Regular readers of this blog may have noticed a lack of new content in a long time. This has not been because of a lack of work on my part but because the new work has been taking so long. Over the last year and a half I have been working with Uday Schultz on an in depth report on what a deinterlined NYC subway system would look like as a precursor to the next futureNYC Subway proposal. At the beginning of the Covid-19 shutdown I was able to get much of the report written but with ridership at a third of pre-Covid levels and a long recovery ahead of us I felt that the timing was not right for such a report (nor, for that matter, a proposal to expand the subway.) When light begins to shine at the end of this tunnel I will revisit releasing both the report and next future subway vision. But for now I felt that a specific proposal needed to be brought up.
The Port Authority of New York and New Jersey have been pushing the Federal government to fund their proposed AirTrain to LaGuardia Airport. This year the Draft Environmental Impact Statement was released and the comments were mostly in favor due to the support of Governor Andrew Cuomo and labor groups. There has been some push back from local residents and online transit groups but I hazard to guess most New Yorkers don’t even know this is happening.
In a post which is seeing the rounds again I speculated that by extending the Astoria Line (presently the N/W but as proposed the R train) to LaGuardia Airport instead of the AirTrain the Broadway Line (N/Q/R/W) could be see capacity expanded and service delays eliminated. While this proposal still stands I have become more convinced that the chips are stacked in the favor of the AirTrain. Governor Cuomo knows how to pull political leavers better than anyone and he knows that the FAA will fund a people mover to the airport but not actual transit. As such the AirTrain has been preordained while a more useful subway extension was never a real option. Where, then, does this leave fixing the bottlenecks of the Broadway Line?
For those who are new to the issue: the Broadway Line has a series of extremely limiting bottlenecks which reduce overall system capacity and are the cause of daily delays:
- N trains merge from the express to local track between 42nd St and 34th St. This merge limits service to Astoria and Forest Hills
- In Brooklyn at DeKalb Interlocking the N/Q and B/D trains all mix to serve different branches. The track layout at this interlocking causes all trains to slow down to the point that many trains have to stop on the Manhattan Bridge and wait. Reverse branching also causes delays to ripple down multiple subway lines.
- The R train is forced to use the Queens Blvd Line (and merge with the N/W in the 60th St Tunnel) due to a lack of direct yard access in Brooklyn.
The gist of my original R train proposal was that by extending the Astoria Line to LaGuardia Airport a new yard could be built along the way in an area which is now just a ConEd parking lot. This would eliminate the need for most of the merges on the Broadway Line and allow the streamlined local and express services to run more trains. At the same time I proposed deinterlining the DeKalb Interlocking so that B/D trains would run along the Brighton Line and the N/Q trains would run along 4th Ave. Now, assuming no new yard will be built in Astoria, we must find another way to detangle these lines. Fortunately a new way forward is not only simpler but cheaper.
The DeKalb Interlocking deinterlining shall remain as it is probably the worst choke point in the entire B Division of the subway. In this new proposal N/Q trains will run on the Brighton Beach Line and B/D trains the 4th Ave express. The major change is that now all 4th Ave local trains, the R, will be rerouted along the West End Line, currently serviced by the D train. W trains will be extended from Whitehall to 36th St and along the West End Line to Bay Parkway.
Just south of 36th St a new set of switches from the local to express tracks will be built to allow the B train to switch to local from 45th St to 95th St in Bay Ridge. The D will take over the N along the Sea Beach Line to Coney Island. The B, now serving the Bay Ridge Line, will have direct yard access to the Concourse Yard in the Bronx and the R/W will now use the Coney Island shops.
Early subway builders designed their systems to have as much flexibility in running trains as possible (and, unfortunately, as cheap as possible). The tracks features numerous switched so that trains can freely move around the system. This is great for emergency changes and maintenance but bad for everyday service as in the case of the N switching between express and local near Times Sq. In many places builders placed holes in the tunnel walls incase the need for additional switched were ever needed.
Along the 4th Ave Line in Brooklyn space was left south of the 36th St station for three full switches on all tracks (*53a in the map above.) While I have not found any official reasoning for these provisions, it is possible that they were designed if 36th St had to be used for a temporary terminal or if stations further south needed different levels of service (especially if the proposed Staten Island Tunnel had been built).
Astute readers will see that adding new switches will add a new merge. But this merge is limited to one line so delays won’t ripple throughout the system the way they do with current reverse branched lines. This merge can be factored into the running times of the B and D trains better than trying to work out merges between the N and D and B and Q trains.
