Heads Up: Neighborhood meeting on HSR for residents of the Willow Glen neighborhood in San Jose, on Wednesday, April 8 at 1292 Minnesota Ave from 7-9pm. Resources at the end of this post.
The fair city of San Francisco has probably had more than its fair share of coverage on this blog. In just the past couple of months, we've already had quite a few posts recently on the controversial new Transbay Terminal Center (see posts 1, 2, 3). In addition, Clem Tillier has published a post on the Focus On: SF Transbay Transit Center on his Caltrain-HSR Compatibility Blog.
So why revisit this issue yet again? Well, because Quentin Kopp is once again saying that 4th & King would be good enough for him - knowing full well that this will unleash howls of protest from SF city officials and residents, in no small measure because CHSRA consistently marketed the starter line as going from downtown San Francisco to San Jose Diridon, LA Union Station and Anaheim ARTIC.
Note, however, one flaw in the snippet from the Examiner: supporting a five-minute headway on the HSR line down the peninsula is not quite the same thing as actually running a train literally every five minutes. It just means that the signaling and other technology has to cope with an HSR train following another within the space of five minutes, something that may well happen at certain times in a timetable supporting multiple service classes (e.g. express, semi-express, semi-local, local). Recovery from unexpected delays and other off-design conditions may also require operators to minimize headways at certain times. Safe operations of sections of a line with moderate speed limits are possible with as little as 2.5-3 minute headways, so 5 is actually conservative.
The number of platforms is a potential issue a number of decades down the road, maybe. On the other hand, the poor design of the connection between the DTX tunnel and the platforms, the so-called "throat", is a real issue right now because Caltrain will also use the tunnel and downtown station. It's a bottleneck even an expensive three-track tunnel cannot fully resolve. Moreover, according to Clem, the current design of the DTX tunnel features curve radii so small that they effectively preclude the use of Japanese shinkansen train designs for the California system.
The current design features 6 long platform tracks, some slightly curved, accessed via three-track tunnel out to Caltrain's existing 4th & King terminus station. The six tracks would be accessible via three island platforms, one for Caltrain and the other two for HSR. The split is partly a result of Caltrain's decision to stick with low platforms even at the Transbay Terminal, where no freight train will ever go.
The reason for the ongoing brouhaha over this one station is really quite simple. CHSRA has raised a very late red flag on the design of the DTX tunnel and train box, claiming it does not support the 12 HSR trains per hour (tph) each way that he thinks it should support, in addition to the 10tph Caltrain is hoping to run during weekday rush hour in 2025. I've argued that (a) any sane operator would anyhow choose to terminate some HSR trains further south long before reaching 12tph and, (b) that four platform tracks is anyhow enough for 12tph if you enforce some pedestrian flow control.
Tellingly, the red flag came - or rather, was shown in public - only after the Transbay Joint Powers Authority (TJPA) decided to apply directly for a slice of the HSR funds in the stimulus bill, rather than route that request through CHSRA. TJPA had previously been under the impression that 4tph would be sufficient for HSR operations. While that's true for the early years, perhaps even for several decades, CHSRA thinks its job is to deliver an infrastructure with enough capacity for the next 50, even 100 years. There is some method to that madness, as expanding a constrained downtown station decades after initial construction is usually extremely difficult/expensive or downright impossible.
While the timing suggested a political motivation, there does appear to be a real problem after all: the so-called throat, in which the six platform tracks have to converge to the smaller number in the tunnel, is very tight and by all accounts poorly designed. It constrains throughput for no good reason, Clem included a much improved version designed by Richard Mlynarik in his post.
Still, with a minimum curve radius of just 150m (~500ft), even Richard's improved version would effectively preclude the purchase of proven, off-the-shelf Japanese high speed trains - those need at least 280m, preferably more. SF real estate development should not drive HSR vendor decisions with statewide implications. To my mind, this is a red flag well worth raising - though CHSRA could and should have done so much sooner and publicly if TJPA was dismissive of this particular issue.
Like any building, the aesthetics of the Transbay Terminal have their fans and their detractors. You can't argue about taste. What matters more to transportation engineers is functionality, which the current design might meet, sort of, for both CHSRA and Caltrain, but only after some modifications.
