When it comes to the new LaGuardia AirTrain officials have lost sight of the larger picture. Extending the subway instead would have a system wide benefit. Here’s how.
As I have been writing about my ideas for fixing and expanding the NYC subway for some time now I often forget how far down the rabbit hole I’ve gotten. Often times online I talk with assumptions made but which I have to realize I’ve never fully explained to others. A big issue with the subway network is bottlenecks, areas where trains merge which cause delays. Alon Levy has for years been promoting his plan to deinterline the system, or rather remove merging trains all together so that each set of tracks serves one train line. The biggest obstacle to this is that each subway system (the IRT, BMT and IND) were all designed to work as interlined systems. Being last the IND was designed with more interlining than the rest. While interlining strives to provide a rider at a particular station a one seat ride option to various locations via multiple lines it does so at the cost of capacity. Each time a train merges it delays the line. This may not have been much of an issue in the past but as the system faces growing ridership (or it was before the whole thing started to go to shit) the need for capacity becomes ever more important. The MTA recently released their Fast Forward plan which aims to have Computer Based Train Control (CBTC) signals installed on more lines over the next decade. Currently only the L has such a system in place while the installation on the 7 train drags on due to software issues. Even with a new plan at hand installing CBTC requires lines to be shut down to do the work and will take time. Deinterlineing is a matter of scheduling trains and could be done for far less, much quicker.
The first place to start is the Broadway Line. Built in the 1920s as part of the Dual Contracts the Broadway Line was the trunk line which connected Astoria and Flushing to Bay Ridge, Flatbush and Coney Island. Modern for its day the line suffered from poorly designed interlockings (the areas where trains switch between tracks). The interlockings were built to allow any train to switch to any track. DeKalb Ave station was the hub, at the time the largest subway station in the world with 6 tracks, 2 local, 2 express, and 2 super-express. DeKalb was a giant sorting machine taking in trains from Bay Ridge, West End Line local and express (then the T but today the D), Sea Beach (N train), Culver local and express (F) and Brighton Beach local and express(Q). Mixing all these trains together and sending them different ways into Manhattan required trains to switch in front of one another often and caused frequent delays. The folly of this was seen right away.
The worst issue for DeKalb was the Manhattan Bridge. The bridge was built with two sets of tracks, one on the north and one on the south sides of the bridge, but because the city hadn’t developed a plan for who was going to operate these new subways the tracks were built before the connecting subways were. Eventually it was decided that the BMT would operate the bridge tracks as part of a new loop subway system. Trains using the north set of tracks would run to midtown and the southern tracks would immediately curve south and connect to the massive Chambers St station continuing south along Nassau and Broad Sts, connecting with the Montague St Tunnel back to Brooklyn. Much like at DeKalb, the Chambers St interlocking was all at the same grade so trains coming from Williamsburg would have to wait as downtown trains passed in front. This meant that almost every train along the entire BMT system would be delayed by just one train switching.
To tackle this problem the city undertook the Chrystie St Connection, a multifaceted construction program designed to fix the DeKalb interlocking and reroute 6th Ave express trains along the Manhattan Bridge. The tracks to Chambers St were severed and the tracks to midtown rerouted to the south side of the bridge so that the north tracks could be connected with Chrystie. At DeKalb the interlocking was expanded to have flying junctions, allowing trains to change tracks at the same time without having to wait for one another. This expansion required the destruction and abandonment of the Myrtle Ave station, a local station between Canal St and DeKalb Av. Today only a trace remains which can be seen on trains as they ascend the bridge. While the Chrystie St Connection was a massive success in Brooklyn it was still built to encourage interlining.
But while the city was removing one bottleneck they had created another. After World War 2 Queens was the fastest growing borough but all of the subways there were built before the war. No new transit was built in Queens after the war, save for a one station extension of the IND to 179th St. The IND had been built with the intention that riders would take local trains to the next express station and then switch to express trains. As originally designed the Queens Blvd Line used the G for all local traffic and the E for express. With the addition of the 6th Ave Line the F was added but it became clear very soon that riders wanted a one seat ride. Express trains were packed and locals were empty.
To remedy this the city looked for ways to add a new local connection to Queens Blvd. As part of the IND Second System a new East River tunnel was to be built between E 79th St and Broadway in Long Island City (such a tunnel was eventually constructed at 63rd St but was not connected to the Queens Blvd Line until 2001!). A cheaper option was found in the 11th St Connection. The BMT Broadway Line was built with 4 tracks to 57 St where 2 tracks turned east to Queens and 2 tracks stub ended pointing in the direction of Central Park West. Before the IND built the 8th Ave-Central Park West Line the BMT was vying to build and operate it. 57 St was designed so that the 4 track trunk line would continue north while 2 local tracks would swerve east to Queens. Because the northern extension was never built all trains either ran to Queens or terminated at 57 St. The 60th St Tunnel connected the subway in Manhattan to the elevated Astoria and Flushing Lines at Queensboro Plaza. (Something forgotten today is that service on both lines was split between the IRT and BMT per the Dual Contracts.)
