Redundancy is Overrated

The night before last, a Northeast Corridor Amtrak train derailed in Philadelphia, killing seven people. For some overviews of what happened, see Vox and Huffington Post. I am not going to talk directly about the accident here; it appears to be the same kind of derailment as on Metro-North a year and a half ago. Instead, I’m going to talk about the general issue of redundancy, which I saw people bring up in response to the train shutdowns that followed the crash. This is not the first time I hear about this; redundancy figures prominently into the list of benefits touted for new rail tunnels across the Hudson, allowing Amtrak to shut down the existing tunnels for repairs. Even before Amtrak proposed the Gateway project, transit activists talked about redundancy as a positive feature, for example Cap’n Transit. In this post, I am going to explain why, in public transportation and intercity rail, redundancy is in fact far less useful than other investments for the same amount of money.

First, let us list the various high-caliber rail networks of the world. In high-speed rail, the biggest networks are those of China, Japan, and France. None of them has redundancy, in the sense that there is more than one way to get between two cities on high-speed track. JR Central is building a second line from Tokyo to Osaka, but this is because the existing line is at capacity, running about 14 trains per hour into Tokyo at the peak; redundancy is a minor consideration. In regional rail, the busiest networks do have some redundancy, in the sense that if one line is shut down then people can take a parallel line, but this is because these networks are so busy that in most directions there’s enough demand to fill multiple lines. In Tokyo, which has the largest regional rail network, the parallel line is usually run by a competing company, so within each company’s network there’s little redundancy.

The reason for this non-redundant operation is simple: building new rail lines is expensive, while maintaining them adequately so that they don’t break down is cheap. Amtrak thinks that the Gateway tunnel will cost $16 billion. The program to repair the damage the preexisting tunnels suffered in Hurricane Sandy is $700 million, which assumes an accelerated construction schedule in which the tunnels will be shut down one track at a time, but conversely also includes work in the worse-damaged East River tunnels and not just the tunnels across the Hudson. This is a one-time repair after salt water intrusion, not annual ongoing maintenance. New Hudson tunnels are a necessary project for capacity reasons, but whatever benefit they have for redundancy is a fraction of their cost.

For high-speed rail, too, the costs of maintenance are far smaller than those of construction. The average maintenance costs of a single route-km of HSR are about €100,000 per year, versus €20 million for construction (see PDF-p. 9 of a study by Ginés de Rus about HSR between Stockholm and Gothenburg). With this amount of maintenance, there need not be any closures or disruptions in service.

Consider the Northeast Corridor, more concretely. To guarantee redundancy everywhere, so that train accidents do not disrupt the line, is to restore some passenger service along the former Baltimore and Ohio and tie-ins. Between Philadelphia and New York this means the West Trenton Line; between Philadelphia and Washington this means the CSX freight line. This also requires new Hudson tunnels. The cost of each of these elements is in the billions, and for the most part, with the exception of the new Hudson tunnels the transportation benefit is very low, especially south of Philadelphia, where there aren’t enough people to justify a second commuter line. Between New York and New Haven, there are no good alignments for a second route except for short bypasses; that’s what makes constructing HSR there so difficult.

Redundancy is a good feature of networks where failures are frequent and unavoidable; for such systems, redundancy is useful, as is the concept of failing gracefully. Rail transit is not such a network. It is both possible and desirable to reduce accident rates to levels approaching zero. Natural disasters remain hazardous, but are extremely infrequent, and at any rate when a deadly earthquake strikes, there are higher priorities than providing alternative passenger rail routes.

This is not to say that redundancy has no uses. Dense subway systems are redundant in the sense of providing multiple routes through the city – although, at the peak, they’re usually all very crowded. This makes it possible to shut down lines off-peak for maintenance; New York and London are both notorious for weekend service changes, and Paris shuts down short segments of lines for maintenance for a few weeks at a time (see for example here). But small subway systems manage to make do with just ordinary overnight shutdowns, and Copenhagen even runs trains 24/7, shutting down one track at a time at night and using the driverless operation to run trains on single track. It’s just more convenient to have more options, but not necessary.

The upshot is that when a subway or mainline rail network chooses where to lay additional lines, it should ignore all needs of redundancy, except possibly as tie-breakers. The benefits are there, but do not outweigh the cost of building less optimal lines. The operator should instead invest in systems, worker training, and maintenance regimes that ensure high reliability, and expand the network based on ordinary criteria of expected ridership and capacity needs. There’s no need to worry about failure, and it’s much better to design the network not to fail in the first place.


  1. the0verheadwire

    What about in 1989 when the Loma Prieta quake took out a section of the Bay Bridge for months and BART was all people had? Seems like travel redundancy is important. Or am I missing the point?

    • Ian Mitchell

      Wasn’t the LIRR constructed as an alternative route between Boston/Providence and midtown manhattan that didn’t require going through New Haven?

      • Alon Levy

        No, the LIRR was the first route – it avoided New Haven because in the 1830s people thought it was impossible to build a railroad in the hilly terrain of coastal Connecticut.

        It wasn’t about Midtown, either. Midtown was exurban then. The LIRR’s terminus was in Brooklyn, with a ferry connection to Lower Manhattan. What is now called the LIRR Main Line was not the original main line, and only became so designated later, when Midtown’s importance grew.

