Today, Caltrain runs 98 trains per weekday, with a peak rush hour capacity of 5 tph (trains per hour) per direction, or 6400 seats per peak hour, counting both directions. Off-peak capacity is 2 tph per direction, and weekend capacity is just 1 tph per direction.
Caltrain has about 37,000 boardings per weekday. At 98 trains/day x 640 seats/train, this figures to a 58% load factor (the average proportion of occupied seats) assuming all trips go the full length of the line. In practice, the average load factor is likely under 50%, although the most popular rush hour trains do approach 100% (i.e. standing room only).
Caltrain plans to grow its ridership, independently of high speed rail, by using electric rolling stock that can operate faster and more frequently. Their Caltrain 2025 plan calls for a peak capacity increase from 6400 seats/hour to 11000-16000 seats/hour by the year 2025. Assuming trains remain at the same seating capacity, this implies an increase in peak traffic from 5 tph to 9-12 tph per direction. Total capacity would roughly double, to about 200 trains per weekday.
Is this a reasonable estimate? It is useful to compare with BART's current traffic through the transbay tube. At peak times, BART runs about 20 tph in each direction, in 10-car trains that seat 700 and often carry crush loads of over 1000 people. Considering that BART serves a much greater geographical area than the peninsula, their 28000 seats/hour peak capacity in 2009 seems at least consistent with Caltrain's lower bound of 11000 seats/hour, achieved 15 years later. The verdict on Caltrain ridership projections: plausible (update: see comments section for a dissenting verdict: highly optimistic)
High Speed Rail Capacity
The California High Speed Rail Authority has published various ridership estimates over the years, the latest of which is contained in their 2008 business plan. The chapter on ridership and revenue estimates that by 2030, the peninsula corridor will see 8 - 10 tph per direction during six peak hours (6-9AM and 4-7PM) and about 6 tph per direction off-peak. That's a total of roughly 220 trains per weekday or 1300 trains per week.
Interestingly, there are no trains that skip San Jose or terminate at San Jose, which neatly sidesteps the rivalry between San Francisco and San Jose for the title of most important Bay Area destination.
If you assume the average HSR train seats about 400 people, that amounts to 27 million seats a year. That compares to 20 million seats a year currently offered by airlines between the Bay Area and the LA basin. (Southwest Airlines alone flies nearly 100 flights daily each way, although each one seats only about 130 people). Compared to airline traffic alone, the HSR figures sound optimistic, although most of the ridership is claimed to come from automobile traffic on the 5, 99 and 101 freeways, rather than airline traffic. Accounting for traffic growth between 2009 and 2030, which would double in that time span at an annual growth rate of just 3.5%, the verdict on HSR ridership projections is: plausible. (update: see comments section for a dissenting verdict: ludicrous)
Peninsula Traffic in 2030
Let's add it all up. Since peak hours for Caltrain and HSR will coincide with morning and evening rush hours, it is conceivable that peak traffic on the peninsula could reach 20 tph in each direction by the late 2020's (9-12 tph Caltrain and 8-10 tph HSR). Off-peak traffic would be closer to 10 tph in each direction (4 tph Caltrain and 6 tph HSR). Total weekday traffic would be about 420 trains (including both directions), an increase of four times over today's traffic levels over a span of two decades. To an observer at any given place along the peninsula corridor, there would be a train passing by on average every 1.5 minutes at rush hour, and every 3 minutes off-peak.
While the 20 tph figure is similar to that achieved by BART on two tracks, the peninsula corridor will be very different: it will support five different classes of service, each running at a different average speed.
- HSR express, SF-SJ in 30 minutes (non-stop)
- HSR limited, SF-SJ in 34 minutes (1 intermediate stop)
- Caltrain Baby Bullet, SF-SJ in ~45 minutes (4 intermediate stops)
- Caltrain limited, SF-SJ in ~55 minutes (10 intermediate stops)
- Caltrain local, SF-SJ in ~66 minutes (15 intermediate stops)
In San Francisco, there is a different implication to this 20 tph peak figure: since all trains must run at the same speed in the final approach to San Francisco, as dictated by the tight curve radii in this area, they would no longer need to overtake each other. Theoretically, running 20 tph through this speed-restricted area could be done on fewer than four tracks-- possibly just two! With homogeneous speeds, BART does it every day through the transbay tube, meshing together trains from four far-flung lines into a 20 tph trunk with just two tracks. Similarly, Amtrak and New Jersey Transit run up to 24 tph each way on the two tracks under the Hudson river. In both cases, there are efforts underway to increase the number of tracks, but it does give you a feel for the maximum throughput of two tracks when traffic is perfectly homogeneous.
Terminal Turnback Capacity
When a train reaches the end of the line in San Francisco, it needs to be cleaned, serviced and provisioned for its next trip. From arrival to departure, these operations typically take about an hour on other rail systems. If time is of the essence, which it will be in San Francisco, the turnback time might be reduced to about 30 minutes, very low by rail standards and commensurate with airline operations. The Transbay Transit Center is planned to have just six platform tracks shared by HSR and Caltrain, a very small number by international rail terminal standards. If a train occupies a platform track for just 30 minutes at a time, the maximum turnback capacity of a six-track terminal is 12 trains per hour.
This highlights another interesting feature of San Francisco rail operations: overall capacity into and out of the city will be constrained by terminal turnback capacity, and not by track capacity. You could build a six-track approach into San Francisco, but you still couldn't squeeze more trains in and out of the city.
