Sepulveda Pass and LAX Transit

Since both of these topics have been in the news a lot lately, it’s about time for a look at north-south transit on the Westside.

Sepulveda Pass in Context

Before we start laying out transit lines, we need to understand the urban context of the area in question. This is especially important for rail transit and other high capital cost projects, because bad decisions will haunt us for a really long time. So…

First, in terms of network design, Sepulveda Pass is a world-class bottleneck, right up there with San Francisco Bay and the Hudson & East Rivers. Your reaction to that, might be, “well, duh”, but we need to realize the implications for network design. Jarrett Walker goes into more detail in Chapter 4 of Human Transit but the chief points here are that (a) more deviation from straight routes is acceptable at bottlenecks and (b) bottlenecks are natural locations for transfers between parallel transit lines.

Second, in terms of engineering and cost, Sepulveda Pass is a very challenging and expensive area. We’re looking at a 7-mile tunnel from Westwood to Sherman Oaks, hundreds of feet deep in the middle. Vertical access between the tunnel ends is difficult at best for ventilation, and impossible for a station or emergency exit. This suggests that within the current planning time frame, we’re only going to get one shot at transit through Sepulveda Pass, so we’d better do it right and get a ton of capacity out of it. In 106 years, New York has managed to build only seven crossings of the Hudson, to connect all of Manhattan, the Bronx, Brooklyn, & Queens with all of New Jersey.

Therefore, any tunnel through Sepulveda Pass should serve multiple transit lines on both sides, and provide as much capacity as possible. The stations at each end will be natural transfer points between the lines. It doesn’t make sense to go to the expense of building a tunnel that long if all it’s going to serve in the Valley is one LRT line on Van Nuys. The tunnel should serve at least two lines on both sides, and probably more. We don’t have to actually build all of the lines right away; the important thing is that the piece from Westwood to Sherman Oaks is built properly at the outset. This is probably a great location for one of Alon Levy’s large diameter TBM tunnels, with four tracks running from Westwood to Sherman Oaks, since there won’t be any stops in between.

This also means that the tunnel should serve rail traffic exclusively and have no auto lanes. In addition to having a lower capacity, auto lanes have much heavier requirements for ventilation. There’s also the question of what facilities auto lanes would connect to at each end, since it’s not like there’s a ton of spare capacity kicking around on the 405, the 10, and the 101. (Note: you could argue for a bus tunnel, using dual-mode buses, with exclusive lanes on arterials for the rest of the lines. I’m not going to get into a bus vs. rail analysis here, since the quality of the ROW is more important than the technology.)

Transit Lines Through Sepulveda

Now that we’ve established what the facility through Sepulveda Pass should look like, we can lay out some transit lines to go through it. In my mind, the logical candidates for north-south transit in the Valley are Reseda, Sepulveda (Valley section), Van Nuys, and maybe Balboa. (Anything east of Van Nuys, at least IMHO, is a future north-south line to connect to La Cienega.) There’s also the potential for east-west lines on Venutra, west to Warner Center and east to Burbank. On the Westside, the north-south candidates are Lincoln, Bundy/Centinela, Sawtelle, Sepulveda (Westside section), Westwood/Overland, and maybe, as a stretch, Avenue of the Stars/Jefferson.

Personally, I’d leave Ventura alone as a separate east-west project. Sawtelle is too close to Sepulveda (Westside), so it doesn’t make the cut. The appeal of the Avenue of the Stars/Jefferson route is that it would serve Century City directly from the Valley, but the resulting line has such poor overall geometry that it wouldn’t be very useful for anyone not going to Century City, so I think it’s out as well.

The other intriguing option, which has been suggested by Henry Fung elsewhere, is having the Westside Subway Extension turn north in Westwood and go to the Valley. Assuming the other lines would be LRT, extending the Purple Line would create some technical challenges (including differing vehicle width). I’ll leave that for a future post focusing on that alternative. This option would take care of Century City.

Here’s a rough plot of these options with reasonable stop spacing:


Don’t worry too much about the stop spacing for now; we’ll take a closer look at that in future posts. Remember, the important thing at the outset is to serve the right area and choose logical overall route alignments. You also might guess from this graphic that I’ve got some changes in mind for Metrolink in the Valley. Yet another topic for yet another future post.

