08 August 2010

Alternatives Analysis Analysis, Part 3

The Supplemental Alternatives Analysis for the peninsula corridor reveals some key engineering choices being made. That tunnels would be taken off the table was never in doubt: tunnels are nothing but trouble. The only tunnels now left are those that have a mountain. Well, almost... as we'll see below.
  • Freight grades: we were told all along that 1% is the limiting gradient for heavy freight. Not so in Palo Alto, where there is a 2% grade shown in the track profiles (page 15). Recall that the steepness of a grade has no impact on passenger comfort, frequent references to roller-coasters notwithstanding. (This question of dynamics seems to confuse civil engineers, who deal mostly with statics.) If a freight train can handle a 2% grade in Palo Alto, it can certainly handle a 2% grade anywhere else; that's the concept of ruling grade. Considering that steeper grades would greatly reduce the footprint of any elevated section, for example, the massive Mary Avenue rail overpass in Sunnyvale shown on page 18, why are 2% grades not the rule?

  • Rail yard: the rail yard in Brisbane, the closest-available 100-acre parcel near San Francisco, is on a direct collision course with the redevelopment plans for this area. This may end up as the largest land transaction in the peninsula HSR project. Curiously, the yard is planned on the east side of the tracks, opposite the historical site of the Southern Pacific rail yard.

  • Berms be gone: Belmont and San Carlos get their grade separation berms removed and replaced by a gratuitously tall viaduct, with the tracks 20 feet higher than today's berms. That doesn't go over well with the mayor of Belmont. Prediction: when all's said and done, Belmont and San Carlos will be begging for their berms to stay essentially as they are.

  • Blank-check engineering: Millbrae gets some ridiculously massive civil works in order to shoe-horn a fourth track under the ill-configured station. The plan features a 2.5-mile long tunnel (see track profiles, page 8), diving 75 feet below grade in order to duck under the Hillcrest Blvd underpass. Cross sections are also provided (see pages 13 and 14). There are two design options that are likely to be far cheaper and less disruptive to construct:

    1. Convert the western-most BART track to Caltrain (see Option D in this diagram). While this requires minor surgery on the existing BART station, and may require an exception to side clearances (e.g. a steel crash barrier to separate BART from Caltrain), no digging would be required.

    2. Put the Millbrae BART platforms underground. The tunnel already exists, emerging right before the station.

    When the alternative is a 2.5-mile tunnel with an underground station mezzanine, which would likely require partial demolition of the existing station structure all in the name of leaving untouched the under-used BART parking tracks, one is left wishing that these issues had been thoroughly examined and addressed in the AA.

  • Slimmer sections: Many cross-sections have gone on a diet, narrowing down from the previous elephantine proportions to more realistic dimensions. The four-track retained-fill berm is slimmed down to 78 feet (wall to wall), the elevated viaduct is down to 78 feet, the trench is down to 76 feet (fence to fence). The at-grade alignment remains at 93 feet (fence to fence), but only because the fences are set back 11 feet from the (perfectly safe-to-touch) overhead catenary masts. The eminent domain panic set off back in March was premature: the tracks will mostly fit within 80 feet, a width already available along 88% of the corridor.

  • Fatter bridges: box-beam elevated structures have a depth of nearly 10 feet (measured from the bottom of the bridge deck to the top of the rails), compared to 5 feet as routinely practiced today, for example on the Belmont - San Carlos grade separations. That means future grade separations must lift the rails at least 25 feet above the road surface, compared to about 20 feet today, with attendant increases in the size of the bridge approaches. Five feet taller, especially when a sound wall will rise another 4 to 8 feet above rail level, makes a huge difference to visual impact.

  • PAMPA sandwich: The finalized application for the next federal funding increment includes a description of the project phasing plan for the peninsula. Phase 1: 4 tracks Redwood City and north + PTC. Phase 2: 4 tracks Mountain View and south + electrification. That leaves Palo Alto - Menlo Park - Atherton (PAMPA) with a two-track at-grade gap, left until such a time as there develops overwhelming pressure to fill it in.

  • VTA buried: The light rail spur in Mountain View is buried under the Stevens Creek, to make room for the expansion of the corridor. That doesn't strike one as the cheapest way to do things.