Solutions to this Transbay turnback constraint include:
- Turning some trains, most likely Caltrain services, at another station. The existing Caltrain terminal at 4th & King streets might be retained for this purpose.
- Turning some HSR services at San Jose. This has the added benefit of reducing traffic on the peninsula.
* Caltrain 9-12tph: extremely implausibleReplyDelete
I'm aware of no comparable long-distance suburban corridor anywhere in the world offering this excessive level of service (except of course where multiple branches converge on city approaches), and especially none for minor CBDs like San Francisco's (SJ's doesn't even figure, nor do intermediate ones like PA's.)
Yes, there are numerous of S-Bahn/RER lines with sub-10-minute (> 6tph in regular intervals) and even sub-6-minute around the world, but these serve dense urbanized corridors and distances a fraction (a half to a third) of SJ-SF, which would count as an intercity run (and possibly an international one!) in most places.
SF-SJ simply isn't the Rhein-Ruhr nor metro Tokyo nor Seoul nor London nor New York. Get real!
Locally, one need only compare with BART's Richmond-Berkeley-Oakland-SF (8tph x 700 seats, 4 of which at peaks are transfers transfers to standing room only C line trains at MacArthur and thus are somewhat theoretical) and BART's Concord-Walnut Creek-Oakland-SF corridor (6tph x 700 seats peak, pretty much at capacity due to poor, over-seated train design) to see this, also bearing in mind the 8-10 lanes of parallel 101/280 capacity versus the 5 lanes of 80 over the Bay Bridge.
Another point is that it is always preferable (from a public investment and public utility point of view, as well as from a operating reliability perspective) to run higher capacity trains than it is to foolishly throw more trains at the "problem", once given a show-up-and-ride (3-4tph) base level of service. Caltrain can double deck (NOT required at present or for the next 20 years!!!) and double length (trains in the 300+m range, ie longer than any suburban trains anywhere in the world) to provide capacity far in excess of reasonable forseeable demand, and at an immensely lower cost (capital, maintenance and direct operating) than if wasteful extra track and wasteful extra trains and wasteful extra crews were used.
It is extraordinarily hard to imagine justifying more than 8tph at the peak of peaks on Caltrain any time in the next several decades.
It is readily possible to imagine consultants and agencies seeking to justify excessive capital spends, of course, and it is readily possible to imagine local consultants and staffs positing 19th century operating practices in order to "justify" massively inefficient capital boondoggles.
* HSR 8-10tph: fraudulent -- utterly implausible.
Compare Tokaido Shinkansen.
Compare total intra-California air traffic capacity.
Also bear in mind that the indicated San Francisco-San-Jose-Los Banos-Sacramento trains are not going to happen. Our highly ethical and highly professional and highly successful and utterly uncorrupt and overall just tippy-top world-class transportation authorities have ensured that this is a completely uncompetitive service: no train operator in the world would take it on, and any that did would go broke. It's dead. The indicated 17 trains over 6 hours (over 4tph!) shown here simply won't be operated.
SF-SJ-Los Banos can't possibly see or justify more than 4tph long distance any time in the next several decades. Just consider any other HSR corridor in the world!
Also what goes for Caltrain capacity goes doubly here: HS trains go to 400m (coupled trains) and double deck long before they go to insanely close service intervals.
3tph x 400m x HSR double deck is over 3500 seats per direction per hour: that's over twenty 737-800s worth.
10tph HSR along Caltrain? That's SEVENTY full 737s in an hour.
The CHSRA numbers are wildly inaccurate, or, to use the technical term, fraudulent.
(The rent-seeking, politically-motivated consultants who come up with these numbers are, of course, not the first to do this sort of thing, and it isn't the first time they've done so. This is just the way of the world.)
* If HSR doesn't terminate trains in SJ (unlikely, given the immense and successful political program to maximize service inefficiency and penalize any part of Northern California that is not controlled by the people who brought you VTA Light Rail), then it makes great sense to allocate empty peak hour seats SF-SJ (there must many, since everybody will be getting on and off at Diridon "not yet dead" Memorial Station) to local traffic, accepting Caltrain fares for local travel on though-running long-distance services in the same way as the most successful transportation operator does.
Building infrastructure in order to support carrying around empty seats at peak hours when a trivial administrative agreement can improve load factors would be insane. And we surely don't want to plan anything insane.
* The étant donné in the blog posting of "five different service classes" in the peninsula ("HSR express" (0 stops), "HSR limited" (1), "Caltrain Baby Bullet" (4), "Caltrain limited" (10) and "Caltrain local" (15) makes less than no sense.
At a realistic 2-4tph (nor well above that), there is no difference between 0 and 1 stops for HSR.
There is simply no need to run a class of "Caltrain limited" trains. The only service types Caltrain needs are expresses (6 intermediate stops: Mission Bay, Millbrae, Hillsdale, Redwood City, Palo Alto, Mountain View) and locals (20 intermediate stops, assuming fully justifiable permanent closures of Broadway, Hayward Park, Atherton) -- where do you get 15?!). By arranging a cross-platform transfer local<->express at Hillsdale for every express train, the need for "limited" service in eliminated. Timetabling cleverness always beats throwing public money down black holes of operating inefficiency and construction boondoggling.
So for the next 20 years or more, an embarrassingly rich mixture services on the Caltrain corridor, far better than anything else on the continent, would be at most 4tph HSR (3 more realistic), 4tph Caltrain local, 4tph (2 more realistic) Caltrain express.
