06 December 2011

Holiday Required Reading

HSR Done Right

Sometimes, it's useful to look beyond the peninsula for context on what works best locally.  Here's a graphic from Richard Tolmach (in the latest TRAC Newsletter) that pretty much says everything that needs to be said about the California High Speed Rail Project.

As can be observed, the route that Tolmach and other organizations (including the plaintiffs in the Atherton lawsuits) have been advocating for years is very different from the route that the CHSRA is stubbornly advancing through the environmental clearance process.  In the Bay Area, the notable departure from the official plan is that HSR would branch off from the peninsula rail corridor at Redwood City, head over a new Dumbarton crossing, and zoom across Altamont Pass along the SETEC Alignment.

What does this have to do with anything peninsular?  Plenty, as it turns out.  Read on.

Caltrain's Blended Analysis

Caltrain recently published its analysis of the "blended" proposal, where Caltrain and HSR would share the peninsula rail corridor using less ambitious and expansive infrastructure than the four-track viaducts originally envisioned by the CHSRA.  This analysis concludes that it's feasible to run mixed Caltrain / HSR operations, although Caltrain service would be bunched up (with irregular skip-stop service patterns limited to six trains per hour) and HSR would need to slow down (about 40 minutes, rather than the planned 30 for SJ-SF) and be limited to 4 trains per hour.  On page 46, the document mentions that
The increased two-track shared use corridor distance from Whipple Avenue to San Jose Diridon, makes it very difficult for a 110 mph train to leave San Jose without encountering delay prior to reaching the overtake, and for a southbound HSR train to keep from being delayed by the Caltrain train it follows after the overtake.
Translation: sharing tracks should be done for the bare minimum distance, and certainly not 50 miles from SF to SJ.  Branching HSR off the corridor in Redwood City is a scenario that was NOT analyzed because it runs against Pacheco orthodoxy.  There is little doubt that it would make for an operationally superior solution (as computed by our free service pattern generator) with more Caltrain service, more Caltrain expresses, better transfer opportunities, easy-to-memorize clockface service patterns, and 125 mph HSR speeds... better in every way than the best scenarios LTK could come up with given the flawed assumptions of the study.

Speaking of better service planning...

The Swiss Take On California

Switzerland arguably has the most advanced, integrated and optimized rail service planning in the world.  The Swiss rail operations consultancy SMA+Partners supported a doctoral thesis analyzing the California rail network (including HSR) from an operations perspective.

Ulrich Leister's thesis (see executive summary) "applies a lean and rational approach to planning that is network and schedule-based.  A precise computer model is used to test different ideas such as infrastructure layouts or train types.  Gradually, the schedule is refined and optimized until the required rolling stock and the minimal amount of infrastructure needed to operate all the scheduled trains is determined."

This operations-first approach will likely come as a breath of fresh air to readers bewildered by our local experts' cost-maximizing ways.  A full copy of the thesis will be linked here as soon as it is made available.  Note in the network diagram at left that the Altamont route is identified as operationally superior, which will come as a surprise to CHSRA consultants who stubbornly insist Pacheco is the only way to go.

The Japanese Take On California

About a year ago, the East Japan Railway Company gave the CHSRA a peer review of their operations and maintenance approach.   Section of this document addresses mixed service with other rail carriers.  It is reproduced in full below, with links added to relevant articles that echo the exact same points on this blog.
Based on JR East's experience of operating conventional train and Shinkansen train on the same track, following three aspects should be carefully considered.
First, the timetable should be carefully planned. The shared operation segment is likely to be the bottleneck of the high speed train timetable since delay in the conventional line will affect the entire high speed trains network. Therefore, if transport capacity is required, 'parallel' timetable (that is, High Speed Train and conventional train operate at the same speed) or increase the capacity of the commuter trains and reduce the frequency will be the solution. To establish a more flexible timetable, additional facilities will be required both in high speed train and the conventional lines. For example, siding tracks are required in stations in this segment, commuter train vehicles with good acceleration should be implemented, speed restrictions on curves should be reduced, more signals should be allocated, etc.
Second, rolling stock should be taken account. If the High Speed Train vehicle width is different from that of conventional trains, platforms must be trimmed, and/or boarding steps must be installed either on the high speed train or on the commuter train. These boarding steps may exceed the loading gauge at some areas, so they should be stowed away while the train is running. The difference in height of the doors of the rolling stock should also be taken into consideration. Finally, compatibility of Automatic Train Control system for high speed train and conventional train should be considered. Since the safety equipment is indispensable for either train, multiple safety equipments must be installed on the rolling stock, and radio communication system must also be shared. These must be switched at the border station. Preventing malfunction both on the wayside and on-board is also important.
All this good advice has clearly fallen on deaf ears.  For example, platform interface coordination is not even remotely on Caltrain's radar, and the HSR project is actively working against it.


  1. Clem,

    I'm disappointed to not see a more in-depth response to Caltrain's -- LTK's -- "analysis of blended operations". Not only on the four overtake scenarios which were simulated; but on the "CBOSS uber Alles". The LTK report finally spells out two points which (one can only assume) are what Caltrain staff see as the key justification for CBOSS. And they're danmingly weak.

    As I've said elsewhere, I honestly think you'd get a much better handle on how the decision-makers arrived at Pacheco, if you spent some time looking at the difference in funding and route decisions, between French LGV and German high-speed ICE lines.
    The first is a strong central government. The second is a Federal sytem, where compromises are made to get buy-in from local states and cities. I'm not defending that, just observing, and that CAHSR seems much, much closer to the second than the first.

    As for the JR East proposal: one might charitably usefully translate a Japanese
    ".. should be carefully considered"
    into UK-english as
    ".. you're bloody stupid if you don't".
    I'm not qualified to judge a US colloquial equivalent. What would yours be?

  2. It's a fine study within the constraints of their assumptions, but misses important aspects of the bigger picture. There's not a lot of value in micro-analyzing the LTK results. They used a fancier train performance calculator than most people have access to, but it's not the tool that's important, it's what you do with it.

    Specific working assumptions that make no sense:

    1) all Caltrains during rush hour are skip stop expresses. Such a hard to memorize and irregular service pattern is not conducive to building ridership, and harks back to a regular 8 to 5 commute rather than today's modern fluid up-to-minute work environment and lifestyle. Big stops are well served (if irregularly) and stay big, while small stops are forever underserved and stay small. This service paradigm is not the only one out there, but Caltrain has latched onto it as the only solution.

    2) There's no such thing as a Caltrain express. Overtakes are HSR only on extra special private HSR tracks. I've pointed out repeatedly that Caltrain-Caltrain overtakes enable both faster express service AND more frequent local service. That doesn't compute at headquarters in San Carlos.

    3) Since the study was coordinated with CHSRA, all the crazy infrastructure (such as double-deck Diridon and the six mile viaduct to Lawrence) is included. Fail.

    4) Again since the study was coordinated with CHSRA, any Altamont scenarios are strictly verboten, with all the operational headaches this entails.

    So, basically what LTK did was narrowly constrained by these assumptions, and the study did not answer the real question of what blended operations could look like on the peninsula corridor under a lean, integrated, service-oriented, infrastructure-minimizing paradigm. It solidly answered the question of what blended service looks like under a crippling set of operational and infrastructure assumptions.

  3. I think some of those recommendations from JR are things they had to learn the hard way themselves. The Shinkansen network was originally designed to be entirely separate from the mainline network and no thought at all was put into interoperability, because why would they ever need it? Well, it turns out that they did, for the mini-Shinkansen lines that run on conventional lines that were converted to standard gauge.

