to get a nice even schedule pattern.

      There is one other problem that deserves separate discussion: overtake takes a lot of space because of safety margins. By the time you account for all the margins of safety required to assure there are no collisions you have greatly limited your total TPH. If you have enough demand to think about running two classes of service don’t you have enough demand to want to run more TPH than you can get with timed overtakes?

      If you are still thinking about this, make sure you reserve enough space so that in a few years when you need to make changes you can build that forth track. Someone will thank you.

      Running single track, with only one class of service might make sense to make a marginal line work. This is one class of service that is just barely going to work out, and even with growth won’t fill more than a few TPH. You will probably find it cheaper to build a station than a full speed overtake at each overtake point, and add stop for some area that otherwise never deserves a station. You are still limited to what you plan upfront, but since this is a marginal area you have more confidence the next generation won’t curse your lack of foresight.

      • adirondacker12800

        they get more than 30 trains per hour on the Flushing Line of the NYC subway. Local and express service, in the peak direction, east of Queensboro Plaza, on three tracks.

  4. Russell.FL

    Electrification north of Croton is a tricky puzzle. If the plan were to only electrify up to Poughkeepsie, then continuing the third rail makes sense, as the top speed of third rail is 100 mph generally speaking, and the top speed on the upper Hudson is 90 mph I believe. But if the long term plan is to electrify to Albany, that gets a lot more difficult. Do you electrify with Catenary starting at Croton? Do you have the transition point at Poughkeepsie? The latter could make sense as it allows Metro-North to use third rail only equipment and then also not have catenary in the most scenic part of the Hudson Valley. You wouldn’t have third rail all the way up to Albany, since that would cap speeds at 100 mph when there are already sections at 110 mph, and the goal should really be to get those parts up to 125 mph. So what equipment would Amtrak use? The dual modes max out at 59 mph on third rail, and switch to diesel mode as soon as they’re out of the tunnel on the west side of Manhattan. I suppose they could use something like the Eurostar Class 373 that could use the third rail sections. All this makes one think that the status quo will likely stay for the time being.

    • Matthew A da Silva

      I guarantee it’ll be catenary for the simple reason of needing to have a DC substation every 3-5 miles for the 40+ miles to Poughkeepsie is not cost effective. Add in the Amtrak intercity case and the fact that Metro-North already has infrastructure to maintain the M8s and its a slam dunk for catenary north of Croton.

      Amtrak can wire up the Empire connection as well and use MUs or trainsets with both underrunning third rail and catenary abilities – ditching their use of LIRR’s third rail.

    • adirondacker12800

      If you maintain the track for a maximum speed of 90 mph the maximum speed for the trains is 90. That makes sense for Metro North because the service speed of their trains is 90. The rumor on is that they maintained the express tracks for 110 way back when the Turboliners were running and the geometry is better than 90 in places,. That costs money no one wants to spend.

      Even with under-running third rail like Metro North uses, third rail is a bad idea in winter. I’m sure they have decades worth of accounting information from the New Haven Line for cost, on time performance etc and the lack of a substation every mile or so. There are 11kV feeders under NYC subway platforms to supply the substations. DC traction made a lot of sense in 1900, not so much in 2021. 25kV/60Hz north of Croton. North of Highbridge would be even better. The M10s or M12s could be tweaked a bit to have a slightly higher max. speed when on AC. But then if they stuff another 2 million jobs into Manhattan the Second Avenue Subway needs to have local and express service and the Tenth Avenue subway needs it too. Might want to run subway express trains to Yonkers instead of Metro North trains. Third Ave in the Bronx can have the Third Ave subway trains all to itself. I haven’t thought about where the four more tracks under the Broadway/BMT should go.

      Amtrak’s cockamamie plan to serve White Plains Airport, Danbury, Waterbury and Hartford is fabulous from my point of view because the high speed trains to and from Albany could also use it. Twice an hour to Buffalo and Ohio, twice an hour to Toronto, twice an hour to Montreal and twice an hour to Saratoga Springs with one of them making all the stops between Albany and Poughkeepsie, it’s too many trains to be sharing tracks with Metro North Expresses.

