Setting Speed Zones

At the Boston meetup two days ago, I was asked about what tools I use to generate timetables, for example for my New York commuter rail posts. The answer is that I use speed zones and then run this code on them – but then the question is how to figure out speed zones. I hope that this sequence of steps will help advocates who are interested in rail modernization.

Generating curve radii

The most difficult element to fix on mainline rail is the right-of-way geometry. Most other things that can restrict a train’s speed can be fixed with more modern maintenance, but right-of-way geometry doesn’t change without physical construction, often in constrained areas – if they weren’t constrained, the curves would have been built wider in the first place.

The best case scenario is that there exist track maps with exact curve radii. I have these for large chunks of the Northeast Corridor, but not all. For example, here is Metro-North (with thanks to The Korot). Curves on such maps are denoted as circles or bumps deviating from a line, with the direction of the circle indicating the direction of the curve’s curvature. On this and other American maps, the radius is listed in degrees, and the cant (see the section below) in inches.

To convert the radius from degrees to more usual units, set one degree to be 1,746 meters, and note that degrees measure curvature and not radius, so a two degree curve has half the radius of a one degree curve. More precisely, the formula is that degrees measure the change in azimuth over 100 feet; 100 feet are 30.48 meters, and converting 30.48 from degrees to radians gives 1,746.37536… meters.

For example, on the Metro-North chart, let’s look at Harrison, New York. It’s on PDF-p. 24 of the chart; Harrison is sandwiched between two curves with opposite orientations, with the platforms on tangent (uncurved) track. The curve just west of Harrison has radius 1° 58′ 30″, which is 884.24069… meters; the curve just east has radius 2° 2′ 15″, which is 857.11674… meters.

Converting curve radii to speeds

The formula for the speed of a train, in SI units, is

\mbox{speed}^{2} = \mbox{radius} \times \mbox{lateral acceleration}

I wrote about lateral acceleration, cant, and cant deficiency two and a half years ago. In short, lateral acceleration, in m/s^2, is the centrifugal force coming from the action of the train rounding the curve at speed. For the purposes of the formula, it is measured in the horizontal plane. To reduce the centrifugal force felt by the passengers (for comfort and safety) as well as that felt by the train body (for safety and maintenance costs), the tracks will typically be banked so that the inner rail is lower than the outer rail, which is called cant or superelevation, and is written in units of distance, such as mm or inches.

The speed of a train on canted track is typically higher than the perfect balancing speed, where the force of gravity counteracts that of centrifugal force; thus, on a fast train there is a residual force pointing to the outside of the train, which can be written down as lateral acceleration in the plane of the tracks (in m/s^2), but is more typically written down in the same units as superelevation, representing the additional superelevation required for the speed to perfectly balance, which is called cant deficiency or underbalance.

The conversion rate between cant (or cant deficiency) is the track gauge measured between the middle of the two rails, divided by the gravitational constant (9.8 m/s^2). Track gauge is typically given as inner rail to inner rail; standard gauge is 1,435 mm inner rail to inner rail. The relevant quantity to superelevation calculations is a few cm more; on standard gauge, it’s taken to be about 1,470-1,500 mm, so the conversion rate is 1 m/s^2 = 150 mm of cant or cant deficiency. The two quantities, cant and cant deficiency, are additive.

The American track charts that I have specify the actual cant. However, the values tend to be too conservative. Again with the example of Harrison, the slightly wider western curve has 5″ cant and the slightly tighter eastern curve has 4.125″ cant. Regulations for maximum cant depend on the country and maintenance standards. The absolute maximum cant I am aware of on any standard-gauge railway is 200 mm on the Tokaido and Tohoku Shinkansen. The reasons not to raise cant further include maintenance difficulties and the risk of a train running at lower speed or even stopping on the track. On lines that are not captive to just high-speed trains, the highest cant I am aware of is 180 mm, in Germany, and this is rare; 160 mm is more common. The American limit is 7″, but frequent inspections are required at that point to ensure that the tracks don’t get bent out of shape to produce higher cant; 8″ is a do-not-exceed level, and in practice track irregularities may lead to exceeding it if there isn’t regular track maintenance.

In practice, raising the cant is usually easy – it can be done with a track geometry machine automatically. However, in one case, it is not: that of S-curves, which have exactly the shape implied by the letter S. Our example of Harrison has two reverse curves in close proximity, but is not an S-curve, as there are hundreds of meters of tangent track between the two curves. Other places do have S-curves, and there, the maximum cant must be lowered somewhat; regulations vary on this, but in Europe, the maximum change in cant is 30-55 mm per second depending on the country (there’s a secondary regulation on mm per meter, but on the Northeast Corridor, the binding rule is mm/s, not mm/m). For example, if we take 45 mm/s, and 180 mm of cant, then it takes four seconds to reverse a curve; note that it is four and not eight, because half of the increase in cant, called a superelevation spiral, is within the curve. Harrison’s current curves impose a hard limit of about 150 km/h, at which point the hundreds of meters of tangent track make it trivially easy to have full superelevation. However, other places, most infamously among Northeastern railfans Elizabeth, there is an actual S-curve, forcing lower cant and lower speeds.

