Sunday, July 29, 2012

Three Approaches to Suburban Transit: Washington, San Francisco Bay, and Toronto

North American cities were transformed by the arrival of suburbanization and the automobile. Postwar rapid transit development was forced to cope with this new reality. Built in metropolitan areas of comparable size, Washington’s Metro, San Francisco’s BART, and Toronto’s subway took very different approaches, and were met with starkly differing levels of success. BART decided to embrace the automobile, building many of its lines in highway medians and surrounding its stations with large parking fields and garages. Washington chose to promote development around many of its stations, including suburban business centres like Rosslyn, in order to generate walk-in traffic. The planning of the Metro went hand-in-hand with land-use planning, which sought to build high-density "urban villages" in the suburbs and direct suburban office development to the areas around stations. Toronto’s subway, by contrast, supplements traffic from adjacent development with a frequent and extensive network of buses to feed its stations, dramatically increasing their catchment area while still providing pedestrian- and development-friendly stations that aren’t swimming in a sea of parking.

Relying on the American Public Transit Association for U.S. figures and the TTC for Toronto (unlinked trips for easy comparison, 2008 figures are most recent directly comparable statistics), we can see the different results quite clearly:
  • San Francisco BART: 366,200 riders per day, 2,192 daily riders per km. 
  • Washington Metro: 950,300 riders per day, 5,554 daily riders per km. 
  • Toronto subway and RT: 1,176,000 riders per day, 16,800 daily riders per km. 

Washington’s combination of park-and-ride with intensive transit-oriented development at many of its suburban stations is clearly more successful than the Bay Area’s almost exclusively auto-oriented model. It has been remarkably successful at directing development to station neighbourhoods. Both systems make extensive use of highway rights-of-way, but Metro leaves the highway much more frequently to serve the heart of suburban business districts, while BART tends to stick rigidly to highway medians outside the urban core of San Francisco, Oakland and Berkeley. Several studies have been written comparing the different approaches of BART and Metro, but none have included the Toronto model: a suburban rapid transit system as backbone embedded within a dense network of high-quality bus service. Almost all of Toronto’s major suburban bus routes operate every ten minutes or better until the late evening. The same cannot be said for Washington or San Francisco suburbs, where buses routinely stop at the dinner hour and run every hour or less on weekends. A good suburban bus network greatly extends the catchment area of a subway station, which is critical in a relatively low-density area. They can also move far more people than any park-and-ride lots could bring. Even better, ridership generated by feeding the subway can justify frequent bus service that is also used for local intra-suburban trips.

El Cerrito del Norte, a BART station (Image: Google Maps)
Bethesda, a Washington Metro station (Image: Google Maps)
York Mills, a Toronto subway station (Image: Google Maps)

The difference between the systems is also illustrated by the ridership at a typical suburban station. El Cerrito del Norte is located in the suburbs of the San Francisco East Bay and is surrounded by large parking lots and garages. Its average weekday ridership is 7,633 (Source). Bethesda station on the Washington Metro is located at the heart of a dense, urban office and residential neighbourhood. Its average weekday ridership is 10,765 (Source). Though it is largely surrounded by forest and low-density single-family homes, York Mills has a ridership of 27,260 (Source), largely due to its connecting bus routes. In fact, York Mills is busier than all but two stations on the entire Metro network. A station with significant high-density residential and employment nearby will have a higher ridership than one surrounded by parking, but truly busy stations in the suburbs need frequent connecting buses.

Despite Toronto’s far smaller rapid transit network, it transports more people than Washington Metro and far more than BART. The Toronto model of high-quality bus service feeding rapid transit stations is clearly the most successful approach to providing transit in a suburban area.

Friday, July 27, 2012

The Rising Cost of Rapid Transit Construction

This post is shared with Urban Toronto.

Nearly every North American who travels to Europe returns with awe at the extent of their subway and intercity rail networks. Part of the explanation for their vastly more impressive infrastructure is greater investment, but another large part comes from a cost of construction that is often a fraction of that in many cities on this side of the Atlantic. The Province of Ontario has invested an unprecedented amount of money in transit as part of its MoveOntario 2020 program, but Toronto is getting far less than other cities have built with similar funding. Transit capital costs have been rising at multiples of consumer price inflation for decades and if that course continues, major transit projects will become unaffordable no matter how generous the government of the day.