    • Onux

      Just the opposite, the entirety of Metro-North should be electrified as catenary, removing the third rail. Third rail is good for subways because it gives you better clearance if you are in a tunnel, but catenary (and as Adirondacker notes, 25kv/60Hz) is far superior for mid or long distance. The Metro-North tracks need to be shared with intercity trains (on the Hudson and New Haven lines) and Metro-North is all above ground except for the last 2.75 mi into Grand Central. The need is prioritize inter-city and eventually HSR service over hundreds of miles, not account for a few miles in a tunnel.

      There are vertical clearance issues in the tunnel under Park Ave, I believe, but there are ways to address this (ballastless track can lower rail height and overhead rail in place of wire can be closer to the top of the tunnel). Any future mainline rail tunnels (Gateway, Penn-GC connection, etc.) will be guaranteed to have the clearance for catenary.

      • adirondacker12800

        Supposedly the clearances in Grand Central are too low for high voltage. It’s probably unwise to have trains doing modern electronic legerdemain south of Woodlawn or Highbridge-ish. If something decides to get confused, it will sooner or later, resetting things doesn’t affect the other lines. The usual suspects can figure something out, like they do with the M8s, where the lower voltage DC is going into the magic. No reason why it couldn’t be from a pantograph instead of third rail but the third rail is already there. Meh.
        There is never going to be a Penn Station – Grand Central connection. It will cost too much money for the few people it will serve.

        • Car(e)-Free LA

          TFW you think a Penn Station-Grand Central connection in the heart of the biggest city in the United States is too expensive and will be underused while thinking a causeway across Long Island Sound is a practical and useful idea…

          • adirondacker12800

            It’s the biggest city in the U.S. with a complicated extensive subway system. There are alternatives for getting from 61st Street and Third Ave, other than using the Lexington Ave Subway to Grand Central, that get to Penn Station. Or 52nd and Madison. If you are starting out in Brooklyn why would want to go to Grand Central? Though if the trains to New England run through Queens someday when the the LIRR goes to Wall Street running through to New Jersey people in Brooklyn could arrange themselves to get on the half hourly train that stops in Brooklyn instead of schlepping to Manhattan. There are almost as many people in Brooklyn as there are in Baltimore and quite a bit more than there are in greater Stamford. 2.8 million in metro Baltimore, 2.5 million in Brooklyn and 2.2 million in Queens. And 0.9 million in Fairfield County Connecticut. The people in Stamford will still be able to get to Boston or Washington D.C. on the half hourly trains that takes the slow route between New Haven and New York.

          • Car(e)-Free LA

            The obvious utility is New Jersey-Midtown East Service. Which is an absolutely enormous traffic volume–and serving it is way better use of money than working on your weird priorities like serving Islip to Boston or whatever.

          • adirondacker12800

            Extend East Side Access to Secaucus where people from the former “Erie” lines with one seat rides to Wall Street or Penn Station can change for Grand Central. Send a third of them to Broad Street Newark where people from the former DL&W lines can change from their one seat ride trains to Wall Street or Penn Station and two thirds of them to Penn Station Newark where people from the former PRR and CNJ lines can change from their one seat rides to Wall Street or Penn Station. Without building Son of East Side Access under Madison Ave so the few people on expense accounts in the few blocks around Grand Central don’t have to take a cab to Penn Station.
            It’s not Cleveland where everybody wants to get to Public Square. Or even Chicago where everybody wants to get to the Loop. Sending trains to Wall Street keeps people out of Midtown. Which is a bigger destination than anything other than Midtown or the Loop. They can spend the money building Son of East Side Access there.

          • Onux

            How are you going to “Extend East Side Access to Secaucus” without the “cost too much money” you stated for the GCT to Penn connection? You are aware that ESA sits far beneath GCT, and any deep tunneling to take it across Manhattan and then across the Hudson and then across the Palisades to Newark would be far more extensive then connecting existing GCT and Penn tracks.

