On February 28, 2020, Metrolinx release a Preliminary Business Case for the Scarborough Subway Extension, and an Initial Business Case for the western extension of the Eglinton Crosstown LRT from Mount Dennis to Pearson Airport.
The Eglinton West Extension IBC is not as blatantly skewed as the Scarborough study in that it acknowledges the LRT plans and uses these as a starting point. The SSE study simply pretends that LRT does not exist and touts “benefits” of the subway versus a bus network.
The idea of a line along Eglinton West has been around for a long time.
1972 ‘GO Urban’ and Rapid Transit Plan on Eglinton: Eglinton corridor was part of Province’s and TTC’s ‘Intermediate Capacity Transit System’ (ICTS) Network Plan (in which the present Scarborough RT was a part of)
1985: ‘Network 2011’ and Eglinton West Plan: TTC Report identified Eglinton West as busway corridor as part of Metro Toronto’s rapid transit network plan (in which the present Sheppard subway was a part of)
2007: ‘Transit City’ and Eglinton Crosstown Plan: Eglinton Crosstown LRT (spanning from Pearson in the west to Kennedy in the east) was part of the City of Toronto’s surface rapid transit expansion proposal
2010: Crosstown LRT Project Approval: City of Toronto sought EA approval for surface LRT alignment from Kennedy to Pearson Airport Area boundary, one year after the City approved the full-length Eglinton Crosstown alignment
2012: Eglinton West Segment Deferment: Metrolinx undertook Crosstown LRT construction, with Mount Dennis-Pearson Airport segment deferred due to funding constraint
2016: Eglinton West LRT IBC: City of Toronto and Metrolinx co-published Eglinton West LRT’s first IBC and recommended surface LRT option, and the City approved funding for preliminary planning and design works
2017: Grade Separation Review: City of Toronto approved arterial and midblock stops along Eglinton West and conducted grade separation study to address community concerns
2019: Surface Option’s Affirmation: City of Toronto in its report maintained its preference for surface LRT based on fine-tuned benefit-cost analysis. [p 9]
As an historical note, a mid-1960s TTC plan to serve the airport with LRT linking southeast to the Bloor Subway and northeast to the Finch corridor was stillborn thanks to the 1972 GO Urban scheme.
We have arrived at a preferred subway option by provincial fiat:
In December 2017, the City of Toronto conducted further studies on additional grade-separated options based on inputs from the local community who are concerned with the at-grade LRT’s traffic impact. The number of options for the Toronto Segment were revised to four (featuring at-grade and below-grade alignments with frequent arterial and midblock stops, and a mostly below-grade alignment with either a single stop or multiple arterial stops) and re-evaluated using traffic model updates and additional metrics recommended by community representatives.
Nonetheless, the City of Toronto in early 2019 released a report re-confirming their preference for an at-grade LRT due to its cost-effectiveness in meeting all of the city’s project and policy objectives.
Subsequently, the Province’s 2019 Budget announcement included the extension of the Eglinton Crosstown LRT to Mississauga as one of the four budgeted rapid transit projects with an underground alignment. [p 27]
This is a second “Initial” study, but it is in the context of Premier Ford’s strong preference for subways. The usual caveats apply about the rough level of cost estimates and comparisons between options.
There is a fundamental conflict in the analysis in this study triggered both by the Premier’s distaste for “streetcars” and by Metrolinx methodology.
The surface option carries the greatest number of weekday riders, but one of the subway options carries the most “new” riders. In other words, the surface option does a better job of serving existing demand while not drawing as much new demand, while the subway option leaves some existing demand quite literally “out in the cold”, but (according to the modelling) shifts more people to transit from autos. The machinery of value assignment rewards the new riders more because they allegedly represent less auto use, and they tend to be longer-distance riders who save more travel time.
At no point does the study explain why carrying fewer total riders at much greater expense constitutes a valid planning outcome.
The Four Options
Five options are under consideration including a “Business As Usual” scenario in which service is provided by the 32A Eglinton West Bus. Unlike the Scarborough study, this is a base case that actually exists and provides a valid starting point for comparison.
All four of the LRT options include a leg between the Renforth Station and the airport that is lifted from past studies of this route, but this is a placeholder. The airport access will be affected by plans for a major surface transportation hub at Pearson Airport and this could affect the link’s alignment and stations.
All options will have construction impacts on the neighbourhoods around the line, although these vary depending on the type of construction (surface or underground) and the number of stations.