The Inflation Problem
The existing Sheppard subway was criticized at the time for its high cost relative to previous projects. The 5.5km line opened in 2002 and cost less than $1 billion including a complex interchange at Yonge. Rob Ford’s recently proposed plan to complete the line from Downsview to Scarborough Centre would add 12.7 km of tunnel for a cost estimated by the TTC at $4.7 billion. That is an increase from $172 million per kilometre on the original project to $370 million projected for the current extension project, not including the seemingly inevitable escalations once detailed planning and construction begin. If subway construction costs had been held to CPI inflation, construction today should cost $221 million per kilometre.

The comparison with CPI inflation is not entirely fair, as various elements of construction, particularly energy intensive materials like concrete, have seen significant price increases. Nevertheless, continuing on a path of steady increases far above CPI inflation is not sustainable. The recent project’s planning studies alone were expected to cost $250 million to $300 million, close to a third the cost of the entire original Sheppard project, before a single shovel hits the dirt.

International Comparisons
Subway construction costs differ to a remarkable degree between cities, even within countries or economic areas. Torontonians searching for a city that spends more than their own need to look no further than New York, which holds the dubious distinction of most expensive construction costs in the world. The tiny 2.4km Second Avenue subway project is costing more than $4.5 billion, or $1.875 billion per kilometre. Many reasons have been cited, ranging from the catch-all of New York’s "uniqueness" to unusually difficult construction conditions. More probable explanations include byzantine union work rules and even the open secret of Mafia involvement in the New York construction industry. But just because New York does it worse than Toronto doesn’t mean that many cities don’t do it better.

The most often cited example of cheap subway construction is Spain. The most recent Madrid Metro extension cost $84 million per kilometre. There are differences from the Toronto subway, including single tunnel bores and somewhat smaller trains, but the differences are not fundamental to the operation of the system and certainly do not justify a four-fold difference in cost. Lest readers question whether passengers are forced to board trains in dank caves, many Spanish subway stations were designed by celebrated international architects, including Norman Foster. It is not decoration and design that add significantly to the cost of a station. Instead, it is the choice to build deep caverns and vast mezzanines, along with project management and other ancillary costs. Though not a subway tunnel, Barcelona was able to build a 5.8km high-speed rail tunnel running directly beneath the streets of the city centre, including a complex underpinning of Gaudi's Sagrada Familia, for €179.3 million (C$222 million).

A Madrid Metro station. Source: FCC Construction 

Other cities are more expensive than Madrid and Barcelona, but they are still far more economical than Toronto. The U5 extension in Berlin, for example, is costing about C$242 million per kilometre despite being built around multiple existing routes and through the most densely built part of the city, a far more challenging setting than Sheppard or Vaughan. The Paris Metro Line 14, opened between 1998 and 2003, was also built through the historic heart of the city; it cost about $173 million per kilometre.

We do not have to go to Europe to see substantial cost differences. The Montreal Metro extension to Laval was built from 2002 to 2007 and cost $143.2 million per kilometre, including a river crossing and fairly lavish stations. Though its trains are somewhat smaller, that difference surely doesn't justify a doubling of costs down the 401. The Canada Line in Vancouver cost $107.9 million per kilometre. It is fully grade separated and automated, the majority of the line is underground, and it includes two major water crossings. By contrast, the Eglinton Crosstown line is only partially grade-separated, including long sections where it is little more than dedicated lanes in the middle of the street. Despite its substantially lower standard of construction, it is costing more than two-and-a-half times as much per kilometre.

De la Concorde Station, Montreal Metro. Source: IBI Group

Potential Savings
The Canada Line is an excellent example of a completely different approach to rapid transit construction. Rather than being managed by the local transit organization, which may or may not have recent experience with similar major projects, it was built in by a private consortium with substantial international experience that consortium designed, built, and continues to operate the line. It was a successful example of a public-private partnership (many are not so successful), particularly because the private partner was responsible for most cost increases and was therefore very inclined to limit costs.

Some cost reductions are not without impact. The most infamous case in Vancouver was the long underground section along Cambie Street. The businesses on the street suffered for several years as the street was dug up using the somewhat more disruptive, but significantly cheaper, cut-and-cover method of construction. This was the method used to build the original Yonge and Bloor lines, but it has since fallen out of favour in this city. While a long section of dug-up street could be very harmful to the many businesses along the street in the downtown core, it would be much less problematic on many of Toronto’s broad suburban streets. A possible mitigation effort would be to develop better compensation arrangements in advance that would sustain businesses for the duration of construction.

