Larkspur’s SMART station: Answering the critiques

Last time, we examined how the station got to be placed where it is. The gap in building the first and second segments gives activists a window to try to change the mind of SMART staff and board members, and by the looks of things they’ll need the extra time.

After seven years, the planned site of the Larkspur station is pretty well set in stone, at least if you ask the agency. Whenever asked to move it, SMART has taken the position that the location is final.

This, to put it mildly, is frustrating.

How can the concerns raised by Larkspur years ago and those raised by SMART be addressed?


When Larkspur first voiced opposition to a ferry terminal station, the city council was opposed to the project entirely and objected on three grounds: glare, aesthetics, and a desire to avoid renegotiating Marin Country Mart’s planning documents.

The concern over glare is so odd it hardly deserves mention. Sun glints off parked cars in the ferry terminal and all over the neighborhood. Adding a train would not increase glare.

A train viaduct in Berlin. Image by Jarrett Walker.

A train viaduct in Berlin. Image by Jarrett Walker.

Aesthetic concerns deserve more of a mention. Larkspur argued that views of the Bay would be blocked by an elevated structure and the neighborhood would be marred by a rail viaduct.

Though the vista is dominated by the ferry parking lot, viaducts are very rarely attractive things, at least in the United States. Since the train would run through the parking lot of a major shopping center and across the field of view of some of the stores, SMART should take a page from Germany and incorporate the shopping center into the viaduct itself.

A huge number of trains in Berlin run on elevated tracks, often running right through the city. Unlike the loud and grungy viaducts in Chicago or New York City, these have been integrated into the city by becoming buildings in themselves. Cafes, shops, and restaurants have taken up residence beneath the rails. In essence, the viaducts are very long buildings with trains running on the rooftop.

Marin Country Mart wants to emulate a maritime village, something vaguely European. By using the viaduct as buildings, Marin Country Mart could emulate something actually European. Though it would require cooperation from the SMART board, it would ameliorate the aesthetic concerns of the neighbors and add value to the shopping center’s owners.

Marin Country Mart brings us to the third objection: planning. Larkspur officials in 2006 did not want to revise the Planned Unit Development plan that governs the shopping center, something that would need to happen if SMART extends a viaduct through their property, as two buildings would have to come down on the edge of the property.

But if both parties are willing to renegotiate, there’s no reason why Larkspur couldn’t amend the plan. Extending SMART through the shopping center makes a more direct connection between the train and shopping. That means value added to the center, especially if the viaduct can be made up as nice as the ones in Berlin.


Now that opposition in Larkspur has passed, SMART itself stands opposed to an in-terminal station. As far as I can tell, it’s mostly intransigence. It’s not in The Plan, so therefore shouldn’t be added to The Plan. But publicly, SMART will likely say that it’s an issue of cost (too high) and ridership (won’t change much). These are things we can assess, though intransigence might go a bit deeper.

For cost, elevated rail structures like this one typically cost about $70 million per mile to build, including stations. SMART would need to build a viaduct 2,200 feet long, or about 0.4 miles. Multiply that against the average cost per mile and one arrives at $30 million. Let’s add in $150,000 for building demolition, $100,000 for EIR amendments, and a generous $1 million for land acquisition, for a total cost of $31.3 million. That brings the cost of the whole system from $724 million to $755 million.

This is a bargain, especially for a project of regional significance. If SMART extends to the Larkspur terminal, it could transport a significant number of ferry riders. If it transports even a tenth of them (540 per weekday), the project will cost about $96,000 per trip, not counting the people who will occupy the now-freed spaces in the parking lot. The Greenbrae Interchange Project, in contrast, will add meaningful capacity for about 825 trips* in the peak hour at a cost of about $173,000 each.


SMART staff have dug in their heels on this project, but that’s not to say they can’t be persuaded or forced to come up with a good plan. However, will will take time.

The first thing you can do is understand the costs involved, as above. While the numbers in this post are estimates, SMART has not studied the issue in depth; they know just a hair more than I do about potential costs and ridership. Until there is a proper study, we cannot know for certain how much it will cost, nor how much benefit those monies will buy us.

