Peak Factors and Intercity Trains

In contrast with Reason’s fraud, CARRD’s Elizabeth Alexis makes a more serious criticism of the XpressWest plan: there is a prominent peak in travel from Southern California to Las Vegas on Friday afternoon and Sunday afternoon, and this means that there will be a lot of ancillary costs associated with peaks, such as extra rolling stock with low utilization rates. More ambitiously, she compares it to commuter trains’ peaks, and uses this to argue that commuter rail-style subsidies may be required. The reality is quite different – intercity trains just cost less to run per seat than local trains, and although the Southern California-Las Vegas travel market may have a stronger peak than most, the difference with high-speed services around the world is (at most) one of degree and not kind.

First, let’s look at how much actual peaking there is between Southern California and Las Vegas. XpressWest’s Environmental Impact Statements include an analysis of current travel patterns (as of 2004) and a ridership projection. This is contained in the ridership forecast in appendix F-D. Table 16, on PDF-page 55, claims that present auto traffic on Friday is 2.03 times as high as on other weekdays and 1.48 times as high as on the average day, including both low-use days and the weekend peak. On Sunday, the numbers are 2.53 and 1.84 respectively. The ridership projections assume that the annual-to-Friday ridership ratio will be 236 (the annual-to-weekday ratio on urban transit systems in the US appears to be about 300). Of course, it is unlikely that traffic is evenly distributed on the peak days – most likely it clusters in the afternoon peak.

However, the same is true, if only slightly less prominently, on existing HSR. For some evidence of this, read SNCF’s proposals for HSR in the US, linked on The Transport Politic, which explain that by rotating trains for maintenance during weekdays SNCF can have near-100% availability for the weekend peak. On PDF-page 195 of the California proposal, it says,

To cater to weekend traffic peaks, train maintenance operations are scheduled to take place between midday on Mondays and Thursday evening and at night.

By timing maintenance in this way, approximately 80% of the fleet can be available in the week (between Monday noon and Friday noon) and as much as 98% at weekends.

This does not mean the peak-to-base traffic ratio on the TGV is 98:80. It is normal on local and regional trains to have both more capacity available for the peak and more crowding. On the TGV all passengers must reserve a seat, but SNCF can instead institute peak pricing. For a random example, I tested Paris-Lyon tickets on October 10th (a Wednesday) and the 12th (a Friday). In both cases, frequency is hourly in the morning and early afternoon and half-hourly in the afternoon peak – but the fare was €25-30 on Wednesday versus €60-89 on Friday beginning at 5 pm. And with only two intermediate stops, both quite far from Paris and in very small towns, the LGV Sud-Est is not a good commuter route. Routes with significant high-speed commuter traffic are different: in the off-peak most Paris-Tours trips require a transfer, and there are only two direct TGVs before the afternoon peak, at 7:34 and 1:40 again on 10/10, and two direct low-speed intercity trains; in the afternoon peak, this rises to half-hourly direct TGVs and additional low-speed trains, and the fare on the two most expensive peak TGVs is €59 versus €15-20 in the off-peak.

In contrast, let us now look at the subsidized local services, both in France (for comparability with the TGV) and in the US and Japan (where schedules are easy to obtain). In Japan, we can use Hyperdia to find the peak-to-base ratio; three heavily used lines in the Tokyo area that I specifically checked – Yamanote, Chuo Rapid (to Tachikawa), and Tokaido Main (to Odawara) – have about twice as much inbound frequency in the peak hour, 8-9 am, than in the afternoon and evening off-peaks. In the US, BART, which is similar in function to European commuter trains, runs 24 trains per hour through the Transbay Tube and the central San Francisco subway at the peak, 16 in the midday off-peak, and 6 in the evenings and on weekends. New York’s subway schedules show a peak-to-midday ratio of about 2, with slightly reduced traffic in the evenings and on weekends. Paris runs 30 tph in the peak on the RER A (in the peak direction) and 20 on the RER B, and 18 and 12 respectively in the midday off-peak; this makes for a lower peak-to-base ratio than on the TGV, but does not lead to profitability.

Elizabeth’s problem with running strongly peaked HSR is that it would have a lot of empty trains, and this by itself would require subsidies. This sounds reasonable, but the actual difference between the profitability of intercity and local trains is not seating utilization. Taiwan HSR had 46% seat occupancy in 2009; it made a profit before interest. The Sanyo Shinkansen averages about 35 actual riders per car (compare car- and passenger-km on PDF-page 19); the 16-car sets that run through from the Tokaido Shinkansen average 83 seats per car, and the 8-car sets that run exclusively on Sanyo average 71. I do not know the seating occupancy on Japanese commuter trains, though it likely averages well over 100%, but in New York, subway cars average 28 passengers, a seat occupancy of about two-thirds. For an alternative measure, taking seating capacity into account, New York subway cars average about 1.5 seats per linear meter, versus 1.4 on the Sanyo Shinkansen.

Nor is the issue a difference of fare – PDF-page 18 of the Sanyo factsheet establishes an average fare of about $0.20 per passenger-km – and unlike on the TGV, fares do not vary based on time of day. Just the operating expenses of the New York City Subway are $0.21 per passenger-km. Those on Sanyo are far lower, judging by JR West’s profitability after depreciation and interest. Something else here is going on: intercity trains can control costs better, perhaps because they have less legacy infrastructure and labor to deal with, or perhaps because faster trips mean that the trains and their operators are more productive.

Of course any operator should strive to reduce the peak-to-base ratio, and doing so can result in meaningful gains in productivity. Vancouver’s busiest bus, the 99-B, benefits strongly from a bidirectional peak; it has not eliminated the peak, but by avoiding unidirectionality, at least the reverse-peak buses don’t run empty.

For XpressWest, it means it is strongly favorable to go after the Las Vegas-to-Los Angeles market, which the Victorville terminus ensures the trains will not serve at all due to passengers’ different responses to transfers at the origin and destination end. So far its plan is to just wait for California HSR to open a Palmdale-Los Angeles link; it has Victorville-Palmdale as a second phase, with plans to either run through-trains to Los Angeles and San Francisco or (worse, and unlikely) make people transfer at Palmdale. This is not enough, and although California is committed to building through Palmdale, it may not have enough money for it; the current budget is $15 billion to complete Bakersfield-Palmdale-Sylmar, which requires $9 billion in outside, presumably federal funding.

At the risk of heresy, let me propose that XpressWest build a medium-speed link, above ground, through Cajon Pass. High speeds are not possible anyway because of the grade, so they might as well compromise on other design standards, build curves of radius 1 km (146 km/h with the currently proposed cant and FRA waiver-free cant deficiency, 160 km/h maximum with unambitious European cant and cant deficiency, 200 km/h with tilting trains and high cant) and not 4 km, and keep everything above ground.

The risk of cost escalation is still higher than for building in the I-15 median north of Victorville, because environmental and geological work may sow that a tunnel is needed in any case. But given that XpressWest can make a profit on Victorville-Las Vegas alone, why not spend a few millions on studying Cajon Pass, and if it proves affordable then build to San Bernardino and if not then not? Independently of what California HSR does northwest of Los Angeles, a route to San Bernardino is already enough to make XpressWest independent of traffic congestion, reduce the need for a large parking lot in Victorville, and raise the number of Las Vegas-to-Los Angeles travelers from zero to small. And beyond that, electrifying and double-tracking Los Angeles-San Bernardino and running through-service cannot be done under present FRA regulations, but is feasible given enough waivers and then the project would provide bidirectional service.

Reason Releases Fraudulent Report Criticizing XpressWest

In response to the forthcoming FRA loan application by XpressWest (the rebranded Desert Xpress) for its high-speed rail line from the edge of the Los Angeles metro area to Las Vegas, Reason published a report claiming the project would fail. Coauthors Wendell Cox, who cowrote a fraudulent report about Florida HSR, and Adrian Moore, argue that costs will be higher and ridership lower than expected, leading to operating losses and bankruptcy. I still have some doubts about XpressWest’s business plan, but Cox and Moore skirt or ignore the real problems, and instead choose to attack it using numbers that are distorted and at times completely made up.

The smoking gun that something nefarious is going on is the attempt to remodel ridership in terms of competition with cars and planes. In table 2 on PDF-page 20, the report shows door-to-door travel times by the different modes to Las Vegas from various origins in Southern California, including Victorville itself, Riverside (80 km and a mountain pass away), and Los Angeles (130 km away). The assumption, which is for the most part correct, is that passengers drive to the airport or train station and need to factor in congestion, and the explicit assumptions on access time are spelled out in table A-1. The zinger is that while station and airport access times are computed by taking the free-flow Google Maps travel time and adding a congestion cushion, the assumed door-to-door travel times for people driving assume free-flow travel – and even this required me to pick a particular (albeit reasonable) location on the Strip that is closer in than the Google Maps point labeled Las Vegas.

