Geary metro, purpose-built for SF
Part 2: East Bay connection, downtown alignment, and the yard
In Part 1, we discussed that the Geary metro is a project with strong potential, for which funding for an alternatives analysis should be identified right away. The Geary metro should feature many stations, and should rely on high frequency to provide capacity and limit station costs.
Part 2 covers the Geary metro’s connection to the East Bay, its alignment in Downtown San Francisco, and the location for its maintenance and storage yard.
The SFCTA’s initial study of the Geary metro does not make any final determinations, but suggests that using BART technology for both Geary and the second transbay rail tunnel would increase ridership and avoid the need to locate a train yard within San Francisco.
Instead, future study should conceptualize the Geary metro as a purpose-built line within San Francisco, which will allow it and the future transbay tunnel to be better optimized for their respective operations and therefore to better serve more riders. The alternatives analysis should focus on identifying a downtown alignment which provides excellent transfers to longer-distance service: BART at Market Street and regional rail at Transbay station. A yard in San Francisco will be needed for the Geary metro regardless of its train technology, and there are good candidate sites for one.
Table of contents
How to connect to the East Bay
Purpose-building the Geary metro as a line within San Francisco with a high-quality connection to the future transbay regional rail tunnel will allow both projects to best serve riders.
The route the Geary metro takes downtown should be a primary focus of further study. For the best connections to BART, Muni Metro, and Caltrain, and the best station access from the surface, planners should strongly consider a shallow, cut-and-cover crossing of the Market Street subway.
San Francisco cannot avoid the need to site a yard for the Geary metro within its borders by using BART technology. Luckily, wide-open golf courses are located at the ends of the potential west side alignments, at the most operationally appropriate place for a yard.
How to connect to the East Bay
SFCTA’s study compared ridership along the Geary/19th subway for a San Francisco-only project and for a project that includes the future transbay tunnel. The study found that a line within San Francisco only “would serve up to 182,000 daily riders,” and that continuing to the East Bay would add 76,000 (on regional rail) or 131,000 (on BART) riders along the Geary portion of the combined service (pages 40–41).
However, the initial study did not examine the scenario that should be the focus of the next phase: a dedicated San Francisco metro that provides a high-quality transfer to BART and to the future transbay regional rail tunnel. It makes sense that SFCTA found that adding a new transbay rail tunnel into the analysis adds ridership, but trains from the Geary metro don’t have to physically continue into that tunnel to serve riders on it.
Geary has the ridership to justify a one-seat ride to the East Bay. However, taking advantage of the fact that transit riders can change between vehicles will allow Geary to be designed as the best-possible local metro and the transbay tunnel to be designed as the best-possible branching regional rail system. This will result in lower costs and higher benefits for both projects.
The advantages for a purpose-built San Francisco metro line that does not have to cross the Bay include:
- A purpose-built metro can have shorter trains, and therefore more, smaller, less-expensive stations.
- A purpose-built metro will be capable of handling tighter curves and steeper grades, enabling better station locations, better transfers to BART and Muni Metro, and reduced cost and property impacts.
- A purpose-built metro can run more frequently at peak — 40 trains per hour — and off-peak, with automation decoupling service frequency from the operating budget. The future transbay tunnel will likely max out at 24 trains per hour, much of which will go down the Caltrain corridor.
- A purpose-built metro can be procured as a standard, turnkey, all-in-one system with the latest technology from one of the major manufacturers, likely Hitachi, Alstom, Siemens, or CAF. There will be no integration difficulties, customizations, or incompatibilities with existing train control systems or new platform screen doors.
The advantages for a second transbay rail tunnel that does not have to continue down Geary Blvd include:
- Maintaining the good choice to build the second transbay crossing as an electrified regional rail tunnel will provide frequent service to San Francisco along new East Bay corridors, and avoid further complicating BART operations.
