Whenever you hear someone cite the travel time of a proposed transit line, your first reaction should always be: “Yes, but at what frequency?” Often, that fact is missing from these soundbites.
There’s a nice example in today’s Transport Politic. Speaking of the proposed Gold Line Foothills extension, which if built will someday extend from Los Angeles to Montclair:
The commute times from the end of the [proposed Gold] line will be a serious problem: once the second phase is built to Montclair, downtown will be a full 75 minutes away, making daily commutes difficult to envision for many people. Even in traffic, that trip takes a total of 70 minutes by car.
Nevertheless, getting people from the San Gabriel Valley into downtown may not be the major goal of the project. Metrolink Commuter Rail can cover the distance between Montclair and Los Angeles in an hour along the San Bernardino Line, though that link is near carrying capacity.
Now, if all you care about is commutes during the peak hour, then and only then do you have all the information you need. Only during the peak commute period, and in the peak direction, can you expect anything called “commuter rail” to serve you anywhere near your desired time of travel.
However, if you care about supporting denser and more urban redevelopment, or helping people own fewer cars, or being relevant to the lives of service workers, or serving students who travel all day, then you’re missing a crucial fact: the proposed all-day frequency.
Frequency is best described by headway, the elapsed time between consecutive trips on a line. The headway is also the maximum waiting time, and half of it is the average waiting time. Add those to the much-advertised travel time (technically called in-vehicle travel time or IVTT) and you have a sense of how long a real-life trip will be. Those are the realities that will ultimately drive ridership.
This makes all the difference in the comparison above. The Gold Line can be expected to run every 15 minutes or better all day, which is pretty much the industry standard for light rail, though some agencies run less service during the late evening. Meanwhile, the Metrolink commuter rail line in this comparison runs every 60-180 minutes midday.
To get a quick sense of how useful a service will actually be to everyone, not just peak commuters, I always check a sample trip around 12:00 midday. The commuter rail headway is 120 minutes at that hour, so for an average trip time I take half of that and add it to the advertised in-vehicle travel time:
Average travel time = Headway / 2 plus in-vehicle travel time
Commuter Rail: (120 minutes)/2 + 55 minutes = 115 minutes.
Proposed LRT: (15 minutes)/2 + 75 minutes = 83 minutes.
Big difference. And that’s just the average. The maximum travel time is …
Maximum travel time = Headway plus in-vehicle travel time
Commuter Rail: 120 minutes + 55 minutes = 175 minutes.
Proposed LRT: 15 minutes + 75 minutes = 90 minutes.
North American commuter rail lines are typically very, very infrequent midday, if they run at all. The Metrolink line in question here (the San Bernardino line) is hourly in the mid-afternoon but has a two-hour gap mid-day (which affects the above calculation) and a gap of almost three hours in the morning. These gaps happen even on a line that is “near carrying capacity” because Metrolink shares its tracks with other uses, including freight and the occasional Amtrak train, and is also constrained by single-track segments. So it tends to flow a lot of service one way during the peak but has trouble serving reverse-peak or midday trips.
The real lesson here, of course, is that when you hear someone cite a commute-hour in-vehicle travel time for a proposed service, you should ask: “Is the commute all that matters?” Journalists often think so. But if you’re taking a longer view, and wanting a project to (a) be part of an integrated network and (b) be useful all day so that (c) it can trigger redevelopment and (d) enable people to own fewer cars, then the commute-hour travel time should be worthless to you. Ask instead about the travel time around 12:00, and always remind them that in the real world, travel time includes wait time. So what’s the frequency?
I think you’re right that it’s essential to include mentions of frequencies when you’re discussing travel times.
That said, I won’t take back my argument that the primary goal of this rail extension — if there is one — shouldn’t be to connect far-off places like Montclair to downtown L.A. If that’s your priority, you’d be better off improving the commuter rail system so that it is the one providing every 15-minute frequencies, rather than investing in the very expensive effort of building a new light rail line.
On the other hand, if your goal is to connect places like Montclair with places like Pasadena, the new light rail line makes a whole lot more sense, simply because it’s filling a new, previously uncovered travel corridor. The question is whether the market is big enough to make light rail justifiable here.