In terms of sequencing, the consensus is that Van Nuys is the top priority in the Valley. On the Westside, Sepulveda (Westside) and Westwood/Overland are only ½ to ¾ of a mile apart, so whichever of the two is built first, the other should be built last. I’d do Sepulveda (Westside) first, if only because it’s more centrally located and spreads the wealth. It’d be useful to some future users of the Bundy/Centinela and Westwood/Overland lines, whereas those two lines wouldn’t help each other’s riders much. The argument for Westwood/Overland first is that it’s closer to Palms and Culver City, which are denser than Mar Vista, and it’s a good enough argument that you could probably talk me into it.

I’d sequence the lines as follows:

  • Van Nuys and Sepulveda (Westside)
  • Reseda and Lincoln
  • Sepulveda (Valley) and Bundy/Centinela
  • Balboa and Westwood/Overland (no Westwood/Overland if Purple Line is extended)

The benefits start immediately with the first line completed, and are amplified as additional lines are finished. To the south, future phases could extend the lines out Florence, Manchester, Century, Hawthorne, Sepulveda/PCH… more than enough possibilities to leave for a future post.

LAX Transit

Note that all these lines naturally converge near Sepulveda & Century, right at LAX’s front door, and would serve far more people than any LAX rail transit proposal on the board now. So in addition to serving LAX, basically at the future Terminal 0, this project would directly serve a couple million other people who might or might not be going to LAX. In other words, this plan would follow one of Jarrett Walker’s main principles: be on the way!

You might have noticed in the first graphic that I didn’t show any connection from the new lines to the Green Line & Crenshaw Line. Clearly, you’d want to provide that link somehow.

Here’s one option for an initial build with two lines. In this scenario, the Crenshaw Line would take over the Green Line’s route south of Aviation/Imperial, and the Green Line would be extended a mile west to meet up with the new Reseda-Lincoln Line. This gives the combined Reseda-Lincoln-Green Line and the Van Nuys-Sepulveda (Westside) Line front door access to LAX, with decent geometry and without making any through passengers go out of their way.


I’ve violated my own rules on stop spacing in El Segundo, going to half a mile to provide a Green Line stop at Maple and a Sepulveda Line stop at Mariposa. With this level of transit service, easy access to LAX, the 105, and the 405, there’s no reason El Segundo’s business district couldn’t become LA’s third downtown.

Here’s an option for full build with four lines.


In this case, I’ve routed Balboa and Bundy/Centinela together, and the line could be extended out Florence towards South LA, HP, Bell, Cudahy, Bell Gardens, and Downey – all dense cities that should get good transit use. I’ve also shown the Purple Line being extended to the Valley and up Sepulveda.

Again, don’t worry too much about the specific stations and routings – we’ll go into more detail on each option in the future.

To Bore or Not to Bore

The decision to tunnel is one of the biggest ones that must be made. Tunneling results in faster speeds and more reliable operations, but the higher cost can push project completion further into the future. Obviously, we’re tunneling through Sepulveda Pass, but on either side, it would be possible to do full tunnels, surface running with selected grade separations (like Expo Line and Crenshaw Line), or full surface running.

Any surface running segments are dependent on the ROW of the arterial roads. Contrary to popular conceptions of LA, the north-south arterials on the Westside are somewhat underpowered. Lincoln and Sepulveda (Westside) are two lanes each way with a center turn lane. Bundy/Centinela and Westwood/Overland are cobbled together, with pieces where the second travel lane is only provided during peak periods by sacrificing on-street parking (always a sign of desperation).

The situation is much the same on Reseda and Balboa, which have the same basic configuration as Sepulveda (Westside), but with more generous proportions. On Balboa, there are three lanes each way with no parking. Sepulveda (Valley) and Van Nuys are wider, at least three lanes each way with a center turn lane plus parking. In many places they’re even wider (presumably where there used to be transit ROW in the middle).

I was going to go into detail and compare grade separation options for each branch, but that would make this post much too long. Now that we’ve got the basic framework set up, we can come back and give each branch the attention it deserves in future posts.


21 thoughts on “Sepulveda Pass and LAX Transit

  1. Alon Levy

    I’ll bite: why does there have to be a tunnel? The ruling grade through the pass is if I remember correctly 5%, which is within the capability of LRT.