I've thought about this some more since my last post on it and have come to the following conclusion: the train portion of the design is suboptimal because planners interpreted the intent of SF voters (prop H, 1999) very narrowly as bringing the trains into the building itself. However, nobody except the developers really cares about that. IMHO, the voter intent was that the trains should stop within walking distance of the financial center, preferably also close to the bus terminal and such that an option of a second transbay tube to downtown Oakland is created.
If you subscribe to this more liberal interpretation, it is no longer essential to run tracks into the basement of the Transbay Terminal building. Rather, the focus can shift to a design that better meets the requirements of modern railway operations:
- large platform number (more than six)
- no platforms dedicated to any one operator
- all platforms full length (400m, i.e. 1/4 mile)
- all platforms straight
- all platforms level boarding (see Caveats below)
- high pedestrian flow capacity between trains and city streets
- minimum curve radius 280m (vendor flexibility, screeching noise)
- minimal cross-blocking of trains entering and exiting the terminus station
- all platform tracks run-through if second transbay tube is ever constructed
Fortunately, it's actually possible to meet these requirements in San Francisco, though it requires divorcing the design of the Transbay Terminal building from that of the railway station. The two would be linked via city streets, possibly a short underground pedestrian passage. If that's good enough for linking to BART and SF Muni subway, why the absolute requirement to put standard gauge trains into the basement? The upshot is that the Transbay Terminal should go ahead as planned, but without a train box or concourse level.
The following map and topology diagram show the concept I've come up with: a Central Station for San Francisco underneath Mission Street. Detailed design and construction would be far from trivial: many SF city buses use Mission to reach downtown, so tearing open that street would be quite disruptive during construction, even if it is done one block at a time. In addition, the street is quite narrow, around 85 feet including both sidewalks. That's only enough room for three tracks, two side platforms and narrow escalators/flights of stairs - which implies a lot of those would be needed to secure sufficient pedestrian flow capacity. Finally, there is almost certainly plenty of plumbing, ancient sewer mains, power mains etc. lurking underneath street level.
View Larger Map
The first thing you'll notice is that the DTX tunnel runs down 7th and Mission Streets, a very different route from the one currently proposed. I'm advocating just two tunnel tracks side-by-side, shared by non-compliant HSR and equally non-compliant Caltrain EMU trains, protected by appropriate signaling. Any legacy FRA-compliant trains need to terminate at 4th & King anyhow, since they're diesel-powered. Nevertheless, track sharing is only possible if FRA approves both Caltrain's request to switch to non-compliant EMUs and CHSRA's intent to buy proven, off-the-shelf but non-compliant bullet trains capable of cruising at 220mph.
One fringe benefit of the changed tunnel route is that no platform tracks are lost at 4th & King. Only a single 90 degree turn is required, in an area where real estate is probably less expensive than in Rincon Hill or downtown. In combination with the less complex excavation methods required for two single-track curved tunnels, this ought to permit a sufficiently generous curve radius. The map shows the maximum feasible radius of ~325m, there are various length constraints along Mission Street. A more detailed design would seek to optimize the radius as tunneling under existing buildings is expensive. Note that I have assumed the curve would be level; a helical curve would permit gradients of less than 3.5% in the vertical elevation change sections at either end.
My point is that the radius can be made large enough even for off-the-shelf Japanese high speed trains. This also eliminates any screeching noises from the long-wheelbase trucks required for stability at high speed. Crucially, it also increases station throat throughput because speed limits are proportional to the square root of curve radius. With appropriate tracks superelevation, it should be possible for outbound trains (non-tilt types) to take this corner at speeds as high as 60 mph, assuming the curve radii and speed limits for the various points along the way also have appropriate values.