Of course average waiting time is half the headway when you’re looking at random arrivals at a train station. When headways exceed 10 or 15 minutues, passengers tend to observe timetables and time their arrival at the station a few minutes before the train’s departure. You seem to include ‘schedule delay’ as part of waiting time, but I don’t know if people expect to be able to undertake long journeys at the drop of a hat without reference to timetables.
I don’t know the nature of development along the commuter line, but if there was a string of strong shopping and employment centres along the line, this would be more likely to generate the sort of spontaneous to and fro demand throughout the day that would make frequent service more viable across a wider span of hours.
That Gold Line extension also includes a service facility and storage yard in Irwindale*. That will help deployment times by putting cars out closer to the terminus.
*Per a Pasadena Star-News article – “Gold Line panel rejects Valley voting challenge” (Mary Bender, 26 Aug. 2004) retrieved via ProQuest.
Re : Frequency
I was in San Francisco to do some shopping yesterday (27 Mar.’10). I chose the direct option of waiting for a less frequent (15 min.) #44-O’Shaughnessy at the Forest Hills Stn. Your classic, little-old-lady-out-shopping crossed the street from the station building. I volunteered the information that she had just missed a #36-Miraloma headed north. She sort of shrugged and pulled out something to read. The problem is that the #36 is a community service route with a thirty (30) minute frequency on weekends.
My trip got delayed due to one or more missing #44’s. I wound up taking a #43-Masonic to Masonic + Geary and transferring to a #38-Geary (very frequent service). This was one of those connections that Jarrett likes to talk about.
Metrolink timetables are so full of unpleasant compromises! First of all, there’s always a gap in service after the morning rush. That’s because they need some time to get trains to and from the maintenance facility, which is near Downtown, so a lot of maintenance actually happens during the day, between the peaks, because that’s when trains are actually there. The new Eastern Maintenance Facility would alleviate some of that problem, at least for the lines that it serves (which is four of the total seven). Then, there’s the almost complete absence of reverse-commute service on the San Bernardino Line: this is because much of the line is single-track, and there’s a single-track segment that’s too long to allow 20-minute headways in one direction while accommodating any service in the other direction. They’re working on eliminating some of those bottlenecks, but it’s slow going due to the generally slightly underfunded nature of Metrolink, especially in LA County.
I looked for confirmation of the single-tracking of the Metrolink commuter service. I found these articles :
“San Bernardino, Calif., Rail Service to Add Second Track.”
Press-Enterprise (Riverside, CA) (7 Aug. 2002)
“Plan for Gold Line to airport off track” (opinion column)
The Sun (San Bernardino, CA) (20 Oct. 2007)
The first describes the funding of a track upgrade project for the line. The second, an anti-Gold Line opinion column, mentions in passing that Metrolink San Bernardino line may still have a single track segment. I didn’t find an announcement of work completion other than what’s below.
Metrolink articles :
Preparatory work in Claremont (21 May 2003)
Work complete in Claremont (11 Dec. 2003)
Very frustrating site to search. What’s missing is a route-by-route clustering of service announcements and press releases.
An excellent point about frequency. This is a point often overlooked when debating between mode technologies.
When discussing this line the easiest assesment would be the Gold Line should have been Commuter Rail but the ridership would not pan out on a railroad corridor you’d have to rebuild, because you’re sacrificing frequency for demand.
Here’s a consolidated page that shows the connections between Metrolink’s San Bernardino line and the Gold Line’s extension (Phase 2B). This also answers the opinion column’s call for an airport shuttle bus.
You can look at the satellite view on Google Maps and you’ll see that the San Bernardino Line is mostly single track. The longest stretch of double track on the entire line is from Pomona to Montclair. The biggest single-track bottleneck is between Pomona and Covina, but there’s enough single track that reverse commute trains will either be very slow (due to waiting in siding) or reduce the capacity of the line in the peak direction. Some bottlenecks can be fixed fairly easily, but others, notably the long single track in the middle of the freeway, would be much harder and more expensive to fix.
All this talk of frequency, and still no mention of Kenneth? 🙂
FYI – Metrolink’s Eastern (Area) Maintenance Facility is located in Colton / San Bernardino (tricky border salient) roughly two miles west of the S.B. station. Off to the east is San Bernardino Int’l. Airport (ex-Norton AFB).
