30 December 2008

Slow Traffic Keep... Left?

In all the existing plans for running high speed rail up the peninsula to San Francisco, the additional pair of tracks required for HSR traffic is invariably shown in the center of the right of way, with Caltrain local service running on the outside pair of tracks. The typical Caltrain station would still have two outside platforms, except there would now be four tracks through the station, like the existing Bayshore and Lawrence stations.

Intuitively, it's easy to accept this configuration without questioning it for even a second, because we are all familiar with freeways, where slower traffic keeps right and faster traffic passes on the left. (photo credit: mojoey) That paradigm is so deeply ingrained in our minds that we tend to think high speed trains should obviously operate the same way, right?

Not so fast-- with nearly every Caltrain station from San Jose to San Francisco about to be totally redesigned from a blank sheet of paper, it's worthwhile to question all assumptions, and in particular the assumption that HSR should run on the inside pair of tracks, i.e. slow-fast-fast-slow when enumerating the tracks from one side to the other. (Credit to Richard M for raising most of the following ideas in previous comments).

What if Caltrain local service ran on the center pair of tracks, with HSR and express trains on the outside tracks, i.e. fast-slow-slow-fast? What would be the pros and cons of each approach? For HSR? For Caltrain? For communities abutting the tracks?

Operational Flexibility

Anyone who commutes on Caltrain knows that periodically, an incident occurs that puts one track out of service for a few hours. While it is tempting to ascribe this to grade crossing accidents and equipment breakdowns, both of which would be alleviated by electrification and grade separation, the fact remains that service disruptions can and will happen on occasion.

When one track for local commuter trains is shut down, service is typically cut over to the other commuter track for a short stretch around the incident area.
Caltrain has the option of switching tracks at over a dozen crossovers, spaced every few miles along the peninsula. Trains can temporarily run the "wrong" way and make their usual station stops on the other platform track. To the extent possible, this minimizes delays and inconvenience to passengers.

With HSR in the mix, it gets more complicated. If HSR runs down the middle pair of tracks, cutting over local commuter trains from one platform track to the other platform track requires crossing over both HSR tracks and thus waiting for, or delaying, traffic on those center tracks. Temporarily running on the "wrong" platform track would involve a complex, coordinated sequence of moves that disrupt service on all four tracks. In addition, waiting passengers would have to dash to the opposite platform in order to catch their train.

If HSR ran on the outside pair of tracks and Caltrain commuter service on the inside pair, a disruption on one of the commuter tracks would not conflict with HSR service. To switch to the other platform track, locals would simply cross over to the adjacent commuter track. Under this scenario, Caltrain stations would have a single island platform in the middle of the right of way, located between the center pair of tracks. Passengers would not have to switch platforms to catch their train on the other platform track, since the platform tracks would serve each side of the same platform.

What happens if instead, one of the express tracks is disrupted? It's not nearly as bad: a train can simply be routed to the adjacent track. Unlike local service, platform access is not required for HSR and Caltrain express services, which would stop only at key stations with four (rather than two) platform tracks. Therefore, the requirement to switch to the other platform track does not apply; any track will do.

In short, the fast-slow-slow-fast track configuration provides great flexibility for dealing with service disruptions on any given track. On the other hand, the slow-fast-fast-slow configuration causes a big mess that disrupts all four tracks, whenever one of the local tracks is knocked out of service.

One other interesting fact is that in certain locations along the peninsula, Caltrain service could theoretically run in both directions on a single track without trains ever conflicting with each other. (Hello Atherton! Take note!) This naturally depends on the density and scheduling of Caltrain traffic, but it opens up the possibility of having only three tracks in certain locations, as dictated by the operational service pattern (Hello Atherton! Service pattern, not your back yards!) Switching to a three-track configuration from a four-track configuration is much simpler if you go from fast-slow-slow-fast (4) to fast-bidirectional-fast (3)... a simple turnout, and presto. On the other hand, trying to neck down from slow-fast-fast-slow (4) to any combination of 3 tracks invariably requires expensive flyovers to avoid frequent fouling of HSR traffic by local trains.

Operational Flexibility - advantage: fast-slow-slow-fast

Turnbacks (added 18 Jan)

The option of turning some commuter service back in the other direction before reaching either end of the line provides additional flexibility to tailor service patterns to passenger demand. This falls under the broad umbrella of operational flexibility, but merits a brief mention. For example, some Caltrain service could be turned back at Mountain View, heading back north where most of the ridership demand currently exists. Similarly, future Fremont service could be turned back at Redwood City. When a train is turned back, it needs to switch from one local track to the other local track. In a slow-fast-fast-slow configuration, this move requires fouling all four tracks in both directions. The fast-slow-slow-fast configuration, on the other hand, allows turnback tail tracks to be placed in the center of the right of way between the local tracks, with turnback moves having zero operational impact on any other track.

Turnbacks - advantage: fast-slow-slow-fast

Station Design

For the slow-fast-fast-slow track configuration, commuter (Caltrain local) stations would
have two outside platforms, like the existing Bayshore and Lawrence stations. For the fast-slow-slow-fast configuration, commuter stations would have a single island platform, like the existing Belmont station.