    As far as Caltrain operations go, one thing I noticed from looking at ridership statistics is that their least popular train in the rush hour is the one that makes all local stops north of Redwood City. Judging from that, it seems like much of Caltrain's market is longer distance commuters looking for a quick ride between SF and stations like Mountain View and Palo Alto. This implies that most trains will be fast, and you might be able to get away with relatively few overtakes. Another option is just slowing the HSR trains down at the height of peak hour, and speeding up the Caltrain expresses. The fewer different speeds of train there are, there more capacity you get, even if you have local trains getting overtaken.

  4. I think some of those recommendations from JR are things they had to learn the hard way themselves.

    Certainly, but that doesn't mean we have to make the same mistakes.

  5. But we're Merka, dammit! We don't need to listen to a stinking thing anyone else says!

    (Kiwi, a good translation would be: "If you don't do this, you're f**king yourselves over."

  6. "This blog accepts that HSR will run between San Jose and San Francisco on the Caltrain corridor. Many other alignment alternatives have been considered and formally eliminated, and it is not the purpose of this blog to reopen decisions that have been years in the making. While the Altamont Pass alignment would likely have been superior to the Pacheco Pass alignment, we let bygones be bygones. This blog will focus on issues and decisions that are ahead of us and that we can realistically try to influence."

    So I take it you will be revising your About This Blog page?

    While Altamont makes a lot of sense, some of the other proposals you feature in this post are just plain daft. Are you now in favor of an I-5 alignment? Can you cite any other examples in the world where a high-speed line has no stops for 300 miles?

    I think TRAC are naive to think that modern passenger rail operations will ever be possible on freight owned tracks, but at least they have an inkling of a plan for providing service to Fresno and Bakersfield. The SMA plan shows no stops whatsoever between Modesto and Palmdale and no mention of how the intermediate cities will be served.

  7. fJon- no stops is better than a number of minor stops none really justifying a HS station, especially when alignment issues will make them P&R stations anyway. The great thing about a direct (faster) routing is that it would make suburban stops in LA and SF more palatable.

  8. Can you cite any other examples in the world where a high-speed line has no stops for 300 miles?

    Add a stations along I5, say near Los Banos, somewhere between Coalinga and Kettleman City, and at the intersection with 99. Put a decent size parking structure at each. Where traffic demands, transit systems could eventually provide connecting service to Merced, Madera, Fresno, Hanford, Visalia, and Bakersfield. Not an ideal solution, but preferable to driving, or even flying, from these cities to the Bay and LA areas. And, of course, preferable to no HSR at all...

  9. Both you guys are trying to solve a problem you just created. The solution to the problem of not having anywhere on I-5 worth stopping at is not to route the line along I-5. Instead, route the line where people actually live. For the ~20 miles added distance to go through the Central Valley rather than along I-5 you gain ~3 million potential riders. Well worth the effort in my opinion, especially when compared to the Tehachapi detour, which adds ~40 miles compared to Tejon but gains just ~500,000 potential riders .

  10. Trying to get through cities at 200+ mph is kind of silly, but not serving the cities altogether is also not very smart.

    What's the compromise? Probably a greenfield alignment. Ag would of course complain, but the bottom line is that farmland is cheap, and the mitigations required to allow farming to continue are much cheaper than the mitigations required in cities. Depending on how much additional track would have to be build/how difficult the alignment is/how willing cities are to commit to TOD etc/how big the cities are to begin with, some cities might justify a station loop (build for much lower speeds and only a two track station, no 60 foot viaducts or wide curves through built-up areas).

    In other cases cities might be difficult enough to bypass/enough trains would be stopping that it would make sense to slow the express trains down to 150 through the city (costing a couple of minutes in most cases).

    In all other cases a greenfield station would be adequate. A downtown station certainly incurs certain benefits but at times the additional cost just doesn't justify them.

    I will outline what I actually had in mind for California, but I will stress that this would have to be studied by someone who actually had the resources to do it right and I'm really just guessing.

    a=bypass+station loop, b=slow down express trains, c=greenfield station

    Stockton: b
    Modesto: a or c (station loop relatively easy but might not be justified)
    Merced: a or c (as above)
    Fresno: a
    Hanford: c (if any station)
    Bakersfield: b (if Tehachapi) or c (if Grapevine)

  11. I really like the idea of a "bay hub" in Fremont as a connection between north and south services. I'm guessing it would be located at Shinn St. where BART crosses over the UP. It's unfortunate to see that the latest dumbarton rail plans and altamont corridor plans being conducted in a vacuum. Both projects would work well together, but they don't even connect. Pathetic.

  12. I'm certainly in favor of station loops (like the one proposed for Gilroy but not carried forward) to help reduce impacts to downtown. I'm also in favor of stations on the edge of cities if necessary (e.g. near a freeway interchange) although this is a less desirable solution than a downtown loop. And yes, if the line does go through downtown it may be necessary to slow down express trains as they pass through.

    What I am not in favor of is routing the line nowhere near major population centers so that anyone from the central valley who wants to use the train has to drive an hour to get to a station. Fresno, for example, is 1hr 15mins from the closest point of I-5 on low-capacity roads. No one in their right mind will drive that far to get to the station. All that will happen is the creation of more sprawl as farmland is eaten up for development near the park & ride stations.

  13. @Jon: opinions change, and mine certainly has in the three years I've looked into the nitty gritty details. Accordingly I have revised the text you quoted, which no longer reflected my editorial stance.

    I kind of opened a can of worms by posting the Tolmach map... let me point out that I don't feel very strongly about the I-5 routing and I agree that an alignment that comes closer to the major population centers in the Central Valley is probably better.

    Tolmach does have a point that the routing is crooked.

  14. How much time would be added if express trains were slowed to let's say 300 km/h when entering the CV cities? Discounting time lost in acceleration and deceleration (calculating that out is beyond my technical abilities right now), I get a total of 1 minute and 20 seconds difference for traversing Fresno and Bakersfield at 300 km/h versus 350 km/h. I would not be surprised, especially if significant time is saved by a Tejon alignment, if the Authority makes a deal with the CV cities to plan for slower speeds (and the decreased environmental impacts it entails).

  15. You need to build the line with population in mind. Running HSR down the I-5 median will be fast, but it will destroy ridership, which after all, the majority of critics keep complaining about. As for the route through Palmdale or to Santa Clarita, which ever makes more financial sense, stick with that, but keep in mind that Palmdale can offer future connections to Las Vegas via Desert Express.

  16. @Clem- Well, it's your blog and your choice what to focus on. Personally, I come here for the excellent technical analysis. The zombie Altamont vs. Pacheco debate is just plain boring.

    The SMA document you linked to is quite amateur. For one thing it contradicts itself by excluding a Bakersfield stop from the route map but then discusses service to Bakersfield later on in the text. And there is no mention of Fresno at all. Why would you stop in Bakersfield and Modesto but bypass Fresno?

    All it is is someone's academic paper. The fact that SMA supported it doesn't really mean anything- it's far from a "Swiss take on California". The routing opinions most likely originate from the student rather than from SMA. If SMA were paid by CAHSR to do a routing analysis I'm sure their opinions would be different- that's the nature of consulting.

    @Jarrett- the 'Bay Hub' would probably be located here at Paseo Padre Pkwy, just a little north of the planned Irvington BART station at Washington Blvd.

  17. "If SMA were paid by CAHSR to do a routing analysis I'm sure their opinions would be different- that's the nature of consulting."

    The nature of consulting in California is that the lead consultant creates the agency, the lead consultant tells the agency how to vote, and the lead consultant tells the subconsultants what answers to provide if they wish to continue to lead full and happy lives.

  18. "For one thing it contradicts itself by excluding a Bakersfield stop from the route map but then discusses service to Bakersfield later on in the text"

    I don't see Sunnyvale on the map.

    The only conclusion to be drawn is that the Swiss Navy plans to wipe that city the face of the earth under the New World Order.

