Tag Archives: transit

The Money’s in the Infrastructure

This is just a short thought on the economics of transit capital costs and operations, which has been bouncing around in my head in the wake of service reductions at many agencies.

As any transit planner would tell you, reducing service can lead to a vicious cycle, where less frequent and therefore less convenient service causes ridership to drop, which becomes justification for further cuts. This is bad enough in a generic bus system, but at least in that case there’s little capital infrastructure that goes to waste, since most city buses just run on regular streets that are already there.

However, for something like rail transit, it’s truly crazy to cut service (unless you’re forced to by maintenance needs) due to the relative magnitude of capital investment. A brief example shows why.

Consider a 10-mile rail line built at a cost of $150m/mile, a total of $1.5b. Spread out over 30 years at 3%, the cost of construction is about $76m/year. If we run the line for 20 hours a day (4am – midnight) with 6 minutes headways, with reasonable cost per revenue-mile, it costs about $24m/year to run the line. This assumes 320 weekday equivalents, representing slightly reduced service on the weekends. The total cost is about $100m/year, of which capital costs are 75% and operating costs 25%.


Now, let’s impose an austerity plan on the line, reducing service to 16 hours a day (6am – 10pm), cutting frequency in half to 12 minutes, and further reducing weekend service to get weekday equivalents down to 300. The operating cost is reduced by over 60%, but the capital costs cannot be changed. The total cost is about $85m/year, only a 15% reduction from the base plan. As a result, passengers will get much less useful service and some will quit riding the line altogether, further worsening the financial position. And the austerity plan will likely reduce the efficiency of labor and equipment usage, again cutting into the savings.

We can see how illogical this is by considering some simple analogies to driving. Once you have bought a car and committed yourself to monthly auto loan and insurance payments, you can’t cut your costs very much by not driving. In fact, not driving may worsen your position by depriving you of employment. The capital costs are large relative to operating costs.

Likewise, if Caltrans is short on money for maintenance, it would be silly to try to rectify that problem by simply closing one or two lanes on the freeway. No one would ever suggest this because it would be considered intuitively obvious that closing freeway lanes constructed at great capital expense to save a few dollars on maintenance is not in the public interest.

The wrinkle here is that transit agencies often don’t pay for the full cost of capital projects, or don’t account for it out of the same pot of money. In that case, there really should be some mechanism to ensure that adequate operating funds are secured so that the public’s capital investment isn’t wasted. I believe the FTA requires recipients of New Starts funding to demonstrate this in a finance plan, but the enforcement may not be there. The FTA has sometimes demanded that funds be returned when not used for the capital improvements promised – see ARC and Cleveland for examples – so maybe service spans and frequencies should be spelled out in funding agreements as well.

Autonomous Vehicles and Human Factors

Over in VC-land, we are told that autonomy will change traffic in cities from circuit switched to packet switched and from TDMA to CDMA. We will take this to mean that (a) riders may use several vehicles over the course of a trip and (b) several riders and several uses (passenger, freight) may use a vehicle for portions of a trip as the vehicle travels between two points. The alternative interpretation, that different portions of vehicles will be sent via different routes and different vehicles will simultaneously occupy the same space on the road, would be a bit too fantastical.

Astute readers have no doubt noted that we already have packet-switched CDMA transportation in cities: fixed-route transit service, where people board and alight at many points along the vehicle’s route, and users may be forced to transfer one or more times between different vehicles to reach their destination.

Looking past the annoyance of everything being turned into tech jargon, this is worrying because it is nothing more than a tech industry reframing of the mistaken “water through a pipe” philosophy of early traffic engineers. The idea that human users can be routed through a transportation network like packets of data across a communications network is akin to the idea that drivers move through a transportation network like water through a pipe. Understanding the human factor is one of the hardest learned lessons of traffic engineering, indeed, something many engineers still struggle to do.

People do not act like water in a pipe or packets of information on a network. Transfer penalties are real and vary with the person, the weather, the trip purpose, and other factors. Many people are not willing to share a ride with one anonymous person – note that the practice of “slugging” arose where HOV restrictions required three people rather than just two.

It is also not certain that ride sharing will replace single occupancy vehicles as autonomous vehicles become more prevalent. None other than Randal O’Toole provides some reasons why: people like having their own car with their own stuff in it, people don’t like the idea of a stranger using their personal property, and if autonomous vehicles reduce the cost and annoyance of car ownership, more people may choose to own. You’d be crazy to pay to keep a car in Manhattan & drive it around, but what if you could have the car drive you around and then go park itself in New Jersey when you don’t need it?

