Central Subway, Muni / SFMTA, San Francisco

Central Subway: Visionary Project or Colossal Boondoggle?

This is Part 4 of a five post series on the Central Subway project. Click here to navigate the table of contents for these posts.

In the last few posts, we introduced the Central Subway project and its potential alignments and stations. The MTA is fond of explaining why San Francisco so desperately needs the Central Subway, and the offered justification goes down the standard checklist of factors that usually come into play when rail upgrades are studied, notably: (1) faster, more reliable, and more comfortable service; (2) potential to increase ridership and decrease congestion; and (3) land use potential. However, this is an expensive investment: $648 million for the initial operating segment, and $1.2-1.4 billion for the subway tunnel. Are we getting our money’s worth? The purpose of this post is to highlight some of the potential difficulties with the Central Subway:

  • Chinatown Stub End: An often-noted problem with the Central Subway is the unnatural terminus in Chinatown. Although the 30 and 45 lines tend to empty out after leaving Chinatown and moving on to North Beach, the subway would enjoy better ridership with even just a one-stop extension into North Beach. As it stands, the Central Subway covers the minimum distance necessary to help relieve sardine-can loads on the 30 and 45 lines (between Market Street and Columbus Avenue), but it fails to address the rest of the corridor.
  • Inconvenient Transfer at Market Street: The Union Square/Market Street Station will be the crucial transfer point between the T-Third and all of the BART and Muni trains that operate in the Market Street subway. Under Alternative 3B, the MTA projects that by the year 2030, anywhere from 9,500 to 12,300 riders will transfer every day between Union Square/Market Street and Powell Street Stations, but the transfer will not be as clean and direct as it should be. As I remarked in the last post, the Union Square/Market Street Station will be centered at O’Farrell/Stockton, but the station will be considerably longer than its 200-foot platform in order to facilitate transfers. However, the distance between the T-Third platform and nearby transfer points is nontrivial, whether it is Powell Station to the south, or the 38-Geary to the north. Combined with the depth of this station (about 90-100 feet below ground), walking and riding escalators and elevators to execute what should be an easy, direct transfer will actually require a nontrivial chunk of time, a burden which is especially difficult for the elderly and disabled to bear. In the end, whatever time savings you gain by riding the subway might be lost while getting to and from the deep tunnel. In addition, there is no easy transfer between paid fare zones. A transfer from the T-Third to another Muni Metro line requires exiting the fare zone at Union Square/Market Street and subsequently re-entering the fare zone at Powell. This is counterintuitive because the entire trip is carried out within the Muni Metro system. In short, the entire design of this crucial transfer point was not thought out carefully, and the difficult transfer actually degrades the connection between downtown and the Third Street corridor. Under the current configuration, T-Third riders have full access to Market Street. Under the Central Subway configuration, T-Third riders may access all of Market Street, but only via this inelegant transfer.
  • Exaggerated Ridership Projections? According to MTA projections, just shy of 100,000 people will ride the T-Third daily by 2030, after the Central Subway has been in operation for 14 years. Of those, 2/3 would be riding the full long line, and 1/3 would ride the short loop between Chinatown and Mariposa. Without having full detail on the models it is difficult to know for sure, but I think there are good reasons to be suspicious of these projections, in part because any such projection will be based on yet more projections about the growth of jobs and population that will occur in Mission Bay and on Third Street in general. In addition, MTA has predicted that by 2030, as many as 89% of riders at the 4th/King station will be using that station to make a transfer to Caltrain — in other words, over 17,000 rides daily. However, the ridership projections for the T-Third do not take into the account the plan to extend Caltrain to the new Transbay Transit Center. It is very fair to say that once Caltrain directly serves downtown, a significant number of people who now transfer to Muni Metro at Mission Bay will simply stay put and ride Caltrain to its new terminus at Transbay. Assuming that funding is obtained to carry out the Caltrain downtown extension, the T-Third ridership projections are artificially buoyed by at least this one significant factor, but really, the numbers are quite high in general. The 2030 projection for just the T-Third line is over 60-70% of the total Muni Metro and F-line ridership in 2007. Even in 2030, in which year the MTA projects that there will be over 320,000 rail line riders daily, the T-Third alone would account for over 30% of the total Muni rail ridership.