To achieve optimal performance more deinterlining will be needed on the Broadway and 6th Ave Lines. The N will be rerouted to 96th St-2nd Ave with the Q and could even be eliminated for a <Q> service instead. The A and B trains will swap along Central Park West so that B/D trains will stay on the express track the entire way between 145th St and 36th St in Brooklyn. R trains will move to Astoria along with the W.
In the final deinterlining report these additional changes and more will be explained in depth. For now the focus is just on DeKalb and 4th Ave.
Local vs Express
The big winners of this switch will be Sunset Park and Bay Ridge riders who will see more service and a quicker ride on the express tracks. Sea Beach riders will see slightly more service as well. The most obvious problem to this plan is that West End Line riders will go from having an express train to a local. But there is more to a trip than just which train goes where. Total travel time isn’t just how long the train takes but also how long a rider must wait for a train, or the frequency.
D trains run every 6 minutes and take 32 minutes from Stillwell Av to Atlantic Av via the 4th Ave express tracks for a trip total of 38 minutes. R trains run every 6 minuets at peak but usually every 10 during the day and take 3.5 minutes longer than express trains between 36th St and Atlantic Ave.
With a much higher frequency the overall travel time is the same. By increasing service to every 3 minutes a West End R train via the 4th Ave local tracks will get a rider to Atlantic Av in the same 38 minutes. While not necessarily making the trip faster it will not make the overall trip any slower.
Additionally, the West End express track can be used for peak rush hour service which will shave 12 minutes off travel between Stillwell Av and 9th Av. Assuming 10 minute headways for peak express trains a rider can still get to Atlantic Ave in 38 minutes.
Service Increase
The Bay Ridge Line will see an increase of up to 50% over existing levels of service and riders will save 5 minutes each way to Atlantic Ave:
- Present: R trains at 10tph, 28 minutes local 95th St to Atlantic Ave
- Proposed: B trains at 15tph, 23 minutes express 95th St to Atlantic Av
The Brighton Beach Line will see and increase of up to 66.6% over existing levels of service:
- Present: B and Q trains at a combined 18tph
- Proposed N and Q trains at a combined 30tph
The Sea Beach Line will see and increase of up to 66.6% over existing levels of service:
- Present: N trains at 9tph
- Proposed: D trains at 15tph
The West End Line will see and increase of up to 80% over existing levels of service:
- Present: D trains at 10tph, 12 minutes express 36th St to Atlantic Av
- Proposed: R trains at 18tph, 15 minutes local 36th St to Atlantic Av.
Building a new switch instead of a full subway extension and yard would have the highest return on investment of possibly any transit investment. Though with MTA costs still out of control it’s hard to know just how much this will cost. Proposed new switches at Rogers Junction on the IRT Eastern Parkway Line have seen cost estimates between $300m and $1.6b. But there no switch provisions exist. On 4th Ave the space exists and will require far less construction.
The new routes would change how riders will have to think about commuting but the benefits far out weigh any short term inconvenience. Deinterlining the Brooklyn branches will also allow the network to better respond to demand changes and will cut down on delays. The service levels proposed above are the maximum possible in the current system and can be easily dialed down if needed. CBTC signals will allow more service when installed but aren’t required for this switch.
The MTA and NYCT are using the current downturn in ridership to work on maintenance backlogs and get work done quicker. Recently a rebuild of the E train tracks along the Archer Ave Line was finished ahead of schedule and work on the Rutgers St Tunnel (damaged by Superstorm Sandy) is moving along quickly. Installing a new switch isn’t as simple as placing new tracks but it is far more simple than planning and building a new line. A post-Covid world will see fewer riders going to Manhattan but more within their own borough and more throughout the day. The system we need for tomorrow is not the one we’ve inherited. Now is the time to start thinking about how to stabilize and turn the ship.
What are the differences between this, and just running the B and D on the Sea Beach/West End, and the R and W to Bay Ridge?
Also, what would you planned for overnight services (if they were to still exist after this)?
If you run the R and W to Bay Ridge only then they don’t have direct access to a maintenance yard. That’s why, currently, the R goes to Forest Hills (and the W comes from Sea Beach). The original idea I put out there built a new yard in Astoria but this plan simplifies things so you don’t have to.
Overnight is a good question. B could do what the R does now, turn at Whitehall, or could get weird and turn at 34th St. Since late night service doesn’t need the kind of frequency that day service needs you can short turn trains in odd places without affecting service as much.
You mention de-interlining CPW, but gloss over some issues maintaining service on the rest of the Eighth Avenue Line.
1. Maintaining express service to Inwood is essential to keep travel time down. Swapping the B/C back to their original service patterns—sending the B express to 207th and the A local to 168th—should be sufficient. The C could take over the B’s current route or run to 168th off-peak, whichever is more efficient for operations. This affects yard assignments, but I expect it’s feasible.