Another example of ignoring human factors is some presentations of automated intersections. For this to work, if pedestrians and bikes are to be permitted at all, they would have to behave in a perfectly predictable manner. Of course, we don’t – we stop to take out our phones, we stop to look at things, bike chains slip off gears. If full automation and vehicle-to-infrastructure communication are achieved, these intersections could prove useful for junctions of limited access facilities, but they won’t be popping up in cities. (And they will likely be more conservative than presented in these simulations, due to the need to allow for mechanical failures, unexpected pavement conditions, and so on, but that’s another issue.)

I don’t mean to suggest that autonomous cars won’t have any impact on cities. The improvements to safety alone from eliminating human error, inattention, and bad behavior will be well worth it. But if you’re waiting for the paradigm shift of changes being hyped in some of the press, I wouldn’t hold my breath.

Work Windows

With WMATA’s unplanned weekday shutdown as the acute trigger and LA Metro’s work on the Red Line & Blue Line as the chronic one, here’s a short take on urban rail and work windows.

Working on any railroad is dangerous. This is especially true of rail transit, which has higher train volumes and more constrained spaces. In tunnels and aerial guideways, there might not be any space at all to clear the tracks other than what’s provided for emergencies. Working on a railroad is also more complex than working on a roadway. Trains can only switch tracks where crossovers are provided. With roadway construction, we can usually close lanes anywhere we want, and lane closures can modified relatively easily.

As a result of these factors, it’s difficult to do construction and maintenance work on urban rail transit systems except when the track in question is completely shut down. This means that agencies are left with two choices: single tracking and full closures. Single tracking, which is what LA Metro has been doing evenings and late nights on the Red Line, means that all trains in both directions share one track through the work zone. Full closure means, well, full closure.

Single tracking lets agencies maintain some level of service, but capacity may be quite low, as you obviously can’t send a train in one direction through a single-tracked section until the train going in the opposite direction has passed. Capital cost decisions made long ago will determine the capacity. For example, the Red Line doesn’t have any crossovers between Union Station and Westlake/MacArthur Park, and there are three stations in between (Civic Center, Pershing Square, 7th/Flower). If it takes 2 minutes including dwell time to travel between stations, it’s about 8 minutes eastbound from Westlake to Union. Allow 3 minutes at Union for turnaround and 8 minutes westbound from Union to Westlake, and there’s your 20 minute late-night headway. Published schedules say 7 minutes, but I think it’s reasonable to add a little extra to account for passenger confusion with trains sharing one track. Note that if one train shows up late, following trains will have to be held, adding to unreliability. To improve on this, costly capital improvements to add more crossovers in underground tunnels would be needed.

Single tracking also reduces the efficiency of the work being performed. It limits workers to one section of track, making it more difficult to perform concurrent work or sequential tasks, as it’s difficult to have more than one or two crews working in an area.

Full closures solve these problems, because you don’t have to worry about maintaining service. This is part of why so many systems close down for at least part of the night. However, even with nightly closures, it is difficult to accomplish work efficiently. For example, if the work zone is difficult to reach, a substantial amount of time may be lost to setup and breakdown time. For example, the only way to bring equipment into the Red Line is at the portal by Union Station. If you are working between Universal City and North Hollywood, you have to bring equipment all the way out there from Union, at low speeds.

Because startup and breakdown times are fixed, costs do not scale linearly with the length of the work window, whether it’s single tracking or a full closure. For example if the last train is at 9pm and the first train is at 4am, you have 7 hours. Of that, the first hour will likely be spent waiting for permission to enter, confirming power is shut down, and getting set up. The last hour will likely be lost to breaking down and clearing the track, probably well before 4am to make sure there are no issues. Operations isn’t going to let you plan to get out at 3:59am. This leaves you 5 hours to work.

If the last train is at 11pm, the window gets cut from 7 hours to 5 hours, and effective work time gets cut from 5 hours to 3 hours. Much less work will be accomplished, but labor costs will not necessarily go down; you can’t call someone out to work in the middle of the night and only pay them for 5 hours of time. In addition, the 2 hours of work that you lose are the most effective hours, because construction work tends to get more efficient as workers get more familiar with the site and each other. You lose 2 hours of the worker swinging the hammer but keep the 1 hour of figuring out where it needs to be swung.

Of course, if you’re some smarmy New Yorker who inherited 4-track subway trunk lines, you can close down tracks for long periods of time, like a weekend or weeks, and maintain service 24/7 using the other tracks. Sure, it might be inconvenient to have no express service or to have to use express and ride back on the local, but it’s still service.

