New York’s Awful Grade Separation
After Rick Scott rejected the Florida high-speed rail funds, a bunch of states as well as Amtrak applied for the redirected funds. The money has just been redistributed – see breakdown here. New York State applied for $300 million to grade-separate a junction between Amtrak and the LIRR, which it spins as an important capacity upgrade and which some online commenters have misinterpreted as a speed upgrade. Let me dispel the myth here.
The track map of the LIRR (link scrubbed for copyright reasons) shows clearly that, in the westbound direction, the junction has no conflicts. Amtrak trains (blue) using the northern tunnel pair to Penn Station have no conflicts with any other trains, except for other trains using the same tunnels. This is not a grade crossing, but a simple switch. In the eastbound direction, trains using the northern tunnels do have an at-grade junction with LIRR trains (purple) – but only trains going to a track farther north than the tunnels to Penn Station, and those all stub-end at Hunterspoint Avenue or Long Island City.
There aren’t a lot of trains going to Hunterspoint or Long Island City: at the peak, only 5 per hour, and of those one uses the Montauk Line, so we’re really talking about 4 trains per hour; Amtrak never runs more than 2 trains per hour to New York from the east. To put things in perspective, the 3 and 5 train on the subway have more than 10 trains per hour each and have a similar conflict in Brooklyn. What’s more, Hunterspoint’s main use is that it has an easy subway connection to Manhattan’s East Side, so once East Side Access opens and LIRR trains can go to Grand Central, traffic there will go down even more, making the flat junction even less relevant than it is today.
So the $300 million the state applied to has no relevance to either Amtrak or LIRR traffic. The only use is to let Amtrak use the southern tunnel pair to Penn Station without conflicts. Since Amtrak can already use the northern tunnels without any conflict apart from the one mentioned above, it is a pure nice-to-have. It would be good for operational flexibility if the tunnels were at capacity, but they aren’t: total LIRR plus Amtrak traffic into Penn Station peaks at 37 trains between 8 and 9 am, where the capacity of the tunnels is about 50 – and as with Hunterspoint traffic, Penn Station LIRR traffic will go down once East Side Access opens.
Pedestrian Observations 
Now you just have to go back and find all your comments on other blogs and publish them as a prequel.
+1
Would that money be enough to replace the Acela fleet with some German tilting DMUs?
Very useful, Alon! But if someone gave you $300 million to be spent on intercity rail anywhere in New York State, where would you spend it?
To be honest, I don’t think there’s anything pressing to be done with intercity rail in New York State. But the most bang for the buck is probably track upgrades to let the Acela run at at least medium speed (~100 mph) in New York City, and/or partially straightening the S-curve in New Rochelle.
A new fleet would require removing the FRA’s regulations; if they were removed, then best bang for the buck would be procuring new trains – not necessarily ones with very high tilt, but rather ones with good acceleration and maybe mild tilting. (And there’s no point in getting DMUs – the NEC is electrified.) If I worked for Cuomo I’d try to work with the MTA on getting the regulations revised, for commuter operations more than for intercity operations, but the only people who can really change the rules by fiat are on the federal level.
Er, oh yeah, somehow I always forget the NEC is electrified…
Alon, I can tell you speaking from Upstate that there is a LOT of pressing stuff to do with intercity rail in New York State.
Mostly west of Schenectady, though. Eliminating the single-track bottleneck near Lyons would be excellent, though that’s tricky because it’s in the middle of some sort of wildlife refuge.
Well, eliminating the Albany-Schenectady bottleneck is a pretty big thing. I’m honestly not sure what to do with Empire at current funding levels. Maybe electrifying Empire South is best, yeah (and unlike the Illinois line, this one is at least in principle good for 125 mph, which is actually HSR); it also doubles as a commuter rail improvement.
Personally, from a different perspective, I’d say the best bang for the buck would be electrifying the Hudson line with overhead all the way to Albany, re-electrifying Penn at 50Hz while they’re at it to avoid the need for frequency changes on the train. But then I think major improvements on the Empire Corridor would be the best bang for the buck. 🙂
Alon:
You very clearly do not understand anything about the operations of this segment of railroad or the East Side Access project.
First, the two bypass tracks are a necessary part of the staging of construction for East Side Access and provide a future conflict free route for both Amtrak and Port Washington trains from Lines 1 and 2.
Second, Amtrak rarely uses the northern tunnels into Penn Station because they are full of LIRR trains. Amtrak uses the southern tunnels into the station because those lead to the tracks which access the Amtrak concourse – tracks 5 to 16. Tracks 17 to 21 are only accesisble from the LIRR concourse. The only tracks with common access are 14 to 16. When Amtrak does use these tunnels, it is usually only Line 4 and when it does it causes conflicts at C tower. The use of Line 3 by Amtrak is almost impossible due to requiring a crossing movement at A tower against all outbound trains for NJ.