    It gets worse, too. I don't see the bus stop in at Haslen Al, Schiessegg anywhere in
    this so-called "map" of the fatherland. Dilettante catrographers! Geographical ignoramuses! Amateur planners! Begone!

  19. ...I can't even begin to list the ways in which your comparisons are invalid. Ever heard of a strawman argument?

  20. the subconsultants what answers to provide if they wish to continue to lead full and happy lives.

    Except when they are in Millbrae and points south and want to get to the airport. That hasn't worked out the way they were proposing.

  21. Sometimes the unions interfere. Nothing that couldn't be solved by pouring more concrete.

  22. That Tolmach diagram is quite the nice piece of graphical persuasion, isn't it?

    I don't drink the I-5 Koolaid, but seeing the map of California presented that way does do its job of making one think again, or think more closely, or scrutinise one's own assumptions and prior analyses.

    Good stuff.

  23. The CHSR Authority’s planned Sylmar-Metro-Link-Station to the I-5_SR152 intersection area rail distance is 308 miles. The shortest road distance between these two points is 238 miles with 99% of that distance being close to I-5. This 70 mile shorter distance would save 20 minutes when averaging 210 mph within a high speed section. A 20 minute savings by only increasing the maximum CHSR speeds along the currently planned route between Sylmar and the I-5 – SR area would require a 285 mph maximum speed. Both scenarios assume a 4.76% reduction in average speed compared to maximum speed due to mountainous terrain climbing and curve delays plus similar schedule pad percentages.

  24. You're including the Palmdale detour in your comparison, which accounts for about 40 of the 70 mile difference. Here are some very rough calcs for SF-LA distances. These numbers are only indicative, but they do roughly match up to your numbers and Tolmach's numbers.

    Altamont, I-5, Tejon: 401 miles
    Altamont, CA-99, Tejon: 420 miles
    Difference: 19 miles

    Altamont, I-5, Palmdale: 457 miles
    Altamont, CA-99, Palmdale: 466 miles
    Difference: 9 miles

    Pacheco, I-5, Tejon: 388 miles
    Pacheco, CA-99, Tejon: 415 miles
    Difference: 27 miles

    Pacheco, I-5, Palmdale: 444 miles
    Pacheco, CA-99, Palmdale: 461 miles
    Difference: 17 miles

    Using I-5 instead of CA-99 saves 9-27 miles depending on the other routing choices. By the same estimation, using Tejon instead of Palmdale saves 46-56 miles, and using Altamont instead of Pacheco adds 5-13 miles.

  25. An April 2011 Trains Magazine European high speed rail/air modal split chart on page 31 shows that rail takes 90% of the total rail plus air traffic for two hour rail trips but only 55% rail patronage for 3 hour rail trips. Cutting San Francisco to Los Angeles CHSR run times from 2:38 to 2:18 may increase SF to LA ridership significantly. If the CHSR SF to LA ridership demand curve exhibits a similar ridership sensitivity to rail trip times as the Trains’ rail/air modal split chart shows a 20 minute time saving could result in a SF to LA ridership modal increase of 15% which would result in a rail ridership increase of (0.70/0.55)(100%) = 127% or a 27% ridership increase. (The forgoing analysis assumes that the total number of corridor travelers remains constant after a service improvement. A significant improvement by a major travel service provider, such as High Speed Rail, would increase the total number of SF to LA travelers. If total corridor demand was completely elastic a 15% modal increase would result in a (1+ 0.15)(.0.7/0.55)(100%) = 146% or a 46% ridership increase.)
    The exceptionally low population density along the I-5 north of Santa Clarita will sharply reduce Central Valley CHSR I-5 line sound mitigation expenses. The CV local branch 110 mph speeds would not require extensive high sound walls or wide separations from parallel freight rail tracks. Current railroad freight right-of-ways could be used without an unacceptable safety hazard. And few property takings would be required in order to proceed.
    No significant air service will draw off CV passengers. A parallel 110 mph local branch SR99 alignment through Bakersfield and Fresno should be connected at both ends: the I-5 – SR99 divergence and the SR152 – I-5 intersection. The CV local line’s high speed connections to both SF and LA would permit CV origin trips to be made in half the time required for similar automobile trips.

  26. You do realize that I-5 is 50 miles from Fresno? Building 50 miles of 110 MPH tracks isn't going to be cheap. And that Fresno is 110 miles from Bakersfield. Building 160 miles of "cheap" 110 MPH track isn't going to be cheap.

  27. It would be a whole heck of a lot cheaper than building HSR as currently planned.

    For HSR, track, systems and electrification (if even needed for a 110 mph corridor) account for about ten percent of the overall budget. Peanuts. The other 90% is grade separations (no thanks), viaducts (no thanks), fancy buildings (no thanks), as well as utility relocation and ROW acquisition (not a problem if triple-tracking an existing freight corridor).

    You could probably build a non-grade-separated 110 mph corridor along CA99 AND a 220 mph corridor along I-5 for the same price as what they are now proposing. These corridors would connect at each end, not East-West across the valley.

  28. Building a mile of 110 mph track is not all that hard when you already have an 80 mph track in place. In fact, you don't even have to do any grade separations, just a few curve easings, install PTC (which is mandatory anyway), and inspect the track a bit more frequently and to higher tolerances. And guess what: there's already a perfectly good railroad linking all the Central Valley cities. Two of them in fact.

  29. NJTransit lays new track for 5 million a mile. That includes prepping it for ACSES. 320 miles is a is just over a billion and half. How many of those two and three track grade crossings are you going to be able to make into 4 and 5 track grade crossings? A glorified Amshack by the side of the tracks is going to be a lot cheaper than the HSR stations in downtown Fresno or Bakersfield but you have to build transfer stations at the end of the 110 MPH tracks which are going to be almost as costly as the downtown stations. The real estate is going to be cheaper out on I-5 but the elevators, escalators high level platforms etc are going to be almost the same cost. Or were you thinking of stringing catenary for 160 miles so the 220 MPH trains could mosey along at 110? .....
    So, just how much does 160 miles of 110 mile an hour track cost?

  30. John:

    1. The air/rail mode split is fairly linear. See PDF-page 8 here and figure 2-4 here.

    2. The travel distance difference between I-5 and CA 99 on Google Maps is 35 km (via Altamont and Tejon). Thus at 350 km/h the travel time difference is 6 minutes, not 20 minutes.

    3. The advantage of doing everything on one trunk line is that frequencies add up. LA-CV and CV-SF trains combine to form LA-SF local trains, in addition to express runs, and this improves service, especially early on when ridership is just starting to build up. Would you rather have a train that does LA-SF in 2:30 coming every hour, or a train that does LA-SF in 2:36 coming every hour plus another train that does LA-SF in 2:55 coming every hour?

  31. @ John Bacon

    I'm not sure how much you think sound mitigation is going to cost, but it's just a blip on the overall budget.

    I would favor reducing speed from 350 km/h to 300 km/h through Fresno and Bakersfield in lieu of building more sound walls.

  32. The problem with slowing any part of the line down is that the rest of it will have to be sped up to compensate, because the total end to end travel time is set by law, and based on a fairly optimistic engineering estimate, not a worst case one, much less an actual detailed design.

  33. I'm not sure how much you think sound mitigation is going to cost, but it's just a blip on the overall budget.

    Duh. It is an externality.