  • Continued Use of High-Floor LRVs: The Central Subway has been designed under the assumption that Muni will continue to use its bulky, high-floor Breda cars, despite an increasing trend across the United States to use low-floor vehicles for new light rail projects. High-floor cars complicate and slow down service, while making the system inaccessible to the disabled except at certain locations. A universal system-wide conversion from high-floor to low-floor is admittedly a large proposition, but it is a poor idea to pour more money and resources into building new infrastructure for high-floor vehicles, when a long-term goal should be to convert Muni Metro into a universally low-floor system.
  • A Rat at 4th and King: You don’t need to tell any Mission Bay rider or resident what a mess the intersection of 4th and King turned into after the T-Third was launched in April. The intersection of 4th and King features a mass of transportation alternatives: entry onto and exit off of Interstate 280, the terminus of the Caltrain commuter rail line, frequent heavy pedestrian activity due to its proximity to the ballpark, the crossing of two surface light rail lines, and median rail stations forcing dozens of people to cross intersections unsafely in order to rush to meet their trains. Faulty signal “priority” forced crush-loaded trains to wait for literally minutes on end while single-occupancy autos sped with relative ease onto and off of the freeway. As the interests of pedestrians, transit riders, and drivers competed, this intersection quickly became one of the most dangerous and poorly managed intersections in the city shortly after the T-Third launch, and the best solution was constant human supervision of the intersection. T-Third trains currently turn from 4th onto King, but when the Central Subway opens, they will cross the whole width of King Street when moving north or south on 4th Street. The MTA plans to operate both a short line and a long line on the T-Third track; while the long line would travel the entire distance between Chinatown and Bayshore, the short line would turn around at Mariposa, providing additional service to downtown, South of Market, and Mission Bay. At peak, both lines might run on 5-minute headways, with the possibility of adding another short loop for peak trips. In other words, the planned headways are actually quite short, which will likely cause trains to bunch up once they emerge from the tunnel. According to one version of Alternative 3B, the trains would run in mixed flow with automobiles on 4th between Brannan and King, creating further potential for backup and bunching. Moreover, by 2016, the E-Embarcadero line will be in operation, terminating at 4th and King, and the N-Judah will run all the way to Mariposa Station. With this increase in rail activity, and no promise on the horizon to execute any substantial upgrade to the traffic signals, it does not seem like we are out of the water yet on the 4th/King issue.
  • The 9X Factor: One of the primary motivations behind the Central Subway is to provide a rail link between Visitacion Valley and Chinatown, a key component of the “Connecting People, Connecting Communities” slogan for the T-Third. Still, we should not forget that these two neighborhoods are already connected via the 9X, a popular express bus that runs locally on both ends of the line but uses Highway 101 for the middle portion. How will the T fare, when compared to the 9X? Understanding that Muni schedules should be taken with a barrel (or so) of salt, let’s take a quick glance at them in any case, just to have a basis for comparison. A trip from Washington/Stockton to Bayshore/Arleta on the 9X (this trip covers almost all of the future T-Third route) takes roughly 35-40 minutes, depending on the time of day. According to the current schedule, a T-Third trip from 4th/King to Arleta takes about 28 minutes. The MTA projects that under Alternative 3B, trains traveling in the Central Subway would take 6.3 minutes to travel from Chinatown to 4th/King. Even if we assume there is no delay at 4th/King station (which is unlikely, for reasons explained above), the trip from Chinatown to Arleta on the T would take about 35 minutes. Even a small delay would push the trip over 35 minutes, and riders may still have to spend additional time walking to and from a T station, while the 9X bus has more stops and provides more direct service; in that sense, the 9X could actually be more convenient than the T. In the final analysis, T-Third trains, even using a Central Subway, will prove no faster than the 9X bus. The difference, of course, is that we will have spent $2 billion to build the whole T line, with no time savings for the Visitacion Valley rider market that was especially singled out as benefiting from this project.

