Category: Regional Rail
Why the 7 to Secaucus Won’t Work
Bloomberg’s expressed support for the now $10-billion proposal to send the subway to Secaucus is generating buzz and speculation about the ability to secure funds. Missing from this discussion is any concern for whether more people would actually transfer at Secaucus than do today. The instinct is to say that this provides a better connection to most of Midtown, but the transfer penalty literature suggests otherwise.
One important thing to note, writes Reinhard Clever, is that for commuter rail, downtown-side transfers are much more inconvenient than suburb-side transfers. Suburban commuters will drive to a park-and-ride, but balk at a transfer at the city end. Clever’s example is Toronto, where commuter rail riders tend not to transfer to the subway at Union Station but only take transit to jobs that can be reached from the station by walking. This problem is what doomed the Austin Red Line. For all its flaws, ARC offered a one-seat ride from the Erie lines to Penn Station.
Another thing to note is that suburban commuters routinely change trains at Jamaica today, but not at Secaucus. I’m not aware of a study on the transfer experience, but I am fairly certain that the difference is that at Jamaica the transfers are timed and cross-platform whereas at Secaucus they are not. Transferring at Secaucus today involves going up steps, passing through faregates, and going down steps, with no guarantee of a connecting train. The literature is unanimous that passengers will spend more than one minute of in-vehicle time to avoid a minute of transfer or waiting time: the MTA uses a factor of 1.75, the MBTA 2.25, Houston METRO 3.5-4 (last two from pp. 31-2 of Clever’s thesis). None of this is going to change if people are instead made to transfer from a commuter train to the subway, except perhaps that the subway train is going to be less crowded because it won’t be carrying commuters from the Northeast Corridor and Morris and Essex Lines.
Both issues boil down to the same fundamental: not all transfers are created equal. Within urban rail, people transfer all the time. Perhaps the disutility of getting up while changing trains is not an issue when passengers do not expect to find a seat in the first place. Regional rail riders transfer as well, when the transfers are easy and there’s no additional waiting time – in fact, setting up a timed transfer on a highly branched regional line increases the frequency on each branch, so any disutility from transferring is swamped by the more convenient schedule. What people don’t normally do is ride a regional line that gets them almost to their job, and then take urban transit for the last mile.
Commuters on the Erie lines can already make an uncoordinated transfer involving passing through faregates at two locations: Secaucus, and Hoboken. Some, but not many, already take advantage of this to get to jobs near Penn Station or in Lower Manhattan. The contribution of the 7 to Secaucus would then be to create a third opportunity for a transfer to 42nd Street. While 42nd is closer to most Midtown jobs than Penn Station, the heart of Midtown is in the 50s. At Queensboro Plaza more inbound riders transfer from the 7 to the N/Q than the reverse, emptying the 7 by the time it gets to Manhattan: the MTA’s crowding estimate as reported by the Straphangers Campaign, has the taken at the entrance to the Manhattan core, ranks the 7 the least crowded subway line at rush hour. Thus, although the 7 to Secaucus would add to the number of jobs served by a two-seat ride, many Midtown jobs would require a three-seat ride, no different from transferring to the E at Penn Station.
Therefore, good transit activists should reject the 7 to Secaucus as they did ARC, and I’m dismayed to see NJ-ARP‘s Douglas John Bowen throw in his support behind it as an ARC alternative. Before anything else is done, the Secaucus faregates should be removed, and the platforms should be remodeled to let passengers go directly from the Erie platforms to the NEC platforms. Here are better candidate projects for adding a pair of tracks under the Hudson:
1. ARC Alt G. Despite the ARC cancellation, it remains the best option.
2. Hoboken-Lower Manhattan. This doesn’t give Erie commuters a one-seat ride to Penn Station, but compensates with a one-seat ride to Lower Manhattan, and a two-seat ride from the Morris and Essex Lines to Lower Manhattan. The Manhattan terminal should not be more than a two-track stub-end with short tail tracks and the potential for a connection to the LIRR Atlantic Division. With about 50 meters of tail tracks and a platform with many escalators, the Chuo Line turns nearly 30 tph on two tracks at Tokyo Station. It’s an outlier, but given the extreme cost of building larger stations in Manhattan, the response should not be “They’re different, our special circumstances won’t let this happen,” but “how can we have what they have?”. Modern signaling and punctuality are critical, but, as the Germans say, organization before electronics before concrete.
2b. Jersey City-Lower Manhattan. The same as option 2, but with somewhat less tunneling in Manhattan and a lot more tunneling in Jersey. The main advantage is that new underground stations at Journal Square and Exchange Place would serve more jobs and residents than a station in Hoboken. It may be cheaper due to reduced Manhattan tunneling, or more expensive due to less maneuvering room coming into Lower Manhattan. It also forces the Manhattan platform to be east-west rather than north-south for a far-future cross-platform transfer with Grand Central and Staten Island.
3. The L to Secaucus, or to Hoboken. This has all the problems of the 7 to Secaucus plus more – 14th Street is at best a secondary CBD – but it conveniently replaces the L’s current low-throughput terminal with another. Ideally the L should only be extended a few hundred meters west, to the Meatpacking District, but if such an extension has large fixed costs, the incremental cost of extending the L all the way could be low enough to be justified by the benefits of a Secaucus extension, which are low but nonzero.
Passenger-Miles Are Overrated
One of the pushbacks I got about my post on road boondoggles is that I didn’t control for passenger-miles of travel, and the number for car subsidies is much lower when one divides it by the appropriate number of trillions. This is not the first time I hear people talk about passenger-miles as a measure of inherent worth, but it doesn’t make it any better.
Passenger-miles don’t vote. They’re not a unit of deservedness of subsidy. They’re one unit of transportation consumption. They’re like tons of staple as a unit of food production, or calories as a unit of consumption. We don’t subsidize food based on cents per calorie.