@anonymouse – Did those jokers grab a railroad ROW for their freeway ?
Sorry, it’s actually southwest based on the address I found at TrainOrders.
Generally is someone is going to only be on a single route. Whether that be on a train or bus. If the frequency is under 15 minutes. Most people will just walk out their door and catch the the next train/bus. Above 15 minutes people will start to look at the schedule. Then plan to leave to catch a certain scheduled train/bus.
Where frequency or headway plays a critical role is when someone needs to do a transfer or transfers along their journey. If the first route is running every 5 minutes but the second route only runs every 30 minutes. It doesn’t matter how often that first route runs. Your second route is what matters based on the low frequency. You just miss the transfer and your looking at a 30 minute wait.
First, the problem with the Gold Line is that the communities it serves are so far away that TOD is unlikely – the only direction in which transit would be even marginally competitive is toward downtown LA in the peak hour.
And second: okay, so LA is completely neglecting increasing commuter rail frequencies and instead building slow light rail lines at high cost. That’s not an improvement – it’s cost maximization.
Some quick notes on the comparison of these two services:
San Bernardino Line of Metrolink not only has single-tracking through two chokepoints (Pomona-Covina and the long stretch in the I-10 median west of El Monte) but its stations are mostly not in CBD locations, because except for Pomona, the cities served are mostly post=WW2 development and thus never oriented to rail. So a lot of the stations are isolated Park-and-Rides, not in good areas for TOD even if you could somehow achieve rail transit frequencies.
Gold Line Foothills extension will be much better on that score. Except for Duarte as noted, most of the towns served do have some kind of CBD or CBD remnant near the tracks, which could be a catalyst for further town-center redevelopment.
Also, Metrolink (S.B.) and the Gold Line are lining up to be complementary. I would not be surprised in about eight years (2017 + slippage) to see people heading outbound on the Gold Line to connect to an inbound Metrolink train. And has anybody published an estimate of the cost of double-tracking the line from San Bernardino to Union Stn. ? Even uglier is the prospect of all of the extra agencies that would get involved in making changes to I-10.
The I-210 traffic already shows that there is tremendous demand along that freeway corridor. On most mornings it backs up worse than I-10, and the carpool lane on I-210 is worse than useless. Each of the cities along the Gold Line has important job centers that are within a 10 or 15 minute walking distance from the train. For example, in Arcadia, you have Methodist Hospital and Santa Anita Park, for Monrovia you have the Huntington Drive industrial corridor, for Duarte you have City of Hope, a nationally recognized cancer care center, for Irwindale you have light industrial and office parks along Irwindale Avenue, and for Azusa you have Downtown Azusa and Citrus College/Azusa Pacific University and the new town of Rosedale at the last stop. Rosedale, in particular, is a hybrid walkable yet auto oriented development that is kind of New Urbanist lite, but it has been killed by the housing crash. When the Gold Line opens, that area will instantly fill up again.
Also, I highly doubt that many people will be taking the Gold Line daily to go to Los Angeles. Express bus service will likely continue to operate that can make it Downtown in 50-60 minutes, as a one seat ride. But the trip to Pasadena will only be 30-40 minutes on the Gold Line from APU, which is half the time of the Foothill Transit local bus today. Similarly, the number of people taking the Blue Line from end to end is very small, especially outside of commute hours, but it has tremendous utility for intermediate trips not served well by the express bus network.
Also, part of the issue has to do with Metrolink’s high fares, especially for short trips. The minimum Metrolink fare is roughly $4, whereas the Gold Line would charge MTA’s base fare. Even if some form of distance based structure were imposed on the Gold Line, because of length, your short trip would still cost no more than the base bus fare of $1.80 (effective 2012).
On the other hand, it is an open question whether cities like Glendora, San Dimas, and Claremont will embrace density to the same extent that Azusa and Monrovia have. These cities have very conservative politics and a strong historic preservation community, which may oppose additional density. Therefore, rather than immediately building the section to Montclair after the line to Azusa opens, it would be prudent to wait a few years and see what the development climate is in the further cities, and also see if parking demand is high at the end of the line stations. If the lot in Azusa fills up to capacity in the second year of existence, then that would be reason enough for an extension, if only to get to a place where they can put another parking lot.
Fares are based on service patterns. If Metrolink decided to evolve out of the steam era, it could cut its payroll and fare collection costs, run rolling stock that weighs 30 tons per car instead of 100, and avoid deadheading so many trains, all of which would allow for lower fares. Essentially, the problem is that the Gold Line is doing on a greenfield line what Metrolink could do on existing track at lower cost. The only difference: Metrolink would be able to offer more reliable express service.
The TOD issue doesn’t really mean much. This isn’t the Blue Line; none of the towns along the Foothills even approaches the importance of Long Beach, and there’s no anchor at the exurban end of the line. This line would work wonderfully as a loss leader for developer interests in those towns, most of which would be car-oriented. A hub-and-spoke transit pattern rarely works for anything other than CBD service, even in cities where the spokes are half as long as the planned line.
On par with Long Beach, There’s not one large anchor like Long Beach however there are a collective series of anchors that will help the off-peak directional ridership of this line.
However there are a few that are along this corridor on an existing right of way -to Azusa- that are growing and densifying around the current anchors that will be served with the Foothill extension.
Arcadia – Bus shuttle to Santa Anita racetrack and Arboretum close to redvelopment of Downtown Arcadia.
Monrovia – its engaging in aggresive redevelopment around the old Monrovia depot and Downtown Monrovia.
Irwindale station would function primarily as Park-Ride feeder station since it is surrounded by mostly industrial uses and can be easily developed into the large park-ride facilities that would help facilitate trips that are going toward Pasadena and Downtown LA.
Duarte – City of Hope Medical Research facility and Cancer Hospital that houses thousands of employees (over 3000 and growing)
Azusa – Azusa Pacific and Citrus College campuses would be served to this station that would link these students to jobs and access to Pasadena and beyond. Both schools have over 10,000 students and faculty on any given day both of these schools are engaged in an aggressive campus expansions.
The fact is that you’re not going to find a lot Montclaireans commuting to downtown LA. We may WANT to see that happen, but it doesn’t reflect reality, and instead it feeds into transit opponents main talking points. I’d rather see this money spent inside LA, improving transit in already dense neighborhoods that need something better.
Vancouver could exhibit a similar issue on a line to construct toward Coquitlam:
the suburb is connected to the CBD by a commuter train running only 5 times a day.
The region is planning to extend its Skytrain network to reach this city with high frequency service, albeit offering a longer ride than the commuter train (and with one transfer to DownTown)…
But the idea is that, as a commuter, you haven’t to worry so much to miss the commuter train…
if you do, not a big deal, you can still reach your destination in 40 mn instead of 30mn…
and may be shorter because the Skytrain line having more stop could bring you closer to destination.
the actual situation is not that great, and running more Commuter train requires a market which is simply not there. so the addition of the Evergreen line (equivalent to the Gold line extension) could be the best thing to do.
Look, I’m generally happy that the “light rail” of the last couple decades is reviving the old “interurbans” that fulfilled largely the same function, but I’m mad as hell that inner-city needs have been largely ignored.
Buffalo’s MetroRail is a very early and reasonably successful example–in light of the city’s otherwise decline–of an idea that should have been replicated across the country. The original idea in the 70s included what we would recognize today as a traditional suburban light rail, but the suburbs didn’t want it, so it got cut off at the city line and became more of an urban subway instead, running down the city’s main street (Main Street) and connecting with all the city’s crosstown bus line. It’s greatly improved local transit in the city–precisely where improvements were needed. There are wishful plans to extend lines to the suburbs as was originally planned, but the general poverty of the region–and the continuing distrust between city and suburb–means that it won’t happen.
Alon: how could Metrolink cut its payroll? There’s one engineer and one conductor per train, and you need the conductor because there has to be somebody to deploy the bridgeplate for wheelchairs. The ticketing system is POP, in line with European best practices. The only way to get rid of the conductor would be to get the CPUC to change the clearance requirements, then rebuild all the platforms for level boarding. This is very much outside Metrolink’s control at this point.
As for the Gold Line, keep in mind that the project has two distinct phases, one to Azusa and the second the rest of the way to Montclair. The former makes sense, both in terms of transportation and urban planning, as well as in terms of Metro’s operational needs (that maintenance facility in Irwindale). The latter does not make nearly as much sense, and is also much less likely to get built anytime soon. In fact, for quite a while Metro was opposed to even the first phase and refused to agree to fund the operations after it was built.