  • Viaduct vengeance: A FIVE MILE, 60-foot tall viaduct is planned in Santa Clara, an area of the corridor that is (a) already fully grade separated and (b) mostly wider than 100 feet. These people don't do subtle: their vocabulary is reinforced concrete, shoring walls and outrigger bents. This area of primarily commercial and industrial properties is less likely to oppose such a gratuitous structure. Operationally, the HSR viaduct needlessly prevents Caltrain overtakes anywhere between the Santa Clara and Lawrence stations. All pain, no gain... or from their point of view, all profit.
Clearly, there is a long way to go before context-sensitivity is achieved.


  1. The Westernmost platform and track @ Milbrae needs to go. There is simply no need for it. It would significantly reduce the cost for the tracks with 2.5 miles less of tunnel just because of the station.

  2. It's not just the grade but the length of that grade. A short 2% grade means that the train has less time to lose its speed, and if it's short enough, the whole train won't be on that grade all at once. A back of the envelope calculation shows that a 2% grade rising 25 feet is 1250 feet long, or about 25 freight cars. So only half the gravel train will be on it at once. And it seems to me (though I admit I know nothing about this subject) that the rollercoaster problem for freight trains would be worst where the grade goes from falling to rising without a level section, and worse yet when the train is passing over both a crest and trough simultaneously.

    The change to freight grades and ROW width just goes to show the process of progressive refinement of the design that happens as the project progresses toward construction. I'm sure we'll see quite a few more changes. As for Millbrae, just to be gratuitous, I'll suggest some dual-gauge track might be in order. Just to make things interesting.

  3. As a means of selling elevated structures to the communities, I find it curious that the HSRA doesn't seem to be contemplating any elevated designs with architectural character (subjectively speaking, of course), such as ones with arches and interesting facades. Germany's brick faced elevated 4-Track system into Berlin (2 tracks Stadtbahn1 commuter rail and 2 tracks ICE HSR) is often given as an example that best reflects our situation. I wonder how the cost of such structures would compare to the proposed trench or elevated viaduct alternatives? Somewhere in the middle? In terms of versatility of design, would the proposed elevated designs allow the space to be filled in with architecturally interesting features as a community saw fit or would it not be allowed? Would it even make sense to approach the design in a phased manner?

  4. Does Supplemental AA have commenting periods? I would hope CA4HSR mentions the few things such as gradients, Milbrae dilemna, and architechtually pleasing elevated designs. These need to be pointed out. Architectually pleasing solutions along with attempts to dampen noise along the tracks might win the support of those who are currently opposed. However, this could be their punishment for not inputing on a compromise solution.

  5. Those are some quite fascinating drawings by the way. I like how they have the locations of the curves down to a precision of 3 millimeters, but don't show the curve radius. There's also an interesting solution for the curve just north of the San Jose station, consisting of basically building one long gentle curve all the way from Taylor to Julian, which greatly increases the speed there and also leaves the CEMOF untouched, but does require demolishing a couple of houses and a bunch of industrial buildings. Of course if the upper deck at the Cahill Street station isn't built, then there won't really be any way for the HSR to bypass the station at high speed, so the curve north of the station won't constrain the speed any more than the station throat already will.

  6. Political_i
    For many, their input of a cut-and-cover type of trench was considered a compromise compared with a.) a tunnel or b.) not at all along the Caltrain corridor. Why do you think it appropriate to suggest those folks should be punished further with a plain Jane elevated structure?

  7. Clem, I'm not sure I understand what direction the HSRA is leaning with regards to cat pole configuration. Since they say an elevated structure would be 80' wide, I assume they are talking about S-1, an outside pole configuration, which is actually dimensioned at 78'6". As I recall, a center pole configuration is about 10' wider. When it comes to stations, the Supplemental AA shows a center pole configuration, STA-7.

    Can you tell me, is it easy to go from a center pole configured elevated station (STA-7)to an outside cat pole configured elevated straightaway (S-1)?

    If width is an issue, why are they considering a center pole configured station? I note, the engineers have reduced the STA-7 width from 139' (if memory serves me correctly) down to 124'4" in the AA Supplement. This is a step in the right direction, but in Burlingame, there is only 120' between the existing station and the curb at Carolan. A tight fit. I think the outside pole config for a station was only 112' in the AA.

  8. @ Ted - or you could decrease the station platforms from 32 feet wide per platform down to 20 feet and save approx. 24. That should get the Burlingame station footprint to fit where it currently is.

