Normally, I would save detailed reviews like this to a general article looking at the Queen route over several months and configurations. However, a deputation at the June 24, 2013 TTC Board meeting is worth comment now while the issue is fresh in Commissioners’ and management’s mind.
A regular attendee of these meetings complained that he had been severely hampered in attempting to use the Queen car late in the afternoon of May 25 to travel westbound to Long Branch. As I have recently received the vehicle monitoring data for several routes for Mar 2013 from the TTC, getting an overview of what was happening was quite straightforward. It is not a pretty picture.
The Service Chart shows the entire day’s operation as a time-distance diagram for each car. Long Branch Loop is at the top, Neville Loop is at the bottom. Cars follow diagonal paths alternating in direction as they traverse the city. Short turns are evident where an individual line turns around at a location other than a normal terminus (e.g. Kipling). Delays and long layovers are evident where one or more lines runs horizontally for a period of time.
The space between lines on the service chart shows the space between vehicles, but this is plotted in a different way for many locations on the headway charts. Each dot on a chart represents the time between a vehicle and the one that preceded it (the “headway”), and a trend line shows the overall behaviour. Where pairs or triplets of cars travel close together, then one or more dots will be near the zero line, and the large gap in front of this bunching will be shown by a single dot well above the average.
The common pattern here as on all routes I have studied is that vehicles overwhelmingly travel with uneven spacing and this shows up in the regular rise and fall of headway values. On the central part of the route, the trend line tends to lie near the scheduled headway because all of the cars pass through these points, just not at the regular spacing we would prefer to see. The further out one goes, the more the effect of short-turns (including the scheduled one at Humber), and the higher the trend line values become.
The Service Chart
From the start of service until about 10:00 am, things run fairly normally although a small amount of bunching is evident later in this period.
- Just before 7:50, a car leaves Humber a bit late and close in front of a through car from Long Branch. Eventually, on its return trip, this car (coloured black) catches up with its leader (pale orange) a bit west of Russell Carhouse and runs close behind it to Humber.
- At about 8:20, a car (light yellow) leaves Humber with a Long Branch car (blue) close behind it. They travel together all the way to Neville.
- At about 8:53, a car (turquoise) enters service from Russell eastbound to Neville right behind another car (orange). They travel more or less together on their westbound trip to Humber. Our turquoise friend then proceeds on to Long Branch and has a layover of 25 minutes. It is somehow appropriate that this car returns east in a gap that widens progressively enroute to Neville.
The first two are simple examples of how the merging of two services appears to be unmanaged with no attempt to space the service even at a logical dispatching point such as Roncesvalles eastbound.
By 10:00, the service is starting to get ragged.
- Three cars arrive at Neville almost together at 10:10, and they proceed west across the city in a 20-minute gap. Two of these are cars that caught up to each other on the trip east from Humber, and the third came out of Russell just before 10:00 to join the parade. The gap gets wider and wider as they make their way to Yonge where it is 30 minutes. A car appears out of Spadina to split the gap, and another at Roncesvalles. Of the three cars that left Neville, one turns at Humber, one at Kipling and one goes through to Long Branch
- Eastbound service is also continuing to show pairs of cars, and short turns with substantial layovers at Woodbine Loop begin to appear.
- Congestion eastbound from Bathurst appears at around noon and continues through the afternoon.
By mid afternoon, there is even more bunching of service and ongoing short turns at Woodbine, Sunnyside and Kipling loops. Cars re-entering service from short turns often do so as part of pack, not in the gap between cars, and so offer almost no relief in service quality. The operator may be back on time, but the riders see little benefit.
This continues until about 5:40pm when there is a holdup westbound at Bathurst Street. Five cars accumulate here and two divert around the obstruction via Spadina and Shaw (turquoise and orange lines disappear at Spadina and re-appear later at Shaw on the chart). During this period no service is operating to Long Branch. A car leaves Humber westbound at about 6:00pm, and nothing goes out on Lake Shore until just before 7:00pm.
During this period, the CIS data show one bus entered service as a 501, but its GPS info was so badly off-route, that the filtering I normally apply to strip out rogue data completely eliminated it from the chart. Upon review, it appears that this bus made three round trips between Bathurst and Humber Loop (although the GPS trajectory was more to the northwest). There may have been other shuttle buses, but they do not appear in the data.
Through the early evening, cars continued to run in bunches with a parade of six cars heading east from Humber around 8:00pm. A separate sextet made its way west from Neville at the same time after a long delay at that location beginning just after 7:00pm. A nearly 45-minute gap was filled by two cars short-turned at Woodbine.
To put this another way, at about 8:00 pm, most of the Queen cars were at Humber or Neville with a few left out on Lake Shore.
By late evening, the large bunches have dispersed, or some of the cars have gone out of service, but pairs of cars are still common. Short-turns at Woodbine continue until well after midnight putting large gaps in service to Neville.
Out at Long Branch, there are three cars in the loop from 11:15pm even though the scheduled headway is 18′. Two of these leave together and travel all the way east to Russell, eventually accompanied by the third car which has caught up. One of them short turns at Russell and the other two continue together to Neville.
This illustrates a complete collapse of anything remotely resembling “management” of the service, and is a common problem seen in analyses of other route behaviours evenings and weekends. The problems cannot be attributed just to “traffic congestion” but show an abdication of any attempt to maintain properly spaced service.
Westbound headways from Neville start off well-behaved within the TTC’s target of 3 minutes plus/minus to the scheduled headway. By 10:00 as we saw on the service chart, things come apart and headways much worse than the scheduled values with a range commonly over 10 minutes wide are the norm. By early evening, we see the echo of the delay at Bathurst as well as the holdup at Neville itself (followed by many short headways). Erratic service continues for the rest of the evening. (Note that 301 night car service is not included with these data.)
The situation at Woodbine is almost identical with the times shifted slightly later.
By Coxwell, we see the effect of short turns at Woodbine Loop (Kingston Road) in reducing some of the wider gaps, but the sawtooth pattern continues. This continues more or less unchanged west to Yonge. The late morning 30 minute gap is quite evident in the chart.
The enormous gap westbound from the Bathurst delay shows up at Humber Loop at 18:46, and the one from Neville at 21:13.
Things are much worse on Lake Shore where cars routinely run in pairs and the nominally 10-minute service actually operates over a range of zero up to 30 minutes. The large gap from Bathurst, thanks to service turning at Humber, is over an hour wide west of Humber Loop. Even without this one big gap, the service is appalling. A rider might see a car the moment they arrive at a stop or wait more than double the advertised headway.
Eastbound headways leaving Long Branch are fairly well-behaved until about 11:00 am and they lie within a 6 minute band. By the afternoon, however, headways are more variable and service more erratic.
East of Humber Loop, the headway swings are not as bad as on Lake Shore, but the persistent pairing of cars leaving Humber is evident. The pattern of headways continues right across the city (step back for forward from page to page for an “animation”) with the gaps getting wider (peaks in the chart getting higher) as the gap car falls further and further behind due to overcrowding and long dwell times.
East of Russell Carhouse, the headways widen considerably because of short-turns. The evening is particularly bad with nine headways over 20′ wide when service is supposed to operate on a 9′ spacing.
This may be a particularly bad example including two long delays on the same line at roughly the same time. However, these incidents alone do not explain what is going on with a failure to properly space the service that is running more-or-less normally.
This is a weekend. The TTC is not going to get a reserved lane anywhere along Queen for Saturday operations, and they need to spend much more effort on managing what service they have to better service this long, complicated route.
Despite two major delays on the Queen route on May 25, there were no TTC E-Alerts issued for these incidents or any other on the 501 for that day.