  2. anonymouse

    All of the HSR networks of China, Japan, and France have a reasonable amount of redudancy: not within the HSR network itself, but between HSR and conventional rail. In France, this means that TGV trains can be rerouted onto conventional lines if there’s a problem on the high speed line. In Japan, the trains can’t inter-operate, but passengers can relatively easily transfer onto the conventional network and back to bypass a blocked section. Here, the rail network is sparse enough that there’s generally just one route, and if that is broken, it’s up to buses to pick up the slack, which they may or may not have enough capacity to do easily.
    Even with that, though, there’s plenty of resiliency/redundancy. Amtrak could, if they really wanted to, terminate at least the trains coming from Boston/Springfield at Trenton, and if they asked nicely, maybe SEPTA could even let them extend some trains from DC or Harrisburg to West Trenton, which leaves a relatively small gap that’s relatively easily covered by a bus bridge.

    • Alon Levy

      Japan doesn’t actually make use of this redundancy, though – JR Central does not shut down the Tokaido Shinkansen and tell people to ride the Tokaido Main Line.

      • orulz

        Though they don’t shut down the Shinkansen for routine maintenance (they accomplish that overnight), Service disruptions do indeed occur on the Shinkansen. Suicides, external reasons like a structure fire too close to the line, and yes – occasionaly even breakdowns. In those cases, they do advise people of possible detour routes. The only portions of Shinkansen track not closely duplicated by a completely separate, high-capacity, high-quality, double-tracked, electrified mainline are the “Mini” Shinkansen to Yamagata and Akita, and a few years from now the Seikan tunnel.

      • Andrew in Ezo

        JR Central as a rule does not provide for redundancy in the case of line shutdown- typically if an incident occurs along the Tokaido Shinkansen, the whole line is shut down until normal operational status is restored, and full refunds are given to passengers. However, with the Yurakucho fire last year adjacent to the ROW, it was estimated that JR Central lost approx. $30 million in revenues that day, so the company has begun rethinking their policy. This month (the 28th) a drill will be held around midnight at Shin-Yokohama station simulating a service disruption on the Shin-Yokohama-Tokyo stretch of the line. A turnback operation will be simulated at Shin-Yokohama, where a formerly Tokyo bound “up” service will crossover to the “down” line to re-commence service to Nagoya/Shin-Osaka. Naturally, in an actual disruption, passengers bound for Tokyo will be directed to get off at Shin-Yokohama and transfer to legacy lines for the last leg of the journey.

    • F-Line to Dudley

      That is exactly what they’re doing. Shuttle bus is in effect West Trenton-Trenton with SEPTA and NJ Transit cross-honoring tickets:

      FWIW…NJT has a study collecting mold in a file cabinet about restoring commuter rail service on the West Trenton Line, which stopped running to West Trenton in 1982 and stopped thru-running to Philly/Reading Terminal in 1981. Would fork off the Raritan Valley Line as diesel, and use the ALP-45DP dual-modes for Penn: $219M startup estimate for stations + capacity upgrades to the otherwise pretty well-maintained CSX line. SEPTA didn’t show any reciprocal interest in a joint venture running all the way to Philly, so it was strictly for NJ in-state constituencies. And obviously hasn’t been acted upon, or is likely to anytime soon.

      But even if it did happen and NJT could lend Amtrak some ALP-45DP’s to thru-route an uninterrupted Regional via Center City and W.T. to NYP in a service disruption, the chaos it would inflict on SEPTA commuters may make the cure worse than the disease. There’s no way they’re set up to handle the full force of a Regionals schedule at rush hour on the two-track SEPTA Main Line without a lot of commuter trains having to get bumped. And that’s 3 SEPTA lines, and 2 of their 3 highest-ridership lines that would feel it. Their signaling density isn’t predicated on mixing commuter + intercity traffic, and electrical capacity on the ex-Reading half of the system is really thin for threading big honking Regionals through Center City to the Main Line. The NJT dual-modes might not even have enough running room to fire up the diesel engine before crossing the Amtrak-to-SEPTA phase break and overstraining the SEPTA grid. Meaning somebody’s trains have to get canceled or delayed to keep the Wayne Junction substation from collapsing under load.

      You get the point. SEPTA-specific lines aren’t nationalized infrastructure; they have to answer to the needs and service levels of their own commuters first and foremost. So ‘possible’ re-routes on their backs aren’t a reasonable real-world expectation when that entails requiring SEPTA to maintain SEPTA-owned infrastructure way over-capacity for SEPTA’s commute-only needs of the day…for sole sake of sacking SEPTA commuters with full-on Amtrak priority once every freak event where the NEC goes tits up in that area. Frame it more in terms of an outside authority deeming “let’s sacrifice Greater Philly for the sake of Greater D.C. and Greater NYC” and it seems a lot less acceptable on-spec.

      That, unfortunately, is going to be the case on a lot of redundancy ‘possibles’ up and down the NEC. Cure worse than disease in terms of aggregate disruption at local level for some noble notion of a few days of ‘greater good’ at regional or national level.

