This is the second of two articles reviewing the operation of a comparatively short downtown bus route to see how it behaves in comparison to longer and busier routes. The first article covered 6 Bay for the same period.
The schedule for 94 Wellesley was adjusted in September 2015 to provide added running time, and to improve service so that the eastern part of the route was in the “10 Minute Network”.
The scheduled service in effect for January 2016 on 94 Wellesley is shown below (click to enlarge).
The 94 Wellesley bus carried 9,400 passengers per weekday as of April 2014.
For the purpose of this analysis, the route is divided into segments so that behaviour within each of these can be isolated. The screenlines (boundaries) used were:
- Bloor at Rosedale Valley Road bridge: this is just west of Castle Frank Station loop.
- Wellesley at Sherbourne St.
- Wellesley just east of Wellesley Station.
- Wellesley at Bay St.
- Wellesley at Bathurst St.
- Ossington just south of Bloor St. and Ossington Station loop.
As with the analysis of 6 Bay, this post begins with a look at New Year’s Day, 2016. Operating conditions on that day were ideal with no congestion, light passenger loads and benign weather.
Headways westbound from Castle Frank Station lie in a reasonably narrow band especially at midday and early evening. One bus goes out of service missing a westbound trip to Ossington at 1:30 pm. It is replaced by a new bus for the next trip at 2:40 pm. This causes a double headway to make one round trip on the line.
However, other wide gaps are simply the result of uneven spacing of one bus which runs consistently off schedule during certain periods. Notably there is a wide headway westbound at 14:15, 15:25 and 16:37. Each of these corresponds to the same bus which takes a long layover and then leaves immediately ahead of a 94B Wellesley Station short turn. (See blue line on the service chart linked below.) Presumably this behaviour corresponds to a shift by one operator.
By the time the service reaches Sherbourne Street, the pattern of headways is more exaggerated as gaps between buses widen, and vehicles running close to their leader catch up. This widening of gaps continues west to Yonge.
Only alternate buses are scheduled to continue west of Wellesley Station to Ossington Station, and the headways leaving westbound vary widely. This is an example of how a point that could easily be used to re-establish regular service (an off-street loop) does not, and the headway pattern that evolved east of Yonge is compounded to the west by the absence of half of the service. (Note that some of the widest gaps during the afternoon are caused by the missing bus.)
By the evening, all service is scheduled through to Ossington, but the headways remain uneven. One bus is consistently operating on a short headway followed by one on a long headway.
Eastbound service from Ossington is much better spaced than the westbound arrivals, but there is still a range of over 6 minutes for much of the day. This part of the route has scheduled service at a level where unreliable headways can lead to long waits for buses.
By the time the service reaches Bay Street, some of the gaps have opened up a bit compared to departures at Ossington.
Leaving Wellesley Station eastbound, the service is better spaced although wider swings do emerge later in the evening. As seen elsewhere, the gaps widen as buses travel east to Castle Frank.
The service chart appears to be well-behaved with buses criss-crossing the city mostly without incident. Variations in headways are not as obvious on this scale, but is is possible to cross-reference the gaps from headway charts with individual buses in the service chart.
The early afternoon missing bus leaves Castle Frank at 13:30 (mauve line). Previously it has been on the Ossington branch, but it turns back from Wellesley Station and then goes out of service at Castle Frank. A replacement bus (green line) appears later and leaves Castle Frank at 14:40.
Terminal layovers are adequate if not generous for all trips, and so being “late” cannot be explained by bad scheduling, at least for New Year’s Day. This is an example of the effect of generous running times coupled with an “on time” standard that accepts a swing of 6 minutes as an acceptable range.
Service leaving Castle Frank Station shows a consistent pattern over the month with headways that are barely inside the 6 minute line during the daytime and on weekends. For weekday evenings, the standard deviation of headways (a measure of the dispersal of their values) is highest and this is visible in the wider “cloud” of data points during that period. This is a problem common to many routes – substantially less reliable service during the evenings.
West of Bay Street, the 94B service has dropped out, and the scheduled service is less frequent. The SD values here are higher too, and the typical range of headways exceeds the TTC’s 6 minute standard.
Service leaving Ossington Station eastbound is at half the level of Castle Frank Station, and punctuality is more important on the wider headways. Even so, weekday headways are quite irregular. Oddly enough, the weekend service is somewhat more reliable than on weekdays, particularly on Saturdays.
Service eastbound from Wellesley Station shows the same problems seen elsewhere, particularly with weekday service. Despite the fact that the subway loop would be a convenient dispatching point to re-establish on time behaviour, headway values are widely spread with the worst being during evenings every day of the week.
The service chart for Monday, January 18, 2016 is fairly typical for this route. There is little congestion on the western portion of the line, but to the east the effect starts at around 7:40 and continues until well after the “rush” hour. The chart shows clearly that this is mainly a traffic problem, not long dwell times at stops, because the lines creep westward in small increments as buses pull through slow-moving traffic. If the problem were confined to stop locations, there would be horizontal segments (stationary vehicles) followed by movement comparable to that we see during uncongested periods.
Some congestion effects are visible eastbound between Bay and Sherbourne in the afternoon peak, but no to the same extent as the westbound delays in the morning.
A few short turns of buses occur at Spadina and at Bathurst.
(Note that spikes and gaps in some lines on the charts are caused by GPS errors, and they are more common on some vehicles than on others.)
The MonthLinks charts below show the travel times from just west of Castle Frank Station and just south of Ossington Station, both ways. This eliminates the effect of terminals layovers to provide a comparison with scheduled running times (see service summary above). Generally speaking, these times are within the amounts allowed for in the schedules. Travel times are fairly consistent as shown by SD values in the 3 minute range, slightly higher during peak periods, and the fairly narrow spread of data points around the trend lines.
The following two charts show values for individual segments of the route on Monday, January 18. These can be read together with the service and headway charts above for a detailed view of the line’s behaviour on that date.
- The westbound link from Castle Frank to Sherbourne roughly doubles in travel time between for the morning peak and a period thereafter. After noon, the travel times settle down and there is only a small rise for the PM peak.
- Sherbourne to Wellesley Station shows an AM peak increase, although it is slight smaller and short-lived than the one in the previous segment.
- Running times from east of Wellesley Station to Bay Street vary somewhat due to varying dwell times in Wellesley Station, but the trend line shows no peak period effects.
- From Bay to Bathurst there is only a small rise in PM peak times.
- From Bathurst to Ossington Station, again there is only a small peak effect. The four spikes in evening values show up as layovers by buses before reaching Ossington Station (see service charts).
- The eastbound link from Bloor and Ossington to Bathurst shows a doubling of travel time during the AM peak.
- The peak effect is even more pronounced between Bathurst and Bay, and this segment also has some PM peak increase.
- From Bay Street through Wellesley Station, travel times are higher in the PM peak.
- From Wellesley Station to Sherbourne shows a small PM peak increase.
- From Sherbourne to Castle Frank there are small, short lived peak increases.
If the schedules were inadequate, it is likely that we would see many very short times to make the loop through terminals because there would be little time for a layover.
At Castle Frank, the minimum circuit at the terminal requires about 3 minutes, and this probably represents trips with a short time for passenger service at the station. However, most times for this circuit are well above this lower bound except during the PM peak and, oddly enough, on Saturday evenings.
As with the eastern terminal, the shortest circuit time is about 3 minutes, but many trips take substantially longer, with many values well above 10 minutes. This appears to be a preferred location to take a layover compared to Castle Frank Station.
There is little question that scheduled running times on 94 Wellesley provide enough time for trips during peak periods, and that the schedules are quite generous at other times. However, as we have seen on other routes, when there is extra time, and the “standard” for headways gives great leeway, then the actual service provided can be quite erratic.
The “Catch 22” for the TTC is that padded schedules lead wither to excessive layovers, or dawdling operations because vehicles are almost always early. The new schedules may eliminate short-turns and reduce overtime costs, but a better approach to headway reliability is badly needed.
For a description of the methodology behind these analyses, please read Methodology For Analysis of TTC’s Vehicle Tracking Data.