For examples, the travel times by car given from Victorville, Riverside, and Los Angeles are 2:56, 3:47, and 4:20. Those are approximately equal to the free-flow travel times to the Palazzo on the Strip. Needless to say, traffic is not free-flow in Southern California. As of this writing, on Friday at 4:15 pm Pacific Time, Google Maps gives me a travel time of 4:23 from Los Angeles to the Palazzo free-flow but 5:13 in current traffic; figure the extra 50 minutes make it 5:10 over the 4:20 given in the study. The door-to-door travel time for a train from Los Angeles is given as 5:04 to Vegas and 4:04 from Vegas, the difference coming from not needing to budget as much time for the possibility of traffic and arrive extra-early. In other words, including realistic rush-hour conditions, driving is not 14 minutes faster than the train on average in each direction, but 36 minutes slower.

In addition, the report slightly overstates the train’s travel time, as 1:40. The environmental impact statement claims, on PDF-page 39 of FEIS chapter 2, that 150 mph electric trains (the alternative that has since been selected) will take 1:24. While this is an ambitious average speed for this top speed, it is achievable for a nonstop train. Subtract 16 minutes from train time and now driving all the way from Los Angeles is 52 minutes slower than the train. As an additional check on the model, Cox and Moore assume travelers must arrive at the train station 20 minutes before departure, in addition to the congestion cushion. This is not observed in HSR systems in such countries as France and Germany, where open station design means people can arrive a few minutes before departure. Figure 5 minutes and now driving is 1:07 slower than the train.

Let us now step back and examine the general argument of the report. Cox and Moore argue the following: there is a tendency for costs to escalate (as examined by Bent Flyvbjerg) and for ridership to fall short of predictions (they call it the International Average Error Forecast but supply no reference and give no indication of the computation involved, and given the above zinger regarding travel time nobody should trust this). The ridership model has flaws, and a series of sanity checks argue that ridership will fall far short while costs will escalate. It is therefore better, they claim, to expand I-15 instead to deal with rush hour capacity.

At every step of the way, the report makes substantial errors. Cox seems aggressively uninterested in the actual causes of cost escalation and ridership shortfalls, following Flyvbjerg’s note in his original paper that cost escalation can come from many sources but it is fairly certain that there will be some cost escalation in a megaproject.

We can do better, and examine recent HSR projects. In Spain, some meet projections and some do not. For example, the Madrid-Barcelona corridor was 25% below projections in 2010, and appears to have fallen farther behind in 2011 – but in 2008 the line was only 4% behind projections, and with a deep recession and 20% unemployment, Spain can be excused for having less economic activity than projected at the height of its bubble. Likewise, in Taiwan and South Korea the HSR lines have fallen far below projections made in the 1990s, when their economic growth was extremely fast – but even those projections failed a sanity check: Korea thought it would get more HSR riders than the Sanyo Shinkansen, which looks reasonable based on city sizes until one remembers that the Sanyo Shinkansen also connects to Tokyo at one end and the KTX does not; Taiwan had estimated similar ridership, even though its largest city, Taipei, had not many more people than the Sanyo Shinkansen’s distant-second largest city and only one third as many as Sanyo’s largest, Osaka. In contrast, French lines tend to overshoot projections, as can be seen in the above link for Taiwan.

In all cases it can take a few years for ridership to build up: Taiwan took 2 years to achieve profitability after depreciation but before interest (and is now profitable even after interest after a refinancing at a lower interest rate), which Cox and Moore spin as “The project suffered an accumulated loss of two-thirds of its private investment in the first 2.5 years of operation.”

Las Vegas did have a bubble, and is slowing down now, although it is nowhere near the level of depression Spain is in. The report in fact mentions that growth in hotel rooms and travel to Las Vegas has stalled (although part of it is due to the national recession, rather than a Nevada-specific crash). It comes close to mark, but even here it fails to note possible similarities and differences with case studies of shortfalls. However, since the report attacks not just projected 2035 growth but also base-case ridership for 2012, it does not deserve this charity, even as here it skirted a real problem rather than completely missing it.

To criticize the actual model, on PDF-page 34 Cox and Moore attack it for surveying a sample of 400 people and asking them if they would ride the train. They attack the general approach of stated-preference, without giving any reference for why it is bad (they include one sentence of criticism), and then offer the following platitude: “It would seem that a prediction of ridership using a ‘less than trainload’ sample would be insufficient on which to make multibillion dollar decisions.” This is not serious analysis; this is the same criticism that led people to disbelieve that George Gallup could forecast elections by polling just a few thousand voters. The relevant paragraph from the ridership model that they could does mention that 400 riders means they results are “less precise than the reported analysis indicates,” but the same passage says later, which they do not quote, that the problem comes from having polled only 51 air travelers, where they would like 150-200 people per mode. Fortunately they polled 300 drivers, and it is auto/rail mode split forecast that is hard, while air/rail is a fairly straightforward function of travel time – see figure 1 of an EU air/rail report.

Now, in lieu of the ridership model that the report criticizes, it offers sanity checks. These are normally a useful check on wildly inaccurate estimates, and if done in the 1990s would have made it clear Taiwan was not going to have 180,000 riders a day, and even its present-day traffic of 110,000 is a miracle. Cox and Moore offer two sanity checks. First is the aforementioned comparison to car and airplane travel time; that one can be disposed of due to fraudulent numbers. Another is a comparison to the Acela between New York and Washington. If the Acela only gets 2 million riders per year, they argue on page 35, how can Victorville-Las Vegas get 9 million?

Of course, people who have taken Amtrak know that the Acela is only about one-third of the ridership on the Northeast Corridor, and the time travel difference between Acela and Regional trains is small enough that the distinction is one of branding and service class. Amtrak claims on PDF-page 41 of its Northeast Corridor Master Plan that 70% of the corridor’s riders (of whom there are 11 million) are on the New York-Washington segment, so that’s already nearly 8 million, not 2 million. Further, the Acela is a slow train – its average speed, 130 km/h south of New York, is not much better than that of the legacy express trains that the TGV replaced; the average speed of the Regional is worse. To argue that XpressWest is just like Acela, Cox and Moore do not offer a serious model of the effect of access and egress times on ridership, but instead issue platitudes about a train that stops 40 miles outside the city.

To see how professionals model ridership, see for example Reinhard Clever’s thesis (the relevant pages are 26-33) as well as a short note of his regarding last-mile connectivity. Transfers, he argues, are less onerous at the origin end of the trip than at the destination end: if they must transfer, 55% of riders prefer to do so at the origin end, 22% in the middle, and 22% at the end. Likewise, commuters in auto-oriented suburbs of transit cities (the example given is Toronto) drive long distances to park-and-rides, but balk at transferring from the city-center station to the subway. Normally the origin end is likely to be the smaller city, but in the case of XpressWest, Las Vegas is the destination rather than the origin. As a result, it is unrealistic to expect significant ridership from Las Vegas residents traveling to Los Angeles (and XpressWest is not assuming any), but quite realistic to expect riders to go in the opposite direction.

Finally, the cost overrun projection is fraudulent. As Cox did in the report about Florida, on PDF-page 40 he is comparing a simple line in a freeway median to the Central Valley segment of California HSR, a line with substantial viaducts and grade separations. To his credit, he no longer includes the 11-point rubric of his Florida report, which overemphasized relatively small components of the cost like electrification and underemphasized civil infrastructure. Instead, the report just says it’s unrealistic to expect cost to be lower than in the Central Valley, without further explanation except that the Central Valley is flat; the need for plenty of grade separations and viaducts is not mentioned.

This could be attributed to a simple mistake, but in fact footnote 76 argues based on the simplicity of the terrain and the ample space in the median that widening I-15 will be cheap, only $1.6-2.5 million per lane-km ($2.6-3.9 million per lane-mile) in both directions. No connection is made with the fact that a grade-separated median is not available to California HSR. In fact California is planning to widen Route 99 from 4 lanes to 6 at $6 billion (PDF p. 22); it is unclear to me how long of a stretch of 99 is under consideration, but the full length including segments north of Sacramento is 640 km, of which about 240 appears to be already 6-lane, which would make the cost $15 $7.5 million (it would include freeway conversion, but the same issue with grade separations is true of California HSR and has been the primary driver of cost overruns in the Central Valley). The construction cost difference between the Central Valley and XpressWest is a factor of 2; perhaps it’s Cox and Moore who, in assuming one ninth to one sixth one fifth to one third the per-km cost of CA 99’s Interstate conversion, are lowballing costs for their own favored project, and not XpressWest. (Update: I misread the footnote, and the cost contained therein is $1.6-2.5 million per unidirectional lane-km.)