The second thing you can do is start to lobby boards, commissions, and SMART staff. Since a ferry/train connection is a project of regional importance, TAM, SCTA, and MTC should rank it high on their list of congestion mitigation projects. $31.3 million is a pittance compared to what is doled out in a given year, and this is a critical link in the North Bay’s transportation infrastructure. Residents of San Francisco and Sonoma have leverage as well, as the station will effect the usefulness of their own transportation systems.

Golden Gate Transit needs to push SMART to improve access, too. This will directly benefit GGT’s ferry business and increase the value of their park and ride lot, should they ever decide to lease it to developers.

Find your SMART, SCTA, MTC, TAM, and GGBHTD representatives and tell them you want SMART in Larkspur.

*This is the number of new northbound cars that will be accommodated on the freeway. The project won’t add any southbound or HOV capacity that will be used.

About David Edmondson
A native Marinite working in Washington, DC, I am fascinated by how one might apply smart-growth and urbanist thinking to the low-density towns of my home.

19 Responses to Larkspur’s SMART station: Answering the critiques

  1. Richard Hall says:

    If any section needs to be elevated it’s the section around downtown San Rafael as this train is going to cause immense backups onto 101 four times an hour during rush hour – for the benefit of a tiny handful of train passengers. Sorry, but practicality (of allowing 101 to flow) trumps aesthetics around Larkspur. Really solutions to both are needed – the coming of the train – wholly unjustified economically – is the real problem, and had it been determined that the line promised could not be anywhere near delivered for the 1/4c sales tax this project should have stopped. Unfortunately powerful forces caused a “go, ready set” approach to lets’ go build a train, instead of stepping back and recognizing improved bus services were the real economic solution – without the devastating environmental, aesthetic, traffic and economic effect.

    A better solution to both aesthetics and 101 congestion would not be to impose a very uneconomic train on the region in the first place. It’s simply not the right solution.

    History is about to repeat itself…only this time it will make those in San Jose seem responsible by comparison…

    • Nathanael says:

      [removed to comply with comment policy] You’re just wrong. There’s no other way to put it. The numbers will prove you wrong very quickly. Roads are uneconomic, trains are economic.

  2. Park car Santa Rosa, SMART to Larkspur, ferry to Giants game!

    Park car Santa Rosa, SMART to San Rafael, bus to Giants game?

  3. Richard Hall says:

    Can you elaborate on the math where you’re getting…
    – $96,000 per trip for SMART
    – $173,000 per trip for the Greenbrae interchange

    Genuinely interested – spreadsheet attachment most welcome.

    • Sure thing.

      The spreadsheet you wanted (XLSX), which simplifies all the below. Feel free to make your own assumptions, including how many people will take the ferry now that there are some parking spaces available.

      Basically: Greenbrae will add capacity going north and south. However, it will only relieve congestion going northbound in the evening, reducing a projected 20 minute trip to a 3-4 minute trip. Southbound traffic will see a trivial amount of congestion relief. HOV lanes in both directions will also only see a trivial amount of relief.

      The northbound relief comes because an average of 825 more cars per hour will be able to be accommodated. This capacity will only be used once per day, in other words, by 825 people. Divide $143 million by 825 and you get $173,000. There are safety improvements heading south, but comparing that to SMART would require a more complicated analysis of crashes per VMT that I don’t really want to get into. See here (PDF) for more details.

      The SMART numbers are more difficult to come by. The agency hasn’t made a ridership survey of an in-terminal station, and so I needed to estimate.

      Current ridership estimates 213 trips per day (PDF) at the Larkspur SMART station, or about 4 percent of current ferry ridership. Models typically don’t distinguish between a 15 minute sit and wait transfer and a 15 minute walking transfer, though they are perceived differently by real riders, especially if the environment is hostile to walking, like Larkspur Landing.

      In other words, I had to estimate how many more people would take SMART if they didn’t have that wretched walking transfer. To be as cost-effective as Greenbrae, SMART would need to attract 180 more trips per day to Larkspur, but that felt low.

      Larkspur ridership is growing quickly despite the lack of transportation options available. If ridership grows at an average rate of 1 percent per year (far slower than in the past few years), it will have about 6,700 trips per weekday by 2035.