  9. Correction - the side platforms are now 26 feet instead of 32 from the first typical cross section. Regardless, you can reduce the platforms (assuming the elevated STA-7 drawing) to 20 feet or a little less and save 12 feet. That would reduce the footprint to 112 feet, which would fit.

  10. For many, their input of a cut-and-cover type of trench was considered a compromise compared with a.) a tunnel or b.) not at all along the Caltrain corridor.

    One would have to be clueless or in complete denial of economic realities to believe that there is zero chance of any at- or above-grade alignment. One cannot fail to provide any input on a likely alignment choice and then complain about the details chosen for that alignment (well, you can always complain, but no one will listen).

    That said, I still think that "architecturally pleasing solutions along with attempts to dampen noise" should be pursued. There is no point in punishing the silent majority with a noisy or ugly structure. Unless said silent majority does not actually exist (as our trackside friends would claim). In which case, sure, choose the cheapest structure possible, since the trackside guys claim that all at- or above-grade solutions are equally bad.

  11. "you can reduce the platforms ... to 20 feet or a little less and save 12 feet. That would reduce the footprint to 112 feet, which would fit."

    SFFS "Football Island" elevated with a generous standard 30 foot island platform is 99 feet ROW edge to ROW edge.
    (3'6" + 10'8" + 15'0" + 5'4"
    + 30'0"
    + 5'4" + 15'0" + 10'8" + 3'6"
    using their very own section dimensions.)

    Create your own objective conclusions about the skill and ethics of the Peninsula Rail Program's transportation professionals, for whom 99 is a subjectively larger number than 124, and who are quite willing to claim so in their professional Alternatives Analyses.

  12. what do you mean by 2 track at grade GAP in Pampa? - I don't understand the term 'gap'.

    There are two tracks through this section in the ARRA application, and if CHSRA is submitting to Feds presuming state matching via Prop 1A funds (which they are) - then that implies this QUALIFIES for prop 1A funds - which means it fits ALL the parameters of Prop 1A... Which means this 2 tracks works for HSR on all parameters. So CHSRA implies, its not a gap, its a solution.

    Which begs the question then, why build 4 tracks north of this, or south of this, anywhere between SJ-SF - because narrowing to two tracks for this 15 mile stretch just becomes a bottleneck - meaning 4 tracks surrounding it are a waste...

    It doesn't appear that CHSRA is considering this a gap - but a viable solution. But this viable solution makes the 4 tracks they are building in and out of PAMPA pretty much useless.

    so the real question is - how is two tracks a real solution for this PAMPA section, If its not, then how can they present it as such to the Feds. (Are they lying?) And if it IS, If it works there, why are 4 tracks needed anywhere between SF and SJ -they are just going to get negated by two track bottleneck through PAMPA, aren't they?

  13. @ arcady

    "Of course if the upper deck at the Cahill Street station isn't built, then there won't really be any way for the HSR to bypass the station at high speed, so the curve north of the station won't constrain the speed any more than the station throat already will."

    The design south of Diridon already constrains the speed so much that the design north of Diridon is kind of made irrelevant.

  14. HSRA has got to drop that costly and stupid and slow S-bend viaduct connecting Tamien with Diridon. The only sensible thing is to run at grade on the existing Caltrain alignment right through the SJ Gardner neighborhood. Sure, educate, mitigate and make it as painless as possible and practicable for the Garnder folks ... but that's the only sane and reasonable way to go. Why should Gardner get any more or better (from their NIMBY perspective) mitgation than, say, Burlingame or San Mateo or ...

  15. The plan is poor in several areas:

    * Millbrae - There's no need for a tunnel or elevated structure. A road overpass can be built for Central Street (or the crossing be realigned). At the Millbrae station, the platforms don't all have to be aligned at that spot. It is possible to keep the current northbound Caltrain platform and all other platforms to be relocated south of the overpass, where there's plenty of space.

    * Belmont/San Carlos - to do anything other than keeping it on the existing grade is silly. Elevated structure won't help make the city look any better (in case of Belmont, all you see would be the back of auto shops and warehouses). The construction impact is huge (probably worse than the original construction).

    * Atherton - there's no reason to elevate or depress the tracks at that town. That town only has two road crossings with limited traffic (2 lanes wide). It may be acceptable to close both crossings or just lower one if not both of those.

    Basically trains should stay on the existing grade as much as it can unless doing so would significantly improve traffic and access by grade separating a number of closely spaced crossings simultaneously.