No other argument is presented that costs will run over, except that according to Flyvbjerg they might. Since the projected costs are well within California’s per-km cost if one omits the viaducts, tunnels, and grade separations, we can assume that costs are likely to stay under control. In fact the cost escalations on international HSR lines have typically come from heavy tunneling, which is less predictable than at-grade construction. The at-grade lines in France have stayed within budget. In Norway the 50% cost overrun of the airport train was centered on a difficult tunnel. German lines run over too, but have significant tunneling as well, and the recent overruns in Korea (subtracting the first phase, comparing cost projections from 2010 and 2000 shows a 40% overrun) were in the nearly-50%-in-tunnel second phase. But in Japan, as far as I can tell recent Shinkansen construction is on-budget despite heavy tunneling, and the same is true of AVE construction in Spain. Tunnels, we can conclude, are riskier than at-grade construction; in fact the biggest risk for at-grade construction, as seen in the California HSR project, is that viaducts or tunnels will be needed due to further engineering or environmental work, and running alongside a freeway minimizes the chance.

Because the study’s attempts to model cost and ridership are so weak, it should not be considered a serious challenge to XpressWest. Cox has had a troubled relationship with the truth in the past, and there is no argument he won’t make, no matter how ridiculous, to argue for the superiority of car travel over rail and mass transit. It’s actually the strong arguments that he fails to make – for example, regarding a possible comparison between Las Vegas and overheated East Asian Tiger economies. (For the record, I think Las Vegas is going to come out solid in such comparison.)

It is in reality quite easy for HSR to make enough money to cover above-the-rail expenses, and even track maintenance is quite cheap at about $125,000 per double track-km, but covering interest expenses is harder. Despite the canard that only the LGV Sud-Est and the Tokaido Shinkansen have paid back their interest, sourced to as far as I can tell just one person and reproduced by Cox and Moore on PDF-page 43, in reality multiple intercity railroads are profitable even including interests. This includes all three main island Shinkansen operators in Japan, SNCF, and DB. The belief that they are not comes from two sources: in Europe, conflation of subsidized commuter lines with profitable intercity lines, which are usually run by the same national railroads, and in Japan, the fact that the government wiped the accumulated operating deficit debt of Japan National Railways after splitting and privatizing it, but not Shinkansen construction debt (see references here).

So if Reason is so wrong, and XpressWest will likely meet both ridership and cost projections, what are my problems? In one word: uncertainty. Projected XpressWest revenue, on PDF-page 54 of the ridership model, is about $500 million per year in today’s money. Long-term inflation-protected federal debt is unusually cheap right now and this could make XpressWest a prudent investment – as of the time of this writing, the US can sell 30-year inflation-protected bonds at an interest rate of 0.5%, or $32 million on a $6.5 billion loan. HSR margins in Europe are low, but in Taiwan the margin in 2009, excluding interest, was 25%, which is enough (that said, despite falling far short of expectations, Taiwan HSR has very high ridership for what it is, and of course lower ridership means lower margins independently of interest rates).

But 0.5% interest is for safe investments, and infrastructure is not a safe investment. The claims that costs would run over and ridership would fall short are probably going to be proven wrong if construction goes through, making the project a success, though not a smashing success. But if the reduction in Las Vegas’s growth proves permanent and not just one recession, or if casino gambling declines, or if station access time proves more important than previously assumed in the model, or one of many other things that could go wrong, operating profits will decline.

This is what Cox fails to understand when quoting Flyvbjerg. Flyvbjerg talks about an average cost overrun – but more than that, he is concerned with risk. Many projects stay within budget or run over just a little, but a few cost several times as much as the original estimate. Telling the Big Digs and East Side Accesses apart from the Madrid Metro extensions is hard, and this is why it’s not appropriate to compute interest rates based on the borrowing costs available to the federal government.

At a riskier rate of return, things are troubling, as Paul Druce notes: he compared revenue estimates to the 30-year T-bill interest rates as of last year (3.75%), and found that operating margins would need to be above 25% until 2031 to maintain profitability. XpressWest is now looking for a larger loan than Paul assumed, but at a real rate of return of 2 or 3%, interest would indeed bite into the cost. If the project is that risky, it should therefore not be funded. That said, European transit projects tend to go ahead with a benefit-cost ratio higher than 1.2, which is certainly true of this project.

So the question is twofold. First, whether it’s sensible to lock in low interest rates and fund projects that would not be able to pay back their loans at the interest rates of a fast-growing economy. Second, how risky the project is. The first question is easier: on a pure cost-benefit analysis, the federal government can afford to lose a few billion dollars on a small number of bad investments, as long as it makes it up with enough successes, and this makes the net financial cost of the project to the government low (but positive, since it bears downside risk but does not benefit from the upside except indirectly through taxes); on top of this, precisely because the High Desert and Nevada are in deep recession, this project has additional economic benefit. The recession won’t last forever, but it exists now and will probably continue for the duration of construction.

I believe the answer to the second question is that it’s of moderate to high risk. The risk of cost escalations is low because the right-of-way is already secured and there is no difficult civil infrastructure. The risk of ridership shortfalls is more substantial – ridership estimates, especially of road/rail mode shares, have an inherent uncertainty, and on top of that the recession could cause permanent damage to Las Vegas. In addition, the strong Friday peak of travel to Las Vegas means that more rolling stock and station infrastructure will be needed relative to ridership than elsewhere, driving down operating margins.

The most troubling part of the project is that growing ridership will require a connection to Los Angeles, and because it requires a difficult mountain crossing, XpressWest is not interested in paying for it. Its current plan is to wait until California HSR opens to the LA Basin, and then link up with a line from Victorville to Palmdale. This is the real cost risk, and not the notion that at-grade rail construction is going to present the same difficulties as urban viaducts and mountain tunnels. In particular, California HSR will need to reconsider how to get from the Central Valley to Los Angeles, and the alternative that links with XpressWest goes through Palmdale, which appears to be more expensive by a few billion dollars than a straighter route through the Grapevine and Tejon Pass.

Since there is no hope for fast enough recovery that interest rates will rise, forcing early investment, it’s fine to wait. I would seriously suggest that the FRA delay decision until after the election, and if the Democrats win control of both the White House and Congress, wait a few more months until there is or is not a federal bill to fund HSR. The important thing to do is avoid biasing California toward an alternative that costs it several billion more dollars for the benefits of the XpressWest operation. Although California seems set on Palmdale, it is feasible that the amount of money Congress will make available for it in six months is enough for an initial operating segment if and only if it switches to the cheaper Grapevine alignment, and then the plan should be to try connecting XpressWest to the LA Basin much later, through tunnels through Cajon Pass. (In fact, if there is any way to get a cost estimate quickly, I would propose that, to see if it’s a reasonable alternative to Palmdale.)

If it’s a yes or no decision then I’m leaning toward yes, but not at any cost. If there is serious competition for other rail projects with higher or less risky benefits, then they should be funded ahead of XpressWest. If the decision biases California against the Grapevine, and the amount of funding available to it (from a separate pot of money, as it’s not asking for an FRA loan) is such that Palmdale would force unconscionable compromises elsewhere, then to protect the more important California HSR project XpressWest should be delayed even at the cost of potentially missing the window in which it can be funded.

But despite my doubts, it’s not a high-speed train to nowhere. It’s a high-speed train from the edge of a large metro area to a major leisure travel destination, and the cost of borrowing is so low that the federal government can expect to make its money back in ordinary circumstances. There is enough cushion against a ridership shortfall that the ordinary uncertainties expected are a small deal, and although a very large shortfall is likelier than for, say, the Northeast Corridor, it’s not probable enough to warrant denying a loan application. If Reason succeeds in canceling the line, it will join Florida HSR as a line that could have had great promise but succumbed to lobbying and fraud.

Are Larger Planes Feasible?

In my previous post, I showed how, in New York, high-speed rail can’t realistically be expected to reduce demand for travel much, and so to decongest its airspace something else is needed. The solutions are to reduce the number of slots, which means either moving them elsewhere (i.e. building relief airports) or increasing plane size. Although increasing plane size is desirable from an operational and environmental point of view, it has problems that make it harder than in Japan, where short-distance domestic flights use widebodies as large as the 747. In contrast, because short-distance air shuttles in the Northeast use very small planes, high-speed rail is a surprisingly promising way to reduce air congestion, despite my original implication.

The key to the plane size problem is this chart of the world’s top air city pairs, with Seoul-Jeju topping at nearly 10 million passengers per year. The chart mainly shows Asian city pairs; Europe and the US are not on the chart. The reason is that the chart considers individual airports, rather than city airspaces; data from within the US shows that there are city pairs that would make the list, all multi-airport. New York-South Florida is close to 20,000 passengers per day, or 7.1 million per year, but there are three airports at each end, and they are fairly evenly matched: the busiest of the nine airport pairs, LaGuardia-Fort Lauderdale, has just 3,500 passengers per day, too few to make the international list.