Third, the existing routes have significant same directional traffic conflicts that are currently resolved by 45 mph turnouts and crossovers. The bypass tracks raise the speed to 60 or 70 mph for this section of line and eliminate all conflicts. Now, if you forced Amtrak into the northern tunnels, you simply move the route conflict from Harold to either A or C interlockings in Penn Station, where the directional weave with LIRR traffic will be at 15 mph instead of 45 mph, and thus take that much longer to accomplish.
Fourth, during the morning rush hour, LIRR traffic is so frequent that Amtrak single tracks from Gate interlocking to Harold, and runs all westbound traffic through the middle of Harold on the normal eastbound track and the black track on the map known as the Long Island Freight Track. This makes Amtrak only have a weave movement with the Hunterspoint traffic and a merge with LIRR Line 2 traffic, which is less than traffic on Line 4.
Fifth
Andrew, what you’re effectively saying is that Amtrak can’t use the northern tracks because the concourse belongs to the LIRR. Now, look up the slogan “Organisation vor Elektronik vor Beton” and ask yourself what’s cheaper – not having separate and unequal concourses, or spending $300 million on an interlocking.
LIRR traffic in the morning isn’t that frequent. It has less than 40 tph, and the number will decrease due to ESA. It can be done with shared tracks and 2/2 running, and will still be easier than North River Tunnel traffic.
There is no conflict in the interlockings at Penn Station. Look at the two black turnouts on the map right next to the station, in the northern tunnels. If westbound trains use the one further west, it’s conflict-free. The turnouts restrict speed, fine; but let’s compare the cost of fixing a turnout at a station that has 2.5 times as many tracks as it needs and the cost of an elevated rail grade separation.
Nice blog, Alon. It’s going on my RSS reader. I’ve always thought you would be better off blogging rather than publishing so many comments on other sites (though I always look forward to your insights).
Could you do a series on now much transit really costs in Europe / East Asia? You often mention that all our transit capital projects are more expensive, and most transit operations are less efficient here in the USA.
Alon:
When Penn Station was built in the early 1900’s, it was purposefully done so that PRR had access to the middle of the station and Lines 1 and 2, and LIRR to the top of the station and Lines 3 and 4. There are enormous obstacles to just changing it around.
The concourses are on different levels. The LIRR concourse is on the subway concourse level to facilitate movement elsewhere on the island. The Amtrak concourse is near the surface to get people out of the station to taxis and on foot.
The LIRR concourse only access tracks 13-21, except on the one central concourse shared with Amtrak. Amtrak cannot use the higher numbered tracks for service up the Empire line because that tunnel departs directly out of Track 7.
Access to Sunnuyside Yard is mainly via Lines 1 and 2, which have dual track connections in and out of the yard to stage train movements. Lines 3 and 4 have only a single connection which would require big changes at Harold to rectify.
The turnouts at Penn Station cannot be fixed. There is limited distance between the platforms and the tunnels, fixing these interlockings as No. 8 double slip switches forever.
The access to West Side Yard is only possible on a large scale from the upper tracks. LIRR can only have a single track connection from Track 10 to 14. There is no access at all from the tracks below 10.
LIRR service is very frequent and at the practical limits of its signal system and platform arrangements west of Jamaica. Between 7:20a and 9:00a, LIRR runs 10 trains from Port Washington, 9 from Ronkonkoma, 8 from Port Jefferson, 5 from Hempstead, 16 from Babylon, 6 from Long Branch, 7 from Far Rockaway, 1 from West Hempstead, and 1 from Montauk. That’s 63 trains in 100 minutes, and the peak density of 41 trains per hour uses up all the capacity of the East River tunnel signal system.
I don’t understand why you think that $300M that needs to be spent to build East Side Access is such a waste when it is clear that the solutions you are proposing would cost billions. It is the simplest solution to the problem of level crossing conflicts at Harold between Amtrak and LIRR.
The North River tunnel signal system does 25 tph every weekday. So the East River tunnels have at least 9 more tph in them, and that’s without using anything that wasn’t invented here like ETCS. If the tunnels were actually at capacity then the extra flexibility of allowing any train to serve any track would be good, but they’re not even close.
What I’m proposing would cost zero. The turnouts are there. More concourse remodeling would be great, but isn’t even necessary. The Amtrak ticket offices can stay where they are; all that’s required is to reprogram the TVMs to work for all railroads serving the station and not just the agency that controls the local turf, and that’s a good thing regardless of train movements. And there is nothing wrong with making people use the lower concourse for some intercity trains, at least nothing more than than making people use the lower concourse for commuter trains.
What the Pennsylvania Railroad did in the era of frog wars and hostile takeovers isn’t always relevant to the operations of railroads in the modern era. The PRR also locked out competing railroads from Penn Station, but I don’t see anyone suggesting that Morris and Essex trains go only to Hoboken.