Even as a unit of consumption, there are flaws in passenger-miles as a concept, when it comes to intermodal comparisons. The reason: at equal de facto mobility, transit riders travel shorter distances than drivers. It’s very obvious when comparing total passenger-miles in transit cities and car cities (see e.g. page 36 here). Transit is slower than driving on uncongested roads, but has higher capacity than any road. In addition, transit is at its best at high frequency, which requires high intensity of uses, whereas cars are the opposite. The result is that transit cities are denser than car cities – in other words, need less passenger-miles.
What passenger-miles are more useful for is measuring intercity transportation. At intercity distance, mode choice has less influence over travel distance (though, even then, HSR and driving are shorter-range than flying, and thus passenger-miles can overstate the importance of flying over ground transportation). It is also a proxy for revenue, whereas on urban transit the fare is either flat or weakly dependent on distance. As a result, intercity railroads usually cite passenger-miles or passenger-km, and urban transit operators usually cite passengers.
But when it comes to local transportation, it doesn’t work very well. A country’s mode share expressed in passenger-miles is lower than that expressed in passengers, and this is going to make transit and especially walking look much less significant than they actually are.
The Rail-Trail Scam
I recently learned that a writer for the Adirondack Explorer has the following proposal to create a new rail-trail: demolish a line that’s in use by a heritage railroad, pave it over, and convert it to a bicycle trail. The arguments in the piece are your standard hatchet job considering only the costs of rail and only the benefits of the alternative, and are downright uninteresting; what’s interesting is that this is just the culmination of the misuse of the original concept of rail-trails.
Originally, rail-trails were created to preserve railroads for future use. Their mandate includes “to preserve established railroad rights-of-way for future reactivation of rail service.” In reality, restoration almost never happens. The Rails-to-Trails Conservancy’s review of railbanking points to success that a full nine railbanked corridors have had rail service restored, out of 301. The rest have been paved over, and often have enough non-railroad users that any restoration would be politically difficult in practice; I suspect that this is why the Providence Foundation makes no mention of restoring service on the second, now-abandoned track between Providence and Woonsocket in its regional rail study.
Another problem with railbanking is that it focuses on what’s useful as a trail, and not on what would be useful as a railroad later. There are pleasant exceptions, such as the Milwaukee Railroad’s route in most of Washington State, but in Rhode Island, the rights-of-way that have been preserved are those that would be easiest and least expensive to rebuild from scratch: the line to Hartford through West Warwick and Coventry, the line from East Providence to Bristol, and the aforementioned second track to Woonsocket. In contrast, many major pieces of infrastructure were demolished. Downtown Providence’s connection to East Providence was cut and would require new urban viaducts to be restored, and it’s sheer luck that the bridge over the Blackstone estuary is still there. Newport’s only rail connection to the mainline was railbanked but removed, which means restoration would face fewer regulations than starting new service from scratch, but only after rebuilding a bridge from the island to the mainland.
This is not intentional, but it’s neglectful of the needs of any mode other than the car as regular transportation; even bikes only get the nod for recreational use. The document coming out of Rails-to-Trails Conservancy, Railbanking and Rail-Trails: a Legacy for the Future, makes this thinking clear, when one reads between the lines. Here are some touted benefits of rail-trails:
The USDA’s Dietary Guidelines recommend at least 60 minutes of daily exercise for children and teens and at least 30-60 minutes everyday for adults. Trails provide close, safe, traffic-free paths for walkers, joggers, inline skaters [and] cyclists. Rail-trails are also part of a nationwide initiative launched by Congressman James L. Oberstar (D.-Minn.) to create safe routes that will encourage school children to walk and bike to school.
The first sign of utter disregard for alternative transportation as everyday transportation is the touting of “traffic-free paths.” Segregation of different modes of travel into different rights-of-way is the thinking of the traffic engineer and the freeway builder, not of the urbanist. The second is the fact that, in practice, the placement of those trails follows ideal corridors for the needs of trains, not bicycles or pedestrians. One does not use a mode of transportation that averages 60 km/h and loses 2 minutes every time it stops the same way one uses a mode that averages 25 km/h and can stop where you want. You can look at the northern end of the aforementioned West Warwick trail on Streetsview or on satellite and judge for yourself how useful it is for a cyclist’s daily work trip; a train would just blast through at full speed.
There’s already an ideal place for pedestrians are cyclists: the streets. Those are the strange linear alignments used by cars and fronted by actual residences and jobs. Away from urban areas, those are the country roads that go through small towns. A policy that aimed at reducing car use and getting people to use more active transportation would impose walkability and bikability standards on streets, which are where the exact addresses people want to go to are. A policy that didn’t care would turn railroads into recreational trails and greenwash it by saying they’re usable by pedestrians and cyclists. And I think we all know which of the two the rail-trail scam is.
Followup on the Providence Line and Woonsocket Trains
There’s a pretty bad mistake in my post about MBTA-HSR compatibility: the length of the Boston-Providence line is 70 kilometers, not 67 as stated in the post. In my defense, 67 (42 miles) is what the official mileposts say, on Wikipedia and on the catenary poles along the line. In calculating travel times I used a mix of milepost and Google Earth data, leading me to slightly understate the travel time difference between future high-speed trains on the corridor and improved regional rail. The difference is small, but is important for choosing overtake locations.
The correct technical travel times for nonstop 300 km/h HSR and 160 km/h regional trains making all current MBTA stops are 19.25 and 38.75 minutes, respectively. It’s offset by just half a minute from the technical time I originally thought was correct, but more of the difference occurs near Providence than near Boston. The upshot is that the single-overtake option in Sharon is loose in the north, allowing an additional Boston-area stop, and extremely tight in the south, requiring 200 km/h trains and not necessarily allowing regional trains to stop at Pawtucket.