- Sending trains from the tunnel down the existing Caltrain corridor will avoid the construction of a huge, expensive wye below the San Francisco waterfront, which would be required to split transbay trains between Geary and today’s Caltrain corridor.
- Sending transbay trains down the Caltrain corridor will provide frequent, rapid trains to Southeast San Francisco, for metro-level service from Bayshore, Bayview, Potrero, and Townsend to downtown.
- Every train in the transbay tunnel will be able to be the full 240-meter (800-foot) regional length, maximizing passenger capacity through the tunnel.
Geary metro riders will already be making transfers, with buses and with BART and Muni Metro at the Market Street subway. These transfers, and transfers to regional rail trains to the East Bay, will be especially easy with an ultra-frequent Geary metro, which will come so often that there will be no need to bother rushing to catch a train.
The rest of the Bay Area and the state should recognize — and should communicate to San Francisco — that Geary is an important line to build regardless of whether trains from another county physically run on it.
The alternatives analysis for the Geary metro should focus on how to provide excellent connections with BART and regional rail on alignments within Downtown San Francisco, and should dismiss the idea of crossing the Bay.
The downtown alignment
The alignment in Downtown San Francisco will be one of the most crucial decisions for the cost and usefulness of the new line. The Geary metro should:
- Serve downtown.
- Provide a convenient transfer to through-running regional rail and high speed rail at the Transbay Terminal.
- Provide a convenient transfer to BART and Muni Metro.
- Ideally, provide a convenient transfer to the Central Subway.
SFCTA’s preliminary study suggests one option for a BART alignment and one option for a regional rail alignment. The technology won’t be BART — the future second transbay rail tunnel is now planned to be and should be regional rail. For the regional rail alignment, SFCTA notes that “…the required depth and tunnel turn radii would likely require a routing via the Civic Center station area before turning back north towards the Western Addition and Japantown.” Having to cross deep under the Market Street subway in a gradual, swooping curve would result in a poor transfer to BART for regional rail.
For the more flexible, purpose-built metro technology, three potential downtown alignments stand out. These are not the only possibilities. Exploration of downtown alignments for a standalone metro, including refining and evaluating these three, should be a focus of the next phase of study.
Montgomery stub
Providing a convenient transfer to BART and Muni Metro is particularly challenging while crossing the deep Market Street subway. The Central Subway has been criticized for the walk required to connect between its Union Square/Market Street station and the existing Powell BART and Muni station, which is an engineering consequence of this deep crossing.
An elegant solution for creating a good connection with the Market Street subway, then, is to not cross Market Street at all. The Geary metro could terminate along Post at Market, with an island platform level with the Montgomery station mezzanine. This would provide the most direct access possible to Muni Metro and BART, without requiring passengers to travel up-and-over between platforms as they must to reach the Central Subway at Powell station.
Although this alignment has an excellent connection with BART and Muni Metro, its connections with the Central Subway and with Transbay station are less than ideal. Although Transbay has limited draw today, it will become a major node in the transit network once Caltrain Downtown Extension (DTX; The Portal) brings rail service there. It will become even more central once the second transbay rail tunnel reaches it.
The Montgomery stub alignment has another issue, which is that Montgomery BART is already overcrowded as it is. Pre-pandemic, BART studied a major capital project to remedy platform capacity limitations at Embarcadero and Montgomery stations by digging outer “saddlebag” platforms. Downtown ridership has not been as high since the pandemic, and it is possible that BART can get by without ever building these platforms by installing new stairs and platform screen doors. However, if a Geary metro made Montgomery station its BART connection, the station’s current capacity would not be adequate.
The 2016 study estimated that one Montgomery side platform would cost $175 million, but that was low. Ballpark $1–2 billion for the station, and a year of single-tracking BART service. This cost may not be prohibitive for the alignment, but it requires serious analysis and collaboration with BART.
Deep crossing at Powell
The standard method to cross the Market Street subway would be in deep tunnels, as the Central Subway boring machines did in 2014. The natural place to do so would be along Geary Blvd itself, as SFCTA’s study shows for its BART alignment.