2. The harder one to solve is the merge south of Columbus Circle. There are no switches, or even provisions for switches, from the local at 59th St to the express at 42nd. This pushes the E onto the express to maintain service levels, and until the Cranberry Tunnel gets CBTC, you’d have to do some clever scheduling to get the A (or C) and E to fit. Of course, building switches at 50th St is potentially also worth the investment, but the cost would certainly exceed the switches at 36 St.
What a nice elegant solution. Right now West End has about 5% fewer riders per station than Sea Beach but giving it Broadway service with Union Square and Times Square should turn that around. I’m looking forward to your in-depth workups of the other two bottlenecks.
Couldn’t you achieve similar results without adding a switch?
(B)(D) trains run local on 4th Ave to Bay Ridge and West End respectively from the north tracks on the Manhattan Bridge.
(Q) trains run express on 4th Ave via the Sea Beach Line from the south tracks of the Manhattan Bridge
(N)(R) trains run on the Brighton Line (one express and one local south of Prospect Park) and to the Broadway local via the Montague tunnel.
The (Q) will be entirely deinterlined like the (1)(6)(7) and (L) allowing close to 30 tph.
This also means that DeKalb Ave, unlike in your proposal, is served by both 6th Ave and Broadway trains, so passengers can transfer between the two if they so desire.
I see two big problems with this plan:
D trains would have to run on the 4th Ave local between Atlantic and 36th but still have the same frequency. This just makes trips longer with no benefit or offset.
Now the Broadway local and express will interline again in Brooklyn. The whole point is to eliminate any mixing of lines.
Hi man, happy to see you posting again.
Actually (N) trains would be local in Manhattan as well, and there would be no (W) train. And the (D) in Brooklyn would be the same as the (R)(W) in your plan.
Very interesting post! I didn’t even think of that switch provision while fleshing out NYTIP. Certainly reduces headaches with the no-new-yard-in-Astoria scenario. I like it!
This would also enhance my alternative option for SBK de-interlining (B/N swap): https://nerdynel.me/2019/02/07/nytip102sbk/
I get the rationale for an A/B swap on CPW, but I prefer the C/D swap. This ensures no one on the northern branches loses express service, although you still have some interlining on Concourse.
While the idea is nice, I do have a concern. You do have the R and W along West End to Coney Island. You have the combined frequency of 24 trains per hour, and the concern with that is that the City Hall curve can only handle 21 trains per hour. With this constraint in mind, wouldn’t you be better off just having the R operate at 20 trains per hour between Astoria and Coney Island?
This is a good point and one I was going to go into but clearly didn’t. In order to keep total trip times between on the now local West End Line to 38min (to Atlantic Ave) the least amount of service you need is 18tph. Therefore anything above that is just extra and can be used for the peak express. Since the Astoria Line has more demand it would make sense to use the lower level of City Hall as a short turn location.
Trust me-you do NOT want the A local in Manhattan. It’s the longest route in the system. Uptown residents will have a fit. Train crrews will have a fit (former train crew here). Also, noy every line has to have access to a yard for every trip. W trains go to and from Brooklyn as “put-ins” and “lay-ups” while the rest of the W Fleet usually lays up on the middle track along the Astoria line.now I could go into more details on this but as I read, I saw one thing you seem to have neglected in your planning: terminal space.
You have 5 lines terminating at Coney island. West End trains end on stub ended tracks. Slow entery. Two tracks. Turn around times won’t be quick enough to “empty a pocket”-free up a track. One thing about our subway IS the interlining. Think about this. Let’s look at Astoria. N and W trains leave Astoria. Riders have a choice of either local or express. Same thing with the Brighton line and in the latter’s case, you can access 6th Ave or Broadway with one train. Lines like the B, C and W are known as Supplement lines. They add to the full line lines (the B supplements the D in the Bronx, the C along CPW, add brighton access along 6th Ave and provides express service for the Brighton. The W supps the N in queens and the R along Broadway. Also, they had an extra line alomg fourth Ave decades ago-the M. Cut due to low ridership. Now maybe these days, extending the W to, let’s say 9th Ave on the west end line, but hope that ridership warrents that(along with the 2-4 extra train sets/Crews needed for that extention of service. Can’t send to coney-not enough terminal space. Plus as someone else mentioned, there’s a whole different service pattern overnights. Broadway only sees the N and Q overnights. So both will run down Brighton overnights? That’s a little bit of overkill in service. Trains average 20 min headways overnights. That’s a train every ten mins when there’s two lines running (examples are the 4 and 6 along lex, the N and Q along broadway, etc).