The design objective should be to permit a train to pull away from any platform at maximum acceleration (~1.1 m/s^2 for Caltrain locals) and maintain that until it either hits the speed limit or has cleared the station's outer throat, i.e. the point at which tracks from the two levels converge. The minimum available distance for acceleration is about 1/2 mile, though trains on the lower platform face a 3.5% uphill climb for half that distance. Either way, any given Caltrain leaving an east platform should be able to clear the inner throat near 3rd Street within 75 seconds and the outer throat within 120. Maximum acceleration for HSR trains is more like 0.6-0.8 m/s^2 so they may take a little bit longer. Note that an incoming train may proceed past the outer throat as far as the inner one while a train is emerging from the same level; this provides a 15 second buffer.
Conclusion: as long as the curve in the access tunnel can be taken at fairly high speed, the elongated nature of the station layout does not appear to present a throughput bottleneck since the maximum time required to clear the outer throat is still less than the absolute minimum time separation imposed by the two-track access tunnel, i.e. 150 seconds at 24tph for HSR and Caltrain combined. Platform dwell times of at least 16 minutes are possible even at this insanely high level of traffic, which SF will almost certainly not attract in the next 50 years - if ever.
The downside is that the Mission Creek outfall needs to be crossed at grade, so a short section of 7th Street would lose two traffic lanes. More significantly, Townsend Street would be permanently severed at 7th unless an overpass were constructed.
Throat And Platform Layout
The second thing you'll notice is that the approach to the station is both long and straight, ideal conditions for designing a throat for minimal interference. In this case, the inbound and outbound tracks need to split into two levels that I call -2 and -3 for reasons that will be come obvious shortly. This split begins immediately east of the curved section. By the time the tracks reach Jessie St West, they need to be stacked 2x2 on top of each other. This ensures that any inbound trains waiting to reach a platform can wait west of 3rd Street on the correct level, without blocking inbound trains on the other level.
Near 3rd Street, the two tracks on each level first combine into a single track and then immediately split via a three-way point into a through track and sidings left and right of it. These sidings are the west platform tracks (2-1, 2-2 & 3-1, 3-2), each with its side platform. A second three-way point near Shaw Alley reconnects these sidings to the through track. East of this point, the through track may descend for a block to compensate for changes in the surface level (at-grade does not mean constant elevation above MSL). Regardless of length, trains always stop on the west platforms such that one end is at Shaw Alley.
A third three-way point at 1st Street provides access to two additional platform tracks east of that location (2-3, 2-4 & 3-3, 3-4). Regardless of length, trains always stop on the east platforms such that one end is at 1st Street. Space should be reserved for a future fourth three-way point reconnecting the east sidings to the through track, for reasons that will become obvious shortly.
Since there is no room for an access platform, the through track between these east sidings can only be used for train storage. Perhaps it could also be used as a tail track for the western platforms if cleaning/provisioning staff board and alight there. However, the only emergency escape route would be at Spear Street, which implies walking the length of the train first.
The layout described above means trains don't block each others' movements any more than necessary. The longest wait states would occur for an inbound train waiting for another to clear one of the east platforms, since that needs to travel up to half a mile first. At an average speed of 20mph along the center track, that would take 90 seconds. The worst-case wait state for a western platform would about a minute. In practice, well-planned operations would minimize wait states or eliminate them altogether by slowing inbound trains down rather than forcing them to come to a full stop. Worst case, there is room for one full-length train between Jessie Street West and 3rd Street.
Operationally, any given train would always stop such that one end ends up at 1st Street, within half a block of the Transbay Terminal building. In addition to supporting regional buses, the building would also contain the customer service counters for train operators, baggage depot, police station, shops, cafes, lounges/waiting areas, bicycle parking etc. Taxi and city bus stops are located on the plaza between Minna and Mission, Fremont and 1st. In other words, the concept of a multimodal hub is preserved even though the tracks are not underneath the building.
The train station would be really a bare bones affair: just the trains and access infrastructure plus some ticket vending machines and restrooms at intersections with cross streets. Mission Street isn't wide enough to accommodate any other facilities on the platform levels.
Strictly speaking, there is no need for a formal concourse level at -1. Since the design uses side platforms throughput, surface streets could be used to access all of them. However, given the anticipated number of passengers, it is safer to provide pedestrian underpasses across Mission Street, even though that means digging a deeper (i.e. more expensive) trench than would otherwise be necessary. Moreover, the desire to provide four full length platforms per level means long walks are inevitable for some passengers on full-length trains. The concourse level would therefore feature moving walkways in both directions along its entire length, interrupted at the eight locations at which there are exits to the surface on both sides of Mission Street.
To either side of these central moving walkways would be regular ones, about 10 feet wide each. To either side of those would be the baggage screening areas, if Homeland Security requires any. Airport-style check-in is possible but usually (a) conductor(s) on board the train will inspect passengers' tickets. Note that security procedures are pointless unless they are implemented at all HSR stations. In any case, terrorists haven't attacked a long-distance train in Europe since Carlos the Jackal did in 1983. Instead, they've targeted subways and commuter trains as well as high speed tracks. Train stations are also at risk.
Optionally, staff could prevent passengers waiting to board from descending until they get a signal from colleagues below that all arriving passengers have cleared the platform. Unless the fire marshal requires this at all times, it would normally only be done during peak periods to avoid excessive congestion on the platforms. Note that the platforms on the two train levels would feature multiple stairs/escalators to the concourse level. Escalators intended for the disabled, women with baby strollers etc. would connect all three levels. Connections directly to the surface are ok only if Homeland Security decides it doesn't need to enforce access control to the platforms.
Pedestrian Cross Passages
A pedestrian passage underneath 1St Street between the train station and the Transbay Terminal building would not be absolutely essential but nevertheless very useful, given the vehicle traffic on the surface. Such a passage is currently planned as an optional extra between the building and Market Street, but under Fremont Street. Note that in the station design discussed in this post, the building would not need a concourse level for transportation purposes, just stairs/escalators/elevators to the pedestrian passage, if any. Moving walkways along this 200 foot stretch would be convenient for passengers and increase throughput capacity.
Additional pedestrian passages under Main and 2nd Street could connect the train station and indirectly, the Transbay Terminal building to both Embarcadero and Montgomery Street BART/SF Muni. Again, not strictly necessary but useful.
Note that all these underground pedestrian passages would provide shelter against the elements. Unfortunately, they may also attract pickpockets, buskers, alcoholics, drug addicts, graffiti artists, homeless people and others just hanging out for no apparent reason. Bright lighting, ventilation, cleaning and security patrols are needed to ensure passengers will feel safe and comfortable enough to use these facilities. It may make sense to ensure the passages are private rather than public property so security has a legal basis (trespassing) for evicting disruptive individuals.
Phasing And Extension to East Bay
What I've described above is actually the fully built out terminus station. If desired, construction could be divided into two phases to accommodate budget constraints. Phase One would be the access tunnel via 7th an Mission plus the four west platforms, possibly the pedestrian cross passages to the the Transbay Terminal building and Montgomery Street BART as well. Phase Two would be the east platforms beyond 1St Street plus the pedestrian passage to Embarcadero BART. The west platforms could already be in operation at that point.
If and when a decision is made to build a second transbay tube, the Mission Street station design affords a fairly straightforward if expensive connection via Point Alameda and into downtown Oakland up Franklin. On the SF side, the tracks should already be deep enough to avoid conflicts with the descending SF Muni subway tracks under the Embarcadero. There is enough room for the through track at level -2 to descend to level -3 between Spear Street and the water's edge. However, unlike the west throat, there is no need to double-track this new one on the east end: instead, the four platforms on the upper level would be used for inbound and the ones on the lower level for outbound trains (or vice versa). This operations change turns all eight platform tracks into run-through types, thanks to the four three-way points installed on each level.
In addition, the presence of a through track in each direction would permit trains to bypass all of the platform tracks. That could be useful for e.g. high speed cargo trains if a transshipment terminal were built in the East Bay or, for trains supporting a major sports or other event on Point Alameda or in Oakland. Normally, however, trains would dwell for a considerable amount of time in SF. Even at 24tph for HSR and Caltrain locals combined, having four platforms each way available means dwell times of close to 20 minutes would be feasible.