Even if we assume that the travel time and ridership projections are correct, and that the project will be completed with no cost overruns — flying in the face of past experience that advises us to consider all of the above very cautiously — even then, the $2 billion total cost for the T-Third initial operating segment and the Central Subway extension is disproportionately high, compared to the service it offers, especially considering its problems and limitations. In some sense, Chinatown looks like a textbook case for a subway, since it is a dense, heavily trafficked neighborhood with so much surface street congestion that buses running in mixed flow are simply not doing the job, but that does not mean that this particular subway is the answer.

The Central Subway is clearly expensive, and the price tag is often dismissed by project supporters because half of the funds are of federal origin. But the mere existence of federal matching funds does not, on its own, turn a flawed project into a good project. If we are going to invest $1.2-1.4 billion in a subway tunnel, we should be as certain as we can be that the money spent is a valuable investment that will offer great returns in the future. In particular, we should be certain that this investment presents the opportunity to reduce operating costs and to increase the efficiency and quality of service. What we have here, though, is an expensive project that does not increase efficiency, nor does it effectively address the needs of riders in the corridor. As much as we are told to believe that the Central Subway will deliver long-overdue transit improvements, its numerous built-in flaws cast doubt on these claims.



23 thoughts on “Central Subway: Visionary Project or Colossal Boondoggle?

  1. This is why you don’t cut political deals with transportation infrastructure. I like the idea of the Central Subway, especially since by 2030 we should be able to extend it to the wharf which is ultimately where it should go imho. But this project should have been secondary to Geary. During the time they were building Geary and engineering the CS, they could have easily just widened the sidewalks in Chinatown (because its so crowded you have to walk on the street) get rid of on street parking and have bus only after the tunnel. However this street needs a subway eventually. It’s the most congested street I’ve ever been on honestly and the sardine tins are not a fun ride either.

    Posted by The Overhead Wire | 16 November 2007, 3:25 pm
  2. The project itself in concept is not controversial. The problem is that the implemenation sucks and the cost is too high.

    Regardless of what the operating plan they assume in the EIR, the actual operating plan can be totally different. Muni would have the most reliable operation by running it as an independent line between Chinatown and 4th Street/Brannon Station, and leave everything else as is.

    Regarding to high floor/low floor issue, it is too late for this project to consider using low floor LRVs. If Muni were to build a new LRT system from the ground up, it should definitely be considered. VTA had to spend years upgrading the platforms to make low floor LRVs to work. Boston, which is an old system with low platforms, also have problems switching to low floor LRTs. There would only be more problems if Muni were to switch now, or even before the T-Third line opened.

    Having high floor cars are okay. Los Angeles’ system is high platforms with high floor cars.

    Posted by Andy Chow | 16 November 2007, 4:28 pm
  3. Andy, as to independent operations: the short loops are as good as we’re going to get, I think. For one, it would not make sense to end at 4th/Brannan; you definitely need to go as far as King. And once Mission Bay is built out, it will make sense to go even further. The service needs to reflect ridership.

    On the low-floor point: I know that this project is not going to be converted to low-floor. You shouldn’t interpret this post as “recommendation” list necessarily. If this were a list of recommendations, I would have only one: stop this project now.

    That said, I do not agree that high floors are “okay” when most of your line runs on street level with no platforms. The older section of the Muni Metro system functions more like a bus, not modern light rail, and the accessible spots are often very widely spaced. Whatever is in place needs to be consistent. Universal, system-wide high-floors would be okay, because the system would be fully accessible. The half-baked solution we have now is not fine. But really, this issue should’ve been addressed before Muni built 18 brand new high-floor stations for the T, as this was the perfect opportunity to do it right. More poor planning.

