In a recent post, I looked at the time needed for 509 Harbourfront cars to make their way between Union Station and the CNE. These times are extremely well-behaved and show the benefit of having a line both in its own right-of-way and on a route where traffic congestion is almost unknown (at least in December).
What this analysis didn’t talk about was headways.
There is no way to put this gently. The service on Harbourfront is appallingly bad because some operators treat the schedule as little more than wallpaper, and nobody in Transit Control seems to care about the quality of service. It’s a little shuttle, it will look after itself. Meanwhile, as we have seen in the Spadina car analyses, the attitude seems to be that short turns of Union-bound cars are just fine, maybe because that wonderfully reliable 509 will handle the demand.
These lines are supposed to show what the TTC can do when we get traffic out of the way. What they demonstrate is that the TTC doesn’t give a damn about running proper service in an area where the population is growing and the transit system hasn’t got long to establish its attractiveness before they all buy and drive cars.
Where will this leave us in the eastern waterfront? What does it bode for Transit City?
Here are the running time charts showing times westbound from Union to Exhibition, and eastbound from Fleet/Bathurst to Union. Note that departure times from the CNE are not reliably present in the CIS data and “CNE to Union” times vary a lot because of layovers at Exhibition Loop.
I am not going to give you many charts to pore over because the information you need can be gleaned from overviews and a few examples.
Because the running times along the route are quite predictable and the route is short, the headway seen at any point is representative of the route as a whole. There isn’t time for cars to change their relative spacing either due to variations in demand or traffic interference.
Here is the headway summary for December westbound at York Street.
If you look at the first three pages (weekdays before Christmas) as well as the fifth page (all weekdays), you can see a huge scatter in the headway values particularly during the pm peak and early evening. The plus or minus three minute rule is nowhere to be found here, and the number of very short and very long headways shows us that many cars are running in pairs.
Page four (weekdays after Christmas) are not quite as bad, but there is still a lot of paired running. Weekends are generally better behaved.
To get a sense of what this looks like on a service chart, here is a sampling of days through the month.
On Friday, December 1, one car Run 3 (yellow) consistently runs close to its leader, run 1 (black). Run 4 (turquiose) only pulls close up to run 2 (purple) on its running in trip leaving Union just after 9 am. Service through midday is well behaved with two cars shuttling back and forth. However, by the pm peak, things getquite ridiculous. Run 3 (brown) develops a fondness for run 2 and stays nose-to-tail with it until running in at about 19:15. Meanwhile, run 4 (green) spends much of its time following run 1.
On Tuesday, December 12, the peak period service is much better behaved. (Note that some of the strange behaviour around 14:00 is due to missing CIS data.)
Friday, December 15 brings us back to pairs of cars running together in the PM peak.
Saturday, December 16 shows well-behaved off-peak service. Note that there is a block of CIS data missing on this day in the early afternoon. Sunday, December 17 is similar.
Monday, December 18 shows a weekday with reasonably well-behaved service.
Clearly, the style of peak period operations depends a lot on who is operating the line. Some days, the operators run properly spaced service, other days, there may as well not be any peak period extras. Line management is nowhere to be seen because the pairs of cars are allowed to run back and forth making wide gaps where riders expect “frequent service”.
Finally, a look at the terminals. Although times at the CNE loop are a bit dodgy (generally, CIS does not “see” a clear departure time), we can measure the round-trips from nearby locations where CIS data is reliable.
At the east end of the line, the round trip time from York to Union and back is quite consistently at the 6-minute line plus or minus a few minutes. This indicates that although there are some layovers at Union, they are fairly consistent and not extraordinaily long.
The data for Bathurst to the CNE is quite another matter. The average of round trips lies in the 10-to-15 minute range on weekdays, and higher on weekends. The considerable scatter in the data shows that cars get long layovers at the CNE on some trips and probably have more running time than they need overall. The higher weekend numbers show that the layovers (and hence the excess running time) are even longer on those days.
The 509 looks like a very cushy route to operate: almost no traffic to worry about, lots of layover time and little demand to provide well-spaced service during the peak. This is not an ideal example of how our new premier services should operate, and it gives the TTC and its “LRT” programs an ill-deserved black eye for riders.