The City of Toronto Planning Department has published a set of demand projections for various combinations of the (Downtown) Relief subway line, SmartTrack, and the proposed northern extension of the Yonge line to Richmond Hill.
This document makes interesting reading because it shows both the status of the evolving master transit plan that went into the modelling, and the vital point that additional capacity into the core area is essential to prevent complete gridlock on the subway system. Both SmartTrack and the Relief Line are essential to a future transit network.
That said, the report raises several issues in part by what it does not talk about, specifically some of the network configurations that have already been presented in various studies.
Alignment Options for the Relief Line, and Other Model Variations
This question of the Relief Line’s alignment is subdivided into two parts: what is the scope of the line, and which route will it take to link Danforth to the core area.
The big options include:
- A “little J” route from Yonge Street to Danforth
- A “big J” route from Yonge Street to at least Don Mills & Eglinton, possibly beyond
- A “little U” from Danforth to Bloor West via downtown
- A “big U” with northern extensions of one or both arms of the “little U”
Work has focussed on the “little J” because that is the scope for a Relief Line so long discussed, and approved for study by Council. Therefore the model numbers do not show any effect of taking the “little J” further north to intercept more traffic bound for the Yonge line. This has already been reported by Metrolinx as a very beneficial extension to the RL.
Within all of the options, there are permutations of a Danforth to Downtown route:
- The north-south segment could lie on either Pape of Broadview.
- The Don River crossing could be at Queen, or further south to allow the line to serve the Unilever site.
- The route into the core could be via Queen or King.
A northern route via Queen makes for a simpler river crossing, but the southern route picks up a major new employment district. The King Street route into downtown also attracts more riders than a Queen route.
City Planning staff have erroneously talked of a King route as if it could only exist as part of the southerly Unilever site alignment, when their own study clearly shows the option of a route crossing the Don at Queen, and then veering into King Street. The more northerly crossing is preferred because it will be easier to build under the river at a narrower point.
The following permutations were modelled to see how they would perform:
- Broadview to Queen
- Broadview to Queen to King
- Pape to Queen
- Pape to Queen via Unilever
- Pape to Queen to King
- Pape to King via Unilever
Of these, the two most promising were the Pape to Queen options with the only variation being whether the line ran to downtown via Queen or via King after crossing the Don at Queen Street. For this article, these are the only two whose demand projections I will discuss.
Further east, there is the question of the Scarborough Subway extension and SmartTrack. Model runs were performed with three variations:
- No SmartTrack
- SmartTrack on a 15 minute headway (4 trains/hour)
- SmartTrack on a 5 minute headway (12 trains/hour)
The SmartTrack cases used a modified land use plan that assumed SmartTrack itself would cause growth that would not otherwise occur. This causes increases for the Relief Line’s projected demand when it is matched with a the lower level of SmartTrack service (4 trains/hour) because the latter does not attract as much riding as the Relief Line.
All model runs used a Scarborough Subway (SSE) with its original three stops, not the “optimized” version serving only Scarborough Town Centre. The disconnect between what is modelled and what is proposed indicates that some of the plan’s elements have changed very recently. The model is supposed to catch up to the plan in future iterations.
None of the SSE or ST figures are included in this report, and so we cannot see how the model divided up demand between them, albeit with the “wrong” station configuration.
Finally, the Richmond Hill extension was added to the model networks to see how it would affect demand on the critical downtown segment and Bloor Yonge Station.
All of these numbers must be taken with awareness of the limitations on what has been modelled, notably:
- With the RL ending at Danforth, the potential benefit (and hence RL demand) of the “big J” is unknown.
- The five-minute service on SmartTrack, identified in a previous study as essential to attract riders, may not be physically possible given constraints on sharing the network with GO.
- It is unclear whether SmartTrack will actually operate at no fare premium above local TTC services, another essential component of making this service attractive to riders.
- The effect of SmartTrack in the downtown segment, including the degree to which it would duplicate an RL at the Unilever site, depends on the ability to operate frequent ST service.
