For the period October 18 to November 21, 2009, the TTC will experiment with splitting the line in one of two ways. The specifics have not yet been decided.
- Split the route at Humber Loop into the original 507 Long Branch and 501 Queen routes.
- Create two overlapping routes with Long Branch cars operating east to Church or Parliament, and Queen cars operating west to Bathurst or Shaw.
Implementation of an overlapped service is constrained by the number of available cars as well as by track layouts and the location of electrified switches for turnbacks.
There are arguments to be made for either configuration, but the first one does have the problem that if Queen service is disrupted in the west end, the link to Long Branch will be lost as most Queen cars will turn at Sunnyside. Conversely, an overlapped route will force anyone travelling from west of the overlap area to east of it to transfer.
The worst possible service design would be to have a Long Branch and a “Beach” car scheduled to travel together on the common segment to ease connections between them. Aside from the difficulty of making this work on the street, it would defeat the purpose of having a better headway for riders in the busy central part of the route.
Where To Split the Route
The following section is for the benefit of those readers who insist on intricately detailed analysis without benefit of actually looking at the routes in question.
There are limited options for turning back cars because the track and loop layouts derive from decades-old requirements and, in some cases, routes that have not been around since at least the opening of the BD subway in 1966, if not longer.
- Sunnyside Loop. This loop, just west of Roncesvalles carhouse, is a common short turn for the Humber service. The loop entrance switch is electrified.
- Dufferin Street westbound (via Dufferin, King and Shaw to Queen). This loop involves three left turns, one at the busy Queen/Dufferin intersection under the railway viaduct. The switches involved are electrified, and the likely layover location, northbound on Shaw, is not too busy. There is no transit priority to aid the left turns at any intersections.
- Dufferin Street westbound (to Dufferin Loop). Instead of an onstreet loop, this would take the Queen car to the existing loop at the West Entrance of the CNE grounds where there is an operators’ washroom and place for recovery time. The trip time would be slightly longer than the around-the-block versions, but would establish 501 service both ways as far as Dufferin.
- Shaw Street westbound (via Shaw, King and Dufferin to Queen). This loop involves three right turns, a preferable arrangement for traffic, although there is no priority signal for the left turn at Shaw. Also, the west-to-north switch at King and Dufferin is not electrified. Layovers on Dufferin are possible, but probably more difficult than on Shaw.
- Bathurst Street westbound (via Wolseley Loop). This involves only one left turn from Bathurst east onto Queen, and all switches required are electrified. Cars can lay over in the loop, although excessive running times could produce a queue that would block Bathurst Street.
- McCaul Street westbound (via McCaul Loop). This loop is already used by the 502 Downtowner service.
- Church Street eastbound (via Church, Richmond and Victoria). This loop was used by the Long Branch downtown service for many years, and is still fairly regularly used (in both directions) by 501 and 502 cars. Layovers would probably be taken on Richmond Street. The main problem with this as a terminus is that a car short-turned eastbound at McCaul would completely miss the connections with the subway. The car and operator would be on time, but empty.
- Parliament Street eastbound (via Parliament, Dundas and Broadview). This loop is used from time to time for Queen short turns and diversions, and all of the electric switches needed for it (in either direction) already exist. Layovers would have to occur on Parliament, although this could block 504 King cars using the same route.
Aside from equipment availability, the main problem with any route design is that there’s a good chance many cars will not reach their planned destination. For that reason, it is important to choose turnbacks that would maintain service over a reasonable portion of the line even if service were short turned.
This would require Queen/Beach cars to be scheduled to go at least to Dufferin if not Roncesvalles so that the common short turn would be no further west than Bathurst or Shaw respectively. Similarly, Long Branch cars should be scheduled to operate to Broadview via Parliament so that they would short-turn at Church and retain service to the core area even if service were disrupted. (A comparison here is with 502 Downtowners that short turn westbound at Church and completely miss the heaviest outbound stops on the route.)
An Alternative Proposal
A scheme that has not been examined in detail is a restructured route taking the Long Branch service to Dundas West Station, with a downtown extension during peak or daytime periods. This would, in a sense, make the 507 a mirror of the 502/22A Coxwell services with downtown services weekdays, and service north to the Bloor subway weekends and evenings. This scheme was proposed by me some years ago, but always seems to be lost in the shuffle when options are discussed.
Yes, I know, Roncesvalles is closed to streetcar traffic until late 2010. We should be discussing service designs now, not after it reopens.
This arrangement would give a Queen route from Neville to Humber, and a Long Branch route from Brown’s Line to Church via King during peak periods. Off peak service would turn north to Dundas West Station.