Finally, the maximum cant deficiency depends on the track, the train, and the regulations. Traditionally, American regulations limited most passenger trains to 3″ of cant deficiency, which is ridiculously conservative; in the 2000s, a waiver allowing 5″ in some cases was derided as the “magic high-speed rail waiver” because it applied not just to higher-speed tracks but also to lower-speed lines that had through-service to higher-speed tracks. Since then, FRA regulations have changed, and now the practical limit in the US, with extensive testing, is 6″, or 150 mm cant deficiency, at most speeds. European limits tend to be around 130-150 mm; high-speed trains are at the lower end of this range unless they are tilting trains, which nearly all trains are not. Cant deficiency, like cant, requires its own superelevation spiral on S-curves, but the limits are in practice looser than for cant, and in some cases trains can change superelevation abruptly, with no spiral, for example on switches.

The upshot is that in the typical case, the most aggressive assumption should be 180 mm cant, 150 mm cant deficiency, for a lateral acceleration in the horizontal plane of 2.2 m/s^2. Most lines will not have this pair of aggressive assumptions: 180 mm is only viable when it’s guaranteed that trains will not stop on a canted curve, which is a reasonable assumption on a reliable high-speed line and even on a German high-speed line. Moreover, if there is any freight on the line, superelevation must fall drastically: slower trains would be at cant excess, and freight trains have high center of mass (diesel locomotives and double-stacked containers both have higher center of mass than electric passenger trains) and therefore have tight cant excess limits. Aggressive assumptions are viable on the Northeast Corridor and on controllable commuter lines with no or almost no freight, such as the LIRR, but not everywhere else.

Finding curve radii

In some cases, curve radii are spelled out in a chart. In others, they are not, and must be figured out. Our program’s schedule writer, Devin Wilkins, tells me she has just found a track chart for SEPTA, but otherwise, I have no such charts south of New York. There, the dirty, imperfect method of estimating curve radii must be used.

For that, I use Google Earth. Nowadays, Google Earth Pro is free, and comes with a circle tool. In theory, I can fidget with the radius of a circle until I find that it approximates the arc of a curve well. This requires paying special attention to how the drawn curve compares with not just the broad outline of the curve but also the exact arc of each track or even each rail: the drawn curve should be at the same relative position to the rails, such as following one rail of one track, or right in the center of one track, or right between the track centers, and so on.

In cases of uncertainty, it’s also possible to use Google Earth line tools, which state the azimuth of each line. If I can find the exact start and end points of each curve, and the azimuths of the tangents on both side, then I can draw the chord with a line tool, verifying that its azimuth is the exact arithmetic mean of the azimuths of the two tangents; if it is not the mean, then either I made an error (more likely) or the curve is not a perfect circle (possible but less likely). The radius of the curve is approximately the length of the chord times 180/pi divided by change in azimuth; more precisely, the radius is

\mbox{chord length}/(2 \times \arcsin(\mbox{difference in azimuth})/2).

This method is error-prone, especially for short, sharp curves. Computing the start and end points of the curve will always have errors, and if the change in azimuth is small, then these will lead to large errors. The circle tool method suffers from the same drawback: it’s easier to use it to estimate the radius of a curve with 60 degree change in azimuth than that of one with 10 degree change.

Over time I’ve gotten this method down to the point that my errors from what I later find with track charts such as that of Metro-North are fairly small, and not very biased in the larger-radius direction. But it takes time and practice and ideally you should avoid it for short, sharp curves.

Update 10-28: Ari Ofsevit has a third method, using chord lines.

Other speed limits

Speed limits on intercity trains mostly come from curves. But there are other things to keep in mind, not all of which are fixable:

  • Tunnels increase air resistance to the point that unless the tunnels are constructed with large enough radius to have a lot of free air (modern tunnels are, legacy ones aren’t), or unless the trains are pressurized, the speed limit has to be lower just to avoid popping passengers’ ears at entry and exit.
  • Switches generally have low speed limits – they have tight curves and no superelevation – making all complex junctions and major stations slow.
  • Terminal stations have another set of speed limits coming from the bumper tracks. American limits are very conservative – 10 mph where a ramp down from 40 or 50 km/h is more normal in Europe – but even 50 km/h is not 200 km/h.
  • Some pieces of infrastructure are so shoddy that they limit the dynamical axle load of the train, which is derived from both static axle load, which is a function of train mass, and speed. As usual, American limits on this are conservative, assuming high static axle load and fixing a low speed on some very old bridges, instead of permitting lighter trains to run faster. But it is sometimes a real problem.

Boston Meetup and Consultants Supervising Consultants

The meetup was a lot less formal than expected; people who showed up included loyal blog readers (thank you for reading and showing up!), social media followers (same), and some people involved in politics or the industry. I don’t have any presentation to show – I talked a bit about the TransitMatters Regional Rail program and then people asked questions. Rather, I want to talk about something I’ve said on social media but not here, which I delivered a long rant about to the last people who stayed there.

The issue at hand is that the only way that seems to work to deliver complex infrastructure projects is with close in-house supervision. This is true even in places where the public-sector supervisors, frankly, suck – which they frequently do in the United States. It’s fine to outsource some capabilities to consultants, but if it happens, then the supervision must remain in the public sector, which requires hiring more in-house people, at competitive salaries.

Why?

The reason is that public-sector projects always involve some public-sector elements. This is true even in the emergent norm in the English-speaking world and in many other countries that take cues from it, in which not only is most work done by consultants, but also the consultants are usually supervised by other consultants. The remains of the public sector think they’re committed to light-touch supervision, but because they, by their own admission, don’t know how to do things themselves or even how to supervise consultants, they do a bad job at it.