  3. Adirondacker12800

    You consider building another set of tracks when the current set of tracks are nearing capacity. Frequently building the new set someplace else makes sense. Spreads out demand.

    Between New York and New Haven, there are no good alignments for a second route

    Sure there is, the LIRR to Yaphank, up the wide parkway to Shoreham and across the Sound to New Haven. From New Haven the obvious alternate is via Springfield. Philadelphia to New York via West Trenton is almost a no brainer. Someday NJTransit is gonna wake up and decide that SEPTA should be hauling Pennsylvanians to Manhattan not NJTransit.

    7 million people on Long Island would generate a lot of passengers. So would the people along the old Reading/CNJ line through West Trenton. At the 2010 Census :
    Buck County 625,255
    Montgomery County 799,873
    Mercer County 367,508
    Somerset County 323,441
    Just over 2 million people. Half million less than metro Baltimore. Or about the same as Hartford/Springfield. Who get stuck in worse traffic. Big park-n-ride at the intersection of I-287 would make it real easy to use for most of Morris County… IIRC NJTransit was estimating half a billion to re-lay a second track and upgrade the signaling. It may have included eliminating the few grade crossings.

    The B&O more or less built it’s new ROW between Philadelphia and Baltimore ten, twenty blocks west of the old ROW the PRR was refusing to share, so even if there were a lot of people to serve it wouldn’t be worth it. Connecting the Camden Line to the Penn Line where they cross over might be a good idea since it would be relatively cheap.

    …..ridership continues to creep up by 2040 the tunnels into Manhattan are at capacity again. Digging a tunnel from Brooklyn to Jersey City shifts traffic out of Midtown and as side benefit provides some redundancy. Digging a tunnel from North White Plains to Rahway … would be nearing capacity as soon as it opens….

  4. johndmuller

    In the long run the NEC itself is the backup route, all we need is the Main Line.

    Meanwhile, there might be a few baby steps taken to ease recovery from some sorts of intermediate level of disruptions. That is to say, something more disruptive than the current crash damage, where service was interrupted for 1 to 3 months. This kind of time frame would not be easy to just wait out, and a longer time frame, say of a year or more, could probably be dealt with by building a temporary rail bypass, if the political will were there.

    For these smallish disruptions, a little emergency planning could go a long way. Building strategic connections, like Adirondacker12800’s Camden/Penn connection could help ameliorate incidents between Baltimore and DC. Constructing some of Amtrak’s proposed High Speed Route between NYC and Massachusetts could provide some relief in that sector; even rebuilding/upgrading small sections, like Brewster to Danbury could allow for limited through service or at least the ability to relocate equipment stranded on the wrong side of an incident such as the collapse of one of those hundred year old bridges in Connecticut.

    • Alon Levy

      Yeah, that’s the problem with redundancy-oriented thinking: Brewster-Danbury is like priority #50 for mainline rail lines. It’s one of those “could be useful, but, assuming away New York’s construction cost premium, there are $50+ billion’s worth of more important projects in the region” situations. To say nothing of the fact that buses exist and will beat any roundabout legacy rail alignment.

      In areas that are less remote than Brewster-Danbury, there’s the whole question of whether it’s useful to have two routes rather than one. In many cases, it’s more cost-effective to realign a legacy line than to bypass it (and in many others it’s the reverse). For example, on Twitter, Ari Ofesvit and I were discussing ways to ease the curve at Frankford Junction. I’ll defend the proposal I made at the end, to realign Frankford to a wider S rather than bypass it over Torresdale, which would require several kilometers of urban el.

      If the old bridges in Connecticut are likely to collapse, train service over them should be shut down immediately. That’s the worst of redundancy-oriented thinking: spending money to make accidents less disruptive to the people who don’t die in them rather than to prevent those accidents in the first place. This cascades to things that do not involve deaths, too. US railroads spend extra money on crew costs to excessively pad their schedules rather than on reducing delays. The MBTA takes the cake in incompetence: it doesn’t construct full-length high platforms on lines with (minor) freight service because a freight train with a flat wheel might sway outside its loading gauge, hit the platform, and derail, whereas with a one-car mini-high it could hit the platform safely.

      • Adirondacker12800

        which would require several kilometers of urban el.

        The Market-Frankford El is in the way. Whatever gets done has to consider the traffic across the bridge to New Jersey.

      • F-Line to Dudley

        No, Alon, that is not true. Please do not twist around the perfectly clear explanations you were given on that ArchBoston discussion about the physics of freight cars around high platforms for sake of a prefit narrative.

        — Flat wheels are irrelevant. Every train car in the world, from a heritage trolley to metro systems to first-world HSR all the way up to the most gigantic freights…will develop flat spots from steel friction as a normal part of day-to-day operation. If you have ridden trains enough times anywhere in the world any time in your life, you have ridden on flat wheels. Full-stop.

        — Long railcars and tall railcars have different suspensions allowing for different levels of lateral movement because of the way the axles on long railcars take curves and the way tall railcars have higher centers of gravity. A bi-level passenger car is going to sway more than a single-level for exactly the same reason. And, yes, the padding at the edge of a full-high platform is because a bi-level can sometimes *gently* and *inocuously* bump the edge, especially if the platform is on a light curve or at the immediate end of a steeper curve.

        — The loading gauge on these wide-load lines–AAR Plate F or higher–is the ruling loading gauge of the line. The railcars are compatible with the loading gauge; the full-high platform is not compatible with the loading gauge. No one is running afoul of the safe loading gauge.

        — A full-length high platform (800 ft. / 9 cars per MBTA’s default new-build platform specs; 12 cars for a NE Regional new-build spec; your results may vary on other commuter RR’s) presents too large a target for the lateral harmonics of an entire freight consist of multiple tall cars (say, a train of autoracks). And that’s why the strike risk is too high, and why those platforms are not compatible with the ruling loading gauge.

        — The single-car mini-high with retractable edges ARE compatible with the loading gauge. So they end up the preferred alternative when freight passing tracks are not available on an exempted wide-clearance route, and gauntlet tracks (which tend to have too many ops compromises of their own) are not an acceptable alternative. The mini-highs achieve full ADA compliance with adequate safety by only having to deal with the lateral sway of 1 individual car at a time. If struck, it is almost always limited to the carbody between axles and not the edges or areas around the couplers (i.e. if there’s a strike, that’s far and away the safest place to hit). The retractable edge is designed to be semi-disposable and cheaply replacable, and only requires 1 advance crewmember to flip/reset a lever. Why do the mini-highs still occasionally get struck when their edges are retracted? Because you cannot assume that a freight train comprised of railcars that could have originated from any of the 48 states, or nearly every province of Canada or Mexico, won’t have one single car in it with a top-heavy load. You could also not make that assumption in some places in Europe, either, though with their far lower freight volumes traveling far shorter distances, and rail networks less chewed-up by mass abandonments means they have far more options for workarounds than East Coast U.S. There are, however, individual lines on the continent that have to juggle this same issue of heterogeneous loading gauge traffic all the same. It is not wholly unique to the Eastern U.S.

        — It is not the passenger agency’s “incompetence” that results in the very presence of freight-clearance mini-highs. Wide-load freight routes are a permanent federally-protected exemption that can only be waived voluntarily by the freight carrier. The MBTA could not, if it wanted to, make a case that so-and-so’s freight traffic is ‘light’, therefore they need to remove the exemption. The exemptions are protected by the full force of interstate commerce law, and it would take a Supreme Court case to reverse that. That is a century-plus of reaffirmed caselaw, not administrative incompetence. And in the case of the eastern railroads, the states and Amtrak getting nearly free public ownership title to these lines in the 1970’s when Penn Central, Boston & Maine, etc. went bankrupt was a vanishingly small price to pay for giving the bankrupt freights a breadcrumb of perpetual Plate F+ exemptions in the surrender terms. So…you know, perspective.

        — Yes, you can theoretically pay “go away” money. MassDOT did just that on the inner Worcester Line, albeit with a significant revenue motive for the state’s coffers tied up in the freight traffic relocation. But some freight companies make more revenues over X years than one-time “go away” money would offset, and aren’t going to be fazed by it. NJT has this problem in a few places; it could never make a deal good enough to boot the freights, because booting the freights means the freights have to take the NEC instead. In other cases, blowing wads and wads of “go away” pure paper transactions is simply too frivolous a use of resources to merit drawing away from far more meaningful enhancements.

        — This is the price the East Coast pays for stepping up its level-boarding coverage to a large scale in the post-ADA era. NJT, SEPTA, MARC, and Metro-North west-of-Hudson have far more conflicts with Plate F-or-greater routes than the relatively meager number of affected MBTA stops that have no workarounds with passers. And far more cases where mini-high installations have to go in-play for lack of available room to do passing tracks, as those agencies churn through accessibility upgrades.

        You can quibble with the workarounds–“go away” money, how far a transit agency should stretch itself to find passing track room, etc. But please do not misrepresent the basic facts of what a freight clearance route functionally and legally is for purposes of furthering an argument about the stupidity of American transit policy. Throwing darts blindly at a wall would ding something stupid about American transit policy with no effort; there are innumerable ways to do that without resorting to factual inaccuracies.

        • Adirondacker12800

          The foamers just love to go down to the station and watch the freights go past the high platforms at Roselle Park NJ. So much that they take videos of it. And post them on YouTube.

          Union NJ too but that’s harder to find because YouTube tends to spit out stuff relating to Union Pacific.

          • F-Line to Dudley

            Roselle Park’s got a gauntlet track on the eastbound side. That’s how the over-wide freights pass the full-high. The consecutive Roselle Park and Union stops on the Raritan Valley Line each have those setups because they’re sandwiched between where the Lehigh Line freight main converges then diverges. Those are the only gauntlets in existence in East Coast level boarding territory.

            The only other ones in the U.S. are on isolated level boarding systems: Hegewisch, Hammond, and East Chicago stations on the South Shore Line, and Hall-Nimbus, Tigard, and Tualatin on the Westside Express Service DMU line.

          • Alon Levy

            Most Upper Hudson Line stations have high platforms, no bypass or gauntlet tracks, and (a little bit of) freight service.