    If there is a tunnel without stations, then the benefits of large-diameter TBMs are reduced. One neat thing that can be done with large-diameter TBMs, which I doubt has ever been done, is putting all four tracks in one tunnel. It’s possible to stack tracks 2*2 in one tunnel if there are no stations, it’s just that the main advantage of large-diameter tunnels isn’t four-tracking but being able to build crossovers and stations without blasting caverns.

    1. letsgola Post author

      That’s… a really good question. The project has been so consistently pitched as a tunnel from Westwood to Sherman Oaks, I just started from there. Dammit!

      From Sunset to Mulholland, you have a consistent grade of 3.0-3.5% for about 4 miles. From Ventura to Mulholland, it’s about 5.0% for 2 miles, with mile in the middle being about 6%. Those grades are within LRT capability. Do the long distances at that grade make a difference? The Transit Coalition claims a 130′ tall aerial structure would be needed… which really isn’t much taller than the way cool direct HOV-HOV ramp from the 105 east to the 110 north. (All grades approximated from Google Earth.)

      A fully elevated or at-grade route via the 405 would favor building the Sepulveda Line in the Valley first, instead of Van Nuys, since it’s much easier to hop over from the 405 ROW to Sepulveda. Everyone seems to want Van Nuys first in the Valley. If you go via the 405 ROW, any route over to Van Nuys is going to have some rough geometry. One thought would be to go elevated or at-grade via the 405 ROW part of the way, say to Bel Air Crest, and then tunnel from there to Ventura/Van Nuys. That would save you about half of the tunneling.

      The idea with the large diameter TBM would be to put all 4 tracks in one tunnel – build it once and be done. It may turn out that doing two 2-track tunnels would be more cost effective due to less excavated material; the question w/ that would be if you need a bigger launching pit. I’m not sure we’ll really need 4 tracks from the Valley to the Westside. However, if the cost premium for one 4-track large diameter TBM tunnel isn’t that much over one 2-track tunnel, maybe you build the large tunnel and only put in 2 tracks for the time being.

      Anyway, Part 2 of this series is going to be a closer look at the trunk from Westwood to Sherman Oaks… so it looks like you just expanded my analysis 🙂

      1. Phantom Commuter

        Just a brief comment in regard to Florence. Are you aware that there is a nice, wide, former P.E. ROW on Randolph Street in the same corridor ? It serves the same transit friendly communities and continues all the way to Orange County. It was used by the former Whittier and Yorba Linda P.E. lines, both of which had a lot more passenger service than the Expo Line ever did. Check it out on Google.

      2. letsgola Post author

        Huh, I’ve seen the ROW but I assumed it was always just a freight ROW. It is a really nice corridor… maybe better as a substitute for Slauson than for Florence, but there would be a ton of demand either way!

  2. keaswaran

    Do you think it’s a reasonable assumption that at some future point a rail line will be allowed to continue down Sepulveda, tunneling under the runways? My understanding was that a lot of the current discussion about LAX transit has been structured by the assumption that this just isn’t going to work. Of course, most of what lies between the junction of Lincoln and Sepulveda, and the planned Crenshaw line station, is Parking Lot C, which will hopefully be completely restructured anyway, so there might be a way to bring all these lines together at the Crenshaw station. That, of course, would require either turnarounds or extra tracks or something there to accommodate the capacity of having all these lines in one place. But do you think that would necessitate a major change to the vision you have here, if the lines can’t go down Sepulveda under the runway? Or could it just be a friendly modification of the same basic plan?

    1. letsgola Post author

      The Transit Coalition’s plan has something more or less like that. My main complaint would be that it adds two miles and four 90-degree turns to an otherwise arrow-straight route down Sepulveda. That’s not a minor deviation or geometric concern; it’s the kind of thing that makes the route much less appealing to people who aren’t going to the airport or transferring at one of those stations, and I’d expect that to be a majority of people.

      In general, my view is that you should never let regulatory or bureaucratic issues cause major design compromises. After all, they are problems that we have created for ourselves. From a technical point of view, tunneling under the south runways parallel to the Sepulveda Tunnel without disrupting air traffic is an easy task. As Alon Levy says: organization before electronics before concrete. This is an issue that can be solved with organization. Perhaps LAWA would be less concerned about a tunnel right next to the existing Sepulveda Tunnel than they would be about a tunnel through the Central Terminal Area and under the north runways. In any case, the LRT options currently on the table are bad options for reasons other than LAWA opposition, and conversely, LAWA concerns should not be allowed to stop better options from moving forward.