  34. Alon: Accepting your implied suggestion I have downloaded the Brazil TAV Project Figure 2-3 on European Rail/Air mode split compared to rail trip times. The Rail/Air mode split ratio does decline at a constant rate of 17% per hour of additional train journey times between 3.5 and 4.5 hours. But in the 2:30 to 2:45 train journey time range the Rail/Air mode split ratio decline rate is 62% per train journey hour. Therefore a 10 minute time savings from 2:40 to 2:30 and would result in a 10% increase in ridership. The current CHSR Authority run time goal between San Francisco and Los Angeles is 2:38. Therefore given the current SF to LA minimum run time estimates a moderate reduction in the SF to LA rail distance would have a significant impact on run times and ridership. Compared to the CHSR Authority’s current plans a continuous I-5 corridor alignment between Sylmar and SR152 will save 62.7 miles, and 18 minutes train run time for SF to LA express trains averaging 209 mph through the Central Valley.
    The CHSR Authority’s Map with Mile Posts from TM-1.1.8 under the Caltrain HSR Compatibility Blog’s of March 5, 2011 posting “The Prescriptive Framework – Update” lists the Sylmar Milepost at 425.8 and the CP Divide milepost near Chowchilla at 160.1. The 2005 Rand McNally Road Atlas lists the mileage along SR152 from SR99 to I-5 at 43 miles. Therefore the current CHSR Authority’s rail mileage estimate between Sylmar and the I-5, SR152 intersection vicinity is 308.7 miles.
    (425.8 – 160.1) + 43 = 308.7 miles
    From the same Rand McNally Road map the I-5,SR152 intersection is listed as exit #403. The I-5, SR14 intersection is listed as exit #162. Subtracting those two exit numbers plus a Google Earth aided mileage estimate of 5 miles between exit #162 and the Sylmar Metro Link Station gives a total I-5 corridor mileage of `246 miles.
    (403 – 162) + 5 = 246 miles
    The mileage reduction of the I-5 corridor vs. the CHSR Authority’s Mile Post Map’s distance is 62.7 miles.
    (308.7 – 246) = 62.7 miles

  35. From Peter: I'm not sure how much you think sound mitigation is going to cost, but it's just a blip on the overall budget.
    I would favor reducing speed from 350 km/h to 300 km/h through Fresno and Bakersfield in lieu of building more sound walls.
    Aerodynamic resistance sound energy is proportional to the cube of the speed. Combine that physical reality with the observation that human auditory sensitivity quickly rises as sound frequencies rise within the sub 500 Hz range. Higher frequencies which tend to move in straight lines can be significantly attenuated by sound walls. But a TGV noise source chart indicates that at least 20% of its sound foot-print comes from current collecting pantographs 12 to 18 feet above track level. Would one need to build sound walls at least 18 feet high through densely populated areas for 220 mph trains?
    Would adequate sound insulation from 220 mph trains through urban areas be expensive? Consider the following notice to residents adjacent to BART’s Silicon Valley Extension for 129 km/hr trains:
    “The Residential Noise Insulation Program (RNIP) will provide noise insulation for the upper floors of residences that may experience potentially severe noise impacts from the BART Silicon Valley Berryessa Extension (SVBX). The residences are identified through studies that determine the noise levels attributed to the SVBX Project that may not be reduced with existing or proposed sound walls. If your residence is found to potentially have an impact, this program will install acoustical windows, doors, and other improvements at no cost to you.”
    Google Earth urban sprawl measurements for both Fresno and Bakersfield appear to have a diameter of 20 to 23 km. In the Central Valley population density along the I-5 corridor is a small fraction of the population density anywhere close to a direct line between Bakersfield and Fresno. Wouldn’t a 220 mph I-5 corridor CHSR alignment parallel to a 110 mph SR99 corridor regional service line connected to the I-5 220 mph alignment south of Bakersfield and west of Fresno with few high sound walls needed for either line be cheaper than the present combined Central Valley non-stop and regional service along one track-way approach?
    Suppose the CHSR single track-way scheme is actually built through Fresno and Bakersfield. If noise mitigation efforts prove inadequate for Fresno residents could adjacent Fresno property owners force CHSR trains to slow down with a successful lawsuit?

  36. It would be a whole heck of a lot cheaper than building HSR as currently planned.

    For HSR, track, systems and electrification (if even needed for a 110 mph corridor) account for about ten percent of the overall budget. Peanuts. The other 90% is grade separations (no thanks), viaducts (no thanks), fancy buildings (no thanks), as well as utility relocation and ROW acquisition (not a problem if triple-tracking an existing freight corridor).

    You could probably build a non-grade-separated 110 mph corridor along CA99 AND a 220 mph corridor along I-5 for the same price as what they are now proposing. These corridors would connect at each end, not East-West across the valley.

    I would bet that these two proposed lines would be considerably cheaper than the current proposal of a full-build HSR line along the CA99 corridor, which is essentially a low-density sprawl of housing and industrial farming activity (ie, lots of grade separating to be done).

    The I-5 corridor would allow the highest speeds without any NIMBY opposition and delay, and relatively few expensive grade separations would be required, making it distinctively more affordable. The I-5 corridor could easily meet the 2:40 SF-LA requirement, connecting the two metropoles in California in the most direct manner. Also, when you're racing at top speed between major markets, you really don't want to stop, especially for low-travel-demand places like Fresno. Check the puny CV airline markets for profiles of meek CV longer-distance travel demand.

    The additional 110-mph CA99 line would also be much more affordable than the sure-to-be-bitterly-fought 220-mph HSR line in the CA99 corridor. Few, if any, grade separations are required, lowering its cost and its profile to local opposition. An accommodation with the freight rail firms is far more easily reached, as such a scenario could be made to appeal to the freight's interests. 110-mph service along an upgraded San Joaquin route also serves the Central Valley better, with many more local stops. Just look at the existing San Joaquin ridership profile: Central Valley users take the train to go up and down the Valley considerably more than going to the Bay Area or LA.

    Not only much cheaper, both lines could be built over a much shorter time horizon, primarily due to less project complexity and much less local opposition.

    Two lines would indeed be far better than a single massive construction boondoggle line. Of course, the redeeming feature of the current proposal (to special interests) is that lots and lots of concrete will be both designed, re-designed, and then finally (maybe) poured, all at enormous capital outlay.

    It's amazing how the simple optimal solutions get overwhelmed by the stupid solutions...

  37. @ John Bacon

    No one has come up with a number yet to compare the costs of the current plan to I-5 PLUS Upgraded San Joaquins, so I can't tell you which is cheaper. I don't see how it would be very much cheaper, honestly.

    You talk about how noise increases with speed. I do not disagree. My point is that a speed reduction would be cheaper than installing more sound walls through Fresno and Bakersfield. That says nothing about how many soundwalls would be needed for 300 km/h service, just that it would be less than for 350 km/h.

  38. Peter, HSR building is all about finding a straight-as-possible, unobscured path for trains. It's very similar in concept to a limited-access highway, and the construction techniques are similar. The sort of opposition it kicks up is also similar. The main cost factors boil down to building grade separations and finding straight ROW paths.

    The I-5 corridor is already remarkably straight and has very few obstructions for hundreds of miles. This means few grade separations to build and easy ROW acquisition. No NIMBYs for hundreds of miles...

    The CA99 corridor has many obstructions for HSR building. The area is not densely built-up, but it is built-up sufficiently to require many expensive grade separations. It has enough local residents and local interests to create significant opposition. The freight rail ROWs are adamant against passenger rail speeds above 125mph, so ROW acquisition is only easy with speeds below 125mph. Acquiring an entirely new ROW along the CA99 corridor is both expensive and disruptive of local interests.

  39. @Peter re noise:

    Richard Mlynarik has argued often that 350kmh running is not going to happen — or at least will never be sustained anywhere in the world in the medium to long term — for compelling obvious energy consumption reasons ... as well as noise, maintenance cost, train cost, no economic advantage in very small time saving, etc.

    But that's not nearly what speed reduction to mitigate noise in urban areas needs. That is like 250kmh even in some of the less lovely Central Valley city places with freeways and industries ... 160kmh or below in more affluent people's back yards. Real world experience.

    On the other blog people get excited how great it is to have elevated high speed trains cutting through small Japanese or Chinese cities right next to apartment buildings, but should we expect or want such a bad result here? And we do ignore lots of reasons (lawsuits!) to think it won't happen. Don't make waves!

    What to learn from all HS lines of Europe is very fast trains go around cities. Maybe some trains stop at a peripheral station (OK as compromise) or maybe the stopping trains run at lower speed on a side loop track into the downtown (more cost more ideal), but nobody except a rail fan wants a train at 270, 300, 350kmh anywhere close to where they are living or doing business. Fact!

    Long comment, sorry. But mitigating noise by slowing from 350 to 300 isn't a mitigation, because I don't think the train will run at 350 anyway and because 300 is still LOUD. I have done the rail fan fun activity of standing on platforms and by very high speed tracks. Whoosh! Fun for me, but just a hobby, and then I can go home away from airports and freeways and HS tracks and steel mills.

  40. @ Caltrain First

    I don't disagree with you. I'm just saying that no one has come up with anything other than conjecture that building BOTH an express I-5 AND an upgraded San Joaquins is cheaper than the current plan.

  41. John:

    My previous comment got eaten, so let me try again. You're conflating two issues - I-5 vs. CA 99, and Tejon vs. Palmdale. Tolmach loves comparing the lengths of 5-Tejon and 99-Palmdale to make it look as if I-5 is the only rational option, but in reality more of those 62 miles of extra distance come from Palmdale and not from 99. Moreover, the extra distance coming from Palmdale is at lower speed, about 200 km/h give or take, because of the grades; the short extra distance coming from 99 is at full speed, which is why it matters so little.

    Caltrain First:

    Like John, you're conflating two issues. Noise and urban construction costs can be avoided by skirting urban areas, LGV-style. That's what Clem's been arguing for: a Fresno station a few km west of downtown, and a Bakersfield station either east or west of the urban area. (For the record, I think this concept works well for Fresno because the station would be reasonably close to the city, but not for Bakersfield.)

    If you want to pretend the farmland acquisition cost of skirting urban areas is more than a third-order term, go ahead. But urban impacts are not an argument for I-5.

  42. You could probably build a non-grade-separated 110 mph corridor along CA99 AND a 220 mph corridor along I-5 for the same price as what they are now proposing.

    Citation needed. I have a feeling the freight railroads don't want lightweight EMUs anywhere near their existing track and will demand a 100 ft buffer just as they did for HSR. A derailed freight train will fuck up a lightweight EMU travelling at 110mph almost as badly as a HSR train travelling at 220mph. If they do demand a 100 ft gap the cost savings will be small compared to HSR, the only real difference is you can make some curves tighter and use quad gates for grade crossings. This is all just speculation until someone gets the freight railroads onboard and comes up with a detailed study.

    Alon, I'm glad you at least can see the apples-and-oranges nature of the I-5 vs. CA-99 comparison. The big detour is to Palmdale, not Fresno and Bakersfield.

  43. @Jon: at 110 mph, you can use existing Amtrak equipment which is 100% compatible with freight trains. You don't even need to electrify. Use some of the savings to implement a kick-butt cross-platform transfer to take the sting out of the two-seat ride where the CV line meets the high-speed line.

  44. A derailed freight train will fuck up a lightweight EMU travelling at 110mph almost as badly as a HSR train travelling at 220mph.

    If you stop abruptly at 110 your guts don't splatter quite as far as they would if you stop abruptly at 220. But they still get splattered. And you are equally dead.

    at 110 mph, you can use existing Amtrak equipment which is 100% compatible with freight trains.

    Hard to do if the freight train ahead of you is toddling along at 55. Start building passing sidings and grade separations because the CPUC and the FRA will not approve 4 track grade crossings and you are spending almost as much as if you built 220 MPH track.

  45. at 110 mph, you can use existing Amtrak equipment which is 100% compatible with freight trains.

    Under FRA rules, the Amtrak service would require a 50% operating subsidy. Perhaps if the FRA allowed for mixed HSR/freight traffic, there is opportunity; otherwise, I don't see great cost/benefit for both an I5 HSR and CV conventional line.

    An I5 alignment alone would actually be "good enough" for a lot of CV trips. Bakersfield is only a 30 minute drive to I5, comparable to the travel time for East Bay residents to get to an HSR station.

    Of course, an I5 route is not so great for Fresno or Merced, where drive time would be more like 45 minutes. But if the cheaper I5 route makes funds available for Sac and Stockton service, then this is still a reasonable engineering trade-off.

  46. @Drunk Engineer: I'm not sure anyone really knows what sorts of operating subsidies a non-FRA compliant intercity rail service would require in this country, as compared to an FRA-compliant one. Just keep in mind that imaginary trains can be much cheaper than real ones, since they don't have to have all the expenses of actually existing.

  47. Just to add to the list of options for CV alignment: Generally following CA-99 while bypassing cities on the way. Stations will be either greenfield or downtown with track connections to the trunk line (probably only in Bakersfield, maybe Fresno) This will avoid most of the noise issues while offering better station access compared to I-5 alignment
    (but worse than going through the cities).

  48. at 110 mph, you can use existing Amtrak equipment which is 100% compatible with freight trains

    Ah, you want to use FRA dinotrains and force a transfer. I guess that's doable, but I thought we were trying to build a modern passenger rail network? My assumption was that it would be something like the Caltrain blended plan, or the upgraded Altamont Corridor plan, with non-compliant electric equipment being able to move between the HSR mainline and the slower Central Valley line. That would be more attractive, but in the current regulatory environment probably not much cheaper than full fat HSR.

    Of course, an I5 route is not so great for Fresno or Merced, where drive time would be more like 45 minutes.

    Wow. How fast do you drive?

    But if the cheaper I5 route makes funds available for Sac and Stockton service, then this is still a reasonable engineering trade-off.

    If we're talking cost/benefit, we should start by eliminated the spending that produces no operational benefit rather than the spending that does. As Clem has repeatedly demonstrated there are several billion dollars to be saved through value engineering on the peninsula alone.

  49. My assumption was that it would be something like the Caltrain blended plan, or the upgraded Altamont Corridor plan

    The Caltrain corridor is only 50 miles long. The high speed trains won't be very high speed. If there is a freight train moving at 55 every hour or so and a conventional passenger train moving at 110 every hour or so and you want to blast through at 220 between Stockton and Bakersfield there's going to 2 or 3 conventional trains to pass and 4 or 5 freight trains to pass. Unless you want to four track lot of it everything has to run with Japanese like precision all the time. Even the Japanese would tell you that it wouldn't be a good idea.
    ..and if there's a train that gets you from Stockton to Bakersfield in a little over an hour why would anyone take the more expensive train that takes 2 hours?

  50. A "shared corridor" could mean that the public pays for extra passenger train track and track access within the freight RR corridor.

    Not saying this is a good idea! But this is what it might mean.

    Build a new FRA track next the to freight track for running FRA passenger trains only. (Maybe it is "only" 25 feet separated, not 100 feet or more they demand for HSR.)

    Maybe under emergency the freight RR gets to use the passenger track, and maybe in exceptional circumstances the FRA passenger train gets to detour onto the freight track, but normally they run parallel and separate with crossovers locked to normal (straight ahead) position.

    This is sort of like UPRR versus Caltrain from Santa Clara past Blossom Hill. (Not like UPRR versus Capitol Corridor, where the state pays and pays the RR and gets the very short end of the stick.)

    With more than two trains an hour unlikely, and no freight interference, even Amtrak might run a schedule on a single track with scheduled passing loops.

    The advantage of this is that everything is terrible crappy FRA, which is easier for the freight RR to get its head around. Also, some real benefit to the freight RR of emergency detour track. With separate HSR in the corridor, all the freight RR gets is a some small money up front and lots of liability. Maybe less track separation and real estate cost.

