Commuter Rail New York

Regional Rail for New York City – Part II

» A proposal for New York regional rail improvements.

This is the second post in a two-part series by guest author Alon Levy. This part describes a proposed regional rail network. The first part explained the reasoning behind the proposal and the general principles underlying it.

In the first part of this proposal, I argued that in order to improve its regional rail system, New York should through-route train services and build new infrastructure with an eye toward making more through-routing possible rather than extending the service area. In this article, I will explain how such a system may be phased. The ultimate plan is presented in the following regional rail map:

This is a long-term plan, which I do not expect to be feasible to construct before 2030. It includes not only through-routing and extra connections through Manhattan, but also extensions of electrification and some new service.

New Connections

The centerpiece of this plan is a set of new expensive tunnels in Manhattan, which would allow through-running of all services and considerably improve travel times to both Manhattan and the main secondary job centers. These include four major investments:

  1. A new pair of tunnels paralleling the North River Tunnels between New Jersey and Penn Station.
  2. A link between Penn Station and Grand Central.
  3. A north-south tunnel linking Grand Central, Lower Manhattan, and Staten Island.
  4. An east-west tunnel between Hoboken and Flatbush Avenue through Lower Manhattan.

The new construction may be summarized in the following map:

Penn Station access is the most important part of the system, entailing four-tracking the North River Tunnels and linking Penn Station to Grand Central; this was Alternative G in the major investment study for the ARC project. It would be cheaper to do this than to build an entirely new deep-level station underneath Penn Station as is currently planned. Completing this option instead of the current ARC project would increase operational flexibility, as the deep-level station is blocked to the east by Water Tunnel 1 and will only be able to reach Grand Central after Water Tunnel 3 is completed around 2020; in either case, it will not be able to continue through the East River Tunnels to connect to LIRR service. Note that both the four-tracking and the Penn-Grand Central link are necessary. Penn Station currently suffers from two major bottlenecks, train capacity in the North River Tunnels and pedestrian flow in the station concourses.

Despite the Penn-Grand Central connection, some Metro-North New Haven Line trains should serve Penn Station directly through the East River Tunnels. This would relieve capacity in the tunnels north of Grand Central, as well as provide service to Queens and the eastern Bronx. In addition to stations at Coop City, Parkchester, and Hunts Point, there should be a major transfer point at Sunnyside, as well as new stations in Astoria and Port Morris flanking the Hell’s Gate Bridge between Queens and the Bronx. The stations in Astoria and Port Morris would not only provide regional rail service to those neighborhoods, but also allow transfers to the proposed Triboro Rx circumferential line.

Giving Lower Manhattan regional rail access is almost as important. The MTA is already building the Fulton Street Transit Center, which would sit atop twelve subway lines and PATH; it is only necessary to bring regional rail there.

There are recurrent plans to extend the LIRR from Flatbush Avenue to Lower Manhattan. However, in order to make full use of the Fulton Street station, Lower Manhattan should be connected not only east to Brooklyn, but also south to St. George on Staten Island, north to Grand Central, and west to Hoboken.

Direct regional rail service to Manhattan from Hoboken would instantly transform the Hoboken Division of New Jersey Transit, giving its riders a one-seat ride to Manhattan. Running under Greenwich or Hudson Street in Manhattan, it could also serve the West Village with an intermediate stop at Houston Street.

Staten Island suffers from inconvenient travel to Manhattan and downtown Brooklyn, by any mode of transport. A tunnel diving under Lower New York Bay would be expensive but well-patronized, as it would take close to 100% of the Staten Island-to-Manhattan commuter market, which in 2000 had 53,000 daily users and has only grown since. Continuing this link further north to Grand Central would ease commutes to both Midtown and Lower Manhattan from both the Harlem Line and Staten Island, further increasing ridership. To serve the neighborhoods on the way, there should be intermediate stops at South Ferry, which most trains could skip, and 14th Street-Union Square, where all trains should stop.

Note that despite the through-routing, some regional lines may remain diesel-operated, in which case they should terminate at either the current Hoboken Terminal or Long Island City.

In addition to the major projects outlined above, Amtrak and the MTA should work to rebuild the Empire Connection along the West Side of Manhattan with third rail electrification and double-tracking, enabling electric Hudson Line trains to use Penn Station. Extra stops at 62nd, 125th, and Dyckman would also provide service to the Upper West Side, Morningside Heights/Manhattanville, and Inwood.

New Lines

In addition to the lines already used for commuter service, in two cases abandoned passenger rail lines should be restored: the West Shore Line on the New Jersey side of the Hudson, and the North Shore Branch in Staten Island. The communities those lines run through have inconvenient commuter rail, but are likely to see much more demand once links to Fulton are constructed; they are also quite dense by suburban standards. Even now there are proposals to operate light rail service on both lines, to serve growing transit needs. This may be used as an incremental solution until direct service to Manhattan is offered.

Of the two lines, the North Shore Line would be easier to restore. The line sees minimal freight traffic at its western end, and is double-tracked. Stations along the corridor still exist, though as they are too closely spaced for commuter rail, some may need to be removed.

The West Shore Line is single-tracked with many grade crossings, and would require substantial rebuilding. Current plans call for diesel service to North Tenafly, about halfway from the Hudson/Bergen county line to the New York state line. For full commuter rail, the line would need both electrification and double-tracking, as well as a new junction from the line to the Hoboken access tracks. There are two branches, separating in northern Hudson County and joining just north of the state line; New Jersey proposes service only for the eastern branch, but the western one should be used as well, as it serves denser suburbs, with a similar number of people working in Manhattan. Those commuters currently drive over the George Washington Bridge, which is used by almost 300,000 cars every day, or take the Lincoln Tunnel buses; rail service to Manhattan would provide them with faster, congestion-free commutes.

There is also a substantial market of commuters to Manhattan far past North Tenafly, and the West Shore Line should be electrified and rebuilt for passenger service as far north as practical. On the map the northern terminus is given as Newburgh, but it may be built in stages. The greatest demand is south of the state line, but there is still substantial demand in Rockland County as far north as Stony Point.

Fixed Routes

The routes depicted on the map are not the only way to through-run service. Any LIRR line, most NJT lines, and with the new connections, all Metro-North lines, would be able to serve Penn Station. The new routing through Fulton would likewise permit most lines to serve Lower Manhattan. This may tempt regional planners to offer ad-hoc through-routing, so that trains from a given line on one side of Manhattan may be sent to multiple destinations on the other side. Such an option would provide more one-seat rides, but frequencies to each destination would be low, especially at the local stations. Consequently, routes should be fixed and predictable. The fixed route principle borrows from the RER, where not only is each line its own system, but also each line’s branches on one side are permanently matched to branches on the other side.

Based on this principle, the New Jersey Transit should allocate resources away from constructing new junctions to permit one-seat ride. The current ARC plans include a direct Penn link from the Erie lines—the Pascack Valley, Main, and Bergen County Lines. This connection should dropped in favor of transfers at Secaucus, unless the Hoboken-Fulton tunnel proves impossible to build.

The maps represent my best estimate of how to match demand. This is easiest to do on lines that have a high market share and are unlikely to see an explosion in demand. With lines that are likely to see much more service in the future, this requires more guesswork. In addition, I had some extra considerations, such as ensuring all three Manhattan destinations—Grand Central, Penn Station, and Fulton—would be accessible from Jamaica.