However, most cities don’t have 4-track subways. Despite what some people might want you to think, it’s pretty clear that you don’t need a 24/7 subway to be a “world-class” city:

  • Tokyo: about 5am – midnight
  • Seoul: about 5:30am – midnight
  • Shanghai: about 5:30am – 11:30pm (last train usually departs outer terminal before 11pm)
  • Hong Kong: about 5am – 1am
  • Paris: about 5:30am – 12:30am (2am weekends)
  • London: about 5am – 12:30am
  • Singapore: about 5:30am – midnight

Look, zero late night rail service does not have to mean zero late night transit service. Traffic is minimal between midnight and 5am just about everywhere; there’s no reason that service can’t be provided by buses during those hours. Riders might be better off with more frequent bus service in lieu of very long headways on the rail network to allow single tracking, and it might not be very expensive if it increases the efficiency of construction work. I strongly believe that good 24/7 transit is necessary in large cities, because there are always people that need service. It just doesn’t have to be on steel wheels.

From Glendale to Downtown LA and Back

Consider this post to be, um, sorbet, a palate cleanser before the long-promised meatier course – a course for which your impatience with the chef is no doubt growing.

Living in Palms, commuting to downtown was easy: I could take my early af carpool, or I could walk to Culver City station and take the Expo Line. Driving to downtown that early, there’s practically no traffic on the 10. But did I mention it was really early? The Expo Line, with 10-12 minute headways all day long, about a mile from my apartment, was the natural transit choice, unaffected by the whims of the traffic deities. If something disrupted rail service, like drivers behaving badly, the Venice bus routes (33/733) were a solid backup, even if the lack of bus lanes on that wide ROW west of Crenshaw got frustrating.

In Glendale, the length of my commute is the same, within less than a mile. I still have the option of the crazy early carpool, the one that lets me start tweeting when the rest of the West Coast is still dreaming. There’s still no delay driving at that time of day, but the background traffic on the 5 is significantly larger; on the 10, there’s nothing but ocean to the west, while on the 5, there’s a lot of long distance north-south traffic.

On the other hand, there’s no rapid transit to Glendale, so the transit options aren’t as good. The closest Metro bus route to me is the 94/794 on San Fernando Rd. As has been discussed on Twitter, the split between a local and rapid here is not particularly helpful, because the headways on both are large enough that you’re better off just taking whatever comes first. The 94/794 is nearly 30 miles long, about twice as long as many Westside bus routes, which makes it even harder to regulate headways. Lastly, the 94/794 uses Hill St downtown, which adds a lot of delay when traffic is stacked up getting on the 110.

You can try to skip past downtown congestion by taking the Gold Line to Lincoln Heights/Cypress Park, and taking a short walk to the 94/794 stop at Ave 26 and Figueroa. However, if Union Station isn’t one end of your trip, that means two transfers, and two transfers can add a ton of delay. Odds are, of course, that Union Station is not one of the ends of your trip.

Today, I finally tried taking Metrolink from LA Union Station to Glendale. The train left on time and it was a fast 10-minute ride to Glendale Station, which is near the southern end of the city by Los Feliz Blvd and San Fernando Rd. Even with zero traffic, you’d be hard-pressed to compete with that time by car. Thanks to Art Leahy and Mike Antonovich, the fare currently sits at a very reasonable $2; before the Antelope Valley Line pilot program, it was $5.50. Honestly, that kind of speed is probably worth $5.50 and I’m just a cheapskate.

Again, though, if you have to transfer, that advantage starts to rapidly dissipate. I happened to be at Union Station today; for most people a Red/Purple Line ride would be tacked onto the end, but service there is frequent enough that it’s not a big deal. At the Glendale end, I had to wait for the 94/794, and the last 2 miles of my trip ended up taking more than twice the time that the first 8 miles took. Glendale runs a bus, route 12, from the Metrolink station up San Fernando Rd; Glendale routes 1, 2, and 11 would also arguably be viable for my trip. The overarching problems with any of these transfer options are the potential for a long transfer delay and infrequent or non-existent service during off-hours.

Two final options that would serve my commute would be Metro bus route 92, and Metro bus 180/181/780 to a transfer to the Red Line. I haven’t had occasion to try these; to be honest, the traffic on Los Feliz Blvd scares me a little bit regarding the latter.

Meanwhile, the Metrolink tracks paralleling San Fernando Rd offer an intriguing possibility. But more on that another time.

Driverless Cars and Driverless Trains

Updated with a note on platooning and some input from R Winston Kappesser (@ronaldkappesser).

There was some back and forth on Twitter today on the potential of driverless cars and their impact on rail infrastructure like transit and high speed rail. In that context, here’s a civil engineering perspective on the technological issues and potential impacts.

Stuck to You Like Rubber on Asphalt, or Steel on Steel, or Something

First, we have to understand the technological differences between rubber-tired vehicles and steel-wheeled ones. That starts where the rubber hits the road or where the steel hits the steel.