This doesn’t directly affect Woonsocket trains, for which my example schedule is based on Google Earth lines and should be considered accurate given the assumptions. However, in a comment, I’ve been linked to a 2009 Providence Foundation study of the feasibility of a regional train to Woonsocket, under present FRA regulations, achieving similar trip times to those I propose but with fewer stops. The service proposed is very good relative to the regulatory and organizational environment it has to deal with – the projected cost per rider is about $25,000, fairly low by US standards.
The Providence Foundation study also includes a timed transfer at Pawtucket between Woonsocket and Boston, something I did not originally think of. Since the exercise on this blog assumes organizational competence on the MBTA’s behalf, we can choose an overtake option that makes this work optimally with short turnaround and transfer times. We should also include fare integration in the scenario, something that doesn’t currently exist even just between the MBTA and Amtrak. Under some HSR operating scenarios, it could charge the same fare as low-speed rail on the same corridor and have integrated ticketing, making a Pawtucket transfer less useful than an HSR transfer at Providence. Under others – for example, an HSR fare surcharge as currently practiced on the Shinkansen or ICE – it is not possible, and while integrated ticketing is still possible and desirable, cost-conscious commuters would need a solution not involving intercity trains.
It turns out that a single-overtake option does not accommodate Pawtucket transfers well, even if a Pawtucket stop could be squeezed into the schedule. Consider the following 200 km/h schedule north of Providence, with the 7% pad, rounded to a half-minute:
| Providence | 0:11:30 |
| Pawtucket-Central Falls | 0:15 |
| South Attleboro | 0:18 |
| Attleboro | 0:22:30 |
| Mansfield | 0:28 |
| Sharon | Arrive 0:33:30, Depart 0:37:30 |
| Canton Junction | 0:40 |
| Route 128 | 0:44 |
| Readville | 0:46:30 |
| Hyde Park | 0:48:30 |
| Ruggles | 0:54 |
| Back Bay | 0:56 |
| Boston South | 0:58 |
It’s possible to replace Readville with Forest Hills; the point is that there’s room in the schedule for it. The times above were chosen to make :00 the symmetry axis – i.e. southbound regional trains leave Boston at :02. Moving the symmetry axis is possible but requires giving up through-service to Warwick – the timetable would be too tight. Under this schedule, southbound regional trains would arrive in Pawtucket at :45, and HSR trains would arrive immediately after, at about :48; thus, southbound Woonsocket trains would arrive at the earliest at :50 and :20, timing them to just miss the northbound connection to Boston. Clearly, under such an option, the only way to provide satisfactory Woonsocket-Boston service is to connect to HSR at Providence.
The two-overtake schedule looks much better. It’s a tighter fit for Woonsocket trains between the faster MBTA and HSR trains, but once they fit, the transfer works well. Consider the following 160 km/h two-overtake schedule, with four-tracking between Readville and Route 128:
| Providence | 0:07 |
| Pawtucket-Central Falls | 0:10:30 |
| South Attleboro | 0:13:30 |
| Attleboro | Arrive 0:18, Depart 0:22 |
| Mansfield | 0:27:30 |
| Sharon | 0:33:30 |
| Canton Junction | 0:36:30 |
| Route 128 | Arrive 0:40, Depart 0:41 |
| Readville | Arrive 0:43, Depart 0:45 |
| Hyde Park | 0:47 |
| Forest Hills | 0:51:30 |
| Ruggles | 0:54 |
| Back Bay | 0:56 |
| Boston South | 0:58 |
Southbound MBTA trains arrive at Pawtucket at :49:30 and southbound HSR trains pass by Pawtucket at :44. Southbound Woonsocket trains have a window of about 1.5 minutes – they can arrive at Pawtucket between :51:30 (after the MBTA) and :53 (before the next HSR) to fit in on the same track pair used by the MBTA and HSR – but within that window they have a convenient transfer: 2.5-4 minutes to the next northbound MBTA train, at :55:30. Note that even in the off-peak, when MBTA trains would come every 30 minutes rather than every 15 minute, this works – we can just shift the slots used by MBTA and Woonsocket trains. Earlier arrival is good for the entire turnaround schedule for Woonsocket trains, which based on trip times would “like” to arrive at Providence at :58 and at Pawtucket at :51, though, if the Mineral Spring stop for Woonsocket trains is dropped, then :52 arrival is very comfortable at all ends.
The inclusion of Woonsocket service also favors ant6n’s proposed no-overtake schedule, in which Boston-Providence trains run at 200 km/h and skip stops near Boston and let Stoughton trains provide local service, and trains run every 20 minutes. It’s tight if MBTA trains stop at Pawtucket, but gives Woonsocket trains ample time for anything. Assuming a Pawtucket stop can be squeezed, for :58 Boston arrival northbound regional trains would depart Pawtucket at :27, i.e. southbound MBTA trains would depart at :33 and HSR would pass by at :35, right on their heels. Woonsocket trains could be slotted anytime between :37 and :47:30, where :41 would be optimal for their own turnaround times and :45-46 would provide the shortest robust connection.
Blackstone River Regional Rail
Following up on my proposal for improving regional and intercity rail service between Providence and Boston, let me propose a line from Providence to Woonsocket, acting as an initial line of a Providence S-Bahn. The basic ideas for how to run a small-scale regional railroad, as usual, come from Hans-Joachim Zierke’s site, but are modified to suit the needs of a line with a larger city at one end. It is fortunate that the road connecting the two cities is not a freeway, and takes 24 minutes, allowing good transit on the same market to be competitive.
RIPTA’s bus route 54 goes from Providence to Woonsocket, generally taking 53 minutes one-way, with a few express runs taking as little as 39; the frequency is about half-hourly both peak and off-peak. A regional line would effectively railstitute it. Lincoln Mall, which is on the bus route but not near the rail line, would be served by a branch bus with timed connections to the train. See map here, together with some proposed intermediate station locations. Depending on the stop pattern, additional buses could be replaced, most readily route 75.
To avoid degrading service, frequency must be at least half-hourly. Of course, complete fare and schedule integration with the buses is non-negotiable: the fare on the train should be the same as on the buses it’s to replace, and transfers should not cost extra money.