Option 1: Surface LRT
Option 1 is the Transit City Eglinton West line. There is a short section of tunnel at Weston Road to extend the current Line 5 westward through a section of Eglinton with buildings very close to the roadway, but the line then emerges to run on the surface to Renforth. (The text accompanying the diagram in the report erroneously states that the portal would be east of Weston Road when, in fact, it would be to the west.) This arrangement has an approved Environmental Assessment in place.
Lands were reserved along the north side of Eglinton for the Richview Expressway originally planned to run east to meet Highway 400 (aka Black Creek Drive). Some of this land was sold for housing, but the strip needed for LRT expansion of the road was retained. Transit signal priority will be required to ensure good travel speeds for a surface LRT, although Toronto has an uneven record in giving true priority to transit. The road’s existing capacity, including provision for turns, would be maintained.
Option 2: Nine Stop LRT Subway
Option 2 contains the same stops on Eglinton West as in Option 1, but the route is completely underground from Mount Dennis to just east of Renforth Station. The tunnel would pass under three rivers at the Humber, a Humber tributary west of Royal York, and Mimico Creek west of Martin Grove.
The Humber Ravine is a particularly difficult area because of the grades required to get from Mount Dennis down below the river. Options 3 and 4 avoid this by crossing the Humber on an elevated structure, then diving underground west of Scarlett Road beyond the river’s floodplain.
The underground stations pose challenges because of high pressure gas lines underneath most sites. This affects both Options 2 and 4.
Option 3: The Express LRT Subway
Option 3 treats the extension as an express route to the airport district with connections to surface routes only at Jane and Kipling. The portion of the route over the Humber River would be on an elevated structure to avoid the complexity of going under this river.
The report talks about sustainable development with prosperous residential and commercial communities. This is difficult to achieve on a line with infrequent stops.
An express LRT from Renforth to Mount Dennis might look nice to a “regional” planner, but it does not address all of the demand in its corridor. Metrolinx’ focus is on “regional” travel including a coarse network of bus routes with 10 minute headways. How this is supposed to make any dents in auto travel demand is a mystery, but that is a separate study.
Option 4: Six Stop LRT Subway
Option 4 is a variant of Option 2 with three intermediate stations – Widdicombe Hill, Wincott and Mulham – dropped to reduce costs. Also, this option uses the elevated Humber River crossing from Option 3. (The report’s text includes the correct number of stops, six, in the section title, but claims there are seven stops in the body of the description. Editing errors like this do not inspire confidence in the care taken in preparation.)
This is the preferred alternative from the evaluation process.
The projected all day ridership numbers are intriguing on a few points.
- Option 1 has the most riders at 42,500 indicating that a line with the most stations will do best at attracting demand.
- Option 2 has fewer riders at 36,500 similar to Option 4 at 37,000. This will reflect the interplay between fewer stations but faster travel in Option 4.
- Option 3 despite being an express line has the lowest demand at 23,000.
The low change in demand for Option 3 shows that there is a substantial potential demand along the route, and in a later section this turns out to be transfer traffic from intersecting north-south buses.
As with so many other rapid transit studies, the importance of feeder services is not well understood by planners for whom the word “express” is though to be the key to “success”.
There is a rather odd contrast between the ridership prospects of the Eglinton alternatives which show Option1 best, and a statement later in the report.
When comparing against BAU, Option 4 has the highest increase increase in boardings for both Toronto’s Eglinton Crosstown LRT and Mississauga’s Transitway BRT, with a 23% jump in combined weekday boardings. [Footnote: Option 4 increases Crosstown and Transitway boardings of 119,000 and 7,000 by 23% and 16% respectively.] This increase demonstrates the value of a complete travel network, as existing rapid transit lines become more useful when the network is made seamless with fewer transfers and slow mixed-traffic links. [p 40]
The increases cited above translate to about 27k for the Crosstown and about 1k for the Transitway, a total of, say, 28k. This is far more than the number of new transit trips cited in Table 9 below suggesting that a good chunk of the projected new boardings are trips diverted from other routes, not net new transit trips. This is an important distinction when talking about congestion reductions and the like.
An obvious question to ask is whether an option is “better” because it carries more riders (even if a lot of them are on transit already) or because it attracts more “new” riders to transit.
Moreover one must be careful not to double count boardings when looking at riders transfering from MiWay to TTC. They are still only one person, and one cannot count their boardings on both systems as if each is a separate new rider.
All ridership projections include the then-current fare structures for 2019 including the GO/UPX/TTC co-fare that is about to be discontinued. This could affect the model’s distribution of trips between all-TTC and GO+TTC route choices.