When it comes to transportation, friction is both our friend and our enemy. We need some friction; otherwise, when you hit the gas your tires would just spin in place, or your rail wheels would just do something like this. Friction between the rubber tires and the road’s asphalt or concrete surface is what keeps cars and buses from flying off the road at corners. It’s what turns the tractive effort of a big honkin’ locomotive into forward motion.

On the other hand, too much friction wears out your car’s tires and makes your car run less efficiently. For trains, friction management is a critical part of track and vehicle maintenance. If there’s too much friction, the rails and wheels will wear out faster, and the train will use more fuel, increasing maintenance and operating costs. Friction management is so important for railroads that locomotives are equipped with sanders, so that the engineer can drop sand on the rails to increase friction on upgrades, while sharp curves are equipped with greasers to reduce friction between the wheels and rails. There’s an entire sub-industry built around friction management.

In general, the coefficient of rolling resistance between rubber and asphalt is about an order of magnitude larger than that between steel and steel. This means there’s proportionally more rolling resistance between your car and the road than there is between an Expo Line train and the rails. The very low rolling resistance on railroads is part of why trains are so much more efficient at long-haul freight than trucks. The coefficient of friction is also lower for steel on steel than rubber on asphalt.

Can’t Stop, Won’t Stop

That efficiency comes with a cost, though, in braking performance. Trains can’t brake as fast as rubber-tired vehicles. How much worse? For a 70 mph design speed, Caltrans Highway Design Manual requires a stopping sight distance of 750 feet. For a 70 mph design speed in territory with cab signals, the standards used by Amtrak and many commuter railroads require a safe braking distance of 4,942 feet. For high-speed trains, the stopping distances for purposes of rail signal design can be in excess of 2 miles.

(Note: some vehicles, notably LRT vehicles and some high-speed trains, have electromagnetic track brakes that use electromagnets to “grab” the track, allowing the vehicle to stop much more quickly. These brakes are used for emergency only; safe braking distances for railroad signal design are calculated assuming no track brake is used.)

The other major difference between rubber-tired vehicles and steel-wheeled ones in this regard is the ability to steer. A person operating a rubber-tired vehicle has the ability to take evasive action to steer the vehicle away from a hazard, while a train engineer is obviously helpless to do anything other than brake.

Design Evolutions

These technological realities have resulted in a different evolutions of civil engineering design standards.

For cars, design is predicated on the driver being able to see further than the distance needed to stop the car. The design of vertical curves (changes between upgrades and downgrades) is governed by the need to ensure the ability of the driver to see over the top of the hill, or for the car’s headlights to illuminate enough of the road ahead of a sag curve. At horizontal curves, vegetation and other obstructions on side of the roadway must be cleared far enough back from the edge of the road to allow the driver to see around the curve. In a safe design for autos, the driver will always be able to see further than needed to stop the car.

In contrast, with the exception of low-speed streetcars, for trains it is simply impractical to design the track such that the engineer would always be able to see further than the distance needed to stop the train. Horizontal and vertical geometry of the track is controlled by vehicle performance and passenger comfort. Safety is ensured by the signal system providing the safe operating speed to the engineer, and in some cases enforcing that speed, based on the locations of other trains (or perhaps more accurately, based on information that sections of track ahead of the train are not already occupied by other trains).

Note the fundamental difference here. For cars, safety is based on the ability of the driver to passively gather information about conditions on the road. For trains, safety is based on active collection of information on the locations of trains, and active dissemination of instructions to trains that it is safe to proceed.

Driverless Technologies, and Others

This means that the interfaces and impacts of driverless technologies will be different for cars and trains. For cars, passive decentralized technologies (i.e. the car just gathers information, and doesn’t communicate with other cars or with a central control center) will suffice. For trains, centralized control is a necessity.

For cars, it will be a huge improvement for safety simply for driverless cars to more reliably and consistently do the things that we currently rely on human drivers to do. This will have some positive impact on practical capacity by reducing accidents. If driverless car technology allows cars to follow each other more closely than they do today, by eliminating the component of following distance related to human reaction time, that will increase road capacity.

For example, you may have already figured out that, despite the stopping sight distance being 750 feet at 70 mph, cars on a freeway flowing at 70 mph don’t actually space themselves 750 feet apart. At that rate, a freeway lane would only move about 500 cars per hour, but the actual capacity of a freeway lane is about 2,200 cars per hour. If you have a driver’s license, you may (hopefully) remember the “two second rule”, that you should leave about 2 seconds of travel distance between yourself and the car in front of you. At 70 mph, that’s a little over 200 feet – less than the stopping distance, and acceptable only because you can see further than just the car in front of you, and you have time to swerve out of the way if needed. Part of that 2 seconds is an allowance for human reaction time; if driverless cars allow that component to be eliminated, they will increase capacity.

On the other hand, safe design for trains is based on maintaining at least the stopping distance between following trains. At 70 mph, a train should never be less than 4,942 feet behind the train in front of it. In practice, the distance will always be larger due to the impact of grades and the use of fixed signal blocks. The engineer’s reaction time is portion of that stopping distance, but it’s not much. Driverless train technology has been around for decades, but the primary appeal is reducing labor costs, not increasing capacity.

If the goal is to increase capacity on rail transit, communications-based train control (CBTC) will probably offer more benefit than driverless technology, because it will eliminate the capacity waste caused by fixed signal blocks. CBTC should also allow railroads to take advantage of better braking performance available in newer rolling stock. The combination of CBTC and driverless trains would allow many transit systems to greatly improve service by increasing capacity and reducing labor costs, thereby allowing the agency to provide more service.

Lawyer Up

A big unanswered question, in my humble opinion, is the liability implications of driverless vehicle technology.

For cars, what will be the standard for safe following distance? At present, we allow drivers to follow each other at less than safe stopping distance. Will driverless cars follow the “two second rule” or will they be allowed to follow more closely? If there’s a rear-end collision, who is liable? Note that some of this must have been decided implicitly or explicitly by the people who have operational driverless cars, like Google.

For trains, at present, railroad signaling is based on the premise that the train in front of you is at a stop, and therefore you must be able to stop too. If you implement CBTC you could argue that if you know the position and the speed of the train in front of you (Heisenberg be damned), you should be allowed to follow more closely. On the other hand, if the train in front derails for some reason, it’s going to come to a stop very quickly, and any following trains that are less than the safe braking distance behind are hopelessly screwed. There’s not a consultant in business in the country today that’s going to sign off on allowing trains to follow each other at less than the safe braking distance, and I doubt any agencies would do it either.

Therefore, for trains, I really think the capacity improvements are going to come from CBTC, not driverless technology.

The Future is Uncertain

I hope this post doesn’t sound like it’s down on transit. For one thing, any improvements available to cars will be available to buses as well. And nothing is going to change the simple geometric advantages that transit enjoys in dense areas.

Predicting the future is hard. If you’re out there predicting the doom of the car every year like James Howard Kunstler or the dominance of self-driving cars in 2020 like Randal O’Toole, you’re probably going to end up looking foolish.

A conservative approach would be to continue investing in cost-effective transit improvements, including CBTC and driverless technologies, where warranted. Automated car technology should, at the very least, result in a considerable drop in the number of people killed and injured by cars, and for that alone, it should be welcomed.

Update: Platooning

A quick note on platooning, which is the idea of having driverless cars follow each other very closely, perhaps only inches apart. This would greatly increase road capacity, but it would absolutely require vehicle to vehicle, and perhaps vehicle to central control, communication, as opposed to passive information collection systems currently being used by the likes of Google. I think that getting such a system operate reliably and safely will be more difficult in practice than many people expect. I don’t think we’ll be seeing it any time soon.

Sepulveda Pass Transit, Prelude to Part 3: Should Cars Be Part of a Tunnel Through the Pass?

The next installment of Sepulveda Pass & LAX Transit is going to look at the pass itself. In the opening post to the series, I dismissed the idea of tunnel lanes for cars offhand, as impractical due to lower capacity and increased costs for things like ventilation.

I wanted to give justifiable reasons for leaving cars out of the mix in the post detailing the section through the pass. In doing research on freeway tunnels, it became apparent that this task would include so much material that it would detract from the structure and flow of the post on the pass itself. That post is intended to focus on tradeoffs between bus and rail, and alignment alternatives. So, I decided to break out the detailed look at highway tunnels as an appetizer.

Context for a Sepulveda Pass Freeway Tunnel

A tunnel following the alignment of the 405 from Wilshire to the 101 would be almost 8 miles long, making it the 10th longest roadway tunnel in the world, though four longer tunnels are currently under construction.

The rarified company for such a tunnel consists entirely of Alps tunnels, Norwegian single-tube two-lane tunnels, and Chinese freeways through mountainous areas. There are a few good analogs for a long, urban freeway tunnel: the A86 in Paris (6.2 miles), the M30 in Madrid (6.2 miles, though not continuous), and the Yamate Tunnel on the C2 Shuto Expressway (6.8 miles). For other urban highway tunnels, we have Brisbane Airport Link (4.2 miles) and Sydney’s Lane Cove Tunnel (2.2 miles).

If these sound suspiciously like the same countries have much lower per-mile transit construction capital costs than the US, well, more on that later.

These tunnels basically fall into two categories: tolled urban tunnels, and mountainous national network highway tunnels. (The M30 is urban, but free.)

National Highway Network Tunnels

The mountainous national network highway tunnels generally offer an enormous improvement over previous routes. They may also be what we might call “transportation equity projects” in the US, by which I mean projects that probably do not make sense on a cost-benefit basis, but that the nation feels obligated to build in order to provide high speed transportation to more remote regions. These tunnels all face geologic challenges related to deep rock tunneling, including areas with thousands of feet of overburden (depth of rock above).

The Alps tunnels fall into the former category. This includes the St Gotthard (10.5 miles), Arlberg (8.7 miles), Frejus (8.0 miles), and Mt Blanc (7.2 miles) tunnels, among others. Some of these tunnels have tolls, which can be quite high – for example, Arlberg is 9 euros, while Mt Blanc is 41 euros. These tunnels bypass treacherous mountain routes and form critical economic links between France, Switzerland, and Italy.

The Norwegian long tunnels are equity projects. The two longest tunnels are on the E36, which connects Bergen to Oslo. Both the Laerdal (15.2 miles) and Gudvangen (7.1 miles) tunnels are single bore, one lane per direction. Even granting very favorable geology, the construction costs for the Laerdal Tunnel are impressively low – $157m in 2014 dollars, or just over $10m/mile. However, the traffic volumes are also very low, just 1,000 vehicles per day. If you assume 30 year bonds and 5% interest, that’s about $15 per vehicle to recover costs – and remember, the route has several tunnels.

The Chinese freeways offer huge improvements in travel time. Many have speculated that China is overbuilding infrastructure; I honestly have no idea if the G65 route from Xi’an to Chongqing warrants a freeway, though Ankang and Dazhou are sizeable cities. The use of tunnels on the G65 seems gratuitous; a cursory look at the terrain averted by the 11.2 mile long Zhongnanshan Tunnel  suggests that the freeway could have easily climbed higher into the mountains and used a shorter tunnel. On the other hand, when you’re rolling in capital and you can throw down dual-bore two-lane freeway tunnels for $42m/mile, why not?

Taiwan’s longest freeway tunnel, the 8.0 mile dual-bore Xueshan Tunnel, is both a significant route improvement and an equity project. It connects Taipei to the relatively undeveloped eastern side of the island (Yilan, Hualien, and Taitung combined are barely 1 million people). The project took 15 years to complete, at a cost of $3.25b or about $405m/mile.

Urban Toll Tunnels

In contrast, the urban toll tunnels are all shallow tunnels, mainly through sedimentary deposits, with some shallow bedrock (limestone in the case of the A86, tuff for Brisbane Airport Link). These projects face challenges related to navigating around and protecting existing infrastructure and buildings.

The A86 is a dual-bore tunnel in an innovative configuration that takes advantage of the ability to increase capacity by restricting one tunnel to low-clearance passenger vehicles. One tunnel is two decks with two lanes each way for passenger vehicles; the other is a single lane in each direction and permits trucks. The project was completed in 2011 for 2.2b euros, about $2.94b or $474m/mile. (Note, if you believe LA Metro’s cost estimate of about $5b for the ~5 mile dual bore tunnels for the 710, that puts French highway tunneling costs at about half of American costs.) Tolls are about $3-$10.

Costs for the M30 freeway tunnels are, like Spanish transit costs, almost unbelievably low. The project built 99 kilometers (61.5 miles) of freeway, including 56 km (34.8 miles) of tunnels, for 3.9b euros. The entire project was completed in just under 3 years, between September 2004 and mid 2007. Adjusting to 2014 US dollars puts the cost at $6b, or about $98m/mile. Costs for different tunnel segments are summarized below.


So for cut and cover, Spanish costs are around $100m/mile. For bored tunnels, about $225m/mile. As stated previously, the M30 is toll free.

The Yamate tunnel  was completed in 2007 and is approximately 100 feet deep, with 70% bored. I could not find a cost for the project. The toll is about $9. The tunnel is part of the Shuto Expressway network, which was privatized in 2005 as the Metropolitan Expressway Company. The roads and debts are owned by a separate holding company.

Less encouraging for the proposition of a privately-funded tunnel through Sepulveda Pass is the experience of two privately built and operated urban toll tunnels in Australia.

The Brisbane Airport Link cost $4.8b, including a grand total of 9.3 miles of tunnel  among many other components. Tolls were expected to be about $5 by the end of 2013. The project was troubled from the very start; the bond sales went poorly and values collapsed. The Product Disclosure Statement forecast traffic volumes of 193,000 per day at opening, rising to 291,000 by 2026. In reality, traffic volumes were about 86,000 in August 2012, and 48,000 in February 2013, when the operator went into receivership. The traffic projections were so egregiously wrong that Arup, the consultant that prepared them, has been sued by investors.