What this means is that if airlines offer any frequency, it’s harder to provide service with larger planes. Harder does not mean impossible, but this is nothing like the huge travel volumes between Haneda and Japan’s other major domestic airports. Larger planes soak up passengers very quickly: despite being the world’s busiest airport pair measured by seats flown, Tokyo-Sapporo has 23 flights per day, with 767s and 777s, compared with 60 for New York-Boston, mostly regional jets.

The other issue is competition between airlines. Tokyo-Sapporo is a duopoly between ANA and Japan Airlines. The busiest routes in the US have more companies, and if they don’t, then they’re dominated by a low-cost carrier, which will stick to narrowbodies to maintain fleet uniformity. The American competition, including the presence of low-cost carriers, lowers the fare: a random check of a roundtrip between Tokyo and Sapporo in early December gives me a fare of about $900 roundtrip, versus $80 one-way for New York-Chicago for the same check, or $171 on average.

However, the competition also means that if each airline wants to offer high frequency on its own, it must fly smaller planes. Even a plane every two hours works out to about 8 departures per day per direction; if the plane is a 787, it’s nearly 4,000 passengers per day in both directions. The busiest single-airline, single-airport pair in the US is American flying LaGuardia-O’Hare, at 2,400 passengers per day; this excludes connecting traffic, but connecting traffic will not by its own make the difference between LaGuardia-O’Hare and Tokyo-Sapporo.

To ordinary travelers the choice of airline doesn’t matter too much: there’s no difference between having two airlines each with flights that leave on the hour, and having each airline’s flights depart every other hours so that they overlie and create hourly frequency. At 6,300 passenger per day on all airlines, JFK-LAX has enough traffic as it is to run fifteen 787s per day in each direction. But other airport pairs not dominated by low-cost carriers, including those to South Florida, could only support three to five 787s.

More speculatively, good transit access to airports – including commuter rail through-running to allow easy travel from New Jersey and Westchester to JFK and from Long Island to Newark – could reduce the difference between Newark and JFK for the average traveler. This means that Newark and JFK could be lumped together. Business travelers may still want their hourly flights out of LaGuardia, but the rest could do with a flight out of each of JFK and Newark every two hours, alternating.

The problem is that it requires a massive rise in the transit mode share of airport access, because it is impossible to drive between JFK and New Jersey in a reasonable amount of time. That said, a political environment that taxed jet fuel to incentivize larger planes would also tax gas and induce a mode shift toward transit. In either case, LaGuardia would be outside this system, since connecting it to mass transit is expensive, and has little benefit other than airport travel; in contrast, commuter rail through-running is not only cheaper but also useful to people traveling to the Jamaica and Newark CBDs, who outnumber air travelers.

So on the busiest routes larger planes are feasible, but nontrivial. The final question should be how useful this exercise is. Each of New York’s three main airports has about a thousand aircraft movements per day – five hundred per direction. There are about 110 daily departures to Chicago, Miami, and Los Angeles, combined. Consolidation into larger planes can realistically cut about a third, or 3% of aircraft movements – a bit more at JFK, a bit less at the rest on account of low-cost flights. Fort Lauderdale and Palm Beach add another 50 between them, but they’re dominated by JetBlue. A few additional thick markets like San Francisco and Orlando add a bit more, but it can’t amount to more than 5% of the total.

In contrast, there are more than 40 daily flights to Washington, more than 60 to Boston (including Providence and Manchester), and nearly 30 to Philadelphia. Adding in the other cities within 3-hour HSR radius gives us about 300 departures per day, 19% of the aircraft movements. Most of those would see O&D air travel disappear, and even at the outer margin of the radius they’d see air travel greatly diminish. Connecting flights would also decrease, because of the relative ease of air/rail connections. Philadelphia would have no reason for an air connection to New York if people could take a train to the airport that were faster than flying; the same is true of Boston and Washington, though Boston is far enough and has no easy air/rail connection, so it might retain a handful of daily flights.

Although I could weasel and say that everything is needed – larger planes, relief airports, and substitution of short trips by HSR – the reality is that those are not equally significant. Not even close. My previous post’s analysis of New York’s air market papered over a large difference between the share of passenger traffic and the share of aircraft traffic that can be substituted by HSR, coming from the use of regional jets on short-range flights. (By the way, this is especial to New York; in California most short-range flights are run by Southwest and use 737s, and so at LAX, the share of short-range flights among both passengers and aircraft movements is the same, at 21%.)

So as it turns out, a significant portion of New York’s air traffic can be replaced, helping decongest the airspace. The total is close to a quarter, of which nearly 20% comes from HSR replacing the air shuttles, and an additional 3-5% could come from consolidation of domestic thick markets into less frequent flights on widebodies.

High-Speed Rail’s Role in Decongesting Airports

One common argument for building HSR is that it will help decongest airports, by displacing high-volume short-distance flights. This can result in a permanent reduction in air travel, reducing environmental impact, or a diversion of capacity to longer-distance flights, or perhaps a combination of both. The question is then how much air travel can be diverted.

The main source I’m using for this is the Office of Aviation Analysis’s master table of all lower-48 origin-and-destination city pairs with at least 10 passengers per day (table 6, 3rd quarter of ’11). The data is less than perfect, because passengers connecting from a domestic flight to an international flight count as O&D passengers, but for our purposes it is good enough.

As a first filter, we can see that out of a million passengers per day, 206,000 are flying distance of up to 500 miles, and 390,000 are flying up to 773, the New York-Chicago distance. Those 39% of travelers constitute a much smaller portion of emissions than 39% but a larger portion of planes. Furthermore, not all can be realistically moved to trains: at the upper end of this range, HSR can compete with air but not decimate service the way it can on shorter trips, and on top of that many city pairs are not located on any realistic HSR corridor.

So as a second filter, let us construct a table, by major city (i.e. the top 7 O&D cities minus Las Vegas), of what the total volume of travel is to HSR-viable markets:

City <2.5h <3h <3.5h <4h <4.5h <5h
New York (153386) 7.4% 10.7% 15.7% 17.6% 20.6% 32.2%
LA Area (132556) 11.6% 26.4% 26.4% 26.4% 26.4% 26.4%
Bay Area (103752) 0% 18.1% 18.4% 18.4% 30.5% 33.3%
Chicago (103540) 9.5% 16.5% 16.7% 19.9% 22.8% 34.1%
Was.-Bal. (97234) 5.4% 16.7% 22.5% 23.2% 29% 31.3%
Boston (75329) 8.7% 21.3% 23.3% 26.7% 28.6% 31.8%

Although HSR can get nontrivial mode share against air even if it takes 5 hours, it does not reduce air traffic at this range, but instead induces demand. So although HSR can produce competition for almost a third of the air traffic coming into the largest US cities, it cannot divert as much air traffic. Meaningful diversion occurs at much shorter range, perhaps 3 hours, and even that diversion is incomplete. When the 3-hour Eurostar opened, Paris-London air traffic was permanently halved, from 4.3 million per year before the Chunnel opened to about 2 million after; once the travel time was further reduced to 2:15 with the opening of High Speed 1, it further decreased, to about 1.3 million on the dominant Heathrow/CDG airport pair.

What this means is that for decongesting airports, the meaningful column is the second from the left, for trips up to 3 hours. We immediately see that HSR can only have a small effect on New York, but conversely can do a great deal in Los Angeles. New York is at a further penalty since the hub system ensures it will remain an international gateway, and so traffic between two different cities still needs to pass through.

For New York, the best things that can be done then are to use larger planes on domestic flights, and find relief airports. In Japan, the domestic flights use widebodies, sometimes even 747s, and this has enabled Tokyo-Sapporo to grow to become the world’s highest-capacity air city pair. In the US there are more airlines and the city pairs are less thick, but there is still room for larger planes than 737s and 757s. In the other direction, faster LIRR service could turn Islip into a better relief airport, but it would still have to overcome the stigma of being too far. HSR could also turn Philadelphia into a reasonable option: using the Airport Line and a freight corridor to the west to bypass some of the Wilmington Line’s curves and reduce travel time should be considered as a full build-out option, and would also put PHL about 45 minutes away from New York.

The New York versus Los Angeles difference is not too surprising once we consider where their respective second cities are located. San Francisco is 700 km from Los Angeles, Boston and Washington are 350 km from New York and Philadelphia 150. Elizabeth of CARRD tells me that on LA-SF the current mode split is 50% air, 50% car. The situation in the Northeast is different – making reasonable assumptions on seat occupancy, even on NY-DC and NY-Boston more people take a bus than fly.

