Home Penn Station A call for through-running to address Penn capacity

A call for through-running to address Penn capacity

by Benjamin Kabak

Last week, Manhattan Borough President Scott Stringer released an appropriately balanced call to develop a plan to relocate Madison Square Garden so that track capacity beneath Penn Station could eventually be increased. A close reading of the plan revealed it to be controversial in its suggestion that MSG be granted a ten-year reprieve before facing pressure to move but also more concerned with rail capacity than with the aesthetics of a Penn Station headhouse. It was, in other words, about planning for the future and not atoning for the decision to allow the original Beaux Arts building to face the wrecking ball fifty years ago.

Not everyone seemed to appreciate the nuances of Stringer’s report. Now, I’m not referring to those who disagreed with it; there was, after all, plenty with which one could disagree. Rather, I’m referring to writers such as The Post’s Bob McManus who published a screed against Stringer’s proposal on the grounds that it was too stuck in the past. Not actually paying attention to Stringer’s report, McManus said that we should all get over Penn Station and that Moynihan Station is not a well-thought-out plan.

McManus is not incorrect on either front, but Stringer wasn’t proposing rebuilding Penn Station or adopting Moynihan as the only solution. Rather, as part of the ULURP process, Stringer wants to see the city formulate a long-term plan for the Penn Station area that focuses first on rail expansion that involves moving MSG because the support pillars block such an expansion. It’s one plan, but it’s not the only plan. I wouldn’t expect an editorial writer in The Post to pick up on nuances or spend 20 minutes reading an 18-page double-spaced report though.

What about another plan? Over the weekend, Crain’s ran an editorial from Bob Previdi calling for a different solution for Penn Station’s capacity concerns. He wrote on through-running:

Finding the $15 billion or so for Amtrak’s Gateway project has proved difficult. But there is a way to spend roughly half as much while still doubling rush-hour train traffic. It involves taking a regional approach in how we use Penn Station.

Today, many Long Island Rail Road and NJ Transit trains terminate at Penn Station by discharging their passengers, loading more passengers and heading back in the direction they came from. This causes too many at-grade movements (where trains cut across each other’s path) within the station. It slows everything down, limiting service and inconveniencing pass-through passengers.

If trains simply kept going in the direction they came from—with NJ Transit trains continuing to Long Island and LIRR trains along New Jersey routes—we could streamline operations and expand the number of destinations each railroad serves. For example, NJ Transit customers could reach JFK airport, Citi Field and the U.S. Open tennis tournament, while Long Islanders would be able to reach Newark Liberty airport, MetLife Stadium and the Prudential Center.

Achieving this would require changing equipment and updating operating procedures, but this concept is not new or radical. It’s been done in the U.S. and around the globe, including in London and Paris. Heck, even Philadelphia did it: Way back in 1984, the old Reading Terminal and Suburban Station were combined into the Center City Commuter Tunnel.

This too is an approach worth embracing because it focuses on rail capacity and costs. Amtrak’s plans seem to focus on rail capacity as long as costs are no limit, and plans from folks bemoaning the lack of Penn Station believe funds should be spent only on restoring the past. “If we are concerned about what we can afford—and how we can leave some funds to fix the existing station—then New York’s and New Jersey’s elected officials should insist that the agencies find a way to make better use of the existing track, tunnels and yards that support Penn Station,” Previdi writes.

Maximizing dollars has been a theme of mine with regards to the MTA and Port Authority. We’ve seen capital projects become fiscal sinkholes as costs and construction delays spiral out of control. We’ve seen one agency spend billions on a headhouse with less regard for rail service and another sacrifice design and flexibility to keep costs under control. Maybe for Penn Station the initial answer lies in through-running. At the least, it’s worth a closer look.

You may also like


Berk32 April 3, 2013 - 12:35 am

now THIS is a idea I can get behind…
of course it will never happen – but one could dream

Nathanael April 3, 2013 - 6:09 pm

We’ve been talking about this for YEARS. Literally.

NJT can run through onto the New Haven Line today. Running on to Jamaica or Port Washington would require stringing catenary.

Running LIRR to NJ would require reestablishing the third rail, but it’s less valuable. The West Side Yard means that LIRR doesn’t actually perform as many reversing actions in the station itself.

AG April 4, 2013 - 1:12 am

yeah – isn’t that what they did a couple years ago by offering New Haven Line riders a one seat ride to the Meadowlands? what ended up happening with that?

Bolwerk April 4, 2013 - 1:38 am

It happened, but as I recall it’s a union clusterfuck and therefore probably not very suitable to daily service.

Alon Levy April 4, 2013 - 6:29 am

The West Side Yard complicates tunnel approach arrangements, though. It makes it harder for the LIRR to use the southern East River Tunnel pair, and as a result New York is spending $300 million on a flyover to let the LIRR keep using the northern tunnels without conflicting with Amtrak’s use of the southern tunnels.

It also makes it harder to quadruple the North River Tunnels by building the new tunnel pair to the north and connecting to the northern East River Tunnel pair; both the official ARC/Gateway plans and the proposed IRUM alternative locate the new Hudson tunnels to the south of the preexisting tunnels, but I believe this is due to the history of this project’s belonging to NJT, which uses the southernmost Penn tracks, and an alternative placing the new tunnels to the north should be investigated. For one it allows vanilla quadrupling, without a Grand Central connection or anything else, without also causing interlocking conflicts.

Henry April 3, 2013 - 12:49 am

I mean, at some point there will need to be a Gateway, because 4 tracks leading into 2 will always lead to that sort of bottlenecking, and Amtrak is already through running. I doubt a through running NJT would like to give up its slots.

Also, considering that MNRR and LIRR don’t even run compatible electrification systems (under-running vs over-running), somehow working NJT into that mess seems awful.

The use of Jamaica as a terminus for NJT trains should be considered, however. Perhaps they could use the ROW through Forest Park that they stopped running trains on this year.

Ryan April 3, 2013 - 8:06 am

Do you mean the Montauk Line? Last I heard, they were running two trains per day in each direction between Jamaica and LIC.

Phil April 3, 2013 - 10:11 am

One of those two trains was rerouted down the mainline last summer, and the other was removed from the schedule right around when Sandy hit.

The LIRR would have had to install PTC along the lower montauk if they kept using it for passenger trains. Because it isn’t cost effective for them to do so they transferred the line over to NYAR and is now only used for Freight.

Henry April 3, 2013 - 9:24 pm

So have NJT purchase the Lower Montauk, spend a couple hundred million to grade separate and install PTC, maybe a couple intermediate stations and a terminal at Jamaica, and voila – dual mode NJT trains can run to Jamaica Station, connecting New Jersey riders to JFK and Queens, with some trains turning at Sunnyside and some at Jamaica. Heck, since LIRR no longer uses the line, you could convert Montauk to NJT electrification and have a very frequent connection between JFK, Jamaica, and points west of the Hudson (EWR and Secaucus, for transfer to Hoboken trains and the like).

marv April 3, 2013 - 9:34 pm

Perhaps even have service on the old rockaway beach LIRR branch to the howard beach/jfk station.

Rob April 3, 2013 - 10:24 am

The electrical systems on two of the MetroNorth RR lines (Hudson and Harlem) are under-running, just like LIRR. It is only the MNRR New Haven Line that is over-running.

Maybe some thru-running could be accomplished with LIRR trains coming into Penn Station, and then continuing along the under-used Amtrak tracks along West Side of Manhatttan to the Bronx and Westchester.

The benefits of this partial and incremental approach could be:
(1) Clearing up a bit of rail capacity within Penn Station by thru-running;
(2) With additional stations on the Upper West Side and Harlem, allowing some LIRR and MNRR passengers to avoid Penn Station, lightening human platform crowding;
(3) Giving one-seat access to LIRR and MNRR points. Westchester Mets fans would be very happy.

Ryan April 3, 2013 - 11:30 am

Yeah, but the Empire Connection is not electrified for part of its length. Even the electrified parts have overhead catenary.

Hank April 3, 2013 - 2:56 pm

The Empire Connection is electrified with over-running third rail from Penn Station through the tunnel under the Javits center. It doesn’t need to be electrified the rest of the way.

Ryan April 3, 2013 - 3:54 pm

The segment north of 205th Street, to the junction with the Hudson Line, is not electrified. So, for the LIRR, only the C3s can run through the connection

Nyland8 April 3, 2013 - 4:07 pm

Uh … no. It loses electrification when it emerges from the tunnel – at roughly 123rd/124th Street. It runs diesel north of that.

Ryan April 3, 2013 - 8:16 pm

Exactly my point, only C3s can operate since they’re diesel.

Gorski April 3, 2013 - 11:44 am

Well, running Hudson line trains down the Empire Connection to Penn is definitely under consideration. Throughrunning would be awesome, but the third-rail scenario makes it a little more difficult. The New Haven line runs under catenary from a little south of Pelham through New Haven, but is under-running third-rail from that point south to GCT–ie, the contact shoe runs under the rail. The Harlem and Hudson lines are also under-running third-rail (which would otherwise cause problems on the shared track to Grand Central).

LIRR, on the other hand, is over-running, with the contact shoe on top of the rail, similar to the subway or the PATH. (That’s why LIRR uses the M7 cars and Metro-North uses the M7A–they aren’t interchangeable.) So the form of electrification used through the Empire Connection would need to be worked out.