Trains on most lines would have to be capable of drawing power from both third rail, used on Metro-North and the LIRR, and overhead catenary, used in New Jersey and Connecticut. The New Haven Line trains are already dual-mode, so this would not be a significant obstacle to combining services.

The Complete System

The red line provides local through-service on the Northeast Corridor, complementing the intercity service provided by Amtrak. Its two ends are already the two busiest commuter lines in the country, each with about 110,000 weekday rides. Extra ridership would come only from new transit-oriented development in the suburbs. This system includes the North Jersey Coast Line and the New Canaan and Danbury Branches, but not the Waterbury Branch, which is unlikely to see much ridership. About half the trains should serve Grand Central, while the other half use the East River Tunnels.

The blue line, combining the Harlem Line with the Staten Island Railway, would see a quick rise in ridership: the Staten Island-Manhattan rail tunnel would likely cause a major modal shift, bringing daily rail ridership to 100,000 on that segment. The line’s northern end would be Southeast, with no further electrification due to low demand; there is already relatively little demand in northern Westchester County, so most trains would terminate at White Plains or North White Plains.

The purple line connects the Morris & Essex Lines and the Montclair-Boonton Line to the LIRR Main Line and some of its branches. There is more demand to the east than to the west, so some LIRR trains would use East Side Access and terminate at Grand Central. All trains serving stations beyond Great Neck and Ronkonkoma should be so diverted, as those segments are single-tracked; through-routing requires relatively symmetric service, which requires two tracks. The single-track section from Farmingdale to Ronkonkoma could be double-tracked, but further east there is no demand; on the Port Washington Line, there is no room for double-tracking beyond Great Neck. The total weekday ridership on all lines in the system is currently 150,000, and would only increase with suburban TOD.

The green and brown lines form a system combining the Port Jefferson, Raritan Valley, and Hudson Lines; the Hudson Line, which is the busiest of the three, can enter Penn Station from both directions, helping match ridership. All lines should be fully electrified early as their diesel sections already see many commuters. Total ridership is currently 100,000, but would markedly increase with TOD in Yonkers and Tarrytown and electrification on the Raritan Valley Line.

The yellow, orange, and aqua lines run on the LIRR’s Montauk and Atlantic lines through Fulton and the NJT’s Erie lines, including the West Shore Line. The aqua line to West Hempstead would run exclusively to Grand Central. The Montauk Line would be matched with the West Shore and Pascack Valley Lines, the Long Beach Line with the Main Line, and the Far Rockaway Line with the Bergen County Line. I project relatively high West Shore traffic, as its cities have more Manhattan-bound commuters than the other Erie lines. Electrification on the Port Jervis Line is provided only as far as Harriman, as beyond that the line is slow and circuitous and cannot compete with cars. Total ridership is currently 90,000 on the LIRR side and 33,000 on the NJT side, but direct connections to Manhattan and West Shore Line service should raise and equalize the numbers.

The goal for the complete system should be a doubling of ridership. Currently, about 400,000 people use the three commuter rail systems every day. This compares with a total market of 750,000 people, consisting of traditional and reverse commuters, as well as people who live in the suburbs and work in an Outer Borough with good rail access or in a suburb on the other side of Manhattan. Although with TOD we could see a higher suburb-to-suburb modal share even without crossing Manhattan, most riders would still be traditional commuters, and much of the initial increase in ridership would come from the new connections to Staten Island and the Erie lines.

Fares and Fare Control

In order to speed up service and save labor costs, the regional rail system should mimic rapid transit and have fare gates instead of conductors who collect tickets. This is not unusual for mainline trains—even intercity trains sometimes have fare gates, for example on the Shinkansen. A rider would have to swipe at both entry and exit, as fares would remain distance-based.

The fare itself may be kept at its current levels. However, the fare systems would be unified, so that one could travel from one suburb to another on one ticket. As this would involve travel in both the peak and the off-peak directions, such a combined ticket should not cost more than a ticket from the further of the two suburbs to Manhattan. This would provide a discount to people making use of through-routing, who are essential to the system’s full success.

The fare gates may be placed either before or after station retail. At major stations such as Penn Station and Grand Central, the retail should be outside fare control, and through-tickets should allow one to swipe out and back in within a time limit without paying an extra fare. However, at transfer stations with little origin and destination traffic, such as Secaucus and Tonelle Junction, any retail may be placed within fare control.

What Can Be Done Now

Without new construction, it is already possible to through-route some services, increasing the range of the affected secondary job centers. The fully electrified New Haven Line and Northeast Corridor Line can be combined to form a single regional route, as can the Morristown Line and the LIRR Main Line, which are electrified deep into suburbia. Such a system would require moving dual-voltage trains from the New Haven Line to the Morristown-LIRR line and AC-only trains from the Morristown line to the combined Northeast Corridor line, but would not need new rolling stock.

Once the Empire Connection is electrified, Hudson Line trains may continue eastward onto the LIRR as well; as an added bonus, Amtrak could then operate electrified Empire Service trains as far north as Croton-Harmon, reducing fuel consumption and improving the run time. Note that if Amtrak electrifies the Hudson Line with catenary as part of a high-speed rail program, then it would make sense to add catenary to the portion of the Hudson Line between the junction with Amtrak and Grand Central, to enable trains to run from the Raritan Valley Line to the Hudson Line without requiring the use of third rail.

Finally, it is relatively cheap to build infill stations, and many, especially Tonelle Junction and Sunnyside, would provide a big bang for the buck.

Construction Priorities

The complete system would be lengthy and expensive to construct, which means it would have to be broken into phases. The following phasing would ensure the highest-priority items are built first.

  1. Four-tracking the North River Tunnels, ideally with the Grand Central connection. This would complete the Northeast Corridor, Hudson-Port Jefferson, and Morris and Essex-LIRR lines but for extending electrification.
  2. Tunneling from Staten Island to Grand Central; this would complete the Staten Island-Harlem line.
  3. Restoring service to the North Shore Branch and West Shore Line. This is a low-cost stage that can be done within the current stub-end framework of commuter rail; however, at this stage the West Shore Line should be electrified and connected to Hoboken.
  4. Electrifying the Erie lines, and building the Hoboken-Fulton-Flatbush tunnel. With a restored West Shore Line, the tunnel is a greater priority, but electrification would be necessary immediately afterward to make full use of the new connection. At this stage, electrification need only cover the nearest and densest suburbs—say, Ridgewood on the Main and Bergen County Lines, and Orangeburg on the West Shore Line, immediately north of where the two branches meet.
  5. Fully electrifying the Hudson-Port Jefferson-Raritan system. This phase can be done immediately after phase 1 if there is money for electrification but not tunneling.
  6. Electrifying the Danbury Branch and the full North Jersey Coast Line. Ocean County is among the fastest growing in the region, and this phase may need to come earlier if the overall construction schedule lags.
  7. Double-tracking the LIRR Main Line from Farmingdale to Ronkonkoma.
  8. Extending service and electrification elsewhere; the greatest demand would likely be on the Montauk Line, which would eventually be electrified to Speonk, followed by the Montclair-Boonton Line, the West Shore Line as far as Stony Point, and perhaps the Pascack Valley Line. Remaining lines on the map are marginal, and lines not on the map would likely never have enough demand to justify electrification.

70 replies on “Regional Rail for New York City – Part II”

I’ll be honest for a moment. Sunnyside from the MNRR New Haven line would be a HUGE personal need. Add to that the idea of connecting service to Jamaica and the airtrain and I’d get back to taking the train to the airport.