Update: Anonymouse in comments brings a good point about the distribution of short-haul travel within airport systems: there is often proportionately more of it at the secondary airports. Providence actually has less short-distance traffic than Boston and Midway is about even with O’Hare, but in California, much more short-distance traffic is at the secondary airports.

The five LA-area airports between them have 27.5% of their domestic traffic within 3-hour radius, but this splits as 21% at LAX, 35% at Long Beach, 37% at Santa Ana, 40% at Ontario, and 63% at Burbank. The three Bay Area airports between them have 19% of their domestic traffic going to LA and a total of 35% within 5-hour train radius, but this splits as 14% and 29% at SFO, 27% and 48% at San Jose, and 35% and 57% at Oakland.

Notes about the table:

1. The transfer penalty is set at 20 minutes, for city pairs that have no reason to ever have a one-seat ride. Both low- and high-speed connecting services are included, including HSR trains running through to the legacy network; I am not proposing new HSR tracks to Green Bay.

2. Instead of making hard alignment decisions, I simply ignored everything that would be controversial. The change in numbers is trivial. For example, neither South Bend nor Fort Wayne is included; both combined have only 2,000 daily air travelers anywhere in the lower 48, and only a handful of dozens to each of the cities in the table.

3. The travel times are full-build, so, for example, the Northeast Corridor is 1:30 Boston-New York and 1:30 New York-Washington, rather than the slightly higher travel times that should be aimed at initially. Average speeds range from 240 to 300 km/h on high-speed lines (higher in the Midwest, South, and flat portions of the West, lower in the Northeast and the Californian mountain crossings), and 100-130 km/h on upgraded legacy lines.

4. For US-Canada travel, we use T-100 data for international flights (data from September 2011). The data quality is poor since small planes are excluded, causing an underestimate in traffic on such markets as New York-Toronto, but conversely many of those flights would be double-counted because international-domestic transfers count twice. We can assume that the two effects (ignoring international flights outside Canada, and counting domestic-international transfers) cancel out, which is equivalent to assuming that exactly half of international travelers connect domestically.

5. The full list of cities included in each entry in the table is:

New York:
-2:30: the Northeast Corridor, Hartford, the Empire Corridor up to Rochester, Pittsburgh, Richmond, Burlington, Montreal.
2:30-3:00: Buffalo, Raleigh, Portland.
3:00-3:30: Toronto, Ottawa, Cleveland, Norfolk, Greensboro.
3:30-4:00: Charlotte, Toledo, Fayetteville, Lynchburg.
4:00-4:30: Greenville (SC), Greenville (NC), Columbus, Detroit, Roanoke, Nantucket, Columbia (SC).
4:30-5:00: Atlanta, Chicago, Dayton, Cincinnati, Wilmington (NC), Savannah.

Los Angeles:
-2:30: Las Vegas, Phoenix, Sacramento.
2:30-3:00: San Francisco, Tucson.
(This is where my exclusion of unrealistic corridors has the most effect. HSR could connect Los Angeles with Portland and Denver in 5 hours, Salt Lake City in 3:30, and El Paso and Albuquerque in 4:30. But the population is too sparse for the overlapping short trips that make comparably long corridors in the eastern half of the US semi-reasonable.)

Bay Area:
-2:30: the entire Central Valley.
2:30-3:00: Los Angeles.
3:00-3:30: Palm Springs.
3:30-4:00: —
4:00-4:30: San Diego, Las Vegas (assuming a Grapevine and Cajon alignment, which is the worst assumption; if the connector is between Victorville and Palmdale, as officially planned, then it’s about 4:00, and if it’s between Mojave and Barstow, it’s 3:45).
4:30-5:00: Phoenix.

Chicago:
-2:30: the corridors to Minneapolis, Detroit/Cleveland, Cincinnati, and St. Louis; Grand Rapids, Louisville, Dayton, Green Bay, Columbus.
2:30-3:00: Nashville, Pittsburgh, Buffalo, Kansas City, Toronto.
3:00-3:30: Chattanooga, Rochester.
3:30-4:00: Atlanta, Harrisburg, Syracuse.
4:00-4:30: Ottawa, Philadelphia.
4:30-5:00: Montreal, Albany, New York.

Washington-Baltimore:
-2:30: the Northeast Corridor up to New York, the Southeast Corridor down to Charlotte, Fayetteville, Norfolk, Lynchburg.
2:30-3:00: Boston, Hartford, Albany, Pittsburgh, Greenville (SC), Greenville (NC), Roanoke, Columbia (SC).
3:00-3:30: Atlanta, Wilmington (NC), Burlington, Cleveland, Savannah.
3:30-4:00: Montreal, Syracuse, Toledo.
4:00-4:30: Charleston, Birmingham, Jacksonville, Detroit, Columbus, Rochester, Chattanooga, Asheville, Portland.
4:30-5:00: Dayton, Cincinnati, Buffalo, Daytona, Ottawa. (Orlando is very close and some alignments put it just under 5 hours, but not all do.)

Boston:
-2:30: the Northeast Corridor down to Philadelphia, the Empire Corridor up to Rochester, Burlington, Montreal, Hartford, Portland.
2:30-3:00: Washington, Buffalo, Harrisburg.
3:00-3:30: Toronto, Ottawa, Erie, Atlantic City.
3:30-4:00: Cleveland, Pittsburgh, Richmond.
4:00-4:30: Raleigh, Toledo.
4:30-5:00: Norfolk, Greensboro, Detroit, Columbus, Dayton.

Vancouver’s Busiest Buses

Translink has a list of performance metrics per bus route here. Those include ridership, boardings per revenue-hour, crowding measured as a percentage of available seats, and operating cost per unlinked trip. Since the numbers are only given per route, without a single table or chart as one could find for Providence or New Haven, here are the busiest routes, per weekday:

1. 99 – 54,350
2. 20 – 27,900
3. 9 – 25,300
4. 41 – 24,800
5. 16 – 21,250
6. 8 – 20,150
7. 3 – 19,950
8. 49 – 19,700
9. 135 – 19,600
10. 25 – 19,300

The full sanitized data for daily and annual ridership, excluding minibuses and night buses, can be found here. I’ve verified that excluding minibuses and night buses doesn’t change the rankings in the top 50 routes.

Although Vancouver’s buses more or less run on a grid, the grid isn’t very clean. Some lines, like the 9 (Broadway), 99 (Broadway), 3 (Main), and 41 (41st), run more or less straight north-south or east-west, bending only at the ends, but many others do not. The 16 follows a broad U-shaped route, serving Arbutus on the West Side, feeding into downtown, and then going east on Hastings and then south on Renfrew. Multiple routes use Broadway for just a few blocks, to orient themselves to the correct north-south street. Others are L-shaped.

This makes it hard to figure out what the busiest corridors are (Vancouver has enough ridership that the 15-minute frequent network extends too far down to give us the busiest routes). Broadway is clearly the single busiest – if 99 and 9 are considered express and local versions of the same route, then Broadway has nearly 80,000 weekday bus riders, compared with 55,000 on 1st and 2nd Avenues in Manhattan, without counting buses that serve small segments of Broadway along their trip. Not counting buses that zigzag, the next busiest are 41st Avenue (41), Hastings (135, 160), Main, and 49th (49).

But this partial interlining does exist. So how busy is Hastings, anyway? If we add the buses that go on inner Hastings – 14, 16, 20, 135, and 160 – we get 90,000 weekday riders. But the 14 and 16 have half their route on the West Side, and the 20 turns south on Commercial; those are not just Hastings buses. The same problem happens on Main (the 8 partially runs on it), and 4th (west of Granville it interlines the 4, 7, and 84, and west of Macdonald also the 44, totaling 40,000 riders).

This doesn’t mean Hastings has more people riding the bus on it than there are taking the Millennium Line. I doubt it’s even close – the 16 and 20 have long north-south legs with connections to the Expo and Millennium Lines, so people from Fraserview and most of the Renfrew corridor are probably not traveling anywhere on Hastings. But most likely, whatever fraction of 90,000 Hastings has, it is probably the second busiest corridor, or maybe the third after 4th.

The obvious problem here is for SkyTrain development. Broadway is almost certainly getting rail, and judging by how far lesser-used corridors are getting SkyTrain extensions, Hastings should get one too. 4th is half a kilometer north of Broadway, but Hastings is 2 km north of the Millennium Line. Hastings’ distance to the West Coast Express is shorter, but it is an active freight line, with active port industry to its north, and often parks separating it from the street grid to the south. Frequent, frequent-stop commuter rail is still possible, but half the station radius is wasted on water, and the freight traffic is such that it might require too much multi-tracking to be cost-effective for the potential ridership.