      Edit: sending the Lincoln & Sepulveda lines over to Aviation might make things worse for people riding those services to LAX too. Instead of a stop right at Sepulveda & Century, you might end up with a stop closer to Sepulveda & Lincoln or 98th & Vicksburg – further away from the terminals.

  3. cph

    1. I think these ideas are neat, If the tunnel is ever built, the routing will probably be Van Nuys (maybe starting at Sylmar Metrolink and running along San Fernando Road a bit), then the tunnel to Westwood (maybe with a stop on the UCLA campus as well as at Wilshire/Westwood)

    2. South of Westwood is tricky, but doable. As you noticed, the streets on the Westside are narrower than the ones in the Valley; at least some tunneling (like on the East LA Gold Line) will be required in some places…

    3. Now as far as any rail line “going to LAX”–remember, LAX has what, 8 terminals? The expectation (of those who don’t really know better) is that the light rail would have a station at each terminal, just waiting to whisk passengers away from the baggage carousel. That would never happen due to the high cost-benefit ratio. Metro could consider one or two stations in the terminal loop, but that would also require expensive tunneling. A people-mover, connecting to the rail system just outside the terminal loop (say, near the Lot “C” bus station) probably gives more bang for the buck…

  4. Pingback: Compete With Arterials, Not Just Freeways | Let's Go LA

  5. Pingback: LAX Transit Part 2 – APMs and Ground Side Access | Let's Go LA

  6. Pingback: Downtown in a Polycentric Region | Let's Go LA

  7. Pingback: Sepulveda Pass Transit, Prelude to Part 3: Should Cars Be Part of a Tunnel Through the Pass? | Let's Go LA

  8. Pingback: Sepulveda Pass Transit, Part 3: Mode and Alignment Through the Pass | Let's Go LA

  9. Dan Golding

    Just a question from a Bike Commuter who would do it more often if it was easier (the hills) and less deadly in LA. Has anyone seriously considered ski lifts/gondolas to the Skirball from both directions? Or through any of the passes?

    1. letsgola Post author

      I don’t think anyone’s looked at it. Gondolas have been successfully applied in cities in a few cases (Roosevelt Island, Medellin, Portland) but the issue in an urban setting is that they have lower capacity than rail. (That said, a tunnel through Sepulveda Pass doesn’t do much for anyone trying to get to Skirball.)

  10. Ryan

    Any possibility of a Sepulveda Pass line running down Westwood Blvd. after the planned Wilshire/Westwood connection? It’s a popular business corridor that also sits walking distance from numerous residential neighborhoods, and is already more of a “destination street” than Sepulveda in that area.

    You’d keep the connection with the Expo line, and could follow the natural curvature of the street when it turns into Overland (via National Pl.), go down Overland (another large street in close proximity to neighborhoods) before turning right at either Venice or Washington and then going down Sepulveda all the way to LAX.

    If you consider that any Sepulveda pass line has a connection to Wilshire/Westwood, then to continue straight down Sepulveda, you have to make a 0.5-0.8 mile turn anyways – why not do it later on? The stretch on Sepulveda between Wilshire and Venice isn’t very attractive for travelers. Westwood may be too narrow, but I’m curious to know your thoughts

    1. letsgola Post author

      I’m torn between Sepulveda and Westwood/Overland as the first north-south line on the Westside; that will probably be the next post in the series. Westwood is a really good corridor north of Pico, and Overland would be more useful for Palms, which has a lot of UCLA students. OTOH, stops at Sepulveda/Santa Monica and Sepulveda/Pico would serve West LA & Sawtelle, and there’s at least plausible development potential there. Palms will be served by two Expo Line stops. And the thought of dropping the major east-west/north-south transfer between this line and the Expo Line at Expo/Westwood, an area doomed to be single-family residences for the foreseeable future by Rancho Park NIMBYs, sends me into an irrational fit. (It’s also bad policy since ideally transfers should be located at major development nodes, and Sepulveda/Pico is better.)

      Anyway, it’s something I want to look at in more detail soon!

  11. Pingback: What Does Induced Demand Really Mean? | Let's Go LA

  12. Pingback: How Does Your Grid Grow? | Let's Go LA

  13. Pingback: Commuter Rail Infill Stops | Pedestrian Observations

  14. Pingback: Future Los Angeles Metro Investments | Pedestrian Observations

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 )

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