Rumor has it that the piles for the Central Subway’s Union Square/Market Street station were held short where two tunnel boring machines along Geary could pass through. I have not been able to verify this in public documents, but renderings suggest it is possible:
If the Geary tunnels were able to pass over the Central Subway before diving under BART at Market, passenger access time from the surface would not be too bad. The Union Square/Market Street mezzanine also serves as a ready-made pedestrian tunnel between a station along Geary and BART/Muni at Powell. However, the transfer would still require a two-block walk between Geary’s Union Square station and Powell.
A deeper station closer to Market Street may be possible, such as along Ellis Street. This and the station along Geary should be explored in the next phase of study.
Shallow crossings
If we assume that:
- Crossing the Market Street subway is necessary to provide a good connection to Transbay station,
- The Geary station at Market Street should be shallow to provide a good connection to BART and Muni, and
- Avoiding the expense and disruption of “saddlebag” platforms at Montgomery is desirable,
then an alternative to explore is crossing the Market Street subway at a shallow level, instead of deep, at Powell or Civic Center station.
There are many options for a shallow crossing, which should be fully explored in the alternatives analysis.
The option along Leavenworth, 7th, and Howard is attractive. It can provide an ideal transfer to BART and Muni Metro off the end of the Civic Center platforms, as shown in the watercolor rendering above. It can also provide a direct connection between the Geary terminal platform and the train mezzanine at Transbay, as shown in the “How to connect to the East Bay” watercolor. Lastly, it would provide a good mezzanine-to-mezzanine transfer with the Central Subway, while acting as a one-stop shuttle from San Francisco’s central high-speed rail station to each doorstep of the convention center.
In addition to convenient transfers, shallow tunneling would provide rapid access from the street to the stations, and cost-free addition of multiple entrances at each station. However, it would require longer sections of utility adjustment and temporary street impact than would deep tunnels with only the stations built cut-and-cover. The next phase of study should define and present the advantages and impacts of the shallow, stub-end, and deep tunnel alignments, so that we can agree on the best overall.
Get InfrastructureWeak’s stories in your inbox
Join Medium for free to get updates from this writer.
The yard location
The trains that run on the Geary metro will need to be stored and maintained at a yard.
SFCTA’s initial study commits to consider a new yard for the Geary metro within San Francisco, but shows a preference for relying on an existing BART yard outside the city’s borders.
However, San Francisco will not be able to avoid providing train storage for the Geary metro adjacent to the new line, and using an existing or expanded BART maintenance shop would be costly, controversial, and require significant tradeoffs.
The standout option is to build a new yard on an existing golf course. This will require political will, but will improve Geary metro operations and enable it to use the latest world-standard metro technology, saving substantial construction cost.
The space required
With 90-second headways and two trains in cleaning or on live standby at the end of the line, the number of trains in active service for various Geary metro alignments would be as follows:
- via Howard to Lands End, 20-minute runtime: 29 trains
- via Howard to Daly City, 31-minute runtime: 44 trains
Adding a 15% spare ratio for trains out of service for maintenance brings the total number of trains requiring storage and maintenance to 34 or 51, depending on the alignment.
To reduce the space needed for storage tracks in a yard, the Geary metro could be designed for trains to park along its mainline service tracks. For example, if a train were parked at both platforms of each station overnight, and cleaned by the overnight station cleaning crew, that would store the majority of the line’s trains. Or, trains could be parked end-to-end between crossovers on one of the two tracks, to preserve overnight maintenance access via the other track.
This practice, known as “mainline stabling,” is uncommon but not rare. It is typically implemented only on automated lines, so that travel from the parking space to the yard does not have to be arranged for train operators. As of the latest public information, Vancouver’s automated Skytrain system was storing 33 trains on its main line overnight:
Even with mainline stabling of a portion of the train fleet overnight, there will still need to be a place to maintain the trains and to store spare or peak-service-only trains.