    Posted by Eric | 16 November 2007, 4:51 pm
  4. Eric, you are right about the older portion of the metro system functioning more like a bus system. Muni really need to modernize the system. I can’t believe that they still have trains stopping at stop signs and stops every block! In fact, some buses are faster then the J line…

    Posted by fs77 | 17 November 2007, 3:05 am
  5. As a CalTrain rider, I’m of two minds. The Central Subway should make it a little more convenient to get from 4th & King to downtown, and that’ll be ok.

    On the other hand, unless the CC is operated a whole lot better than any part of Muni is now, it’s still going to be a pain to make this connection. Expecting the passengers of 5 car CalTrain trains to cram on 2 car LRV trains is pretty crazy, too.

    I feel like to some extent the CC is being going to be pushed as a (very lame) substitute for a CalTrain downtown extension, giving the MTC an excuse to put off the downtown extension indefinitely.

    I’m also kind of appalled that nobody is studying the possibility of combining a Muni and CalTrain subway. Are we really going to build two subways through SOMA, more or less in parallel, just blocks apart? It really seems like there must be some way to combine these projects (a la the Muni/BART subway under Markey) and save a lot on both.

    Posted by Nick/295bus | 17 November 2007, 3:18 pm
  6. Nick, I feel very similarly on the combined subways. At the end of all this, we’ll not only have Caltrain under 2 St, Muni under 4 St, but also Muni on the Embarcadero. And, after all that, we still won’t have rail under Geary. I’m not sure, but I’ve assumed that the reason this hasn’t been studied more carefully is a combination of the planning timelines and funding issues and, well, cooperation. It’s also possible that the current approach for the Caltrain DTX is easier in terms of tunneling, and dealing with the basements of existing buildings. If you combined them, you’d have to get over from 2nd or 3rd to Stockton — and that would be another version of Alternative 2 for the Central Subway. But it’s a good point.

    The Central Subway should make it a little more convenient to get from 4th & King to downtown, and that’ll be ok.
    Indeed, but is it $1.4 billion worth of okay?

    Expecting the passengers of 5 car CalTrain trains to cram on 2 car LRV trains is pretty crazy, too.
    How about 1 car LRVs?

    I feel like to some extent the CC is being going to be pushed as a (very lame) substitute for a CalTrain downtown extension, giving the MTC an excuse to put off the downtown extension indefinitely.
    Unfortunately, though, the Central Subway is in no way a valid substitute for the Caltrain DTX.

    Posted by Eric | 17 November 2007, 3:31 pm
  7. There was an idea floated years ago — like 15 years ago now, which should give you some idea of just how this project came to be such a fait accompli — to turn Caltrain into, essentially, RER or S-Bahn. The basic concept would be a single EMU tunnel from 4th & King under 3rd, Kearny and Columbus. You might have stations at Mission, California, and Broadway. You might then go northwest as far as North Point, or west as far as the Presidio, but that’s part of the problem with both this and the long-term T concept–why? Your demand really peters out past North Beach. There were some other problems as well, most notably that regional rail is not really the sort of service you want to provide to the Northeast; you’d get a one-seat ride from the Peninsula, but not very frequently, and so would do little to relieve Stockton.

    Another alternative would be to keep the T on the surface until roughly Mission. This would save you several hundred million dollars, and the impacts would largely be on traffic, not transit.

    Posted by Steve | 19 November 2007, 12:12 pm
  8. You might then go northwest as far as North Point, or west as far as the Presidio, but that’s part of the problem with both this and the long-term T concept–why? Your demand really peters out past North Beach.
    This sums up very nicely why I don’t really think rail is the right fit for this corridor. The intensely high demand portion of the 30/45 line is so short, and it’s really just in this core section of the route that we need increased capacity — as compared to Geary, where the crowds are sustained for a higher percentage of the route. The current design of Stockton is such that increasing capacity in the high demand section may offer more seats, but won’t add any benefit in terms of speeding up the ride. Dedicated lanes would at least address that issue.

