In previous articles of this series, I reviewed the St. Clair line’s operation on New Year’s Day 2010 as well as the first week in January. This article turns to the last week of January when TTC had implemented a new line management strategy to address the problems of service reliability.
Monday, January 25: This day’s operations are much more smooth than what we saw in the first week of January, and two changes are immediately obvious.
Service is departing on fairly regular spacing eastbound from Lansdowne. Although bunching is still present, very large bunches with associated wide gaps do not form up.
There is a regular pattern of short turning cars at St. Clair West Station eastbound. The effect is to restore the extra running time needed for operators to stay on time, but to do this on an organized basis. Given the route’s short length, it is not necessary to short turn every trip because the accumlated slippage can be corrected by less frequent short turns.
There is one small service delay westbound at Spadina just after 1720. This produces a small bunch of cars that persists for one trip, but does not induce a major parade. Service is fairly regular once the peak period cars have left service around 1940.
Note that only two of the St. Clair shuttle buses are GPS equipped, and these vehicles are on the line for only part of the day.
One thing that is particularly evident on this chart is the locations where cars sit in one place for extended periods. This shows up as a row of horizontal notches at certain locations.
Avenue Road (80): Delays here are mainly westbound except during peak periods.
Oakwood (360): Around 0800, you can see cars stopping at two distinct locations (the orange and turquoise lines travelling down to the right are particularly good). First is the nearside eastbound stop line where cars await a clear traffic signal, followed by the farside service stop. Where westbound cars show a delay, it tends to be only a single event because this is a nearside stop.
Dufferin (420): Delays here are not quite as marked, but it is still possible to distinguish the difference in position between eastbound and westbound cars. The westbound cars stop further east indicating that these are generally delays caused by the traffic signals, not by stop service on the farside of the intersection.
Earlscourt (453): Some delays are visible here, notably westbound in the afternoon.
In a future article in this series, I will look closely at specific intersections to distinguish between nearside and farside delays to determine how much is contributed by the traffic signals and how much by stop service time. A great advantage of the GPS data and the frequent data sampling is that this level of analysis is finally possible from the monitoring data.
Tuesday, January 26: The service is not quite as well regulated compared to Monday, although a similar overall management pattern is evident. On a number of occasions, cars manage to run as pairs for at least one round trip suggesting that service spacing is not as closely monitored on this day. At 2135, service is held bothways west of Arlington (300), and this does not clear until almost 2450 (0050 on the following day). Cars remaining east of the delay operate a shuttle service between St. Clair West and Yonge Street.
Wednesday, January 27: This is similar overall to Tuesday’s operation. A short delay eastbound near Alberta at 1530 holds three cars, but this cluster is sorted out in the course of the next trip.
Thursday, January 28: There is a bit more bunching on Thursday, and some evidence of short turning rather than spacing cars to deal with the problem through the midday.
Just before 1600 there is a delay at St. Clair Station and this accumulates six cars. They eventually leave westbound starting at 1631. Meanwhile service has been turning back from St. Clair West to maintain a regular headway to Lansdowne. Another delay forms eastbound from Lansdowne starting at 1630 when half a dozen cars queue up. Eventually the first two of these move off together (possibly coupled), then sit at St. Clair West, and finally run out of service.
Friday, January 29: At 1055, the first of three cars holds eastbound at Wychwood (260). Westbound service at St. Clair West has been turning back or held since 1035. The three eastbound cars are let past whatever is obstructing service, but through running does not resume until about 1115.
The remainder of the day shows the by-now familiar pattern of cars getting short turns every so often to restore them to their scheduled time.
Rather then presenting location-by-location charts for each day in the week, the charts here contrast the headways as seen at a few key points during the first and last weeks of January.
Looking at each pair of charts, the scatter of data for each day is not as severe in week 4 as in week 1. This shows that even for locations east of St. Clair West where much short-turning occured, there was still a more reliable headway with the line management scheme used late in the month. (Although I have not shown weeks 2 and 3, they much more closely resemble week 1 than week 4.)
There is some shift in the position of the trend lines on the charts reflected the better overall headways provided in week 4.
Link Time Charts
These charts show the trip times over the route, excluding the terminal loops, for weeks 1 and 4. For trips in both directions, there is a noticeably better time in the PM peak than in week 1. This suggests improvements in traffic flow, possibly due to traffic signal adjustments. It could also be due to better service spacing resulting in less overloading of vehicles carrying large gaps.
When I reviewed the details for various parts of the route, the main improvement shows up in the segment from Bathurst to Oakwood. There are small changes in other segments contributing to the total, but a good deal of it comes in this portion.
This section is particularly thick with traffic signals, some of which were not exactly friendly to transit vehicles when the route began service.
Oddly enough, there is also some improvement between Yonge and Tweedsmuir, the long-established part of the right-of-way.
The contribution of each section of the route to the link times for the PM peak hour (1700 to 1800) is shown in these charts.
During this period, the scheduled round trip time is 49 minutes including recovery time. The charts here cover only the portion from Yonge to Lansdowne each way and do not include time spent running through or laying over in the terminals. Even on the best of days, about 44 minutes is consumed between Yonge and Lansdowne both ways, and the schedule is tight. This was corrected in mid-February.