      It seems reasonable that an in-terminal SMART station could capture 10 percent of current ridership, or about 540 trips per day. If I’m right about 2035, that will be just 8 percent of ferry trips, or double the proportion estimated to use SMART to get to the ferry. In other words, I estimate that 327 (540-213) additional people will use SMART to get to the ferry if it’s in the terminal than otherwise. Divide $31.3 million by that and you get $96,000 per trip.

      Yes, this is a very rough estimate. To get better numbers, I’d need to know where Larkspur riders come from; how many drive; how many people have a final destination around the ferry terminal; and an accurate assessment of a walking transfer penalty.

  4. Richard Hall says:

    Dave – fantastic. I love the fact you have this in an Excel! This gives a great framework/basis for dialog.

    Interestingly the ferry’s own advertising (timetables) promotes that ferry riders need to arrive 20 minutes before departure in rush hour to ensure they get a seat on the boat. I’ve witnessed this first hand “missing the boat” even when on time having been a frequent ferry user. So you might make transit time on the ferry nearer to 40 minutes (walk + wait) before then adding train and ferry and then time at the other end taking MUNI in the city. I find it’s possible to drive to even the deepest points in San Francisco (Potrero) from San Rafael in under an hour. With the train/ferry/MUNI combination transit will be nearer 2 hours. So how attractive will this really be versus driving? (A bus could do all this in one journey without wait/transfer times).

    While you say that the Greenbrae interchange adds capacity for 825 additional people, this doesn’t seem to consider that that interchange is currently routinely causing backups for all 170k 101 northbound commuters (please do correct me if this figure is incorrect) in the evenings adding 10 minutes, sometimes 30 minutes to their commutes. 825 seems to significantly under-represent this impact.

    What seems to be missing is that 101/Greenbrae interchange is already overly in demand – people are already using it. The train is based on hypothesis that people will use it (but as mentioned above with the transit times being so long I challenge this).

    • Don’t forget to use the reply button (when available) when replying to a specific comment.

      Transit, at least in this particular situation, isn’t meant to serve all trips, only a subset of those trips. I’m surprised by the popularity of the ferry given its limited usefulness as a transfer point. For the vast majority of trips from one arbitrary point in Marin to another in San Francisco, the ferry – and therefore a stronger SMART connection – will not be useful.

      The Greenbrae project adds capacity for 825 more cars, which unlocks the whole interchange for the rest of the 5,000-8,000 cars per hour that would otherwise be stuck in traffic. Perhaps there other, cheaper ways to add this corridor capacity. In other words: are some of those 5,000-8,000 trips within the subset of trips that can be taken by transit? I’m sure those who choose to drive have perfectly acceptable reasons to choose that mode. Can we perhaps solve some of the problems that prevent them from choosing transit (or biking, for that matter) for less cost per trip the interchange project? If we can, we wouldn’t need to spend $143 million.

      If an insufficient number fall within that subset of possible transit users, then we can either broaden the transit network to bring enough trips into the galaxy of feasible transit trips or add road capacity to serve those trips, whichever is cheaper per vehicle accommodated.

    • By the way, I wasn’t suggesting the Interchange project should be replaced by this proposal, only that it is more cost-effective at adding capacity (in general) than Greenbrae. If every SMART/Ferry-riding commuter I project would have driven to SF (and therefore from SF), that gets us only about one-third of the trips needed. We don’t know how much congestion that would mitigate, however, because the impact study explicitly excluded transit options from its alternatives analysis.

  5. Stephen Nestel says:

    I don’t believe that even with a timed transfer door to door, will the SMART-Ferry connection would be able to compete with convenience/transit time with a motorcoach . See to look at the real world efficiencies of buses and the VTA.

    • Richard Hall says:

      If we could get off the topic of the SMART, I mean insane train – we might all find a lot of common ground. The train’s adverse traffic impact, ludicrous economics, railroading of environmental and residential sensitivities (eminent domain!) just present an impossible hurdle to overcome. It’s not helping that the train throws in a massive helping of high density housing without any forethought about water supply, schools, traffic impact, actual projected population growth (or in fact shrinkage according to the California Department of Finance projections).