Fewer service classes, long trains, higher seat occupancy = better capital utilization and lower operating costs while still providing a nearly optimal service plan. It is a win, win, win.
So where's the problem?
(Answer: San Francisco turnback, but even that could be managed if anybody but imbeciles were working on the Transbay Terminal fiasco. That's another and longer screed. But given the existing Caltrain/TJPA/PTG disaster in SF, terminating trains in SJ or operating even fewer than 8tph Caltrain is a given. Amazing, suicidally depressing, simply incredible, but true...)
I think we're still in the very early stages of Caltrain and HSR getting to grips with how each operator's plans to serve the SF peninsular corridor will affect the other.ReplyDelete
The SF peninsula is one of very few areas whose population is expected to grow very slowly in coming decades. SF county may even see a decline.
In that context, five separate service levels in the SF-SJ corridor strikes me as overkill. In particular, why not consider operating just a single Caltrain service level with e.g. 6 intermediate stops alongside the two for HSR. There's nothing in the rule book that says such semi-local trains must always stop at the same stations. Caltrain's existing baby bullets don't, though it does complicate setting up a timetable.
There's also no reason to maximize the number of trains per hour. As Richard points out, fewer, larger trains would ease traffic management and reduce the number of noise events for residents near the tracks. HSR stations are supposed to have 1320' platforms (or at least, land reserved for future platform extension).
Long trains are cheaper to operate and, they alleviate the problem of space constraints at SF Transbay Terminal: if you manage the pedestrian flows via mandatory seat reservations and markings on the platforms and, you hire more cleaning staff, then you can turn around a 16-car single-level train just as quickly as one half as long. Trains have many more access doors than aircraft do and, there's no cargo/baggage loading to consider.
It would certainly be cheaper than another alternative that has been suggested: extending the DTX tunnel to a second set of platforms just south of Clementina St. to form a one-way loop track. This second set of platforms would be connected to the the Transbay Terminal via an ~800ft underground pedestrian passage featuring moving walkways.
The highest capacity trainsets currently in service are the 8-car TGV Duplex (545 seats, top speed 300km/h) and the 8-car E4Max (817 seats, top speed 240km/h). Both are bi-level designs, and both are sometimes combined into 16-car mega-trains to maximize line throughput.
Another option is to combine an 8-car passenger trainset with an 8-set high-speed cargo trainset (for mail, packages, air cargo containers, perishable foods, fresh cut flowers etc). Some thought would need to be given to where those would be decoupled and driven off to a siding for cargo handling. In SF, that could happen at 4th & King if a platform is sacrificed for the purpose.
I like the idea of forcing all trains to travel at the same speed in the approach to SF. The DTX tunnel would be easier and cheaper to construct if the number of tracks could be reduced from three to two. Additional bores for the tunnels between 4th & King and Brisbane could be omitted. However, FRA would have to approve this arrangement since UPRR still runs a couple of trains a day in the corridor.
I also like the idea of terminating some late evening HSR trains in San Jose so they can continue to SF first thing in the morning to support business travelers heading north to catch an early flight out of SFO or a meeting in SF. There would be value in letting these earlybird HSRs stop in PA/RC as well if using the HSR equipment this way would let Caltrain get away with buying fewer EMUs. Early morning HSR trains out of SF should probably also stop at SFO, PA/RC and SJ before continuing as express trains to LA. There may be value in an analogous arrangement for early morning HSR and Metrolink service between LA and Anaheim.
Personally, I would have preferred an HSR station at Santa Clara instead of SJ Diridon and a route past Milpitas and through Altamont instead of Pacheco. Based on CHSRA's figures, that would have added only ~8 minutes to the SF-LA line haul time but made SF-Sac express trains time-competitive with driving on I-80. Improved connectivity for existing Amador and Central Valley towns plus Castle Airport would have been gravy on top.
However, now that the BART extension to Santa Clara has been approved and a new Dumbarton rail bridge is less likely than ever, the new challenge is to develop the Gilroy station into a transit hub that also serves northern San Beneto and Monterey plus southern Santa Cruz counties. That may mean limousine buses rather than commuter rail - it would take investment in track maintenance and bypasses to extend either Caltrain or Amtrak CC service to Hollister and Salinas, respectively.
Just to comment on the last part of your post, Rafael, TAMC tells me that they're still on track (heh) to start running the Caltrain extension to Salinas in 2011. No word yet on what impact the economic crisis is going to have on those plans.ReplyDelete
Thanks for your counter-analysis. As I stated at the outset, these are certainly upper bounds.
While it's true that SF-SJ is not New York City, there does exist 10 tph long distance suburban commute service today between Trenton, NJ and New York City. Not only is it 10 tph, but many trains are 10-car double deckers, with more than double the seating capacity of a typical Caltrain.
HSR at 10 tph already happens on a routine basis in France. LGV PSE operates at 12 tph peak. I don't know as much about Japan, but your chart does show that the Tokaido line operates at 12 tph peak. (Most lines have a theoretical max capacity of 15 tph, although it's impossible to use all of it due to intermediate stops)
So, I suppose the argument is more about actual passenger demand than technical feasibility (and yes, demand is where everything should start from)
Regarding train sizes... In France, they grew frequency first and then train size (double deck) when they ran into track capacity limits. Ditto in the New Jersey commute mentioned above. In the airline world, average airplane size has been decreasing, despite higher costs, in order to offer more frequency. It's not so clear-cut that you would necessarily maximize train capacity before maxing out track capacity, so assuming 1000+ seats per train is probably no more realistic than my assumption of 400 seats.