            If you are going to build all that infrastructure, why would you do it only to give people on the Erie/DLW/PRR lines *transfers* to Midtown East, followed by another route to Long Island which in a world with through-running they already have via Penn? If you four track from Newark to the Hudson, build two new Hudson tunnels, and connect Penn to GCT you can give those same people a one seat ride to Penn AND GCT, plus continued service to Harlem, Westchester and other places people want to go. Plus Hudson/Harlem/New Haven line riders can get to Penn, or Newark, and other places.

            Moreover, Penn-GCT allows trains from south of NY on the NEC to go north to Albany, etc., which means connecting Phila, etc. to upstate NY, not just Boston.

            Hoboken-Wall St.-Brooklyn is a fantastic idea, but irrelevant to GCT-Penn, except insofar as to note there are probably more jobs located within a similar radius of GCT as there are of a future Fulton St commuter train station.

          • adirondacker12800

            No station between Grand Central and Secaucus?
            All of Long Island and all of New Jersey getting one seat rides to Grand Central, Penn Station and Wall Street is too much capacity. Somebody somewhere is going to be changing trains. It’s faster than schlepping on the subway, they’ll do it. They do it now at Jamaica, Newark and Hoboken.
            East Side Access is already there. Tunnel is relatively cheap. Station caverns with stairs and elevators etc. are very expensive. It could swoop across 35th Street without much fuss. Two station-less tunnels to Grand Central instead of four to Son of East Side Access to mindlessly give people in New Jersey a one seat ride to Grand Central. It means there isn’t third rail splattered all over New Jersey, just to Newark.

          • Onux

            Connecting GCT to Penn would also involve two stationless tunnels, except they would be shorter, and cost less because they wouldn’t be as deep, and wouldn’t go under the Hudson. Even an ARC alt G plan would be shorter because you wouldn’t need to tunnel past Penn and it’s platforms/yards, you would use them as part of the ROW.

            Yes transfers are part of any good system, but there is no point putting two tunnels across the Hudson where EVERY trip through them is a transfer from NJ. Not is giving people in Newark the option of Penn-Jamaica and GCT-Jamaica when then could have a GCT-Westchester option.

            Forgot to mention extending from ESA would lock your plan into the fixed low clearance of the 63rd St tunnel, which means M9 size equipment only, ever, no possibility of Amtrak, bi-levels or anything using catenary. Foolish for such an expensive and important piece of infrastructure as crossing the Hudson to Manhattan.

          • adirondacker12800

            NJtransit runs 21 trains an hour now. . It’s why they are building two new tunnels to New Jersey. To get all the trains they are might be running two days after Gateway opens isn’t enough capacity to send them all to Grand Central on two tracks.
            The few dozen people an hour who want to go from Jamaica to Newark can get there though Wall Street. The only reason I can think of wanting to go through Grand Central from Jamaica to get to Newark is if I was feeling especially masochistic.
            Yes it would mean something M7-ish to go through the restrained tunnels. Vaguely like PATH trains going through their restricted tunnels. During regular service the Port Authority is a bit vague about service between Newark and the World Trade Center. It’s “every 3 to 5 minutes”. Every six from Journal Square or Hoboken on the Uptown line. They could fudge a little and have the schedule to Penn Station Newark, from Grand Central, read “every 5 minutes” and the schedule for Broad Street Newark be “every ten minutes”. I bet they could even arrange the schedule through Broad Street so that the express from Gladstone has a timed transfer that is a bit longer than the timed transfer the local from Montclair would have. Though there are three tracks at Broad Street. They could both have the same timed transfer in the peak direction! Where people from the Gladstone train to Penn Station and the Montclair train to Wall Street have one too!

        • Onux

          I’m not sure what “modern electronic legerdemain” is supposed to mean; there is nothing deceitful about 25kV overhead wire, it is a system in widespread and reliable use across the world, including in the US.

          My understanding is the clearance issues come from the Park Ave Tunnel, not Grand Central itself, however even if there are clearance issues in the station they can be addressed in the same way (lower track, raise conductor). Lots of stations around the world have high voltage service underground, so it is not an insolvable technical issue.