Sydney’s Lane Cove Tunnel is much shorter than the others discussed here, but I’m including it as a representative toll tunnel project. The tunnel cost $1.1b, or $575m/mile in 2014 dollars, and was built by the same consortium as the Brisbane Airport Link. The tunnel is part of the M2 freeway, with a toll of about $3, and the only alternative is a surface arterial. Nevertheless, the operator went into receivership in early 2010, and Transurban bought the assets for just $630m.

The travel time savings and traffic projections claimed by the two Australian tunnel projects should have been suspicious from the start. The Brisbane Airport Link claimed it would attract up to 291,000 total entries a day; for reference, the George Washington Bridge, with its 14 lanes of traffic, carries about 275,000 vehicles a day. (Note: point volumes and total entries are directly comparable; the Brisbane Airport Link has an intermediate entry/exit point, so it’s impossible to say what point volumes are on the link). The Lane Cove Tunnel claimed it would save up to 17 minutes of travel time, which at 50 mph over its 2.2 mile length implies an average speed of just 7 mph on alternate routes.

As one more data point, for large bore tunnels, this document reports a range of costs, all much less than the approximate $1b/mile suggested for projects like Sepulveda Pass. (Note, the document also suggests that LA will be able to achieve French costs for the 710.)


Sepulveda Pass – Somewhere in the Middle

Our little pass between the LA Basin and the Valley certainly isn’t in the same league as the mountain tunnels, as a look at profiles easily shows.

Zhongnanshan-profile Xueshan-profile MtBlanc-profile Laerdal-profile Sepulveda-profile

(Note: open each image up in its own tab to see them at the same scale.)

It’s also not an equity project since we already have freeways through the Santa Monica Mountains. On the other hand, it’s longer and deeper than the other urban tunnels. I didn’t even bother creating profiles for those tunnels, since they wouldn’t register at that scale.

A real assessment of the geology is obviously well beyond my abilities. For now, we can just acknowledge that at the ends, tunneling would be through alluvial deposits like the other urban tunnels, but in the middle, it would be through the complex folded and faulted sedimentary rocks of the Santa Monica Mountains. The mountainous tunnels are all through igneous and metamorphic formations like gneiss, which tend to be extremely competent bedrock, except for the Xueshan Tunnel. This tunnel traverses heavily folded and faulted sandstone of varying quality, with the easternmost  2.0 miles being very poor. The high permeability of the ground traversed by the Xueshan Tunnel made groundwater an enormous problem. This tunnel might therefore be the best analog for the Sepulveda Pass tunnel.

Auto Tunnels – Technically Feasible, But Are They Worth It?

My initial instinct was to write off an auto tunnel through Sepulveda Pass as impractical. However, upon review of long freeway tunnels, it’s pretty clear that the tunnel would be technically feasible and that the costs are not preposterous. Let’s assume an A86-type setup, where one tube would be two decks for low clearance vehicles (i.e. passenger cars) and the other tube would be for transit vehicles.

If it can be built for Xueshan Tunnel type costs, it would cost around $3b. At 5% for 30 years, the tunnel would need to generate about $16m per month in revenue to cover capital costs. Assuming 20 weekday equivalents per month, that’s about $800k per day. If the tunnel attracts 100,000 users a day (about a third of the current traffic volume through the pass), the average toll would be $8, though obviously it would be varied at different times of the day, depending on congestion. This is in line with rates charged by Metro Express Lanes on the 10 and the 110 high-occupancy toll (HOT) lanes, which max out at $1.40/mile, or about $11 for Sepulveda Pass.

However, the overall numbers from Metro Express Lanes aren’t encouraging for a Sepulveda Pass auto tunnel. Net revenue over the pilot year of the program was $19m – certainly making the project able to cover its own $210m capital cost (if it were required to do so) with enough money left over to spend on expansions or other improvements. Unfortunately, that’s an order of magnitude short of what would be needed to fund capital costs for a Sepulveda Pass tunnel, and it’s being collected over almost three times the distance.

The travel time math isn’t hard to do. At 60 mph, a tunnel would get you from Westwood to Sherman Oaks in 8 minutes. Even if the freeway were only averaging 20 mph, the tunnel would save you 16 minutes. Are there 100,000 people going through Sepulveda Pass every day that value their driving time at $30/hr? The experience of the Australian toll tunnels should make us think hard about that.

It seems likely that the finances for a tolled auto tunnel don’t pencil out, unless the project can be delivered for considerably lower capital costs. The topography of the pass is really not challenging for passenger vehicles. If express HOT lanes through the pass are desired, the first alternative should be converting existing lanes. If that is not possible, any new lanes should probably be built on the surface through further freeway widening, or as an elevated facility like those on the 110.