The BART or Caltrain yard idea
It would be possible to operate the Geary metro with shorter (perhaps 4-car or 90 meter) BART or BART-like trains at high frequency, and to send these trains to BART’s shop in Hayward only for required maintenance. However, this would not solve storage of Geary metro trains.
BART’s Colma yard, near the Daly City connection with the 19th Ave alignment for Geary, is fully used to stage trains for its morning and peak service. In fact, BART has identified that its existing yards, including Colma and Daly City, will already need to be upgraded for the higher service frequency that will be enabled by its current train control upgrade project (BART Metro: 2030 and Beyond briefing book, Page 27). If the Geary metro is built compatible with BART so that it can use its facilities, it will still have to have its own substantial tracks to store the trains it does not need off-peak, positioned close enough to the line to roll them in and out at the necessary times — a yard.
Connecting the Geary metro with the BART tracks could avoid the need to build a maintenance shop within San Francisco, which would save some of the cost and size of a yard. However, to do this the Geary line and its trains would have to be built to the BART standard, which is unique compared with rail lines around the world. This customization and reduction in competition between major manufacturers would certainly raise the line’s cost more than the savings of an avoided maintenance shop.
BART’s maintenance capacity at its maintenance facility, in Hayward, is also limited. For the Valley Transportation Authority (VTA)’s San Jose extension, BART has required the construction of a new maintenance shop and yard at the end of that extension, to accommodate its additional track length and trains. BART has resisted VTA’s attempts to delete or defer this maintenance facility. It is likely that the project budget for a BART-compatible Geary metro would still have to at minimum contribute to building additional maintenance capacity at the Hayward yard, if room for that could even be found.
Taking Geary metro trains to the Hayward yard would also either take up track capacity slots from regular BART service during the day, or would take up time from BART’s overnight maintenance window. VTA’s Diridon BART station is closer to the Hayward yard than Downtown San Francisco (37km/23mi vs 42km/26mi).
Otherwise, the Geary metro could have an out-of-service connection with Caltrain at Transbay station, to use a yard along that line. That would require the Geary metro to be built with 25kV power, which requires greater electrical clearance than 750–1500V metros do — and accordingly larger tunnels. Caltrain’s small maintenance facility and yards are already fully utilized. However, across San Francisco’s south border, the Brisbane Baylands are slated for a high-speed rail light maintenance facility and yard. Adding space and facilities to this yard for trains from the Geary metro would certainly not be to the liking of the City of Brisbane. This idea is worse than the BART one.
It makes sense that San Francisco is attracted to the possibility of siting the storage and maintenance of trains for a new metro line outside of the city itself. However, it will still need storage tracks adjacent to the Geary metro, the design and operational constraints required to perform maintenance somewhere else along a BART or Caltrain line would be significant, and the expansion or construction of new facilities outside the city’s borders will still cost money and cause controversy in whatever city land is sought. The next phase of study should therefore quickly get around to eliminating this idea, and explore yard options within San Francisco.
Expand a Muni Metro yard
With aggressive off-peak and overnight storage of Geary metro trains along the line itself, it may be possible to perform maintenance and store spare trains at an existing Muni Metro light rail yard. This could be accomplished with an out-of-service track connection, such as to the westbound Montgomery station track at a stub-end Geary station along Post.
It would be possible, but not ideal, to operate the Geary metro with an additional order of Muni’s current Siemens S200/LRV4 light rail vehicles. A 90-meter, four-car Muni Metro train loses substantial space to cabs and couplers. It also has features not needed for the Geary metro, such as sliding stairs and sand dispensers to climb 10% grades. However, with Muni’s new train control system and platform screen doors, driverless operation with the LRV4s on Geary should be possible, for good all-day frequency.
