A high-frequency SMART corridor, revisited

In this week’s IJ Forum, guest and transit skeptic Richard Hall brought up something I didn’t expect: the argument that transit frequency equals freedom [1]. This is something that transit advocates have been harping on in Marin and elsewhere for years, and one of the key factors that is likely to hold back SMART from reaching its potential. Though I addressed this point back in 2012, with the train opening up this year it is worth revisiting how SMART might be able to reach higher frequencies, and how much it might cost.

There is blood in the single-track stone

Once operational, SMART’s trains will run in both directions simultaneously despite having only one track. It can accomplish this through the judicious use of passing tracks, meaning a strategic stretch of the corridor will be double-tracked so trains can pass one another.

At full build-out, from Cloverdale to Larkspur Landing, there will be 4 such passing tracks, each of which is 12 miles long, which will allow trains to run every 30 minutes in either direction. (This 30-minute time is called the “headway” in transit-speak.) To double the frequency to 15 minutes, we would need to double the number of passing tracks to 8. To double frequency again to 7.5 minutes, we would need to double the number of passing tracks again to 16.

At this point, the permanently single-tracked Puerto Suello Tunnel between North and Central San Rafael becomes the choke point. Any further increases would need to widen the Puerto Suello Tunnel and basically double-track the whole system. This would likely overkill for the foreseeable future.

Using the existing construction costs as a guide, the cost for the passing tracks on a 15-minute headway system is $60 million. The cost for a 7.5-minute headway system is twice that: $120 million [2]. If we include the cost of the trains needed to run such frequent service, the total cost would be $120 million and $307 million, respectively [3].

We can squeeze blood out of this stone if we are willing to spend the money for it.

Of course, ridership at this point becomes a problem. Is it worth it to run trains this often? Definitely.

There are almost 43,000 jobs and over 19,000 people within a half-mile of SMART’s stations. 2,237 people live and work within the station areas [4], but commute trips account for only about 20 percent of all trips, and this doesn’t include people who might use the park and ride services. Ballpark figures, using a ridership model [PDF] from the Transportation Research Board put the ceiling of SMART’s potential ridership at about 25,000 [5]. SMART itself estimates its ridership will be around 4,500 trips per day, which the model agrees with.

We need to dive a little bit into some wonky economics for a moment using these numbers but bear with me.

When a transit operator alters some aspect of a transit service, they of course also alter the ridership. Increase the vehicle speed and ridership goes up. Increase fares and ridership goes down. This is called the elasticity of demand: how much does a given change affect the demand for that transit service. The first example mentioned references the in-vehicle time elasticity because it indicates how much people change their habits based on how much time they spend in a vehicle. The second example is the price elasticity because it deals with the price of using the service.

Headways alter the access time elasticity, which is more valuable to people when it comes to transit. As much as you may enjoy your walk to the train, you probably enjoy the fact that you can work or sleep on the train en route a little bit more.

Access time elasticity among commuters is 1.28, which is pretty high. Decreasing average wait times by 1 percent by decreasing headways by 2 percent increases ridership by 128 percent [6]. Doubling SMART’s frequency to 15 minutes, then, will get us a 64 percent* increase in passengers: from 4,500 trips per day to about 7,380. Doubling frequency again could get us to 12,100 trips per day. This, of course, does not count the number of people who would use SMART for non-commute trips too, and it also ignores the effects of improving on the very infrequent service SMART is planning on providing in the middle of the day.

I would argue that a 15-minute frequency would absolutely be worthwhile on the basis of the commuters alone (the cost per new commute trip is $42,000, half the cost per trip on the baseline SMART system). A 7.5-minute system is not as cost-effective based on new commuters, but is significant if we include non-commute trips.

It may also be a viable alternative to a wider Novato Narrows. The traffic congestion there is in part due to a 15 percent increase in the number of cars travelling through to Central San Rafael, or roughly 500 more vehicles per hour [7]. Diverting 3,690 trips per rush hour (half of 7,380) would alleviate that congestion, at least until drivers fill up the space again thanks to induced demand [8].

So Hall is quite right on this point: transit frequency influences how people travel and how many people use the system. I am unsure whether he knew just how influential frequency can be, but no matter. SMART would do well to examine the effects of increasing its service frequency.

* Average wait times are half of the headway: if you arrive at a random point during the inter-train period, your average wait is half the full headway. Doubling frequency, then – a 100 percent increase – decreases average wait times by half that, or 50 percent.

Note: Given how off-topic any comments section can get in a SMART-focused article, I have turned them off for this post.

Works Cited

[1] Traffic in Marin, IJ Forums (San Rafael, CA, 2016).

[2] David Edmondson, “High SMART Frequency on the Cheap,” The Greater Marin, August 8, 2012.

[3] David Edmondson, “Can SMART Double-Track?,” The Greater Marin, August 6, 2012.

[4] Center for Economic Studies, “Longitudinal Employer-Household Dynamics” (United States Census, n.d.).

[5] Daniel G. Chatman et al., “Making Effective Fixed Guideway Transit Investments: Indicators of Success” (Washington, DC: Transportation Research Board, 2014).

[6] Arthur O’Sullivan, Urban Economics, 8th ed (New York, NY: McGraw-Hill/Irwin, 2012), 295.

[7] David Edmondson, “The 101 Corridor: Transportation Myopia in Practice,” The Greater Marin, January 13, 2013; Caltrans, “Traffic Volumes on California State Highways” (Sacramento, CA: Government of California, 2014).

[8] Connor Jones, “The Street Economics of Induced Demand,” The Greater Marin, December 21, 2015.