Rail Transit

Basics: Why Aren’t the Buses Timed to Meet the Trains?

Short answer:  Because the buses are timed to meet each other, and this is harder than it looks.

Long answer:  If you’ve used public transit in an area that has infrequent trains, including the suburbs of many cities, you’ve probably wondered why the bus and train schedules aren’t coordinated.  Why didn’t they write the bus schedule so that the bus would meet the train?

First of all, let’s gently note the bias in the question.  Why didn’t you ask why the train wasn’t scheduled to meet the bus?  We assume that because trains are bigger, faster, and more rigid, they are superior and buses are subordinate. You’ll even hear some bus routes described as “feeders”, implying that they have no purpose but to bring customers to the dominant mode.

But it’s rare for an efficient bus route to have no other purpose than feeding the train. Public transit thrives on the diversity of purposes that the same vehicle trip can serve.  At a busy rush hour time, you may encounter a true feeder bus that’s timed to the train and will even wait if the train is late.  But most bus services carry many people locally in their area, on trips that don’t involve the train connection.  For these networks to work, they have to connect well with themselves, and this is harder than it looks.

We’re talking here about infrequent bus routes (generally every 30 minutes or worse) and infrequent trains.  When frequency is high, no special effort is needed to make the connection work.

Pulse scheduling. Buses of many lines are coordinated so that buses meet at the same time each hour, allowing fast transfers despite low frequency.

Infrequent transit networks have a huge problem.  There’s not just a long wait for the initial bus or train.  There’s also a long wait for any connection you may need to make to reach your destination.   We often combat this problem with pulse scheduling.  At key hubs, we schedule the buses to all meet at the same time each hour or half hour, so that people can make connections quickly even though frequencies are low.  We design the whole network around those connections, because they are so important to making the network useful.

That means that the whole schedule has to have a regular repeating pattern.  As much as possible we want this pattern to repeat every hour, so that it’s easy to remember.  We even design route lengths to cycle well in this amount of time, or multiples of it.

If the train schedule has a similar pattern, we will certainly look at it and try to match our pattern to it.  But the timing of a pulse determines the schedules of all the routes serving that point.  Sometimes we have lattices of interacting pulses at several points, which can make an entire network interdependent.  You can’t change any of these schedules without changing all of them, or you lose the fast connections between infrequent bus routes that makes suburban networks usable.

Sometimes, an infrequent trunk train service will also present a repeating hourly cycle in its schedule, and if so, we’ll look at that and try to coordinate with it.  But at most this will be possible at a couple of stations where the timing works well, because of the way the local bus schedules are all connected.

More commonly, especially in North America, we face an irregular regional rail or “commuter rail” schedule, where there may be a regular midday pattern but there’s often no pattern at other times.  The pattern may often shift during the day for various reasons that make sense for the train operation.  All this is toxic to timing with the local bus network.  Local bus networks need that repeating hourly pattern to be efficient and legible.  For example, if at 1 PM the train pattern suddenly moves five minutes earlier, the bus network can’t adapt to that without opening up a gap in its schedules that will affect lots of other people.

Usually, the regional rail network and the local bus network are part of different transit authorities, which makes this an even bigger challenge.  A particular problem in multi-authority region is that different authorities may have different schedule change dates, sometimes baked into their labor agreements, and this prevents them from all changing together at the same time.  But the core problem isn’t just institutional.  Merging the authorities won’t solve it. No efficient bus system – working with sparse resources and therefore offering infrequent service – can make timed connections with a train schedule at every station, and especially not if the train schedule is irregular.  It’s just not mathematically possible.

The best possible outcomes happen when the rail and bus authorities have a relationship that recognizes their interdependence rather than one based on a supposed hierarchy.  That means that the rail authority recognizes that the local bus authorities can only connect with a repeating hourly schedule pattern, and tries to provide one.  It also means that rail schedule changes are made with plenty of warning so that there’s time for bus authorities to adapt.