A single platform has several advantages for passengers: there is no choice of which platform to use, making station access less confusing. A single platform is more pleasant, because it is typically about 30 feet wide rather than 15 feet. It is safer, because there are fewer opportunities to be isolated from other people. It is cheaper to provide all the station amenities, such as benches, shelters, ticket vending machines, elevators, information signs, platform lighting, etc. since they are shared for both service directions.

Ease of Station Access - advantage fast-slow-slow-fast
Platform Comfort - advantage fast-slow-slow-fast
Platform Safety - advantage fast-slow-slow-fast
Cost of Station Furnishings - advantage fast-slow-slow-fast

Impact to Abutters

On stretches of track between stations, both track configurations require about the same amount of land. The fast-slow-slow-fast configuration puts high speed traffic 15 to 20 feet closer to abutters' back fences. Whether this makes a material difference to noise levels remains to be evaluated, since high speed trains are generally quieter than any other train type at speeds of just 125 mph.

The fast-slow-slow-fast configuration also requires more land for the approaches to commuter stations, because the outside fast tracks must slew aside to clear the center platform in the station. On tracks built for 150 mph operation, tracks can be slewed by the necessary amount (about 15 feet) in a run length of about 1200 feet before and after the platform itself. These dimensions require an extra four triangular strips of land, 15 x 1200 feet, at each of four corners of the station area; in total, just shy of 1 acre.

Noise Impact - no clear advantage
Land Impact - advantage: slow-fast-fast-slow

Passenger Comfort

In a slow-fast-fast-slow track configuration, HSR trains would run straight down the peninsula without care for Caltrain infrastructure. In the alternate scenario with fast-slow-slow-fast island platforms, HSR must jog around each and every one of about twenty Caltrain station platforms. However, comfort for HSR passengers would not be adversely impacted even with the somewhat frequent twisting and turning of the high speed tracks. The key to comfort in curves is managing the lateral acceleration and its time derivative (also known as "jerk"), and the dimensions mentioned above account for suitably benign track geometry.

HSR Passenger Comfort - no clear advantage

Freight Service

Union Pacific runs freight trains over Caltrain's peninsula tracks, to serve various industrial customers as far north as the port of San Francisco. There are roughly two round trips per night. Service to industrial branch lines could continue essentially unchanged under the slow-fast-fast-slow track configuration. If the local tracks were moved to the center for a fast-slow-slow-fast configuration, the occasional freight train would have to cross over the outside HSR track to reach the customer. Given the massive difference in axle loading and speed, it is unlikely that running freight trains over any HSR tracks would even be feasible.

Freight Service - advantage slow-fast-fast-slow

Impact to Caltrain Stations (added 02 Jan)

HSR on the peninsula will require the vast majority of Caltrain improvements made in the last decade to be demolished and rebuilt. Dozens of platforms, buildings, etc. will have to be moved to make way for four tracks. However, Caltrain has two stations, Lawrence and Bayshore, which could be used essentially as-is for slow-fast-fast-slow operations. Converting these two stations to fast-slow-slow-fast would require demolishing and rebuilding the platforms, modifying the pedestrian access, and realigning the tracks. On the other hand, the Belmont station is in the opposite situation: it could be used as-is for fast-slow-slow-fast operations, but would have to be entirely rebuilt for slow-fast-fast-slow. Whatever approach is chosen, there will be nearly total reconstruction of every Caltrain station, and the issues with Bayshore, Lawrence and Belmont will be lost in the noise.

Impact to Caltrain Stations - no clear advantage

Future Infill Stations (added 13 Jan)

Commuter stops that are added at a future date, as development fills in around the tracks, are termed infill stations. With a slow-fast-fast-slow configuration, they are easy to build by simply adding a pair of outside platforms and suitable pedestrian under/overpasses. With fast-slow-slow-fast, realignment of all four tracks is necessary to provide clearance for the new central island platform. Communities along the Caltrain right of way are already well developed, so future infill stations are unlikely to be numerous.

Future Infill Stations - advantage: slow-fast-fast-slow


To pull together all these pros and cons, it is useful to pause first and remember that Caltrain's most valuable asset is its land. As such, HSR ought to be considered a tenant on Caltrain's property, which means giving proper consideration to Caltrain's need to provide effective commuter service. That is why operational flexibility should be weighted heavily. Putting all the factors together in a table, and assigning weights, one can build a trade study to identify the optimal solution. The nice thing about a trade study table is that arguments about the best solution can be reduced to arguments about the factors, and the weights assigned to each.

FactorWeightScore (S-F-F-S)Score (F-S-S-F)
Operational Flexibility35%01
Ease of Station Access4%01
Platform Comfort4%01
Platform Safety4%01
Cost of Furnishings4%01
Noise Impact10%0.5
Land Impact15%10
HSR Passenger Comfort5%0.5
Freight Service10%10
Impact to Caltrain Stations5%0.50.5
Future Infill Stations4%1


Weighted Total


One can quibble with the weights assigned to each factor, but the conclusion seems pretty emphatic: the Caltrain tracks should be placed in the center, flanked by HSR on the outside. In short, Slow Traffic Keep Left!