The most dreaded request is “study everything.” It’s so easy to just add more scenarios, more possibilities, more caveats. It’s the bane of collaborative documents (ask me how I know). In the Northeast Corridor timetabling project I’m doing with Devin Wilkins, I could study everything and look at every possible scenario, with respect to electrification, which projects are undertaken, rolling stock performance profile, and so on. It would not be doable with just me and her in a year or so; I would need to hire a larger team and take several years, and probably break it down so that one person just does Boston, another just does Philadelphia, a third just does Washington and Baltimore, several do New York (by far the hardest case), and one (or more) assists me in stapling everything together. The result might be better than what we’re doing now, thanks to the greater detail; or it might be worse, due to slight inconsistencies between different people’s workflows, in which case a dedicated office manager would be needed to sort this out, at additional expense. But at least I’d study everything.

Because I’m doing this project for Marron and not for an American public-sector client, I can prune the search tree, and do it at relatively reasonable expense. That’s partly because I’m the lead, but also partly because I know what I’m doing, to an extent, and am not going to tell anyone “study everything” and then dismiss most scenarios after three months of no contact.

The behavior I’m contrasting myself with is, unfortunately, rife in the American public sector. And it’s the most common among exactly the set of very senior bureaucrats, often (not always) ones who are there by virtue of political appointment rather than the civil service process, who swear that consultants do things better than the public sector. There’s no real supervision, and no real narrowing of the process. This looks like an alternative to micromanagement, but is not, because the client at the end does say “no, not like this”; there’s a reason the consultants always feel the need to study everything rather than picking just a few alternatives and hoping the client trusts them to do it right.

It’s telling that the consultants and contractors we speak to don’t really seem happy with how they’re treated by the public-sector client in those situations. They’re happy when interfacing with other private actors, usually. I imagine that if I hired a larger team (which we don’t have the budget for) and gave each person a separate task, they’d be really happy to have come up with all those different scenarios for how to run trains in the Baltimore-Washington area, interfacing with other equally dedicated people doing other tasks of this size. When consultants are supervised by other consultants, only the top-level consultant interfaces with the remains of the civil service, hollowed out by hiring freezes, uncompetitive salaries, and political scourging; the others don’t and think things work really smoothly. This, I think, is why opaque design-build setups are so popular with the private consultants who are involved in them: by the time a country or region fully privatizes its supervision to a design-build consultant, its public sector has been hollowed so much that the consultants prefer to be supervised privately, even if the results are worse.

In contrast, the only way forward is a bigger civil service. This means hiring more people, in-house, and paying them on a par with what they would be earning in the private sector given their experience. As I said at the bar a few hours ago, I’m imagining someone whose CV is four years at the MTA, then five at a consultant, then four at the MBTA, and then six at a consultant; with these 19 years of experience, they could get hired at a senior engineer or project manager position, for which the market rate in Boston as I understand is in the high $100,000s. For some things, like commuter rail electrification, there are unlikely to be any suitable candidates from within the US, and so agencies would have to hire a European or Asian engineer.

With competitive salaries, people would move between different employers in the same industry, as is normal in American and European industries. They could move between public and private employers, because the wages and benefits should be similar. They’d pick up experience. An agency like the MBTA, with its five to six in-house design review engineers, could staff up appropriately to be able to supervise not just small projects like infill commuter rail station, which it built at reasonable cost on the Fairmount Line, but also large ones like the Green Line Extension and South Coast Rail, which it builds at outrageously high costs.

I’m Giving a Talk About Regional Rail in Boston

I haven’t been as active here lately; I think people know why and ask that you find other things to comment on.

I’m in Boston this week (and in New York next week), meeting with friends and TransitMatters people; in particular, I’m giving a talk at the Elephant and Castle on Wednesday at 6 pm to discuss regional rail and related reforms for Boston:

What I keep finding on these trips is that public transportation in the US is always worse than I remember. In Boston, I had a short wait on the Red Line from South Station to where I’m staying in Cambridge, but the next train was 13 minutes afterward, midday on a weekday. The trip from South Station to Porter Square took 24 minutes over a distance of 7.7 km covering seven stops; TransitMatters has a slow zone dashboard, there are so many. A line segment with an interstation a little longer than a kilometer has a lower average speed than any Paris Métro line, even those with 400 meter interstations; in Berlin, which averages 780 meters, the average speed is 30 km/h.

In New York, the frequency is okay, but there’s a new distraction: subway announcements now say “we have over 100 accessible stations,” giving no information except advertising that the MTA hates disabled people and thinks that only 30% of the system should be accessible to wheelchair users. There are still billboards on the subway advertising OMNY, a strictly inferior way of paying for the system than the older prepaid cards – it’s a weekly cap at the same rate as the unlimited weekly, but it’s only available Monday to Sunday rather than in any seven-day period (update 10-24: I’m told it’s fixed and now it’s exactly the same product as prepaying if you know you’ll hit the cap), and the monthly fare is still just a bit cheaper than getting weeklies or weekly caps.

The MTA 20 Year Needs Assessment Reminds Us They Can’t Build

The much-anticipated 20 Year Needs Assessment was released 2.5 days ago. It’s embarrassingly bad, and the reason is that the MTA can’t build, and is run by people who even by Northeastern US standards – not just other metro areas but also New Jersey – can’t build and propose reforms that make it even harder to build.