With the decline of rush-hour commuting due to increased working from home, transit demand is even more all-directions and all-the-time.  It no longer makes sense to just assume that one trip – say, the commute to the big city – is superior to another, like the local trip to a grocery store or retail job.  All possible trips matter, and we get the best transit network when authorities coordinate to provide the best possible connections for all of them.

 

 

 

 

 

 

 

Los Angeles: A Major Expansion of Rail Transit Access

Los Angeles’s Regional Connector is open.  It’s a small piece of subway — less than two miles, with three stations, but it utterly transforms the rail network of Los Angeles, making trips across downtown much faster.  Here’s LA Metro’s quick diagram of the change:

It’s simple: In the network as it existed until this weeekend, light rail from north and east, called Line L, only came to the northeast edge of downtown, while the two lines from the south and west (Lines A and E) only came to the southwest edge. Traveling across the center thus required making two transfers, using the Line C & D Subway.  The Regional Connector rearranges these lines to that all services flow across downtown and out the far side. Much faster trips across downtown mean greater access to opportunity for many people across the city and beyond.

I’ve written before about the difficulty of getting regional focus on these core-city projects.  LA Metro did a good job with this one, starting by branding it the Regional Connector.  It may be in downtown Los Angeles but it’s not for downtown Los Angeles.  It’s for the entire region.

When Buses Are Free but Trains Aren’t

I’m in Bloomberg CityLab with a piece on the dangers of applying free fares to buses but not to trains in the same city.  Key quote:

When we encourage people to get off trains and onto parallel buses, both kinds of transit lose. Buses are smaller, so they run out of capacity at a lower level of ridership. This requires the transit agency to put out more service — buses that could have been used in areas away from train lines where they provide the only mobility.

I go on to challenge the “buses are for poor people” assumption, which both equity advocates and wealthy elites often agree on.

To many, buses and trains symbolize positions in the class struggle. Equity advocates and wealthy elites often agree on this view. The equity advocate will say that we need to focus on buses because poor people use them, while I have heard elites argue (often in private) that we should neglect buses for the same reason. Most obviously, when developers and other elite urbanists argue that transit-oriented development requires rail, they are understandably privileging the view of people who are in the position to buy market-rate urban real estate. Those fortunate folks are especially likely to say, openly or not, that they would never ride a bus.

But in an urban transit network that’s trying to give everyone the greatest possible access to destinations, rail and bus services work together, and many trips involve both. You use a train for longer trips along corridors with high demand. To travel in lower-demand areas, or to make many shorter trips, you take a bus. Sometimes you ride the subway for part of the trip and a bus for the rest. An efficient and therefore liberating urban transit network encourages people to think about the total network, and to use buses or trains according to which is better for each part of their trip.

Anyway, the whole piece is worth a look.

Why Are US Rail Projects So Expensive?

We’ve known for a long time that the US pays more than most other wealthy countries to build rapid transit lines, and especially for tunneling.  If the incoming Biden administration wants to invest more in transit construction, then it’s time to get a handle on this.

The transit researcher Alon Levy has been working on this issue for many years, has generated a helpful trove of articles is here.  Alon’s work triggered a New York Times exposé in 2017, focus on the extreme costs (over $1 billion/mile) of recent subway construction there.

But while the New York situation is the most extreme, rapid transit construction costs are persistently higher than in comparable countries in Europe, where they are tunneling through equally complex urban environments.

Now, Eno Foundation has dug into this, building a database of case studies to help define the problem.  Their top level findings:

  • Yes, US appears to spend more to build rail transit lines than comparable overseas peers.
  • This difference is mostly about the cost of tunneling, not surface lines.  The US pays far more to tunnel 1 km than Europeans do, even in cities like Rome where archaeology is a major issue.
  • Needless to say, the type of rail doesn’t matter much.  Once you leave the surface, either onto viaducts or into tunnels, any cost difference between light rail and heavy rail is swamped by the cost of those structures.  (This is true of bus viaducts and tunnels too, of course)
  • Remarkably, stations don’t seem to explain the difference in rail construction costs.  European subways with stations closer together still come out cheaper than US subways with fewer stations.