    …the impacts would largely be on traffic, not transit.
    I’d love to see how this goes down. Maybe not quite so bad, though, if you ran trains northbound on 3rd, and southbound on 4th.

    Posted by Eric | 19 November 2007, 2:09 pm
  9. The point of having an independent operation is to have the line operate in an manner similar to real subways, with longer trains, high speed and shorter headways. The existing system is built not to be reliable, where trains operate in areas with low speed and traffic congestion, and trains converge in congested areas.

    It is preferable to force transfers to some passengers but have a very reliable operation. The Central Subway isn’t just going to be used by those living in the Bay View. A more reliable Central Subway will benefit others who transfer from other Muni lines and BART at Market St.

    Unfortunately the 4th & King intersection is so messed up that there’s nothing much could be done about it.

    Also, by the time the Central Subway is supposed to be open, the T line would have been operated for 9 to 10 years. Eliminating the current configuration in sake for direct Central Subway service is similar to the original T line service change, where the N line no longer served Caltrain. The N line operated for almost 9 years from the opening of the extension to the T line opening. That’s why I don’t believe the operating plan outlined in the DEIR/S. I am more concerned with the TEP process.

    Posted by Andy Chow | 19 November 2007, 2:40 pm
  10. Once again: the service has to reflect the ridership, and 4th/King will be way busier than 4th/Brannan — a few times over, at least until Caltrain is extended. We’ll have to wait and see actual ridership, but I think that calling it “some passengers” is trivializing the issue. Given the projections, under TEP standards, I don’t think there’s any way they would recommend cutting off the line at 4th/Brannan, and either requiring a transfer, or forcing people to hike down to Caltrain. Likewise for the northbound direction. Also: don’t forget the future Mission Bay ridership.

    If a service isn’t convenient, people won’t ride it. If people don’t ride it, it doesn’t matter how smooth the operations are. It’s all about striking a balance. Turning some trains at Brannan is one thing. Forcing all Caltrain-bound passengers to transfer one station before King, every single time, is quite another.

    4th/King problems could be easily addressed with an aerial structure, but I doubt we’ll see that in my lifetime. That said, even a pedestrian tunnel would alleviate some of the issue.

    Posted by Eric | 19 November 2007, 2:52 pm
  11. 4th/King problems could be easily addressed with an aerial structure, but I doubt we’ll see that in my lifetime. That said, even a pedestrian tunnel would alleviate some of the issue.

    What kind of aerial structure are you thinking of? For pedestrians? Trains? Cars?

    4th & King to me has always seemed a place where 99% of problems would disappear with some excellent, accurate signage (I know, that doesn’t exist here, but bear with me) and signal priority for the T – no need to even give it to the N.

    Posted by Chris | 19 November 2007, 3:19 pm
  12. Sorry I wasn’t clear about that: aerial track would be great, as you’d get the same quality of service you would in a tunnel at much less cost. But then we’d probably hear Avalon Mission Bay residents complaining about their “destroyed views.” Also, here, it might be a bit awkward with the bridge to the south and the tunnel to the north. A pedestrian flyover at least would be much shorter, but a sufficiently wide, clearly marked pedestrian tunnel could definitely do the job.

    If we had legitimate signal priority to carry the T through, it would solve a lot of problems, perhaps especially when the trains go north on 4th Street before entering the Central Subway — assuming a semi-exclusive right-of-way between King and Brannan. But right now, when riders can use either of 2 platforms to get downtown, we still have the issue of people running back and forth across multiple lanes of traffic in order to catch the train that comes first, an issue further complicated by the fact that NextMuni frequently spits out incorrect predictions as to whether and N or T will be the first to depart from 4th and King.

    Posted by Eric | 19 November 2007, 4:41 pm
  13. I think the line could end at 4th & King, as long as there will be a separate platform that is not shared by either the existing T or the N, nor cross or share tracks with these lines either.