As with so many of Metrolinx’ studies, the primary benefit lies in the imputed value of travel time savings for riders. In this case, the comparison is complicated by the interplay of travel time and demand. Option 1 has more all day ridership, but Option 4 saves more time for its less numerous riders, thereby winning out. Also, an underground line has the greatest benefit for riders who travel its entire length whereas a surface route is more supportive of demand and potential development along the way.
The chart below shows the travel time savings for each option. The differences seem large except when one looks at the total time as described in the text, not the delta values in the chart. The starting point is an 82 minute journey via the Crosstown from Yonge to Mount Dennis, then a TTC bus to Renforth, and finally a Miway bus to Square One. Of course it takes forever, and nobody in their right mind would attempt this unless they had no other choice. The surface LRT would get this down to 73 minutes, and the six-stop subway option to 68. This is still a long journey, and we don’t know how many people would attempt it. The chart is misleading because it exaggerates the proportionate saving over the entire trip.
The shortest trip of course is for the express subway Option 3 which speeds through Etobicoke almost as if it were not there to ferry people to and from the airport and Mississauga. These are the same savings numbers shown in Table 9 above.
The chart below shows the total travel time values rather than just the deltas as above.
From a financial perspective, none of the options will attract enough new ridership and fare revenue to offset the increased operations and maintenance costs. In a rare bout of clarity, the report notes that “this is consistent with most other rapid transit improvements of similar nature” [p 4]. Rapid transit, and indeed most transit, does not pay for itself out of the farebox. However, on a purely financial basis, Option 1 with its lowest cost would be preferred. Its capital cost is estimated at $2.128 billion (2019$, NPV) with a return on investment of 0.14. That is a very low figure, but it is the best of the lot.
An important point here is that all options are costs including the link from Renforth Station to the airport, but the value of this link is not stated. Therefore it is difficult to assess the cost estimates for the Mount Dennis to Renforth section in isolation, or in comparison with previous studies.
Part of the evaluation includes access to homes and jobs. The metrics used are an 800m radius from stations for homes, and a travel time of 45 minutes to jobs. A subtle effect shows up in the evaluation of Option 1 which has only a 4% increase in access to jobs within 45 minutes, compared to 11-18% for the other options. The relative isolation of the airport means that jobs located there include a long travel segment that must be substantially shortened for these to change significantly in accessibility from other parts of the city. The surface operation of Option 1 and its additional stops to not make enough of a difference to affect the metric. This would be more obvious if Metrolinx published travel time gradient maps showing where the 45 minute boundaries lie for each option relative to the airport.
The dichotomy between regional and local planning shows up under the first Metrolinx Strategic Objective: Connect More Places with Better Frequent Rapid Transit.
The cumulative network effects will help to improve the overall connectivity within the region, particularly for Etobicoke and other suburban areas straddling the Eglinton Ave and Transitway corridors. The proposed investment will result in uninterrupted rapid transit connectivity that stretches from Toronto’s Golden Mile (Scarborough) in the east all the way to Mississauga’s Erin Mills Town Centre in the west, covering Midtown Toronto and Downtown Mississauga, both of which are important emerging urban growth centres in the GTHA, and Pearson Airport.
A well-connected frequent rapid transit network improves travel times between key regional destinations to enable transit users to get where they are going faster and to access further destinations. Of the four options, Option 4 will have the largest impact on regional transit accessibility, with 13,000 new transit trips and 140,000 person-minutes of morning peak transit travel time savings for the GTHA transit network as compared to the BAU scenario. [p 38]
In contrast, Option 1 will have the least impact, with 9,000 new transit trips and 60,000 person-minutes of morning peak transit travel time savings for the GTHA transit network. [p 39]
The statement about new transit trips omits that ridership on Option 1 is projected to be higher than for Option 4, although travel time savings would be less because of the surface alignment.
It is all very well to talk about major centres, but there are a lot of them, and moreover there is a lot of space in between. Travel time savings vary depending on where riders are going (their origin is not mapped here) and the large benefit is to employees in the airport district. However, this is spread over a large area, and good “last mile” links will be needed to make the simulated benefit actually useful.
There are some oddities on this map where a travel time increases are forecast implying that the demand model has a few kinks in it. A little piece of downtown Toronto is oddly a longer trip for riders after the LRT opens implying that the Eglinton West bus somehow has a far-reaching benefit on a corner of the city.
There is a mildly amusing discussion of stations and accessibility.