Most of us have known this for a long time — though I admit to being surprised by the last point.  But it’s good to see a respected institute like Eno building out a database to make the facts unavoidable.  If you want more rail transit in the US, it simply has to be cheaper.

San Francisco: A Forbidden Fantasy Comes True

Around 1989, when I lived in San Francisco, I spent too much time in little rooms with transit advocates (and some transit professionals who could not be named) complaining about Muni Metro, the combined surface-subway light rail system.  It looked like this and still does, except that the T line was added more recently.  Note the r0ute letter names in the lower left.

The segment with 3-5 lines on it, from Embarcadero to West Portal, is the underground segment, which carries the heaviest loads through the densest part of the city.

It had always been wildly unreliable.  The five lines that ran through it (J, K, L, M, and N) always came in sequences of pure arithmetic randomness: N, J, M, K, J, N, N, K, K, M, N, J, L.  (Finally, my “L”!  But of course, after such a long gap, it’s crush-loaded and I can’t get on.)

Four decades after the subway opened, lots of things have been fixed: longer and better trains, better signaling, an extension downtown that helped trains turn back more efficiently.  But none of this touched the true problem:  The core Metro subway carries five lines, all of which deserve to be very frequent.  But they can’t all be frequent enough because they all have to squeeze into one two-track subway.  The other part of the problem is that they all have surface segments at the outer end, where they encounter more sources of delay, causing them to enter the subway at unpredictable times, and in an unpredictable order.

In those small rooms in the 1980s, we all knew that there was only one mathematically coherent solution.  Some us drew the map of this solution on napkins, but we really didn’t need to.  The map was burned into our minds from our relentless, powerless mental fondling of it.  Of course it was politically impossible, so impossible that if you valued your career you would wad up that napkin at once, burn it probably, and certainly not mention it outside your most trusted circle.

At most you might let out the pressure as a joke: “You know, we *could* turn the J, K, and L into feeders, and just run the M and N downtown. And then we’d have room for a line that just stayed in the subway, so it was never affected by surface delays.”  Everyone would titter at hearing this actually said, as though in some alternate universe such a change could be possible.

Now, the impossible is happening.  Without fear or shame, I can finally share the content of that forbidden napkin, because it looks like San Francisco is actually going to do it.

 

The two busiest western lines (M, N) will still go downtown, the others (J, K, L) will terminate when they reach a station but you have to transfer to continue downtown.  M trains will flow through as T.  Finally, a shuttle (S) will provide additional frequency in the subway, immune to surface delays.  As always, asking people to transfer makes possible a simpler, more frequent, and more reliable system.

You may detect, at San Francisco’s tiny scale, a case of the universal “edge vs core” problem.  Like many, many US rail transit systems, Muni Metro had been designed to take care of the edge, people who lived on one of the branch lines, rather than the core, people traveling along the subway in the dense inner city.  The new system finally fixes the core. But the edge folks benefit from a reliable subway too.  What’s more, in the future it may be possible to run the surface segments of J, K, and L more frequently, because their capacity will no longer be capped by the need to fit down the subway with four other lines.

All that in return for having to transfer to go downtown if you’re on the J, K, or L.

Let me not make this sound easy.  These transfer points, West Portal and Duboce Portal, are a little awkward, because they were never designed for this purpose.  You have to walk from one platform to another, crossing at least one street.  There are valid concerns from people with mobility limitations, which will have to be addressed with better street and intersection design.  Plenty of people won’t like it.

But the transit backbone of a major city will finally function.  And for those of us who’ve known San Francisco for decades, that’s a forbidden fantasy come true.

 

 

The Fishing Pier Problem in Public Transit Equity

Photo: Heditor6, Wikimedia Commons

Have you ever heard people say: “They got that transit improvement in their neighborhood!  We deserve to have one in our neighborhood!”

But does this demand always makes sense?