When all is said and done, your tally still amounts to 12 tph peak. So they're only off by 66 percent-- I would say par for the course, as far as public transportation estimates go.
I've been thinking about the Transbay terminal limitation, and I was wondering if they could use the planned tail tracks to help with the turnaround times. As a disclaimer I'm a computer engineer, so I think of things like using pipelining in order to maximize throughput of a system.ReplyDelete
The obvious turn around method is to have a train come and go from the same platform. A train arrives, unloads, gets cleaned / re-provisioned, then leaves all in the same location. In this case you have 6 platforms, so your throughput in trains/hour is 6 platforms * 60 / turnaround time in minutes. So if your turnaround time is 30 minutes (very low), you get a throughput is 12 trains per hour (meaning 12 trains can leave within an hour). A turnaround time of 45 minutes drops you to 9 tph.
Now what if instead of having each platform act individually, you had platforms act in pairs (probably the 2 sides of the island platform) in conjunction with a tail track. Note that currently there are only 2 tail tracks planned, but it shouldn't be a huge change to get 3. In this pipeline there would be 3 stages. Trains would come into one side of the platform and unload the passengers. As passengers are unloading cleaning / re-provisioning personnel and equipment would load on. The train would then move into the tail tracks where the personnel could work cleaning/provisioning while not taking any platform space. Once done the train would move to the other side of the platform to unload the personnel and load passengers.
This plan has two benefits. The obvious one is that you get 33% more throughput for the same turnaround time. Lets assume a 45 minute turnaround breaks down into 15 minutes of unloading passengers/loading personnel, 15 minutes of cleaning, and 15 minutes of unloading personnel/loading passengers. Since you can have a train arriving into a platform pair every 15 minutes, you get 3 pairs * 60 / 15 = 12 tph. If you can cut each step down to 10 minutes for a 30 minute turnaround, you get 18 tph.
The other advantage is that each island has one side that is always arrival and the other side is always departure. This means you can optimize things like escalator placement / direction and overall passenger flow through the station.
Again I have no idea if the logistics for this would work at all. I'm just proposing it as a fairly cheap way to increase the throughput of the station without having to expand to 8 platforms or cut turnaround times to be insanely quick.
* I don't think one ought to design SJ-SF infrastructure by taking the greatest number of regional and local trains per hour one can find on some line somewhere in the world and adding it to the greatest number of high speed trains per hour you one finds elsewhere.
So I don't think the CHSRA-"data"-derived schematic diagram showing more than 9tph of HS traffic into SanJosicso has even a passing relation to the real world.
This is on the order of the number of trains per hour per Shinkansen line operating into Tokyo at peak hours, for God's sake!
[BTW I do like the Netzgrafik convention of indicating 1tph by one graph link between nodes, eg this amazing one (CH). Line thickness suggesting trains per six hours doesn't work for me. But your site, your immense effort, your call!]
Note also that the Japanese 12tph shown in my (immensely time consuming to produce, I do hope you enjoy it) Tokaido graph include commuter trains (see in particular the cleverly-placed flights of Kodamas from Mishima, Shizuoka, Hamamatsu into Tokyo in the early AM) which the world transportation authorities at CHSRA, MTC and CCJPB have assured us cannot run and do not run and will not run on any CHSRA track. (That was their entire "technical" justification for killing Altamont, after all.
See the testimony of Gene "blatant perjury in the service of currying political favor and future employment is no vice" Skoropowski for just one example.)
[Another aside: that Shinkansen line is all double track except for very brief passing loops right at stations. The operating discipline is breathtaking. There are some links to Youtube cab rides here for your distraction. And I subsequently found a track map on page 7 of this that I ought to use to decorate my graphic timetable sometime. (I don't plan to update the timetable data -- far too much work, and October 2007 is good enough to illustrate any number of points.)]
So no "commuter" trains on CHSR into SJ or SF. No Sacramento-Altamont-Livermore-Fremont "commuter" traffic.
So where the hell does anybody find the bums to put on seats to fill 9, 10, 11 or 12 HS trains per hour from the Central Valley into SanJosisco? (To repeat myself, because it does put things in perspective: 10tph is over seventy full 737 aircraft full per direction per hour!)
4tph is super generous. Look at Madrid-Barcelona (not even 2tph base today, perhaps doubling over the next decade), and then compare Madrid Atocha to SF Transbay or Barcelona Sants to San Jose Saint Diridon. Does not compute!
* Re generosity, note I even gave up 4tph for HSR (which amounts to "show up and go") even despite my suggestion that double traction and double decking are the preferred operating solutions when traffic gets denser. (Note they run 400m double AVE trains in Spain even in the presence of 30 minute or longer headways.) Yes, LGV-Sud-Est has more HS traffic than I posit for SF-SJ, but, well, SF isn't Paris (shock!) and there are no avoiding lines (e.g. LGV Interconnexion Est) circling around SF and then joining up in a high speed junction a bit south of Belmont. We just don't have the local traffic generators, and we don't have the interlining from elsewhere. So I reject your LGV-Sud-Est analogy. (If you were talking Fresno-Palmdale with LA-Sacto traffic interlined I'd give you some credit.)
* Nor do I think that rail traffic levels on the NJT NEC line (incidentally, I get up to 12tph vs your 10, but like typical old US "commuter rail" is is very, very peaky and unsustained) has any relationship to demand into SF or SJ.
That would mean San Bruno suddenly becoming Elizabeth, New Jersey, just for instance. Or Palo Alto morphing in 15 years into New Brunswick.