          • F-Line to Dudley

            The Hudson Line does not handle over-wides, and never did…not since Day 1 of the third rail electrification and installation of those full-highs. The N-S Putnam Division from Brewster to University Heights (abandoned by Penn Central 1970) was the east-of-Hudson clearance route into New York, and the preferred clearance route PERIOD into New York until the Central enlarged the tunnels on the west-of-Hudson West Shore Line in the early-60’s. The Putnam was abandoned because all that traffic shifted west-of-Hudson.

            The Lehigh Line is one of the busiest intermodal lines in the entire country today, and pools shared CSX and Norfolk Southern traffic (a.k.a. “Conrail Shared Assets, North Jersey Div.”). That short amount of Raritan Valley Line overlap between freight-only Lehigh segments is as mission-critical as it gets.

            Clearance routes are the exception, not the rule. This is not an all-or-nothing thing. Piecemeal ADA on the vast majority of *non*-clearance lines that are A-OK for 100% full-highs–freight or no freight–is a legitimate institutional and funding problem. I have no disagreement on that point. But it is not, and never was, a problem in the limited places where legitimate co-existence is non-optional for the health of the transportation economy. It is a sliding scale. And it’s a disingenuous misrepresentation to lump it all together into the same black-and-white “everything is terrible there; everything is great tthere” narrative. You’re undermining your own arguments, Alon, by continually going to that well treating all factoids as equally damning indictments. Transportation networks are not that simplistic…not operationally, not in the number of masters they have to serve for the basic functioning of the regions they serve. You know this…you routinely articulate this when it’s anywhere but the U.S. So why does it keep going down this rabbit hole on irrelevant stuff like this?

          • Alon Levy

            You really don’t need to say “you’re undermining your own argument, Alon” ten times. Saying it more times won’t make it less wrong. You think it’s fine for railroads to let trains sway outside the structure gauge/build platforms inside the dynamical envelope, I don’t. And you think the reliability margins are lax enough that it’s okay to have retractable platforms one car long rather than eight cars long. That’s the difference (it’s really the same issue of how reliable you want the system to be).

            At any rate, the barrier to full high platforms everywhere in the US without oversize freight is not funding. Caltrain’s reconstructed nearly every station in the last 20 years, but never bothered to raise any platforms during these projects. In the Boston area, the Fairmount Line upgrades were $6 million per station at some stations, and included more than just raising platforms. The recent cost overruns on South Coast Rail – let alone the total cost overruns – would pay for this at every MBTA station without oversize freight. It’s just a matter of priorities. The commuter railroads just don’t care about level boarding, except in the context of providing de minimis compliance with the ADA.

          • F-Line to Dudley

            “You really don’t need to say “you’re undermining your own argument, Alon” ten times. Saying it more times won’t make it less wrong. You think it’s fine for railroads to let trains sway outside the structure gauge/build platforms inside the dynamical envelope, I don’t. And you think the reliability margins are lax enough that it’s okay to have retractable platforms one car long rather than eight cars long. That’s the difference (it’s really the same issue of how reliable you want the system to be).”
            Giving real-world citations of the complexity of railroad operations on a sliding scale of importance to the question at hand is fundamentally different from pounding the “first-world countries don’t ____” desk without substantiation or dilineation. This isn’t about opinion or intensity of belief, it’s about substantiating an argument. Without resorting to logical fallacies. Flat wheels are not an operational hazard in the real worldwide world; they happen around the world right under passengers’ noses every day. Legally-protected clearance envelopes and legally-protected accessibility requirements have a need to coexist in *very small overlaps* in the real worldwide world (<– the *emphasis* the linchpin of that statement). Consist-long harmonics vs. an 8-car retractable platform are outside the legal and regulatory-vetted safety and reliability tolerances while single-car harmonics vs. a 1-car retractable platform are inside legal and regulatory safety tolerances. Those are neutral factual statements or neutral real-world conditions. Not duels of personal beliefs. Not the thin red line between civilization and savages. Merely the arcane, mundane–and, frankly, dreadfully boring–day-to-day world of mainline railroad operations. No more, no less.

            Believing harder than anyone else that these things are not true, believing harder than anyone else that these things are equal and interchangeably awful to exclusion of relevance and degrees of difference, and believing harder than anyone else that a perfect transit abstraction is more righteous than minor facts that don't fit a narrative…does not make it any more relevant to improving transit in the real world. Nor does clarifying that those facts exist signify equal-and-opposite devotion to the righteousness of N.I.H. If there is a counterargument of suggested improvement to make in these areas, by all means make an evidence-backed argument. Simply denying that neutral conditions exist as if this were a battle between good and evil is a giant waste of time and energy. You're better than that.