Imagine a city on a lake or ocean, where neighborhoods near the water tend to be wealthier than those inland.  Suppose the city has a plan to build several fishing piers, but all the proposed piers are in the wealthy neighborhoods on the waterfront.  Isn’t that unfair?

No, I think you’d say, because fishing piers only work if they’re on the water.  If you were concerned with equity, maybe you’d propose a program that helps inland people get to the waterfront fishing piers quickly.  But you wouldn’t support an inland elected official’s battle to get a fishing pier on dry land in their neighborhood, because it wouldn’t be useful for fishing.

In short, the point isn’t to equitably distribute fishing piers.  It’s to equitably distribute the ability to fish.

In the transit business, when a cool new thing is created somewhere, you always hear the rest of the city say: when do we get that cool thing?  You’ll hear this about everything, from subway lines to Bus Rapid Transit to little vans that come to your door.   Enormous amounts of money get spent trying to act on this principle.

The typical pattern goes like this:

  • Cool new transit thing x is introduced, and deployed in one or two places in the city where it makes sense.  (Let’s build a fishing pier on the waterfront!)
  • Other neighborhoods demand the same thing, often claiming it’s unjust or inequitable that they don’t have it.  (Why don’t our inland neighborhoods have fishing piers?)
  • Often the cool new thing is actually built in those other neighborhoods that demand it, but it doesn’t work well there, because the geography is wrong for it.  (A fishing pier is actually built inland, extending across a patch of grass, but nobody uses it.)

The marketing of cool new transit things can make this problem worse.  The more you put out the message that light rail or BRT or microtransit or “Metro Rapid” is cool and different and better than “ordinary” buses, the more mad people will be if their neighborhood just gets ordinary buses.  That leads to political pressure to bring the cool new thing to a place where it just doesn’t work very well, which in turn leads to the cool new thing failing, just as an inland fishing pier will fail.

You’ll get the best transit mobility if we use the tool that works with your geography, even if it’s different from what works in other places.

So perhaps it makes no sense to equitably distribute any cool transit thing.  It makes sense to equitably distribute the ability to go lots of places quickly on transit.

How would our transit debates be different if we did this?

 

 

One Less Barrier to Expanding US Urban Rail Transit

 

Caltrain between San Francisco and San Jose is one of many urban “commuter rail” lines that really need to be high frequency rapid transit lines. Now that’s a little more likely. Photo: Lucius Kwok.

Here’s some good news for people who want more rapid transit service in US cities, and soon.

In the US, all passenger rail services that could potentially mix with freight are governed by the regulations of the Federal Railroad Administration (FRA). This applies not just to Amtrak but also, critically, to “commuter” rail lines, crucial rail transit services that run on freight railways.

If cities wanted to rapidly upgrade their rail transit systems, the cheapest way is often using upgraded commuter rail rather than building new lines.  Many major cities have large networks of radial commuter rail lines [typically originally freight lines] which, if upgraded to run every 15 minutes or better all day, would effectively become metro lines, on the cheap.  You’ll find this level of service in many major metro areas overseas. Toronto’s Smart Track plan is exactly this idea.

The problem, as always, is frequency, which in turn is a problem of operating cost.  Most US commuter rail systems are far too infrequent to be useful for anything but 9-5 commuting, even though many of them run through dense urban fabric where the demand is there for all-day frequent service.

The Obama FRA, responding to several freight rail disasters, had proposed a rule mandating two-person crews, and had quietly inserted language extending this to passenger rail, even though passenger and freight rail present very different safety issues.   Those requirements would have made commuter rail service too expensive to run frequently enough for it to be useful, and would have persisted regardless of whether technological developments improved the safety outcomes of one-person crews.

The Federal Railroad Administration has just announced that it will stop requiring two-person crews and preempt state requirements to do so.  If this were a genuine safety issues, I’d be alarmed, but it really isn’t. The new FRA position liberates transit agencies and other local governments to negotiate the right solution with their unions in the context of what’s technologically possible.