The goals of this scheme are:
- Sever Long Branch service from the busy Queen route, especially its most congested section from University to west of Bathurst, while continuing to provide direct service to downtown during the peak period. Rather than the occasional 508 Lake Shore car, all 507 Long Branch trips would operate through to downtown.
- Return the Queen service to a line of manageable length (roughly an hour one way) so that it will not require extended recovery times.
- Preserve an overlap between Queen and Long Branch routes to ensure that connections between them are maintained even if a Queen car is short-turned.
- Provide additional service on Roncesvalles during off-peak periods when disruptions on 504 King can result in irregular headways.
An optional change for this route structure would be to operate all (or most) 504 service and all 507 service with ALRVs, and switch the 501 back to CLRVs. The purpose of this scheme is:
- Recognize that the King route is the heavier of the two and should routinely have the larger cars assigned to it. The TTC often talks about the problems of traffic signal management with short headways, and yet it operates the shortest headway on King, not on Queen.
- The Queen route with its potential for disruptions both in the Beach and on Queen West needs the maximum number of units available. This gives the greatest flexibility for short turns because the gaps they cause are shorter on close headways than on long ones. CLRVs are also a bit more sprightly than the ALRVs, and this is useful on a congested street like Queen.
How Many Cars Would This Require?
In January 2008, TTC staff reported on Queen route issues, and this included comments about reallocation of the CLRV and ALRV fleets. No further update on this scheme has been brought for discussion at the Commission.
From the May 2009 schedule, the AM peak requirements are
- 31 ALRVs on Queen
- 38 CLRVs plus 7 ALRVs on King
- 3 CLRVs on Lake Shore
- Total: 41 CLRVs and 38 ALRVs.
(It is possible a few CLRVs could be freed up if the 502 and 503 services were consolidated as one route, but that’s not part of this discussion.)
During the PM peak, Queen uses 31 ALRVs, while King and Lake Shore use 41 CLRVs between them. Therefore, as a starting point, without adding any more to the fleet requirements, we have 38 ALRVs and 41 CLRVs.
Anyone who looks at these routes will know that Queen often operates with CLRVs on ALRV headways, and ALRVs can be found on King in the PM peak. There does not seem to be any well-coordinated pattern, but rather whatever is available goes out on the street.
The Queen service from Humber to Neville operates at a 5’10” combined headway on a 130 minute round trip in the AM peak, 5’40” and 138 minutes in the PM peak. Converting on a 3:2 ratio for CLRVs, this would change the headways to about 3’25” in the AM and about 3’45” in the PM. This would use 38 CLRVs assuming no change in recovery time.
The King service from Broadview to Dundas West Station operates in the AM peak on a 4’00” headway and a 108′ round trip. Overlaid on this is a 4’00” headway of 16 trippers producing a 2’00” headway eastbound during the peak hour. Some of these trips are supposed to operate with ALRVs. In addition, three Lake Shore cars form part of the eastbound peak service.
In practice, the trippers are not well-integrated because the 504 cars originate at Russell Carhouse and tend to depart unreliably. They may or may not make it as far as Dundas West, but are short turned, and may not fit well into the eastbound headway. Moreover, the combined 2-minute headway almost guarantees that cars will run in pairs due to effects at traffic lights. The Lake Shore cars do not operate on reliable times.
If the King service operated with ALRVs on a 2:3 replacement ratio, the headway of the base level of service would change to 6’00” and would require 18 ALRVs. Overlaid on this would be one hour’s worth of trippers on a 6’00” headway, or another 11 cars.
A Long Branch service to Church would require a bit over an hour each way, or 13 cars to provide a 10’00” headway. The current service (on paper) is one ALRV every 10’20”.
Total requirements (AM peak) would be 38 CLRVs (Queen) and 42 ALRVs (King and Long Branch) compared with current schedules of 41 and 38 respectively. Considering that an ALRV rebuild program is about to start, we don’t know what actual fleet availability will be by late 2010, but it should certainly be better than today.
Please do not send endless comments with other configurations. My purpose in working through the calculation is simply to show that it is possible, and we do not need endless discussions of fine tuning the headways.
TTC staff have not yet published their proposal for a split operation on Queen, and this issue will not come before the Commission again before the trial service is implemented. We will know the details probably by early September when the schedule designs for October are available.
The trial is worthwhile, but I am concerned that it may fail if there is insufficient overlap between the two halves of the route.
As for the Long Branch proposal, a full implementation would have to wait for the reopening of Roncesvalles, although an interim scheme could bring a “507” service east to Roncesvalles Loop overlapping a 501 “Humber” operation. Similarly, the interchange of ALRV and CLRV fleets between the King and Queen corridors can be performed at any time.