As in the case of Zierke’s proposal for regional rail in southern Oregon, this is impossible under FRA regulations. Unlike the case of MBTA-HSR compatibility, getting a waiver here is difficult, since RIPTA is a small agency and can’t afford to conduct the studies required for a waiver request. In addition, north of Pawtucket, the line is an active freight line owned by the Providence and Worcester Railroad, and passenger service with high platforms (low-floor equipment is ruled out by the high platforms at Providence) may well require a new passenger-dedicated single track, raising capital costs by tens of millions of dollars.
Nonetheless, in a regulatory environment more favorable to passenger rail, such a line can succeed. Travel time of about 25 minutes, comparable to driving, is realistic. The length of the line is 25.5 km, and could still support a minimum speed of about 90 km/h even in its curvier northern half. The technical travel time is about 15 minutes plus 1 minute per stop. To ensure one-way travel time remains well under 30 minutes, enabling two trains to provide half-hourly service, there’s a maximum of about 9 stops. The map above includes 7 stops I believe are necessary for the line’s success, and a few optional locations. The explicit assumption for the following schedule is 90 km/h speed north of Lincoln Junction and 120 km/h south of it. Together with 7% padding, we obtain:
| Woonsocket | 0:00 |
| East Woonsocket | 0:02 |
| Manville-Cumberland Hill | 0:06 |
| Albion | 0:09 |
| Lincoln Junction | 0:12 |
| Valley Falls | 0:16 |
| Pawtucket-Central Falls | 0:19 |
| Mineral Spring | 0:21 |
| Providence |
0:24 |
| Providence | 0:26 |
Trains meet south of Lincoln Junction, requiring at a minimum two tracks at and south of the station. If trains leave both ends simultaneously, then they stop at Lincoln Junction within 2 minutes of each other, making timing the connecting bus easier.
This meshes with the sped-up trains to Boston well. Travel time from the junction with the NEC in Pawtucket is 7:30 minutes, versus 3 minutes on a 200 km/h intercity trains. Under the one-overtake option, intercity trains arrive in Providence 3 minutes after regional trains from Boston, giving the DMUs an ample window to make local stops (8 minutes with a 2-minute headway and 15-minute Boston service), even with the flat junctions at the split in Pawtucket and at Providence Station. Under the two-overtake option, Boston regional trains arrive about 5 minutes after intercity trains assuming no additional stops in the Providence area; adding the same three stops made by Woonsocket trains to the Boston trains would turn this into 9 minutes, and the DMUs would have a window between the intercity and regional trains, combining to provide intense local frequency between Providence and Pawtucket.
In other words, capacity constraints at Providence do not exist under this service pattern, answering concerns raised in comments on a post Greater City: Providence. The post itself has important ideas for pleasant development near Providence Station, which is currently urban renewal hell. The only drawback of railstitution is that Kennedy Plaza is closer to the jobs of downtown Providence than the train station, and even with the trip time cut from 53 minutes to 26, it’s essential to provide easy pedestrian access from the station to nearby city destinations.
Modern DMUs have fuel consumption similar to that of buses and are maintained in the same shops, so with higher speed RIPTA can expect similar or lower operating costs and higher ridership. If a passenger-dedicated track is not required, then 9 high platforms, a passing siding, and 4 DMUs should suffice; capital costs would be very low, especially relative to ridership, and may well receive federal support. Based on Zierke’s German examples, daily ridership in the low to middle thousands would be good but realistic; 10,000 would be a miracle and 2,000 a bust.
(With thanks to Jef Nickerson for the idea.)
MBTA-HSR Compatibility
There is going to be major investment in the Northeast Corridor, and several possibilities, including Amtrak’s NEC Master Plan, call for running trains at higher frequency and somewhat higher speeds than today on the Providence Line, and assumes electrification of commuter service. Since the line is already being used by the MBTA, which according to Amtrak is limiting the number of intercity train slots for capacity reasons, this calls for a good measure of schedule integration, based on the principle of organization before electronics before concrete.
Amtrak’s Master Plan calls for three-tracking the entire Providence Line south to Attleboro (one viaduct excepted) instead, at a cost of $464 million – $80 million in Phase 1, $384 million in Phase 2 – in addition to money spent on unnecessary expansion at South Station. Such a cost is excessive, suggesting that better MBTA-HSR compatibility is required. Full-fat HSR programs go even further and avoid the Providence Line in favor of a greenfield alignment or an I-90 alignment, instead of making use of the existing high-speed track in Rhode Island and Massachusetts. To reduce costs, a better plan would four-track short segments for passing sidings, and time the overtakes. The principle is similar to that of the blended Peninsula plan in California, in the version proposed by Clem Tillier.
In many ways, for example the metro area populations involved and the current ridership level, the Providence Line is similar to the Caltrain line. The main difference is that the Providence Line has fewer stops and therefore can expect higher average speeds. In addition, the Providence Line is straighter and passes through less developed areas, so that even today Acela trains plow it at 240 km/h, and about 330 km/h is possible with true high-speed trains and higher superelevation.
In Switzerland, trains run as fast as necessary, not as fast as possible. In this context, this means running just fast enough to meet a good clockface schedule. Boston-Providence travel time on the MBTA today is about 1:10; for a good takt, this should be cut to about 55 minutes, allowing hourly service with two trainsets and half-hourly service with four.
For the purposes of schedule symmetry and avoiding switching moves at high speed, passing segments should have four tracks rather than three when possible. Costs should be controlled by making those passing segments much shorter than the three-tracking Amtrak proposes.