Rob April 3, 2013 - 6:45 pm

ok. Thanks for the clarification.

Hoosac April 3, 2013 - 11:50 am

“The electrical systems on two of the MetroNorth RR lines (Hudson and Harlem) are under-running, just like LIRR. It is only the MNRR New Haven Line that is over-running.”

I may be mistaken, but I was under the impression that the Hudson and Harlem lines are under-running, whereas the LIRR is over-running. I believe that the New Haven is under-running in third-rail territory too, until it switches over to catenary. Used to be, anyway.

Ryan April 3, 2013 - 1:48 pm

Yep, the New Haven line is under-running south of Pelham.

Joe April 3, 2013 - 2:35 pm

I think that’s correct. Though I suppose it wouldn’t be impossible to design a flexible third rail shoe that could switch positions (same voltage either way), it would certainly require a significant equipment change.

Alon Levy April 4, 2013 - 6:33 am

My understanding is that the M7s are designed to be modifiable between over- and under-running third rail, but only in shops rather than on the fly. An on-the-fly option should be investigated, though; the LIRR and Metro-North should figure out how much it costs to modify the M7s (or M8s, even) such that they can switch while in revenue service. Once that’s possible, it’s just a matter of putting over-running third rail on one side of the tracks and under-running third rail on the other, and having trains change while dwelling at Penn Station.

Ryan April 4, 2013 - 3:35 pm

I believe that the M8s were built with contact shoes for both over-running and under-running third rails, but my understanding is that it currently only uses under-running third rail.

Hank April 3, 2013 - 2:53 pm

You are confused on what is meant by ‘over-running’ and ‘under-running’, and mixing in ‘catenary’.

The Hudson and Harlem lines, like the LIRR and the NYC Subway, are powered by 600V-750V DC from the third rail. On the LIRR and the Empire Connection (as well as in Penn Station itself, and through the Hudson River tunnels, and the subway) the contact shoes that pick up that current ride on top of the rail. On the MNCR lines, those shoes contact the bottom of the third rail. These systems are not compatible without modifications to the equipment.

Amtrak, NJT, and the MNCR New Haven line use an overhead catenary (overhead) wire to provide AC current to the trains (anywhere from 12.5kV to 25kV) MNCR’s New Haven line trains can operate on both the overhead wire OR the under-running third rail (which they use from Westchester into GCT). Amtrak has locomotives capable of operating on the over-running third rail in the tunnel portion of the Empire Connection into Penn Station.

For NJT trains to go to Jamacia or LIRR trains to go to Newark, both would need cars similar to what Metro-North uses on the New Haven line, or each would need to install the other’s electrical system.

Nathanael April 3, 2013 - 6:11 pm

You’re confused, Rob.

The New Haven Line has overhead wire, also known as overhead catenary; so does Amtrak and New Jersey Transit. (There are three different electrification systems on this, but that’s another matter.)

The rest of Metro-North has third rail. But it’s weird. It’s referred to as “under-running” because the electrically active part of the third rail is on the BOTTOM — a sort of hook extends under the rail from the train. The top of the rail is not electrically active. (For safety!)

LIRR has third rail, but it’s different. It’s called “over-running” because the electrical pick up is on the TOP of the third rail, just like on the subway.

Rob April 3, 2013 - 6:46 pm

You’re right. Sorry about that.

Alex C April 3, 2013 - 12:52 am

Can we do all three? New Penn Station, new tunnels and through running? I mean, I know actually improving public services is a tough sell for politicians, but damn that would be something.

Bolwerk April 3, 2013 - 11:51 am

In reverse order? Through running is cheap and relatively easy. New tunnels are much more important than a new Penn Station.

Actually, I’m still not convinced a new Penn Station is very important at all.

al April 3, 2013 - 2:53 pm

They need to address signalling and electrification issues. How do you add the cables and 3rd rail in the North River Tubes? You’d need that for LIRR – NJT territory service. A dual surface 3rd rail shoe with a height actuator could work for MNRR – LIRR service. New Haven MNRR and NJT catenary is the easiest to do as both largely run on Northeast Corridor.

Rob April 3, 2013 - 5:29 pm

In my NY days, there WAS 3rd rail in the tunnels, left over I suppose from DD-1 operations. Is it not still there? Or has it been removed in recent yrs?

Nathanael April 3, 2013 - 6:12 pm

How far does it reach? Manhattan Transfer? Newark? It really does little good to run LIRR trains to Newark only to have to reverse them in Newark, which would also block traffic going the other way.

Bolwerk April 3, 2013 - 6:49 pm

Even a stub track off the NEC somewhere would help.

In any case, I think third rail does extend into Jersey, though I have no idea where.

SEAN April 3, 2013 - 6:56 pm

You could run NJCL trains toNew Haven or NJCL trains to Poughkeepsie as a start.

Rob April 4, 2013 - 2:42 pm

I was not addressing what good it would or would not do; only the question of “How do you add the cables and 3rd rail in the North River Tubes?”.

Alon Levy April 4, 2013 - 7:10 am

There is third rail in the tunnels, but it only extends to about the tunnel portal. But having it continue all the way to Newark Penn costs in the low tens of millions.

Nathanael April 7, 2013 - 12:22 am

Which is useless because of the lack of grade-separated reversing tracks at Newark Penn, moving the chokepoint over there.

Sigh. Don’t extend third rail, please. Extend catenary.

political_incorrectness April 3, 2013 - 1:05 am

This is what needs to happen, but it seems to be dealing with a universal issue within American transportation systems that ends up doing a disservice to the customer. Territorialism I like to call it. Where each agency wants to keep its turf, rules, etc. If we can find a way to get over that, it would be a monumental breakthrough.

Also when it comes to Penn. Long dwell times for trains do not help. Quick turn around’s would also help in reducing the time the platform is occupied. Moynihan Phase I with additional access should allow for quicker turns on the commuter trains. I like the idea of widening the platforms at Penn but that will take some serious feat of engineering to complete that.

Brian April 3, 2013 - 1:11 am

Merging NJT and LIRR train services to improve the regions transportation? Would never happen, makes too much sense

Frank B April 3, 2013 - 1:27 am

Surely there must be some precedent for this; the Long Island Railroad was once part of the Pennsylvania Railroad; certainly at one point or another there must have been trains traveling through Manhattan to Jersey, rather than all of them terminating at Pennsylvania.

Ryan April 3, 2013 - 8:08 am

The LIRR was owned by many different companies before it became the LIRR.

I don’t think LIRR ever went through to Jersey, though.

pete April 3, 2013 - 10:08 am

It did, a very long time ago, http://en.wikipedia.org/wiki/P.....C_MP54.27s

Nathanael April 3, 2013 - 6:13 pm

Bizarrely, even under Pennsylvania Railroad ownership, they never really unified PRR and LIRR operations.

DF April 3, 2013 - 2:23 am

I think a little more elaboration would be useful to those of us who are not that knowledgeable on the subject as to what problem through-running is meant to solve. To me it seems like it might be worth trying for the small number of people who are actually traveling through but could only be a marginal improvement to the system as a whole.

Let’s take an AM rush hour NJ Transit train arriving in Penn Station.

One aspect of capacity is the limit on trains that can go through the tunnels. I don’t think this is meant to be affected by through-running.

Another aspect of capacity is the number of people within Penn Station (and the time needed for trains to stay there for those people to get off and on). But, through-running or not, Penn Station would be the only Manhattan station on the line, and almost everybody entering Manhattan would still be getting off there. Of course some people are looking to continue to Jamaica to connect to JFK, for example, but likely these are only a tiny percentage of rush-hour commuters (i.e. people traveling at the time the capacity constraints matter). (This makes it different from the Philadelphia situation, for example.)

Another aspect is how efficiently the trains use the tracks. If the traffic management now is truly such that the trains are staying in the station for an unnecessarily long time, then getting them out by through-running would help. But it is not the through-running per se that would help, but rather getting them out of the station as soon as possible. Unless there are a nontrivial number of through passengers, sending the trains to Sunnyside Yards until they are needed seems like it would work just about as well.

If all Previdi means by “doubling rush-hour train traffic” is that a train entering NYC with 1,000 people then leaves Manhattan for Jamaica with 50 people still on it, thereby “doubling” the number of train movements in and out of Manhattan, then I think “doubling” gives an overly rosy picture of the real effect through-running would bring.

Bolwerk April 3, 2013 - 12:17 pm

Put simply, through running improves queuing efficiency. Instead of trains blocking each other reversing/switching at Penn, they are distributed to points east/west and reverse in calmer places, simplifying movements and, yes, probably allowing a bit more capacity to be squeezed out of the North River Tubes. It also allows better use of the space already there, perhaps by combining platforms.

Compare how the ~20 tracks at Penn only manage about twice the passenger movement of the 6 (well, 8 if you count the shuttle) subway tracks at GCT.

Ryan April 4, 2013 - 3:40 pm

Actually, the shuttle is triple-tracked, so that’s 9 subway tracks at GCT. Nine tracks is slightly less than half of 21 (the number of tracks at New York Penn), by a factor of 3 tracks.