    The platforms on 4th Street by King is not long enough for 3 car service, as well as other stations further south. If that’s one of the primary reasons to justify shorter platform for the Central Subway (the other reason is obviously cost), then it is not worth $1.6 billion.

    Posted by Andy Chow | 19 November 2007, 4:55 pm
  14. Why is there not more discussion about turning Stockton street north of the Stockton tunnel and south of Columbus Ave. into a pedestrian and transit mall? An area with wide sidewalks, transit right-of-way and no cars allowed! There is no need to have cars travel through this area.

    Posted by Charles | 20 November 2007, 3:13 am
  15. @ Charles: Merchants who will raise hell if their thoroughfare becomes “less accessible” to potential customers? (Though, ironically, becoming much more accessible.) A lack of political will and leadership? Sometimes this town isn’t as progressive as we’d like. Also, the last time Chinatown became less accessible to drivers (when the freeway collapsed in the earthquake) there was protest… and well, at least we don’t have the freeway anymore, but we do have this far-from-perfectly-planned subway tunnel instead.

    That said, I agree with you (if you haven’t seen it already, check out the last post of this series).

    Posted by Eric | 20 November 2007, 8:27 am
  16. good stuff Eric thanks for the blog

    This is not totally on topic but is related to the T and coordinated transit planning; in a better world the idea of equity for poor southern neighborhoods would have been a coordination of the electrification of Caltrain to downtown with better local bus connections to the Bayshore station

    The T we ended up with is such simply embarrassing

    Posted by Zig | 26 November 2007, 11:37 pm
  17. Zig: sounds pretty on topic to me.

    If we’re willing to be open-minded about alternatives to tunneling under Stockton, the benefits of BRT on Third Street, as opposed to rail, become clear. The sluggishness of the T-Third is largely due to an excess of stops, and yet, in the absence of any local bus, we’ll continue to hear complaints from riders about stops being spaced too far apart, right alongside comments that service is too slow. Running different versions of local/express bus service in dedicated lanes would address both concerns.

    For fast pass holders, it would require a separate fare, but as you mention, for the really fast trip to downtown, you’d have a 10 minute ride from Bayshore to Transbay on electrified Caltrain.

    Posted by Eric | 27 November 2007, 8:02 pm
  18. in this perfectly integrated fantasy world I don’t see why intra-city commuters couldn’t ride Caltrain with a fast pass

    If the point is moving people it makes sense

    Posted by zig | 28 November 2007, 6:17 pm
  19. in this perfectly integrated fantasy world I don’t see why intra-city commuters couldn’t ride Caltrain with a fast pass

    Oh, for sure. After all, we already have it with BART, so it makes sense to implement something similar for Caltrain too, once the corridor becomes more viable for city residents. I didn’t intend to ding your idea — just pointing out another necessary bit of coordinated effort between agencies.

    Posted by Eric | 28 November 2007, 7:17 pm
  20. Why does everyone always say “rail under geary”. Its pretty obvious from CC alignment 3B that if it ever happens we will have rail under O’farrel and rail over Geary.

    Posted by duncan.idaho | 20 January 2008, 10:44 pm
  21. It is not believable that the 9X can run on 101 at 35 minutes at rush hour. The also assumes no increase in future automobile congestion. This seems optimistic at best.

    The whole point of transit is to move more people at peak traffic times.

    Posted by bernaldaddy | 7 November 2021, 10:44 pm


  1. Pingback: Subway Dream (San Francisco) « Transbay Blog - 20 December 2007

  2. Pingback: The Northwest Subway « Transbay Blog - 18 March 2009

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


RSS Feed Facebook Twitter Flickr

Archives by Month

Archives by Topic

Archives of all blog posts, organized by topics and themes. Click here for more.


Links to some of our favorite urbanist and transit blogs, websites, advocacy groups, news sources, and government agencies. Click here for more.

If you are interested in California water issues, you may want to check out my other blog on that topic.

Copyright © 2007-2021 Transbay Blog.
%d bloggers like this: