Updated March 23 at 11:20 pm: After this article was published, one commenter noted that the cleanliness and attractiveness of the system is another vital aspect of “good service”. In my article, I had concentrated literally on “service”, not on the physical condition of vehicles and stations. However, I do agree that a run-down, dirty system does not inspire confidence, and the TTC is looking decidedly shabby.
Tess Kalinowski has an article in the Star on the issue of station condition on the south end of Yonge Street where the BIA is doing a running audit to track progress on fixes. The TTC has its own internal monitoring, and loves to trumpet improvements as a good news story, but the truth is that there is backlog of repairs that gets longer by the day. This topic is worth a post in its own right, but I wanted to add the link to the Star’s article here as it fits right in with the comment.
In all the talk about who should run the TTC, or whether transit should even be provided by the public sector, one important question is rarely answered. What sort of transit service do we as a city want? What is “good” service?
“Good” is a relative term depending on your viewpoint. If your job is wrestling with municipal or provincial budgets, your outlook will be to restrain growth in costs and to limit expectations of service quality. This runs head on into schemes to redirect growth in travel from autos to transit. More passengers almost always mean more subsidies over and above any inflationary growth.
If you are a transit rider (or thinking of becoming one), you don’t want service that is barely acceptable. You want service that retains your loyalty and that you would recommend to others. Riding should grow because transit attracts customers, not as the “least worst” of options.
There are three essentials in transit service: reliability, frequency and connectivity.
Service must show up when you expect it. This must happen not just at your local stop, but wherever you travel on the network. Ideally, service should be frequent enough that you don’t worry about the schedule, and waits should not be a major part of your trip.
Frequent service should arrive frequently, not in bunches of vehicles separated by long gaps, and vehicles should go where they are expected to go. Whether the vehicle is on time matters less than whether service is regularly spaced.
Where low population or job density dictates less frequent service, reliability is essential. This is particularly true on branching routes which need to be on time so that customers can plan to use specific trips.
Connectivity has two elements — provision of diversity in travel options and, again, reliability. One seat rides are impossible for every trip, and transfer connections are a fact of life in large transit systems. These connections must work, not be left to chance. Trip planners and real-time information displays won’t make up for the frustration when someone just misses a connection.
On smaller systems, buses can meet regularly (say half-hourly) at one or more nodes, and almost all transfers occur at these nodes. This simplifies network design, but circuitous trips do not attract riders. This design is not practical for a large busy system like the TTC, and a challenge for smaller systems (or portions of large ones) is how to make the transition from a nodal network to something more like the TTC’s grid.
The TTC’s own planning standards recognize that the presence of a transfer connection is a major deterrent to riding, and penalizes transfers where they might be created through a route reorganization. The tradeoff lies in providing the most direct route for the most people while not so gerrymandering a network that it is useless to anyone who isn’t following the “typical” peak trip.
How comfortable should a trip be? How should a system trade off passenger demand with vehicle size? On a network like the TTC, routes vary in their travel patterns. Some are highly directional and vehicles will run with very light loads in the counter peak. Some have good bi-directional demand, or may replenish their loads at major demand points including transfers with other routes. Looking at a single point on well-used routes rarely gives a good picture of how many riders the route actually carries, or how riders travel on that route.
All of this brings me to the question of service standards. The cost of providing transit service is strongly linked to whatever standards a system might have, but the relationship must be understood both in the context of local conditions and the role of a transit system in a city.
As the TTC went through the funding cutbacks of the 1990s and the slow recovery of the 2000s, many “standards” existed only on paper. Transit agencies (indeed any organization being measured for “performance”) may lower their standards to meet actual operations and budgets rather than striving to attain hard-won improvements in goals for their system. Transit that is only “good enough” quickly declines to transit that cannot retain and attract riders.
The TTC’s Service Standards fall broadly into three categories.
Service designs are based on the average load at a peak point over the peak hour. This value is not the same as the engineering design capacity of a vehicle.
- If a vehicle is loaded at crush capacity, it has no remaining circulation space to accommodate movement of passengers at stops, and stop service times become quite long. Everyone knows what happens when buses, streetcars or subway trains with no room at all arrive at a stop. They stay there for a very long time.
- There will always be variations in demand over a peak hour due to slight changes in vehicle spacing, arrival times of connecting vehicles and other time sensitive events. Designing service to be “full” 100% of the time means, in practice, that some vehicles will be overloaded.
- The inability or unwillingess of the TTC to manage vehicle spacing, combined with effects of short turns, leads to bunching where the average load may appear low, but most of the passengers are on one vehicle and experience far more crowded service than averages would suggest.
In practice, TTC service standards call for all passengers, on average, to have seats during off-peak periods, and for vehicles to be moderately crowded, again on average, at peak. These standards are:
- Buses: Offpeak 35-39, Peak 47-51 (lowered from 52-57 as part of the Ridership Growth Strategy in late 2008)
- CLRVs (short streetcars): Offpeak 46, Peak 74
- ALRVs (long streetcars): Offpeak 61, Peak 108
- SRT (4-car train): Offpeak 130, Peak 220
- 6-car Subway: Offpeak 500, Peak 1,000
When subway line capacities are quoted, the typical train capacity cited is 1,200 passengers or more, and this yields a theoretical 36,000+ line capacity based on 30 trains/hour. However, service cannot reliably operate at that level of crowding as anyone who rides packed rush hour trains will attest.
Most of the time, a transit vehicle is not travelling at a route’s peak point or direction, and it will not be full. Routes with diverse origins and destinations may do better overall, but transit vehicles cannot, by definition, be full all of the time. Any attempt to design service to such a goal is counterproductive. The real question is to get good use out of the vehicles that are on the street.
All door loading improves passenger distribution inside vehicles (compare subway trains to buses and streetcars), and enables a higher standard capacity because internal circulation for stop service is simpler. When the TTC moves to all-door, low-floor loading and self-service fare collection with the new streetcars, this will improve vehicle utilization and allow a higher average load to be carried relative to the theoretical capacity.
Improving route speeds with reserved lanes and priority signalling can shorten trip times and contribute to a feeling by riders that “we are getting somewhere”. However, the capacity of a line is a function of the number of vehicles per hour past a point, not the speed those vehicles are travelling. A 10-minute headway means only six buses an hour even if they travel like rockets.
For some transit planners, the main benefit of “transit priority” is reducing the round trip times and hence the number of vehicles on a route, rather than on allowing for more frequent, and hence more attractive service. This was a flaw in the original plans for the 512 St. Clair line and this outlook undermined the project’s credibility.
The TTC’s current standards call for all surface routes to run during normal operating hours (nominally 0600 to 0100 the following morning, except Sundays when service starts at 0900). The maximum headway permitted is 30 minutes.
On the subway, the maximum headway permitted is 5 minutes. Relative to surface routes this provides a much higher quality of service and, coupled with generally faster travel times, contributes to the “I only want a subway” outlook that many riders have. If the TTC were as generous with surface operations, these might become more popular, or at least more credible as alternatives.
A scheme to move to a 20-minute maximum headway was proposed a few years ago, but the cost of doing this system-wide is prohibitive, and the TTC feels that the money would be better spent on targeted improvements.
The TTC plans to introduce a “Transit City Bus Network” in the fall of 2010 which will identify a core set of routes for maximum headways of 10 minutes during normal service hours. I understand that the originally proposed network will be revamped before this actually comes into operation, and we should see a new map for the proposal later this spring.
Any discussion of service standards must consider both frequency and loading standards. In some neighbourhoods, the headway needed to produce a seated load would be so wide that nobody would wait for the bus to arrive. Indeed, short routes can be quite productive without achieving seated loads because they don’t carry people very far and the cost per trip is low. Doubling the headway won’t produce twice as many passengers per bus because many riders won’t wait that long.
As things stand, the TTC applies its headway standards over the entire system and this brings 30-minute service to areas where one might question its worth. With the introduction of the 10-minute core network, the idea of different standards for different parts of the network will be in place. This could equally allow for a relaxation of standards in those areas such as industrial subdivisions with little or no demand outside certain peak travel times. However, such a change must be implemented with care to avoid starving the system in residential areas at the margins.
The TTC considers the access distance to transit service as part of its evaluation for new routes or revisions. Potential new passengers should lie no more than 300m beyond existing transit services (or 200m if there is a “higher-than-average” proportion of seniors).
This has important implications for route speed and for new rapid transit proposals. If stations are 1000m apart, it is obvious that some potential riders cannot lie within circles of 300m radius, let alone the extra distance required because the street grid does not allow straight-line access between home and the station. A surface bus may be provided, but on a very wide headway up to the policy maximum of 30 minutes, and this can lead to a very substantial degradation in service for those who fall outside of the immediate area of stations.
A missing part of the service standards is a recognition that 30 minutes is simply too long to wait for a bus that is providing an “alternative” to a rapid transit line.
Another big gap in the standards is any recognition that wide headways pose a problem for those riders who may be on the margin between using “conventional” and Wheel-Trans services. The TTC speaks of the importance of allowing riders to use the conventional system as much as possible, of the high cost of providing Wheel Trans and of the challenges for handling non-critical Wheel Trans trips. However, these concerns are not reflected in the service standards.
When the TTC looks at a potential service change, they assess each component of a rider’s trip with different weights to reflect the perceived (and often real) differences in travel and convenience. These weights are:
- 1.0: In vehicle time
- 1.5: Waiting time
- 2.0: Walking time
- 10.0: Each transfer
If we can cut the average waiting time for a vehicle by 5 minutes, this is equivalent to shortening the ride by 7.5 minutes. A walk that is shorter by 5 minutes is equivalent to 10 minutes of saved travel time. Putting it another way, a longer walk or wait for transit may be tolerable if the service is faster, but it must also be reliable to ensure that the waiting time factor does not kick in. Similarly, a transfer to an infrequent surface route poses a greater obstacle than between two frequent services.
In the debates about the future of our transit system, a key question is the quality of service we expect. “Quality” has many components as I discussed here, and it is important that these are well-understood. If there is a demand for broad cutbacks in transit subsidies, where should they be made? If we plan to improve transit, what sort of changes should we make to system-wide standards? If there are proposals for outsourcing or privatization, what standards should we demand?
I have deliberately stayed away from questions of wage rates, working conditions and system management in this article to concentrate on service quality. Once we know what quality we want, then the question turns to how we can obtain it at a cost (fares, subsidies or both) that we as a city and as riders are prepared to pay.
In the next article, I will turn to the question of fares and fare collection.
For more details about the TTC’s service planning protocols, please refer to Service Improvements for 2008 on pages 7-10.