  1. It wouldn't surprise me if there is a long term agreement with UP that will prevent the FSSF alignment from working at all; but I do agree that it is the best solution to the problem

  2. Another advantage of FSSF would be reduced relative speeds on the center tracks. That means they could be spaced a little closer together without any risk of windows blowing out on passing trains.

    Wrt UPRR traffic - it should be possible to construct the short sections involving turnouts for freight trains to support higher axle loads. In particular, they might be built on concrete slabs rather than conventional ballast to minimize the risk of geometry degradation over time.

    Safety then basically becomes a matter of signaling, given that FRA does permit mixed traffic if there is guaranteed time separation. With only two daily round trips, the impact on HSR service should be acceptable, especially if freight trains are fairly short and run at predictable times of day.

    If UPRR needs to run double-stacked flatbed cars in the peninsula corridor and an FSSF config is selected, all four catenaries must be installed high enough to avoid envelope problems. European and Japanese high speed trains are designed to run on tracks with low catenaries. I'm not sure exactly how large the difference is, perhaps the pantographs can easily adapt, at least at the speeds being considered for the Caltrain corridor.

    Finally, some thought should be given to how HSR and Caltrain will co-operate in the existing short tunnel sections in San Francisco, where only 2 tracks are available. Afaik, the DTX will have three tracks and the tunnel in Santa Clara will only be used for HSR. See CHSRA's Google Map for details of the proposed alignment implementation.

  3. Rafael,

    The closer spacing on the inside tracks is canceled out by the wider safety envelope on the outside tracks. It all works out to the same overall width.

    If UPRR installs ETCS Level 2 or whatever full-authority modern signaling system in their locomotives, I don't see why time separation would be required. It is only necessary when using antiquated human-in-the-safety-loop systems still in common use in the US.

    (I cringe when I see a Caltrain conductor, between idle banter with the assistant conductor, acknowledge signal aspects to the engineer... what the hell is the safety value of that ancient and hallowed practice, besides pissing away valuable fare revenue?!?)

    Did I forget any other factors?

    BTW another thanks to Richard Mlynarik for the good ideas behind this post.

  4. "Given the massive difference in axle loading and speed, it is unlikely that running freight trains over any HSR tracks would even be feasible."

    Well, if the turnout is at a *non-superelevated* -- i.e. dead straight -- and also dead flat -- location on the HSR track, it can be built as a high-durability diamond on a slab. The freight trains would cross the HSR track, never actually having the opportunity to run *on* the track.

    It can also be done with turnouts, but that's more difficult to maintain because it involves high axle weights on *moving parts*; I suspect that diamonds are the way to go for crossing the HSR track at grade with a freight train.

    Safety can be guaranteed by signalling (freight will probably only be running with time-separation from HSR anyway!) and then there's no *real* problem. But there is an FRA regulation problem.

    Now, let's think about the freight traffic integration.

    There are three major freight branches: at Redwood Junction, South San Francisco (a complex bundle of branches), and at Pier 90. There also appear to be a few more freight sidings in the Bayshore area.

    Notably, *all* of the existing freight branches and sidings appear to be on the east side of the mainline. I may have missed one, but I don't think so. (The Railroad Avenue branch looks disused.)

    Most of the branches and sidings are in the area from South San Francisco to Pier 90. A single freight-only track along the east side could serve this entire section -- which appears to be all the freight north of Redwood Junction. Then a single flyover or diveunder could link it into the slow tracks south of South San Francisco. This would be ideal (but see the tunnel problem below).

    Alternatively, FFSS would have serious advantages for freight access in this area. The disadvantage for Caltrain switching from local to express probably outweighs this, but it's worth thinking about if there isn't room for a dedicated freight running track. (But again see the tunnel problem below.)

    I see no problem with this layout from a passenger point of view: S. San Francisco and Bayshore don't actually need express platforms, while Bayshore, 22nd Street, and 4th and King all seem to have room for platforms on all four tracks (and probably even extra bypass tracks too). There would have to be a change to a different layout south of South San Francisco, though -- one flyover or diveunder.

    However, the constraining factor is the two-track tunnels in southern San Francisco, which must carry Caltrain express, local, and Baby Bullet, freight for Pier 90, and HSR. I really think these two-track tunnels are going to be a major, major chokepoint. This doesn't seem like a feasible permanent solution, but where to add additional tunnels? The choice of new tunnel location and design would *determine*, essentially by itself, whether SFFS, FSSF, or FFSS was the most suitable solution for the area.

    Perhaps these tunnels should be the focus of the next blog entry?

    Meanwhile, the Redwood Junction freight branch is at the same location as the Dumbarton branch.

    Access to the Dumbarton branch may have to be grade-separated *anyway* to avoid southbound Dumbarton passenger trains conflicting with all northbound traffic (and also with southbound expresses in SFFS layout). A flyover here could double as a flyover for freight access.

    A flat layout here could cause delays to all traffic -- at the least, an extra track would be needed for Dumbarton trains waiting to cross the northbound traffic (so as not to delay southbound traffic), and I don't think there's comfortable room for enough tracks here.

    FSSF would allow a flyover or diveunder to be constructed for the northbound fast track only (crossing all the turnouts). Southbound Dumbarton trains would have a clear path except when northbound expresses *and* locals were running simultaneously! This would also provide a clear path for southbound freight trains provided they weren't running at the same time as Dumbarton trains (or as northbound expresses *and* locals).

    SFFS would require a more complicated, curved flyover or diveunder from the southbound local to the Dumbarton line, which would not double as a freight route (southbound freights would have to wrong-rail or cross both fast lines).

    There is one more disadvantage to the FSSF arrangement for stations, however. With a side platform arrangement, it's possible to "cheat" and use nearby roadway underpasses or overpasses for handicapped access to the "far" side (no elevators needed). With island platform arrangements, you *have* to put in a new overpass or underpass.

    I think it won't be done for station reasons -- namely, that a number of stations already have four-track layouts, will not get HSR stops, and have side platforms. Specifically, Lawrence and Bayshore. These will not all need to be reconstructed, and avoiding their reconstruction will probably be the overriding consideration.

  5. Some good thoughts there, anon

    I will react immediately to your last statement, that SFFS may be adopted just because Bayshore and Lawrence are already built that way. Building new island platforms in those locations and shifting the tracks would cost on the order of $10-20M at each location. When we are considering an investment of $4.2B ($4200M) on the peninsula alone, it is NONSENSE to allow $10-20M expenses to swing the major architectural decisions for a $4200M project. That would be penny wise and pound foolish.

    If Lawrence and Bayshore are torn up, that's OK-- they are worth chump change in the grand scheme of this project.

    I will definitely address the tunnel issue in a future post.

  6. Another thought occurred to me. For SFFS, you would have to tear out the island platform in Belmont. This is a far more expensive piece of infrastructure than Lawrence or Bayshore, which are built on flat land and do not involve grade separated underpasses. FSSF and SFFS may actually be a wash, as far as demolition of recent Caltrain improvements.

  7. FSSF south of Redwood Jct.
    HSR NB (and SB?) dive under at Redwood Jct.
    FFSS north of Redwood Jct.
    Entirely new tunnel(s) for HSR leaving the old tunnel(s) for Caltrain etc.

    A diamond crossover is a good compromise. But does everyone who has or wants a freight spur that is not located along a freight siding get a diamond(s)? Freight needs customers. Is a customer charged for an additional diamond?

    There are heavy sand and gravel cars that shuttle back and forth regularly (weekly). Occasionally some of the old cars hammer on flat tires under the weight, beating up the nice new tracks installed a few years ago. Eventually the cars are serviced and the flat wheels replaced. It would help if the sand and gravel customers are along the FFSS so that they do not have to cross HSR even on the diamonds.

    I rode a Caltrain once that was developing a flat tire and asked the conductor who mentioned the service interval. I recall the trucks are serviced on the order of every few weeks. Seems like an emergency stop could cause a flat.

  8. "But does everyone who has or wants a freight spur that is not located along a freight siding get a diamond(s)? Freight needs customers. Is a customer charged for an additional diamond?"

    I think an assumption is being made that there will not be new freight customers outside the "freight zones" which already exist. This is probably accurate because outside those zones most of the areas are no longer industrial.

    With the customers concentrated in clusters, the idea is that multiple freight customers would share one crossing of the high-speed line (whether a diamond or a flyover).

    There are already long freight leads.
    However, they're parallel to the mainline around South San Francisco and Pier 90 (they're quite another direction at Redwood Junction). There are severe operational difficulties with shallow-angled diamonds, of course, which might mean turnouts if space is constrained.

    My instincts still prefer FFSS for the northern end of the alignment, as it solves all these problems quite simply. Indeed with the ROW width at Bayshore the existing station could be left intact and the high speed lines run to the west of it (entering a tunnel portal, of course). A new tunnel alignment could also be much straighter. But of course for any new tunnel there's the easement acquisitions....

  9. 1. The first point is that any talk about freight is the tail wagging the dog.

    It is totally insane to propose to screw up the daily operation of >140 Caltrain regionals and a few dozen HSR (no, there will never be 8 HSR per hour in this real world -- do the arithmetic!) in order to accommodate a couple of relic freight trains a day. Sure, freight on the rails is better than a few score extra trucks on the road -- I'm about the biggest road-hating fuel-efficiency-loving enviro-weenie on the planet --, but at what cost?

    Crazed, self-serving, self-perpetuating but otherwise irrelevant bodies such as the Port of San Francisco seem to believe they have legal or institutional standing to sabotage regional rail service (perhaps taking a clue the likewise crazed, etc, VTA that has "successfully" sabotaged rail service in the whole of Northern California for the next several decades through its BART/Pacheco fraud), but the questions any competent funding or planning authority ought to ask include: "where's the money?", "to what benefit?", "to whose benefit?" and "what are the comparative costs of alternatives?"