            "At any rate, the barrier to full high platforms everywhere in the US without oversize freight is not funding. Caltrain’s reconstructed nearly every station in the last 20 years, but never bothered to raise any platforms during these projects. In the Boston area, the Fairmount Line upgrades were $6 million per station at some stations, and included more than just raising platforms. The recent cost overruns on South Coast Rail – let alone the total cost overruns – would pay for this at every MBTA station without oversize freight. It’s just a matter of priorities. The commuter railroads just don’t care about level boarding, except in the context of providing de minimis compliance with the ADA."
            I don't disagree with any of this. What does that have to do with dynamical envelopes and freight clearance routes? These are not part of the same argument. Nor does the fact that freight exists on non-clearance routes–occasionally in large volumes–bind the argument about half-assed ADA compliance on non-clearance routes to the legal means of ADA compliance on a clearance route. They have nothing to do with each other. Make every station on a non-clearance route level boarding. Please…yesterday. Nothing freight or passenger will ever bump those more than will harmlessly bounce off the wood/plastic/rubber bumper (as they do on 51" platforms in the first-world. Yes, they do; it happens. No, really.). Clearance routes: that's the tougher question. They exist, and they can't not exist. So what are the feasible options? Is there room for a passer? What happens with those couple of tough outliers that can't be modded…can the schedule smoothly absorb an unfavorable dwell at a mini-high at those outliers if everything else went level? How dense does traffic have to be before the mini-high dwell is any sort of worry, and which lines have low enough density ceilings that it'll never make much sense to get too concerned? What's the pain threshold for installing a gauntlet; what can we learn from NJT's Raritan Valley Line installations to help inform those decisions?

            Those are relevant questions. Caltrain doing past-tense station rehab reaching back to Pete Wilson's first term in office on a line that's still at-present and for a minimum of the next half-dozen years in low-level territory with low-level boarding equipment…MBTA boondoggles yet-to-be-constructed (and probably now placed on hold)…and MBTA projects with full-disclosure "more than just raising platforms"…are not relevant to the question of barriers to full-highs "everywhere". Clearance routes vs. non-clearance routes are. What is this unfocused rant even about? If there's a case (there is) to be made about stepped-up ADA compliance to the full spirit of the law, at least cite the fodder that's halfway relevant to the point. Not every object within arm's reach is equally damning or equally to the point.

          • Alon Levy

            Yes, those are relevant questions. But it’s equally relevant that American commuter rail agencies, with the exception of the LIRR and Metro-North, do not have any interest in level boarding. Even when the incremental cost of providing it zero, they do nothing. That’s the US passenger rail environment, which goes a lot to explain my complete lack of faith in “regulatory-vetted safety and reliability tolerances” that point in certain directions.

            So in this case, consider various solutions that offer level boarding even with extra side clearance. They’re not being studied in the US, because the agencies don’t think level boarding is important. Given that they don’t, why is it so obvious that platform strikes are inevitable? Why is it so obvious that it’s not possible to just buy out the carcasses of Pan Am when Pan Am goes belly up? Etc.

            I also don’t get the “Pete Wilson’s first term” dig. I especially don’t get why it’s there in a paragraph that accuses me of ranting without focus. If what you’re saying is that it’s in the past and isn’t relevant to the future, then I disagree; it’s still the same railroad culture.

            The bit about flat wheels is, as I recall, due to an example you gave on ArchBoston of something that could lead a freight car to hit a high platform and derail. At least the way I understand your comments, those trains aren’t actually oversize. Their static dimensions are within the normal Plate F loading gauge, which still leaves 3″ between the train and a high platform. They just sway a lot, and this is something that can be fixed with better maintenance standards – after all, the Shinkansen manage to pass platforms 7 cm away at 200 km/h! In the link above to Clem’s presentation, he mentions slab track as something that could reduce sway, allowing trains to go at their normal full speed within 3″ of high platforms (although, I’d note that the Tokaido Shinkansen uses ordinary ballasted track and achieves the same).

            As for mini-highs and schedule risk: okay, now we’re talking. The schedule risk that each low platform poses is approximately the same, modulo issues like traffic (Beverly is a bigger risk than, say, River Works). So for the most part, it’s linear – each low platform adds a similar amount of dwell and required padding to the schedule. The extra dwell of a low platform is normally 15 seconds – 45 vs. 30 – but it’s possible to push 20 at minor stations with level boarding. There’s the extra schedule risk, which I don’t know quantitatively. But, shared segments between likely overtakes, i.e. Mansfield through Canton Junction, need to have high platforms because the schedule’s more fragile there than elsewhere.

            The other issue is that the optimal door placement is at the quarter points, which makes trapdoors impossible, which means the trains can only serve a relatively narrow range of platform heights, too narrow to encompass both 200 mm and 1,220 mm. As with electrification, converting the last low platform to level boarding has extra benefits beyond the usual marginal benefits of converting one more low platform.

          • Adirondacker12800

            Roselle Park’s got a gauntlet track on the eastbound side. That’s how the over-wide freights pass the full-high.

            Yes they do. What’s more reliable a couple of switches or retractable steps? Which is cheaper to maintain? Cheaper to build? Which one gets passengers on and off the train faster so the train can leave the station quicker, getting out of the way of the next freight train bearing down on it?

        • Alon Levy

          I’m not complaining about the existence of freight; I’m complaining about the way the MBTA builds platforms inside what the industry appears to accept as the dynamical clearance envelope of trains, on the grounds that it’s not a big deal if the freight trains hit a short platform. And if passenger trains bump the platforms… no. Just, no. It’s not normal or inevitable practice – the subways of the world manage without it, the high-platform HSR operations manage without it. It doesn’t matter how the physics work; advanced first-world railroads design a dynamical clearance, don’t build obstacles inside of it, and don’t let the trains go outside of it. The incompetence is not in not kicking out freight; it’s in assuming that platform pounding is inevitable and level boarding is an ADA-based luxury. It comes from the same place as “Asians don’t value life the way we do” and “the US has heavier trucks than Europe” and other flights of fancy that come out of national ignorance about the rest of the world.