Then as also mentioned here, speed restrictions. Yes, City Hall on the broadway line can only handle so many due to the speed restriction.
The issue with DeKalb isn’t with the interlining or switches, it’s dispatching, priority scheduling and delays. But the service patterns are to cover most frequent needed routings. Dekalb sees two broadways, a 6th Ave line and one from each skip dekalb (both of which are 4th Ave express). But if you wanna see a good example of bad dispatching, spend an hour at Astoria during the rush hour. If you had an N train leave astoria more frequently on time, it will get across the bridge on time, then the D won’t have to wait for it to enter the dekalb bypass before it can itself. Get a better superintendent for the R to make sure it’s allowed to go ahead of another train instead of waiting since that other line’s superintendent did a better job of lobbying for it’s priority over the R. Or how about a better communications system where radios work better so City Hall Tower can confirm you’re a Q train instead of being mistaken for an N causing either a wrong line up or a delay while dekalb resets the switches to the proper routimg for it. The list goes on and on before major routing and infrastructure changes should be made.
You bring up some very important points, especially about ops. What I was afraid of, releasing this as a stand alone post, was speculation and misunderstanding of what I’m proposing because much of the details for other improvements and swapped train lines are still only in the final report (terminal ops and capacity is a big issue I touch on). This new switch is part of a larger networking of the system that will streamline services and allow for additional capacity. Removing merge points will also cut down on delays. I’m a big proponent of Ops>Electronics>Concrete and I talk about this in the report. But defending interlining because “that’s how we’ve always done it” isn’t much of a defense. Obviously right now is not the time to be talking about how to add 40% capacity through more efficient routes. But one day riders will come back. Why do things they way we’ve always done when people are riding the subways differently now (and in the future)?
What I’m proposing here and will propose in the final report is going to change how people will have to think about how they commute. I will argue and work to create a better system with deinterlining but some people just don’t like change until it happens and they see how things are improved.
Can the Brighton Line as signaled today handle three services or would it have to be upgraded to CBTC? Its an interesting idea to keep BMT with the Brighton LIne in Brooklyn and the IND except for the R train as IND on the fourth ave line. Same question on train throughput is for the 4th ave line.
Well these lines handled additional train lines in the past so you don’t need CBTC. But the point here is to have fewer lines with more service on the remaining lines.
I do like what I see here. The plan will have three services running independently through DeKalb. Broadway expresses direct to Brighton lines, 6th Ave expresses direct to 4th Ave express, and Broadway locals through the tunnel to the 4th Ave local. This alone would provide every Broadway line with access to the CI Yard and allow the 6th Ave trains to access the Concourse yard.
In fact most theorists on the web forums (or websites) who propose deinterlining DeKalb propose doing the above or possibly sending the Broadway express to the 4th Ave express and the 6th Ave express to Brighton (but still having the Broadway locals to the 4th Ave local). The above is more popular, because there is a preference for Broadway express for Brighton riders, a preference for Sunset Park riders to the 6th Ave express (especially the Grand St stop) and that the most popular lines will all have direct midtown express service via either Broadway or 6th Ave. The only real novelty here, in my view, is sending the locals to West End instead of Bay Ridge. Because of the yard issues, this is needed, otherwise it likely makes more senset to split the 6th Ave expresses between Sea Beach and West End instead of between Sea Beach and Bay Ridge, as your plan presents.
While it is troubling for those on the West End to lose their express line, the increased service should not cost too much time as they would not have to wait as long. Plus, the West End riders do have numerous opportunities to transfer to an express service toward Manhattan.
While it was mentioned in your article, it is key to emphaasize the true benefits of deinterlining DeKalb, because we can remove the trains blocking each other just before hitting Manhattan. And it is very key that since 6th Ave stations are very close to Broadway stations in most of Midtown, most riders should be largely ambivalent as to whether their home station serves 6th Ave or Broadway. And if that is in fact the case, a plan like this is only a plus, since one can easily add more service under such a plan.
In your schematic map (the last one in the post) it looks like 3 lines stop at DeKalb. Maybe you want to edit that map so that the station only serves the N/Q and R/W? Just shorten the oval so it does not overlap the B/D. I think that would increase the clarity of the plan substantially. Thanks.
Good catch.
I like this proposal very much. But two questions: Is the B going to be moved to Concourse Yard? The B is currently based out Coney Island, so if it replaces the R at Bay Ridge and stays based out of CI, then like the pre-1987 Astoria R, it too will have to deadhead to CI Yard for storage and maintenance.