Eric April 3, 2013 - 3:59 pm

Paris commuter rail has through running, NY doesn’t.

Châtelet–Les Halles station in Paris serves 493000 passengers a day (counting only commuter rail) on 4 platforms.

NY Penn Station serves 300000 passengers a day on 11 platforms.

That’s all you need to know.

Ryan April 3, 2013 - 9:53 pm

That’s not a very good comparison.

Here’s a better example which has through-running, this time between subway and commuter rail.

Shibuya Station in Tokyo serves 2400000 passengers a day on 14 tracks (counting the non-through Ginza Line), served by 8 platforms.

By contrast, Gare de Châtelet – Les Halles serves 493000 daily with 7 tracks/4 platforms. And NY Penn Station serves 300000 passengers a day on 21 tracks/11 platforms.

Ryan April 3, 2013 - 9:57 pm

Besides, FRA rule prohibited through running between subway and commuter rail long ago. See this for more details.

Alon Levy April 4, 2013 - 7:16 am

Nobody is talking about subway-commuter rail through-running in New York. The RER doesn’t do that: the Metro is completely incompatible with the mainline network, while the RER is compatible modulo platform heights. The same is true of S-Bahns. Japan’s experience with through-running is useful for questions like what platform width is required, what realistic capacity is, when it’s appropriate to have subway-style seating on commuter rail, and how to handle the agency boundaries, but at least to me the primary model was traditionally Paris.

Nyland8 April 3, 2013 - 11:42 pm

Does Châtelet–Les Halles have two different railroads with two different electrification protocols in two different states on either side of their through-running?

Or don’t I need to know that?

Alon Levy April 4, 2013 - 7:36 am

Two different railroads, yes. RATP runs the RER A. It also runs the RER B south of Gare du Nord, one stop north of Chatelet; SNCF runs the RER B farther north, with a change of drivers. SNCF also runs the entire RER D, which shares the Gare du Nord-Chatelet tunnel with the RER B though not the station tracks. RATP is a city agency that also runs the Metro and SNCF is a national one that runs the intercity train network.

The electrification systems are compatible. That said, France does have a history of dealing with incompatible systems, since the postwar intercity electrification uses 25 kV whereas the older lines south of Paris use 1.5 kV DC, which means that TGVs have to be at least bicurrent. Since other European countries don’t always use these electrification systems, some TGVs are tri- or quadri-current, and European locomotives (for freight) and EMUs that are expected to run between multiple countries are multicurrent. All of this is with catenary, but a few lines also need a special third rail for tunnels where there’s not enough clearance for any catenary. Japan, too, has bicurrent lines, like the Tohoku Main Line.

The upshot is that the major manufacturers are intimately familiar with the need for multicurrent trains, and will happily supply them at little extra cost. And this is on top of the fact that the NJT equipment is already multicurrent and can run through to Metro-North territory.

Nyland8 April 4, 2013 - 8:39 am

Thanks for that comprehensive reply.

Subutay Musluoglu April 4, 2013 - 7:53 am

In Paris you have mostly two voltages – 1.5 kV DC on the sections built and managed by RATP, and 25 kV AC on the sections that are part of SNCF’s national rail network, mostly in the suburbs. There are a few exceptions. Unless I’m mistaken, the voltage changes on the Paris RER vary – on Line A the change between 1.5 kV DC and 25kV AC takes place well away from the city center west of La Defense. On Lines B / D it occurs just north of Gare du Nord. The core of the system, including through Chatelet les Halles, is 1.5 kV DC, which corresponds roughly to the parts of the system over which RATP has management responsibility. Line C voltage varies, and I’ve forgotten where the change occurs. Line E is 25 kV AC throughout.

The other thing that takes place in Paris is an engineer change, where drivers of RATP hand off to SNCF and vice versa. Except for the occasional strike or other labor action, this has generally worked well for decades until more recently, when it became obvious that it was contributing to delays on the densest part of the system. Recent agreements between RATP, SNCF, and the various unions is now allowing for this practice to be phased out.

Nathanael April 3, 2013 - 6:17 pm

Reversing (outgoing) trains from one platform cross the path of incoming trains to another platform. This slows everything down. If you just run in, stop, and run out all on one track, you can push though a lot more trains. This is why the subway can handle so many passengers. Notice that the terminal stations on the subway have *extra platforms* because of the delays involved in reversing.

Now, at Sydney Central Station in Australia, the main station is set up with a gigantic “thresher” of diving and rising bridges so that incoming trains come in from “above” and outgoing trains go out “below” (or maybe it’s the other way around, I don’t remember). So there, the reversing trains don’t block anything. But there is no station in the US with that arrangement, and it would be very hard to build.

There are other costs to reversing: by FRA regulation, a brake test has to be done (which takes minutes), the driver has to get out of one end of the train and walk to the other end, etc.

Bolwerk April 3, 2013 - 10:41 pm

I’m not quite clear on what the implications of the brake test rule are. It would seem it might be possible for a second crew member to be avoided if someone were permanently stationed the terminal to do the brake test.

Nathanael April 7, 2013 - 12:23 am

The implications are mostly delay. The brake test is stupidly slow. It means you can’t just “pull in – pull out”… the way the Subway can reverse at the ends of lines.

D.R. Graham April 3, 2013 - 2:27 am

In all honesty MSG has to move. The thought of the Garden relocating has been brought up so many times over the past two decades that it might be finally time to sit Dolan and his managing group down and discuss a serious relocation plan within the city.

In turn you can serious gut the track and platform layout of Penn and really get something good going. Or else start figuring out a way to throw in a new level of tracks but that is likely to cost way more than ESA is currently running under two levels of tracks at Grand Central.

Berk32 April 3, 2013 - 8:03 am

the City missed its chance to move MSG.

It would cost the city/state/etc too much to move it now that MSG has done its renovations

Jeff April 3, 2013 - 8:29 am

They did sit them down and asked them to move, and Dolan did consoder it. Politics got in the way and killed the proposal though. So Dolan started his renovation and the window of opportunity was lost.

Kevin April 3, 2013 - 11:20 am

The operational lifespan of modern arenas is pretty short.
Just planning station improvements and a MSG replacement will take a dozen years.
Which will be about the same time the recently renovated MSG will be deemed to be non longer modern enough.

marv April 3, 2013 - 3:48 am

assuming that we write off facilities over a 50 year period, including both construction costs+interest expense, how much would each desired improvement cost per expected passenger trip? What would be the time savings per trip? then let us try to figure out which improvements would be worth the cost.

Perhaps it would be cost and operationally effective to run trains back and forth from secaucus, to Jamaica via Penn Station. Almost 30 t/hr could be run while still allowing room for Amtrak. To allow of boarding/bebarking (commuter trains take longer than subways due to narrower isles and fewer doors) 4 or even 8 tracks at Penn Station could be used for this service whiles still leaving plenty for Amtrak.

Service could be increased on all branches (as needed) both from Jamaica and NJ, as the river tunnels would not longer be the constriction point. I wonder if the time switching trains would be justified by the increased service and monies saved.

AG April 4, 2013 - 1:23 am

“assuming that we write off facilities over a 50 year period, including both construction costs+interest expense, how much would each desired improvement cost per expected passenger trip? What would be the time savings per trip? then let us try to figure out which improvements would be worth the cost.”

well unfortunately – most elected officials don’t use that type of sensible thinking.

Larry Littlefield April 3, 2013 - 7:15 am

You have to consider the differences between commuter rail and the subway. On the subway, passengers get on the first train and empty the platform. On commuter rail, they want to get on one specific train, and that increases dwell time and decreases capacity dramatically.

The super cheap option would be to eliminate direct NJT service to Penn Station, and run shuttle trains with no seats back and forth to Secaucus. And add to the rail relay and yard capacity there so more trains could terminate there.

But that would be a lower quality ride, with two or three changes (including the subway) rather than one or two. And in any event, I still think we need another track to Penn, if only to provide redundancy and eliminate single tracking during maintenance.

Bolwerk April 3, 2013 - 11:39 am

When there is a major difference, you’re doing it wrong. Yes, we’re doing it wrong. But, with NYCTA often having up to four services per platform direction, people are certainly skipping trains they don’t need. Indeed, regional commuter rail is probably less likely to serve multiple trains.

Rob April 3, 2013 - 12:13 pm

I think the RER commuter trains in Paris have quicker on-off by passengers. If we did thru-running with Metro North and LIRR, maybe this would work?

Nathanael April 3, 2013 - 6:19 pm

Yep. Every train from NJ should be continuing eastward.

Nathanael April 3, 2013 - 6:20 pm

(Well, um, except perhaps for some of the long-distance Amtrak trains. It makes sense to terminate the train from Florida at Penn Station.)

Bolwerk April 3, 2013 - 7:59 pm

Ideally, those would go to the airport somehow. But airlines also can’t play nice with Amtrak, even if it’s in their benefit.

Nathanael April 7, 2013 - 12:24 am

Well, they do actually go to the airport — Newark Airport, the oldest of the commercial airports in the New York City area.

(The other two are frankly due to the arbitrary political division between New Jersey and New York.)

Henry April 7, 2013 - 12:45 am

LGA has never had easy rail access on any ROW, and the most obvious one, the Grand Central Parkway, has a height restriction due to landing paths.