I see people discuss the slate of expansion projects in the assessment – things like Empire Connection restoration, a subway under Utica, extensions of Second Avenue Subway, and various infill stations. On the surface of it, the list of expansion projects is decent; there are quibbles, but in theory it’s not a bad list. But in practice, it’s not intended seriously. The best way to describe this list is if the average poster on a crayon forum got free reins to design something on the fly and then an NGO with a large budget made it into a glossy presentation. The costs are insane, for example $2.5 billion for an elevated extension of the 3 to Spring Creek of about 1.5 km (good idea, horrific cost), and $15.9 billion for a 6.8 km Utica subway (see maps here); this is in 2027 dollars, but the inescapable conclusion here is that the MTA thinks that to build an elevated extension in East New York should cost almost as much as it did to build a subway in Manhattan, where it used the density and complexity of the terrain as an argument for why things cost as much as they did.

To make sure people don’t say “well, $16 billion is a lot but Utica is worth it,” the report also lowballs the benefits in some places. Utica is presented as having three alternatives: subway, BRT, and subway part of the way and BRT the rest of the way; the subway alternative has the lowest projected ridership of the three, estimated at 55,600 riders/weekday, not many more than ride the bus today, and fewer than ride the combination of all three buses in the area today (B46 on Utica, B44 on Nostrand, B41 on Flatbush). For comparison, where the M15 on First and Second Avenues had about 50,000 weekday trips before Second Avenue Subway opened, the two-way ridership at the three new stations plus the increase in ridership at 63rd Street was 160,000 on the eve of corona, and that’s over just a quarter of the route; the projection for the phase that opened is 200,000 (and is likely to be achieved if the system gets back to pre-corona ridership), and that for the entire route from Harlem to Lower Manhattan is 560,000. On a more reasonable estimate, based on bus ridership and gains from faster speeds and saving the subway transfer, Utica should get around twice the ridership of the buses and so should Nostrand (not included in the plan), on the order of 150,000 and 100,000 respectively.

Nothing there is truly designed to optimize how to improve in a place that can’t build. London can’t build either, even if its costs are a fraction of New York’s (which fraction seems to be falling since New York’s costs seem to be rising faster); to compensate, TfL has run some very good operations, squeezing 36 trains per hour out of some of its lines, and making plans to invest in signaling and route design to allow similar throughput levels on other lines. The 20 Year Needs Assessment mentions signaling, but doesn’t at all promise any higher throughput, and instead talks about state of good repair: if it fails to improve throughput much, there’s no paper trail that they ever promised more than mid-20s trains per hour; the L’s Communications-Based Train Control (CBTC) signals permit 26 tph in theory but electrical capacity limits the line to 20, and the 7 still runs about 24 peak tph. London reacted to its inability to build by, in effect, operating so well that each line can do the work of 1.5 lines in New York; New York has little interest.

The things in there that the MTA does intend to build are slow in ways that cross the line from an embarrassment to an atrocity. There’s an ongoing investment plan in elevator accessibility on the subway. The assessment trumpets that “90% of ridership” will be at accessible stations in 2045, and 95% of stations (not weighted by ridership) will be accessible by 2055. Berlin has a two years older subway network than New York; here, 146 out of 175 stations have an elevator or ramp, for which the media has attacked the agency for its slow rollout of systemwide accessibility, after promises to retrofit the entire system by about this year were dashed.

The sheer hate of disabled people that drips from every MTA document about its accessibility installation is, frankly, disgusting, and makes a mockery of accessibility laws. Berlin has made stations accessible for about 2 million € apiece, and in Madrid the cost is about 10 million € (Madrid usually builds much more cheaply than Berlin, but first of all its side platforms and fare barriers mean it needs more elevators than Berlin, and second it builds more elevators than the minimum because at its low costs it can afford to do so). In New York, the costs nowadays start at $50 million and go up from there; the average for the current slate of stations is around $70 million.

And the reason for this inability to build is decisions made by current MTA leadership, on an ongoing basis. The norm in low- and medium-cost countries is that designs are made in-house, or by consultants who are directly supervised by in-house civil service engineers who have sufficient team size to make decisions. In New York, as in the rest of the US, the norms is that not only is design done with consultants, but also the consultants are supervised by another layer of consultants. The generalist leadership at the MTA doesn’t know enough to supervise them: the civil service is small and constantly bullied by the political appointees, and the political appointees have no background in planning or engineering and have little respect for experts who do. Thus, they tell the consultants “study everything” and give no guidance; the consultants dutifully study literally everything and can’t use their own expertise for how to prune the search tree, leading to very high design costs.

Procurement, likewise, is done on the principle that the MTA civil service can’t do anything. Thus, the political appointees build more and more layers of intermediaries. MTA head Janno Lieber takes credit for the design-build law from 2019, in which it’s legalized (and in some cases mandated) to have some merger of design and construction, but now there’s impetus to merge even more, in what is called progressive design-build (in short: New York’s definition of design-build is similar to what is used in Turkey and what we call des-bid-ign-build in our report – two contracts, but the design contract is incomplete and the build contract includes completing the design; progressive design-build means doing a single contract). Low- and medium-cost countries don’t do any of this, with the exception of Nordic examples, which have seen a sharp rise in costs from low to medium in conjunction with doing this.

And MTA leadership likes this. So do the contractors, since the civil service in New York is so enfeebled – scourged by the likes of Lieber, denied any flexibility to make decisions – that it can’t properly supervise design-bid-build projects (and still the transition to design-build is raising costs a bit). Layers of consultants, insulated from public scrutiny over why exactly the MTA can’t make its stations accessible or extend the subway, are exactly what incompetent political appointees (but I repeat myself) want.