But what is the real likelihood that East Side Access will allow MNRR into Penn instead of just increasing service on the LIRR? And then to have the brains to build a station there.

I think the ideas are brilliant. I really do. Which is why I’m so skeptical

I agree with Daniel M Laenker about Staten Island. A subway connection from Brooklyn off the R line and/or Triboro Rx would make more sense for providing transit to Staten Island.

IMO, “Regional Rail” to SI would only make sense if two conditions are met:
1. The SI route doesn’t end in SI, but continues past Port Ivory or Tottenville over to NJ, providing more commuter capacity across the Hudson.
2. Express tracks can be built on the SIR to make it so through trains don’t take forever to make their way through SI.

The idea of through-routing commuter trains is definitely a no brainer, but it’s not as necessary in NY as it is in Paris or London or Tokyo, since the NYC Subway already has multi-track express lines, which play part of the role of the RER in Paris, “Overground” in London, and JR in Tokyo. I think that all the subways in those cities are local-only.

I’m not too sure on the specifics of its implementation, but your plan makes sense for the most part. My one complaint is that you left the ESA Grand Central Terminal as a stub, in spite of the fact that it has tail tracks which will make it very easy to extend. IMO, THAT’S what should be extended downtown, and on to Staten Island

I wish I had time to go through all of these. Most of them seem like a good idea, however, there are a few qualms, most notably with Staten Island, which I will discuss first. A tunnel under the harbor is a bit much. It would be better to either extend the PATH south through Bayonne to SI (and thus be able to serve central SI as well), or to extend the existing SIR branch west from Tottenville to Perth Amboy, where one can transfer to a NJT train (which would be a big plus to commuters from southern SI, who would still be somewhat screwed by this system). The North Shore branch should also definitely be done (as heavy rail) and also extend west to NJ to serve Elizabeth and Roselle.

Also, remember LIRR and Metro-North use different third rails; the LIRR uses 750V top contact and Metro-North uses 700V bottom contact, so this would be somewhat difficult to do, since you’d need multiple types of contact shoes on each car. Plus, the post does not take into account phase gaps in the catenary, which exist in several places:

1. At New Rochelle for trains using Hell Gate
2. At Kearney for trains going onto the Lackawanna lines from the NEC
3. Somewhere on the NJC

The New Haven MUs can’t run through phase gaps, and neither can NJT’s Arrow series of MUs. Though if the Arrows were equipped with third rail shoes (LIRR), they could through run from the Lackawanna lines to the Atlantic Branch via your Fulton tunnel, or from future electrified Erie lines, assuming they’d be 25 kV, 60Hz. And locos of any kind are unacceptable for commuter service; MUs are far superior.

Also, this map left out the West Trenton line, which would be good for not just commuter rail service to southern Hunterdon and northern Mercer counties, but also, built as a four track line, would be good for relieving intercity congestion between NY and Philadelphia, and two of the tracks would be reserved for high speed rolling stock, which would allow the Acelas to run at 200 mph between Somerville and West Trenton (from West Trenton to Philly the line is electrified for commuter rail service).

Orulz, the subways in London and Tokyo are local only, but the local lines have interstations of about 1.2 km, compared with 700 meters in New York and 500 in Paris. The regional lines have both local and express tracks, at least in Tokyo.

The choice of which of the two Grand Centrals to connect with Staten Island was difficult. I decided to go with the subsurface station because a) the Harlem Line is the best ridership match for what the SIR would become if it went into Manhattan, and b) it would provide Metro-North with service to Lower Manhattan, which the LIRR would already have via Brooklyn. Hudson and New Haven Line commuters would still have to transfer, but they could make a cross-platform transfer at Harlem-125th, instead of go from subsurface Grand Central to deep-level Grand Central.

Also, the SIR already has a mixture of local and express service.

Adam, the main contention of this series is that capital construction dollars should be spent on systemwide connectivity, rather than new lines. A new West Trenton line may make sense if the NEC and Raritan Valley line get too clogged, though. But I’m skeptical that it’ll be a better ROW for the Acela than the existing NEC, which is almost perfectly straight through New Jersey and only needs constant-tension catenary and a Metuchen bypass to permit 200 mph operation.

You’re right about the westward connections from SI, though. Those will be especially useful if Woodbridge ever gets commercial TOD…

Having now fully read the second article (hadnt when first comment was made) I can agree with the above about SI. I just don’t see that as a priority. An eventual connection, of course. I do like the idea of running GCT south to Fulton to start.

I also think the country as a whole should realize that overhead catenary is the way to go and we should start to phase out the other modes or at least not allow further construction of them.

To run new catenary around the system would allow those trains to run through while not tearing up the existing 3rd rails would allow those locals to continue to run. But eventually everything should end up with 1 electrical system. Not three.

Electrifying, and a common electrification would be the priority for NYC and the entire country.

Speaking of connectivity and Amtrak, what’s it going to take to get Amtrak to add Secaucus to its NEC routes? A stop there would make many towns in New Jersey a two-seat ride from any other NEC station, but even the slowest regional-service Amtrak trains blow through Secaucus while making stops at less useful stations (New Carrollton comes to mind).

Chris, SI doesn’t look like a priority because right now Staten Islanders usually end up driving. The transit modal share from SI to Manhattan is probably sub-50% (the ferry’s ridership is 60%, but it includes tourists). Get it up to 100%, which is where it’ll be if there’s a direct rail tunnel, and you’ll get a line that competes with the Connecticut and New Jersey NEC lines for highest commuter traffic in the nation.

It’s expensive, but not unaffordable – the cost estimates for a two-track Brooklyn-Jersey tunnel, which should cost about the same, are on a par with those of ARC. And it’ll drastically shorten Staten Islanders’ commute, making them more supportive of rail improvements in general. Right now Borough President Molinari complains every time they hike bridge tolls to pay for the subway.

I don’t think it’s the second highest transit priority – that would be Phase 2 of SAS – but it should be the second highest regional rail priority.

As for Metro-North to Penn, it’s actually officially planned to go online as soon as ESA goes online. With through-routing it can go up even sooner, since right now the problem is track capacity at Penn Station itself, rather than rolling stock or East River tunnel capacity.

I’m with just about everybody here, great plan but give Staten Island a conncetion to NJ and Brooklyn. I can see that you’ve included the SI harbour tunnel with the line to Lower Manhatten, and that just doesn’t work for me. While that is needed yesterday, lets focus on the North Shore line, Brooklyn, Jersey and maybe some faster pedestrian ferries for Staten Island (hmm, hydrofoils in the harbour probably aren’t a great idea, but would solve the speed if not the transfer problem).

All in all, I’d rather see the funds go to TRX than a harbour tunnel for now, and when regional service to SI is justified (or more to the point the rest of the system is complete) wouldn’t a line parralelling (more of less, presumably tunnelled anyway, so deviations for actual destinations make sense) I-278 be almost as fast and get more riders along the way?

I’m thinking of something along the lines of a loop through Lower Manhatten, along the Brooklyn waterfront, across the Verrazano Narrows, the top of Staten Island (linking up to an extended Bayone HBLRT line), across to Jersey at Elizabeth and either terminating at EWR or returning to Manhatten.

Actually, Avon, unless the center express tracks on the NEC were restricted for Acela’s use, the Acela could only travel at a 150 MPH top speed, not 200 MPH (because of the ever-anti-progress Federal Railroad Administration), even with constant tension catenary. And using the two center tracks would be very dangerous because then local service would have to be cut out if anything happened. Ideally the high speed tracks would be on one side and the local tracks on the other. A post about this was made on a blog dedicated to the section of the California HSR project that runs along the Caltrain ROW, which would take a form very similar to what we have on the NEC.