          • fjod

            There’s no need do that though. High speed trains exist that can run on AC and DC; high speed trains also exist that run on third rail and catenary. You do not need to change the electrification on a whole line just because AC-only trains are going to run on one part of it.

          • adirondacker12800

            They take whatever the A.C. supply is, convert it to D.C., convert the D.C. it to variable voltage variable frequency polyphase to run the motors. it’s not 1915 anymore with the New Haven or the PRR running D.C. motors on 25Hz. And it can run backwards sending the braking energy to the lighting and the HVAC or even back to the supply. The ACS64s do that and it’s going to save hundreds of millions of dollars over their lifetime.


            Quite popular outside of railroading, software changes the characteristics instead of high maintenance contacts, switches and relays etc. If you have a newer refrigerator, likely that. Or a newer high end window air conditioner. The more efficient versions of central air conditioning. Slightly difference paradigm. It’s how the fans in your computer change speed.

            I suspect what the electrical engineers did for the M8s was to design the intermediate D.C portion to use nominal 750 volts D.C. and that gets injected in the middle of the of the sophisticated wizardry. You don’t want to do that changeover in the middle of a busy main line. Things get confused occasionally and have to be reset. It can happen in the vicinity of the merge/diverge from the main line where it disrupts less traffic. The third rail is already there. Meh.

            The conversion from 60Hz to 25Hz for NEC and SEPTA isn’t 100 percent efficient. Everything except the M8s can now run on 25Hz or 60Hz. The conversion to 60Hz can happen in chunks. Wilmington to Washington D.C has the least traffic and the longest stretches that can be high speed cheap. Perhaps that.

          • Onux

            “There’s no need do that though. High speed trains exist that can run on AC and DC”

            The issue isn’t AC vs DC but high vs low voltage. Running low voltage carries a performance penalty, which is why all new build HSR systems use 25kV. Metro North rail is 750V, no high speed train runs *at high speed* on 750V (even if they can do so to access legacy route or stations). You can install equipment to run on both, but this carries a cost, capacity, and performance penalty, since you have to buy, fit in, and carry around equipment for both phases/voltages/etc. Why not just settle on one (the one that is required for HSR performance)?

            “high speed trains also exist that run on third rail and catenary”

            Third rail cannot carry very high voltage because, being closer to the ground, the voltage would arc to earth. See above on the need for high voltage for high performance. All modern HSR systems use catenary, and if they have third rail shoes it is to access legacy track, not for use on high speed lines.

            “You do not need to change the electrification on a whole line just because AC-only trains are going to run on one part of it.”

            But the AC-only trains should be running on basically ALL parts of it. The entirety of the New Haven and Hudson lines should host HSR (to Boston and Albany/points beyond respectively). It wouldn’t be HSR, but the Harlem line should be the start of a Regional service to Danbury/Waterbury/etc. For interoperability, electrification needs to be standardized, just like rail gauge. 25kV is the standard for modern HSR, so it should be the standard for all mainline rail services, including commuter/regional rail like Metro North.

          • adirondacker12800

            Harlem line should be the start of a Regional service to Danbury/Waterbury
            One of the rumors about why they re-electrified the New Haven line at 12.5 kV instead of 25kV is that the clearances aren’t good enough. If there are four inverters on the car doing their magic, switching between 12.5kV and 25kV is nearly trivial. Or even two. You probably don’t want one because with two if one goes PfffFfT the other can let things limp along. Like the lights and the HVAC.

            There’s not a lot of demand out there. There are these things called automobiles for places few people want to go. They are especially good at the times even fewer people want to go there. Alternately I could drive to one of the stations, hope to find parking and schedule an Uber to wherever I’m going. Instead of just driving there on I-84. Google maps says it’s 37 miles from Brewster to Waterbury.

            The few people out there rarely get the urge to shop in Waterbury’s Walgreens instead of the CVS in town. What are they going to Waterbury for? Or what is in Mount Kisco that they can’t get in Danbury?