    The disadvantage is FRA.die.die.die. Too heavy, too slow, terrible fuel efficiency, over staffed, terrible operating cost, level boarding much less likely, bad operating culture (less than one day late is on time isn't it?), institutionalized sub-mediocrity, and likely Amtrak monopoly. Set "expectations" to "rock bottom".

    I think it is not a good idea. But it is an idea.

  51. Note that a good number of FRA trains have level boarding whereas most mainline trains in Germany or France or Switzerland do not. Yes, there are actually some things that the US does better in railroading. Not very many, mind you, but some, and actual level boarding is one of them. In Germany, for example, it's just two steps into the ICE, and not level boarding by any means. Also, it's not like all American trains are grossly overstaffed. From what I saw in Germany, the intercity trains had two conductors, which is fairly typical of Amtrak too, and the commuter trains (S-Bahn in Dresden) had one conductor, which is certainly doable under current regulations, because Metrolink does it.

  52. On the NEC, there are three conductors per train. And that's a short train, i.e. 8 cars - I'm not sure how long the 2-conductor trains are in Germany, but in Japan there's just one conductor even on a 16-car train.

    Dwell times at most intermediate stations are 1.5-2 minutes, regardless of level boarding, since the stations without level boarding are less busy. More annoyingly, there are no plans for implementing level boarding at stations that do not currently have it.

  53. The San Joaquin Valley’s landscape is radically different than anywhere else where high speed rail now operates. The SQV’s western edge has a far more stable base needed to support HSR railway tracks and almost nobody around to object to the considerable noise 220+ mph trains will produce. Compared to the middle of the San Joaquin Valleys’ high population density there are very few people living along the SQV’s western edge. Separate San Joaquin Valley track-ways are well justified in this special case.
    But in most cases concentrating all trains serving a corridor along one track–way has overwhelming advantages over multiple separate track-ways along a travel corridor. Even one non-stop track in each direction has such enormous capacity that the operation of more than one track in each direction except in close-together station sections is rarely practiced or justified. From an efficient use of rail technology point of view running CHSR trains along the full length of San Francisco Peninsula along a mostly four track system between San Francisco and San Jose makes sense. An express track can accommodate at least one 400 meters long 200 km/hour train per minute while simultaneously observing today’s conservative rail safety margins. Because of high express track capacity both CHSR and Caltrain services can share an express track with little mutual interference. (Safe, optimized for maximum theoretical capacity train separation control systems that accommodate the foregoing capacity description for rail transit is likely to be available at moderate cost in the foreseeable future. The current electronics improvement revolution plus the world-wide tendency to build high traffic density transit railways train control suppliers are likely to provide numerous safe competitive train control system choices at moderate cost.) Because of the possibility of using alternate paths around delayed trains when parallel local and express tracks are frequently connected to each other both train reliability and scheduled speeds for both CHSR’s and Caltrain’s trains can improve due to the reduced need for schedule padding. Also numerous cross-platform-transfer-stations will significantly increase the number of convenient travel options for both CHSR and Caltrain riders.
    The fact that two major airports are close to the Peninsula rail right-of-way results in a strong incentive for San Joaquin Valley residents to use the CHSR in order to connect with long distance air service. Plus Silicon Valley is now the fastest growing industrial area in the United States. Their work force is relatively young, fairly well compensated, and inclined to travel often. These strong travel to the peninsula incentives plus a local population inclined to travel are further reasons for a CHSR peninsula approach to San Francisco.

  54. John, you keep saying things like "An express track can accommodate at least one 400 meters long 200 km/hour train per minute," and they're not true. The relevant ETCS 2 factsheet states that the minimum headway for nonstop 200 km/h track is 1:37 and the minimum for nonstop 300 km/h is 2:30.

    But that's minimum headway between two trains, not maximum frequency. Richard tells me that Switzerland's Mattstetten-Rothrist line, with a headway of 1:50 at 200 km/h, doesn't actually run 32 tph. It runs 7 trains in 4 minutes twice per hour instead. The need for short headways comes not from the need for high tph count, but from schedule integration with connecting lines.

    Likewise, we know what the maximum achievable headway is at high speed because the Tokaido Shinkansen is at it. Local stations have two stopping tracks and two express tracks, and express stations have four stopping tracks. And the maximum frequency is about 14 tph, at 270 km/h.

  55. @Alon: from my recent (couple months ago) experience in Germany, the 2-conductor trains were about 8 cars long, not unlike the typical NEC trains. The particular ones I was one were actually not DB stock, but rather international trains operating from Poland to the Czech Republic with Czech stock, and from Budapest to Berlin using Hungarian stock, but in both cases, the onboard employees were German. I think there was also a food cart attendant in addition to the conductors.
    The commuter trains, which had one conductor, ranged in length from one(!) to four double-decker push-pull cars hauled by an electric locomotive.

  56. As Jarret Walker would tell you, branching divides frequency. The three major cities of the bay area are San Francisco, Oakland, and San Jose, and this plan puts them on three separate branches, meaning no train could serve even two of the three. They'll each need their own local trains and express trains to LA, putting high capacity demands on the center of the network, reducing frequency for each of the cities, and generally making it harder to fill the trains.

    If you're going to create a new bay crossing, why put it there? Is serving Redwood City that important? Just so the trains still share more than half the Caltrain corridor? If you want to separate HSR and Caltrain, put a crossing between SF and Oakland and build an HSR corridor from Oakland to SJ, then through Pacheco. Now you need a lot fewer train service patterns and so service is more frequent and convenient and more likely to be profitable. As a bonus Caltrain could be extended to the east bay.

  57. Alon: This response is intended to conform to the notion that train frequency per track estimates should not exceed the maximum trains per unit time possible when constrained by transit industry safety related maximum credible brake performance and minimum train separation norms. According to the Transit Cooperative Research Program, sponsored by the Federal Transit Administration, Report #13 on Rail Transit Capacity on page 77, Table 7.1 lists transit industry norms for system and vehicle characteristics. It lists moving block safety distance as 50 meters = 164 feet and maximum credible braking rate as 75% of 1.3 meters/sec^2 = 3.2 feet/sec^2. Common sense tells you that the maximum credible braking rate for a give railway track section should be reduced by the steepest downhill acceleration opposing braking action present anywhere along a given railway track section where a maximum assured braking rate standard is being established. Along Caltrain’s San Francisco to San Jose section the steepest downhill segment is the 3,000 foot distance between University and Embarcadero Avenues where the railway descends 15 feet. The acceleration (a) reducing maximum braking rates along this section is: a = (g)sin[arc tan(rise/run) = (32.2 feet/sec^2)sin[arc tan(−15/3,000)] = − 0.16 feet/sec^2 where “g” is the acceleration due to gravity. (While a = g(rise/run) will produce essentially the same answer as a = (g)sin[arc tan(rise/run) at low angles the simplified formula for a 6% grade will produce a 0.18% error. Note: In this context “a” must never exceed g as is always true for the formula used here.)
    Conforming to rail transit norms maximum assured electrified Caltrain braking would appear to be 3.20 – 0.16 = 3.04 feet/sec. But a Transit Research Cooperative Research Results Digest on Improving Methods for Increasing Wheel/Rail Adhesion states that pine and cedar leaves combined with frost or drizzle will produce the worst rail braking conditions. Under these conditions attempting to achieve braking rates above 0.83 meters/sec^2 = 2.72 feet/sec^2 pose a rapidly increasing wheel slip risk. A conservative approach for setting worst case braking performance standards along Caltrain’s remarkably constant elevation profile would be to set its assured braking standard to below 2.72 – 0.16 = 2.56 feet/sec^2 or to a more convenient 2.50 feet/sec^2.
    (BART’s assured braking rate standard under the worst case track contamination conditions mentioned above is 2.20 feet/sec^2. BART’s out-door braking standard allows them to run down a 3.106% slope and still be within transit industry assured braking rate norms.)