  16. @ Reality Check

    I think that the displacement of the church is a good enough reason to not use the Caltrain alignment between Diridon and Tamien. I'm not aware of any similar community impacts on the Peninsula, so I'm not sure that they're similar enough.

  17. Adirondacker1280009 August, 2010 20:43

    I think that the displacement of the church is a good enough reason to not use the Caltrain alignment between Diridon and Tamien.

    If it's the church I think it is, it's well outside the ROW. It's a house that has been converted. The site is constrained. Make them a generous offer and they would be willing to move.

  18. Question for the peanut gallery: What good are any of these federal stimulus tracks?

    Start with a tunnel several miles long from a station that doesn't exist in San Francisco down to Bayshore? No use for anybody except a bat habitat or a mushroom farm until two or three billion from somewhere more goes up in smoke digging to Transbay Terminal.

    Four tracks Bayshore via an underground station at Millbrae with platforms incompatible with Caltrain, continuing with either no stations or no stations that can't be used by Caltrain to past Redwood City? Maybe guess they could run express trains if they put in lots of cro$$over$, but there's no way they can do Clem's super cross-platform express-local transfer, because the HSR tracks have HSR platforms and the Caltrain tracks have freight platforms.

    And isn't Caltrain broke anyway? How are they going to run trains on the new tracks or anywhere? The electrification report said their operating costs will go UP! Maybe they buy electric trains but they have no money to run them so they keep the diesels chugging along on the surface after all? Maybe use the long San Francisco tunnel to store the electric trains?

    Four tracks Sunnyvale to San Jose, which are either up in the air or underground, either way not connecting with the Caltrain-freight tracks, and terminating up in the air in San Jose at two platforms that Caltrain can't use? WTF?

    Maybe this, not Merced, is the HSR Test Track?

    Trains run from the end of a tunnel with no station in San Francisco to San Jose and back? Testing CBOSS for five or ten years until it works?

    I don't get it.

    Is there there a plan?

    There must be a plan.




  19. South of the San Jose station, the curvature does not justify Caltrain's current 35 mph speed limit. Something like 50-60 should be possible for HSR. But in that case, the station throat becomes the limiting factor and I imagine speeds there being limited to 35 or so, which is still better than the current 10 mph or so. And yes, that's pretty low, so a tighter turn on the north side would be acceptable. But their plans don't show turn radii so I can't tell you what speed they're proposing there.

  20. So we now have confirmation that the Authority was engaged in "strategic mis-representation" in stating there was no funds available for tunnels on the Peninsula.

    In fact, the Plan does pay for several miles of tunnels on the Peninsula. They are just in the wrong locations where they are of no benefit.

    Hopefully, this makes the job of 'nimbys' much easier. Just take the ~1 billion that would be spent on Millbrae tunnel + Bayshore tunnel and re-program to more useful location.

  21. Anonymous, Regarding why 4 tracks north of RWC. You have to wonder if it is because the HSRA has concerns about the Altamont Pass lawsuit? Until that is sorted out, maybe they don't want to look at 4 tracks south of RWC, but don't tell the FRA.

  22. That SJ Gardner area church would be the Word of Faith Church on the corner of Delmas and Fuller ... and it's no reason not to use that route vs. the insane, costly "iconic" and train-slowing S-bend viaduct. Buy 'em out ... or shift the existing tracks as far north as possible and shoehorn the other two right up against the church. It's only a church for Chrissake! A little junky looking one at that. Not like it was an important or valuable building or anything like that.

  23. Think about what the cost of relocating the church would be.

    Now think about what the cost of that bridge over the freeway interchange will be.

    Now make your decision.

  24. Wow, way to screw up the Millbrae station more than it already is.

    The viaduct between Redwood City and Belmont seems is completely gratuitous.

    Not even going to comment on the massive viaduct (or the equally ridiculous deep bored tunnel) between Santa Clara and San Jose. Is this really the best they could do?

  25. Did anyone else notice how the elevated "east" alignment into Diridon in the track profile link that Clem has at the top shows that they are finally no longer restricting themselves to Caltrain's relatively tight turns north of Diridon? I'm not sure how much time that would save but it has to be considerable. Anyone have any idea what that turn radius might be?

  26. Using some basic geometry, I worked out that the curve radius of the "eastern" alignment is 640 meters approaching Diridon from the north. That would allow speeds of up to 120 km/h (according to wikipedia) approaching Diridon. The western alignment is only approximately 260 meters. How much time would the larger radius save?