          • Nathanael

            FWIW MBTA’s policy is currently to build full-highs at every station. This has been true for at least a few years now. I can’t speak to their *prior* policy.

          • Alon Levy

            I know, I know. But they do not seem to care much about rebuilding mini-high platforms as full highs, and on the contrary build platform infrastructure that makes transition harder, like the South Attleboro overpass.

          • F-Line to Dudley

            — Passenger trains do bump the platforms. Do. They do. In the first-world, they do. On many systems. Every day. For a variety of reasons. For example, the undercarriage airbag of a subway car can mid-run develop a leak and fall out of alignment, causing a wobble that may scrape the plastic or wood platform edge. That’s one of the more mundane examples. Air leaks the same way in the Eastern hemisphere as it does in the Western. It is not a problem. Nor are nearly all other kinds. Passengers on the train rarely notice it, and that is why you never hear about it. There is not an angle in every…single…micro-…detail…of basic rail operations that prefits your narrative indictment of U.S. rail policy, Alon. This is in the same category of complete irrelevance as flat wheels.

            — The “industry”–or the evil empire of FRA, railroads, railcar manufacturers, and standards bodies–decides that mini-high platforms are the preferred solution for ADA compliance on high-boarding cars (in absence of passing tracks) for safe coexistence with the dynamical clearance envelope of routes designated Plate F or greater. Not the MBTA; that does not prefit your narrative, either. All regions outside of the East with 8-inch platforms do not have to deal with this. Nor do the non-clearance routes (including the NEC), which in passenger territory vastly outnumber the clearance routes. But it is an inevitable compromise on the East Coast. Routes overlap with other routes that has to interface with 3rd rail territory around NYC, so level high-boarding cars are a necessity.

            And all of the bigtime dedicated freight corridors do a great convergence to the big ports and Interstate trucking corridors on that last several dozen miles before the coast. Short-distance overlap with commuter traffic and commuter stations is unavoidable as freight lines merge/diverge. Just look, with all the Google Maps action on this NEC accident, at all the crisscrossing freight lines within a 5-mile radius of Frankford Jct. right in the heart of the 5th largest city in the U.S. There’s nowhere else to put those. Some very active spurs are only reachable with sojourns on a shared passenger line.

            And for the record, very few number of freight trains run cars that are Plate F in commuter territory. You’re talking autorack trains, gigantic intermodal trains traveling 200-500 miles, super-heavyweight loads of coal, and so on. Trains that super-long, but with loads hyper-consolidated into just a couple big runs per day (or, most often, the dead of night). Rarely on the last-mile locals. And never tanker trains; the biggest rail tankers on the continent don’t even swing close to a full-high. On the MBTA the Fitchburg Line Ayer-west, the Haverhill Line Ballardvale-north, and the Worcester Line Framingham-west (but overwhelmingly Westborough-west) are the only places where the retractable platform edges get raised on a daily basis during passenger hours. Lowell Line…the overnights are the only one, and the T just retracts the edges at after the last train when it empties the parking lot fareboxes. Franklin Line…maybe once or twice a week depending on what customers are scheduled to unload at Readville Yard on the outskirts of Boston on a given day. The situation is similar on MNRR, NJT, and SEPTA: most often on the routes that overlap the intermodal freights that have to get to the coasts, infrequently or after-hours everywhere else.

            — ADA is not a “luxury”. It is federal law upheld multiple times over by the Supreme Court, and backed by all kinds of public transit-specific caselaw. It is required that new-construction public infrastructure comply with the law. This is what they have to do to comply with federal law. That does not prefit your incompetence narrative, either.

            “Flights of fancy” are citing irrelevant day-to-day operational realities of railroads worldwide, and unchangeable court-reinforced federal laws, as narrative-building fodder about U.S. transportation incompetence. And treating as if it is equally relevant and mutually supportive of all other narrative-building fodder about U.S. transportation incompetence. I don’t understand why you would feel the need to do that, since the U.S. is virtually DROWNING in fully legit and relevant pants-on-fire transpo planning incompetence. It weakens and dilutes, not strengthens, a fully legit argument that needs to be shouted from the rooftops. A little more focus would go a long way.

          • Alon Levy

            Where did I say ADA is a luxury? No, I said that most (#NotAll) American mainline railroads view level boarding as a luxury and only provide it because the ADA makes them, whereas here the operators understand how important it is for reliable commuter rail operations and move toward providing it even without such a law.

            That’s where the mini-highs come from. It’s a minimal solution to satisfy the letter of the ADA without actually providing level boarding. It has nothing to do with clearance: San Diego has low-floor trains and 550 mm mini-highs, even though the part of the freight load that’s oversize will clear a 550 mm platform.