JFK has never had direct rail access, despite numerous sad attempts to establish it with AirTrain running (?!)

Honestly, if there was a place that could fit a mega-airport relatively close to the city, it would’ve been Fresh Kills. The other three are surrounded by environmentally sensitive areas, snobby landowners, water, or a combination of the three. Floyd Bennett Field would be a nice location for an airport, if they would reclaim some land (which they won’t).

AG April 9, 2013 - 9:47 pm

Floyd Bennett Field is part of the Gateway National Recreation…. it’s federal parkland… it won’t be an airport again… especially as they are trying to “heal” Jamaica Bay.

Henry April 3, 2013 - 9:00 pm

Any comparison to RER is pointless, because RER has multiple center-city locations to drop off passengers, whereas any service entering Manhattan will only have one – either GCT or Penn.

Supposedly, the PRR was looking to also build a Fourth Avenue station when constructing the railroad tunnels to Penn, but this never happened. (Fourth Avenue is today’s Lexington Av, I believe.)

Woody April 3, 2013 - 10:31 pm

Well, 4th Avenue got renamed Park Ave. It’s halfway between 3rd and 5th Aves.

Lexington is actually 3 1/2 Ave, and Madison is actually 4 1/2 Ave.

Frank B April 4, 2013 - 12:24 am

And, 4th Avenue still existed below South of Grand Central for some time afterward; but luckily some politician thought “Everyone’s mail should go to the wrong place!” And created Central Park SOUTH. Unfortunately, this leads to some aggravation in getting your mail delivered…

And yet somehow we ended up with a 6 1/2 Avenue… Go Figure.

Ryan April 4, 2013 - 7:56 am

And Vanderbilt Ave is actually 4 1/4 Ave. Makes no sense.

Nathanael April 7, 2013 - 12:25 am

“any service entering Manhattan will only have one – either GCT or Penn.”

That’s what “alternative G”, the tunnels connecting Lower Level GCT to Penn, was going to fix. ANYway…

Alon Levy April 4, 2013 - 7:43 am

Ew. What you’re proposing re shuttles kills ridership, since riders are exceptionally averse to transfers near the CBD (origin-end transfers are fine). It’s pretty bad even if the transfer is done perfectly, which it can’t because Secaucus only has one island platform for a cross-platform transfer, which needs to be used by trains in opposite direction.

And there’s no need for it. The subway routinely has two services sharing platforms. On top of that, the subway doesn’t have fixed clockface schedules, since even the branches run frequently enough for show-up-and-go frequency. The commuter rail branches, like commuter rail everywhere else in the world, are timetabled. So if the schedules are redone so that riders can remember them – “the express to Trenton leaves at :07 repeating every 20 minutes, except at rush hour when it repeats every 10 minutes” – then it’s not a problem because riders will know when to go down to the platform. On top of that, Penn has so many tracks it’s possible and desirable to schedule trains such that no two successive trains use the same platform.

Jeff R April 3, 2013 - 7:19 am

When the connection for LI trains to stop at grand central won’ttis free up Penn Station?? I see it pointless to put any money into expanding Penn station when the volume will drastically decrease. In fact when world trade center opens fully most NJ rail will travel threw therre . thus another decrease. I see Penn station as a during relic ..

Ryan April 3, 2013 - 8:04 am

Actually, the East Side Access is intended to shorten the trip between LI and NYC. So yes, there will be freed-up capacity.

There’s no mention of a NJT station ever being planned at World Trade Center.

Benjamin Kabak April 3, 2013 - 8:08 am

A few points:

1. When WTC options, it will have no impact on New Jersey Transit. It’s for PATH trains only and won’t boost rail capacity at all at that end. It certainly won’t impact the number of trains into or out of Penn Station. It’s a PATH station.

2. When ESA opens, there may be space for additional capacity, but you have the MTA angling to bring Metro-North into Penn and LI politicians fighting to keep LIRR volumes at their current level. Either way, Penn won’t suddenly be empty or have much, if any, capacity for more trains.

Ryan April 3, 2013 - 9:11 am

3. The LIRR would want to keep their Penn slots, so while there would be freed-up capacity theoretically, that capacity will then be used by extra LIRR, NJT, and Amtrak trains.

4. MNR will not want to operate in both Penn and GCT, so NJT would take even more slots off the freed-up capacity to run trains from Port Jervis and Pascack Valley. MNR would then no longer run trains west of the Hudson.

Benjamin Kabak April 3, 2013 - 9:12 am

Are you just making things up again? I’ve asked you not to do that, and I’m frankly getting sick of you doing this. There’s no basis in reality for these claims – especially considering that MNR and LIRR operate under the same parent agency. Stop. Final warning.

Ryan April 3, 2013 - 11:00 am

I’m not the same commenter as “Someone” who is reputed to make things up. I know my sources. I can prove it.

Benjamin Kabak April 3, 2013 - 11:01 am

Same email addresses, same IP addresses. The evidence isn’t weighing in your favor.

Ryan April 3, 2013 - 11:32 am

Nope, I switched to my personal email.

I use a shared computer. The computer isn’t mine.

Bolwerk April 3, 2013 - 11:43 am

Do you share a toothbrush too?

SEAN April 3, 2013 - 12:57 pm


I suggest you throgh yourself on the mercy of the court. LOL

Through running has been tested under the guise of “Train to the Game,” and by all indications it’s been a success. The next logical step is to expand apon it.

Trains dwell in Penn station for 10 – 15 minutes before departing. The issue that relates to this is the mad rush to a given track once a train is anounced. However if you know where to wait, you can minimize the onslaught.

Ryan April 3, 2013 - 1:25 pm

Nope. 😀

Simon April 3, 2013 - 11:46 am

Neither does this post.

Ryan April 3, 2013 - 1:20 pm

Oops. I accidentally must have commented under “Someone”, who in real life is not named Ryan (unlike I.)

Benjamin Kabak April 3, 2013 - 1:24 pm

This is truly stretching the bounds of believability. Let’s just say it shouldn’t happen again and leave it at that.

AlexB April 3, 2013 - 2:23 pm

Can’t you just delete things that are blatantly false and misleading?

Chris C April 3, 2013 - 6:54 pm

OK so post links to these sources so we can verify what you say is correct?

And in future please include the links with your original posts and if you can’t post links then don’t post your ‘facts’.

Other people can do it and so can you !

Ryan April 3, 2013 - 9:54 pm

I never said it was fact. I’m just predicting it (and indeed, I am a very bad predictor.)

Frank B April 4, 2013 - 12:28 am

I am starting to regret following this thread… I’ve received over 50 emails this evening alone alerting me of new postings.

We’re here to discuss transit for our great city and our region, and how to improve ridership, service, and access for all.

And ‘speculation’ will not help that; We’re here to share knowledge; not guesswork.

Chris C April 4, 2013 - 7:59 am

YOU said – “I know my sources. I can prove it.”

So PROVE it then.

Oh but you can’t because your ‘facts’ are (once again) made up so aren’t ‘facts’ at all.

They aren’t even that good as ‘predictions’ either

Walter April 3, 2013 - 6:43 pm

Pure BS here.

The president of Metro-North has been pushing Penn Station Access very hard, and always crows about the smallest improvements to the Port Jervis Line.

The Port Jervis Line is a commuter rail service in New York State, so it falls under the MTA, and since it serves northern suburbs, it falls under Metro-North. Metro-North doesn’t even run the trains up there, it simply hires NJ Transit to run them. But it’s still an integral service of Metro-North that’s not going anywhere and definitely not being taken over by New Jersey taxpayers.

Ron April 3, 2013 - 10:08 pm

Yup, forget the fact that those lines don’t and can’t connect to Penn!

Frank B April 4, 2013 - 12:30 am

Well… At least you can transfer at Secaucus Junction now; it was far worse years ago…

Bolwerk April 3, 2013 - 11:43 am

The nice thing about through service NJ<->LI is it probably could be used to show LI politicians they’re netting even more service, not only to Penn but to points west. It’s a truly silly thing to oppose.

Ryan April 3, 2013 - 8:02 am

If there were turning loops and third rails in New Jersey, then maybe the LIRR can through-run trains there, and NJT can turn around in Sunnyside.

Eric F April 3, 2013 - 8:40 am

The Penn platforms are too narrow and access is too cramped to permit the discharge of passengers from one train with simultaneous boarding. Occasionally an Amtrak train will share a platform island with an NJT Train. Chaos ensues.

LIRR uses 3rd rail, NJT uses overhead wire, so they can’t use each other’s systems.

Bolwerk April 3, 2013 - 11:45 am

Dual mode equipment for the win.

(Isn’t there even historical – that is, PRR – precedent for service from points south to Long Island?)

Sic Transit Philadelphia April 3, 2013 - 5:50 pm

The PRR’s Sunrise Special ran 1x weekly in the warm months for two decades. Pittsburgh (!) to Montauk Fridays, return Mondays.

Frank B April 4, 2013 - 12:35 am

Thank you! I had asked that before, and I couldn’t get a clear answer.

I knew there must have been some kind of service like that; for the Pennsylvania railroad to own the Long Island Railroad and have no through trains seemed like a complete impossibility to me.