Frankly, I think an extension of the PATH NWK-WTC line south to Newark Airport and Elizabeth which would parallel the NEC on one side and make intermediate stops would be appropriate, then it would turn east onto an existing ROW at the Elizabeth station and then south onto another existing ROW paralleling Arthur Kill (with a stop at the Elizabeth IKEA :D), and then turn onto the drawbridge across Arthur Kill next to the Goethals Bridge, and then would go east along the SI North Shore line to St. George’s. Sounds confusing, yes, but here you have a rail line that would be heavily utilized in both directions (commuters can get to the ferry, Elizabeth, the airport, and downtown Newark).

I like this plan, however..

I agree with the comments regarding the plans for Staten Island. With just 500,000 residents Staten Island will never challenge Westchester, Long Island, Connecticut or the NJ suburbs for transit ridership into Manhattan. But that doesn’t mean a line should not be constructed because that line can be sent beyond SI to bring relief to NJ’s crowded network. You may as well send it up through Brooklyn and connect it to both the North Jersey Coast Line, NEC, and RVL on the southern end if we’re going to spend that kind of money. That way we get a redundant bypass of Newark and the crowded High Line, and maybe knock a few minutes off the trip for anyone south of Metuchen on the NEC. A station in Bay Ridge or Red Hook would serve to reduce crowding on the Brooklyn subway lines and reduce dependence on express buses in that area.

IMHO the political and technical barriers to combining NJT, Metro North, and Long Island Railroad service into one regional system are so great it will never happen. As a NJ resident and regular commuter on the Northeast Corridor I’m not certain I want NJT directly folded into the MTA’s system, with all the potential problems that could create. Incidentally from what I’ve heard from NJT officials the NYDEP is requiring a sign-off on any major work which takes place within a few hundred feet of the first water tunnel. So even a shallow tunnel between NYP and GCT (with all the geotechnical problems that creates) would be unlikely to recieve a go-ahead before 2030 when a deep tunnel would be possible.

That having been said, I do not think we need to condense the networks down to a single system in order to achieve all the benefits a through-running arrangement would provide. If we’re going to undertake a major tunnel running north-south between Lower Manhattan and Midtown then we could easily combine the two new tunnels into Lower Manhattan with that north south tunnel and then connect them to ESA and ARC at the northern end. The junctions at either end of the north south tunnel would provided the LIRR and NJT with their own interconnected loop which would allow extremely high capacity without trying to bite off the political and technical challenges to integrating the networks. Finally by arranging the platforms within the system cross platform transfers could be created to facilitate trips from NJT trains to the ESA terminal at GCT, or from Long Island RR trains from Brooklyn into Jersey City. Metro North would unfortunately be left out of this first stage other than what can be sent into NYP or perhaps ESA via the Hell Gate Bridge or Empire Connection, but a (relatively inconvenient) transfer would be available at GCT for Lower Manhattan connections.

I’m afraid my is quite a bit more basic than yours, but the operational plan would remain quite the same. Thankfully the LIRR network would largely work with the loop thanks to Jamaica. The only line which would create a problem would be the Port Washington Line, and then maybe that could operate on a full time basis to Jersey City. On the NJ side the network could be tied end to end. The Port Jervis/Harriman, Spring Valley and the Morristown and Erie lines could be tied to the NJCL, NEC, RVL, and Midtown Direct lines, with the Waterfront Connector and Kearny Jct allowing some trains to loop on the same line rather than continue on to the north. This is of particular use with the NEC, which sees far more traffic than either of the former Erie or Lackawanna lines.

I realize it’s not quite as ambitious, but it’s the sort of thing that is the result of starting with the same plan you proposed. I like to think that if I can get the ear of some folks it can get built without too many externalities involved in rationalizing operations, power supply systems, and every other element.

re: SI:

First, in terms of the number of commuters to Manhattan, Staten Island is at two thirds of Westchester and one half of Long Island, and growing much faster – and it’ll grow even faster if there’s a quick connection.

Second, yes, it’s possible to connect SI to Manhattan via Brooklyn or Jersey City. But that would turn a 10-minute trip into a 25-minute trip, reducing the time advantage of rail over cars. It would certainly not be any faster than the NEC, making it useless for people traveling from New Jersey to Manhattan. Detouring through Newark would be especially bad – just Newark to WTC is 22 minutes; getting to St. George from there would likely take another 20-30.

Third, Newark Penn is not a bottleneck; it needs no bypass. The focus on redundancy is misguided – commuter rail is never redundant, especially when it involves something as expensive as underwater tunnels.

re: ARC:

The problem with the deep-level tunnels is not that they need authorization to be extended. It’s that they can’t be extended, at all, for the next 10 years – and later on, they can’t meet the East River Tunnels.

A circular line just won’t work. There are too many river crossings for such a line to work as an effective way of transferring from one suburban line to another. The region will have to make do with Triboro in the east and the Hudson-Bergen Light Rail in the west. Round-robin loop lines are a bad idea, too – they make transfers a living hell, again; the BMT tried and abandoned a similar routing involving the subways serving the Williamsburg and Manhattan Bridges and the Montague Street Tunnel.

Connecting SI to the NWK-WTC PATH isn’t to bring people to the city. It’s more so Staten Islanders can have convenient airport access. Extending it to the existing line is only to close a gap. The best way feasible way into the city involves a PATH line going under Kill van Kull. A tunnel under the harbor is just plain not feasible (has anyone ever tried to cross a body of water longways?)

Also, if the BDFV and NQRW weren’t in the way, I would extend the PATH to Grand Central via the (useless) 42nd Street Shuttle (I believe both the IRT and H&M were both built at the same gauge), or maybe even connect it to the 7 subway so there can be easy access across the river for those that live in the immediate suburbs and outer boroughs can get to each other (I know plenty of people from Brooklyn/Queens and Jersey City/Hoboken that have to get to the other side of Manhattan), though then again, that could only work with one of the PATH lines because then you’d have 3 lines on the same track which is unacceptable.

I am not sure the Staten Island tunnel is worth the cost, but the vision makes a lot of sense. I definitely agree with orulz’s post that a harbor tunnel should be connected to the NJCL and the NEC in order to justify the expense.

Looking at the map, you are effectively proposing two new tunnels in lower Manhattan south of the village. These could possibly be combined. I believe the Harlem and New Haven lines can get access to the Hudson Line. If not, they could probably connect it cheaper than the cost of building a new tunnel from GC to Fulton.

The extension of the metro north lines could proceed down the west side highway (maybe even as an elevated or trenched line) before going back underground to extend to Brooklyn. It wouldn’t be as useful as the proposal shown in the maps above, but it would be very effective at providing service downtown at much less cost. The proposal shown above would require tunneling at some point under existing subways and then crossing over some blocks to get to the west side of downtown.

The location of the Sunnyside station shown in the map is preferable to the location the MTA is building it. The MTA station is effectively underneath the 7 train, where it crosses the Sunnyside yards. It would not be able to serve trains headed to or from GC, which is a shame. It is designed to replace the LIC and Hunterspoint stations.

All in all, it’s great seeing this proposal. Very exciting. Great work!!