    1a. If freight trains are going to use our vital, limited, passenger-priority, publicly-owned infrastructure they will have to do so on our terms. If a train has a flat wheel it doesn't enter the right of way, just as it can't if it doesn't talk to the signalling system. If a train fails on our infrastructure and delays our trains it will be at a direct financial cost that persuades freight RRs to consider the comparative costs of deferred maintenance and the magnitude of their marginal revenue. These guys aren't going to be shunting on the mainline or derailing piles of lumber on the public's ten-billion train set -- certainly not more than once.

    Undesirable tenants with poor references, bad credit histories and a history of trashing their living quarters either play exactly by the the rules of the landlord or they get evicted.

    1b. If massive axleweight US freight trains do end up running on (ie screwing up) the Caltrain corridor, it is extremely unlikely that we would see sort of scenario imagined by a commenter above in which some subset of tracks are "rated" for behemoths while others would fail if a freight were mis-routed. That's just not the way civil structures (overpasses, bridges) shared by different tracks are built. (OK, it's possible to imagine a lower-axle-load flyover for a single track at a junction, or perhaps in one scenario to sort out the 100%-VTA-induced FRA clusterf*ck of ACE/Capitols/Dinosaurs between San Jose and Santa Clara, but I suspect this is irrelevant to the Caltrain corridor and way way way down in the construction cost noise anyway.)

    2. There's nothing that says that freights can't or shouldn't use the fast (outside) tracks, the easternmost in particular. (Elsewhere in the world they regularly or even generally do so, especially to avoid passing passenger platforms.) Given the minute amount of freight service, the off-off-off-peak scheduling to which it will be confined, and what it guaranteed to be a contractor-welfare-o-riffic massive overbuild of infrastructure between SJ and SF, there's no scheduling constraint to keep such trains from running on any particular track.

    Freight, should it run, will be neither fast nor frequent, both of which would contribute to rapid infrastructure deterioration and high maintenance costs. The question is whether accommodating even a low level of freight traffic can be economically justified.

    The fast and slow tracks on the Caltrain line are likely to be engineered for roughly similar running speeds anyway and, with a few exceptions which no doubt our caltrain-hsr blog host will eventually cover, the matter of wildly varying track superelevations for wildly varying traffic speeds doesn't seem an overriding factor in track choice.

    3. Diamond crossings with fixed crossings are simply illegal under a competent rail safety regime.

    German regulation for example forbids any fixed crossings at any speed above 230kmh, and very strongly discourages them >200kmh. Steep angle (1:9 or worse) diamonds are illegal above 100kmh, and shallower diamonds must have moving K crossings at least.

    Anyway, except where geometrical constrains forbid it, which is nowhere on Caltrain south of 16th Street in SF (and perhaps near SJ), plain turnouts are always preferred for reasons of flexibility, standardization, and reliability.

    I don't know where the commenter got the idea that this there would a significant cost savings in avoiding a couple point motors, but I assure you that the people behind local HSR construction lose that much change rolling out of their pockets and falling down the back of their living room couches every hour or two. You could just about pave the entire route from SF to SJ with moving point crossings and have no effect on the project's final budget.

    4. Redwood Junction must be grade separated, assuming that the plan isn't to kill all rail service in that corridor forever. (Which, of course, has been blatant VTA/SJ/SVLG/BART policy for over a decade now.) Any sort of flat junction at this point, near the mid-point of the Caltrain line where same-direction overtakes will be occurring in addition to whatever HSR traffic shows up, is an operational disaster.

    Note that there was an almost perfect solution to this issue with Altamont HSR, as the four-track section running south of San Mateo to Redwood Junction with fast tracks on the outside would peel off the outer tracks to the Dumbarton line (the southbound track burrowing under) and two central tracks -- which are all that would ever be needed for non-HSR service north of Santa Clara until the end of time -- continue south along Caltrain.

    Trying to do with with anything like four through tracks along the Caltrain line is going to be an enormous mess.
    Flying or burrowing junctions designed for any non-glacial speed are extremely long and much wider than many seem to believe (a track has to negotiate a wide S-bend and then pass under/over the other lines at a steep enough angle that the bridge/tunnel is constructible); meanwhile the space available at Redwood Junction (even assuming a complete and expensive rebuild of the Hwy 84 overpass) is extremely constructed.

    Good luck!

    If there had been anything like an honest or professional or ethical or legal alternatives analysis by CHSRA and its consultants this issue, and many like it, many of which we'll probably read about on this blog, would have come up and would have been given heavy weighting.

    I see a solution for the Mighty Port of Redwood City (a turnout from the easternmost Caltrain corridor track!), but I frankly don't understand how to make a working Redwood Junction with four through tracks to Atherton.

    We've been screwed over.