          • fjod

            @Onux: you were proposing lowering the track through the Park Avenue tunnel and possibly the whole of Grand Central, at considerable expense. This is not track that you’re going to see trains running at high speeds on, so there is no need to electrify it solely for these speeds. There are no capacity drawbacks (unless you can provide a source for that) and very small performance drawbacks – as far as I can see, the ICE trains designed for 4 (!) different electrification systems have the same speed and acceleration as the German-only trains but a 6% gain in weight. Any cost drawbacks need to be balanced against the huge expense and severe service disruption of closing the stretch of line out of Grand Central, lowering the tracks and installing catenary.

            And given the Harlem Line is already third rail to almost-Danbury and you’re not proposing running high-speed service over it, there is also little reason to re-electrify it. 100mph top speeds for this alignment and service pattern are sufficient.

            Ultimately, if you had limitless funds and institutional bandwidth then maybe you could do this, but I don’t see how it can be anything like a priority given all the more useful and cost-effective improvements you could make.

          • adirondacker12800

            People in Danbury or Waterbury can get to Grand Central without going through Brewster or White Plains. On trains that go to Danbury branch or the Waterbury branch of the New Haven line.

          • Onux

            @fjod: Lowering the track with ballastless (or slab) track is not expensive. It means instead of several feet of rock (ballast) with ties/rails on top, you have a foot or so of concrete with ties embedded, gaining a few feet of clearance. It doesn’t involve excavation, and the rails and ties have to be replaced from time to time anyway. There are machines that automatically remove rails/ties/ballast, and pouring concrete is not unusual; slab track is the standard for high quality installations around the world now.

            The track will not have high speeds, but it should see high speed trains, eventually HSR from NYC to Buffalo or beyond. Ideally (with a Penn-GCT connection) this should be the route for DC-Bos trains, since Penn to New Rochelle via GCT would be 1.5 mi shorter than the current route (with average slower speeds in NY, this could save 1.5% of travel time NY-Bos).

            A 6% gain in weight means 6% more power which means 6% greater energy costs *for every mile travelled the entire lifetime of the train.* Since the GCT approach tunnel is less than 3 mi, that means HSR trains would be carrying that 700V deadweight around for more than 99% of the time they are in service. Given greater upfront cost for a dual voltage train and the 700V is useless for anything else, even a 2% weight increase is foolish.

            Multi system trains in Europe do not have the same speed and acceleration on all voltages. For instance, the PBKA Thalys trains to Paris, Brussels, Koln and Amsterdam produce 8,800kW under 25kV, but only 3,680kW under other voltages, and thus are limited to 200kph in Germany instead of the 300kph in France. My understanding is planned service to Frankfurt had to be cut as a result, but I couldn’t find a link.

            You cannot avoid these costs. The Hudson and New Haven lines have to be 25kV catenary if you want HSR to Albany and Boston. Omitting the last three miles to GCT isn’t saving much. Wiring the Harlem line is an expense, but it means the MTA can standardize a 25kV fleet, saving capital cost by omitting dual voltage from rolling stock, and saving operating costs by being all-electric with no diesel. Although 25kV catenary is not required for Harlem line speeds, it can still be used, supports higher speeds in the future, and allows future longer distance services to use the route with standard equipment (i.e. same trainsets as the NorthEast Regional, or Shoreline East, or anything) – Danbury and Waterbury have service to GCT now, but not to each other, or to White Plains/New Britain/etc.

            All mainline rail services (commuter, regional, HSR, etc.) touching the NEC should be standardized on 25kV 60Hz.

          • adirondacker12800

            Going over the Hell Gate Bridge has approximately zero construction costs.

          • fjod

            @Onux – Hang on, I was assuming it was only Metro-North trains that would be using this section of track; you said above that intercity trains would only need to be sharing tracks on the Hudson and New Haven lines, implying they would not be doing so on the Park Ave tunnel. I don’t think you (or Alon) have referred elsewhere to a Penn-GCT connection being used for HSR, which is the only reason that all the expense to rewire that tunnel plausibly might make sense. But I don’t see why you would want a Penn-GCT HSR connection, for all the reasons others in the thread give.