Finally, the timetables proposed here are based on the following performance assumptions: regional trains have a top speed of 160 km/h, accelerate like a FLIRT (45 seconds acceleration plus deceleration penalty), have an equivalent cant of 300 mm, and dwell at stations for 30 seconds. Intercity trains accelerate like an idealized N700-I, have an equivalent cant of 375 mm, and dwell for 60 seconds. The equivalent cant is by and large unimportant; the acceleration and dwell times for regional trains are. The approach into and out of South Station has a speed limit of 70 km/h through the 90-degree curve toward Back Bay, and 100 km/h to south of the curve at Back Bay; intercity trains are limited to 200 km/h south to Readville and 250 km/h south to the Canton viaduct, and, at the southern end, 225 km/h west of the curve in Attleboro and, curves permitting, 200 km/h in Rhode Island. Regional trains turn in 5 minutes, or 4 at a minimum, and intercity trains turn in 10 minutes at a minimum. Signaling allows a headway of 2 minutes at a speed of 200 km/h and 3 minutes at higher speed, but if a regional train starts from a siding stop, it can follow a high-speed train more tightly initially, say 1 minute, still far higher than a safe stopping distance, since the spacing rapidly increases over time. Grades are ignored; the Providence Line is flat enough that they’re not an issue. Timetables should be padded 7% from the technical time.
With the above assumptions, the technical time for regional trains is 38 minutes with the present stopping pattern, which yields 41 minutes with padding; this compares with 46 minutes for the fastest Acela. Clearly, if Acela service levels remain similar to what they are today – which includes the Master Plan, which calls for a 10% reduction in Boston-New York travel time (see page 40 on the PDF linked above) – there’s no need for passing segments. To raise travel time to 55 minutes, trains should make more frequent stops, and/or run to T. F. Green Airport always. Although the speed profile of regional and intercity trains would be different, the average speed would be the same, and given that the corridor has a small number of trains per hour of each type, this mismatch is no cause for concern. The $464 million Amtrak is proposing would then be a complete waste, and the federal government should spend any money toward this goal on electrifying more MBTA lines and funding EMUs.
However, in a scenario involving a significantly improved intercity service, the best technical time for nonstop Boston-Providence service with a top speed of 300 km/h decreases to about 19 minutes (20.5 with pad), and this makes overtakes necessary. A slowdown to 250 km/h only adds about one minute of travel time, so the operating pattern is almost identical.
If 15-minute service, both regional and high-speed, is desired, then regional trains can be about 11 minutes slower between successive passing segments, since 11 = 15-3-1 or 15-2-2. A single mid-line overtake is theoretically possible: 41-20.5 = 20.5 < 2*11. However, such an overtake would have to be exactly at the midline, and, in addition, there could be merge conflicts at Providence, whose station tracks include two on the mainline and two on one side of the mainline as opposed to one on each side.
It’s still possible, but tight, to have a single overtake at Sharon. The immediate station vicinity would be four-tracked; this is no trouble, since the area around the station is undeveloped and reasonably flat. In addition, there’s more than enough time in the Providence area, making the merge conflict a lesser problem. However, this is very tight near Boston South, beyond signaling capability unless four-tracking extends a few kilometers further north. One way to counter this problem is to slow high-speed trains by making them all stop at Back Bay and/or Route 128, adding precious minutes to the schedule but reducing the speed difference. Conversely, the current weekday pattern of Providence Line trains skipping Ruggles could be made permanent. There is no room for infill stops; the overtake would only add 4 minutes to regional train travel time, so there’s time to run further to the airport at 160 km/h, and even make an extra stop at Cranston.
Another possibility is to have two overtakes, taking advantage of existing four-tracking around Attleboro. The capital costs are similar; it would require four-tracking around Route 128, possibly extending north to Readville if an on-the-fly overtake is desired. The operating complexity is much higher, since there’s one more opportunity for a late train to mess up the entire schedule. However, there is plenty of slack south of Attleboro and north of Route 128 allowing for additional stops. Under this option, the train loses 4 minutes waiting at Attleboro and about 2.5 at Readville, since the overtake is not completely on-the-fly, raising travel time to 47.5 minutes. There’s no time for airport trains, not on the same takt. However, there’s space in the schedule for 5-6 infill stops in addition to Readville; Forest Hills, Pawtucket, Central Falls, and perhaps one more in each of Boston and Providence closer to city center.
In principle, it’s possible to extend this analysis to 10-minute service, with three overtake segments, at Route 128, Sharon, and Attleboro. In practice, this is operationally cumbersome, and the operating profits coming from filling six full-length high-speed trains from New York to Boston ought to be able to pay for four-tracking the entire line, even the viaduct.
Not included in this analysis are the branches. Those are not a worry since north of Readville there are three tracks, and frequencies on the other lines are low. The Stoughton Line is a bigger problem; however, with the three tracks through Boston, it could still be shoehorned. Electrifying it should not be difficult due to its short length, though the proposed Taunton extension would make it harder.
National Low-Speed Rail Network Proposal
With all the focus on high-speed rail and urban transit, it’s easy to forget the low-speed rail that forms the backbone of every good national transit network. Switzerland, whose high-speed infrastructure consists of shared passenger and freight rail base tunnels, has a national rail ridership that puts the rest of Europe to shame. Japan may be famous for the Shinkansen, but the enormous low-speed networks surrounding Tokyo and Osaka are the two busiest in the world. Although intercity travel produces disproportionate revenues, most trips are local, even on mainline rail, and government rail planning should make sure to prioritize regional travel.
While the main intercity routes in the US should be eventually upgraded to high-speed rail, rather than rapid legacy rail, the low-speed network should dominate regional traffic as well as intercity gaps in the high-speed network. This means two traffic classes: regional and intercity. The intercity travel in question is for the most part short-distance – for examples, lines in Michigan fanning out of Detroit and not reaching any future high-speed lines to Chicago, lines in Georgia fanning out of Atlanta, and the portions of Amtrak California that won’t be replaced by HSR. The service level on the intercity lines should be a more modern version of the Regional, Keystone, and Empire South services; the service level on the regional lines should be the same as that of regional lines in Continental Europe.