Yes, removing this requirement is a “conservative” idea that would be unlikely to come from a Democratic administration.  But it removes a significant barrier to providing more useful urban public transit, which leads to all kinds of benefits for equity, prosperity, and the environment.

Visualizing Transit Reliability

Reliability is one of those essential features of transit that you can’t take a picture of.  It’s an overwhelming issue in the lives of transit customers but can seem abstract to others who make transit policy.  And it’s a major issue in many transit systems.

Poor reliability of buses has many causes, mostly having to do with the traffic congestion and other causes of random delay to which they’re exposed.  But when a rail line runs in an exclusive right of way, never interacting with traffic, there aren’t a lot of excuses.

The Miami organization Transit Alliance has done a nice visualization of transit reliability on that city’s rail transit system.  It looks at the system right now and shows how many trains are running late.  It’s important to note here that late does not mean behind schedule.  It means that the maximum wait time is longer than scheduled, by a given number of minutes.  (That’s the only rational way to talk about reliability in high-frequency services.)

Some insitute really needs to create a database of reliability info across many agencies, searchable many ways — and always based on this headway reliability rather than on-time performance.  Most transit agencies now have real time vehicle location feeds, and they are already released in standard formats for use by apps.  Yet we see remarkably few of these kinds of analytics that could help people understand the severity of the reliability problem — even on services like grade-separated heavy rail that have few external causes of delay.

 

 

The Pleasure of Track Maps

If you’ve never seen a subway track map, I suggest you look at this one, for New York, by “radical cartographer” Andrew Lynch. Most track diagrams are not to scale, and look like they’re meant to make to make sense only to insiders.  But this one is beautiful.
nyc track map b

 

 

What’s more, it’s accurate in geographic scale, though of course the separation of tracks can’t be on the same scale as the network.  Still, New York’s subway is both huge and full of details, so this is no mean feat.  Only 22 insets were required, to zoom in on tricky segments.

Gazing at a good track map can give you an appreciation for the heroics involved in moving trains around in this limited infrastructure. Switches and extra tracks are very expensive underground, which is why they are never where you need them to handle a particular incident.  This, for example, is why a track closure at one station may continue through several stations nearby.

Gaze at this piece of the Bronx, and marvel at what a train would have to do to get from Jerome Yard to a station on the Orange (B+D) line.  I presume they don’t have to do this very often, but in a pinch, they can.

nyc track map a

I spent a delighted hour with it.

Portland: Frequent Bus Performance Approaching Light Rail’s

Here's an interesting chart:

Tri Met Ops Cost per Ride

This is a year's trend comparing bus and light rail (MAX) service in Portland's transit agency, TriMet, from the performance dashboard at the TriMet Transparency and Accountability Center webpage.  

The metric here is operating cost per boarding ride.  This is a good overall measure of how effectively a transit agency is liberating and moving people, where down means good.  (I prefer this ratio upside down: ridership per unit cost or "bang for buck," so that up means good. but this is obviously a chart by finance people who always want cost on top.)  This is a "macro" metric.  Practically everything a transit agency does affects it, so it's lousy diagnosis but not bad if you only have bandwidth to convey one measure.

Most American transit data just compares bus and rail, and inevitably shows bus performing worse.  You'll see that here too if you just look at the wide solid lines.  From this we get endless ignorant journalism lamenting the poor performance of the city bus, as though all city buses are basically alike.

What if we separated out highly useful and liberating bus service as a separate category?  That isn't exactly the distinction made here but it's close.  TriMet's Frequent Service network (still being restored, but mostly now back in existence) is the network of all services that are almost always coming soon.  

This chart says two remarkable things:

  • Frequent bus performance is now very close to light rail performance.   
  • The spread between Frequent Bus and infrequent bus is usually bigger than the spread between all buses and light rail.

The lesson is pretty clear:  The "city bus" is a misleading category, and the much-fetishized difference between bus and rail may matter less than whether the services are designed to be useful.  And when it comes to usefulness, no one variable capture that more than frequency.