    5. Four tracks aren't needed north of Bayshore. Of course the sorts of people who get paid to build this sort of junk will of course proposed this sort of junk (and as we know CHSRA has explicitly communicated via its alternatives choice that cost is no object), but operationally it makes not the slightest sense.

    Train throughput on the Caltrain corridor is limited by turnback capacity in San Francisco (together with whatever points south of that trains terminate). The capacity of a pair of plain tracks (up to 20tph with non-exotic, real-world signalling) is far higher than any conceivable real-world train reversal capacity on the other side of that "bottleneck".

    Trains between Bayshore and SF will be travelling at nearly uniform speeds which means that the sorts of local/express speed differentials that eat track capacity will not be an issue.

    In the worst case it might be necessary to close 22nd Street station, but all indications (and the global precedent of competent rail operators) are that by competent scheduling and appropriate choice of rolling stock there is no reason that a subset of trains should not continue to serve this or a nearby stop.

    And no, a transbay rail crossing (a) isn't going to happen and (b) doesn't affect the calculus.

    6. The situation at Belmont is worse than I thought, at least for the sorts of impractical speeds which our blog host and the CHSRA's super-professional consultants are contemplating. While it is about the only remotely correctly-designed station on the entire Caltrain line, it is simply in the wrong place geometrically if trains are supposed to be running through the town at 200kmh or more. There's a big reverse S-bend that takes the existing Caltrain tracks away from the straight line of Old County Road between Belmont and San Carlos.

    Anyway, my money is on never seeing anything much over 160kmh anywhere north of Mountain View, and probably not very far north of San Jose's southern city limits.

    That's how things appear to work in the real world of shared urban/suburban corridors (and of course it's another reason why the corrupt Pacheco scam is a disaster that will keep coming back with more and more consequences).

    CHSRA has sold a bill of goods on the practically, politically, operationally achievable, sustainable speeds on the Caltrain corridor, as it has on numerous other issues. (Not that it matters, because the only important thing is that the cash spigots are turned on full blast.)

  10. @ Richard Mlynarik -

    excerpt from Caltrain's history page (emphasis added):

    "December 1991

    The Joint Powers Board purchased rail lines in San Mateo and Santa Clara counties. The JPB secured trackage rights to Gilroy for another $4 million, with an option to acquire half the right of way in the future. Union Pacific retains rights to operate freight service in the corridor.

    Right of way is purchased from Southern Pacific Transportation Company for $220 million."

    In other words, it appears the relationship between the Joint Powers Board and UPRR isn't one of landlord and tenant. Rather, UPRR - which merged with SP - retained an easement when the ROW was sold.

  11. It might be helpful to this discussion to have a good understanding of the Peninsula's remaining freight service.

    Union Pacific currently operates three freight trains per weekday, all based out of the yard next to the South San Francisco Caltrain station:

    1) SOUTH CITY SWITCHER: Goes on duty early in the morning: switches industries between South City and Pier 96 in San Francisco. Shippers include Granite Rock, Central Concrete and Pacific AgriProducts in SSF; Sierra Point Lumber near the Bayshore station; Dean's Refrigerated Trucking off Carroll Avenue in SF; Darling International, a renderig plant near Pier 96; and the Waste Solutions Group dirty-dirt concession at Pier 96. Famous for being the last freight train serving San Francisco.

    A sizable photo archive detailing this operation can be found here: http://www.macborja.smugmug.com/gallery/2793190_pEyTx#148982392_FrynX

    (Note to Richard: The Port of San Francisco has nothing to do with this operation aside from being Waste Solutions Group's landlord.)

    2) BROADWAY LOCAL: Goes on duty at 5:30 p.m.; switches industries between SSF and San Jose, such as the Port of Redwood City, after the evening rush hour ends. (One such industry is the Unilever plant in Sunnyvale, as seen here: http://www.railpictures.net/viewphoto.php?id=192892&nseq=0 )

    3) MISSION BAY HAULER: Goes on duty at 6:30 p.m.; gathers up all the outbound cars brought in by the other two locals and hauls them to the UP yard in Milpitas, then returns with the inbound cars for distribution by the locals. This train can easily see 60 to 90 cars.

    The point here is that even the Peninsula has some blue-collar industry, and they'll need to have a seat at this table. If there is no provision for freight trains, these businesses would have to either ship by truck or move away from SF and the Peninsula (and take the jobs with them).

    This sort of thing is not impossible. See the article titled "Local freight where it shouldn't be" in the July 2007 issue of Trains magazine.

    (That same issue also has an article about the South City Switcher, from which I drew a sizable portion of the information used above.)

    Of course, passenger trains will and should have priority, as is presently the case on the Peninsula, and the freights will have to work themselves in whenever they can. How long they last will depend entirely on their customers' needs.

    (slight subject change)

    Note to Clem: You're doing nice work here so far. As a native of Burlingame, I'll be interested to see a "Focus on ..." detailing that city. I think some people in town are still ticked off about losing the Broadway station on weekdays. I can't see HSR getting a warm reception.