The standards for the low-speed network should be based on the best industry practices. Because those lines are by definition not the highest-volume routes, it’s important to plan them with utmost care to keep costs under control. Federal assistance should aim to do the opposite of what FRA regulations do today. Instead of encouraging outdated practices, the federal government should on the contrary promulgate a set of good practices, based on what is done in Switzerland and other countries with good regional rail.
This is similar to what the various good roads bureaus did in the early 20th century, creating a unified set of standards. That said, the roads movement should only be an inspiration in the vaguest sense, since in reality US road building was much heavier on concrete than necessary and lighter on organization, leading up to the overbuilt Interstate network. This means that, whereas federal-aid highways are required to meet minimum standards for width and speed, federal-aid low-speed rail should be required to meet minimum standards for schedule and fare integration with local transit, signaling, and punctuality. The German motto, organization before electronics before concrete, rings truer here than for other kinds of transit investment, and agencies that ignore it should not receive funding for concrete before they complete the cheaper fixes.
Scandinavia is of especial importance as a rolemodel, because the lower density of its metro areas forces its regional trains to be faster, as they ought to be in the US. Combined with the wider loading gauge, it means that Swedish and Norwegian orders should be one of the sources of early American rolling stock. The lower speeds of Continental Europe (excluding Scandinavia) are not sufficient for more sprawling American urban areas. Instead, regional trains should have a top speed of about 160 km/h or just a little less, except on branch lines. A good example for the service quality to aim for is the Caltrain-HSR Compatibility Blog plans for trains from San Francisco to San Jose: local trains, stopping about once per 3.5 km, average 59 km/h, and express trains average 85 km/h.
While some regional lines in the US already average 60 km/h or even more, the cost is a very sparse station spacing, such that walking to stations is infeasible, even if the station areas are walkable, which they usually aren’t. For example, the Providence Line from Providence to Boston averages 58 km/h, with one daily late-night train with less schedule padding and another that skips stops achieving 65 km/h; however, the average interstation is 6.8 km, and requires skipping or closing down entirely several urban stations (Forest Hills, Ruggles, Readville, Pawtucket).
Instead of current practices, I would recommend a program of federal standardization based on the idea that transit should be able to compete with driving and provide meaningful transportation at all times of day. Federal action means that a few best practices could be violated: most prominently, rolling stock doesn’t have to be completely off-the-shelf if the federal government can induce transit agencies to combine and buy in bulk. However, the most important of the general best practices – perfect schedule and fare integration, allowing seamless intermodal transfers regardless of which agency operates the vehicles – are as important as ever. This leads to the following set of suggestions, in addition to the aforementioned set of best practices:
1. The main lines, both regional and intercity, should be electrified, with 25 kV 60 Hz.
2. Trains’ design speed should generically be 160 km/h, or a little lower on unelectrified branch lines and regional lines with frequent stops, though the track speed could be lower if increasing it is not worth the extra cost. Acceleration should be high, to allow average speed to remain high even with a few more stops. The ideal train should look like an M-7 with bigger doors from the outside and have the performance of a FLIRT. On unelectrified lines, good choices include the diesel Talent, GTW, Desiro, and Coradia. Bilevel trains are useful only in narrow circumstances in which passenger volumes are very high and the higher dwell times coming from the double-deck configurations are not a major problem; with a few exceptions such as the MI 2N used on the RER, this is practically never the case.
3. Subsidies should still be acceptable for regional services, though relative to passenger volumes they should be lower than they are in the US today; they should not be acceptable for the intercity network, though weak lines within a network could be subsidized by stronger lines they connect to.
4. In urban areas, regional service should function as urban transit and not just as peak-period commuter rail from the suburbs to the city center; therefore, there should be frequent stops in the city, replacing the longer-distance functions of American light rail lines. In-city fares should be identical to those of local urban buses and rail.
5. Regional trains should have just one operator, with the fare enforced with random fare inspections; intercity trains, which have lower traffic, can have one operator and one conductor.
6. There shouldn’t be any distinction between regional, intercity, and high-speed rail stations. High-speed rail should be able to seamlessly run through to lower-speed territory when necessary – for example, surplus Northeast Corridor trains that do not need to go to Boston should serve Jamaica at least (with catenary strung over the LIRR Main Line), and possibly even Mineola, Hicksville, and Ronkonkoma.
7. Construction projects should prioritize lines that serve markets that cars can’t, e.g. travel that passes through CBDs or parallels roads that are not freeways.
8. Signaling should be either ERTMS or ATC. Unless the two systems can be made to talk with each other, the federal government should invite delegations from the vendors, pick one, and mandate it. (And unless Hitachi can provide a convincing explanation for why its vendor-locked system is better, the pick should be ERTMS, which has eight vendors.) It can squeeze amazing capacity out of two tracks and, when enabled, provides absolute crash protection.
9. High punctuality is non-negotiable, especially when timed transfers or overtakes are involved. Trains should be able to stick to their clockface schedule and passengers should be able to rely on transfers even with short connections. Here is a list of ways to maintain punctuality. The ultimate goal is Japan, where, barring suicides and natural disasters, late trains are almost unheard of.
Those requirements are deliberately meant to be as scalable as possible. Although the rolling stock I’m implying is very ambitious for small-scale operations, the advantage of the high top speed is that such operations could piggyback on larger orders by the main established agencies, which could make great use of the extra speed and acceleration and get a more rationalized schedule as a result. The point is to give agencies pricing power coming from pooling together to order multiple thousands of more-or-less identical EMUs.
Although the investment described here is much more intensive than anything done in the US up to now, the true cost is not high. Restoring regional branch lines should be doable for a million dollars per kilometer, bulk electrification of main lines can be done for not much more and has been done on $3 million/km on the NEC, and mainline ETCS installation costs $1–1.5 million/km. It’s comparable to the per-km cost of the diesel-only, single-track, low-platform, commuter-only Lackawanna Cutoff, and if past results are any guide would lead to a sharp increase in transit ridership, measured in hundreds rather than tens of percent.