(Of course our very special PBQD = BART-SJX + Pacheco-HSR friends would have you believe that BART in Milpitas (Milpitas!!!) will see as many passengers as BART at Montgomery in the SF CBD, just as they "promised" similar numbers for BART in Millbrae...)
I can understand this sort of thing coming out of organizations that exist almost exclusively to engage in public works fraud, but expected better here!
As I said, I know of any number of S-Bahn-y lines operating more than 8tph.
But all of them are in corridors far denser and far more developed (in terms of connecting services and land use) than the SF Peninsula is likely to ever be.
I think a perfect analogy for SF-SJ traffic is the local one of the BART R line SF-Oakland-Berkeley.
To think that the Peninsula -- once again, featuring 101+280+El Camino, with the significant population of Daly City safety removed from the rail corridor, and with nearly all of San Jose nowhere near it -- in a decade or two will justify 3 times (12tph vs 4tph) the amount of rail service that BART has built up over the last three decades is, well, unhelpful. (Plus, to repeat, Caltrain capacity per 300+m double-deck train can be higher than BART 215m singles, should it ever come to that.)
So Clem, we're not vaguely in the same ballpark.
10-12tph Caltrain + 10tph HSR (per your very nice blog entry, and perhaps already operating in some Parallel Universe Earth) in five (count 'em!) traffic classes is not remotely similar to what I believe is more realistic, i.e. 6-8tph Caltrain + 3-4tph HSR in 3 stopping patterns.
I suppose our difference in tph estimates can be mostly ascribed to our factor of three different train size assumptions. Your 10 tph assumption fills 70 737s, my 10 tph assumption fills only 24 737s (in 2030). For comparison, today (in 2009) at peak hours, airlines fly roughly 20 737s per hour each way.
As for SF isn't Paris... no, but LA is comparable in size to Paris. SF/SJ/Oakland is much larger than Lyon. Looking at metro area populations, if you squint really hard:
LA = Paris = Tokyo = Madrid
SF/SJ/Oakland = Lyon = Osaka = Barcelona
I wouldn't consider it so far-fetched that HSR traffic in California could some day (2030) rival the traffic on the LGV Sud-Est or Tokaido today (2009). Arguments to the contrary seem to mainly invoke the disparity in feeder networks.
And yes, perhaps five classes of traffic was a bit optimistic.
Your points well taken; I'm glad they're recorded here for all to read.
I'm glad we both agree by analogy that LA is the central and more vibrant conurbation and that SFSJ is a provincial appendage.
But do note, before going too far with the analogy, that population alone is little guide to ridership. Your LA=Paris equation breaks down in the same way and for the same reason that that VTA<<Muni despite the "tenth largest city in the country" jive from down south (and massive corrupt MTC pork allocations to VTA while shafting Muni.) Without the density, without the Metro, without the RER, without the other Grandes Lignes, without the extreme administrative and cultural and political centralisation of Paris (the only other place like it in the world is London), LA, immense population notwithstanding, is going to struggle to generate a tenth of the HS passenger traffic that Paris does (or London, or Berlin, or Seoul, or Shanghai, or Madrid, or ...)
After all one station in tiny Zürich (population 360k) sees more train riders every day than LA's (pop more than ten times as great) entire rail system.
My overall point about trains per hour is that there is always a balance between service levels, train technology and infrastructure, and that by operating one fashion (say 4tph with 400m trains) versus another (say 10tph with 200m single level trains) can result in immense infrastructure and operating savings while having near-zero effect upon service attractiveness.
There is nothing given about 9tph HSR (nor 12tph Caltrain), other than some combination of extreme technical ignorance and/or cooking the books on ridership models at CHSRA, so basing billions of dollars of construction upon and generating lots of legitimate political opposition in reaction to a service plan somebody pulled out of thin air (the charitable interpretation) isn't the way to go.
9tph HSR simply isn't an objective, engineering-derived, justifiable number, and shouldn't be treated as such.
Or: "garbage in, garbage out".
Now if SF were the terminal for 5 different HSR branches to different destinations it would be one thing, but with SF-Sacto and SJ-Sacto dead sans Altamont, we're left with SF-Fresno, SF-Fresno, SF-Fresno, SF-Fresno and ... SF-Fresno. It would matter a lot if SF-Sacto ran every hour rather than every thirty minutes, just as it does for BART SF-Walnut Creek; however it doesn't matter in the slightest whether SF-Fresno is every 30 min, every 20, or every 7 minutes (7 minute headways is what your 9tph HSR means!) in terms of attracting or dissuading long distance riders. There's an economic point of "good enough" frequency for attractiveness of different types of service, and it happens a long way before 10 minute headways for multi-hundred-km inter-city trains.
Fundamentally, running that many low-capacity trains to essentially one destination is scandalously wasteful when the same capacity can be provided at lower cost without compromising service attractiveness.
I make the same argument for hypothetical (ie alternate universe only) 12tph SF-SJ Caltrain; this isn't just about CHSRA insanity.
That's it from me on this.
I'm with you, and I appreciate your patience in laying it out.ReplyDelete
I guess we can take comfort in that operators (be it HSR or Caltrain) will not run more trains than they can sustainably fill based on the real, actual demand out there. Witness current BART service levels between Millbrae and SFO. (zero.)
That is the silver lining for people who live near the tracks... the tph figure will adjust itself.