          • Onux

            Even if only Metro-North trains use the Park Ave tracks/tunnel (NY-Bos uses Hells Gate, NY-Albany uses Empire Connection), it is still worth wiring it (and the Harlem Line) for 25kV catenary, for a number of reasons:

            1) It lets MN get rid of diesels, which are inferior to EMUs in every way (more expensive to buy fuel, break down more often, higher emissions, lower acceleration/trip time). This increases flexibility/reduces fleet size, because right now an M7A on standby for Croton service cannot sub for a broken diesel serving Poughkeepsie. If the rest of the Hudson/New Haven are being given 25kV for intercity service it just makes sense to electrify the rest of MN territory to take full advantage of hundreds of miles of new electrification.
            2) It means cheaper trains in the future. The M8s with their tri-voltage equipment are more expensive than if they had 25kV only. It also means MN can get greater efficiency with larger orders of a single type of rolling stock, instead of M8s for New Haven, M7As for 700V only, diesels with third rail for unwired.
            3) These lines have value for regional service farther out than commuter trains, but not as far/fast as HSR. All nations with effective rail systems run this kind of service (RegionalBahn, InterRegio, etc.). The services should naturally end at GCT, and have to use 25kV since they will ru on the same tracks as the HSR. Making all of MN 25kV means this kind of service is possible. Note that because of the lower frequency of this type of service, it isn’t cost effective to buy custom trains with 700V third rail just for a few miles use. You want to be able to use standard trains in service on other parts of the network.
            4) The Penn-GCT connection is long-term, but not ‘aspirational’ in the sense that it needs to be done. There cannot be RER or Crosslink style through running in the NY region without it. As previously mentioned it also has benefits for a NE HSR network by letting upstate NY access DC-Phila. If there is going to be through running there should be common voltage on both sides of the line. Yes, everyone knows dual/tri-voltage trains are possible, but there is no reason to run them if you can avoid it.

          • adirondacker12800

            The MTA and CDOT have platoons of bookkeepers and accountants who have computers and detailed information going back to when the New York Central’s and the New Haven’s bookkeepers and accountants were doing it with those new fangled punch card machines. When they re-electrified the New Haven they didn’t get the urge to run third rail all the way to New Haven or put catenary in Grand Central.

            Try to keep two things in your head at time. This isn’t a Red Line and a Yellow line that could be combined to make an Orange Line with a Purple Line and Green Line that could be combined to make a Brown Line.
            There are four tracks of Metro North, four tracks of LIRR and someday four tracks of NJTransit. Send all of Metro North to New Jersey the LIRR gets to turn trains around in West Side Yard. Sending Metro North to New Jersey doesn’t get anybody to Wall Street. New Haven line passengers can already get to New Jersey. On Sundays when there is a football game they already do that. The Metro North Train-to-the-Game train magically turns into the train to Trenton and they change in Secaucus to get to the Meadowlands. They don’t need to go through Grand Central to get to New Jersey. Not since 1917.
            There are 8.8 million people in New Jersey, 7.6 million people on Long Island and maybe perhaps a million along the Harlem and Hudson lines? Running LIRR trains to New Jersey likely serves more people. Through Wall Street so they don’t have to get on PATH or go through midtown to get there. The Great Neck local can go to Matawan, The Floral Park local can go to Plainfield, the St. Albans branch local to Summit and the Locust Manor Branch to Ridgewood. It’s not an Orange Line and Brown Line. If I’m on the Great Neck Branch it’s fabulous if I want to get to Newark, Newark Airport, North Elizabeth, Elizabeth, Linden, Rahway, Avenel, Woodbridge, Perth Amboy, South Amboy or Matawan If I want to get anyplace else at all I’m changing trains anyway. Rinse repeat for the other three branches. The Hudson line passengers will be able to get trains to Penn Station soon after the New Haven line passengers. They can change trains in Penn Station too. Harlem Line passengers would be able to get one seat rides to limited number of stations in New Jersey reallllly big Whoopee. You don’t want to send all four tracks of Metro North to Staten Island, there isn’t enough demand. They would be well under capacity south of Grand Central. Curing that problem by only sending two track south of Grand Central means people are changing trains.
            It’s not an Orange line and Brown line.