Hence, the assessment. Other than the repulsively slow timeline on accessibility, this is not intended to be built. It’s not even intended as a “what if.” It’s barely even speculation. It’s kayfabe. It’s mimicry of a real plan. It’s a list of things that everyone agrees should be there plus a few things some planners wanted, mostly solidly, complete with numbers that say “oh well, we can’t, let’s move on.” And this will not end while current leadership stays there. They can’t build, and they don’t want to be able to build; this is the result.

High Speed 2 is (Partly) Canceled Due to High Costs

It’s not yet officially confirmed, but Prime Minister Rishi Sunak will formally announce that High Speed 2 will be paused north of Birmingham. All media reporting on this issue – BBC, Reuters, Sky, Telegraph – centers the issue of costs; the Telegraph just ran an op-ed supporting the curtailment on grounds of fiscal prudence.

I can’t tell you how the costs compare with the benefits, but the costs, as compared with other costs, really are extremely high. The Telegraph op-ed has a graph with how real costs have risen over time (other media reporting conflates cost overruns with inflation), which pegs current costs, with the leg to Manchester still in there, as ranging from about £85 billion to £112 billion in 2022 prices, for a full network of (I believe) 530 km. In PPP terms, this is $230-310 million/km, which is typical of subways in low-to-medium-cost countries (and somewhat less than half as much as a London Undeground extension). The total cost in 2022 terms of all high-speed lines opened to date in France and Germany combined is about the same as the low end of the range for High Speed 2.

I bring this up not to complain about high costs – I’ve done this in Britain many times – but to point out that costs matter. The ability of a country or city to build useful infrastructure really does depend on cost, and allowing costs to explode in order to buy off specific constituencies, out of poor engineering, or out of indifference to good project delivery practices means less stuff can be built.

Britain, unfortunately, has done all three. High Speed 2 is full of scope creep designed to buy off groups – namely, there is a lot of gratuious tunneling in the London-Birmingham first phase, the one that isn’t being scrapped. The terrain is flat by French or German standards, but the people living in the rural areas northwest of London are wealthy and NIMBY and complained and so they got their tunnels, which at this point are so advanced in construction that it’s not possible to descope them.

Then there are questionable engineering decisions, like the truly massive urban stations. The line was planned with a massive addition to Euston Station, which has since been descoped (I blogged it when it was still uncertain, but it was later confirmed); the current plan seems to be to dump passengers at Old Oak Common, at an Elizabeth line station somewhat outside Central London. It’s possible to connect to Euston with some very good operational discipline, but that requires imitating some specific Shinkansen operations that aren’t used anywhere in Europe, because the surplus of tracks at the Parisian terminals is so great it’s not needed there, and nowhere else in Europe is there such high single-city ridership.

And then there is poor project delivery, and here, the Tories themselves are partly to blame. They love the privatization of the state to massive consultancies. As I keep saying about the history of London Underground construction costs, the history doesn’t prove in any way that it’s Margaret Thatcher’s fault, but it sure is consistent with that hypothesis – costs were rising even before she came to power, but the real explosion happened between the 1970s (with the opening of the Jubilee line at 2022 PPP $172 million/km) and the 1990s (with the opening of the Jubilee line extension at $570 million/km).

On Hierarchy and the Civil Service

Some argument on a military history blog with good posts and terrible comments led me down a rabbit hole of talking about different models of how large, hierarchical organizations function.

The origin of this is that I’ve been reading the same things about military organization and the chain of command, in different variations. Here, via Quora, is the best encapsulation of how a multi-layer hierarchy works, on a submarine:

The CO never gives orders directly to watch standers. He works through the OOD or EOOW (officer of the deck and engineering officer of the watch). 99.9% of the time through the OOD. – He/She must re enforce their authority and not hijack it. It will cause confusion. If a Captain gives an order to a watch stander he/she assumes the conn. but not the deck. (See my other postings about this.)

The contrast is with personalistic managerial styles, in which the top person in the hierarchy constantly subverts and undermines their direct subordinates, by changing plans at the last minute, giving orders to low-rank people directly, or openly scolding the direct subordinates as out of touch. Think, for example, about the common cultural stereotypes of middle management; upper management that invoke such stereotypes against middle management is unlikely to be able to implement any long-term corporate culture, because the people responsible for implementing it know that they can be sniped by the top at any moment.

I bring this up here because populist politicians, or politicians in populist systems, love so undermining their own instruments of governance. The civil service is one such instrument, and is what’s relevant for infrastructure – but this undermining also includes other instruments.

For example, in the 2018 Senate race, Beto O’Rourke visited college students to induce them to turn out to vote for him and against Ted Cruz – and let people know that political consultants recommended against this move. Needless to say, O’Rourke uses consultants just like everyone else (and in US politics, at least for the Democrats, all candidates use the same pool of consultants – there’s no separation into, say, ones advising more left-wing and ones advising more centrist candidates). But he still felt it was necessary to scold his own instruments of campaigning.

So this isn’t even about the trend in the US, and in some countries with democratic backsliding, of replacing the apolitical civil service with party institutions. Personalistic politicians (and the US encourages personalism even in non-populists) undermine the instruments of ideological party governance, too.