Sadly, overestimates do lead to terrible sub-optimization of the infrastructure. (see Millbrae SFO above). For HSR, the poor choice of Pacheco is one such example. Expanding the four tunnels into SF, botching the Transbay project, and building a gigantic terminal in SJ may turn into another.
What to do about it?
@ Clem -ReplyDelete
sounds to me that Richard disagrees vehemently on ridership growth through 2030 and with you on the optimal train frequency vs. length.
You mentioned that SNCF started out increasing train frequency first and only started deploying longer and bi-level trains when its trunk line between Paris and Lyon began to saturate.
What you didn't mention is that the available trainsets early on were all single-level. It's possible that Alstom had not yet perfected control of joined trainsets, just as the AGV is its first EMU design (as opposed to dual tractor).
CHSRA has a lot of headaches, but lack of technology options isn't one of them. Therefore, it makes perfect sense to consider options that minimize up-front spending on infrastructure.
In particular, provided FRA is amenable, the network sections in which HSR will anyhow not run at more than 125mph should be thought of as "rapid rail", with HSR sharing a dual-track corridor with slower commuter trains and a small number of freight trains each day, according to a timetable optimized for passenger needs. See this graph for a sense of the maximum speed ranges CHSRA is considering.
Rapid rail corridors should feature a sufficient number of bypass track segments of sufficient length to ensure HSR trains are not forced to slow down on account of commuter trains.
The commuter trains used on rapid rail corridors should be UIC-compliant and leverage the overhead electric infrastructure to achieve high acceleration, so expensive bypass tracks can be kept to a minimum. Caltrain is already planning to do just that for SF-SJ.
Bottom line, CHSRA's strategy should comprise the following aspects in the SF-SJ corridor:
1) Work in concert with Caltrain, UPRR and FRA to formally define requirements for rapid rail as a sub-class of HSR, especially with regard to the signaling safety needed FRA-compliant and non-compliant trains sharing track.
The other sub-class would be what Europeans call very high speed rail (VHSR), i.e. true bullet trains running at well over 125mph on dedicated tracks. Bullet trains can slow down to leverage legacy track in rapid rail corridors.
2) secure enough ROW to upgrade to a full four-track build-out at some point in the future, should ridership projections actually pan out - much depends on how road congestion and the consumer cost of petrochemical fuels develop. Where berms are to be used, consider negotiating easements rather than exercising eminent domain. This allows homeowners to landscape the sloped section and integrate visually attractive sound walls that double as fences (subject to a permitting process with Caltrain in the loop to ensure safe operations).
3) straighten the alignment where needed to deliver the line haul time targets, preferably without modifications to freeways (e.g. use eminent domain to fix the curve radius south of I-380). Once the optimized sections goes into service, the land occupied by the old curve should be sold.
4) construct all grade separations such that they can accommodate a future build-out to four tracks without additional changes to the cross road.
Avoid constructing more tunnel tracks than absolutely necessary. The two-track tunnels Caltrain already has should do nicely for a very long time and, the DTX should not need more, either. It would be smarter to spend more on additional tail tracks, as Peter explained above.
Rapid rail eliminates any need for a tunnel between San Tomas and SJ Diridon. The chicane around the Caltrain maintenance facility can be eliminated by swapping the usage of two tracks east of the building with two west of it.
5) Caltrain just spent quite a bit on bringing its existing tracks up to a state of good repair. An upgrade to rapid rail classification (i.e. up to 125mph) will mean more intensive maintenance and may require changes to the superelevation of some curves, unless CHSRA chooses trainsets with tilt technology. JR East's Fastech 360 development platform did and it also met the 220mph objective.
Afaik, the exact rail/wheel interface geometry and seeking aren't critical issues for bullet trains moving at rapid rail speeds.
6) ensure that all stations where HSR trains may stop have enough land to accommodate full-length (1/4 mile) platforms suitable for level boarding, even if they are initially only built out to half length.
7) construct stations such that pedestrian flow capacity is ample or can easily be increased if the number of platforms is increased from two to four. Also draw up a plan for dealing with transit, bicycle and personal car traffic growth that would occur if the station were expanded, trains run more frequently or longer trains deployed in response to ridership demand.
8) anticipate that additional operators such as ACE or high speed cargo outfits may ask for limited trackage rights on the rapid rail corridor in the future
9) optimize service classes, fare structures, train frequency and train lengths for the corridor rather than independently for each operator. Set trackage fees accordingly.
For example, Caltrain may need to cancel local service in favor of running only semi-locals that stop at e.g. 5 of the 15 locations between SF and SJ, but not always the same ones.
HSR may have to limit itself to two service classes, express (SF-SJ-LA-AH only) and semi-locals (assuming a second peninsula station in RC, a possible one in HanFord and a future one in IRvine):
Passengers traveling between distant secondary destinations would have to transfer once, for everyone else there is direct service. Similar logic would apply to trains service the phase II spurs to Sac and SD once they become operational.
10) share standard gauge storage yards and maintenance facilities where appropriate
Resisting the urge to fully build out everything from the outset means spending less on track construction before the first HSR train goes into operation, which may happen earlier than it otherwise would have. Funding for the SF TT trainbox + DTX tunnel will be easier to deliver.
With rapid rail, Caltrain may save enough on the overhead catenaries (or else generate higher trackage fees) to afford the Dumbarton rail project with its BART intermodal in Union City.
And even though the old bridge is just single-track and contains two swing bridges, it will be a valuable first step toward a future in which suitable ACE trains will deliver infrequent high-capacity rapid rail service between SF and Sac (perhaps down to Merced as well).