  27. No interest? Oh well, I guess I should have given it in feet and mph.

  28. Peter, I know the answer. But it's not rewarding to do the work to try communicate it here.

    PS 600m radius both north and south of a single level shared SJ Cahill Street works very nicely indeed, and has been completely laid out -- but not by Caltrain/CHSRA.

  29. If we can get same height boarding for Caltrain and HSR, they should scrap the multilevel station at San Jose. Waste of money and plenty of platforms for both. This way the ROW should continue as it is today, not that ridiculous bridge and hairpin turns.

  30. @ Richard

    Any guidance as to where I could find out the answer?

  31. Ok, assuming 160 mm cant and 100 mm cant deficiency, the speed through a 260 m curve would be 77 km/h.

    After roughly measuring out the distance, it appears that the 120 km/h curve would save about 20 seconds over the 77 km/h curve, not including accelerating and decelerating.

    That nearly makes up for the San Bruno curve debacle. And yes, it would be preferable to have a single-level station with 600 m curves, so don't bite my head off.

  32. I just save money on my car insurance14 August, 2010 10:08

    Next month FRA will approve modified Stadler GTW DMU for shared freight-passenger tracks with no restrictions:


  33. I'm guessing that one of the reasons for burying VTA Light Rail in Mountain View is so that VTA can complete double-tracking to Mountain View. Without this VTA can't complete its Long-T configuration where they have express trains running between Mountain View and Alum Rock. I don't think they would be able to fit six tracks including platforms between Mountain View Station and Whisman otherwise.

  34. VTA light rail? Express? Ha ha ha!!

    I work near Highway 237, and the best thing that VTA can so is to blow up the stupid light rail that is impossibly slow and instead run some buses from the station.

    No "express" light rail won't change anything. Even without stopping at stupid stations that nobody will ever use, the line is so indirect and so slow that you have to be crazy or desperate to try it. Try it. Once.

    Hopeless hopeless. Don't throw good money after bad. There's no way it can ever work. Look at a map!!

    It would be crazy if the Mountain View Caltrain and HSR station were all screwed up just to make the light rail "work" — when nobody will ever use the stupid light rail because it is slow slow and runs around in circles instead of going anywhere. Look at a map!! No hope. Blow it up.

    The only way the light rail should be underground is when it is dead and buried.

  35. In portions of Sunnyvale, Santa Clara and San Jose the ROW looks wide enough to build 6 tracks at grade if they wanted to (with limited need for eminent domain actions). I have not looked at the plans close enough but would think that would be the ideal scenario (particularly wherever 4 tracks are already provided). Anyone know if that was being considered for any stretches of the SF to SJ line?

    Ignoring of course those areas where increasing to 4 tracks has become a big issue. The 4 tracks would of course remain dedicated to Caltrain to maintain existing operations and capacity with the extra 2 going to HSR. Is there much precedent in other systems around the world for an HSR line and high frequency commute line to share just 4 tracks? I am all in favor of HSR but don't want it to cripple Caltrain's express and limited stop operations either.

    I am guessing the purpose of the current 4 track sections is to allow trains to follow each other more closely in time but not for passing except during breakdowns. I've never noticed a limited or "baby bullet" pass a local. At least not between Redwood City and San Jose.

  36. @ Anonymous 16:19

    Yes, VTA light rail is slow. That doesn't mean that it should simply be shut down. There are a number of things they can do to improve its performance, and a number of them they can implement without any changes to the infrastructure.

    The Short-T, for example, can be implemented without any changes to the infrastructure, as can express trains along Guadalupe, and 20 mph in the Downtown Mall. Those would go a long way towards improving service by a lot, without any additional capital investments.

    @ Brian

    Lawrence, for example, is a station where Caltrain has overtakes. My wife often has to wait on her south-bound local for an express train to overtake her train.

  37. In the current Caltrain timetable, there 1 overtake per hour in each direction during peak hours at Lawrence. Bayshore is almost the same, it's just missing the morning reverse-peak overtakes

  38. Did anyone else notice that the trenches for Menlo Park are currently planned to be around 80 feet deep? That's pretty crazy.

  39. More trouble with civils design.

    Perhaps the authority needs to get a better consultant on civil engineering.

    Arup makes nice designs.