The upshot of all of this is that a city like New York can have large civil service departments, but it won’t really have a strong civil service if the mayor keeps publicly undermining it. If planners know that whatever they work on will get sniped on a whim, they will not give their best; they’ll adopt punch-clock behavior, doing just the minimum until they qualify for a pension. If planners know that whatever they do, they’ll face a glass ceiling and have to answer to an inexperienced aide to a political appointee, they’ll leave as soon as they can. The most talented workers will go to the private sector, and everyone who can leave the city will as services atrophy.

Our Webinar, and Penn Reconstruction

Our webinar about the train station 3D model went off successfully. I was on video for a little more than two hours, Michael a little less; the recording is on YouTube, and I can upload the auto-captioning if people are okay with some truly bad subtitles.

I might even do more webinars as a substitute for Twitch streams, just because Zoom samples video at similar quality to Twitch for my purposes but at far smaller file size; every time I upload a Zoom video I’m reminded that it takes half an hour to upload a two-hour video whereas on Twitch it is two hours when I’m in Germany. (Internet service in other countries I visit is much better.)

The questions, as expected, were mostly not about the 3D model, but about through-running and Penn Station in general. Joe Clift was asking a bunch of questions about the Hudson Tunnel Project (HTP) and its own issues, and he and others were asking about commuter rail frequency. A lot of what we talked about is a preview of a long proposal, currently 19,000, by the Effective Transit Alliance; the short version can be found here. For example, I briefly mentioned on video that Penn Expansion, the plan to demolish a Manhattan block south of Penn Station to add more tracks at a cost of $17 billion, provides no benefits whatsoever, even if it doesn’t incorporate through-running. The explanation is that the required capacity can be accommodated on four to five tracks with best American practices for train turnaround times and with average non-US practices, 10 minutes to turn; the LIRR and New Jersey Transit think they need 18-22 minutes.

There weren’t questions about Penn Reconstruction, the separate (and much better) $7 billion plan to rebuild the station in place. The plan is not bad – it includes extra staircases and escalators, extra space on the lower concourse, and extra exits. But Reinvent Albany just found an agreement between the various users of Penn Station for how to do Penn Reconstruction, and it enshrines some really bad practices: heavy use of consultants, and a choice of one of four project delivery methods all of which involve privatization of the state; state-built construction is not on the menu.

In light of that, it may make sense to delay Penn Reconstruction. The plan as it is locks in bad procurement practices, which mean the costs are necessarily going to be a multiple of what they could be. It’s better to expand in-house construction capacity for the HTP and then deploy it for other projects as the agency gains expertise; France is doing this with Grand Paris Express, using its delivery vehicle Société du Grand Paris as the agency for building RER systems in secondary French cities, rather than letting the accumulated state capacity dissipate when Grand Paris Express is done.

This is separate from the issue of what to even do about Penn Station – Reconstruction in effect snipes all the reimaginings, not just ours but also ones that got more established traction like Vishaan Chakrabarti’s. But even then it’s not necessarily a bad project; it just really isn’t worth $7 billion, and the agreement makes it clear that it is possible to do better if the agencies in question learn what good procurement practices are (which I doubt – the MTA is very bought in to design-build failure).

Different Models of Partial Through-Running

I gave a very well-attended webinar talk a few hours ago, in which a minority of the time was spent on the 3D model and a majority about through-running and related modernization elements for commuter rail. I will talk more about it when the video finishes uploading, which will take hours in the queue. But for now, I’d like to talk about different conceptions of how through-running should work. I was asked what the difference is between my vision (really our vision at ETA, including that of people who disagree with me on a lot of specifics) and the vision of Tri-State and ReThink.

One difference is that I think a Penn Station-Grand Central connection is prudent and they don’t, but it’s at the level of detail. The biggest difference is how to react to a situation where there isn’t enough core capacity to run every line through. Tri-State and ReThink prefer connecting as many lines as possible to the through-running trunk; I prefer only connecting lines insofar as they can run frequently and without interference with non-through-running lines.

Partial through-running

To run everything in New York through, it’s necessary to build about six different lines. My standard six-line map can be seen here, with Line 7 (colored turquoise) removed; note that Line 7’s New Jersey branches don’t currently run any passenger service, and its Long Island branches could just be connected to Line 5 (dark yellow). The question is what to do when there are no six through-lines but only two or three. Right now, there is only one plausible through-line; the Gateway tunnel/Hudson Tunnel Project would add a second, if it included some extra infrastructure (like the Grand Central connection); the realigned Empire Connection could be a third. Anything else is a from-scratch project; any plan has to assume no more than two or three lines.

The question is what to do afterward. I am inspired by the RER, which began with a handful of branches, on which it ran intense service. For example, here is Paris in 1985, at which point it had the RER A, B, and C, but no D yet: observe that there were still large terminating networks at the largest train stations, including some lines that weren’t even frequent enough to be depicted – the RER D system out of Gare de Lyon visible starting 1995 took over a preexisting line that until then missed the map’s 20-minute midday frequency cutoff.

The upshot that whenever I depict a three-line New York commuter rail system, it leaves out large portions of the system; those terminate at Grand Central (without running through to Penn Station), Brooklyn, or Hoboken. The point is to leverage existing lines and run service intensely, for example every 10 minutes per branch (or every 20 on outer tails, but the underlying branches should be every 10).

Tri-State uses a map of the RER in its above-linked writeup, but doesn’t work this way. Instead, it depicts a trunk line from Secaucus to Penn Station to Sunnyside with branches in a few directions. ReThink is clearer about what it’s doing and is depicting every possible branch connecting to the trunk, even the Hudson and Harlem Lines, via a rebuilt connection to the Hell Gate Bridge.