    And I'm with Richard on this: the Pacheco alignment is a monumental mistake. Altamont was always the wiser choice in large part because of the large-and-growing customer base in the Central Valley, not to mention easier access to the state capital. (Full disclosure: I now reside in Stockton, which gets thoroughly hosed by the Pacheco alignment.)

    Just my $2.


  12. I'm not sure that FSSF is quite as great an idea as it's made out to be here. At least, it's not very common in four track lines around the world. From my experience, FFSS and SFFS tend to predominate by far. SFFS has one advantage that I don't think was mentioned here: you can use the express tracks bidirectionally for extra peak capacity, and for things like expresses overtaking each other. The American model for this kind of operation is Metro North's New Haven Line. Another advantage of side platforms is that station access is way easier, since you can just put an access from the street to the platform at any point at grade. On a grade separated ROW, side platforms allow for a narrower ROW profile on average, since you only have to provide enough width for stairs/elevators at the points where you actually have them, rather than along the whole length of the platform.

  13. @anon 21:31

    I submit that the New Haven line is probably not the ideal reference, for the following reasons:
    (a) its configuration evolved over a century ago, well before trains went faster than 100-125 mph, and it is not a clean-sheet design opportunity such as on the peninsula;
    (b) peak traffic is predominantly in one direction (towards NYC in the AM, away in the PM) which results in asymmetrical overtake requirements. This is unlike HSR or Caltrain peak traffic patterns, which are only mildly asymmetrical;
    (c) its speed differentials are not nearly as large as can be expected on the peninsula;
    (d) it's in the United States, hardly the best place to look for examples of how to implement mixed traffic with HSR. For best practices, it's better to start looking around Europe and Japan.

    I would suggest a more relevant example (recent, clean-sheet, European best practices, high speed differential) is the Swedish Arlanda Express, running between Stockholm and its airport. This is a 4-track mixed ROW that supports both local commuter service and 125 mph long distance and airport express trains. From a clean sheet, it was built FSSF with island platforms. If you are in any position to do something about it, it would be worth doing a literature search and contacting the engineers who designed it to discuss whatever lessons they learned. Just some basic due diligence.

    Regarding platform access: since everything is grade separated, under/overpasses are a given. Side platforms do afford additional access, but from one side of the tracks only, so passengers only benefit on one leg of their journey. The additional width used up by access facilities in an island platform is required anyway, since the platform needs to be ~30 feet wide to provide a comfortable passenger environment with sufficient stand-back distance from tracks. That being said, FSSF does require a bit more ROW as already noted.

    I'm interested in your thoughts about this.

  14. Are you talking about 30 foot wide platforms at side-platform stations? Nobody builds them that wide. Having done a quick google maps survey in the Northeast, looking at stations on various parts of the NEC that were re-built relatively recently, a 10-15 foot wide platform appears to be considered adequate for the typical local station. Oh, and if you're looking for examples of mixed traffic with HSR, Japan is definitely not the place to look, given that their HSR network is almost completely isolated from the mainlines, and in fact the northern and southern Shinkansen lines have no track connection between them either.

    Anyway, I think historically, things ended up as SFFS on many lines because the local tracks were originally sidings, which eventually extended and merged into a full four-track system. SFFS also makes the transition from two to four track easy and compact: the express track continues straight, and there's a switch to the local track, which can be designed for 30, 45, 60, or 80 mph. But with FSSF, the transition requires the express track to swing out away from the centerline, before having a switch to the local. And if you want to put a station in, it will probably have to be further away.

    Another advantage of SFFS, incidentally, is that infill stations are trivial to construct, since you don't need to do any track realignment at all. Just put up another platform, maybe even a temporary wooden one, and you're good to go. On the other hand, with FSSF and the line at ground level, it becomes easy to add express platforms to a local station, something quite difficult with SFFS, and which causes problems like Metropark.

    I think in the end, though, it really comes down to the particular conditions on a particular line, such as traffic patterns, speeds, curvature, existing infrastructure, and so on. Which is why we need competent engineers to deal with this problem.

  15. A couple more points that come to mind: Despite the fact that Caltrain's schedule has roughly the same number of inbound and outbound peak trains, there's still a significantly higher number of passengers going to SF in the morning and back in the afternoon. The situation is not too different from Metro North after all: there are plenty of reverse commuters there from NYC to Westchester and the closer parts of Connecticut. Thus, it's not unreasonable to expect that in the future, Caltrain will end up with more trains going in the conventional eak direction than the other way, which means it does make some sense to have sections with a reversible center track for express trains (you'd need high speed switches though). A single track for both directions of local trains, on the other hand, seems like a really bad idea to me, because it suddenly introduces a dependency between the inbound and outbound schedules, which means delays tend to propagate more between the two directions. I think it's realistic to assume a peak service of 6-8 tph express, and maybe 4-6 tph local, which makes all but the shortest single-track local section infeasible. It also, by the way, means that the northern terminal will have to accomodate 14 tph, which has implications for the design of the Transbay Terminal, but it probably doesn't make it necessary to quad-track the line between Bayshore and SF.

  16. "1. The first point is that any talk about freight is the tail wagging the dog."