In Jersey, let’s not forget the “Lackawanna Bypass”! Please adjust the diagrams to reflect the extension of NJT from Lake Hopatcong to Andover, Blairstown, East Stroudsburg, &c. to Scranton.

Alex, thanks for the encouraging words.

The Empire Connection naturally connects to Penn Station and the High Line, but nothing else. You could extend it in a new subway downtown, but then you might as well go under 10th Avenue.

The way the current plan works is that the Harlem Line goes under Madison to Union Square, then dives under the Broadway Line to get to Fulton. The only lines it really crosses are the L and A/C, the same as Second Avenue Subway.

Adam, the Seikan Tunnel’s underwater portion is three times as long as this of the SI-Manhattan tunnel I’m proposing. Closer to home, the proposed Jersey City-Brooklyn tunnel would be only slightly shorter, and would not have any length issues even though it would be used by diesel trains. The cost of a straight SI-Manhattan tunnel would be higher than of a PATH extension, but it would be sure to capture a near-100% market share.

The idea of through-routing commuter trains is definitely a no brainer, but it’s not as necessary in NY as it is in Paris or London or Tokyo, since the NYC Subway already has multi-track express lines, which play part of the role of the RER in Paris, “Overground” in London, and JR in Tokyo. I think that all the subways in those cities are local-only.

Actually, there are three subway lines in Tokyo that have express service (not counting the private suburban commuter lines that interconnect with said lines). The Tozai Line has an express segment from Toyocho east of Tokyo to its terminus in Nishi-Funabashi (Chiba); the Shinjuku Line has express operation along its entire line from Shinjuku to Moto-Yawata (Chiba); and the Fukutoshin Line has express operation along its entire line from Shibuya to Wako-Shi (Saitama). The express trains tend to be noticeably more crowded than the local trains, especially during peak periods.

I agree that a tunnel under New York Harbor to Staten Island might be overkill, unless you connect it to the NE Corridor to serve as a bypass route. Even then, you would face a great deal of opposition from Staten Island residents who prefer their suburban lifestyle in spite of being a New York borough, unless the island’s density radically increases over the next 40-50 years.

If SI residents oppose the idea, it’s their right. I for one think most of them would like a train that would halve their commute, but I may be wrong. New York is full of neighborhood groups that try to dirty their area and avoid infrastructure upgrades in order to make it undesirable for outsiders.

You could extend the Hudson line down the west side highway, and save money by elevating it or putting it in a trench between uptown and downtown lanes, possibly even removing a lane or two to make room. If you connect it to the line you propose coming in from Hoboken, you’ve saved yourself about a mile and a half on new tunnel construction. Similarly, if the extension of the Brooklyn line followed Atlantic Ave to the South Ferry, and met up with the line from Staten Island there, you could save yourself a half a mile of tunneling in downtown Manhattan, even more of a nightmare than midtown. Together, this would save at least a billion dollars.

I assume a tunnel from Grand Central to Fulton St would be very deep, so to a degree, passing under numerous subways would not be that big of a deal. However, finding the land to open the tunnel near Grand Central, dealing with the maze of utilities in the heart of Manhattan, etc. would be a nightmare. I’d think building at the west side highway would not only save on overall tunnel length, it would make for cheaper construction.

You’d have to forgo a new station at Union Square, but then, you would save probably another billion dollars not building a very deep train station underneath a crowded web of stations and passageways. All the passageways and entrances are full at rush hour; it wouldn’t just be a new platform. You’d have to build a whole new circulation network with tons of new elevators, escalators, pedestrian tunnels, etc. It would take a decade just for that station alone. If 14th st is a major goal, it might be cheaper and definitely easier to build a new regional station on the west side highway and extend the L to it. 14th st can be closed off completely and a new transfer station built there in 3-4 years without disturbing one of the most crowded parts of the city.

I did some fast and crappy photoshopping to explain what I mean.

A quibble:
“Amtrak and the MTA should work to rebuild the Empire Connection along the West Side of Manhattan with third rail electrification and double-tracking, enabling electric Hudson Line trains to use Penn Station. Extra stops at 62nd, 125th, and Dyckman would also provide service …”

I’d like to see this happen, with transit-type schedules and new stations at 155th and 181st as well. They might need funiculars, or banks of escalators, to connect to the adjoining heights. But those areas are very densely populated and on the cusp of gentrification, which could add more intracity commuters..

But a station at 62nd is probably impossible now. The tracks through the old rail yards were put inside concrete tunnels, then dirt, streets, and high-rise apartments put atop them. From my eye-balling, the tunnels will only allow two tracks, total. A station might be possible around 73rd, or about 57th, but the Trump development precludes one at 62nd.

A larger problem is the low bridge over the Spuyten Duyvil, the waterway between Upper Manhattan and the Bronx. Doesn’t it have to open regularly for various passing boats? A much higher bridge or a fairly deep tunnel would add engineering challenges and billions to the budget.


Alon —

Nevermind the quibbles, this is great work on the pig picture, and a challenge to us and our politicos to think BIG for our region’s future.

Alex, the routing of the Flatbush-Fulton tunnel is a tradeoff between total tunnel length and underwater tunnel length and depth. Connecting from Brooklyn to South Ferry will involve a longer underwater segment with a deeper profile, which will raise costs, even as it saves on some tunneling in Manhattan.

The routing of GCT-Fulton is somewhat simpler. The construction north of 14th can be done subsurface under Madison. The only east-west obstruction is the Penn Station access tunnels, which have been successfully crossed by four subway lines, with a fifth in planning. From Union Square south the construction needs to be somewhat deeper, say 50 feet underground, the same depth as the 7.

Woody, yes, the single-track bridge over the Spuyten Duyvil is a huge pain. A higher bridge may not cost too much as the approaches would just be over existing rail lines. On the other hand, it’s not critical – it may just be twinned by another similar bridge, which would open for boats but provide the line with two tracks throughout. That should be enough for about 8-10 tph per direction, which would be enough to satisfy ridership demand.

And I thought the Trump development included preservation of the track area in order to permit a station at 62nd in the future.

A station at 62nd, well, it could be. The new apartment towers have been constructed from 72nd down to about 64th or so now. Maybe they will save room under the last few still to go up. I stumbled because you said “double-tracking” the Empire line. I took that to mean four tracks, because most of it under Riverside Park and under the Trumped-up apartments is already two tracks.

I’d have loved to have seen the 2/3 line extended west under 104th to connect with this ROW and head down the West Side to Penn Station or beyond. But the Trump project does preclude that possibility.

Anyway, I was thinking that a higher bridge at Spuyten Duyvil would require long, long approaches to get the trains up from almost sea level to a height to clear all maritime traffic. Isn’t that a huge problem for the Acelas to Boston? Every river, bay, and yacht basin along the Connecticut shore has maritime traffic. So, drawbridges and limited frequency. Or build new high bridges with long, long approaches to keep the grades from being too steep. Yeah, I can see twinning the low bridge to keep within budget.

Actually, rolling stock alterations *would* be needed.

The LIRR has overrunning third rail, Metro-North has underrunning third rail. :-P It’s worth mentioning this.

Disagree on phasing. Step 1 should be through-running at least part of NJT with either Metro-North or the LIRR, to work out the institutional, political, and technical kinks and get a situation where people are *willing* to think in terms of through-running.

Alon Levy said: “A circular line just won’t work. There are too many river crossings for such a line to work as an effective way of transferring from one suburban line to another.”