Option 1 will be the most accessible as compared to other grade-separated options which feature elevated guideways and tunnels that require stations to be built further from the surface. Option 2 will be the least desirable in terms of station access and egress times as it features the most underground stations, which will likely require additional access times as compared to elevated stations.
Nonetheless, Option 2 together with Options 3 and 4 feature mostly grade-separated stations that afford the highest level of protection against severe weather conditions. The greater service reliability offered by these options is ultimately the most important contributing factor towards creating a more positive travel experience for transit users. [p 42]
One might suggest that the Eglinton West planners have a chat with the Ontario Line planners about the joys (or otherwise) of outdoor vs underground stations.
The projected ridership profile for Option 4 is very interesting. The highest AM peak ridership is eastbound to the Spadina subway at Cedarvale (aka Eglinton West) Station, and continuing high over to Eglinton Station at Yonge. However, on the Eglinton West extension, the predominant flow is outbound.
Note that the underlying network for this projection is supposed to include the Ontario Line, but the estimates for Science Centre station “Don Mills & Eglinton” suggest that boardings are more numerous than alightings (i.e.: more “ons” than “offs”).
The number of people getting off at Eglinton Station (over 8,000 in the peak period) should give pause to anyone thinking about “surplus” subway capacity.
No charts for other options were included in the report to show how their design would affect demand on the route.
The Line 5 extension is projected to draw many riders north from Line 2 (Bloor subway) to make their east-west journeys on Eglinton. Figure 15 below shows that the Eglinton West bus loses more riders to southbound demand, especially at Kipling, than it gains from northbound demand. This is expected to shift with the LRT in operation with some riders opting to stay on Eglinton would would otherwise travel south to Bloor.
Comparable numbers for the other options are not included in the report.
LRT vs Electric Buses
There is a passing reference to Electric Buses that warrants comment.
All options will be almost equally capable in reducing the GTHA’s contributions to climate change compared to the BAU due to the LRT’s cleaner energy source as compared to the traditional diesel bus, which the TTC recently aims to fully phase out beyond 2040. Besides, the LRT will likely produce lower well-to-wheel greenhouse gas (GhG) emission even when compared to electric buses with similar carrying capacity due to the former’s superiority in energy-efficient vehicle design. [p 52]
The issue referred to here is that any electric vehicle that includes an energy storage system wastes power on conversions from line to storage and back to motive power. This is not the case for a vehicle directly powered from overhead or third rail.
The initial capital costs are summarized in the table below and are stated in Year of Expenditure (YoE) as opposed to current dollars. Option 1 (surface) is of course the cheapest, although Option 3 saves a lot of money by crossing the Humber on a bridge and avoiding the cost of many stations. That is counterproductive as we have seen in the demand projections.
Option 1 has the highest cost for vehicle (and associated expansion of the Maintenance and Storage Facility) because the surface option has a slower speed (needing more cars to provide a given level of service) coupled with higher demand (needing more frequent service).
The ongoing operating, maintenance and capital renewal costs are also expressed in YoE terms. Because most of these occur in future decades, the cost is inflated relative to current dollars, especially for renewal items like replacement of the vehicle fleet in roughly 30 years.
The revenue effects table shows the distortion that comes by treating existing and net new riders differently. Option 1 has the highest demand, and therefore has the highest total revenue. However, Option 4 gets the most net new riders and therefore has the highest incremental revenue.
There is no consideration for the fact that some of the “existing” riders who come to Option 1 were formerly on Line 1 Bloor and by shifting them, space is available for other new riders.
The numbers above are collected into a single table which, unlike the source values, does include Net Present Value (NPV) adjustments to deflate future dollar values to current day. There are some strange things in this table.
First off, it is the total project revenue which is used, not the incremental value, and so at least here, total ridership takes precedence over new ridership.
Second, the ratio between the YoE values for Total Operating and Maintenance costs shown in Table 20 and the NPV values shown below are huge and well out of proportion to NPV adjustments for other factors. It is not credible that about $11 billion on a YoE basis reduces to $1.3 billion on an NPV basis. Similarly there is a very high ratio between YoE rehabilitation costs and the NPV values.
Unless this is resolved, the values shown below should be treated with suspicion.
In any event, the ROI for all options is low because of the low revenue benefit relative to capital cost, although the value is best for the surface option because it costs least and has the most demand.
Note that this analysis does not include any of the imputed benefits such as the value of travel time, or the alleged reduction in congestion from redirected auto-to-transit trips.