Does Metrolink run its own ROW, or is it subject to FRA regs?
Anonymouse, the engineer can operate the wheelchair lift himself. That’s how it’s done on New York City Transit buses, which are 100% accessible, 100% OPBO.
The conductor is there only because FRA regs require it.
“Fares are based on service patterns. If Metrolink decided to evolve out of the steam era, it could cut its payroll and fare collection costs, run rolling stock that weighs 30 tons per car instead of 100, and avoid deadheading so many trains, all of which would allow for lower fares.”
“Does Metrolink run its own ROW, or is it subject to FRA regs?”
It is shared with Freight track and rights-of-way which makes them subject to FRA regs, so unless all of the lines are owned solely by Metrolink operating lighter equipment (which I agree would make more operating sense) will not happen…yet.
“The fact is that you’re not going to find a lot Montclaireans commuting to downtown LA. We may WANT to see that happen, but it doesn’t reflect reality, and instead it feeds into transit opponents main talking points. I’d rather see this money spent inside LA, improving transit in already dense neighborhoods that need something better.”
Which is a key reason why the line is only going to go as far as Azusa Citrus College where there’s an actual demand. However the reasons why it doesn’t serve the Inner City has more to do with the funding mechanism and the fact that LA County is such a large geographical area. The funding for said projects come from a LA COUNTY Sales tax. If this were a City Sales tax then you’d see a lot more rail (in the City of LA area)
Sales taxes go to all types of projects. Cities have been subsidizing suburban roads with gas taxes and sales taxes for nearly a century. In LA County, chances are that even with projects like the Expo Line and the Subway to the Sea, the city is still a net tax donor, what with all the road widening.
Metrolink could always ask for FRA waivers, like Caltrain. It could point out the outstanding safety record of Japanese railroads, the Caltrain simulations showing that FRA compliance actually lowers safety, etc. And then it could spend money on high platforms and electrification, which would still cost much less than building a new light rail line. This would enable very cheap tie-ins, for example service along the line connecting Union Station with LAX that serves the intermediate neighborhoods and not just airport traffic. (With any luck, such a line would be HSR-compatible, enabling trains that don’t go to Anaheim to serve LAX.)
Alon: okay, so you have a single-track line. The engineer stops, gets out, walks back to the 6th car on the train (a distance of 500 feet, thus 2 minutes), deploys the ramp, lets the wheelchair user on, stows the ramp, then walks back. Now your train is 5 minutes late, and additionally, the train it’s supposed to meet at the next meeting point is 5 minutes late. Now the whole railroad is delayed… it just doesn’t work out, without a massive investment in double tracking everything and building high platforms, and if you’re going to that far, you’ve reached the point point where Metrolink has basically been transformed into BART.
You also have to understand how Metrolink is funded: by contributions from the member counties. They all have their own priorities, and in the case of LA County, Metrolink tends to get a fairly low priority, and thus not all that much funding for infrastructure. The San Bernardino Line, urgent as that is, isn’t even the top of the priority list: I’d put the Van Nuys-Chatsworth double-tracking at the top, because that has potential repercussions as far away as SF.
As for FRA waivers, Metrolink is a rather different operation from Caltrain. Caltrain is a dead-end spur off the mainline network with miniman freight service, and they own their own tracks. Metrolink only owns part of its lines, and the rest runs over some very busy freight lines, including two transcon lines that are full of container trains. Even the branches that they do own (San Bernardino, Ventura, and Orange County) have rather more freight service than can be accommodated with nighttime-only operation. You’d need a different sort of FRA waiver from what Caltrain is trying to get, and that would take even more time. Anyway, it’s not like Caltrain is going to be needing their waiver anytime soon, because they don’t actually have enough money to electrify yet, and it’s hard to say when they will get it.
Going back to the original topic though, I think Metrolink is a perfect demonstration of the tradeoff between capacity and frequency that is inherent in things like commuter rail. It does an excellent job of moving large numbers of people long distances at fairly low cost. Providing the peak capacity of Metrolink’s San Bernardino Line would require a three-car light rail train every 7 minutes, which clearly requires a fully double track line (with at most a few very short sections of single-track). But because Metrolink has much larger trains running at lower frequency, they can actually get away with having a mostly single-track line, at much less expense than light rail, while still doing a fairly good job of serving the regional market, where waiting time is less of an issue and speed matters more.