The issue of separation

The other issue for me – and this is a long-term disagreement I have with some other really sharp people at ETA – is the importance of separating through- from terminating lines. Paris has almost total segregation between RER and terminating Transilien trains; on the most important parts of the network, the RER A and B, there is only track sharing on one branch of the RER A (with Transilien L to Saint-Lazare), and only at rush hour. London likewise uses Crossrail/Elizabeth Line trains to connect to the slow lines of the Great Eastern and Great Western Main Lines, more or less leaving the fast lines for terminating trains. Berlin has practically no track sharing between the S-Bahn and anything else, just one short branch section.

With no contiguous four-track lines, New York can’t so segregate services while keeping to the Parisian norm that shorter-range lines run through and longer-range ones terminate. Any such scheme would necessarily involve extensive sharing of trunk tunnels between terminating and through trains, which would make Penn Station’s schedules even more fragile than they are today.

This means that New York is compelled to run through at fairly long range. For example, trains should be running through on the Northeast Corridor all the way to Trenton fairly early, and probably also all the way to New Haven. This makes a lot of otherwise-sympathetic agency planners nervous; they get the point about metro-like service at the range of Newark, Elizabeth, and New Rochelle, but assume that farther-out suburbs would only see demand to Manhattan and only at rush hour. I don’t think that this nervousness is justified – the outer anchors see traffic all day, every day (New Haven is, at least on numbers from the 2010s, the busiest station in the region on weekends, edging out Stamford and Ronkonkoma). But I get where it’s coming from. It’s just a necessary byproduct of running a system in which some entire lines run through and other entire lines do not.

On the New Jersey side, this compels a setup in which the Northeast Corridor and North Jersey Coast Line run through, even all the way to the end. The Morris and Essex Lines and the Montclair-Boonton Line would then be running to the Gateway tunnel, running through if the tunnel connected to Grand Central or anything else to the east. The Raritan Valley Line can terminate at Newark with a transfer, or be shoehorned into either the Northeast Corridor (easier infrastructure) or Morris and Essex system (more spare capacity) if extensive infrastructure is built to accommodate this. The Erie lines, planned to have an awkward loop at Secaucus, should just keep terminating at Hoboken until there’s money for a dedicated tunnel for them – they’re already perfectly separated from the Northeast Corridor and tie-ins, and can stay separate.

On the LIRR side, this means designating different lines to run to Penn Station or Grand Central, and set up easy connections at Jamaica or a future Sunnyside Junction station. I like sending the LIRR Main Line to Grand Central, the Atlantic lines (Far Rockaway and Long Beach) to Brooklyn, the Port Washington Branch to the same trunk as the Northeast Corridor, and the remaining lines to the northern East River Tunnel pair (with Empire Connection through-running eventually), but there are other ways of setting it up. Note here that the line that through-runs to New Jersey, Port Washington, is the one that’s most separated from the rest of the system, which means there is no direct service from New Jersey to Jamaica, only to Flushing; this is a cost, but it balances against much more robust rail service, without programmed conflicts between trains.

And on the Metro-North side, it means that anything that isn’t already linked to a through-line goes to Grand Central and ends there. I presume the New Haven Line would be running through either via Grand Central or via the Hell Gate Line, the Harlem Line would terminate, and the Hudson Line depends on whether the Empire Connection is built or not; as usual, there are other ways to set this up, and the tradeoff is that the Harlem Line is the most local in the Bronx whereas the New Haven Line already has to interface with through-running so might as well shoehorn everything there into the system.

I’m Giving a Webinar Talk About Penn Station

The model that I’ve been blogging about is going to be the subject of a Zoom webinar, on Thursday 9-28, at 19:00 Berlin time or 13:00 New York time.

The talk will be in conversation with New York Daily News reporter and editor Michael Aronson, who has been very passionate in private conversations with us about improving rail service in the area and criticizing poor project management and high costs. In particular, he may yet save the Gateway Project three years, advancing capacity that much faster.

Specifically, the issue is that the existing tunnels between New Jersey and New York, the North River Tunnels, were heavily damaged in Hurricane Sandy, and require long-term repairs. The preferred alternative is long-term shutdowns of one track at a time, which is not possible until the Gateway tunnel (the Hudson Tunnel Project) is completed and would take a total of three years across both tracks then. The alternative is to do those repairs during weekend shutdowns. It is commonly believed that already there is repair work every weekend, and the timetables through the tunnel are written with the assumption that traffic can fit on a single track every weekend, giving a 55-hour shutdown period once a week. However, Michael found out that over a four-year period ending in 2020, the full shutdown for repairs was only done 13 times, or once every three months, and most of those shutdowns were not for repairing the tunnels themselves; in the following year, no shutdowns were done due to corona, and subsequently, the sluggish pre-corona rate has continued. If the repairs are done every weekend as the timetable permits, then it should be possible to wrap up simultaneously with the completion of the new tunnel, saving those three years of shutdown.

Penn Station Followup with Blueprints

People have been asking about the Penn Station 3D model I posted at the beginning of the week (for a direct link, go here again and use letsredothis as a password). This post should be viewed as a combination of some addenda, including a top-down 2D blueprint and some more comments on how this can be built, and also some graphics contributed by Tunnelbuilder in comments, who sent some Grand Central profiles to me for posting to argue that it’s difficult to impossible to punch through to the station’s stub-end tracks and build through-running infrastructure.