Custom trains compatible with Muni Metro, but designed new for the Geary metro, could be produced, either for Geary’s opening day or in a future upgrade once capacity demanded them. This decision would still impose some constraints on the new line’s design. Muni Metro vehicles are a reasonable width, although a purpose-built metro could go wider for more capacity. Muni Metro’s 600VDC overhead wire is less than the typical 750 or 1500V, so there would need to be a few more traction power stations along the new line. The trains would need to be compatible with the signaling system along the surface light rail track, and may need to comply with other requirements for light rail vehicles, such as turn signals and headlights. The turning characteristics would likely be the biggest constraint — the route to the yard would need to be mapped and the new trains designed with articulation to attempt to handle the existing curves without losing passenger space to couplings between cars.
Of Muni’s two light rail yards, Muni Metro East is the one with potential to expand. It is 6km (4 miles) along the tracks from the Geary metro’s out-of-service track connection at Post and Market. Relying on this location would make daily rollout of peak-service trains more difficult than from a yard along the Geary metro itself.
I do not believe that the design constraints and operational complexity of investing into expanding Muni Metro East would be worth any savings from avoiding similar yard construction closer to the Geary metro. However, this is a better idea than using a BART yard.
A tight yard at Masonic
One of the few larger sites along the Geary metro near downtown is Muni’s Presidio bus yard, and the Target and Target parking garage to its south, at Geary and Masonic Ave.
A total reconstruction is planned for the Presidio yard, and it would be possible to dig down further to add a level for train storage. Housing and retail could be redeveloped on top of two levels of train storage and maintenance shops on the Target site.
However, there is no ignoring that this site is constrained, awkward, expensive, and urban compared to a typical yard. It is also in the middle of the full Geary metro line, which would add operational complexity: trains would either need to “deadhead” from the end of the line back to the yard without passengers, or would need to announce that they were going out of service without reaching the end of the line, making the rider experience more confusing.
This Masonic site may be necessary to use if we build only a “Phase 1” Geary metro project that is limited to the dense, high-ridership area nearest to downtown. This is more reason to manage costs through pragmatic design, and to invest adequately, so that we do not have to phase the project.
Half of a golf course
San Francisco has nine golf courses, and two of them are located right at the ends of the likely west side alignments for the Geary metro (more on the west side alignments in Part 3). The end of the line is exactly where we would want a yard for the most efficient train operation, as it avoids trains deadheading or going out of service partway through their run.
Any of these golf courses could be restored on top of a yard built underground, if desired, for additional expense. Underground train yards are most often built below dense and valuable urban development, as Caltrain plans for its 4th and King yard. However, there are numerous existing underground yards for modern metros below parkland, including Osaka’s Tsurumi and Tsurumi-ryokuchi-kita depots, Tokyo’s Ōjima and Kiba, Singapore’s Kim Chuan, and Brussels’s Erasmus.
Along a 19th Avenue alignment, the San Francisco Golf Club is ideally situated: flat and adjacent to Junipero Serra Blvd, which the metro should loop along to reach Daly City BART. The San Francisco Golf Club is an exclusively private, old-money institution, which could mean behind-the-scenes barriers to making use of its land. An alternative site is the Lake Merced Golf Club to the south, which is also privately-owned. The Lake Merced club still laments its 1960s alteration for the construction of the I-280 freeway.
Along a Richmond alignment, placing a yard at the City-owned Lincoln Park Golf Course may be legally easier than at the private golf clubs along 19th. However, its location on a hill would make the civil engineering more difficult. We could mine a one-block, single-track connection from the Geary metro to the yard under 40th Ave. Excavated soil from the west side of the site would be stockpiled on the east side of the site for construction of the tall maintenance shop, then placed back over the shop for later golf course restoration. The tree screen between the yard and Clement Street would be maintained throughout construction.
The availability of these large, open golf course sites means that siting a yard for the Geary metro is a phone call problem, not an engineering problem. We do not need to constrain the line’s design and damage its feasibility with long platforms or unique technology to enable operationally-inefficient trips to a distant yard, with these sites already right where we need one.