In suburban Richmond, the Canada Line is elevated. This is another approach that was frequently used in earlier decades in Toronto, but has since been almost completely shunned. The York-Vaughan subway extension is a perfect example. The TTC is using expensive tunnel boring machines to dig deep tunnels under largely empty land. Certainly the section north of Steeles could have been built above ground for a massive cost saving. Many suburban business centres, like Scarborough Centre and Metrotown in Burnaby, are built around elevated transit lines. If the buildings were designed from the outset to accommodate elevated rapid transit, there is no reason why it should have been detrimental to the new Vaughan Metropolitan Centre. In fact, an elevated alignment was briefly examined in the TTC’s study, but it refused to consider an above-ground VCC station and therefore dismissed the resulting grades as inconvenient for operations. The study did not even include a cost estimate so that the savings could be balanced against the costs. The Sheppard extension to Scarborough is another place where a partially elevated alignment could result in substantial economies.

Elevated lines are out of the question on pedestrian-oriented downtown streets, but many of Toronto’s suburban arterials simply do not fit that description. Many are lined with the back fences of homes or distant set back buildings. They would see minimal aesthetic impact from an elevated line. 

A Canada Line elevated section in Richmond. Source: 24h Vancouver 

This is a subject that I will be returning to in much greater detail in the future. Doing a detailed examination of why Toronto’s costs are so much higher than those in other major developed-world cities would require detailed analysis of major project budgets, a task that would require a team of experts. Toronto’s current mayor prides himself on his penny-pinching reputation. This is a place where real action on cost control could make a major difference, both to the city’s budget and, more importantly, to the possibility of further rapid transit construction in the future. Instead of taking cost estimates at face value, governments and citizens must seriously examine why they have risen so rapidly and whether they can be reduced.

Comparative Approximate Costs of Subway Construction (in 2012 Canadian Dollars) 

  • New York (Second Avenue Subway): $1.875 billion/km
  • Toronto (Proposed Sheppard Subway Extension): $370 million/km
  • Toronto (Spadina-York Extension): $302 million/km
  • Berlin (U5 Extension): $242 million/km
  • Paris (Line 14): $234 million/km
  • Montreal (Laval Metro Extension): $163 million/km
  • Vancouver (Canada Line): $117 million/km
  • Madrid (Line 2 Extension to Las Rosas): $89 million/km

Saturday, July 21, 2012

The Transit Planning Process in Toronto and the GTA

This article is shared with Urban Toronto.

The OneCity plan flamed out as suddenly as it emerged, but it leaves in its wake a renewed discussion about transit expansion in Toronto and how to pay for it. As a collection of routes, OneCity merely chose to compile virtually every unbuilt plan from the last three decades, among them the remaining lines of Transit City, subways downtown and on Sheppard West, and some form of regional rail on the Stouffville and Weston corridors. Like many transit plans in the GTA, it drew on the planning work of decades ago and then added a few more lines without real study. As OneCity was buried, however, the phoenix of a more serious transit planning process is emerging from its ashes. Toronto council has instructed its planners to prepare a new Toronto Public Transit Expansion Plan, to be partly derived from a series of public consultations that would ascertain what the public really wants in terms of transit expansion. Of course, this new process is not without its flaws, including an exclusive focus on the City of Toronto, reliance on the same city staff that prepared many of the existing plans, and a lack of funding sources. Nevertheless, it is a positive step that could lead to a new era of serious transit planning in Toronto.

Friday, July 13, 2012

The VIA Rail Cuts

The federal government recently announced severe cuts to VIA Rail service in a press release with the Orwellian title of “Via Rail continues its modernization and takes action to better meet customer demand." The reductions will hit hardest on the transcontinental routes and in Southwestern Ontario. The Canadian from Toronto to Vancouver will be reduced to only two trips per week in the Winter, while the Ocean from Montreal to Halifax will be cut back to three trips per week from six. These cuts seem like a worrisome prelude to hiving off these vital national services into a privatized summer-only cruise train like the Rocky Mountaineer. Southwestern Ontario was equally hard-hit with the cancellation of trains to Windsor, Sarnia, Kitchener, and Niagara Falls. VIA claims that riders can switch to parallel GO Transit services and coordination with GO makes sense where they have become direct competitors. However, many affected communities aren’t served by GO trains, either at all or at the affected times.

These cuts mark the end of a relatively friendly disposition towards VIA on the part of the present government. They had continued the “Renaissance of Passenger Rail” investment programme, which included station renovations, locomotive rebuilding, and track improvements in the Corridor. With their new majority government, the Conservatives may be returning to their more traditional views on passenger rail. While in opposition, they introduced multiple Private Member’s Bills to privatize VIA in its entirety. These cuts certainly don’t go that far, but they’re a worrisome step. In 1990, Prime Minister Mulroney’s government cut more than half of VIA service, including the abandonment of passenger service on the historic Canadian Pacific railway that united the country. VIA has never recovered from that blow.

As is so often the case, these cuts take the easy way out of cutting service rather than taking steps to make operations more efficient. VIA’s frequently uncompetitive fares are a major deterrent to riders. A regular adult round-trip from Kitchener to Toronto costs an astounding $62. Even given the cost of parking in Toronto, that is far higher than the cost of driving for a single person, let alone for a whole family. VIA’s schedules are also far too infrequent on many routes, forcing travellers to drive when they would prefer to ride the train. A significant expansion of service would attract many passengers to existing trains, strengthening VIA’s operations in the long term.

According to its 2011 Annual Report, VIA’s “Direct costs per available seat-mile” in the Southwestern Ontario region amount to 19 cents. That means that moving a seat, whether empty or full, from Toronto to Kitchener or Toronto to Brantford would cost $11.40 excluding all overhead and, presumably, most station costs. Current plans to cut trips will not reduce these costs at all. VIA’s revenues average 14 cents per available seat-mile in the region, requiring a government subsidy of 21 cents per passenger-mile. Government subsidy is the rule in passenger service around the world. The exception comes only after massive capital investment to provide service of sufficient reliability and speed to compete with automobile and air travel. If VIA is not going to receive the investment required to make its operating budget self-sustaining, and the government is going to continue cuts to its subsidy, it may have to take a hard look at some of its operating practices in order to reduce costs without reducing frequencies and further depressing ridership. These could include reducing the number of staff on trains, relying on e-ticketing and automated machines at all but the busiest stations, and reducing the staff at major stations to direct people to tracks. This would hurt VIA’s admirable customer service standard, but it may be the only option in the face of continued reductions in the operating subsidy. Further fare increases would be devastating to ridership while more cuts to the already-skeletal network would leave little remaining. Lower operating costs just might permit lower fares, which could attract people back to VIA and finally create a virtuous cycle of ridership and service increases.

Many VIA trains consistently operate well below capacity, yet their fares are as high as peak-period services that are routinely sold out. While VIA has taken admirable steps toward yield management in recent years, there is still much room for improvement. Several years ago, after a strike scare, VIA offered a temporary across-the-board reduction of 50%. The late night train from Toronto to Kitchener, on which less than one of its five or more cars is usually occupied, ran almost full. While it might not work in all circumstances, a dramatic reduction in fares on less busy trips could result in higher, rather than lower revenues. VIA also has a price floor, which makes short trips like Kitchener to Guelph or London to Brantford unreasonably expensive. Such short trips could be a good source of passengers for otherwise-empty trains.

VIA Rail could also adapt its reservation system so that its trips are displayed on an even playing field with airline trips on consumer and business travel search engines. Many companies require business trips to be purchased through services where train trips are not currently an option. In the more distant future, better connection with Pearson and Trudeau airports would permit through-ticketing with airlines and even code sharing with the use of flight numbers on VIA trains. This practice occurs in many countries, including the United States, France, the Netherlands, and Germany.

Some of the hostility to federal funding for VIA Rail in parts of the country is understandable. VIA’s network is clearly centred on Central Canada, and all of Western Canada sees a handful of train trips per week. There are a number of corridors, such as Edmonton to Calgary and Victoria to Courtenay, that are very suitable for intercity rail service. Investing in these lines would give VIA a real presence in Western Canada and would help build a broad national base of support for greater investment in passenger rail.

These cuts are a bad precedent for the future of VIA Rail over the next several years. While it is fortunate that the core Toronto-Ottawa-Montreal corridor is seeing some investment, the cuts to transcontinental trains suggest that they may go the way of train service to Calgary and Regina and VIA looks to be all but abandoning the potentially lucrative Southwestern Ontario market.

Monday, July 9, 2012

The Density Fallacy

Museum Station is adjacent to some of the most important destinations in the city, including the Royal Ontario Museum, Gardiner Museum, and part of the University of Toronto. It is also surrounded by dozens of office buildings, condo towers, hotels, shops and restaurants. Within 500 metres of York Mills station, you will find one office building, a golf course, a forest, a handful of small condominium buildings, and estate homes on large lots.
Which station is busier? It would be quite reasonable to assume Museum, but in fact York Mills sees more than three times as many riders per day. York Mills’ 27,260 riders make it one of the busier stations in the system. (Source) This isn’t an altogether fair comparison, since Museum’s is one of the least busy downtown stations (8,220 riders per day) and it shares much of its natural catchment area with other stations. Nevertheless, this clearly shows that surrounding density isn’t the only factor in determining how well used a subway station will be.
Our current planning process places an enormous amount of emphasis on the number of jobs and residents within 500 metres. That’s no doubt very important, and several subway stations are quite busy based entirely on their local walk-in traffic (i.e. North York Centre). It ignores the importance, however, of connecting bus routes. York Mills is well used because the 95 York Mills is a very busy bus route. Perhaps the greatest strengths of the TTC are the excellent connections between bus and subway and the relatively high quality of its suburban bus service. Riders can easily transfer between their local bus service that serves their home directly and faster subway service that allows them to make longer trips in a reasonable time. Good bus connections dramatically extend the catchment area of a subway line. Intense development around a station is good to have, but a station can be very busy without much in its immediate neighbourhood as long as it connects to a busy bus route.

Friday, July 6, 2012

The CityRail Concept: Real Regional Rail for the GTA (Updated)

This article is shared with Urban Toronto.

The introduction of genuine regional rail in the GTA would be the most revolutionary transit improvement in the region since the opening of the original subway in 1954. Toronto woefully underuses its comprehensive network of rail corridors connecting all of its suburbs with the central city and with each other. Right now, most corridors see about a dozen mammoth trains a day shuttling commuters from vast parking fields to downtown office buildings. Regional rail is an entirely different type of service in the same corridors, more akin to a subway or other rapid transit line than to locomotive-hauled GO bi-levels. There are several key characteristics that define regional rail: frequent service all day and every day; full fare integration and seamless connections with local transit; and electrified multiple unit trains that allow fast acceleration and frequent stops.

CityRail Plan for Toronto, cartography by Craig White
CityRail Plan, cartography by Craig White

Genuine regional rail will not occur as a result of piecemeal improvement of existing GO service. It must be developed through a comprehensive plan that guides all of the improvements required to achieve a desired standard of service: the CityRail concept. GO currently adds trains as demand requires and as freight railways permit. This is fine for a commuter service, but it’s not how rapid transit works. Imagine if the TTC cut back service on the subway to every half hour on Sundays because the trains weren’t running full. CityRail establishes the goal of real rapid transit service throughout the GTA and describes the necessary infrastructure improvements that will be required to achieve it. This clear final product makes the promotion of the concept much easier for transit agencies, both to the general public and to higher levels of government.

The initial CityRail concept involves the creation of rapid transit lines in all of the existing GO Transit rail corridors out to the edge of the urban area. Riders on CityRail lines will enjoy a minimum standard of service on all lines.
  • CityRail would have trains at least every 15 minutes from morning until late night every day, including weekends, so that riders can go to a station without consulting a schedule and know that a train will arrive shortly. This allows the system to be used as a rapid transit backbone in both Toronto and the 905 suburban municipalities.
  • CityRail services would have fully integrated fares with local transit service. Riders would transfer to and from CityRail as seamlessly as they currently do between TTC subways and buses. Fares would be the same for a given trip regardless of the mode chosen. Local transit routes would be routed into regional rail stations as they currently are to subway stations.
  • CityRail trains would be operated with electric multiple units that produce no air emissions and accelerate far more rapidly than existing GO Trains, permitting frequent rapid-transit-style stops without longer travel times than GO services.