Perhaps most importantly, the above would give FRA a regulatory precedent for rapid rail and its combination with VHSR. Rapid rail corridors, whether passenger-only or with a few freight trains, could become very useful in many parts of California and around the nation. That should make it easier to secure federal funds going forward. Sen. Kerry (D-MA) has already recognized the importance of reforming FRA in his HSR bill, which he is expected to re-introduce soon.
I see there's quite the debate going on here. My own opinion is that the estimates of Caltrain traffic are optimistic, and those of HSR traffic are wildly optimistic. I can see Caltrain getting maybe 8 tph at best, with service divided between express trains (with two different stopping patterns as now) and local trains, with 2 tph all stops service. The limited is probably going away in its current form. If anything, it will be replaced by Dumbarton Rail running on the local tracks, but skipping some local stops, which would account for another 1-2 tph. And if ridership keeps growing, they can just lengthen their trains. All the stations are already capable of platforming 6 and maybe 7 car MU trains, and can be probably be lengthened to 8 cars without too much trouble. Incidentally, the longest commuter trains operating in this country are on the Long Island Railroad, with 12 85-foot cars, for a total length of about 310 meters.ReplyDelete
As for HSR, I can't imagine 9-12 tph worth of demand. The single busiest long distance rail corridor in this country, NY-DC, has only 2 tph of through service, and at best another train between NY and Philadelphia. And NY Penn Station has half as many intercity passengers as HSRA is forecasting for SF Transbay, despite the obvious difference in population between NYC and SF, and the fact that NY Penn has three Amtrak corridors radiating from it, whereas SF Transbay only has the one line to LA. So 4 tph would be a more reasonable estimate for peak HSR traffic, and 12 tph shouldn't be too much of a problem on double track between Bayshore and SF. It might not even be necessary to quad-track the whole corridor, but there are definite reliability benefits to having four tracks for as much of the line's length as possible, and local service is very important to expand Caltrain's market beyond the traditional Peninsula/South Bay-to-SF and SF-to-South Bay commutes. When a potential customer wants to go from Belmont to Mountain View, and there's only one train per hour at Belmont, which doesn't stop at Mountain View, that customer is almost certainly not going to bother.
Thinking more about this, it seems to me like in the morning at least, the Caltrain and HSR peaks would not coincide in the SF-bound direction. For an HSR train to get to SF by 9 am, it has to leave LA before 6:30, and I just don't see that being such a popular option. The Fresno-SF traffic can't be that huge. In the AM reverse-peak direction it would be the Caltrain traffic that is reduced, because there's just less commuter traffic in that direction. So the peak useage of the line would be 10 tph.ReplyDelete
Oh and by the way, Richard, the New Haven Line runs 21 Grand Central-bound trains in the peak hour. Of those, only three are from branches, and five go all the way from New Haven (a distance of 72 miles).
Caltrain 9-12tph: extremely implausibleReplyDelete
I'm aware of no comparable long-distance suburban corridor anywhere in the world offering this excessive level of service (except of course where multiple branches converge on city approaches), and especially none for minor CBDs like San Francisco's
This just isn't true, Richard. It's fair to omit NYC since its CBD is so much larger than the CBD of any other US city. But consider Chicago's Metra.
The Metra Electric Line tops out at 20 tph during peak hour - far more than that 9-12 tph that you consider implausible. Metra Electric has two smaller sub-branches though. A fairer comparison is the Metra BNSF Line. This is a 38 mile triple-tracked diesel line (comparable to the 47 mile Caltrain). Metra BNSF maxes out around 13 tph, again above the 9-12 tph that you consider implausible.
Is the potential market for Metra BNSF substantially larger than the potential Caltrain market? Hard to say. In Metra's favor, the Chicago CBD has more jobs than the SF CBD (630,000 in 2000 vs 341,000 in 2000). In Caltrain's favor, the SF CBD is higher density than the Chicago CBD (136,000 jobs/sq mile vs 79,000 jobs/sq mile), the population in Caltrain's catchment area is substantially larger than the population in Metra BNSF's catchment area (Metra BNSF is flanked by the Metra UP West line ~6 miles to the north and the Metra Heritage Corridor line ~6 miles to the south), and electrified Caltrain will be a superior service compared to the diesel Metra BNSF.
Bottom line: given that Metra BNSF currently hits 13 tph, it's very plausible to think that Caltrain could eventually reach a max of 9 tph.
CAHSR is a different story. I could easily see them hitting 8 tph if they run 8 car, single level trains. But why do that vs running fewer 16 car, possibly double-level trains?
For those curious, here is a sample of Metra BNSF line departures from Chicago Union Station. There is only a single branch on the line, but a few of the trains terminate in Downers Grove, which would be equivalent to Caltrain terminating a few trains in Palo Alto. Also note that different LTD/EXP trains stop at different stations (not unlike Caltrain).ReplyDelete
Oh and also note that the Metra BNSF line is at least three tracks for most of the way. The SF-SJ corridor also has some fairly unique characteristics, with major employment and population centers at both ends of the line, and for most of its length, a very geographically constrained corridor, with the bay on one side and the mountains on the other. The Bay Area in general is unlike the typical urban area that sprawls in all directions. Instead, it sprawls out linearly along the few geographic corridors available, and stretches a relatively long distance along those. Oh and for another comparison, the LIRR Babylon Branch has 11 trains in the peak hour, not counting the two through trains from Montauk, with each train being probably 10 cars (ranging between 8 and 12).ReplyDelete
@ arcady -ReplyDelete
for reference, I just checked on SouthWest's web site to see what they offer for weekday daytrips from SFO to LAX (as an example):
Outbound flights at 6:15, 7:05, 8:15, 9:45, 11:45.
