At its meeting on July 9, 2009, the TTC considered a staff report about operation of the Queen 501 streetcar service. This reviews various attempts at line management and scheduling as well as their effects on service as measured by short turn counts as well as the number of wide gaps between cars.
For the period October 18 to November 21, 2009, the TTC will experiment with splitting the 501 into two separate overlapping routes, although the exact termini of the routes is not yet decided.
The Queen car has been the source of many complaints about service quality, and I have written several articles here examining the actual behaviour of the route in detail based on TTC vehicle monitoring data. The most vocal complaints arise in The Beach, and much effort has focussed on that end of the line, but problems also exist on Lake Shore. Concerns to the west emerged during recent public meetings on the proposed Waterfront West LRT line where providing basic, reliable service today was more important to residents than a new LRT service in the distant future.
Broadly speaking, there have been two ways in which the TTC has attempted to improve the reliability of service on Queen Street.
- This is an extremely long line (3 hours round trip from Neville to Long Branch), and operators reaching the end of their trip require some recovery time. The amount of time provided has been increased in an attempt to eliminate short turns as a means of providing breaks, or of operators leaving late on the inbound trip. One side effect of this is that when the route is not congested, cars can have very long layovers, sometimes greater than one headway, at terminals. Cars laying over are not providing service to passengers, a critical concern with the streetcar fleet shortage.
- The original schedule treated all cars as a pool, and vehicles switched regularly between the Humber short turn service and the through operation to Long Branch. If there was any problem with vehicle order or placement due to congestion or other disruptions, this linkage would require short turns simply to get cars into sequence. The Humber and Long Branch runs are now treated as two separate routes and are scheduled to minimize changovers between them.
Passengers (and the streetcar they are riding) really don’t care about the operators’ shift schedules. Short turns performed simply to get an operator on time (or to create a break) annoy passengers who must change cars to reach their destination, and create gaps in the inbound service from termini. Although this may be done with the goal of dependable service on the route’s central section, users of the outer parts of the line suffer from erratic service and long waits for cars.
Two schemes were tried to deal with this problem:
- Step forward. Operators arriving eastbound at Russell Carhouse would swap to a car headed westbound putting them back on time. A pool of spare operators would shuttle cars to and from Neville. In effect, the operators would be short-turned, not the streetcars.
- Step back. In a scheme similar to subway operations, operators would be scheduled to take a break of a few headways eastbound at Russell, and then take over a following car. This creates relief time, but does not address the situation where an operator is behind schedule.
Both of these schemes were labour intensive because aspects of the operation are managed in response to daily conditions, not from a pre-set timetable. Extra supervisors were needed to manage operators and dispatch cars. If the pool of spare operators is signed up as scheduled work rather than overtime, there will be days when this staff will be underutilized. If the pool is made up of volunteers on overtime, it is more costly, and moreover the pool is most needed on days when service is disrupted (e.g. bad weather) and the extra staff might be better used elsewhere, assuming the volunteers are even available.
By contrast, subway operations explicitly schedule the crew changes, and this eventually became the tactic used on Queen. However, the scheduling software has some problems with this concept that need to be corrected.
This arrangement also depends to some extent on geography. Russell Carhouse is about a 25-minute round trip to Neville (when times are good), and it is easy to manage a spare pool of operators for this segment. The same scheme would be much more difficult for locations far from a TTC office. Moreover, the scheme has expended disproportionate resources on one end of the line.
An interesting side-note to this discussion is that in conversation with Bob Kinnear, President of ATU Local 113, I learned that the union has wanted the TTC to unlink vehicle and operator schedules for some time, but that the TTC has resisted.
The report shows the relative effect of various permutations of the schemes described above, although some trial periods suffered from unusual weather or lengthy street closures (the Tamil protest on University Avenue lengthened travel times on the central part of Queen). The overall conclusion was that longer running times reduce large gaps, probably because there is little or no need for unscheduled recovery time, but increases bunching. The latter is probably caused by two factors. When running time is more than needed, some cars will dawdle to stay “on time” while others will catch up. Moreover, there is no incentive to leave a terminal on time because there will always be a chance to catch up later.
The primary benefit of Step Forward crews appears in the PM peak. This is little surprise based on my own analyses of route behaviour. Running times and schedules stay fairly well in step until mid-afternoon, but thereafter congestion through the west-central part of Queen can delay cars. This triggers a need for short turns unless there is some alternative mechanism to keep operators on time independently of the vehicles.
For the period June 22 to July 31 now in progress, the TTC is testing a step-back system similar to that used in the subway although there are various clerical and bureaucratic problems this triggers according to the TTC. This is an odd statement considering that subway operations have run this way for years, and I can’t help thinking that it is the process that needs fixing. One particular item, the handling of waybills and transfers (I won’t bore you with the details), will vanish once the TTC converts to electronic fare media and does away with operators managing fare collection.
Improved Vehicle Monitoring and Management Capabilities
Probably the most important part of the report deals with changes to the Communications and Information System (CIS). This is the system that tracks vehicles as they move around the city. As I have discussed at length in articles reviewing route operations, this system has many flaws because it tracks vehicles based on their schedules, a set of fixed radio signposts and calculations of distance travelled by hub odometers. More recently, with all vehicles now having GPS, the location information is accurate. This greatly improves the quality of data both for real-time route supervision and future “next car” displays.
Route Supervisors are testing handheld displays to determine their usefulness for on-street management, and there will be a report on experiences with these units in January 2010.
Many reporting tools are being developed from CIS data, and several of these will provide similar information and analytical capability to the charts I presented on this site in reviewing operation of various routes.
A customer survey was performed both to determine origin-destination data in anticipation of a route split, and to see whether riders noticed a service improvement with the changes in line management techniques.
Of passengers surveyed west of the Humber River (just east of Humber Loop), about one fifth originated east of Yonge outbound, or were destined for east of Yonge inbound. East of Kingston Road, about one fifth of outbound passengers originated west of Yonge, but about one third of inbound passengers were destined for that area.
The survey showed that riders bound for the outer ends of the line did notice fewer short turns and more evenly spaced service. This was mirrored by statistics of customer complaints during the same periods.
Split Route Operation
What is not seen in the data in this report (which are somewhat inconsistent for Beach trips) is the demand pattern between Humber and Woodbine loops. This is important for determining what proportion of all riding will be forced to transfer if the route is split and depending on the amount of overlap between the two new routes.
However, information from other travel surveys is available in the February 2009 update. This shows that substantial demand originating east of Bathurst is destined east of Church, and therefore a “Beach” car should go at least as far west as Bathurst, preferably further so that it would still reach Bathurst if short-turned.
I will discuss route structure in detail a separate post to keep my proposals and the inevitable comment thread separate from this article.