The second question is, why not use 38th St Yard in Sunset Park as a storage yard for R trains? Then the R can stay as the 4th Avenue Local train from Bay Ridge to Atlantic. Yes, it would require yard-bound R trains to either terminate at 36th St southbound and run light to the yard, or for northbound trains to terminate at 36th on the express track then reverse back into the station and go southbound to access the yard. Or they could just run light from 59th to CI. That’s what the R did for decades prior to 1987. It never seemed to be a problem until then. I’m not sure why it would necessarily be a deal-breaker now.
As I haven’t released the full deinterlining plan little details like this are missing from this post. But my idea is that the B will use 207th and take over the A express on CPW to 207th.
38th St is totally doable but as you outlined it requires complex movements. The whole idea behind deinterlining in the first place is opening up unused or unusable capacity in the system. So if we are going to propose more service then we need to keep all non revenue movements to a minimum.
To add to my point above:
The one hurdle I see is convincing the West End riders that this is better for them than a direct express, which they will be losing. On an ops basis, it certainly is better, since they wait less for a proposed R and then transfer to a  B or  D at 36th – but with all deinterlining plans the “cost” is that the West End riders will need a transfer, when before they could ride directly on the  D, albeit a less frequent  train than what is proposed here.
The basic truth of any deinterlining is a trade-off: More frequent trains and fewer delays due to merging in exchange for an additional likely transfer for a set of passengers. I think it is a good trade-off, but many others do not.
I agree with all of this.
I see your point about non-revenue movements. Even if they’re confined mostly to off hours, they could delay express service and result in some parts of the R line losing service if trains go OOS to get to 38th St Yard. This plan would avoid all that. I look forward to seeing your full plan. And you’re right. We can’t have a future full of, “Well, that’s the way we’ve always done it!â€
I like it, and I also like sending the (B) with the (A) to Inwood as EXPRESS. The (C) will remain the way it is as a LOCAL, however, we will need an additional 8th Avenue LOCAL to terminate at World Trade Center with the (E) from Bedford Park Blvd in the Bronx during Peak Hours so the (D) can go EXPRESS on the Concourse.
The (R) would gain a new terminal in Astoria with the (N), so the (W) has to go to the Upper East Side with the (Q). There will be no need to run the (R) along the Queens Blvd line, especially when the Queens Blvd LOCAL will connect with the Lexington Avenue line at 53rd Street (51st Street (6) Station). The (M) will run Full Time, and I would suggest extending the route as the LOCAL to the (F) to 179th Street, before splitting to Southeast Queens (Cambria Heights) via Farmers Blvd and Linden Blvd (the (F) continues East to Queens Village – Springfield Blvd).
This would support the (B) extension into Staten Island, to relieve pressure on the Staten Island Expressway.
I would let the Franklin Avenue Shuttle extend to Brighton Beach so the (N) and (Q) can run EXPRESS along the Brighton Beach line. We will need to extend the capacity of the Stations at Eastern Parkway and Fulton Street so the Franklin Avenue Shuttle can run 4 car trains.
My Ideas to improve lag on Times Square, Herald Square and Dekalb Avenue (Brooklyn)
N should go via 2nd Avenue, Broadway Express, 4th Avenue Express, vía West End
R should go via Broadway Express, 4th Avenue Express enter Queens via Lexington Avenue 63rd Street, also R goes via Manhattan Bridge. Skips 45 Street and 53 street in Brooklyn.
D train via 4th Avenue Local, vía Sea Beach, stops at 45 street and 53 street in Brooklyn.
Q train via Broadway Local and its new north terminal is Astoria Ditmars Blvd, along with the W train, also Q goes via Montague Street Tunnel.
This solves problems in both Manhattan and Brooklyn.
Explanations of Problems:
I get frustrated every time I have to wait at HSQ or TSQ for the N train to crossover tracks when another train is across the platform.
Sometimes the N train has to wait for the D to skip Dekalb after they both pass the Manhattan Bridge simultaneously.
The same thing happens when the Q train leaves Dekalb and N train skips Dekalb. They are going to pass the Manhattan Bridge, but one has to lag behind while the other goes forward.
Explanations of Solutions:
Eliminate crossovers by having N go to 96 street 2nd Avenue. Also put the R on the express track from 59 street Brooklyn to 57 street 7th Avenue (R can enter Queens Blvd via Lexington Avenue 63rd Street).
Eliminate lag by having North Manhattan Bridge trains B and D stop at Dekalb Avenue and D go 4th Avenue Local, while the South Manhattan Bridge trains N and R skip Dekalb Avenue and go 4th Avenue Express. And lastly put the Q train on the local track from Astoria to Coney Island via Montague Street Tunnel.