Inbound flights at 6:55, 7:55 and 9:55.
Considering that you need more time to get to the airport, check in, go through security and board than you would for a train, I'd argue that there will be a market for very early HSR trains. It might make sense to schedule stops at both Millbrae and Palo Alto/Redwood City plus of course San Jose before proceeding straight to LA and Anaheim. This will make it easier for Caltrain to run its own trains - preferably as a mix of skippers (stops at SFO plus every other intermediate station) and/or SNCF-style semi-locals (run as local in either the first or the second half of the corridor, then non-stop to the final destination. Passengers traveling between secondary destinations may need to transfer once.
Early trains out of LA/Anaheim can run as expresses, provided that Metrolink sticks with its current model of running on freight tracks, even in the Anheim-Lancaster segment of the HSR route. The only exception is Fullerton-Anaheim(-Irvine), where the ROW is too narrow to construct new dedicated HSR tracks.
Where HSR will be at a disadvantage is redeye trips. Because of the greater line haul time, leaving LA as late as 10pm would mean arriving in SF at around 12:45am. Passengers might be ok with that, but the towns the trains run through might not, due to train noise.
On the other hand, the last Caltrain on a weekday arrives back in SF at midnight, bells ringing and horns blaring at every grade crossing (FRA quiet zone, anyone, even before grade separation is completed?) Compared to those horns, an aerodynamic train zooming past at 125mph would be downright quiet.
I don't know how one at 220mph would compare to a mile-long freight train rumbling through the Central Valley at around 11pm. Red-eye HSR (and even later high speed cargo) trains might be acceptable to residents living next to the railroad in those communities, there's no way of knowing for sure until testing at maximum speed begins.
I'd like to point out the elephant in the room to those who haven't noticed.ReplyDelete
HSR SF-SJ=30 minutes
The trains can easily run a 5 hour round trip without being thoroughly cleaned and restocked in LA and SF. So it could come down to having those jobs at one location or the other. In fact there is likely to be provisions only in one location. Minor cleaning and upkeep can be done by onboard staff during the trip. Moat trains make much longer trips between cleanings.ReplyDelete
@Andy, 45 minutes is an estimate based on 100+ mph top speeds and full grade separation. It assumes that all the corridor improvements are made for HSR. Without those improvements, run times will not shrink much from today's 1 hour.ReplyDelete
Richard, the Bay Area is a lot bigger than you give it credit for. It's bigger than Nagoya once you use the same definition for both metro areas.ReplyDelete
It's also less important for HSR to connect to good transit than you think. Tokyo, Osaka, and Paris are all very transit-oriented, but Lyon has a tiny metro system. This hasn't stopped Paris-Lyon HSR traffic from capturing 90% of the air/rail market share. People who fly expect to take a taxi or rent a car anyway; HSR actually improves on that by being in a more central location.
Another solution to the Transbay Terminal constraint would be a second transbay tube.ReplyDelete
Peninsula Rail Traffic - Rerouted
In my example above, I've rerouted the 17 SF-Sac to the Capital Corridor; those 17 SF-Sac trains are now SD-LA-SF-Sac-LA-SD trains. Alternatively, the trains could be routed via Altamont instead of the Capital Corridor. Either way, the second transbay tube is instrumental to reducing the HSR throughput from 9-10 tph to only 6-7 tph.
The recent State Senate hearing provided some information to clear up the TBT picture ... along with much spin on the part of the TBT staff trying to use their present advantage in staffing to get the jump on the CAHSRA in trying to get an inadequate design pushed through as a fait accompli, leaving a massive upgrade bill to a future generation.ReplyDelete
One island Caltrain, 2 islands HSR, because of the different platform heights.
Now, if the HSR is on the "outer bend" side of the train box, that means that Platform Track 1 is not connected to the tail tracks, because it extends out to reach the 400m required for the HSR.
And that prevents HSR from adopting the debark platform / tail for restock and cleaning / embark platform procedure that gets extra station time by taking the train off one platform early.
With 3 minute headways, 12 minutes to debark, 15 minutes tail dwell, and 15 minutes to embark, 2 platform tracks and 1 tail track can cope with 4 tph while offering the 40 minutes station time requested by the CAHSRA. And since the same capacity is not required in the evening, one of the tail tracks could be freed for Caltrain to use for stabling between the evening commuter peak and the morning commuter peak.
So, what stands between the current design and Caltrain 8tph and HSR 8tph?
Well, with the central tunnel track used as access, the outer tunnel track used for HSR egress, and the inner tunnel track used for Caltrain egress, and the HSR going above 4tph by using the debark/tail/embark operation, as long as Caltrain can get by with 15 minute dwell (12 minutes turnover, 3 minutes headway), there's only one thing ...
... a tail track connection for the outermost platform track.
This is not, of course, free ... it may mean shifting the outermost island toward 2nd in order to give it room at the back for the outer tail track to switch to platform track 1 and the inner tail track to switch to tail track 2. Platform tracks 3 and 4 are, in effect, extended to form the tail tracks, with the Caltrain Platform track 5 switching onto the inner tail track. In evening/night mode, two trains can be stabled at platform 6 and two stabled at the inner tail track.
However, while that has some cost, it is much more frugal than any other design I have found to give 8tph Caltrains / other regional rail line and 8tph HSR / other long distance electric intercity.