Walkabout: Naebang, Seoul

If you’re a politician, a big fancy consultant, or a global thought leader, what to do you do when you visit a major city in another part of the world? You probably fly into Shiny New Airy Spaces International Airport, hop on the Global Thought Leader High-Speed Rail Line, and head downtown where you visit a bunch of places with high-end amenities and architectural showpieces like stadiums and convention centers.

Then you come home, pass through an old, congested airport terminal, and stand on the curb breathing exhaust while you wait for a car service to pick you up, since very few US airports have high quality rail connections.

Any wonder that US politicians and global thought leaders are obsessed with airport transit and convention centers?

The problem is that as an international traveler, the way you experience a city’s transit system is very different from the way that ordinary people experience it. Most people riding transit aren’t going to the airport or convention center or stadium very often (unless they work there). The vast majority of transit riders are trying to go about some mundane daily business – going to work, school, or shopping – in the places that the system serves.

So, what should you do if you want to experience a city the way residents do? Easy, just get out a rail map, pick a random stop, get off there, and start walking around. If I have time and the system isn’t too expansive, I like to try to pick one midline stop and one suburban terminal. (In theory, this works just as well – if not better – with buses, but it’s often harder to figure out the frequently and span of service for bus routes on an unfamiliar system.)


You want me to just pick a random stop that I don’t know anything about? Yes. In fact, this really doesn’t work in a city you know, since you probably have preconceived notions about the neighborhoods around most of the transit stops. The goal is to learn things you wouldn’t learn in a structured exploration.

What if I pick a boring neighborhood with nothing to do? Even better; you’re learning about a part of the city that the tourist bureau doesn’t want to show you.

What if I don’t speak the language? Even better! Now you’re experiencing a city the way recent immigrants might.

What if I pick a dangerous neighborhood? This is a potential concern. In general, many Americans (especially a certain generation that rhymes with Maybe Doomers) overestimate the level of danger in cities. Go during the day when it’s light out, and just act naturally.

What if I get lost? You have a smartphone, don’t you? Before you head out, scroll around the neighborhood in your maps app to load a bunch of data into the cache (phone memory) so that you can see where you are even without cell service. Also, there seems to be a lot more free WiFi in some countries, and you can almost always find a local restaurant or coffee shop that has WiFi nowadays.

If the stop spacing on the rail network is reasonable – like say, a mile or so – you can plan to walk between two stations, which helps give you a feel for what the city is like a little further from a transit node.

Okay, enough introduction. Here’s what I saw when I had a free day in Seoul, and a random jab at the map led me to Naebang.

Naebang Station

Seoul is a very dense city, so no matter where you go, you’re going to see some urban environments. I exited Naebang and headed east on Seocho-daero. (Note: streets in Seoul are often just numbered sequentially with a local neighborhood name, which can be a little confusing.)

Heading uphill, there’s this Flatiron-esque building on a triangular lot.


This street is actually pretty wide an auto-oriented, like many modern arterials in Seoul. Neighborhood streets are a different story, as we’ll see later.

Seocho-daero leads you uphill and dead-ends; look back and you’ll be rewarded with some views. Urbanization in Asia is far more advanced than anywhere in the US outside Manhattan.

02-lookingback 03-lookingbackzoom

In the distance, you can see some green hills. Seoul’s dense urban environment is frequently broken up by hills, many of which form nice neighborhood parks. In a way it’s a little like. . .  Los Angeles. Is Seoul a glimpse of a future denser LA? A thought for another time.


Turning around, it’s time to head up into Seoripul Park.


There are some nice, graded paths, but why not scramble up some rocks when you can?


Just a little ways up, I reconnected with the main path through the park.


Seoul is dry in the winter, but has a summer monsoon that dumps a lot of rain. That keeps things very green in the hills.

08-thruleaves 09-overtreetops

So. . . dense. . . and yet the parks provide a real escape from the city.


Alright, now we’re back out into some steep streets fronting up onto the hills. I like the merciless use of overhead power distribution. Also, it does snow in Seoul in the winter, pretty regularly. Next time I go, I want to go in winter and see how people navigate those hills in snow.


Off of the arterials, you can see part of how Seoul achieves such density – really narrow side streets.


Of course, all those towers help too.


I kept walking downhill. Narrow streets and mid-rise development abound.


And hey, I was able to find where this is in Google Maps!

I kept wrapping around the hill on the street, into a neighborhood that looked a little more upscale. Nice sidewalks!


The building on the left was completed after Google street view’s last pass.


Seoul – not scared of funky design.


Oh, and one more thing: all those mid-rises in the upscale neighborhood? They’re all podiums with parking on the ground floor. Ha!