I’m sorry, this comment completely confuses me. How are there any more river crossings in a system wherein NJT and the LIRR use a north-south tunnel connected ARC and ESA to a pair of tunnels under the Hudson and East Rivers than in a plan which attempts to join Hoboken and Flatbush Avenue terminals via Lower Manhattan? I count 4 tunnels with a total of 8 tracks in both plans, with both alternatives handling the same number of peak and off peak direction trains through each river crossing.

I hate to derail this discussion, but I fear some clarification is required, and I believe the diagram I attached to my name could be quite illustrative. I am not proposing a tunnel which circles around downtown NY or Manhattan. I am proposing one tunnel which links Midtown with new tunnels under the Hudson and East Rivers. The NJT and LIRR systems would thus become two interlocked loops, resulting in train capacities equal to through-running arrangements without the political, technical, or operational problems that are bound to doom any effort to integrate the three commuter railroads. I fail to see how this results in transfers being a “living hell” when every station is designed to facilitate transfers on a cross platform basis.

Finally, I have it on extremely good authority that there was virtually zero chance of the NYDEP granting approval to the construction of a new tunnel from Tracks 1-4 before they can shut down the first water tunnel for buttressing in 2025 at the earliest. That, Metro North’s refusal to allow the LIRR into their terminal via the ESA shallow options, and their hostility to NJT ruining their 3 Peak, 1 off-peak direction track arrangement on the Park Avenue line with through-running empty trains effectively killed ARC Alternative G. One way or another NJT will have to wait for the water tunnel project to be completed before we can even think about anything being discussed here. And even then we will not stand a snowball’s chance in hell of getting NJT and the LIRR together, let alone getting Metro North any further than some peak hour service into NYP. The two loop proposal is not perfect, I’ll admit that, but it depends on far fewer factors than any through-running agreement will rest on.

Alon, re SI:

I don’t know what the modal share of SI-Manhattan commutes is precisely, but driving does not take up anywhere near half of it. I’d guess it’s probably about 40-50% ferry, and about 30-40% express bus. I’ve made this point on the comments here before (I think regarding buses using the Brooklyn Battery Tunnel) – there are over a dozen express bus routes connecting all parts of Staten Island to Manhattan. I grew on Staten Island and I used to take them all the time. I know plenty of people who did, or do, the SI-Manhattan commute, and I’ve never met anybody who drives. Sure, it’s anecdotal evidence, but still.

Considering this fact, even though I’d love to see my home borough reap the benefits of a cross-tunnel rail link, I’m not really sure it is worth the expenditure. If the rail tunnel were taking modal share from cars, as it seems you had thought, then it may well be worth the cost. But it will mostly be taking commuters from the ferry and the express bus and putting them on the train – a great thing for the commuters, no doubt, but probably not worth it for the city/state. My favorite SI rail solution is to put the R train over the V-N bridge and make it a skip-stop express, like the ‘diamond’ version of the 7. This would be a more efficient link than NJT Light Rail, which is slower and necessitates a transfer to the PATH. Of course, it would also be worth it to have both. And the North Shore Railroad needs to come back.

Regarding community opposition to a rail link, I think that would’ve been a real problem 20 years ago, but not anymore. Three reasons: a) SIers are sick of having the nation’s longest commutes and endless congestion in the borough, b) the population is more diverse and less provincial, and c) rail transit does not carry the same negative connotations for suburban whites that it did 20 years ago. I’ve been to community planning meetings in SI, and, almost universally, the biggest priority for people is improving public transit options, particularly rail. The borough’s population density, BTW, is about 8,000-9,000/sq mi., certainly enough to support some more rail links.

Wdobner, your loop proposal will increase capacity, but not by enough. Loops of the type you propose have a higher capacity than a stub-end terminal, given regional rail dwells, but a lower capacity than through-routing. They also make it difficult to live in a county on one side of Manhattan and commute by rail to work in a county on the other side. Yes, you could transfer, but you’d have to either make a detour to Union Square or transfer at Penn Station, neither of which is pleasant; there would likely be no coordination of schedules or one-ticket rides as on the LIRR with the Jamaica transfer.

Metro-North’s 3/1 operation is not scripture. It works for a railroad that needs to move 50 tph into Grand Central, but Metro-North can safely shunt 15 of those into Penn Station, 10 from Connecticut and 5 from the Hudson Line. This contrasts with a normal limit of 60 tph for a four-track railroad signaled as rapid transit (or 80 tph if you’re in Moscow). The insistence of keeping things as they are is entirely political and cultural.

Vin, what’s good for commuters is good for the region, in many respects. One of the limiting factors to SI’s growth is bad transportation alternatives. Remove that, and you can pack another 500,000 people there, which means 500,000 fewer people commuting to New York from the Poconos and Catskills. It also helps attract middle- and high-income people to New York City, which means more tax base for city schools, without displacing lower-income people.

Extending the R to SI is a good idea as well, but it’ll help SI-Brooklyn commuters more than SI-Manhattan commuters. Even if you ran the R express on Fourth Avenue, it would take too long to get to Manhattan.

I know very little about New York and have never been to the city, however I have seen an image on google earth of Giants staduim. I imagine it is impossible to get public transport to work to this stadium around the massive car parking facility. Does this new rail plan incorporate this staduim?

… however I have seen an image on google earth of Giants staduim. I imagine it is impossible to get public transport to work to this stadium around the massive car parking facility. Does this new rail plan incorporate this staduim

Service to the Meadowlands train station starts this weekend.

Boris, thanks for the link. It had some good info. And a bit of wrong info, you may have noticed. No way anyone will go from that station to Grand Central. I don’t think it could happen if all of Alon’s wonderful tunnels got built!

The Metro North trains would head into Penn Station after this stop. From there plenty of easy subway and bus connections. Otherwise, the stop itself is currently off the beaten track, so to speak, though I’m sure a couple of bus lines could be rerouted to serve it, probably the Crosstown lines on 65th and 57th Sts.

The area is certainly becoming more dense with high-rises in every direction. But as for getting a contribution from the developer to build the station, I’d bet that train has left the station. I rode past the development on my bike again yesterday, and from 72nd St going south there are already 10 buildings, one still under construction, leaving only three empty blocks to go. Most of the completed buildings are condos, so the developer doesn’t give a damn if they benefit from a new station or not.

And I hope that Cpuncilmember Gale Brewer is working with Metro North to get a few more stations in the pipeline, like one at 86th, and at 125th for sure.

Why electrify the Montauk Branch only to Speonk? Why Speonk? Nobody needs to go to Speonk. Double-tracking and electrifying the branch out to East Hampton would yes be more expensive, but people actually go to the Hamptons…

How much would commuter rail from Grand Central to Staten Island cost? If rail access to Staten Island is important, perhaps extending the 1 train should be considered. That brand new South Ferry terminal is pointing at Brooklyn. Three stops in Brooklyn, and then a tunnel to Staten Island should be achievable at a much lower cost then a MetroNorth option.

The cost of a tunnel from Manhattan to SI is independent of what’s done north of Lower Manhattan.

A connection to the 1 is a problem because it’d make the 1 the longest route in the system, even longer than the A to Far Rockaway, which would invite delays. It would also not match capacity well. At current commuting pattern, the highest reasonable demand for an SI-Manhattan tunnel would be 50,000 boardings per day at SI stations. This is on a par with Harlem Line traffic, making those a good match. The 1 gets the same number of boardings at its stations from 181st Street north only – for the entire line, demand is so much higher that either trains would run two thirds empty in SI, or they’d be packed to the gills in Manhattan.