Greatly appreciated. Now we can point out to the Long Island Politicians they’re actually getting LESS service than they used to. They’ll start support through-running in 5 seconds flat. 😀

Nathanael April 3, 2013 - 6:23 pm

Dual mode equipment is getting to the point of stupidity — and I believe five-mode equipment is genuinely unreasonable.

Amtrak has to handle three overhead wire electrification systems. LIRR and Metro-North have two different third rail electrification systems.

Re-electrification is what needs to be done. The UK has finally decided to start replacing its mainline third rail with overhead catenary (after more than 100 years). The same should happen in New York City: 60Hz 25kV overhead catenary for the win.

Walter April 3, 2013 - 6:49 pm

Can’t have anything more than 12.5kV catenary on the New Haven Line do to clearance issues with bridges over the line. Simply using existing Amtrak locomotives that can operate on different voltages is fine.

Nathanael April 7, 2013 - 12:28 am

Two voltages on the same frequency is easy. Two frequencies is… tolerable. DC third rail and AC overhead is beginning to push it, though it’s been done (with low top speeds on one or the other) in the UK.

Three varieties of AC and two varieties of third rail is just too damn much, especially when one of them (Metro-North third rail) is Unique In The World. There’s a great advantage in replacing that with overhead catenary.

Bolwerk April 3, 2013 - 6:54 pm

It takes time. Meanwhile, isn’t at least some NJT equipment already third rail equipped?

Nathanael April 7, 2013 - 12:29 am

AFAIK, no. Certainly none of it is equipped for Metro-North’s quirky, unique-in-the-world underrunning third rail, which is really the system which should be retired first.

Simon April 4, 2013 - 2:25 pm

What is the advantage of overhead catenary over third rail?

Nyland8 April 4, 2013 - 2:31 pm

Well … for one thing, nobody crossing the tracks will step on a catenary. A lot of these suburban trains cross at grade, are not fenced in and do not have totally secure ROWs.

Walter April 4, 2013 - 9:15 pm

The main advantage of using catenary is that it’s usually powered using alternating current, while third rail is usually direct current.

DC requires multiple substations, roughly one every mile, while the more manageable and efficient AC allows one to just tap into the power grid to power the trains. Amtrak, for example, has only 4 substations between New Haven and Boston, while Metro-North has something like 30 (on the underpowered line that only allows 8-car trains) between New York and Brewster.

This goes way back to Tesla vs Edison, and the fact that we don’t all have direct current generators in our basements shows which came out on top.

Nathanael April 7, 2013 - 12:31 am

FWIW, I believe nobody uses AC on third rail except some isolated subway systems — it’s considered far too dangerous to trespassers.

As a side effect of the advantages of AC, AC can allow for higher-speed trains.

Also, maintaining contact between the train and the third rail becomes a serious problem above about 100 mi/hr, but maintaining contact with the catenary is relatively easy.

Given that the Empire Corridor (Hudson Line) is supposed to get high-speed trains, and that it has a unique-in-the-world form of third rail, it really should have catenary strung over it instead.

Bolwerk April 4, 2013 - 10:20 pm

It’s probably cheaper to both install and maintain, especially over long distances. I suspect it’s more stable at higher speeds too.

It might be less practical in tunnels. Third rail probably takes no extra space over not having it.

Nathanael April 7, 2013 - 12:32 am

This is why third rail is still used in subways (and nowhere else). Overhead wire requires somewhat larger diameter tunnels (even if you use low-profile “overhead rail”, which some tunnels do), third rail doesn’t require any extra space.

Alon Levy April 5, 2013 - 1:15 am

Third rail has a limited voltage, which limits the power that can be drawn. It’s of no consequence to urban rail or low-speed commuter rail, but the very high-powered trains used on high-speed intercity service or fast commuter rail schedules (like the FLIRT, Coradia Continental, etc.) need higher voltage than can be supplied by third rail.

Eric April 3, 2013 - 4:19 pm

There are too many platforms/tracks in Penn Station already. Get rid of a couple tracks and there will be plenty of platform space. on the nearby platforms.

Ryan April 3, 2013 - 10:11 pm

And plenty of delays for other trains.

jim April 3, 2013 - 8:43 am

LIRR and Amtrak already through run. LIRR trains come in from LI in the morning, unload and then run through to the West Side Yard. They make very efficient use of the platform capacity. It’s NJT that’s the problem, despite Previdi’s attempt at false equivalence. And most of the NJT trains that turn at Penn Station, buggering up the other train movements, are deadheading back to Jersey because there’s not enough train storage space east of the Hudson. There isn’t enough demand for counterpeak travel, reverse commutes, for more than a fraction of them to operate as revenue trains. Running LIRR (or Metro-North) trains through to Jersey doesn’t create that demand.

If one wants NJT to stop turning at Penn, then (1) one has to provide more train storage space east of the Hudson and (2) one has to make it worth their while to spend the money to use that storage. They aren’t going to do it just because Previdi thinks it’s a good idea.

Ryan April 3, 2013 - 9:07 am

We do have that. It’s called Sunnyside Yard.

jim April 3, 2013 - 9:19 am

Sunnyside only holds 18 trainsets. Not enough.

Ryan April 3, 2013 - 11:01 am

18 tracks is more than enough, or the trains can use the loop tracks

Alon Levy April 4, 2013 - 7:47 am

You’re forgetting that trains don’t just unload at those tracks; they park there between the am and pm peaks.

al April 3, 2013 - 3:11 pm

They could combine some of the trains between the new MNRR Penn Access on MNRR New Haven with NorthEast Corridor NJT. They can run between Trenton and New Rochelle, or New Haven and Philadelphia. The first service can use M6/M8 while the second pattern can use low platform capable Comet 3/4/5 or Arrow 3.

A far out proposal is for the constituent Northeast Corridor commuter rail agencies (MBTA, ConnDOT/MNRR, NJT, SEPTA, MARC) to operate through run local and express trains with EMU train sets and have AMTRAK run connecting Acela Express and Super Express service.

Nathanael April 3, 2013 - 6:26 pm

“And most of the NJT trains that turn at Penn Station, buggering up the other train movements, are deadheading back to Jersey because there’s not enough train storage space east of the Hudson.”

The first, and easiest, step is to run a bunch of these NJT trains as contra-peak Metro-North trains on the New Haven Line. There’s enough storage space in New Haven and points east.

There are some problems, however: notably, the southernmost four (IIRC) tracks are stub-end tracks which basically can’t head east. It would seem logical to put Amtrak services terminating in NY on them, but most of those have to go to Sunnyside Yard for servicing. The remainder is a mix of Empire Service, Keystone, and Northeast Regionals which are not at the beginning or end of their service days, and that’s not very many trains, certainly not enough to use all of the four tracks.

So in some sense NJT is using the southern tracks “because nobody else is”, and that causes NJT to do reversing movements.

Eric F April 3, 2013 - 8:59 am

I forgot to mention that working out using LIRR crews to work NJt routes and vice versa would lead to union work rule armageddon. The U.N. Security Counsel wouldn’t be able to sort out that fracas.

Ryan April 3, 2013 - 9:06 am

Which is why the LIRR can’t currently have operating rights on NJT tracks.

Bolwerk April 3, 2013 - 11:52 am

Oh, the poor dears. Having to share sounds so much worse than not having a job to go to.

Nathanael April 3, 2013 - 6:27 pm

LIRR unions are whack. I bet NJT and MNRR can work something out, but LIRR? Yeesh. LIRR is an *extremely* old-fashioned railroad.

Bolwerk April 3, 2013 - 6:55 pm

Not sure it was the unions’ fault, but they didn’t work something out when ARC could have gone to GCT.

Tim April 5, 2013 - 9:05 am

Didn’t Water Tunnel #1 have something to do with the reason ARC was so deep when originally designed?

orulz April 3, 2013 - 1:01 pm

The way it works in Tokyo for through running trains is that the trains go through, the crews do not. All this would involve is training the NJT crews to operate the LIRR equipment, and the LIRR crews to operate the NJT equipment.

LIRR crew hands the train off to NJT crew at Penn Station. They then walk across the platform and become the crew of the next NJT train headed in the other direction, easy enough.

The ideal cost is no object end game would be running every NJT train through at least as far as Jamaica and every LIRR train at least as far as EWR would be great, but in a fiscally constrained real world, at the very least, each and every reverse peak LIRR and NJT train should be a through running train.

Bolwerk April 3, 2013 - 1:05 pm

That’s pointless and unnecessary, and we should dispense with these idiotic union fiefs. Amtrak simply needs to refuse to allow through running, and the other railroads can work out how to best share tracks and terminals – and preferably unify fares.

al April 3, 2013 - 3:13 pm

We could fall back on trackage rights regimes.

Nathanael April 3, 2013 - 6:30 pm

That works. No prob.

Ryan April 3, 2013 - 8:23 pm

NYC has more stringent trackage right rules than Tokyo does, I’m assuming.

Alon Levy April 4, 2013 - 7:51 am

It’s suboptimal, but it’s done at places with big track capacity crunches. At Gare du Nord, the RER B+D share four tracks, two in each direction, while running 32 peak tph between them; they merge to the south to a two-track tunnel, but the next station south, Chatelet, has separate platforms, so it’s possible to squeeze more tph than on a line where each tunnel track leads to just one platform track. On the RER B, which has 20 of those 32 tph, they change crews at Gare du Nord.

Alon Levy April 4, 2013 - 7:53 am

Correction: the RER B+D have separate tracks at Gare du Nord – two platform tracks are dedicated to the B, two to the D.

Ryan April 3, 2013 - 1:23 pm

Which brings to mind the fact that even the rolling stock in Tokyo is interoperable, unlike that of LIRR/MNR/NJT.

Nathanael April 3, 2013 - 6:33 pm

“in a fiscally constrained real world, at the very least, each and every reverse peak LIRR and NJT train should be a through running train.”

I looked at the actual schedules. NJT runs a significantly lower count of trains than LIRR, to the point where NJT peak service is approximately at the same level as LIRR reverse-peak service. NJT reverse-peak services is a LOT lower than LIRR peak service.

In order to avoid multiple re-electrifications rather than just one, I’d advise running all the NJT services east, substituting for a small number of peak LIRR trains and a lot of reverse-peak LIRR trains. LIRR would become even “peakier”, while NJT would become more balanced.

This is all assuming some sort of sane re-electrification, of course.

Walter April 3, 2013 - 6:56 pm

Metro-North and Amtrak crews do this already on the few through running Shore Line East trains from Stamford to New London. The Metro-North crew gets off at the platform in New Haven, and the Amtrak crew is there waiting. Takes a couple minutes at most.

And this has nothing to do with union rules. It’s all about FRA mandated employee qualifications on the territories and operating rules. This way the crews only have to know their territory, signals, and rules, and don’t have to carry a bunch of different rule books with them.

Nyland8 April 3, 2013 - 9:03 am

Let me get this straight. The consensus seems to be that NYPenn is overcrowded – i.e. cramped ingress and egress, platforms much too narrow, not enough tracks, etc.

. . . and that somehow turning trains around quicker – i.e. getting them in and out faster – reduces those limitations. Is that correct?

Is the proposal of through-running somehow supposed to mitigate that problem? Then wouldn’t we have people waiting to leave, on an already too-narrow platform, with people that are simultaneously disembarking the train??!!?? OR – you’re not telling the people in NYPenn which platform their train is leaving from until everyone has already disembarked – in which case there is no difference in the turnaround time from what we have now.

If the trains are already leaving the station at rush hour with only a handful of reverse commuters, how does that change with through-running? Wouldn’t the same NJT, LIRR or MNRR train be leaving with just as few people in the reverse commute?

What am I missing?

Off hand, it seems the only advantage to through-running would be to the train manufacturers, who’ll make hundreds of million$, which we’ll have to spend converting everyone’s equipment to run on each others tracks.

Does anybody have any real data on the number of “pass through” passengers Previdi alludes to? And what makes him think there are enough in … say Far Hills New Jersey who are headed for … Little Neck ??? – just to give an example. Wouldn’t MOST of those commuters still be getting off in NYPenn and waiting for their train to … LI, Jersey, etc. ???

There may be many places in the world where through-running is an elegant solution to a transportation problem – but are we sure it applies to NYPenn?

Again … what am I missing?

JJJ April 3, 2013 - 2:47 pm

Turning trains around required crew work – brake checks etc. Continuing straight does not. Having those delays be somewhere else opens up Penn.

Eric April 3, 2013 - 4:58 pm

You can widen the platforms at the expense of tracks which are now unnecessary. For example, if P=platform space and T=tracks, you can go from




and now you have half as many platforms (still enough) but each one has much more space than you need.

Benjamin Kabak April 3, 2013 - 5:00 pm

I’m a little unclear as to the why. One of Penn’s problems is that it doesn’t have additional capacity for more trains. That’s a track issue, not a platform issue. To expand both tracks and platforms, you need to expand the underground footprint — which can’t be done easily (or perhaps at all) with MSG and its support columns still in place.

Nathanael April 3, 2013 - 6:38 pm

Benjamin: think this through.

(1) If you use through-running, you need fewer platforming tracks.
(2) If you need fewer tracks, THEN you can widen the platforms.

That’s the logic.

Henry April 3, 2013 - 8:45 pm

Well, another limiting factor is the fact that Penn platforms take obscenely long to clear out – a good two to three minutes. In those two to three minutes, you can’t dump more passengers onto the platform, because the previous passengers haven’t cleared out yet, and that would lead to overcrowding issues. If you can get people off the platforms faster, then theoretically you can also get trains in and out of stations faster, assuming that they’re going straight through. Thus, wider platforms (and more frequently spaced, widened access points) will lead to less clearing time and more time for a train to clear the tracks for the next one.

Ant6n April 3, 2013 - 6:06 pm

Going from
instead of
Still provides wider platforms, but now every train could be discharged from two sides (some call it the “Spanish solution”). This would allow shorter dwell times, especially with the bi-levels that have lots of passengers per door.

Nathanael April 3, 2013 - 6:40 pm

This is the right thing to do, except there are two problems:

(1) Not enough tracks. Through-running will help, but it won’t allow you to get rid of *half* the tracks.

(2) The damn pillar locations. There’s currently a dense line of pillars between each pair of tracks. So if you go from
you are faced with trains which can’t be boarded from one side due to that damn line of pillars, which will be blocking the doors.

ant6n April 3, 2013 - 7:38 pm

Oh right, I forgot about those damn pillars. Maybe something more creative is needed. But boarding/alighting from two sides should reduce dwell time a lot, together with through-running this means track capacity may actually increase, even if the number reduces.

What is usually a component of through-running is having multiple stations along the downtown trunk-line, distributing the passenger load. Why are there no proposals to put a second station in Manhattan?

Nathanael April 7, 2013 - 12:36 am

“Alternative G” was going to run straight through from Penn to Grand Central Lower Level. It was actually going to REDUCE operations costs while increasing ridership, AND it had the cheapest capital costs of all proposals.

But it was rejected because, according to the study, they were worried about the rich landowners along the way. Not kidding!

Ryan April 3, 2013 - 10:00 pm

Install fences between the damn pillars, then. It defeats the point, but at least each platform has its own dedicated space.

Alon Levy April 4, 2013 - 7:55 am

Re pillars, I checked this a few years ago and with NJT door locations, only one or two doors per train faced a pillar.

Nathanael April 7, 2013 - 12:37 am

That’s enough to be fatal given the relatively small number of doors per car.

Ryan April 3, 2013 - 8:21 pm



rather than
provides the optimal wide platform solution, while still having more than 3 tracks.

AlexB April 3, 2013 - 9:27 am

Through running is a great idea, but the main choke point is trans-Hudson capacity and through-running can’t get more trains (24 per hour existing) from NJ to NY during the morning rush than current capacity allows. I’d be very happy to be wrong about that. The best solution would be to combine through running with a single new “peak direction” track under the Hudson so Newark to New York has a minimum of three tracks all the way, with other elements of the full Gateway project to follow. A single track with no stations could actually be done for a very reasonable amount, probably around $2 billion, and would almost double capacity. Shared between the feds, NJ and NY, it shouldn’t be a big burden. As there would have to be almost twice as much space to store all these trains, they would almost certainly have to continue to points in Long Island where there is more yard space.

Running NJT trains to the LIRR is easy; they’re already “rest” at Sunnyside during the day. If some of those could go to Jamaica and some could go to Flushing, it could be a big help.

LIRR trains stop in the Hudson Yards and they could be sent to New Jersey, with Newark and Metropark being great termini.

Eric April 3, 2013 - 5:00 pm

You are correct, that is the main choke point. But is a single track tunnel really that much cheaper than a dual track tunnel? The fourth track will be needed eventually anyway.

Henry April 3, 2013 - 8:48 pm

Tunneling is relatively cheap – the tunnels themselves are a matter of starting up a TBM or two. The main cost of ARC, Gateway, and everything else after it is the land acquisition cost for a new station – in ARC’s case, a giant hole in the Manhattan bedrock, and in Gateway’s case, demolition of two blocks of Midtown for Penn South.

Ryan April 3, 2013 - 9:59 pm

WIth four tunnels the station can hold 48 trains per hour from New Jersey. But what about the trains from Queens?

Henry April 4, 2013 - 8:31 pm

There are already four tunnels for LIRR trains heading to Penn, and there will be two more for LIRR trains heading to GCT.

LIRR has a total of four terminals to run to now, and five if you count LIC on the recently terminated Lower Montauk. The main constraint on LIRR capacity is the crawl through Jamaica, the lack of a third track past Floral Park (if I’m not mistaken), and the lack of a second track somwhere further down the main line.

Nyland8 April 4, 2013 - 12:08 am

… which brings us back to …

AlexB April 4, 2013 - 11:42 am

Usually each tunnel would be a separate bored tube or separate tubes laid on the bed of the river, so yes, only doing one would be cheaper. Sometimes they bored a much larger tube that can hold two tracks, but that’s not usually cost effective. Doing everything at once is probably cheapest, but I’m trying to suggest a way to get more capacity as fast as possible considering the existing tunnels are maxed out and the trains are standing room only for quite a ways.

AG April 4, 2013 - 1:30 am

That’s exactly what the Crain’s editorial said (well at a higher dollar amount).

lawhawk April 3, 2013 - 9:36 am

Through running addresses a couple of issues and is worth examining in depth.

Consider that once a LIRR or NJT train enters NYP to discharge its passengers during a rush hour, it is either going to the yard in a no-revenue run, or will carry a counter commute with fewer passengers.

Cross-running trains might not necessarily increase capacity, but it would increase flexibility and reduce the number of no-revenue runs through the East River of Hudson River tunnels.

That could increase the efficiency of the system, but there are significant hurdles to overcome – work rules and the kind of equipment run on the systems isn’t compatible beyond NYP or Sunnyside Yards.
Addressing equipment issues first:

NJT uses overhead lines to get into NYP. LIRR uses a 3d rail. While NJT has dual mode locomotives that could run on the overheads into NYP and Sunnyside and then use the diesel mode for the LI segment, LIRR doesn’t have the same kind of capacity as far as I’m aware. They’re stuck having to either purchase dual mode locomotives or sourcing electrics that can convert from 3d rail to overhead for the runs into NJ. That’s an added cost to both agencies but one that involves the purchase of locomotives. While dual level NJT cars can fit through the Hudson River tunnels, I’m not sure that the LIRR cars could do so in a through commute. If they can’t, that would have to be addressed as well (either by running only single level cars or by LIRR purchasing compatible railcars).

Work rules between LIRR and NJT are different, and not sure how they’d be reconciled, but it should also be noted that NJT and MNR have a shared agreement on West of Hudson Service – namely the Port Jervis Line to share equipment and access to rail. It’s that deal that could potentially form the basis for a through service agreement.

Would this save both NJT and LIRR money over the long term? I don’t know – there’s an initial cost to get equipment that can run on both lines (and NJT is still down a number of dual mode locomotives due to Sandy damage), but optimizing and harmonizing service might provide a boon to commuters that are going beyond NYP in a counter-commute fashion during rush hours.

With a through commute, someone who would have had to transfer at NYP to go from points in NJ to somewhere in Queens or LI could then stay on the train to head to Jamaica (for a transfer to JFK via Airtrain for instance). Likewise, someone from LI could avoid the transfer at NYP for a trip to EWR).

Ant6n April 3, 2013 - 6:08 pm

Third rail and overhead could just be extended further into the other agencies territories.

Nathanael April 3, 2013 - 6:41 pm

“NJT uses overhead lines to get into NYP. LIRR uses a 3d rail. While NJT has dual mode locomotives that could run on the overheads into NYP and Sunnyside and then use the diesel mode for the LI segment, ”

This makes it clear what to do: Extend NJT trains to replace LIRR trains. Keep the existing LIRR trains for the LIRR peak; LIRR runs more services than NJT anyway.

Alon Levy April 4, 2013 - 8:06 am

I’d argue it the other way. The M7s are better trains than the NJT equipment: they are lighter and more powerful and accelerate faster than loco-hauled trains. It may well be cheaper to make them multi-voltage, or just retrofit the M8s with LIRR third rail and switch equipment around, and then use the ALP-46s mainly at rush hour and mainly for express runs.

Nathanael April 7, 2013 - 12:41 am

The M8s are OK, and ordering a bunch more M-8s would be smart.

HOWEVER, to use the M8s, you’d have to change out the 25Hz overhead system in favor of the 60 Hz overhead system. Which should be done anyway.

Or you’d have to get yet another captive fleet of different, incompatible M-8s which use a different third rail AND a different overhead catenary. Sure.

The M7s…. well, retrofitting them with pantographs isn’t gonna be possible. Period. Sorry.

Everyone is trying to avoid re-electrification. Re-electrification is *correct*. Everything else will cost more in the long run. We’ve fooled around with this incompatible mess for 100 years or so, that’s *long enough*.

Carwil April 3, 2013 - 9:54 am

For those asking for details, I think these articles are a good place to start:

Rob April 3, 2013 - 10:08 am

Thanks for these links. When I saw the headline on SAS, I thought it was deja vu from the Transport Politic discussions.

Eric F April 3, 2013 - 10:46 am

Effectively, there is no ‘cheap’ solution. The cheapest thing I could think of — which isn’t all that cheap — is to acquire right of way to double capacity between Kearny, NJ and NY Penn but only actually build and operate an additional one track tunnel under the Hudson. If you did that, you’d massively unclog the train routes for NJT and Amtrak and more or less eliminate conflicts. You’d also be able to do maintenance work and respond to equipment problems without seizing up the entire east coast. Even that ‘little’ plan implies rebuilding Portal Bridge and adding trackage in the Seacucus area.

AlexB April 4, 2013 - 11:44 am

I think that’s exactly what should be done and it should start yesterday. The new Penn Station can follow a decade or two from now.

Douglas John Bowen April 3, 2013 - 11:37 am

It’s interesting that “Amtrak’s plans seem to focus on rail capacity as long as costs are no limit,” while not acknowledging that NJ Transit’s all-or-nothing fixation on Access to the Region’s Core isn’t brought up.

Amtrak, at least, didn’t call for a stub-end, parallel railroad lacking interconnectivity.

Other than that, a thoughtful piece.

Benjamin Kabak April 3, 2013 - 11:38 am

What’s the point really, by now? We’ve rehashed ARC to death around here, and the project was killed a few years ago. It ain’t coming back.

Eric F April 3, 2013 - 1:40 pm

Actually, the mess that is relations between the transit systems, whether it be Amtrak denying Portal improvements to NJT after ARC was cancelled, to LIRR and Metro North getting into a political fray over Penn slots shows that NJT’s intuition to have its own separate station and trackage into NYC may have been rather wise.

Nyland8 April 3, 2013 - 2:12 pm

” … NJT’s intuition to have its own separate station and trackage into NYC may have been rather wise.”

If one’s definition of “wise” includes being unable or unwilling to work and play well with others; being provincial and territorial; afraid of someday being subsumed into some kind of sane, regional transportation solution; being myopic; … etc; etc;

While we can all certainly sympathize with the frustrations of working with different interstate and federal agencies, that should never mean that any quasi-governmental entity as big as NJT should simply give up.

Bolwerk April 3, 2013 - 2:39 pm

I’m not sure the inability to play nice was entirely NJT’s fault. MTA sub-agencies had a lot to do with it too, and so did NYC and NYS.

A tunnel to GCT would have been much cheaper, afterall. And I think helluva more useful.

Nathanael April 3, 2013 - 6:43 pm

By all accounts, LIRR is particularly awful about “playing with others”.

Anyone who can’t cooperate with Amtrak is incompetent, given that Amtrak cooperates with so many different railroads I can’t count, so I am inclined not to blame Amtrak. This, however, points a fair amount of blame back at NJT.

The tunnel from Penn to GCT Lower Level should have been done. It was abandoned, *according to the study*, by fear of real estate interests. I am not kidding.

Bolwerk April 3, 2013 - 10:45 pm

MNRR takes a big dump on Amtrak. They didn’t want to allow silly things like tilting on the Acela trainsets – which would help on the curves in Connecticut. IIRC, MNRR didn’t want to share GCT with NJT either. LIRR might be the worst, but none of them seem to play nice with each other.

Sometimes I wish Amtrak could(?)/would throw its weight around more.

Walter April 3, 2013 - 11:36 pm

Metro-North only disallows the Acela tilting because the track centers on the New Haven Line are too close together. A tilting Acela would hit a train on the adjacent track.

MNR and Amtrak seem to have a neutral relationship, where Amtrak would love higher speeds on the Hudson and more slots on the New Haven, but where MNR works with them on many projects, like replacing century-old catenary and rebuilding Shell Interlocking to simplify Amtrak moves to/from Hell Gate.

Bolwerk April 4, 2013 - 1:43 am

I got the impression that MNRR didn’t want to accept the improvements that might change that. But it could have been CT’s fault.

Anyway, it can probably be at least mitigated with proper scheduling.

Walter April 4, 2013 - 1:19 pm

The ROW is largely there, but to widen track centers on the New Haven Line the hundreds of old New Haven Railroad catenary supports would have to be replaced, and the overpasses over the line widened. CDOT and MNR probably decided it’s more trouble than it’s worth.

A better solution would be to upgrade the tracks to allow better conventional speeds, or at least get them back to where they were in the New Haven days. Currently the highest speed allowed is 90mph east of New Rochelle, with 70-75 being typical. There’s no reason straight pieces of rail (say New Haven to Milford or Larchmont to Harrison) can’t be 100mph.

Amtrak should look into bypasses of the worst curves, with the viaduct in Bridgeport being the worst, followed by the slow zones caused by the movable bridges in Cos Cob and Norwalk. Those bridges (and the one in Westport) all have to be replaced anyway.

Added up, Amtrak could shave maybe 10-15 minutes off it’s time on MNR territory, pretty low return for what are huge capital costs.

Alon Levy April 5, 2013 - 1:17 am

Metro-North doesn’t even allow the Acela to do 5″ cant deficiency without tilting, which the Regional does all the time east of New Haven.

Eric April 3, 2013 - 5:02 pm

The problem is not that bad projects sometimes fail, it’s that good projects are never even considered. If nothing changes, the next expansion plan will be as stupid as ARC was.

Henry April 3, 2013 - 8:55 pm

A good project would question why we’re trying to expand rail capacity to a station location that is not convenient to most riders’ destinations without a subway transfer.

A better project would question the wisdom of dumping everyone from New Jersey and Long Island into a single station.

Nyland8 April 4, 2013 - 12:14 am

I’ll second both of those, Henry

Alon Levy April 4, 2013 - 8:10 am

ARC Alt G actually addressed the second question.

AlexB April 4, 2013 - 11:47 am

Does anyone know if it would be possible to extend ESA to New Jersey, allowing NJT to share it with MNRR?

AlexB April 4, 2013 - 11:47 am

I meant, share it with LIRR.

Nathanael April 7, 2013 - 12:43 am

LIRR doesn’t want to share with anyone. They wouldn’t share Penn Station if they could avoid it.

Alon Levy April 5, 2013 - 1:17 am

There’s a TBM stuck at the end of the tunnel because they thought it’d be too expensive to remove it.

Woody April 7, 2013 - 9:53 am

Sorry, this means they’d only need to crank up the TBM and go? … which would be good … OR that the abandoned TBM now blocks the way to any further extension?

Alon Levy April 9, 2013 - 12:10 am

It blocks further construction.

AlexB April 10, 2013 - 11:44 am

OK, if it were removed, could ESA be extended to connect to the existing Northeast Corridor and used as a terminal station for NJT and LIRR?

lawhawk April 3, 2013 - 2:12 pm

In related news, seems that the Municipal Arts Society is sponsoring a design contest to reimagine Penn Station.

In a design competition announced today in the Times, MAS is challenging Santiago Calatrava, Diller Scofidio + Renfro, SHoP Architects, and SOM to rethink the much-maligned train station and the arena that sits atop it, Madison Square Garden. “We’re really trying to unlock people’s imaginations about the very real potential of a new arena and of a new Penn Station,” MAS President Vin Cipolla told the Times.

What makes this competition so intriguing is that there’s a chance something could come out of it. Madison Square Garden’s permit recently expired, and local officials (and archicritics) are asking for it to only be renewed for 10 years so the city has time to come up with an option that would improve the arena and the dungeon-like Penn Station. The whole permit thing is quite a mess, and there are obvious reasons for not tearing down MSG, including a recent billion dollar renovation, but the groups like the Municipal Art Society and the Regional Plan Association are pushing the city to “take the time to get it right.” Can a design competition help make that happen? The four designs will be unveiled at a conference on May 29.

If Calatrava’s involved, it’s going to mean a design that looks quite startling, get the fawning attention of critics, but whose cost will easily be more than twice what the initial estimates suggest.

marv April 3, 2013 - 9:45 pm

When/what hours are 3rd/4th tracks under the hudson needed? Is it 7am-10am and the PM rush or is it all day. How do weekends figure in?

Could one tube of the Lincoln Tunnel be converted for rush hour rail and cars the rest of the time? (by limiting the tube to just cars, there would not be a problem of trucks brushing/destroying the cantinary.)

Could a 42nd street light rail – street running across manhattan but express from the hudson to secaucus (or Newark)and the East River to Jamaica offset the non-direct ride by eliminating the time spent getting out of Penn Station and delivering users closer to their destinations?

Could rail (subway or commuter) on the George Washington Bridge take pressure off of Penn Station?

In short, building a dead end hole in the ground seems like an expensive solution to the capacity issues. Perhaps there are other solutions (out of the box and hole) that can do the job at a far lower cost thus not killing other equally needed projects. (A freight rail tunnel, a rail connection to Staten Island, a LGA link etc.)

Frank B April 4, 2013 - 1:10 am

I say that New York, which despite its political ineptitude still manages to take action, has made all the right moves here.

There are several routes going from both Pennsylvania Station and Grand Central Terminal heading upstate; Metro North train service is actually now being increased to meet additional demand.

Following East Side Access, Long Islanders will be able to switch lines at Jamaica for whatever train they’d like; West Side, 33rd Street, or East Side, 42nd Street.

6 tracks will be available for the Long Island Railroad; 2 of them (In the 63rd Street Tunnel) will be dedicated exclusively for such use, 4 will be shared with Amtrak.

And more and more of our neighborhoods in the 5 boroughs themselves are having new life and investment drawn into them. And we’re building subway lines again. We’ll eventually have a full IND Second Avenue Line, no matter how long it will take; its a necessity.

Meanwhile, NJ Transit is limited to two century-old tunnels that they don’t even own.

NJ Transit will wake up one day to find their access into Pennsylvania reduced by Amtrak once Amtrak realizes they’re making comparatively more money on the Northeast Corridor (as ridership continues to increase for the years to come) and that its ultimately not cost-effective to have their most profitable line and limited tunnel capacity shared with NJ Transit. (Which stuck its foot in it by cancelling its own tunnels; like it or not, Macy’s Basement was in Manhattan, where it needed to be, and not Jersey.)

NJ Transit has made their bed. Their poor decisions will cripple their state economically for years to come. New York has the market cornered on the suburban way of life for the next 20 years or so; people will move upstate and east; they’ll be no access troubles then.

Let’s build up New York! We once built the tallest building in the world in a single year! We are the EMPIRE STATE. Let’s build up our city to keep it the greatest city in the world! Rezone! Repair! Rebuild!

You know what I say?

Let’s build the Triboro RX!

Let’s tunnel to Staten Island!

Let’s restore the Rockaway Line!

Let New Jersey Eat Cake.


Bolwerk April 4, 2013 - 1:47 am

I agree it’s good that New York at least pulls off some projects, but it’s not getting much bang for the money it spedns.

Jersey could feast on feces for all I care, but you can’t just ignore them. And you can’t dismiss the idea that Gateway is good inspite of New Jersey; Amtrak wants it for HSR, but it’s probably the only way to get more capacity either way.

Joseph Steindam April 4, 2013 - 7:09 am

I think it’s a bit silly to suggest that NJ Transit didn’t want the ARC Tunnel. You’re ignoring the major player in the story, Chris Christie, who was looking for a way to taking federal dollars and shore up the State Transportation Trust Fund (which needed money to pay for the Turnpike widening). In a state with arguably the strongest executive authority, Christie was able to kill the project with limited oversight from anyone, including NJ Transit or their riders.

Nathanael April 7, 2013 - 12:44 am

If you want to build up NY, please for God’s sake build the HSR line along the Empire Corridor so I can get from Syracuse to NYC quickly. Thank you.

marv April 4, 2013 - 8:14 am

“penn station is used to park trains”.

unless capacity exists this is crazy (building a tunnel so that this continues makes midtown car parking look cheap.)

yes … a train station is the middle of expensive real estate is for the purpose of pickup and discharging people.

train yards, etc should be where land is cheaper even when it means movements…. even better is when those movements serve passengers as well.

Kennedy airport is basically a bulb surrounded by terminals. By elevating all parking, a train yard/terminal could be built. Acqaduct could better serve ny as train yard than a slot machine location.

This would:
*take pressure off penn station
*take pressure off jfk as shorter flights would now use rail
*take pressure off the roads as rail options would be more desirable

Anto April 4, 2013 - 2:05 pm

Don’t be surprised if LIRR has through-running trains from Jamaica to Secaucus during the week of the Super Bowl. If you don’t believe it’s a possibility, view the MTA Board – Long Island Rail Road Committee Meeting – 3/11/2013 on youtube. Make sure to listen to the LIRR’s president comment about through-running trains.

Ryan April 4, 2013 - 3:38 pm

For reasons discussed above, it’s not going to be LIRR who operates the through running trains. It’s going to be NJT.

Anto April 4, 2013 - 4:27 pm

Absolutely! It will be awesome to board an ALP-45-DP at Jamaica to Secaucus.

Ryan April 4, 2013 - 7:36 pm

Agreed. Though, I wonder if LIRR and NJT would be able to work it out, what with the LIRR using 750V DC third rail, and NJT using 25 kV 60 Hz, 12.5 kV 60 Hz, 12kV 25 Hz overhead catenary.

Anto April 4, 2013 - 8:07 pm

They should be able to work it out with no problem. Currently the ALP-45-DP run all the way to Sunnyside yard under cat during the day. When the the ALP-45-DP exits the east river tunnel heading east, then the engineer would power up the diesel engine for the trip to Jamaica. There is no need for third rail shoes over or under.

Alon Levy April 5, 2013 - 1:22 am

Awesome! Diesel on an electrified line! A 33 t axle load locomotive on a line that mostly sees EMUs with a 14 t axle load!

Nathanael April 7, 2013 - 12:45 am

What can you do when nobody is willing to re-electrify? But yeah, this is what is going to happen in the short run, heavy diesel on an electrified line. Why? Because everyone’s been too cheap to string wire.

Anto April 7, 2013 - 9:05 pm

Currently the LIRR operates several DM30AC dual mode locomotives on Port Jefferson, Oyster Bay, and Montauk Branches. The DM30AC weight is 128 t. Locomotives operate in diesel mode till reaches west of Jamaica, then switched over to electric mode.

Alon Levy April 9, 2013 - 12:09 am

Yeah, that’s pretty stupid, too. The maintenance cost plus the cost of the locos is high; the cost of getting rid of diesels on at least PJ, the busiest diesel lines, isn’t that high.

Leave a Comment