- The relative roles of the Scarborough Subway and SmartTrack in attracting riders is unknown because the now-proposed station layout has not been modelled.
That is a long list of variables. Many of these will be addressed in updated model runs expected in coming weeks, but readers should be careful not to take the current model output as definitive.
Nonetheless, the report concludes that treating SmartTrack and the Relief Line as options is misguided because both will be required to accommodate future demand to 2031 and beyond. Addition of the Richmond Hill extension to the mix will exhaust the Yonge line’s capacity by 2041. This makes further study of the “big J” quite important.
The findings in this Summary Report make clear the importance of the Relief Line. It is apparent that both the Relief Line and SmartTrack will be required in the future to ensure the efficient operation of the existing and proposed future transit networks. Additional work is required to assess the potential benefits of extending the proposed Relief Line north of the Bloor-Danforth subway to Eglinton Avenue and potentially to Sheppard Avenue. [p 3]
The Capacity of the Yonge Subway
For the purpose of these model runs, the Yonge line’s capacity is set at 36,000 passengers/hour. This corresponds to a design service of about 33 trains/hour, or 110 seconds, with a train capacity of 1,100. This is the target for the line once the new Automatic Train Control system is active sometime in 2019. (The TR trains can hold more than 1,100 riders at crush load, but this cannot be sustained over a peak hour because crowded trains take much longer to service stations thanks to onboard congestion.)
Once upon a time, TTC management claimed a much higher target for subway capacity with trains running as often as every 90 seconds. This was impractical and is no longer considered a realistic goal. However, the higher claimed number, plus a drop in system demand with the 1990s recession, allowed the TTC to dismiss the need for a Relief Line as an essential future component in the network for many years.
System capacity is not just a question of how many trains/hour can move along the track, but of capacity for passenger movement at stations. This is important for the ability to clear platforms between trains (a similar challenge exists with future GO/SmartTrack service levels at Union) and especially for transfer capacity at Bloor-Yonge. Diverting demand away from that transfer point not only lessens crowding, but also could avoid the need for a billion-dollar retrofit and expansion at that station.
Demand on the Yonge line is constrained by train capacity, and if it is not possible to board, people will seek other routes or modes. Many of the modelled networks show less reduction in peak demand on the Yonge line than new demand on the SmartTrack and/or Relief Line corridors. This happens because any “relief” allows the Yonge line to backfill with today’s unmet demand. In effect, space on Yonge is purchased by building capacity elsewhere and redirecting traffic.
Relief Line Demand and Effects at Bloor-Yonge
The table below summarizes many of the model configurations for the RL alignments from Pape Station to downtown via a Queen Street crossing of the Don River. Networks without SmartTrack, and with two different ST service levels are also included. The assumed headway on the RL is 3’00” (20 trains/hour).
“Boardings” count the number of people who get on the RL at some location. The Peak Load values vary considerably with an obvious effect where frequent ST service draws demand away from the RL. This observation must be tempered by the unknown potential effect of the “big J” version of the RL that would serve a part of Toronto where ST is not a competing factor.
“Net New Riders” shows an estimate of the riders who are new transit users, as opposed to those diverted from existing routes. It is unclear whether this includes only riders on the RL, or additional riders on the Yonge line taking advantage of new capacity there.
The peak demand south of Bloor Station for many configurations is at or above the Yonge line’s expected capacity of 36,000/hour. The situation is even worse when demand from the Richmond Hill extension is added (see following section). An important effect of both the RL and ST traffic diversion is the reduction of the busy west-to-south transfer movements at Bloor-Yonge. This is important because the TTC has a billion-dollar proposal to expand capacity at this location, a project that might prove unnecessary if transfer traffic through the station can be reduced. That saving would be an offset against the cost of any new line providing that relief.
Not included in the published information is any analysis of transfer traffic at Pape-Danforth station which would take over part of the function of Bloor-Yonge. This will be sensitive to the configuration of RL used because a “big J” would pick up many riders north of Danforth rather than having them board from feeder buses or the Danforth subway at Pape.
