Over this last week, I have been posting plots of bus positions vs. time in this folder. I’ve just updated the plots for the 5 most bunched bus routes run by the CTA: 9 – Ashland, 22 – Clark, 49 – Western, 66 – Chicago, and 79 – 79th. Excluding the 22 Clark, these are also the 4 highest ridership bus routes, which is not a surprise since ridership causes bunching.
I’ve added a right column to the plots to show the average inverse speed of busses on the route, binned every 1000 feet with the standard deviation in the error bars. This shows us the sources of delay on the route and where the delay is variable (either from bunching or congestion). I’ve plotted and averaged over delay instead of speed as delay is proportional to the cost to operate the route and the cost of lost time for customers. Let’s take a look at where and how much delay is occurring.
For the 66 – Chicago, the most delay occur in the River North section of the route. This is where investments to improve bus speeds have the greatest ability to reduce running times. While a dedicated bus lane would allow the 66 to bypass congested vehicles, congestion does not seem to be the greatest source of delay in this section. Rather, signal timing and passenger loading are the leading causes. Even late night OWL trips, when loading delays are small, are consistently delayed by the high concentration of traffic lights in this part of the city. The delay due to passenger loading can be seen in the asymmetry between eastbound and westbound travel times, especially at the State & Lake Red Line stop. The eastbound buses are half a minute faster here as an average of 11 customers load onto each westbound bus. A similar number of customers load at Michigan and at the Blue Line. Introducing signal prioritization and off-board fare payment with all-door boarding at the three ‘L’ stations and at Michigan would reduce travel times by an average of 5-7 minutes per round trip. With an average of 105 round trips per day and an operating cost of $135 per revenue hour, the CTA could save $430,000 per year. A far greater amount of value in the time of CTA customers would also be saved.
Outside of River North, there are notable delays at many major cross-streets where boardings are high. The delays seem to be larger than can be expected from boarding delays alone at both Pulaski and Halsted. It’s not clear to me what’s causing the delay at Pulaski, but the delay at Halsted is largely due to congestion. High variability in bus travel times within 1000-2000 ft on approach to Halsted in both directions indicates that busses are becoming stuck in traffic queues. Queue jumping bus lanes are in order. There is more than sufficient pavement width to implement the lanes and to create protected bike lanes at the same time, although some pay parking would have to be relocated on the approach from the west. This lane would save on average 3-4 minutes per round trip with larger benefits during peak times.
Moving onto the 79th St. bus, very large delays are occurring at two complex car-centric intersections on the route: the six-way Stony Island / South Chicago intersection, and at the Dan Ryan. For many reasons, the intersection at Stony Island requires a complete redesign and improved bus speeds should be one of the goals in the redesign. At the Dan Ryan, the 79 is almost guaranteed to hit a long red light at both State and Lafayette, right at the busiest section of its route. The unnecessarily wide pedestrian crossing distances at these intersections require long signal times that are coordinated for through vehicles. The 79 always has to stop between the two intersections thus missing the green wave. Similar problems exist at other intersections of bus routes with the Dan Ryan. Extended curbs and bus signal prioritazation should be implemented. The large westbound delay at Western is due to the extra distance the bus requires to loop around the 79th and Western terminal in this direction. This penalty will be reduced once the construction on 79th between Western and Ashland is completed. Finally, there is noticeable congestion delay at most major cross-street intersections along the route: Kedzie, Ashland, Halsted, and Cottage Grove. Short queue jump lanes would be appropriate at these locations when the cost of parking removal or relocation isn’t prohibitive.
Now the 22 Clark bus. The top plot is from a Saturday while the bottom plot is from the same Thursday as the plots for the 66 and 79. This route suffers greatly from congestion delay as indicated by the large errorbars along nearly its entire length. Some locations are notably bad: the intersections at Devon, Ridge, Foster, and everything south of Diversey/Broadway. Queue jump lanes would help greatly in these locations. Fortunately, there is room on the street for bus lanes at most of these locations. At Devon and Ridge, Ashland is very wide to where bus lanes could be added without significantly affecting automotive throughput. At Foster, left turn restrictions and moving the stops to farside would have to be implemented to provide sufficient room for a queue jump bus lane. South of Lincoln Park West, a vehicular travel lane in each direction could be reassigned as bus lanes, and the same could be done to LaSalle St south of North with the 22 shifted over one block onto LaSalle. These latter two reallocations of road space would result in a significant reduction of peak period automobile throughput. While the mobility benefits of a lane reallocation are likely to outweigh the loss in automobile capacity, the cost/benefit would have to be looked at in more detail.
While full BRT on wide streets such as Ashland promises to greatly reduce the delays experience by a single bus route, there are many inexpensive projects that can combat transit delays where full BRT cannot be implemented. We don’t need to wait for a BRT demonstration to start making these common sense improvements. We do need to be able to visualize transit performance data in a way that allows us to find these opportunities.
Next week, I’ll return to the Ashland BRT, to examine how time saved by transit customers balances against added automobile congestion.