    Yes, but

    (1) The freight is protected by *legal rights* held by Union Pacific, which cannot be seized by eminent domain, and which Union Pacific would undoubtedly charge four arms and five legs to give up, regardless of its actual economic value.

    (2) Freight rail is massively more environmentally sound -- and pleasant to be around -- than freight trucks. If the tradeoff for HSR was dumping a bunch of heavy gravel trucks on the road, all the environmental groups would drop their support, most of the locals would be very angry. It's politically a nonstarter.

    In fact, the environmental benefit of shifting freight to trains is generally quite a bit larger than the benefit of shifting passengers to trains. Because this is a *few* freight trains versus a *lot* of passenger trains, the balance might weigh in the other direction -- or it might not. I don't know.

    Either way, it would be politically impossible to drop the freight trains. Current freight service *must* be retained.

    (3) Current FRA regulations make it nontrival to deal with freight in the optimal way (just run it at night -- when there are very few passenger trains -- on the same old tracks, with plain old turnouts and crossovers).

  17. "I frankly don't understand how to make a working Redwood Junction with four through tracks to Atherton."

    Dive the east-side (northbound) express track under the entire junction. (Or over it.) Without intersecting anything. Have the turnouts for Dumbarton as a flat junction from the slow lines. This requires FSSF layout. Northbound locals from San Jose would have to divert to the fast track before the junction when a southbound Dumbarton train was coming.

    I see no other plausible scheme, but this one is actually relatively straightforward. No S-curves or anything -- just a rise and a fall to reach underpass/overpasss level on the express northbound track. This requires a lot of length (No express station for Redwood City, perhaps) but not much width.

  18. Added paragraph on infill stations per suggestion above.

  19. Is there any merit to considering reordering the lines part way down the peninsula, especially if any portions end up being double-decked?

  20. Having grown up on the peninsula, and a frequent rider back then on SP and then Caltrain, this is great stuff.

    I'll ask this likley naive question: Why not an FFSS arrangement?

  21. According to the Caltrain newsletters (or maybe it was their meeting minutes, I forget exactly), they have been trying to change all of Caltrain stations to no longer use a single center station for safety reasons -- the reasoning being that with the current set up it is fairly unsafe for passengers to wait in a center station when it is possible that a baby bullet train could pass the station from both directions at the same time at high speed. Even with a 30 ft platform, this still means that you are standing closer than 15 ft (especially if the platform is crowded) to a 70 mph train.

  22. Here's an example of FFSS on the Düren-Cologne line in Germany. The southern, slow tracks are for separated S-Bahn (commuter train) service. The northern, fast tracks are used by regional express trains and international high-speed trains.

  23. I'm not a big fan of the the center stations to the need for tracks to go around them. Even when riding on the newer bombardier cars, such 'kinks' in track can be felt and are more noticeable than turns. The most annoying ones are for Cal Ave station, Belmont and this random kink north of San Carlos. Looks like they're trying to dodge some signal. Was that just bad planning?

  24. Have FFSS with flyovers in places.

  25. That doesn't allow for cross-platform transfers between local and express trains traveling in the same direction.

  26. Besides, flyovers probably aren't cheap and they add to visual impacts. Plus the additional space requirements for ramps.

  27. Ease of Station Access:
    With slow-fast-fast-slow, at least the platform at the "right" side can be reached or departed from without staircases. And there may be able space for staircased or ramps.
    With fast-slow-slow-fast, a staircase is required for every axccess to a station.

  28. How important is a cross platform transfer between local and high speed service?
    In Tokyo Station transfers between adjacent lines involve going downstairs and up the next staircase. The Shinkansen is in another part of the building, and the Chuo Line is upstairs from everything. It seems to work. People transferring to a local line probably aren't critically concerned with a few extra seconds of time.
    I'm in favor of slow on one side and fast on the other.

    1. The problem is, SSFF is good only in case of 4-tracks-all-the-way cases. For lines with 2 tracks segments it results in a lot of conflicting train paths at the places, where 4 tracks turn in 2 and vice versa.

    2. There won't be a slow and a fast line; Caltrain needs to be integrated with HSR completely, despite what transit planners wish. The two systems will share one pair of tracks through San Mateo and the approach to Mission Bay.

    3. SSFF is also very awkward when there are industrial spurts on the FF side.The FF side will need crossings over each track, which is likely to cause a speed lrestrictoin. US freight cars are poorly maintained, with flat wheels, which won't do the crossings any good.

  29. Does it make any sense for the population of the West Bay to have HSR going from LA beyond San Jose? What we really need is a people mover from and to approximately Palo Alto in both directions.
    Light rail in other words every 10 minutes with the infrastructure build to be able to up this to 2 minutes per train.
    I estimate only 1/4 of people coming from LA actually really want to go to SF.
    So have HSP stop in SJ and improve the light rail to European standards. Which means we could save 10-30 Billion by going to a suspended mono rail like the vaunted Wuppertal Shwebebahn from 1898.
    Or some such system.

    Street level rail crossings still need to be eliminated.