HSR Routes: Triangles and Ys

This post partially responds to “The Altamont of X” comments made by Adirondacker, though it is far more general than that.

Whenever a route has to connect three non-collinear cities, compromises must be made between cost and directness. The two basic configurations are a triangle and a Y or T; a triangle is more direct but requires more infrastructure, whereas a Y is the opposite. The purest example of this issue is in Texas; the Interstates connecting Houston, Dallas, and San Antonio form a triangle, but with future high-speed rail, either configuration and many compromises in between are possible. Since not even in Texas is there a pure triangle with equal vertices and nothing in between, each site has its own questions regarding phasing, constructibility, intermediate cities, and relative importance of the triangle’s three sides.

In California, the Altamont vs. Pacheco debate is at least in part a Y vs. triangle debate. Here, the three nodes are Southern California, the Bay Area, and Sacramento. The LA-Sacramento leg is the simplest, because the line would just run straight up the Central Valley. The question is then what to do with the other two. The Pacheco alternative is essentially a triangle: San Francisco-Sacramento service gets an Altamont overlay, or maybe a heavily upgraded Capitol Corridor, and there is wide separation between the Central Valley-Bay Area connection used by trains heading to Los Angeles and ones heading to Sacramento. Altamont is a Y whose branch point is Manteca, with tracks going west to the Bay Area, north to Sacramento, or south to Los Angeles.

The particular case of California, however, favors the Y over the triangle. LA-SF and SF-Sacramento are both important corridors, so being able to serve both more easily is an advantage. Although Pacheco is shorter in distance than Altamont, it is not shorter in time to San Francisco, because more of Altamont is in the Central Valley and less is on the Caltrain corridor; for the same reason, the two options are about even on the cost of LA-SF alone. Altamont is actually a bit cheaper according to the original alternatives analysis, and the recent cost overrun is disproportionately in areas used only by Pacheco, such as the pass itself and the San Jose Diridon complex. Although Altamont has to cross water, a water tunnel parallel to the potential crossing site is currently under construction and so the geology and environment are well-understood. Pacheco’s advantage is just about San Jose: it offers it a faster connection to Los Angeles, and also the prestige of being on the main line rather than on a spur that would have gotten canceled as soon as costs ran over.

The fact that Altamont is no worse than Pacheco at connecting Los Angeles to San Francisco, as opposed to San Jose, is the key here. Altamont has other advantages, but since the biggest advantage of triangles here is reduced to connecting a secondary city better, there’s every reason to prefer the Y.

The same is not true elsewhere. Let us consider three cases: New York and New England, Texas, and the eastern part of the Midwest.

In the Midwest, this is the easiest. The question is how to connect Chicago to Detroit, the options being the I-94 corridor through Michigan, and the I-90 corridor through Indiana and Toledo, which would be shared with a connection to Cleveland. In this case the savings due to picking a Y rather than a triangle are much greater, while, again, the Y does not compromise Chicago-Detroit, but only reduces Chicago’s connectivity to small cities on I-94 in Michigan. Unsurprisingly, there is no longer a debate I am aware of; the SNCF proposal and the Siemens proposal both connect Detroit to Chicago via Toledo.

In the other regions, it is harder. When one leg of the triangle is obviously more important than the other two, it can be useful to have a T, which is like a Y except that one leg is straight and the other two are lengthened slightly more. If Houston and San Antonio swapped locations, it would be obvious that it should be a T. But given that they are where they are, the strongest leg, Dallas-Houston, has nothing significant in between, while Dallas-San Antonio has two intermediate cities in addition to Austin, complicating that kind of T. The Texas T-Bone alignment keeps straight Dallas-San Antonio, the second strongest leg; on this rudimentary list of possible alignments on Keep Houston Houston, a T with Dallas-Houston straight does not even appear. SNCF’s proposal starts with Dallas-San Antonio and is agnostic on whether to extend to Houston as a triangle or a T.

Practically any solution but a triangle would make the weakest leg, Houston-San Antonio, more circuitous, but various compromises that keep it at least competitive are incompatible with making both Dallas-San Antonio and Dallas-Houston straight. The presence of Austin also makes an exact triangle infeasible. Houston-San Antonio on I-10 is 321 km; via Austin, it is 389; via the T-Bone, it is 500; via Dallas, it is over 800, making it completely uncompetitive with driving. The Interstates had an easier time – cars can get from Houston to Austin, Temple-Killeen, and Waco on state roads, and because 1960s’ Texas was empty between the three Triangle cities, construction costs were low.

In the Northeast, there is also an opportunity for a triangle versus Y argument, in the New York-Boston-Albany triangle, but this time the Y is weaker. The problem is that New York-Boston is by far the strongest leg and the first that should be constructed. For that leg alone, the advantage of a shore route through Providence over an inland route through Hartford and I-84 is not overwhelming, but it requires less construction (New Haven-Kingston vs. New Haven-Boston). On top of that, the pure Y would not use I-84 but require New York-Boston trains to go through Springfield, lengthening the trip, and even that would only make the extra construction required even with the triangle. On a high-value, relatively short corridor where every minute matters, this is a problem. The only leg that works either way, Boston-Albany, is by far the weakest.

Meanwhile, the second leg, New York-Albany, would greatly suffer from any such detour. New York-Albany direct is about 230 km. Via New Haven and Springfield, it’s 330, and the average speed is also lower because of unfixable curves between New York and New Haven and several forced station stops. On top of that, although less overall construction would be required at the end, New York-Albany direct requires less tunneling than going through the Berkshires, even with the Hudson Highlands, and also less urban construction through Hartford and Springfield. (Without the Y, New Haven-Hartford-Springfield would be an upgraded legacy corridor, rather than a dedicated HSR line, which would provide similar local functionality but be insufficient for an intercity through-route to Boston or Upstate New York.)

What this means is that just because a Y is preferable to a triangle in one location does not mean Ys are always better. It depends on how it impacts the stronger legs, on phasing, and on very dry constructibility questions. “The Altamont of X” is incomplete; the Altamont Y is special in that the strongest leg is indifferent to Altamont vs. Pacheco, making the benefits (as opposed to costs) a matter of 10 or 20 extra minutes on secondary markets.

Connecting New Jersey to Manhattan, Redux

This post responds to arguments made by Brian in comments regarding how to connect New Jersey regional trains to Manhattan, in addition to the present tunnels to Penn Station; Brian argues for leveraging the Staten Island Railway, including the North Shore Branch, since a Staten Island-Manhattan tunnel should be built anyway.

In my post about the various options for connecting New Jersey to Lower Manhattan, all four alternatives I looked at featured a tunnel across the Hudson from the Hudson County waterfront to Manhattan, differing only in the location of the portals and the route used to get to the New Jersey portal. There are in principle other options, and I’d like to explain why they’re less feasible, and conversely why a connection along the lines I suggested should be one of the top two priority trans-Hudson projects, together with an additional tunnel pair to Penn Station.

First, because Lower Manhattan is the second most important business district in the region, as well as a subway hub, it deserves some connection. More than that, it deserves a connection from as many directions as possible, same as Midtown, and it deserves a connection earlier rather than later. The longer it takes to build a direct commuter rail line to it, the more it will decline in favor of other business districts, which with the exception of Midtown are much harder to serve with transit. It’s likely that if the LIRR, the Pennyslvania, the Lackawanna, the Erie, and the New York Central had all managed to build commuter lines to Lower Manhattan, instead of relying on the subway and the Hudson Tubes for the final connection, Lower Manhattan would not have lost out to Midtown so readily; Midtown would remain more convenient for commuters from Uptown Manhattan, the Bronx, and Queens, but not for commuters from Long Island or New Jersey.

Because of those principles, we get that a connection from the Erie lines to Lower Manhattan is critical. Once we accept that the major New Jersey lines, or groups of lines, need to be connected to both Manhattan job centers, it becomes best to gear the Lower Manhattan connection to the Erie lines, which are the northernmost in New Jersey and therefore wouldn’t intersect a Lower Manhattan connection to another line. The ARC solution of looping trains around Secaucus and connecting them to Penn Station is a fine first step but is inadequate afterward: a Lower Manhattan connection from the Erie lines would intersect the other lines at Secaucus, allowing a transfer, but a connection from any other direction would not allow a transfer from the Erie lines to Lower Manhattan.

On top of this, the cost involved in building such a connection, along any of the four alignments I proposed, is a tunnel across the Hudson, some extra tunneling on the Manhattan or Jersey City side (the farther south the alignment, the more Jersey City and the less Manhattan tunneling is needed), and of course a station in Lower Manhattan. This is quite bare-bones in the sense that any other connection to Lower Manhattan has to incur the same costs of a tunnel across water, and a Manhattan station. Concretely, this means it’s easier to tunnel from Jersey City or Hoboken to Manhattan than from Staten Island to Manhattan, and as such this would be built first, becoming the initial connection from New Jersey to Lower Manhattan.