I've already touched on a number of caveats, such as the consequences for 7th and Townsend Streets and, the need to pass under SF Muni tracks at the Embarcadero. In addition to those, there is a potential conflict with the planned SF Muni Central Subway. That is supposed to dive under the BART tracks at level -3 and, it looks like the chosen alignment (alternative 3B, Fig 2.12 would be deep enough at Mission Street to avoid a conflict with the alternate heavy rail access tunnel described above. However, if heavy rail station under Mission Street is considered, the clearance issue at 4th Street ought to be double-checked before ground is broken on the Central Subway. I don't have a good sense of how the surface grade changes between 4th and Embarcadero along Mission Street. The heavy rail tunnels should be level from 3rd to Shaw Alley and from 1st to Spear Street yet still deep enough to permit a concourse level plus a future extension to Oakland.
Another caveat is that passengers need to know not just the platform but also the car number of their train to descend at the appropriate point. The preferred approach is mandatory seat reservations (cp. TGV in France), the receipt can then include this information. Given the width constraints, pedestrian traffic along the platforms should be minimized. Reservations also ensure there are no standees on the train and that no-one needs to walk far inside the train to reach their seat.
Also on the receipt could be helpful hints, e.g. the nearest BART/SF Muni subway/SF Muni streetcar. Indeed, a good reservation system would ask about your connecting transportation to offer you a seat that minimizes your transfer distance. That could be printed on the receipt, along with an indication of how many minutes to budget for an comfortable transfer incl. security screening, if any. The flip side of seat reservations is that a late change to the platform, especially from one west of 1st Street to one east of it (or vice versa), would cause significant inconvenience for passengers.
Finally, I've assumed throughout that all eight platforms in the design would be created equal, i.e. that all would feature level boarding at the international standard of 3'6" (1067mm). Caltrain is hamstrung by a ridiculous 1948 rule intended to protect freight railroad workers hanging off the sides of cars. Is it really appropriate to perpetuate ye olde tyme practices to the detriment of passenger convenience and throughput capacity, in a station that no freight train will ever enter? IMHO, if Caltrain wants access to any downtown station in SF, it needs to buy EMU equipment that can cope with both level boarding and the lower level (2'1" = 639mm) that the platforms at its other stations are built to. If UPRR and CPUC have no objection, those could be upgraded to level boarding as well over time.
Heads Up: Neighborhood meeting on HSR for residents of the Willow Glen neighborhood in San Jose, on Wednesday, April 8 at 1292 Minnesota Ave from 7-9pm.
The purpose of the project-level EIR/EIS phase is to nail down how the HSR alignment should be constructed in each location, after close consultation with the affected communities. The starting point will be what CHSRA consultants suggested for the purposes of cost estimation prior to the election. This information has been available online since 2007, though you do have to look for it. The Authority is a planning body, i.e. a bureaucratic organization - most of its documents were designed for hardcopy rather than the web.
The objective of the HSR project is to deliver a net gain for the state California, especially for the cities served by stations. The following resources may help you understand what CHSRA has done to date regarding San Jose and the Willow Glen area in particular.
- CHSRA Google Map of implementation details used for cost estimation purposes.
- CHSRA Pacheco Pass implementation details such as height of structures, embankments as used for cost estimation purposes. Willow Glen is on page 4.
- CHSRA Station Fact Sheets, San Jose Diridon is on pp41
- CHSRA Various Cross Section Drawings, San Jose Diridon is on p20
- CHSRA Speed Limit Ranges for the route (NOTE: part of draft EIR/EIS documentation, no longer present in the final version. Possibly out of date, as Altamont Pass and East Bay sections have since been scoped out of phase I and phase II of the HSR project.)
- CHSRA Capital Cost Estimates for the Bay Area to Central Valley section (line item details)
- CHSRA Bay Area to Central Valley Final Program EIR/EIS
- UPRR owns the existing rail right of way between San Jose Diridon and Gilroy, Caltrain and Amtrak have trackage rights on UPRR's tracks. HSR needs land for two additional dedicated tracks, preferably adjacent to the existing ones. See our posts on How Important Is UPRR To California HSR? and Union Pacific Speaks for more information.
- How others do it: first pass analysis of alignment options and local impacts by San Mateo city planning staff, prepared for the city council