          • fjod

            @Onux: I think you can get most of these benefits anyway.
            1) Just electrify to Poughkeepsie. Under these plans, you could combine that with re-electrification from DC to AC from Croton-Harmon south to Spuyten Duyvil to allow you to run trains from Albany etc to Penn station.
            2) I am not sure the savings from standardised trains will weigh up positively against the cost of re-electrifying the tunnel under any realistic timescale. I also think you could run the same dual-voltage stock on all 3 lines, minimising any higher train purchase costs.
            3) I agree, but again I think you have to balance the small-scale cost-ineffectiveness of dual-voltage trains against the large-scale cost-ineffectiveness of re-electrifying. My intuition is that the latter is worse. In coming to this conclusion I note that Thalys trains have 1500V DC capability solely for the last ~15km into Amsterdam and the approaches to Rotterdam C.
            4) RER B and Thameslink quite happily switch between DC and AC and have been doing so for decades so I don’t think this is an impediment to through service. There is indeed no reason to run dual-voltage trains if you can avoid it, but the costs and disruption of avoiding it are worth considering!

          • adirondacker12800

            RER B and Thameslink quite happily switch between DC and AC
            They have been doing it on the New Haven Line since 1907.

      • Michael Whelan

        Why would you bother to re-electrify the Harlem Line and the Park Ave Tunnel when you will never have intercity trains running through Grand Central? Just cut the Hudson Line third rail back to Marble Hill, string up wires from there to Albany, and buy some more M8s for Metro-North (or rather, lighter M8s that comply with the recently-improved FRA requirements).

        Eventually, New York should build Alon’s proposed regional rail system, including the tunnel connecting the lower level of GCT to Staten Island by way of Union Square and Fulton Street. It’s probably easier and cheaper to build that tunnel if you don’t have to clear it for catenary.

        • adirondacker12800

          Grand Central Station goes “ten stories” down under the parts you can see from the passenger concourses. Filled with all sorts of interesting stuff like the 750 volt D.C. electricity supply. And the city went and built subways on 42nd. You need accurate three dimensional models of what is along Park Ave from 59th Street to 40th or so to determine if it possible to send trains across 41st.

          • Jan

            > And the city went and built subways on 42nd.

            Curiously enough, I’ve stumbled across a report on the then new Grand Central Terminal in a German civil engineering newspaper from 1912, which mentions – and also shows in an included track diagram – a planned connection between the lower level tracks and the (pre-Dual Contracts extension) subway.

          • adirondacker12800

            Yes, things have changed since 1910. The tunnels the Flushing line uses were for the trolley cars that were going to service the villages out in Queens. I’m assuming they thought the LIRR passengers would still take the ferry from Long Island City to the terminal in Manhattan and change to the El. Supposedly there is still space reserved for the H&M at Grand Central. So people could get from Grand Central to the DL&W terminal in Hoboken, the Erie terminal, PRR terminal and the CNJ/Lehigh/Reading/B&O terminal in Jersey City. I’ve seen a blueprint of the proposal to quad track the H&M in Jersey City to handle all the traffic. Maps from that time highlight that you can take the H&M to Hudson Terminal, change to the brand new subway to Brooklyn and use the LIRR that way. The original plans for the Sixth Avenue subway sank the H&M down under the subway so it could do that. Under the local tracks so that some day there could be more express trains on that level. Things have changed.
            ……The pedestrian tunnel between Penn Station and Herald Square wasn’t there to be nice to pedestrians. It was so you and your porter could take your bags from Penn Station to the terminals in New Jersey. Or vice versa. With suggestions in the etiquette manuals on how much you should tip him. A lot since he’s going to spend a half hour. Things have changed.

  5. Matthew Hutton

    I seriously think we should just bite the bullet and electrify all our railways – even ones like the far north line.