I waver on whether this should be done before or after four-tracking the North River Tunnels. The tunnels are still extraordinarily busy at rush hour, and even state of the art signaling will only buy a few years before traffic matches the new capacity; moreover, Lower Manhattan-bound commuters can already transfer to PATH at Newark Penn cross-platform or at Hoboken, either of which is more convenient than transferring at Penn Station. On the other hand, people can also get to the southern edge of Midtown on PATH, and direct Lower Manhattan service can justify diverting some Morris and Essex trains from the mainline. It buys at most a few more years of breathing room, but it adds more destinations that can be reached by train, whereas a Midtown solution just adds capacity to an existing destination.

But, now, what of a future Staten Island connection? If a Staten Island-Manhattan tunnel is built, along the straightest alignment, bypassing Brooklyn, then it could provide a second connection from New Jersey to Lower Manhattan. This is the brunt of Brian’s comment: it would require using the bridge from Elizabeth to the North Shore Branch, which is active, and for another access point a new bridge from the mainline to Perth Amboy, but even building the latter bridge costs much less than new tunnels. Here is a map of the alignments.

The problem with using this for through-trains from the Jersey Shore and the Raritan Valley Line, the lines that connect best to Staten Island, is speed. The distance to Grand Central through either Staten Island and Lower Manhattan or the Northeast Corridor and Penn Station is about the same; the distance to Lower Manhattan is several kilometers shorter and one transfer fewer than via Secaucus, but once one connection to Lower Manhattan exists, a secondary connection would have to be justified based on demand to all job centers, of which Midtown is the biggest.

But now the Staten Island connection would have a much lower average speed. It is curvier, independently of all other considerations. The tunnel from Staten Island to Manhattan should also be lower-speed, to reduce the required bore diameter and save money. Since there is no good reason for intercity trains to use this connection – the Perth Amboy connection leads to no intercity line, and the North Shore Branch connection would require building a new junction to the Northeast Corridor, which would be both expensive and curvy – there is no reason to optimize for speed, unlike the case for the Northeast Corridor. So the choice is between one line where express commuter trains could do 160 km/h except maybe in the last few kilometers into Manhattan, and one where they’d do 100 or charitably 130.

On top of that, there are more stations in Staten Island, and also more local demand. Part of it is just bad operating practices in New Jersey – there should be more local stops in Elizabeth – but Staten Island has far more local demand, and so dropping local stops to make it easier to run express trains is less justified. As of 2000, the latest year for which the census data is readily available, Staten Island had 53,000 Manhattan-bound commuters. The relevant intermediate cities on the Northeast Corridor and North Jersey Coast Line – Newark, Elizabeth, Linden, Rahway, Carteret, and Woodbridge – had 10,500 between them. The corresponding numbers of Brooklyn-bound commuters are 29,000 and 1,500, respectively. It makes sense to keep the current stop spacing on the trunk line between Newark and Rahway, or add just one or two stops, but it makes none to not fit a North Shore Branch service with many local stops, which would then slow down longer-distance regional trains.

While the North Shore Branch can’t be widened except with many takings, the Staten Island Railway mainline could conceivably be four-tracked to allow overtakes, and this would make it a more competitive route. But if there is money for that, there is probably money to six-track the remaining four-track gap between Newark Airport and Linden, allowing full separation of local commuter trains, express commuter trains, and intercity trains on the Northeast Corridor except for segments on which the speeds are similar (Newark-New York) or ones where traffic is low enough to fit on existing tracks (south of Rahway).

The problem is really that the North Jersey Coast Line doesn’t have enough traffic to justify two highly separated branches, one through Staten Island and one through the Northeast Corridor. The split I proposed in my regional rail posts is much smaller – trains are only split east of Penn Station, after they begin overlapping with the Morris and Essex Lines, and so it’s possible to time transfers in such a way that people from Long Branch can board any train and be at their destination with just one additional easy transfer. At most this may justify a few peak hour runs; otherwise, even if the Tottenville-Perth Amboy bridge is built, timed transfers at Perth Amboy are almost as good and avoid reducing frequency on each branch too much.

Quick Note: Vancouver’s Transit Revival

I’ve been looking for Canadian mode share numbers that are more recent than 2006; although there was a census in 2011, it apparently did not include such numbers. However, a separate survey regarding commuting was published a year ago, using data from 2010. Mode shares are only included in Toronto, Vancouver, and Montreal, and those are listed separately for the city and the suburbs rather than for the whole metro area, but we can take a weighted average of population; it’s not perfect because the employment rate in the suburbs may be different from in the city, but it’s very close.

The result: Toronto’s transit mode share in 2010 was 22%, Montreal’s was 24%, and Vancouver’s was 21%. The Toronto number is the same as the numbers in 1996, 2001, and 2006. The Montreal number is a bit higher than past-decade numbers. And the Vancouver number compares with 14.3% in 1996 and 16.5% in 2006 (it was 11.5% in 2001, but there was a bus strike when the census was conducted).

Put another way, Vancouver gained 4.5 percentage points of transit mode share between 2006 and 2010. Judging by the opening of the Canada Line and its relatively high ridership, this is indeed plausible and doesn’t have to be a statistical artifact, though I’ll still want to see numbers a few years from now to confirm the new trend. If the trend holds, it’s over 11 percentage points per decade, enough to make Vancouver the metro area with the largest transit mode share by about 2019. It’s a similar rate of increase to what I included in my April Fool’s post for the US at large, intended to be at or beyond the outer limit of what is plausible if everything is done perfectly. Previously, I’d thought 3-5 points per decade were the best possible in Canada and Australia.

This means Translink has made major success with revival, as opposed to merely retaining old mode share by getting people who previously couldn’t afford a car to stick with transit even as they enter the middle class. If instead it is just an artifact of the Canada Line’s opening, then it suggests Vancouver will continue to do well in the next ten years, as the Evergreen Line and hopefully the UBC extension open. The Millennium Line opened in 2002 and so figures into the 1996-2006 increase, but its ridership is 80,000 a day, versus 110,000 on the Canada Line and an estimated 146,000 on the UBC extension and 70,000 on the Evergreen Line.

Low- and Medium-Hanging Fruit

The entire process I try to apply to cost-effective rail construction is to figure out the best places to spend money per unit of time saved. Obviously, this is mainly for intercity traffic – for local traffic it’s more interesting to look at cost per rider – but it’s intercity traffic that benefits most from this kind of optimization anyway.

With the Northeast Corridor, there are definitively low-hanging fruit, such as new (non-FRA-compliant) rolling stock, raising superelevation, improving platform access within present infrastructure, and adding constant tension catenary south of New York. Those are so useful, in terms of cost per benefit to travelers, that they should all be pursued immediately. The more interesting question is what to do afterward. I’ve proposed a few things before, in various posts, but it’s more useful to talk about the general process of determining where to build, i.e. which fruit are medium-hanging and which are high-hanging. I think traditionally this boils down to two parameters:

1. Cost per minute saved, including by improving reliability. This is of course adjusted for demand: New York-Philadelphia minutes are the most important, then Philadelphia-Washington, then New York-Boston, and finally other corridors.

2. Reduction in operating cost. If the rest of the network is based on hourly trains, and you need to squeeze five additional minutes to reduce your travel time including turnaround to an integer number of hours, it’s worth spending the money on it to avoid needing extra trains, or a schedule that doesn’t match up with the rest of the network. (And the same is true if the network repeats every 52 minutes – there’s nothing magical about 60 here.)

However, three additional, less obvious parameters are important:

3. Usefulness to local transit, in terms of speed, reliability, etc. This essentially reduces the cost imputed to intercity trains per minute saved.

4. How low-hanging the fruit becomes if combined with another. The issue is that eliminating two adjacent slow zones in an otherwise fast run saves more than double the time of eliminating just one of the two; another way to think about it is that eliminating the second slow zone saves more time than eliminating the first. This can result in counterintuitive phasing in a constrained funding environment.

5. How high-hanging the fruit becomes if it is delayed. If there is significant disruption to service coming from construction, then it’s better to do it earlier than would be warranted based on pure cost-per-minute-saved calculation.

#3 features prominently in Amtrak’s preexisting planning – in fact, too prominently, with its emphasis on Gateway. It’s a matter of agency imperialism more than anything, but it can lead to good results elsewhere. It’s really points #4-5 that aren’t optimized – either the costs are out of whack, or they are ignored. Washington Union Station‘s remodeling is an example of overemphasizing #5 without considering the cost or the ability to use existing infrastructure more cheaply; Transbay Terminal‘s poor column placement is an example of ignoring #5 entirely.