The ultimate goal of low-speed rail is to make it convenient to use regional transit. With speeds comparable to those of driving, local fares comparable to those of buses, and a frequent, memorable clockface schedule, transit would be a realistic option for many more people in the US than it is now. Every trip should be serviceable by transit, or else people will find it more convenient to buy a car for their irreplaceable car trips and then drive it for other trips. SBB claims that 32.7% of Swiss travel to work is on mass transit; this is higher than the figure for Greater New York, and about seven times the figure for the US.
Some of this is, to quote James Kunstler, Bill Lind, and other supporters of transit who look backward to the industrial era, merely restoring what the US had in the 1920s and 30s, before cars made all but the most traffic-intensive rail travel unprofitable. But the operating practices I’m proposing are modern, in line with today’s labor and capital costs and with innovations in countries that have kept improving their rail systems. Modern low-speed rail shares many characteristics with old local trains, but it’s fundamentally something that’s never really existed in North America. It’s about time to try it.
Uncompetitive Transit
In general, government at all level should be encouraging a mode shift away from cars and toward trains, using legacy lines for regional service outside urban areas. Here is a canonical example of such a proposal, unfortunately completely unofficial, in Medford, Oregon. A key point is that transit needs to provide a competitive trip time, and connect people to where they want to go, or else there’s no point in running it.
Sometimes, it’s impossible given present infrastructure. One example of this, routinely mooted on California High-Speed Rail Blog, is a system connecting to Gilroy and feeding high-speed rail. For the purposes of this discussion, let’s assume that the current FRA regulations and US rail practices have been completely gutted and replaced with Swiss or Japanese practice, and, more speculatively, that the legacy line can be made passenger-primary, despite Union Pacific ownership. The system would connect Gilroy, Santa Cruz, Salinas, and Monterey, using a now-abandoned right-of-way to get to downtown Monterey and legacy lines elsewhere.
The result can be seen on this map. There would be timed transfers at Castroville and Watsonville (running one-seat rides everywhere at acceptable frequency would require too many trains), and several additional intermediate stops, such as Marina, Seaside, Capitola, and Aromas. In terms of pure railroad operations, it could be a well-run system. Unfortunately, it could not be a successful one: the largest and densest city on the line, Salinas, is connected to the others in a very roundabout way. Salinas-Gilroy is 60 kilometers by rail and only 45 by road. Frequent curves would make it impossible to maintain a high average speed. Even a 55-minute trip time, allowing two trainsets to provide hourly service, would be ambitious, though possible with a wide stop spacing and good rolling stock; in contrast, driving takes 37 minutes according to Google Maps.
Monterey-Gilroy and Santa Cruz-Gilroy would be a little more competitive – they’re 50 and 54 minutes by car respectively. However, the markets are much smaller, especially in the case of Santa Cruz, where to get to any regional destination other than Gilroy, it’s faster to drive to San Jose. In addition, Santa Cruz-Gilroy is the hardest pair to get on a reliable clockface schedule: it’s 65 km, and the segment west of Watsonville is 34 with many curves, some of radius going down to about 220 meters, restricting speed even under optimistic performance assumptions to 75 km/h.
Since the congestion level in this part of California is not very high, cars could always beat the train, and for many trips so could buses. Therefore normal origin-and-destination travel would not produce much ridership on such a system. The worse trip time would be tolerable to some high-speed rail travelers if the transfer to high-speed rail were well-configured; however, high-speed travel alone does not generate enough ridership to justify an entirely new rail system, especially at an outlying station such as Gilroy. It would be the high-speed rail equivalent of an airport express.
There occasionally arise such cases, of lines that look good in principle but can’t be made competitive in practice. That is one example. A few more, not all seriously proposed by transit proponents: many international high-speed rail links in general, and some in particular, for example Minneapolis-Winnipeg (it would dominate the market, but the market is so small it’s not worth it). The only thing that can be done is spend scarce transit funding elsewhere. There are enough regional and intercity lines that could work well and no shortage of local transit supporters, some with political clout, who want them. Urban lines, which routinely get the short end of the stick in California in favor of low-performing outward extensions, would clamor for some of the money required to get a Santa Cruz-Monterey-Salinas-Gilroy system up to acceptable performance standards.
Sunnyside Junction Proposal
The in-progress East Side Access (ESA) project linking the LIRR to Grand Central is scheduled to open in 2016, and Metro-North is already studying options to use space vacated by the LIRR to run its own trains to Penn Station along the Northeast Corridor. Thus the basic service pattern will look as in this map. Observe that alongside Sunnyside Yards, there’s a stretch of track between the split between the Northeast Corridor and the LIRR Main Line, and the split between the access tracks to Penn Station and the ESA tunnels.
This should be turned into a new junction station, Sunnyside Junction. At this junction, passengers could transfer cross-platform between trains to Grand Central and trains to Penn Station, just as they do at Jamaica between trains to Penn Station and trains to Brooklyn today. If Metro-North diverts half of its 20 peak New Haven Line trains per hour to Penn Station, and the LIRR diverts two thirds of its 36 peak tph from Penn Station to ESA, then the service to each Manhattan terminal will be about equal.
Since both Manhattan destinations are of high importance, no train should skip Sunnyside Junction, not even peak-of-peak LIRR express trains that skip Jamaica. (Trains should not skip Jamaica, either, but that’s another matter.) Thus off-peak frequency could be assured to be fairly high, comparable to that to Jamaica (about a train every 10 minutes), and peak frequency would be so high that the transfer penalty would be negligible.
An advantage of this setup is that even at the peak, one-seat rides to each destination would become unnecessary. Therefore the interlockings and switching moves could be simpler, and new grade separations should not be necessary. In the off-peak, the transfers should be timed, even across agencies; this should be the first step of good regional rail service. Note that I advocated something similar as part of a comprehensive regional rail plan for New York, but Sunnyside Junction could be built independently of it. Indeed the interlining that minimizes switching moves and conflicts is not the same as the through-running in my original plan, which is based on matching ridership at the New Jersey end to ridership at the Long Island or Connecticut end.