  6. adirondacker12800

    it is possible to run express trains to the Hamptons on the Babylon Branch, but it imposes a real cost on frequency to dense suburbs and should therefore be avoided.

    I’m confused. There are 4 tracks under the East River. I don’t want to attempt to figure out what is going on in Sunnyside. Two tracks head off to New England with a train or two. There’s six tracks through Woodside, Port Washington trains divert off east of there. Four tracks between there and Jamaica where there are 6 island platforms served by 10 tracks. If I remember correctly the platform diagram in Wikipedia doesn’t show tracks that express around Jamaica. Four tracks between there and Floral Park and four tracks between there and Lynbrook. The South Shore is more densely populated and gets more service. But even if it’s two thirds of what left Penn Station they don’t need more concrete to run a few more expresses. They don’t even need electronics. Just someone who doesn’t want it to be a red line, blue line and green line. Or even more electricity because they do it with dual modes. There isn’t any electricity east of Babylon. Until recently not even any signals. I’m confused.

    It’s not three dozen people who want to go to the Hamptons. They are able to fill whole trainloads to standing room only. They don’t need to be on the standing room only commuter train. They are ready to start their weekend and it’s not unheard of for it to get into a very …… jovial party… mood. They can be congenial and have some bonhomie on a separate train. Nor are they expecting to catch the 5:27 express to Suburbia that has been at 5:27 since electrification. They arrange themselves to whatever this summer season’s schedule is. You have heard of the little project to divert some of the people from the standing room only commuter trains out of Penn Station to Grand Central? The are expecting it to be 20 an hour at peak someday. Not three car MBTA trains or even six car SEPTA trains but 10 and 12 car M9s. And they aren’t competing for space with the people who want the Fire Island ferries.

    Someday it’s going to be six into Manhattan There will be three tracks to Hicksville by then and trains can also get to Babylon that way now, through Bethpage instead of Lynbrook. I’m confused.

  7. Magnus

    It’s interesting how many popular leisure destinations that are 50 to 100 miles from Manhattan. They could all be very well integrated with Manhattan with a 200 mph concept that stops no more than every 10th to 15th mile before it is 50 miles from Manhattan. At that distance car trips pass the 1 hour limit, while the 200 mph train still could be less than half an hour from Manhattan, even including stops in places like White Plains, Queens and Newark.

    China have such a line from Beijing, stopping 8 times in 100 miles heading for Tianjin and finally Pacific Ocean harbour city Binhai, making the trip in around an hour.

    I haven’t got much faith in concepts along Hudson or up west, but I think lines to Connecticut, Long Island, Jersey Shore and Delaware seem very interesting. The line could be packed with commuters and leisure travelers at different times ensuring even usage through the week. The concept must be loyal to a average speed of around 100 mph including stops, meaning car travel outside rush hours should take almost twice the time to NYC.

    Possible stations on the four different lines:

    Connecticut Express: Should make it to Hartford in an hour including stops in South Bronx, Woodlawn, White Plains (with connecting buses from west side of Hudson, Rockland County), Stamford, Bridgeport and New Haven.

    Jersey Express: Could reach Brick Township in an hour with stops in Hoboken, Newark, Newark Airport, New Brunswick and Long Branch. Possible extensions to Manahawkin and Atlantic City within the next 30 minutes of travel time.

    Long Island Express: To Southampton in an hour via Brooklyn, Queens (Jamaica), Hempstead, Babylon, Patchogue and Hampton Bays. Could be extended to Montauk via East Hampton, with high-speed ferries to New London in less than an hour.

    Delaware Express: To Philadelphia in an hour serving Hoboken, Newark, Perth Amboy (with connecting ferries and buses from Staten), New Brunswick, Princeton and Trenton. An extension to Wilmington could be interesting as well, maybe via Springfield/Media to activate a part of the Philadelphia region not benefiting from Acela.

    I haven’t calculated the lines but a wild guess should be construction costs of $5-10 billions per line and ridership around 20-40 millions per year.

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 )

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.