The reason I push concrete-heavy improvements between New Rochelle and Stamford, but not between Stamford and New Haven, comes essentially from those three points. The Cos Cob Bridge replacement is good because of points #1, #3, and #5; an I-95 bypass of Port Chester and Greenwich then interacts with it positively because of point #4, and also provides a suitable passing segment between high-speed and express commuter trains. In contrast, the projects east of Stamford don’t interact so positively: they involve constructing various bypasses, at high cost per minute saved, in separate locations so that the same increasing returns do not exist, and generally it’d not difficult to connect the bypasses to existing tracks so that the disruption effect of #5 is not in place.

Troll Rail Projects

In lieu of a real post, I want to discuss a few possible rail projects that are not completely thought-out. By this I mean rail projects that probably have critical constructibility and cost problems, but not obvious ones. They lie somewhere between true trolling – say, transcontinental HSR from New York to Los Angeles – and projects that are difficult and not yet proposed but need to be seriously considered, such as new train tunnels to Lower Manhattan or a Geary subway.

The projects are roughly ordered from most serious to most frivolous. The projects for the Northeast may well be feasible and should be at least considered, and the first was probably originally not done due to agency turf issues. The rolling stock projects are the most speculative – they suggest things to be done by competent rolling stock manufacturers that probably would’ve done them already if they could. The non-Northeastern infrastructure projects are somewhere in between. Make of this what you will. Just, please, do not use any of this as the basis for any alternative proposal, and do not link with a description like “Why have transit agencies not thought of this?” unless you know what you’re doing.

Northeast

ARC-North: the proposals for cross-Hudson tunnels that connect to Penn Station, including ARC Alt G and now Amtrak’s Gateway, would have the new tunnels connecting to the south of the main intercity through-tracks: ARC goes to the southern tracks, currently used by New Jersey, and Amtrak eventually wants to add tracks to the south. I propose that when they eventually build such a project, they build the new tunnels to the north, connecting to the existing northern pair of East River Tunnels; a connection to Grand Central could then be built from one of the two East River tunnel pairs, the one not used by intercity trains.

Right now, the northernmost tracks have the most access points and the southernmost tracks the fewest. The system would take advantage of the reduction in demand to Penn Station after East Side Access opens. In case the present-day North River Tunnel diameter is too narrow to allow for higher speeds, the new tunnel could then be used (also) by intercity trains at 200 km/h while letting commuter trains go to Grand Central without reducing capacity there.

Northeast Corridor to Market East, on the cheap: a short connection between North Philadelphia and North Broad, similar to that proposed for the Chestnut Hill West Line but used for the Northeast Corridor instead, would let intercity trains serve Market East or Suburban Station, in addition to 30th Street Station. Trains continuing down to Washington would probably not want to use such a connection, as it would slow them down because of the sharp turn in the SEPTA tunnel, but trains continuing on the Keystone Corridor would emerge from 30th Street oriented the right way. Right now trains to the Keystone Corridor have to either reverse direction (as they do today) or use a connection that skips 30th Street Station (as the fastest New York-Chicago trains did in the Broadway Limited era). It could be useful for local HSR trains if there ever were HSR from Philadelphia to Pittsburgh.

Philadelphia Bypass: also on the subject of HSR from New York to the Keystone Corridor, if express trains skip Philadelphia, it would be useful to build a bypass roughly along existing freight routes and I-276, starting at Trenton and ending somewhere between King of Prussia and Exton. The cost may not justify this in terms of cost per minute saved on New York-Pittsburgh (and New York-Cleveland, and New York-Chicago).

Providence Downcity Station: using the East Side Rail Tunnel, trains could continue west to Downcity, and then connect to the legacy tracks by hopping over I-95 in Federal Hill. For commuter trains, an underground station at Thayer Street is necessary. This is a pick-your-poison project in terms of takings: there are tradeoffs between curve radius, i.e. noise, and takings, and also between both and centrality. One option would be a curved station over City Hall Park, which would become the new Kennedy Plaza, and then what is now Kennedy Plaza would be landscaped and turned into the new City Hall Park. Another would go straight west, cutting through Citizens Plaza, and have a station elevated over Memorial Boulevard.

To troll even further, trains could use abandoned trackage starting from East Providence and then go to Fall River (reconstructing more abandoned trackage) and Newport (building new tracks through Bristol and over the Mount Hope Bridge).

Old Erie Line Revival: New Jersey Transit’s Main Line trains do not use the Erie Main Line south of Paterson, which is abandoned, but instead go along the Lackawanna’s old Boonton Branch. The right-of-way for the original Erie line is still intact, and serves the center of Passaic better. It might be useful to rebuild the tracks, which would require viaducts, and realign the Main Line. Service on all lines would probably require too many outlets – not even a dedicated tunnel to Lower Manhattan, combined, could be used for all lines serving that part of North Jersey, so some would have to be severed and turned over to light rail (maybe the Northern Branch) or the subway. The old Erie line is actually the best candidate for being part of a subway extension, since it serves dense communities and has a natural terminus at Paterson, where it would probably have to go underground.

Steinway Tunnel Widening: the Steinway Tunnel was widened from trolley loading gauge to IRT loading gauge when what is now the 7 was built. Since the rest of the 7 is built to the wider BMT/IND loading gauge, widening the tunnel is a useful capacity reliever to spend money on. It’s probably supremely expensive – I’m sure the MTA has studied it in the past; it’s also far from the most crowded Queens-Manhattan crossing point. But the cost may compare favorably with other means of providing extra capacity, and it may also be beneficial to let some Flushing Line trains serve Broadway and some Astoria Line trains serve 42nd Street.

West Coast

Subway to Burbank: Los Angeles’s Red Line does not go straight north along Vermont to Burbank, but swerves west to swerve more of Hollywood and serves Universal City and North Hollywood on the Valley side of the mountains. Since Downtown Burbank is a major secondary employment center, soon to be served by HSR, why not extend the city’s transit system in that direction? The Orange Line there should be a no-brainer, but more speculatively, the MTA could find money (another ballot measure, maybe?) and program another a subway branch off the Red Line that serves Burbank, with excessive splitting prevented by a new Vermont subway, or even (to troll further) an entirely new line that follows Western south of the mountains.

San Jose – Almaden Street Station: San Jose has a medium-sized CBD, roughly comparable to Providence or Burbank, but Diridon Station is separated from it by a freeway. Since there’s already a plan to spend large amounts of money of turning it into a multi-level train station, which the local technical activists have dubbed Diridon Intergalactic (or Pangalactic), why not also move the station? Trains could go on an alignment like this, elevated over Almaden, on a viaduct dedicated to Caltrain and HSR so that only four tracks would be needed. It would also bypass the current reverse curve between Tamien and Diridon, obviating the need for an iconic bridge. In a realistic, cost-conscious blended plan this is too expensive, but they should at least compare the cost with both a blended plan and the proposed full-fat business plan before rejecting it.

San Francisco – Embarcadero Station: with Transbay Terminal facing every planning and constructibility problem known to humanity, and the current terminal at 4th and King too far from the CBD, why not extend the trains under King Street and then the Embarcadero and build a station near the Ferry Building? Building this close to water is a nightmare, and the curve from King to the Embarcadero may be too sharp, but at least this connects to BART directly and has no station length constraints. On the third hand, the Embarcadero is wide but possibly not wide enough for three platforms and six tracks.

Rolling Stock

Tilting HSR: tilting HSR trains are either relatively low-speed (the Pendolino is limited to 250 km/h, with a few derivatives capable of a bit more) or relatively low-tilt (Talgos are capable of 180 mm of cant deficiency, and the latest Shinkansen trains have active suspension allowing up to about the same for the E5 Series. However, trains capable of 250 mm cant deficiency and 360 km/h are feasible; this is the main subject of Martin Lindahl’s thesis, which I (and others) have been quoting as a ready source of HSR track standards around the world. That said, probably the only place in the world that needs such trains is the Northeast Corridor, due to its unique combination of long straight stretches, on which very high speeds are possible or could be with minor infrastructure upgrades, and long curvy stretches, on which even major upgrades could not bring up to full HSR standards.

Catenary-free HSR: there’s new technology for catenary-free light rail, which is intended for use in historic city centers with aesthetic opposition to trolleywire. The contactless power supply is buried under the tracks, with each segment activated only when a train is completely above it. Although the technology is still low-speed, it could be useful for HSR. Pantographs generate disproportionate noise at high speeds, and Japan specifically has been squeezing every possible decibel out of low-noise pantographs. Being able to eliminate the pantograph would carry this to its logical conclusion. On the margins, it would also permit narrower rights-of-way, since no space for catenary poles would be needed.