Peak boardings are higher on the RL with the SmartTrack 15′ service as compared to no ST because new developments are included in the land use for those version of the model run. The developments attract riders, but they wind up preferentially on the RL because it would operate at a much shorter headway. Conversely the peak point loads do not change much indicating that the “ST-related” development is adding to demand that is counter-peak, or at least not flowing through the peak point.
An intriguing observation about peak demand on the Yonge line:
There is an apparent shift of the Yonge AM peak load point from south of Bloor station to just north of it with the introduction of the RL indicates that reducing transfers from the westbound Bloor-Danforth line to the southbound Yonge line cannot, in itself fully solve Yonge line over-crowding, which largely results from heavy southbound flows along the Yonge line from points north. [p 10]
In other words, as more capacity is added to the network south of Bloor Street, the congestion point shifts north of Bloor Station. Building new capacity only for riders south of the Bloor-Danforth subway will not address the heart of the problem, the volume of traffic originating to the north.
The Effect of the Yonge Subway Extension to Richmond Hill
When the YSE is added to the mix, peak demand at Bloor southbound increases from 39,600 to 41,600 in the model. However, both of these numbers exceed the line’s capacity, and the added congestion shows up as a modelled drop from 9,900 to 9,100 for transfer movements from the Danforth subway.
Congestion on the subway clearly would limit the attractiveness of the extension, especially for today’s riders of the subway. Although total riding projections only rise marginally, this will almost certainly occur by displacement of riders who now board further south.
Comparison of Bloor-Yonge Effects With Relief Line, SmartTrack 5' and YSE Without YSE With YSE 2031 2041 2031 2041 Peak Hour Demand 31,200 36,000 33,800 37,800 WB-to-SB Transfers 4,900 5,340 3,600 5,200
All of the configurations modelled here show a Yonge line operating above capacity by 2041 when the YSE is included. From Metrolinx demand projections published in June 2015, we know that the “big J” would attract considerably more demand than the “little J” away from the Yonge line.
The Vital Role of Both SmartTrack at Close Headways and the Relief Line
Although the combination of SmartTrack and the Relief Line, as well as planned additional capacity on the Yonge Subway itself, go well toward resolving the capacity problem on the existing network by 2031, the effect is not lasting. Moreover, the entire plan depends on the attractiveness of SmartTrack running on five minute headways and at a TTC fare.
SmartTrack provides comparable relief to the Yonge line relative to the RL, with the magnitude of this relief very much depending on the SmartTrack headway. With 15-minute headways, SmartTrack has relatively marginal impact on Yonge line ridership, while at 5-minute headways, the impact is quite substantial. [p 10]
As discussed above, Yonge line crowding still generally remains a concern, however, due to the heavy volumes boarding the line from the north. The little-J RL options cannot address this problem directly. SmartTrack diverts some of this traffic, particularly at lower headways, but, ideally, more should still be done to keep the loading of the Yonge line during peak periods under capacity. [ibid]
(As an aside, let us not have any shilly-shallying about what a “TTC fare” might be. For the purpose of this study, SmartTrack was assumed to be an integral part of a TTC flat fare network.)
The Relief Line’s ability to attract new riders is constrained by its primary role as a bypass link from Danforth to downtown, especially on a route that avoids the potential new development at the Unilever site.
In general, these results indicate that the RL primarily functions as providing an improved path into the Toronto downtown for trip-makers within its catchment area (e.g., diverting trips from the Bloor-Danforth Subway). It provides an improved level of service (e.g., reduced travel times) for these riders and helps reduce over-crowding on the Yonge line. Its potential for attracting large numbers of new riders to the transit system, however, appears to be somewhat limited. [p 12]
Whether this statement would be true for the “big J” route to Eglinton or beyond is another matter as this would bring rapid transit service to an area that does not have it today. We must await further model runs to see how this works out.