The rebuilt Penn Station blueprint

This version highlights the underlying map of columns (which I flagrantly disrespect in the main block of the station):

The platforms are in magenta. The ochre paths are tracks and areas immediately next to them, 3.4 m wide since the track center to platform edge distance is 1.7 m in the American loading gauge; this leaves an uncolored strip, 1.1 m wide, for generous 4.5 m track centers (German standards allow 4 m). The elevators are in green with black Xes; staircases and escalators are in different shades of red. Partly transparent gray denotes streets and East and West Walkways. Partly transparent dark green denotes West and East End Corridors, the former about two-thirds deep (same as the subway passageways) and the latter one-third deep (same as the subway platforms); the green connection between them is the existing Connecting Concourse, portrayed as changing grade, with potential changes if it’s decided to place East End Corridor on the same grade as the West End. Partly transparent light blue denotes the footprint of Moynihan Train Hall. The scale is 10 pixels = 1 meter, with the black cube helping show scale.

How to build this

The sequence for construction should be as follows:

  1. Madison Square Garden just got a five-year operating permit extension; previously it had always gotten 10-year permits. There is real impetus for change, at least at the level of City Council. This means that there are five years to work on the design and find MSG a new site in the city. In 2028, it should begin demolition, also including Two Penn Plaza.
  2. The superblock between 31st Street, 7th Avenue, 33rd Street, and 8th Avenue should be hollowed out with direct access to the existing concourses. At this stage, East and West Walkways should be built, by a method that is either independent of what is below them (such as a tied arch) or is supported by columns at the middle of the future platform locations. In the latter case, it is necessary to take out some tracks out of service early, as the columns would hit them: tracks 10, 13, and 20 are all aligned near the centers of future platforms.
  3. Temporary escalators and stairs should be dropped from the walkways to the existing platforms, as the concourses between the street and platform levels are removed and the tracks daylit.
  4. Tracks should be closed in stages to permit moving the platforms according to the blueprint. The first stage should be the southernmost tracks, 1-4 or 1-5, because they don’t run through to the east, and in this period (early 2030s), most to all New Jersey Transit trains should be running through to the New Haven Line or the LIRR. If tracks 10, 13, and 20 are closed, then construction of future platforms 4, 5, and 8 can be accelerated, since tracks 8A and 8B are aligned with 19 and 21, and tracks 5A and 5B are aligned with 12 and 14.
  5. After tracks 1-5 are replaced and platforms 1 and 2 are built, or potentially simultaneously, middle- and high-numbered platforms should be progressively replaced. With good operating practices, trains to and from New Jersey can be accommodated on six tracks (four New Jersey Transit tracks, two Amtrak), and LIRR trains using the tunnels under 33rd Street can be accommodated on about six or potentially four (current service fits on four, especially with the high capacity of tracks 18-21). This means that of tracks 6-21, 12 need to be operational at a given time, or maybe 10 in a crunch if there are compromises on LIRR capacity. Tracks 1, 3, 6, 9, 12, 14, 15, 17, 19, 21 are aligned with tracks 1A, 1B, 3A, 4A, 5A, 5B, 6A, 6B, 8A, 8B and may be able to stay in service for the duration of construction, in which case the process becomes much easier, requiring just two stages; in the worst case, four stages are required.

The deadline for this is that the Gateway tunnel (the Hudson Tunnel Project) is slated to open in 2035. The current plan is to then shut down the preexisting North River Tunnels for three years for repairs, but in fact, the repairs can be done on weekends; the New York Daily News found that only 13 times in four years did Amtrak in fact conduct any repairs in the tunnel, even though the weekend timetable is designed for one of the two tracks to be out for an entire weekend continuously. The new tunnel points toward the southern end of the Penn Station complex, and thus the new platforms 1 and 2 need to be in operation by 2035, giving seven years to build this part; the other tracks can potentially follow later, and tracks 18-21 in particular may be kept as they are longer, since the current platform 10 (tracks 18-19) is fairly wide and the current platform 11 (tracks 20-21) has many access points to the Connecting Concourse.

The Grand Central complication

The through-running plan implied in this design is that platforms 1 and 2 should connect to the Grand Central Lower Level, where Metro-North trains terminate (the Upper Level has additional Metro-North tracks, generally used by longer-distance trains). This requires the tunnel to thread between older tunnels, including subway tunnels. The following two diagrams are in profile, going south (left) to north (right); the second diagram continues north of the first one.

It’s possible to punch south (left) of the Lower Level while respecting every constraint, but not all of them at once. Two constraints are absolute:

  • No interference with the 7 train tunnels
  • No interference with the 6 train tunnel (labeled “SB local”)

These can be satisfied easily. However, all other constraints, which are serious, require some waivers, or picking and choosing:

  • Keeping absolute grades to 4%, forcing the tunnel to go above the 7 and not below it (which requires clearing around 15 m in around 150 m of distance)
  • Respecting the Lower Level loop track
  • Respecting the disused Steinway Loop tunnel

If the latter two constraints are waivable, then the tunnel needs to clear around 1.5 m, for 6 m of diameter minus 4.5 m between the roof of the 7 tunnel and the floor of the 6 tunnel, in what looks like 40 horizontal m; it’s doable but with centimeters of slack, and may require waiving the 4% grade (though over such a short length it doesn’t matter – what matters is vertical curve radius, and the vertical curves can be built north of the 7 and south of the 6).