Because a stretch of straight track for this station already exists, all that is necessary is platforms. Because all trains should stop at this station, and the capacity limit lies elsewhere in the system (namely, in the ESA tunnels), it would suffice to have two island platforms and four tracks serving them, and two additional bypass tracks to allow Amtrak to skip the station even at peak hour. If the station became very busy then two additional stopping tracks could be required, and construction should leave space for them.
To ensure the station is well-patronized by transferring passengers, like Jamaica and unlike Secaucus, it should not feature fare barriers or other obstacles between the platforms. Transferring should involve walking a few meters from one track to another, on the same platform. This is perfectly compatible with the current regime of requiring conductors to check every ticket on the train, because Penn Station and Grand Central are both in the CBD and thus the fare to them should be the same. The rationale for the faregates at Secaucus is that fares to Hoboken and Penn Station are different, and conductors would not have time to check that everyone on a train from Secaucus to Penn has a valid ticket to Manhattan; this is irrelevant to Sunnyside.
In the future, the LIRR and Metro-North should consider lowering in-city fare and raising frequency, which could work with more modern operating rules (i.e. proof-of-payment instead of conductors checking all tickets). Seamless fare integration with the subway would open the door to direct Queens-Bronx service; Metro-North is already considering Bronx stops for its Penn Station service. It would also give Queens another access point to Manhattan, slightly decongesting the near-capacity Queens Boulevard subway; the reason I say slightly is that the worst problems are far east of Sunnyside. And frequent service to the rest of Queens and to Manhattan would provide another public transit option to the area.
Unfortunately, the LIRR seems to not make any plans for such a station. It had plans for a station west of the split, serving only Penn Station: see page 13 here. I do not know if such plans will ever materialize in light of ESA’s cost overruns; I cannot find a more recent official reference to them. A cross-platform connection seems to never have been on any official agenda. Fortunately, even now it should be possible to add one, at relatively low cost since this station would be entirely above ground, and with minimal disruption to service since the site is a wide railyard with 6-8 active through tracks.
Good Industry Practices Thread
In contrast to the mismanagement highlighted in the last few posts, there’s a set of best industry practices for good transit. Here is a list of what I believe are the most salient. As far as possible I’ve avoided contentious issues that well-run agencies disagree on. By its nature, the list is open, and you should feel free to comment with your own ideas of what’s more important.
1. Regions should organize regionwide transport associations (the German Verkehrsverbund) with integrated fares and schedules, even across political boundaries. One ticket should be valid on all trips, and transfers should be free even across different operators. Bus and rail schedules should coordinate to minimize transfer time; rail-rail transfers should be cross-platform when possible and timed when possible, even if frequency is high.
2. Schedules should be organized on simple clockface intervals (Takt): instead of complex timetables, the same pattern should repeat every half hour or hour, and should be compressible to a system map. Supplemental peak services should be integrated into the same takt, for example arriving at the half-points or maybe third-points if the peak is very prominent. Minimum off-peak frequency for regional branch lines is hourly; for commuter rail and anything else intended to serve as suburban transit, it’s half-hourly; for urban services, it’s 15 or at worst 20 minutes.
3. If express service is desired, it should be limited-stop and make stops at all major stations, rather than running very long nonstop segments. For a good example, go here and click on the interval timetable links. In addition, the express buses and trains should run on their own clockface schedule, and express trains should have timed transfers to maximize utility and overtakes to minimize the amount of four-tracking required. The practice on Metro-North and other legacy US railroads of having peak commuter trains make a small number of suburban stops and then run nonstop to the CBD should end; not everyone works in the CBD.
4. Boarding should be level. For regional rail, this means at least moderately high platforms are non-negotiable. For surface transit, this means low-floor equipment; high-floor BRT is a feature in Latin America, where it’s a lower-cost replacement for a subway, but in developed countries, the cost of paying so many bus drivers is such that BRT is a replacement for local buses and should be open with many curb stops in outlying areas.
5. All payment should be done on a proof-of-payment basis. Any vehicle, no matter how long, should have at most one employee on board, operating it. The fare should be enforced with random inspections; it pains me to have to say it, but the inspectors should never hold a bus during inspections. This should be done systemwide, even on local buses, as is normal in Paris, Singapore, and every German-speaking city; turnstiles are only worth it on extremely busy trains (nothing in the US outside New York) and maybe also legacy subways that already have them. To discourage fare dodging, there should be a large unlimited monthly discount, as well as unlimited 6-month or annual tickets, so that most riders will be prepaid; the unlimited monthly pass should cost about 30 times as much as a single ride even with multi-ride discounts.
6. Intermediate-grade surface transit – i.e. the BRT and light rail lines providing service quality higher than a local bus and lower than rapid transit – should run in dedicated lanes, except perhaps on outer branches. Bus lanes should be physically separated, and tram lanes could even be put in a grassy median. Except for special cases where one side of the street is much more important than the other, in which case one-way pairs may be defensible, those dedicated lanes should be in the median of a two-way street, when street width permits it, which it does everywhere in the US except the North End of Boston and Lower Manhattan.
7. Intermodal transfers should be painless. Commuter trains should run through from one side of the region to the other, to allow for efficient suburb-to-suburb travel, and the infrastructure should be upgraded to allow for such operations. It should be unthinkable to terminate transit short of its natural destination. Though transfers at the originating end are unavoidable, planners should still endeavor to place rapid transit stops at every walkable place the line intersects, and achieve adequate speed by running better rolling stock. (In contrast, bus stop spacing should be 400-500 meters, rather than 200-250 as is common in North America). Parking lot commuter stations should be rare; they impede reverse-peak traffic, are expensive to provide, and help ensure transit will be used only when there’s no alternative.
Any other important principles for transit, dear commenters?