Two points about load balancing. The first is that a new St. George station would have to be built, and that could be constructed to allow some trains to turn there. The second is that existing demand should not be used in place of projected demand. Staten Island continues to grow. A new rail option that saves time will attract additional riders from other transportation modes. New service will spur on development along this rail corridor. Staten Island will grow into the capacity that is offered by subway service.

As far as the length of the route goes, that is a legitamite concern. It can be mitigated by rebuilding a station, perhaps 125th St, to short turn trains.

I believe that the concept of extending MetroNorth to Staten Island has a lot of merit, but it also has extremely high costs and obstacles.

There is the cost of tunneling from (more likely under) Grand Central to Lower Manhattan. It would easily cost several billion to get to Fulton St.

Let us not forget Water Tunnel #3. The City of New York will not allow any subterranean construction in the immediate vicinity of water tunnel #3 until it is up and running. NJ Transit wants to extend ARC past 6th Avenue. The City has told NJT that they have to wait until the water tunnel is done. It is scheduled to be completed in 2020. Does anyone want to plan on a major infrastructure project in NYC being completed on time before another project can be started?

The existing capacity issue of MetroNorth into Grand Central should not be ignored. There is no additional capacity. MetroNorth might have to construct two more tracks, under the four Park Avenue tracks, starting in the Bronx. At a much lower level, about where the LIRR’s ESA is, MetroNorth could then continue to Fulton St and beyond.

When Water Tunnel #3 is completed, and if the political will to persue extending MetroNorth south of Grand Central is mustered, and if other transportation projects are not deemed as important, the MTA will then have to get through community board reviews. I can see the real estate families of New York killing a major project that disrupts the economic activity of their Midtown assets, without providing any direct benefit. Now compare all of this with load balancing and route length issues, and those seem rather trivial by comparison.

Water Tunnel 3 is on time – it completion date has been set for 2020 for decades now. It’s not like SAS, whose completion time is pushed back three years every year. Even then, the city is being too irrational; construction south of GCT would run under either Madison or Park, both of which are well to the east of Water Tunnel 1, which is under Fifth.

Metro-North wouldn’t be at capacity if it could run trains into Penn Station. The entire point of through-running is that it allows using current capacity more efficiently. Penn Station has capacity issues to the west, but not to the east; with through-running on the NEC, it could be used to relieve Park Avenue’s capacity issues, which would allow slack for new capital construction.

The cost of a GCT-Fulton tunnel depends on how well construction is managed. At typical expensive-city-with-a-lot-of-legacy-infrastructure cost, it should be $500 million per route-km, i.e. $3 billion. At SAS cost, it should be $1.7 billion per route-km, i.e. $10 billion. For reference, a tunnel from Lower Manhattan to SI should cost about $7 billion, on a par with the projected cost for a two-track freight tunnel from Brooklyn to Jersey City.

Just out of curiousity Alon, how do you convert Grand Central to through routing service? I admit I am not familiar with all of the track layouts in Grand Central, but the existing main level and lower levels have dead end tracks. Do you connect Grand Central going south (to SI and Penn) through the new ESA station (175 feet below street level) or do you add another lower level in between the current dining/lower level and the ESA level?

I really like your idea, but I see this as a stumbling block, because I don’t think connecting GCT to this system with the ESA platforms being a good idea, nor do I like the concept of destroing one of the levels in GCT (which is would face serious legal challenges thanks to GCTs landmark status) to connect it to a large system. Obviously this problem exists at the other stub terminals that you mentioned should be converted to through stations, but none of them (with the possible exception of Hoboken terminal) have the architectural beauty of Grand Central.

I think the MTA can just take four of GCT’s lower level tracks and extend them. For example, it can take them out of service, build a trench through them, and continue mining south until it hits 42nd, where it can lower a TBM. It could even continue with the ESA TBM south to Lower Manhattan, but that would best be done with a connection between ESA and the existing station.

Alternatively, the building itself has landmark status, but the food court doesn’t. There’s nothing wrong with saying, “We don’t need 67 station tracks,” moving the food court to where the upper level tracks are, and extending the lower level tracks straight ahead.


Great work, I wish that Mayor Bloomberg and Governor Corzine had a copy of your plan! There’s been some very preliminary talk of building a tunnel between Oyster Bay on Long Island and Rye in Westchester. I wonder if it would make sense to have a branch of the New Haven Line split off at Rye, go through the tunnel, and continue on to Manhattan via the Oyster Bay Line (which you forgot to include on the map). A growing number of Long Islanders commute to Westchester and Connecticut to work (and vice versa)–in the unlikely event that this tunnel actually gets built, it might be useful to have train service between Long Island and Connecticut.


Do you think there is any merit to the idea of running a service that connects the New Haven Line with the LIRR Main Line, via the Amtrak connection that you have already purposed for “Red” service. The line might effective in providing East Bronx and Westchester (and possibly very close Connecticut) faster access to JFK airport. I don’t know if it warrants a full service, and I don’t know if it could even be a full service with Amtrak, half of your “Red” service and this service using the Hell’s Gate Bridge and NEC. But it might be an interesting idea, there is probably some demand for service from Long Island to north of the city, and it would mean not having to transfer in the city.

Anthony, Joe:

Either idea could work – I hadn’t thought about it.

The tunnel I’m not sure about – right now there isn’t a big LI-Westchester commuter market. It could work in the future, as a phase two for through-running from the LIRR Main Line to the NEC, if it gets popular.

The through-running on existing tracks should ideally diverge from the mainlines immediately after major transfer stations. On the NEC end, that’s Astoria, intermodal with Triboro and the N/W; it depends only on if the grades permit a station. On the LIRR end, it’s more difficult – it needs a new connection (but that’s peanuts compared to an SI-Fulton tunnel), and more service to Forest Hills. It also needs a local station as an anchor, preferably one close enough to Broadway/Roosevelt for a transfer.

It looks like New York City should build some type of beltway railroad line that would be 60 miles away from the city core to go around the city in a half ring that would allow people to take the train from suburb to another with out having to go though downtown New York City. It would take traffic off of Penn Station and Centeral Station.

There are some ways of building circumferential commuter rail. The problem is that when the radius is too large, which 60 miles definitely is, ridership isn’t very high. The advantage of commuter rail is capacity, not speed. You could get away with commuter rail serving heavily trafficked job centers, even suburban ones like Stamford and White Plains, but not for much else.

Just a thought: at the point at which you are constructing a tunnel to SI, why not include a stop at Governor’s Island? I know they’re trying to turn it into a bigger destination, I’ve seen posters advertising it as such for some time now. I’m sure adding a station there would significantly boost the number of visitor’s they receive.

Arc125: I actually thought about it… it may be a good idea, or not. It might even be cheaper because of ventilation issues, in which case there should obviously be a stop there.

I don’t think Governor’s Island warrants a stop in the SI-bound tunnel, especially considering the island is amazingly small, and the proposed tunnel would have to be really deep beneath the harbor. Maybe if Governor’s became more than a difficult to access public space, and perhaps the next Roosevelt Island (not that I support such a plan), a station could be justified easier.

If a station is more logical than it seems to me, the design should allow most traffic to skip the station, especially allowing peak direction trains to bypass the station easily. If you’re goal is to speed transit to Staten Island, including a staion to the park might not be beneficial towards that goal.

If we are looking to improve access to underutilized park space in this proposal, we should investigate a stop on the Hells Gate to provide access to the park space on Randall’s Island.