            And no, platform pounding is not normal. Usually higher-speed routes have bypass tracks at stations where trains don’t stop, but normally they don’t. At Atami, there aren’t any bypass tracks for the Shinkansen. It’s near an old curve so the trains don’t go 270 km/h there, but they still go 200 km/h, with only 70 mm between the platform and the train (vs. 4″ in US NEC standards if cars are 10′ 6″ wide). Somehow that’s not led to any pavement pounding disaster, even though hitting obstacles at this speed is often catastrophic.

          • Nathanael

            What year was the aforementioned South Attleboro overpass built? And who built it (MBTA or the city)? And how old was the design approval?

            I do know MBTA policy on this changed fairly recently. I’m not surprised that the totally-inaccessible stations are higher priority than the ones with mini-highs. Building infrastructure which makes it harder to build full-highs would certainly be bad, but I really haven’t seen much of that in the last few years.

  5. threestationsquare

    I agree with your conclusion, but want to point out that there’s actually a lot of redundancy on the Chinese rail network. For example between Beijing and Tianjin, you have a choice of trains on the Jingjin HSR (33 minutes from Beijing South to Tianjin Main), the Jinghu HSR (41 minutes from Beijing South to Tianjin South), or the legacy Jinghu railway (75 minutes from Beijing Main to Tianjin Main). Similarly between Shanghai and Nanjing there are two high-speed lines and a legacy line. Of course, those lines don’t necessarily have the capacity to absorb all of each other’s passengers without inconvenience and delays, but the failure of one is unlikely to leave people actually stranded for long.

    Complaining about redundancy in connection with the Frankford Crash seems particularly weird because that section of the NEC is already redundant. From a passenger’s point of view (if they are for for some reason allergic to buses) one can get from Philadelphia to Trenton NJT by taking PATCO to the NJT River Line. From Amtrak’s point of view, if they really wanted they could run a few dual-mode trains along the CSX and SEPTA lines via Bound Brook and West Trenton, similar to the Inland Route (via Springfield) trains they ran during the New London bridge reconstruction; they’ve chosen not to bother. And of course, the redundant option most passengers will use is the extensive bus service along the corridor, which carries more people than Amtrak under normal circumstances and which is used to handling holiday and similar spikes in demand. New York-DC Megabus fares this weekend aren’t even unusually high!

    Similarly there’s actually a huge amount of redundancy for the Amtrak Hudson tunnels. In a short-term disruption the obvious thing for passengers to do is transfer to PATH at Newark, but there are also 10 lanes of parallel road tunnels that could have their capacity greatly increased with HOV restrictions (as was done with the East River bridges after Sandy). Of course a closure of the Amtrak tunnels would be inconvenient and many people would forego discretionary trips as a result but again nobody would actually end up stranded.

    Situations where there’s actually no redundancy on a major route are potentially more cause for concern, but it’s hard to think of many modern examples. (Singapore-Johor Causeway before the Second Link opened?)

    • Adirondacker12800

      PATH is at capacity during rush hours. They could institute HOV only on the roads but then where are they going to get the buses and where are they going to pick up and drop off passengers?

      • Alon Levy

        They’d get buses by running rush hour frequencies off-peak as well, and they’d drop off passengers either at Port Authority or at various subway stations in Manhattan reasonably close to the Holland Tunnel. Traffic in the Holland Tunnel is infernal, but in emergency reroutes, people can’t be picky.

      • threestationsquare

        In an emergency/disruption situation HOV or access-only restrictions could also be applied to 10 or 20 blocks of midtown west, and around the Holland Tunnel, to use as curbside loading/unloading areas.

    • Alon Levy

      Between Tianjin and Nanjing, there’s just the one high-speed line and the legacy line. It provides redundancy in the sense that the trains can run on the legacy line if something goes wrong on the high-speed line, but it’s low-speed redundancy, not much better than the ability of I-95 to provide redundancy for the NEC.

    • Nathanael

      What you said makes sense. There aren’t that many places where there is *no* redundancy. And in transportation you generally only need a little bit of redundancy.

      BNSF has been hanging onto both routes across the mountains in Washington State because of redundancy. They basically only use one of them… but a shutdown of that tunnel would be such a disaster to their freight system that they hang on to the other one.

      Losing the BART tube across the Bay would probably completely snarl San Francisco for quite a long time, despite the Bay Bridge.

      I’m having trouble thinking of similar examples. Ah — something which took out both I-95 and the Metro-North line between Greenwich and New Haven would be catastrophic. (And they are adjacent.)

  6. Wanderer

    In terms of Caltrain, it’s at least been suggested that they might want rebuild the local station platforms to the same height as any High Speed Rail stations.

  7. po8crg

    The wash-out of the tracks at Dawlish last year is one of the few examples I can think of where redundancy would have made some sort of sense, but even there there were plenty of buses and coaches. The biggest problem wasn’t the long-distance services (which is what the, mostly London-based, media complained about) but the way the western road routes into Exeter gridlocked when all the commuter rail passengers switched to road.

    Even there, it’s not redundancy that’s needed – it’s a replacement route that doesn’t run along the top of a seawall.

  8. Pingback: There’s More Redundancy Than You Think | Pedestrian Observations

Leave a Reply

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

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

Twitter picture

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

Facebook photo

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

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.