To add a new local train for the Queens Blvd Line the city built a connection between the 60th St Tunnel (at 11th St in Long Island City) to the local tracks of the Queens Blvd Line as it entered Queens Plaza station. While the new connection certainly helped riders it unfortunately created a new bottleneck where Astoria capacity was reduced to allow for the new Queens Blvd trains. Even this new service was not enough to keep up with demand. In 2001 the 63rd St Tunnel was finally connected to the Queens Blvd Line at 41 Av which allowed for another Manhattan-bound local train to be added. This, however, required cutting back the Crosstown G train to Court Sq, first only during the day but eventually full time in 2010 .
The Broadway Bottleneck
The bottleneck works like this: the local tracks of the Broadway Line, which run from 95th St in Bay Ridge to Ditmars Blvd in Astoria, don’t have direct access to a storage yard and maintenance facility (in the past local trains were moved from the 36th St Yard off the D or the Coney Island Yard off the N). This means that at the beginning on their trip the trains must run over other lines to even get to their beginning point. If there is a breakdown the train must then be shuttled to the nearest yard along other lines as well. At rush hour when more trains are needed this limits capacity on the local tracks and snarls traffic on other lines. The R had originally ran to Astoria while the N went to Forest Hills but in 1987 they were swapped so that the R could have direct access to the Jamaica Yard. This pattern worked until ridership started to grown and in 2001 the W train was added along side the N in Astoria. Now 3 trains must merge at 60th St with N trains switching to the express track at Times Sq. This creates a bottleneck and limits each service to about 8 trains per hour max. While this works fine along the trunk it forces the branches to run a lower capacity. The switch between local and express tracks at Times Sq, due in part to the high demand on the local 49 St station, also causes delays.
The 60th St Tunnel was the only option for Broadway service but in late 2001 the 63rd St Tunnel was finally connected to the Queens Blvd Line in Long Island City. The 63rd St Tunnel was designed for future integration with the 2nd Ave Subway to allow Broadway express, Q, trains to turn up 2nd Ave while 6th Ave local trains would continue to Queens. The current set up of 63rd St, which is solely used but the F, runs below capacity while the 60th St runs at or above capacity. Now that 2nd Ave is open and the Broadway section of 63rd in operation it would make sense to send the Broadway-Forest Hills trains via 63rd so that more Astoria trains could be run. This would require R trains to merge at 57 St, still causing delays. A fully remedy of the bottleneck requires a new yard off the Astoria Line so that the R trains can run alone on the local tracks while the N is shifted to Forest Hills via 63rd St.
This is where LaGuardia Airport comes in. A proposal was floated in the 1990s to extend the N train from Ditmars Blvd along the Grand Central Parkway to LaGuardia Airport. The problem is twofold but the most publicized issue was running a new elevated line near homes and the local politicians shot it down. Since then the AirTrain in Jamaica has shown how a modern elevated structure can be built where it does not impact the community as much as the traditional subway structures. But the thing no one seemed to notice is that there is already a giant elevated structure over the GCP, the Hell Gate Bridge approach, which would require any line above the highway to soar hundreds of feet into the sky; It’s a non-starter.
But there is another alignment which should be considered. If the Astoria Line was to be extended north along 31 St two blocks it would reach the vast ConEdison lands where much of the land is used for storage and parking. Here is the prime location for a new yard and maintenance facility which would greatly boost capacity and efficiency along the entire system. An extension could then be routed southeast above 19th Av which is an industrial area of low warehouses to 45th St where there is vacant land and parking. Here the line would enter into a tunnel to finish the journey to the LaGuardia Airport terminal. There would be only two residential city blocks effected by the new elevated line and this is something that would need to be mitigated as there seems to be no other way to extend the line (a portal to the subway would require demolition of a block of apartments which would be even worse). The entire extension would be 2.6 miles long and include just one new station. The yard and facilities would be 18 acres, larger than the current East NY Yard used by the J train.
If the MTA was to build the LaGuardia Extension there would be a system wide benefit. The R train would be the sole local train and could run at much higher frequency, close to the L, 6, or 7 trains which also run alone. Bay Ridge riders would see increased service and while they would still not have express service along 4th Ave, a higher frequency R train would mean less waiting and therefore a faster all around journey. The center track on the Astoria Line could be used for peak express service for Astoria Blvd and the Airport which would also pick up riders who would today take a long bus to the crowded 7 train. The N would run through 63rd St to Forest Hills giving the UES more service and eliminating the merge at 60th St. One downside to this configuration is that Queens Blvd riders would lose their transfer to the Lexington Ave Line at 59th St but since the N is express in Manhattan they may not even need the transfer as they would have a quick one seat ride to Union Sq and Canal St. An alternative transfer is already available at Court Sq where E/M train riders can transfer to the 7 to access Grand Central and Midtown East. The improved efficiency would more than make up for the small percentage of riders that would actually be affected.
