Saturday, November 16, 2013

The ignominious D Street light

Much has been written about Boston's Silver Line (including on this page). It's certainly not rapid, but it is rather convenient: I made it from Kendall Square to Terminal A in 25 minutes. The problem? It should have been 23.

I've noticed in the past that the Silver Line experiences long waits at the D Street grade crossing after it exits the tunnel. On my ride yesterday, I decided to find out just how long, by means of Youtube:

The bus gets to D Street, and proceeds to sit there for not 30 seconds, not a minute, but just shy of a minute and a half! This is a major service failure. The scheduled time from South Station to Logan airport and back is 45 minutes, meaning that if the bus loses 1:30 each time it crosses D Street, 7% of the route time is spent waiting for a traffic light. For the SL2 line, which is a 25 minute round trip, 3 minutes is 12% of the total operating time!

There's this thing called "transit signal priority" which could be employed to eliminate these wait times. A sensor could be placed just outside the WTC station (and a similar one on the inbound run) which would be tripped when the bus passed by (there are already sensors which detect the bus and raise gates when the pass, and which close barriers should someone attempt to drive in to the tunnel). This would give 15 seconds to flash the don't walk sign and change the light, allowing the bus to proceed through the intersection at full speed. Traffic would not be dramatically impacted since there light would only be red for a few seconds when the bus passes, and an algorithm could be put in to place to assure the green cycle for traffic was long enough to avoid backups (but not, you know, not 90 seconds when very few cars pass through; see above).

Transit signal priority (TSP) is not very expensive; even at the high bound it costs $35,000 per intersection. Ridership on the SL1 is, give or take, 8,000 per day. This means that in a year, for one penny per passenger, trip lengths could be reduced by more than a minute. This should be implemented immediately.

What's more, this would result in reduced operation costs for the MBTA. Buses cost somewhere on the order of $100 per hour to operate. Even if the average time savings per bus was only 30 seconds, this would equate to 4.5 hours of operating time per day, or a $450 savings. Assuming a $35k cost for TSP implementation, it would pay for itself in 78 days—two and a half months.

The argument could be made that these times would just be built in to schedule padding at the end of the route and savings would only be from the reduced power use related to not stopping and starting. But, especially on the SL2, saving a couple of minutes could be used to decrease the overall route time and increase service, something the Seaport District desperately needs. At rush hour, decreasing the trip time from 25 to 23 minutes would allow headways to drop from 5:00 to 4:36—a capacity increase of 8%—without any additional cost. This would allow for 3 additional round trips at rush hour, or 75 additional minutes of service, which would save the T $125. By this metric, the payback would be $250 per weekday, and take 140 weekdays to pull in to the black. That's 7 months. After that, it's gravy.

(Another improvement: extending overhead wires along the whole of the SL2 route would allow the buses to operate without a change of power twice per trip; combined with the savings at Silver Line Way this might allow service to operate at 5 minute headways with 4 buses—a dramatic savings, albeit one with a higher initial capital cost.)

There is no logical reason that transit signal priority should not be immediately procured and installed at D Street. There is no need for a time-consuming review process; the benefits are clear and any disruption to traffic will be far less than the current disruption to the traveling public. While the Silver Line is still hobbled by a convoluted route system, low capacity, slow tunnel speeds, traffic and a poorly-designed power switch (often requiring the operator to exit the bus and manually raise the trolley poles), this inexpensive change would be a good start to dramatically improve service.


  1. Great post! This is something I think about all the time -- how we can have such an incredibly simple piece of tech can make some [relatively] incredible improvements. The MBTA seemingly refuses to consider signal priority anywhere, however -- it is completely mind boggling. But here you actually worked out the math for the Silver Line and you can show the benefits.

    I'm hoping maybe you'd add up all the red lights on specific Green Line branches next? B needs it desperately, then the E, and finally the C. With the C, the city of Brookline even INSTALLED THE HARDWARE for signal priority. They told the MBTA it would improve service and the MBTA essentially said, "Pft, prove it! And then MAYBE we'll do it."

  2. Yes, it would certainly help on the surface light rail lines. What I really like about the D Street example is that you have a very small fleet to equip (there are only a few dual-mode trackless units) and only one signal which would save a lot of time. If it were successful in reducing trip times (and it would be), then it would be a good proof-of-concept for the surface rail lines.