      For the same money (or when SMART comes back to the table asking for another 1/4c) it would be far better to have a fleet of express, luxury buses that really attract people to switch from cars to public transport with WiFi, executive seating, seat back TVs, express routes… (while seeminly ambitious surely all achievable at a fraction of the cost of the … well the unclever train thing).

      Dave – can we have a post evaluating whether improved buses would have been better than the train? We may find lots of common ground (Otherwise you’re just giving us easy fodder to lay into the SMART, I mean insane train).

      • Nathanael says:

        You’re simply wrong. Studies have shown that trains actually attract people. Your “express, luxury buses”….. are significantly less popular than a BAD train service, where both exist. AND they cost more to operate. [removed to comply with comment policy]

        “Dave – can we have a post evaluating whether improved buses would have been better than the train?”
        I think Dave should make such a post, because the answer is already known: “No, they would have been a hell of a lot worse, and more expensive.”

    • By “motorcoach” do you mean car or bus? As I replied to Richard, the number of trips that are practical by transit is only a subset of all trips, so this won’t work for everyone, but let’s say you live in downtown Petaluma and work right near the Embarcadero BART station.

      Google Maps says that will take about 60 minutes hour southbound, and it’ll probably take about 80 minutes northbound given the constraints of traffic, so a 140 minute commute. GGT Route 74 takes about 90 minutes both ways, so a 180 minute commute.

      With a timed transfer between SMART and GGF, that’s 39 minutes on the train, a 5 minute transfer, 30 minutes on the ferry, and about 10 minutes walking. Your round-trip time comes to about 168 minutes. It doesn’t beat driving, but then it’s an issue whether you want to listen to NPR or read your local urbanist blogger in the morning. Or sleep, for that matter.

      For other trips you might want to stick to Route 74, or transfer to a bus at San Rafael, or transfer to rail downtown. Maybe take a bike instead, or perhaps it will still be easier (or more comfortable for you) to drive.

      • Stephen Nestel says:

        I am talking about a an upscale bus along the lines of what winetours use with limousine like accomodations. It is definitely cool to have a cup of coffee, do email, surf the web, listen to the radio while someone else has to deal with road rage and traffic. I think this is a far better option than waiting at train stops, bus stops and ferry terminals everyday. I believe the people of San Jose figured that out and no longer take the VTA as in the video above.

        • John Murphy says:

          The problem with buses is that they are far more labor intensive. I ride a full 72X and it has what, 80 passengers, per one driver. I get on a Caltrain with 600 passengers and 3 engineers/conductors.

          Then you also have maintainance – more maintainance per passenger mile for a bus than for a train. Not solely because the train is carrying more passengers per train mile, but also because a bus has a lot of parts that wear faster – a train wheel gets orders of magnitude more miles than a bus tire.

          As for coffee? No coffee on the bus – because of the logistics of trash that are solvable on a train where you can move around. Bathrooms? None on a GGT bus. 2 per train on Caltrain.

          • A full 72X has only 57 seats, FWIW.

          • Nathanael says:

            Also, the fuel costs per passenger are higher for a full bus than for a full train.

            And unless you have a busway, the bus gets caught in traffic.
            And if you do have a busway, the busway costs more to build and maintain than a railway.

            Oh — and the bus gives a poor, bumpy ride which makes people motion sick (who are prone to motion sickness). While the train gives a smooth ride.

            And on and on and on…. buses are all very well for providing mediocre-to-bad service to places which simply don’t have enough demand to deserve good service, but they aren’t a substitute for a train.

          • Richard Hall says:

            @Nathanael – even SMART’s highly inflated 2035 ridership figures don’t have full trains; what parallel universe do you live in where these trains will be full? Restated, can I have some of what you’re smoking?

  6. Its Me says:

    After have using both the buses, and BART, I would prefer to use a train.
    The train is a lot more comfortable, as wider, also its seem I am more philosophy relax, rather being crammed in to a narrow bus. The other point, if things in the middle east goes off the deep end, thing will get a little furry for your avenge car commuter.

Leave a Reply

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

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

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s

%d bloggers like this: