NYTIP – transforming commuter rail, part 1a: the harlem line in nyc

UPDATE (07.06.2021): Post updated to correct errors in the Shell Junction diagrams; this update also includes additional commentary on said junction.

Welcome back to my NYTIP series! In this series of posts, which comprise Volume II of NYTIP – transforming commuter rail into regional rail, I will discuss ideas for transforming the disjoint commuter rail systems in the Tri-State Area into an integrated regional rail network. This post addresses the Harlem line – a branch of MTA’s Metro-North Railroad – in The Bronx.

Note: Click any image to enlarge.

Several commuter rail lines serve the Tri-State Area (New York, New Jersey, and Connecticut). In New York alone, MTA’s Metro-North Railroad (MNR) and Long Island Rail Road (LIRR), as well as NJ Transit, provide service; the latter two share a terminal at Penn Station, while MNR serves Grand Central Terminal. Despite the relatively robust coverage provided by each of these railroads, they operate entirely separate services with differing fare regimes and virtually no interconnectivity between them.

Making matters worse, most commuter rail stations within NYC only see limited service – half hourly or hourly in most cases – and the fares are prohibitively expensive compared to local buses and subways. Compared to a $2.75 local bus or subway fare, trips to Harlem – 125th Street or Grand Central from anywhere in The Bronx cost $7.25 off-peak and $9.75 peak. These fares are prohibitive for most Bronxites, given The Bronx’s low median household income – approximately $40,000 per year in 2019 – and high poverty rate. (The off-peak fare is even higher than MTA’s express bus fare of $6.75!) While intraborough fares are only $3.00, service is sparse at all but a few stations.

A similar situation exists on the LIRR, which serves Penn Station in Manhattan and several stations in Brooklyn and Queens. Trips to Penn Station from Queens stations west of Jamaica cost $6.50 off-peak and $9.00 peak; from Jamaica and eastern Queens stations, this increases to $7.75 off-peak and $10.75 peak.

The only attempts to address these obscene fares are the weekend CityTicket, the LIRR’s Atlantic Ticket, and lower fares for certain intraborough trips. Needless to say, these efforts aren’t nearly enough.

There are thus two key problems to address when it comes to NY’s commuter rail infrastructure – infrequent service and inequitable fare regimes. The Regional Plan Association (RPA), in their T-REX report, has the right idea when they suggest consolidating fare zones and substantially reducing fares within the inner urban core. Indeed, if there is any hope of using commuter rail to relieve overcrowded subways and encourage transit trips, it begins with lower fares. A comprehensive report by embattled NYC Comptroller and mayoral candidate Scott Stringer suggests lowering rail fares to match the current $2.75 bus and subway fare for travel within city limits, an excellent starting point.

However, lower fares mean nothing without better service. Both Stringer’s report and the aforementioned T-REX report state as much; while the latter presents a bold vision for such improvement, it actually takes much less to set the table.

On a recent INSIDER post, I stated that regional rail fares and service levels should be predictable. To that end, NYTIP prescribes the following:

  • A flat fare of $3.00 for travel within the urban core (generally all of NYC and possibly some inner suburbs), which matches the current single-ride subway and bus fare,
  • Eliminating peak fares,
  • OMNY integration to allow multimodal transfers, and
  • Baseline service every 10 minutes during peak hours and every 15 minutes during off-peak hours, per RPA’s recommendation, with more service on trunk lines.
    • Baseline service should increase as demand warrants.
    • Generally, local trains serve the urban core and express trains serve the suburbs and exurbs, with limited stops within the urban core.
    • Peak-only service could overlay this baseline, subject to capacity constraints.

Note that this does not preclude fare reform beyond the urban core.

In addition to being predictable, regional rail must be accessible. In addition to ADA, there are several “low-hanging fruit” opportunities to increase both system capacity and reach. As such, NYTIP prescribes a four-step approach to make regional rail accessible:

  • Expand existing stations,
  • Construct infill stations,
  • Remove bottlenecks, where feasible, and
  • Promote network and fare integration.

Let’s begin with MNR.

MNR Harlem Line

At present, MNR’s Harlem line serves seven stations in The Bronx – Wakefield, Woodlawn, Williams Bridge, Botanical Garden, Fordham, Tremont, and Melrose. The New Haven line joins the Harlem line north of Woodlawn, but generally runs express. Aside from Fordham, few trains serve The Bronx. Moreover, most of these stations are not ADA-accessible and have short platforms – in particular, the Tremont and Melrose stations can only platform two cars! The easiest way to encourage ridership at these stations, aside from rationalizing fares, is to lengthen the platforms and enhance accessibility. The figures below illustrate my ideas for doing just that:

[Figs. 1, 2, 3, 4] Station enhancements at local stops on the Harlem line.

I didn’t include Botanical Garden because it is already an ADA-accessible station capable of platforming 8-car trains. I also excluded Tremont for reasons I’ll explain later.

Now, the Fordham station is the only Bronx station served by both Harlem line and New Haven line trains – though the latter only dropped off southbound and picked up northbound until MTA axed this rather nonsensical 19th-century rule in 2019. It can platform 8-car trains, like Botanical Garden station to the north, and is also ADA-accessible. As Fordham is one of the busiest local stops on MNR and the busiest reverse-commuting station in the nation, an opportunity exists to transform the station into a regional hub.

Potential capital investment: Convert the Fordham station to an express stop.

[Fig. 5] Express stop conversion at Fordham station. (Unmodified track map by vanshnookenraggen.)

Express stop conversion at Fordham – along with extra-long platforms – serves several purposes. The first is to allow major service increases without constraining capacity. The second is to allow faster trips from Fordham to both Manhattan and the suburbs (mainly the latter since New Haven line trains already operate express to Grand Central). The third is to provision for future intercity rail (i.e. Amtrak) service, made possible by new connections that I will discuss in follow-up posts. This requires careful modifications to the historic stationhouse, as well as right-of-way (ROW) widening under Fordham Plaza. Note, however, that this conversion is not a prerequisite to regional rail.

To “jump start” regional rail, local trains could initially run with 6-car (510-foot) trains. Then, as ridership grows and more rolling stock becomes available, longer trains of at least 8 cars (680 feet) could run.

A Note on Interlining

The Harlem line, as a four-track trunk line, sees local and express service from both the Harlem and New Haven lines. Generally, locals from each line take the local tracks, and expresses from each line take the express track; however, the New Haven line locals only stop at Fordham, and thus run express on the local tracks in The Bronx. Another option is to pursue a form of de-interlining, where all Harlem line trains run on the local tracks and all New Haven line trains run on the express tracks in The Bronx. However, this amounts to a partial de-interlining at best, since all MNR lines interline through the four-track Park Avenue tunnel.

As of 06.04.2021, I’m still leaning toward keeping Harlem and New Haven line locals on the local track, and expresses on the express track. This allows for predictable service patterns and easier reverse-commuting from the Harlem line.

Infill Stations

The Harlem line contains several abandoned station sites. Stations once existed at 183rd Street, Claremont Park, Morrisania, and 138th Street. Incidentally, reactivating some of these stations would close a large gap in rail service between the Concourse line and the IRT White Plains Road lines. Recently, vanshnookenraggen wrote a post arguing that it wasn’t necessary to build a 3rd Avenue branch of SAS – among other possible subway lines – because regional rail could close the gap at a fraction of the cost.

While I’ve always had a “soft spot” for a 3rd Avenue branch of SAS, I’m open to having MNR’s Harlem line fill this gap through infill stations. As these would be new stations, they’d be at least 700 feet long for 8-car trains. Let’s examine potential infill station locations.

Belmont (183rd Street)

[Fig. 6] Overview of the Belmont station.

The Belmont station’s main purpose is to close the gap between Fordham and Tremont; however, the resulting interstation distances are less than 1/2 mile. Moreover, while the surrounding area is relatively dense, the desired catchment area is several blocks east of the station site (namely, St. Barnabas Hospital on 3rd Avenue and the famous Arthur Avenue corridor – the spine of The Bronx’s Little Italy).

This brings me to Tremont station. The Cross-Bronx Expressway destroyed a good chunk of Tremont’s catchment area; its relative proximity to the Tremont Avenue station – an express stop on the Concourse line less than 1/2 mile west – also limits its draw. Consequently, I gave thought to another idea proposed by RPA – relocating the Tremont station northward to 180th Street:

[Fig. 7] Overview of the Belmont – Tremont station.

The Belmont – Tremont station’s south end is less than 1/4 mile from the existing Tremont station, effectively serving a similar catchment area. It would still connect to crosstown bus service, given current Bx36 bus service and the proposed Bx40/42 reroute via 180th Street in the Bronx Bus Redesign. As of 06.04.2021, I am leaning on replacing Tremont station with the new Belmont – Tremont station.

Claremont and Morrisania

When I first drafted this post, I presented two options for serving the central Bronx – either two stations at the original Claremont and Morrisania station sites, or one station halfway between them at Claremont Village. Initially, I chose the latter:

[Fig. 8] Overview of the Claremont Village station.

I liked this station site because it has a strong catchment area – dense housing and several schools. However, given the Tremont station relocation, I’d like to revisit the two-station option:

[Figs. 9, 10] Overview of the Claremont and Morrisania stations.

The Claremont station serves the north side of Claremont Village, several businesses, and the Bathgate Industrial Park; it also connects to the crosstown Bx11 bus. The Morrisania station serves the south side of Claremont Village and connects to the crosstown Bx35 bus.

As of 06.04.2021, I am now leaning on building both the Claremont and Morrisania stations, given connections to existing transit services and good catchment areas.

Mott Haven

The 149th Street – Grand Concourse subway station on the 2, 4, and 5 lines contains signage directing riders to a proposed MNR station that was never built. A station in this area would not only connect to these subways, but to several bus routes, the Hostos Community College, Lincoln Hospital, and other points of interest. Just south of 149th Street, the Hudson line joins the mainline at grade. This is a significant bottleneck since northbound Hudson line trains foul the entire mainline. Given this constraint, an infill station here would be a major capital investment – more so than the others.

When I first drafted this post, I presented three options – two on 149th Street and one on 138th Street. I initially chose Option 2 – a station at 149th Street served only by Harlem and New Haven line trains to facilitate grade separation of the Hudson line:

[Fig. 11] Overview of Mott Haven Option 2.

However, given that the Hudson line serves Yankee Stadium and the West Bronx, it could provide some relief for the oft-crowded 4 train. Since building the Hudson line platform means retaining the flat junction, I thought of ways to improve it. Here is what I came up with:

[Figs. 12, 13] Overview of Mott Haven Option 4.

Option 4 enhances Mott Haven Junction with two new switches – one for additional flexibility, and another on the Hudson line to allow trains serving Mott Haven to access the northbound track without fouling southbound service. South of Mott Haven Junction, this option proposes the following train movements:

  • Track 1: Southbound flow
    • All southbound Hudson line traffic
    • Harlem and New Haven line locals
  • Track 2: Northbound flow
    • All northbound Hudson line traffic
    • Harlem and New Haven line expresses
  • Track 3: Southbound flow
    • Harlem and New Haven line (primarily for expresses, secondarily for locals)
  • Track 4: Northbound flow
    • Harlem and New Haven line (primarily for locals, secondarily for expresses)

This paradigm does not preclude other train movements, as needed, to preserve regularity. There would still be some conflicts – mainly between northbound Hudson line and southbound Harlem and New Haven line locals, and between some Harlem and New Haven line expresses – but the fouled segments are shorter than at present.

I’ve gone back and forth between this option and Option 2. Grade separation is a worthy goal, but the more I think about it, the more I’m leaning toward establishing a hub for all east-of-Hudson MNR lines at Mott Haven. Giving the South Bronx access to these MNR lines is too good of an opportunity to pass up in my view, even if such requires an imperfect junction. Therefore, as of 06.15.2021, I am now leaning on Option 4 for the Mott Haven station. Penn Station Access (PSA) factors into my thinking. PSA Phase 1 sends some New Haven line trains to Penn Station from New Rochelle, which means these trains never interact with the Harlem line. PSA Phase 2, formerly West Side Access, would send some Hudson line trains to Penn Station, relieving the Mott Haven junction. In my next post, I’ll discuss potential capital investments that can optimize PSA for regional rail.

Eliminating Bottlenecks

The Mott Haven Junction isn’t the only bottleneck to address on MNR. I discuss two additional bottlenecks below.

Bottleneck: Woodlawn Junction

[Fig. 14] Overview of Woodlawn Junction.

MNR’s Harlem and New Haven lines meet at Woodlawn Junction. It’s mostly grade-separated southbound, but flat northbound. Consequently, there are several merging conflicts – mainly due to the way New Haven line trains merge into the Harlem line; Figure 14 illustrates these conflicts.

As with Mott Haven Junction, a capital option exists to remedy these issues.

Potential capital investment: full grade separation at Woodlawn Junction.

[Figs. 15, 16] Woodlawn Junction grade separation.

Full grade separation at Woodlawn Junction requires a modification to just one track – the northbound express track. Beginning at a point south of 224th Street, the track runs downgrade at about 1%. This ensures the new track clears the Bronx River, which is very shallow in this area. A new high-speed switch provides direct connectivity to the New Haven line’s express track, minimizing conflicts; Google Earth measurements suggest northbound express trains can emerge at a 1% grade to meet existing tracks. For operational flexibility, this project retains part of the existing northbound express track as a stub. This is important because there are still a few unavoidable conflicts. Southbound New Haven line express trains either merge with the locals briefly, or foul the northbound Harlem line local track briefly while using the stub to rejoin the express track. In either case, there are fewer conflicts than before; when combined with PSA, the express/local merge is even less of an issue.

Bottleneck: Shell Junction

[Fig. 17] Overview of Shell Junction.

Shell Junction, located outside of NYC near New Rochelle station, is a key choke point on the New Haven line. Delays in this area often trickle down to and through the Harlem line. This single junction limits intercity (Amtrak) service between New York and Boston, and also constrains capacity on the New Haven line. Given both PSA and the FRA’s plans for high-speed rail (HSR), fixing this junction is crucial – so much so, that it’s part of NEC Future (see slide 24). However, grade separation at Shell Junction is a long-term investment that would likely take years to realize. As such, I thought of a short-term fix that could allow the capacity increases needed to enhance PSA, while still lending itself to grade separation later.

Short-term capital investment: Shell Junction Improvement, Phase 1.

[Fig. 18] Shell Junction Improvement, Phase 1. This phase includes station improvements at New Rochelle.

At New Rochelle station, a new platform built in the space between the express tracks serves Amtrak and PSA trains. This platform would be 850 feet long for 10-car trains. The existing platforms would extend to 850 feet so that they can also accommodate 10-car trains; the southbound platform extends an additional distance to replace the west end. This would yield the necessary space for new switches to support full grade separation.

At Shell Junction, the switches connect to the center tracks. While the junction is still flat at this phase, southbound trains traversing the Hell Gate line only foul the mainline at the switch; currently, these trains foul the mainline from a point east of New Rochelle to the junction. An existing interlocking east of New Rochelle can accommodate PSA locals.

Long-term capital investment: Shell Junction Improvement, Phase 2.

[Fig. 19] Shell Junction Improvement, Phase 2.

Grade separation at Shell Junction ensures that both the New Haven and Hell Gate lines can accommodate regional rail (local and express), intercity, and HSR services with minimal conflict. Due to space constraints near Shell Junction – namely, Webster Avenue to the west and New Rochelle station to the east – grade separation requires right-of-way (ROW) widening and a relatively steep gradient. Google Earth measurements suggest that a 1.9% approach gradient – which is the ruling gradient on the Northeast Corridor – should be feasible, if a bit tight. Consequently, at least part of the existing flat junction would remain for freight moves.

(On Transit Twitter some time ago, I learned that grade separation at Shell Junction has been studied for decades, with no success. This underscores the challenges involved, though I still believe such is worthwhile.)

Now, current PSA plans expand the Hell Gate line from two tracks to four, but only in The Bronx. Under NYTIP, the four-track segment extends to a point west of Shell Junction (not shown in Figure 19) to separate local trains from express and intercity trains; indeed, the FY21-FY25 Northeast Corridor Commission plan proposes a total of four tracks over the replacement Pelham Bay bridge east of Co-op City.


The first phase of regional rail under NYTIP consists of several key elements. The first is predictability – manifested through fare parity and service scheduled at regular, frequent intervals. The next is accessibility – manifested through station improvements (including ADA accessibility), infill stations, and eliminating bottlenecks. This post mainly addressed the Harlem line in The Bronx. In Part 1b of this post, I’ll address regional rail on MNR’s Hudson and New Haven lines (including PSA), and Triboro RX provisioning. Until next time!

4 thoughts on “NYTIP – transforming commuter rail, part 1a: the harlem line in nyc

  1. Awesome! And exciting! I look forward to upcoming parts.

    Mott Haven tho. Is there a rule that northbound trains have to use the rightmost tracks and southbound trains use the leftmost tracks?

    What if those were partially deinterlined? Hudson trains only use the left two tracks, and Harlem + New Haven trains only use the right two tracks. That way you can do Option 1. Both would fit the two island platforms at Harlem-125 nicely, and I believe both have equal access points to the upper and lower levels at Grand Central Terminal.

    Also, you mentioned provisions for the Manhattan Spine; where would those be?

    Liked by 1 person

    1. Hey Sam:

      I hope you’re as excited about reading as I am about blogging it! Part 1b is already in the works!

      To your questions:
      1) Manhattan Spine provisions: on Mott Haven Option 1, I reroute the S/B Harlem line local track to create space for a platform. Provision for the southbound Spine would be immediately north of the S/B plat at Mott Haven. Haven’t figured the exact connection from N/B. Provisions from the Hudson line are out of view.

      Mott Haven Option 2 would include provisions for the Spine as well, but at different locations – likely further north of the Option 1 provisions. I don’t have an exact location picked out yet. In Part 1b, you’ll see that Spine provisions from the Hudson line could potentially feed Triboro RX instead; honestly, I’d probably send Harlem/New Haven line trains thru the Spine rather than Hudson line trains if I had to choose a branch.

      Now, regarding your partial de-interlining idea – it’s a helluva good one. I didn’t even think of that! The only issue I could see from such is heavier demand on the Harlem and New Haven – not to mention more total trains at present – than Hudson. Penn Station Access (and later, the Spine) may provide the necessary relief, but for all that, you can potentially squeeze more total TPH now with the grade separation.

      Liked by 1 person

  2. Really interesting read.
    What are you envisioning for service patterns to be in terms of terminals? Would the locals continue on to Stamford and White Plains? Same for Hudson would they turn earlier then Croton Harmon? I had the idea recently of filling in service that would be more subway like, only out to Yonkers, Mount Vernon West and New Rochelle. Turning the trains might be tough without dedicated terminals though.


    1. Hi Peter:

      With the existing network, locals would likely use their current terminals (i.e. Croton-Harmon, North White Plains, and Stamford). However, with thru-running, it may be wise to shorten the intra-city locals. Incidentally, MTA is expanding New Rochelle yard as part of PSA for short turns (but this is mainly due to the flat junction at Shell Interlocking). Other “mini-yards” near Yonkers and Mount Vernon West could be upgraded to handle short turns, but these have the same problems New Rochelle does – at-grade crossings.

      I’m also considering “intermediate” termini (i.e. Tarrytown, Crestwood, Port Chester) for the locals, but haven’t finalized my choice. Either way, you need some type of grade-separated turning infra.


Discuss respectfully. No trolling or threats allowed - violators subject to moderation or ban. Thanks!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s