Yes, but the problem with commuter rail is always the line on the map, which gives nonsavvy people the impression that the market is fully and adequately served.
That’s a good point, but actually… the new MTA map does not have lines for the commuter rail! The old was geographic and definitely had the Metrolink lines, while the new one is more schematic, and only has the stations, not the lines. So, it looks like perhaps the MTA is aware of this trend and is trying to fight it.
Can anyone explain why “commuter rail” lines (and any other operations with poor off-peak service) aren’t combined, more often, with complementary bus service to provide a more complete travel solution? By this I mean buses serving the same stations along the same routes at times when train service is inadequate.
Presumably the roads are less congested during the off-peak, so bus service could provide adequate speeds, and would serve the lower passenger volumes at less cost. Light rail lines often shut down at night for maintenance, and could be replaced by “Owl” service buses.
My theory is that rail tends to be built by “corridor planning” rather than system planning, and the folks who do that don’t appreciate the benefit to the system that would result if they provided the related bus infrastructure such as access ramps and lanes that would be needed for a true seamless solution. They way I would do it is that if you showed up at a station when the trains weren’t running, a bus serving the same route would soon show up and take you where you wanted to go.
Components of the solution are in use in such places as Seattle’s downtown tunnel, where buses serve the same platforms as light rail trains, using a raised strip of pavement next to the platform that lifts the buses to an appropriate height for the rail platforms.
Where rail lines parallel freeways, I would think that providing bus-only access to the stations during off-peak times, while a significant expense, would be well worth it. Where surface streets already serve stations, the capital expense might be quite minimal; perhaps only some signals or yield signs to help buses avoid delays.
These facilities would also keep things running more smoothly during rail service disruptions, when “bus bridges” are required. While “heavy rail” stations would presumably have separate bus lanes, light rail right-of-way at stations can easily be designed so that buses can use portions of it when appropriate.
The problem with the bus-based solution is that rail stations are generally not in places that are convenient for through buses. Going by google maps, a non-stop trip from the San Bernardino station to LA is 1 hour and 4 minutes. The train trip is about 1 hour and 24 minutes, but a bus that makes all the same stops as the train is 2 hours and 16 minutes. When Metrolink does run bus bridges, they generally run the buses in several sections (in this case, it would be something like San Bernardino-Rialto-Fontana-Rancho Cucamonga-LA, Upland-Montclair-Claremont-Pomona-LA, and Covina-Baldwin Park-El Monte-LA), which makes travel between intermediate stations impossible, but that’s the only way to provide a reasonable travel time.
The Commuter rail line in Metro Vancouver. Does run a highway coach. Twice after the last west bound train. Two that go east bound mid day and 3 more after the last east bound train. I think there are two or three of them. They are highway coaches. And the only stops are the stations them selves. Although it isn’t perfect. It does allow someone to stay a bit later after work. They also run them on the weekends when the commuter rail does not run.
Caltrain has a sign by US 101 on the San Francisco peninsula that shows the headway problem badly. It shows travel time estimates and says something like:
DOWNTOWN SF BY FREEWAY: 41 MINS
DOWNTOWN SF BY CALTRAIN: 33 MINS
to encourage people in cars to get off the freeway and take the train. But then it totally throws off its campaign with another line:
NEXT TRAIN AT 7:15
showing that whatever time you saved not sitting in traffic, and then some, will instead be spent waiting for the train. Not the message they were intending to send.
Another factor that is at least important as headway is reliability. A bus that comes every 15 minutes on paper, but in actuality, has random delays of 0-15 minutes on every trip is no better than a bus that runs every 30 minutes, but consistently runs on time.
At one time, the Metrolink Antelope Valley line coordinated with a local transit operator (Santa Clarita Transit) to provide bus service when trains weren’t operating.
Most trains operated from LA to Santa Clarita; only a few could go all the way to Lancaster because of track capacity reasons. Buses met the trains in Santa Clarita for continued trips to/from Lancaster.
Eventually, the tracks were improved to the point that more trains could run, and the bus service was cut back accordingly.