What about extensions of the subway into New Jersey? This kind of routing would work wonders in Chicago too, which has similar problems of stub terminals; Union Station only has one through track and two additional stub terminals in the loop (LaSalle Street and Randolph Street).

1. What could be easier than adding a station at 57th St. for Metro-North and one at Lexington Ave. for LIRR? That would provide a one-seat ride for many people who now have to transfer to a subway, eg. Westchester and Connecticut commuters whose destinations are north of Grand Central, and LIRR commuters whose destinations are on the east side.

2. Some kind of access to NJT trains would be great for people who could use the SEPTA R3 line to West Trenton and then continue to northern NJ and NYC via either the old connection from West Trenton to the NJ Central Raritan Valley line or by extending the Princeton Junction – Princeton Dinky to West Trenton. Now we Bucks County, PA people have to drive to Trenton or Hamilton to get our train to northern NJ and NYC.

Lionel Ruberg

1. There used to be a stop at 59th on what’s now called Metro-North. But the problem with it is that it makes Metro-North less express – it basically reduces it to the 4 with suburban service. It’s further away from most employment than Grand Central, so the one-seat ride factor isn’t large enough.

Lex on the LIRR is not feasible – there’s too much of a gradient.

2. Very long-term, you’re right. But in terms of construction priorities, it’s either West Trenton or Fulton Street, and Fulton Street’s more useful. Maybe a Trenton-West Trenton interurban would be more useful short-term – it would concentrate mainline train frequency instead of dissipating it.

“2. Very long-term, you’re right. But in terms of construction priorities, it’s either West Trenton or Fulton Street, and Fulton Street’s more useful. Maybe a Trenton-West Trenton interurban would be more useful short-term – it would concentrate mainline train frequency instead of dissipating it.”

RiverLine extension? Not a long one, either.

It’s occurred to me that I’ve never actually seen a detailed discussion of what would prevent the construction of new tracks directly beneath the existing Penn Station tracks. Can someone (Alon, maybe?) point me to where that’s all been hashed-out, long ago? For example, do the city’s water mains get in the way, or would it force the construction to take place even deeper than the current proposed 34th Street location, is it because the additional passenger load would congest Penn Station even more, or something else?

Oh, and by the way, is anyone planning any sort of observance of the 100th Anniversary of:

1. The completion of the original Penn Station (August?)

2. The beginning of LIRR service (Sept. 8th.)?

3. The beginning of PRR service (Nov. 26/27th)?


I haven’t seen anyone hash it out, no. But if I had to guess, I’d say that there’s a stability issue. It’s not easy to bore directly underneath an active rail line. Ideally, you’d want to build the line on top with the expectation that you’d build the future line under it; this was done on the IND 6th Avenue Line. If you don’t do that, you might as well do deep boring.

Nowadays NJT also wants the ARC project to be as far away from the original tunnel pair and the original station as possible. The reasons they state always change – sometimes it’s redundancy, sometimes it’s a fear that boring 300 feet from the old tunnels will destroy them, sometimes it’s a deep-underground column support that has historic status, and sometimes it’s real estate deals.

After staying in Jersey City during a vacation to New York, I found the PATH to be incredibly convenient and surprisingly well-used. I find Mr. Levy’s project amazing but the new tunnel to Brooklyn, Hoboken, and Staten Island are beyond unrealistic. Wouldn’t a minor change in FRA regulation regarding freight/transit mixing allow for numerous new jersey trains come downtown.

Major kudos for recommending through-running.

This is a fascinating and far-sighted plan. One suggestion would be to consider routing the Staten Island line through a tunnel under the Narrows to Brooklyn and from there north to Lower Manhattan. Not only would it serve riders in Brooklyn but the tunnel could also be used with the North Shore branch as the long-discussed freight connection to Long Island.

The key point in any regional rail service like this is a basic service pattern on all routes. Riders should be guaranteed that a train will appear every, say, 10 minutes all day long. This is the RER/S-Bahn model and it would massively increase the attractiveness of the service.

In terms of continuing the railway south of GCT, would it be possible to make use of the disused turnaround loop tracks? They connect the south end of the lower level tracks and could theoretically be modified to feed into a continuing southward line rather than looping back on themselves.

Thanks for the compliments. First, I hadn’t really thought of the idea of an SI-Brooklyn routing being used for freight. Thanks for the idea – it sounds interesting. My own assumption in developing this plan was that the few parts of Long Island that are still diesel-only would be electrified, and freight trains would run through Penn Station in the off hours. The LI-Brooklyn option might be better, though – it depends on how much passenger traffic there would be on the North Shore Branch to interfere with freight traffic.

Second, I don’t think it’s possible to use the loop tracks well. Besides being single-track and curvy, the loop only connects to the western platforms, which can’t connect to the northbound mainline track without creating a double-track bottleneck in the station throat. You can see details on Rich E. Green’s track maps.

Third, you’re completely right about the service plan. I discuss this in a coda to this regional rail piece.

North River tunnels are too small for today’s freight. Even off hours it would be moving right in the middle of the country’s busiest and third busiest passenger railroads. Not a good idea. The Narrows, while it looks good on paper, doesn’t have any rail on the Brooklyn side. The existing freight ROW points more or less directly at the freight ROW on the other side of the harbor. It’s out of the way of passenger service. Port Authority’s studies concluded that Jersey City to Brooklyn would be cheapest.

You then bump into the loading gauge of the East River tunnels. If I remember correctly they are going for double stacks. Also if I remember correctly they are aiming for fairly busy which is why the alternatives like using more car floats and the Hell’s Gate bridge isn’t an option. The point of the tunnel is to stabilize or marginally reduce truck traffic across Manhattan through 2030, 2040. You aren’t going to do that with three freights a night through Penn Station. …. New York City garbage alone will fill 200 cars a day…

Again if I remember correctly, it’s high enough from 65th Street to Maspeth where they want to put one of the intermodal terminals.

Alon, great post and love your ides. Just wondering if you could tell me where you got the ridership figures for the individual commuter lines. Thanks

Two sources. First, I looked at the ridership numbers on the existing branches that make up each line, and added them together. For example, see here for New Jersey Transit’s suburban stations; there are also equivalent numbers for Metro-North and the LIRR. For many lines, this is considered the sole source of ridership.

Second, there’s data on commute patterns as of 2000, on the level of both counties and towns – see here. I used this to get a rough estimate of expected ridership on lines that would see new service, such as the Staten Island Railway, or on lines where through-service would serve a large commuter market. That’s where my 750,000 users (=1,500,000 daily boardings) figure comes from.

No, two. The only four-tracking of existing lines that are currently less than four tracks, beyond passing segments, is the North River Tunnels and the LIRR Main Line to Hicksville.

This might sound a little bogus, but what if Path subway was converted to commuter rail. I understand that the rail cars and infrastructure are different, but converting the path into a commuter taillights be.cheaper than new tunnels. The path already connects Hoboken trm to lower Manhattan and midtown Manhattan. It could then be extended to Atlantic trm from the wtc, and to gct,from 33st station. The path seems to be purposely built as a rail system that connects stub end terminal. Why not complete the gaps and convert the infrastructure? It seems like it would save a lot of money. If all the connections were to built by using new tunnels, the path would become completely obsolete. Please let me know what you think, I love local rail systems!!

Invaluable analysis – I loved the details . Does someone know where my company would be able to locate a sample a form form to edit ?

Leave a Reply to AlexB Cancel reply