  58. Using a “moving block” train position detection system and allowing the braking rate to equal the maximum credible braking rate under all realistic conditions of b = 2.50 feet/sec^2 even when track surfaces are subject to wet leaf contamination let’s more than double the moving block safety distance to 380 feet. The total minimum time separation, or headway (H) between 1320 foot trains moving at 200 kph = 182 feet/sec can be determined by the following calculation:
    The time it takes the train to travel its own length plus its minimum separation safety distance: T1 = (L + So)/Vo = (1320 feet + 380 feet)/182 feet/sec = 9.33 sec.
    The time it will take at maximum speed to travel the maximum assured braking rate distance where the maximum operating speed is Vo = 182 feet/sec:
    T2 = [∫VdV/b(Vo)] from Vo to 0 = (Vo^2)/2b(Vo) = Vo/2b = ½(182 feet/sec)/(2.5 feet/sec^2) = 36.45 sec.
    Referring again to the TCRP Report #13 on Rail Transit Capacity on page 77, Table 7.1 on transit industry norms the time lost due to braking jerk limitation is 0.5 sec. For exceptionally safe operation let headway (H) increase due reaction time plus delay due to smooth initiation of braking action be:
    T3 = 2 seconds.
    H = T1 + T2 + T3 = [(L + so)/Vo] + Vo/2b + T3
    H = (1320 + 380)/182 + ½(182/2.5) + 2 = 47.8 sec
    A four aspect 2215 = B foot long per block computer controlled train separation signal system with a 2.5 feet/sec^2 braking rate could support 60 second headways:
    S2 = [(Vo)^2]/2b = [182 feet/sec)^2]/[2[2.5 feet/sec)] = 6,644 feet = 3B ≈ 6645 feet
    T2 = 4B/Vo = 4(2215 feet)/182 feet/sec = 48.61 sec
    H = T1 + T2 + T3 = 9.33 + 48.61 + 2.00 = 59.94 sec.

  59. You've assumed that these very high braking rates could be achieved at speeds well in excess of "transit speed". You've also assumed that no trains ever stop at a station or get diverted to a different track at an interlocking. Both cases prevent headways from reaching a theoretical limit. Speed differences are even worse. Here's the thing: theory is great, but it's... Theoretical.

  60. Clem: Is 2.5 feet/sec^2 a high braking rate? Considering a passenger’s tolerance for high acceleration and braking rates perspective BART’s in-service maximum acceleration and braking rate standards are 4.4 feet/sec^2. BART’s in-tunnel safety braking standard is 2 miles per hour per second which is equal to 2.93 feet/sec^2. BART’s maximum acceleration rates can be easily verified by an observer positioned at the rear of a ten car train. When not approaching a close speed limiting track section, such as the Oakland Wye, the time required to travel from start to the platform end was observed to be consistently equal to 18 seconds whether the train was heavily loaded or near empty or on up or down grades out of the station. The time required to travel 700 feet from a standing start at a constant 4.4 feet/sec^2 acceleration rate is exactly 17.84 sec. Referring to the TCRP transit industry norm Table 7.1 says: “time lost to braking jerk limitation 0.5 sec”. These consistent acceleration observations verify the way electric machinery using a low impedance power source, such as the grid, is usually controlled. Fast acting high-loop-gain but stable negative feedback control circuits can completely eliminate performance variations due to voltage source or load changes. Especially since the mid 1990’s when variable frequency three phase traction motor drives became a transit industry standard these precise performance controls are easily implemented by controlling the pulse widths of the circuits generating the variable-frequency-three-phase traction motor driver currents.

  61. For above 80 mph performance one can refer to Bernard de Fontgalland’s 1980 book “The World Railway System” which says a passenger train’s resistance to motion at 140 kph as equivalent to climbing a 0.6% slope. (Note: The reduction in wind resistance per unit weight by recent streamlining efforts could be largely canceled out by lower train weights per unit train volume and mostly irrelevant because a great proportion of wind resistance stems from bogies extending close to the track-way surface.) Deducting 0.1% for rolling resistance in order to isolate the variable with speed wind resistance deceleration factor (a) opposing train motion the net wind plus rolling resistance deceleration at 200 kph would be:
    a = g{0.001 + sin[arc tan(0.005)(200/140)^2]} = 0.3605 feet/sec^2
    Sufficient wheel/rail adhesion required to sustain a 2.50 feet/sec^2 braking rate would need a surface adhesion driven thrust equal to:
    2.50 – 0.36 = 2.14 feet/sec^2
    In order to sustain safe operations, control maintenance costs, and provide a high level of riding comfort the high speed transit industry has made a strong effort maintain a smooth track structure and minimize the proportion of un-sprung train weight. Long-travel train suspension systems will keep train wheels at a nearly constant force against its supporting rail in spite of the roller-coaster effect of moderate track undulations at high speed. Therefore potential net braking rates may actually rise as speed increases within the speed range now being considered.
    But suppose track/wheel adhesion deteriorated to such an extent that dispatchers elected to reduce express train speeds. What effect would declining maximum speeds have on track headways?
    H = T1 + T2 + T3 = [(L + so)/V] + V/2b + 2
    The minimum headway speed can be found by taking the first derivative of H with respect to speed (V), setting the result equal to zero, and solving for V.
    dH/dt = − (L + so)/V^2 + 1/2b = 0
    V = Square root of 2b(L + so)
    V = square root of 2(2.5 feet/sec^2)(1320 + 380)feet = 92 feet/sec
    At this maximum through-put speed non-stop track headway is:
    H = [(1320 + 380)/92 + 92/2(2.5) + 2 = 18.44 + 18.44 + 2 = 38.9 seconds.
    Note: Following train switch delays can be avoided by scheduling as leading trains those whose first stop is beyond the following train’s first stop. Trains converging on one express track can avoid switch delays by designing converging track switches to operate like a spring frog in the unlikely event the converging points are not reset and locked before the following train arrives.

  62. One reason to minimize proposed excessive and largely redundant expenditures on Central Valley urban railway infrastructure is to have enough cash left of the soon to be available initial $6 billion in order to immediately construct a passenger rail route for a train to somewhere. We should keep in mind the mandates that nearly all the initial $6 billion must be used for a credible part of a future California High Speed System and at least half to be applied to the Central Valley. Seventy percent (70%) of the CHSR Authority’s planned 82.4 mile rail-right-of-way between Gilroy and a Chowchilla connection to the existing Amtrak route to Fresno is on the Central Valley side of Pacheco Pass.
    The Rail distance between San Francisco and Fresno on the CHSR Authority’s Mile Post Map is 191.5 miles. Compare this prospect with Amtrak’s current San Joaquin service between Fresno and Oakland which has a 205 mile rail distance scheduled to be traversed in 4:05. A diesel powered Amtrak San Joaquin Peninsula route service should be able to travel between the 4th & King San Francisco Station and Fresno in 2:50. The foregoing running time estimate assumes a San Joaquin Valley bound express train will require one hour to run between San Francisco and San Jose that includes intermediate stops at Millbrae, Hillsdale, Redwood City, Palo Alto, Mountain View, and Sunnyvale; 30 minutes for the 29.7 miles to Gilroy, and 55 minutes by averaging 90 mph along the 82.4 mile new rail right-of-way connection between Gilroy and the current Amtrak rail route near Chowchilla.( I am assuming an 8 kw/tonne diesel train climbing 800 feet over Pacheco pass will be delayed five minutes.) The remaining 31.4 mile distance along the present Amtrak route from the Chowchilla connection to Fresno should take 25 minutes at an average speed of 75 mph.
    An all-day and early-evening hourly San Joaquin express train frequency to major Caltrain stations should produce dramatically improved off-peak Caltrain Route service and may lead to a sharp increase in peninsula rail use. Reliable hourly service to stations near SFO and SJC Airports may prove to be a major San Joaquin Valley rail traffic generator. A positive response would help support a CHSR and Caltrain route change in order to serve a subway station directly beneath the SJC Airline Passenger Terminal. Rerouting the First street light rail and the Silicon Valley BART extension to the same SJC Airport Station would produce two rail route connections to San Jose’s Golden Triangle Industrial District and increase ridership on all four San Jose rail services. (Note: During the July-September 2011 Quarter BART’s SFO Terminal ridership has risen to 6337 passengers per day; 13% more than a year ago in spite of imposing a $4 surcharge on all SFO BART Station users.)
    There may be enough of the initial $6 billion left to improve the reliability and scheduled speed of Caltrain and CHSR express trains by extending Caltrain’s passing track infrastructure. A 6 mile grade separated four track segment between one mile north of Hayward Park past four stations to a point one mile south of San Carlos would improve service reliability. This center of Caltrain’s frequent local service segment would be particularly helpful in an effort to increase the scheduled time gaps between local and express trains. A local train delay would be far less likely to lengthen express train running times.