Line capacity is not equal on every line due to a number of factors which include terminal space, power, track geometry, speed, and merges which require extra spacing for trains to move between lines. For simplicity, using the existing signal block technology we will say capacity tops out at 24tph or one train ever 2.5min (in fact the only train which comes close to this is the L train). Removing current bottlenecks would open up capacity as shown on this chart.
Note: All service levels are shown at rush hour. Day, night, and weekend would be less. Additionally, this chart shows what is possible but not necessarily what actual service would end up being.
|Route||Current Service||Proposed Service|
|N||9tph||12tph, 10tph to Coney Island, 2tph to Bay Ridge-95 St|
|Q||10tph||12tph, 9tph West End Lcl, 3tph West End Peak Exp.|
|R||8tph||24tph to Whitehall, 14tph to Bay Ridge-95 St|
For service levels this high the Broadway Line would need to be deinterlined in Brooklyn as well. As proposed by Levy the Q and D would swap so that trains would no longer have to merge at DeKalb. The merges as 36th St and Prospect Park would still occur but have far less of an impact as trains would run as set pairs along their entire route; currently if there is a delay merging two trunk lines are backed up but if reverse branching is removed then only one line would be delayed. In this case Q trains would take over the West End Line and D trains would take over the Brighton Beach Local. As DeKalb is removed this would even allow for more frequent B/D service, although the lines would still be subject to reverse branching north of Columbus Circle.
The MTA’s argument against this switch has always been that it would force riders to transfer at Atlantic Ave via long an confusing passenger tunnels. But Levy makes the point that the Broadway and 6th Ave Lines run so close to one another that it seems unlikely that most riders would even need to make such a transfer. On Manhattan to Brooklyn trips riders would more likely just walk to Broadway or 6th Ave, a block apart in most places, so they have a one seat ride home.
R service would obviously benefit the most but ridership is still unbalanced between Queens, Manhattan and Brooklyn which is why 10 trains per hour would still terminate at Whitehall. In Brooklyn the added 6tph would almost double service and the reduced wait times would added to shorter total travel. The added Q service could also be split along West End for added peak express service. Worth considering is splitting off certain N trains at 59 St to create one seat express service for Bay Ridge. This would be similar to present service which has 3 N trains start at 96 St and 3 W trains start at 86 St (Brooklyn) during rush hour.
An obvious fault of this plan is that it still requires reverse branching along Queens Blvd and would add a new merge point at 63rd St. Without a brand new subway in Queens it is not possible to fully deinterline Queens Blvd. As the MTA is currently installing new CBTC signals along Queens Blvd which would add capacity there would still be room between trains for the merges required at the higher frequencies.
The downside of Alon’s plans is that to fully deinterline the system the MTA would need to spend vast sums of money to create new connections, often pedestrian transfers, to replace the one seat rides. Riders would also be forced to transfer more often, although if lines are running at higher frequency then transferring would be a far shorter wait than it can be today. While deinterlining works on paper the average rider often prefers a one seat ride which interlining allows. Where trains can be deinterlined for minimal cost they should be. Where lines cannot be so easily deinterlined they will still benefit from the improved efficiency further down the network.
The MTA gave up long term planning long ago as the bureaucracy, poor spending habits, and non-consistent funding would quash any dreams of expansion. Even basic bottleneck issues have never been addressed. Grand subway expansion plans are left to politicians who more often than not have no planning background nor even understand what they are proposing. Subway projects seem to be thought up in vacuums with little consideration for how the rest of the network will be affected. Worse, because agencies cannot even talk to one another we are left with expensive projects that add no new capacity (Fulton Center, the WTC Oculus) or build duplicate facilities when using existing ones would save billions (East Side Access). Today the Port Authority, at the behest of Governor Cuomo, is pushing ahead with their new AirTrain to LaGuardia which would run to Willetts Point so that travelers could access it via the 7 train and LIRR Port Washington Branch. This new shuttle is indicative of everything wrong with how transit planning has been operated in NYC. The new line would be 2.25 miles long and is projected to cost $1.5 billon but knowing the PANYNJ I’m sure this will ride. The line adds no new capacity to the existing system and in fact relies on riders using the already packed 7 train or navigating Penn Station. Many riders will still opt to use the M60 or Q70 bus lines since they will offer free transfers from the subway while the AirTrain most likely will require a second fare. The AirTrain was chosen over the subway because it was deemed cheaper and more politically feasible. But this is a terrible way to plan. The costs and hurdles of the Astoria Extension are greater for sure but so is the reward and the return on investment is something that needs to be considered in projects like this.