  63. Tolmach discredits himself with his #1 suggestion, which is stupid. He's also developed a record of dishonest misreading of studies, notably with his "Tejon is easy" bullshit.

    Sorry, not reading crap based on the statements of someone that unreliable. Don't discredit yourself by associating with him.


  64. Well, Tejon may be cheaper than the Tehachapis. I can't tell whether it's cheaper than the Tehachapis were thought to be before it turned out that the above-ground Soledad Canyon route was unavailable, but PB thinks it's cheaper than the Tehachapis are now.

  65. The #1 suggestion is one that I agree with. An idea can be separated from the person who advanced it, and must be judged on its own merits. The idea of Altamont + Tejon has great merit. The I-5 idea may also have merit seeing as money doesn't grow on trees. So enough with the Tolmach hate.

  66. Alon: Remind me why Soledad Canyon was eliminated...

  67. Joey: further environmental work showed that there's no above-ground alignment that minimizes wildlife impact. I leave it up to you to determine whether,

    a) This is an honest, unforeseeable complication,

    b) They foresaw the complication in 2008 but ignored it to keep the public cost estimate down,

    c) They could have foreseen the complication in 2008 but did not, or

    d) This complication is crap and they're just trying to drive up the cost of the Tehachapis in order to study Tejon.

    For the record, I'm somewhere between a and c. The powers that be don't want Tejon; Tolmach doesn't have the pull to force an honest reconsideration, let alone a dishonest one; and if the Tehachapis were known to be impossible in 2008, they'd have chosen Tejon and strategically misrepresented that.

  68. If the Altamont route is reconsidered and chosen by the CHSRA over Pacheco would it not make more sense for HSR trains from San Jose headed to LA to backtrack north along the Caltrain corridor to a Redwood City junction station, reverse directions and continue south to LA via the new bay crossing?

    I know it would take longer than a direct run from San Jose to Fremont but assuming the entire Caltrain line from SF to San Jose will be upgraded to CHSRA specifications (4 tracks, grade separated and electrified) I think it should be easily manageable in 15 to 20 minutes. Maybe only about 10 minutes longer than a direct San Jose to Fremont run.

    Caltrain could also offer through train service from both SF and San Jose to what would then replace the need for a separate Dumbarton crossing. It seems to me like that would maximize the utility of the infrastructure investment. Caltrain needs those 4 tracks even without HSR. I know it doesn't quite appease the nimby's in Palo Alto bringing the Altamont alignment up solely as a ruse to keep HSR and 4 track alignments out of Palo Alto, but could be considered a concession since the HSR traffic on the line thru Palo Alto would be reduced to only those trains serving San Jose. Palo Alto could even get an HSR stop if they ever had a "change of heart" on their anti-HSR position in the future. Caltrain would still benefit from the 4-track alignment where it needs it the most towards the middle of the line between Palo Alto and Redwood city. Cross platform connections between locals and expresses could be made at a new much larger Redwood City HSR/Caltrain junction station.

    Personally I got to admit as a south bay resident I am sort of biased in favor of the Pacheco alignment fearing an Altamont route would give San Jose and the south bay substandard or no HSR service. Anyways, by pure cold logic I can still acknowledge the benefits of an Altamont route.

  69. "If the Altamont route is reconsidered and chosen by the CHSRA over Pacheco would it not make more sense for HSR trains from San Jose headed to LA to backtrack north along the Caltrain corridor to a Redwood City junction station, reverse directions and continue south to LA via the new bay crossing?"

    No, for several reasons. But it's worth having thought about!

    * That's a really bad use of very expensive HS trains. Low speed, lots of time wasted, few passengers = poor use of capital resources. You can bet that any commercial HS operator would only provide this service under contractual political strong-arming, not because it makes economic sense.

    * That's no better than doing a HS to Caltrain cross-platform transfer (or in one cunning plan of mine, up a escalator transfer.) In fact the transfer might be faster, since reversing a train is a bit of a song and dance even in the best circumstances.

    * A major problem with Los Banos HSR is that HS trains "clog up" the entire line, not just as little as possible. Unclogging requires expensive extra infrastructure, especially more quadruple passing tracks, but in Redwood City it might even require entire extra platforms so that the stopping and reversing trains doesn't block everything else.

    So in effect instead of building a pair of new tracks from SJ to Fremont, you're building then from SJ to Redwood City, which is more expensive (construction on an "active" railroad is just insanely costly) and of no practical use (Caltrain doesn't need that track, so it's HS-only, so why not build HS-only track separately on the direct route?)

    * Lastly, since PBQD has decreed on all of our behalves that its BART is to be the rail connection between Fremont and SJ (with PBQD's puppets at SJ, Santa Clara County, VTA, BARTD and MTC dutifully acting as instructed and providing the slush funds to their transit-industrial complex overlords), it's really hard to see what non-fictional benefit HS trains directly to SJ Cahill station provide: unlike BART, the station isn't central to "downtown" SJ even, so why take people nowhere and then have them transfer to BART or VTA anyway? A BART transfer station between Fremont and Warm Springs is the best that can be salvaged from this disaster.

    (I've said a million times that if anybody in SJ had the intelligence of a slime mould, SJ to Livermore HS would be under construction today as the first segment of CHSR, and Santa Clara County's taxpayers would be several billion dollars richer and able to pay for all the nice things they promised but say we can''t have because BART takes everything.)

    In short: all downside, no upside. A transfer to BART )(for Milpitas and SJ) or to Caltrain (for Palo Alto, Mountain View, Sunnyvale) will be faster and equally convenient.

  70. A transfer to BART )(for Milpitas and SJ) or to Caltrain (for Palo Alto, Mountain View, Sunnyvale) will be faster and equally convenient.

    But but but everybody else says San Jose - California's third largest subur... city - it too important! Capitol of Silicon Valley, Navel of the Universe, etc. So much of a Very Important Place that BART is going to have a station at the Grand Central of West.... HSR will just have to go there!! .

  71. Yes, I guess it does seem kind of "out of character" for BART to serve San Jose. San Jose is no "Richmond", "Pittsburg","Pleasanton" or "Fremont". You got to be a city of that sort of scale and magnitude to deserve BART service. Sorry, I know this is kind of off-topic but I remember listening to a radio interview with one of the BART officials who mentioned that BART is the largest operator of parking west of the Mississippi. In some respects BART really is just a glorified parking lot shuttle. Silly for San Jose to think they could be as important of a "city" as Fremont!

  72. Anon: why stop at San Jose? If the point of BART is to operate parking lot shuttles to the biggest city around, then I say it should build a dedicated BART-only line to LA, in addition to HSR.

    I'm only half-joking, by the way. When you understand why BART to LA is insane, you'll see why BART to SJ is.