This blog is supposed to be mainly about coding, but for engineering-types, the efficient movement of people around a city is always an interesting puzzle, so though this blog post is off-topic, I don't think it's uninteresting (note: I'm not a transit expert; I'm just a transit fan who likes shouting out like a blowhard).
Well, anyway, now that there's a new(ish) mayor of Ottawa, there's been a lot of newspaper articles about whether the old mayor's light rail (LRT) plans should be continued or scrapped. Personally, I always thought that the old mayor was pretty dumb (or at least naive) regarding transit. It takes decades to put together transit plans and get contracts signed, so for the old mayor to cancel the original light rail contracts when he got elected was stupid. The plans may not have been perfect, but it was better than nothing, and by canceling the contracts, the mayor left the city with nothing for at least another decade (the electric light rail line proposed by the old plan sure would have been useful when gas prices went through the roof!). I always assumed that the new light rail plans proposed by the old mayor would be impractical, and everything I heard about it seemed to confirm that. But with all the talk in the newspapers, I finally took a glance at Ottawa's 2008 transportation master plan, and I can now definitively say that it's a bad plan. It's a boondoggle that will cost a fortune but not actually improve public transit at all. It was obviously designed by a car driver who took a look at some other cities, became impressed with the shiny trains and subways, and decided that Ottawa should get them too, without any thought about money, efficiency, or political realities.
The reality is that this is the age of the car. In the age of the car, trains only make sense in certain limited situations. It does not make sense to copy the transit systems of Europe because those cities were built before the age of the car. It does not make sense to copy the transit system of large North American cities because those cities are larger than Ottawa, and those cities evolved before the age of the car. Ottawa grew big during the age of the car, and it built a transit system appropriate for the age of the car with its Bus Rapid Transit (BRT) Transitway. As a result, Ottawa is now a leader in designing transit systems for the age of the car, and Ottawa can't really look to other cities for patterns to copy because most other comparable cities to Ottawa are behind it, not ahead of it. Ottawa must innovate in order to come up with appropriate solutions to its transit problems.
Cities that developed during the age of the car are notable because people live all throughout the city, and they might work all throughout the city. It's the urban sprawl model of urban development. It is true that cities in the age of the car still have a downtown core where a large minority of the population work, but it's definitely not New York where masses of people commute from the suburbs into downtown. In the age of the car, people on the east side will work in an industrial park on the west side, people in downtown will work at a shop in the south, or people in the south-end will work in the south-west. To be useful, public transit shouldn't take everyone to a central hub or to simply move people between two points--it has to be able to take a person from anywhere and move them to anywhere in a reasonable amount of time. The hub and spoke model doesn't work in the age of the car, public transit needs to provide a network or lattice that can move people from anywhere to anywhere. Buses are the most appropriate transit option in the age of the car because they can easily go anywhere in the city, and they can use dedicated busways to go long distances quite quickly. LRT is an obvious choice if you already have a bunch of unused rail lines sitting around (Ottawa dug up its rail lines a long time ago) or if you have a lot of extra money that you can splurge on indulgent infrastructure (Canada hasn't had extra money for decades). If those conditions don't hold, then you have to be a lot more strategic in when you build LRTs.
So although I love trains, you can't let the romanticism of trains influence your decisions. In reality, trains are expensive. In modern medium-sized cities, it almost always makes more sense to use buses for public transit than trains. As far as I can tell, these are the factors that must be considered in evaluating new transit projects:
- Do the proposed lines go anywhere interesting? In particular, they should go among high density residential, commercial, and industrial districts containing people who are too poor to own cars. Some rail advocates argue that a train station magically makes a place an interesting destination. I think this is false. In the age of the car, a train station can add to the attractiveness of an area, but you can't build something from nothing. If a real estate developer has a choice between building a condominium tower in a place with good amenities and car access (which is good for the 70% of a city's population who are car drivers) or building a condominium tower in a place with nothing except for a train station (which is only good for the 30% of the population who are either prefer a lifestyle where they are within walking distance to amenities or who are too poor to own a car let alone a condo), I think they'll skip over the train station. They may build a slightly taller apartment building in a place if there's a train station there though. It's also important that transit lines go through as many interesting places as possible due to the network effect. A line going from point A to B can carry people from A to B or from B to A. A line going from A to B to C can carry people from A to B or C; or from B to A or C; or from C to a or B (that's six combinations vs. two). A line going between more locations allow for a lot more possible journey combinations, making the line a lot more useful.
- How long is the waiting time? I think the rule of thumb is that transit riders feel that waiting for transit feels twice as long as it actually takes. So if you have to wait 10 minutes for a bus, then need to waiting another 5 minutes for a bus transfer, you'll feel like you wasted 30 minutes of your life even though you spent only 15 minutes waiting. As a result, a good transit line should minimize transfers and have vehicles that come often.
- Capacity: a line should have enough capacity to handle the expected ridership. Trains can potentially carry more people than buses, and they can board passengers more efficiently than buses. Ottawa does have a capacity problem in the downtown area in that the roads cannot handle the number of buses that pass through the area.
- Cost of construction: The cheaper the better. Also, exotic technology should be avoided at all costs. When you build any piece of infrastructure, you end up having to rebuild it every few decades as part of the maintenance. If you use standardized technology, you benefit from lower costs due to mass production since everyone else in the world is using the same technology. If you use exotic stuff, you have to pay elevated prices to rebuild the same custom, exotic stuff every few decades.
- Cost of operation: the most successful transit lines are the ones that make money (in reality, I don't think any transit line in world supports themselves without subsidies or financial wizardry--though they probably pay for themselves indirectly by improving the efficiency of the economy and reducing the need to build highways). Electric trains have a lower operating costs than buses because they require fewer drivers and can hold more people. There's a bit of a trade-off needed to get this lower operating cost for a train though--you save money by running fewer trains than buses, but that means that people have to wait around longer for trains than for buses. It's not clear to me how mostly empty trains or diesel trains compare to buses in terms of costs. Again, exotic technology like hybrid buses, diesel-electric trains, etc. often end up costing more money in the long run.
- Speed: I think the speed hierarchy is bus/streetcar, bus/tram with dedicated lanes, bus/lrt with dedicated lanes and traffic priority, busway/dedicated rail line, bus/lrt with grade separation (i.e. tunnel, trench, or elevated line), automated underground trains. As far as I can tell, trains do not go any faster than buses. The main determining factors in the speed of transit are whether a line is separated from car traffic and how many stops there are on a line.
- Risk minimization through diversification: things go wrong, so it's useful to have a backup plan. With trains, people are always throwing themselves onto the line or whatnot, and that leads to the whole line being backed up. Buses are often useful in that you can reroute them around traffic incidents. With multiple possible routes serving an area, a transit system can more easily withstand problems such as accidents, long construction projects, blackouts, demonstrations, etc.
- Every phase of the project must provide an incremental benefit: The reality of big infrastructure projects is that they often get canceled due to lack of money or a change of politicians. Although politicians can make 10-20 year plans, the plan usually only survives for 3-5 years. If that first phase provides significant benefits, then if you're lucky, the other phases might get funded. So each phase must be nicely self-contained and provide a benefit in and of itself.
So given these criteria, this is why I think the 2008 transit master plan for Ottawa is a waste of money:
- Do the proposed lines go anywhere interesting? No. The proposed LRT will follow the existing Transitway line, so it won't provide access to anywhere new.
- How long is the waiting time? Wait times will be longer, especially during its construction. Anyone crossing the city will need to make two additional transfers.
- Capacity: the LRT should solve the capacity problems on the downtown portion of the Transitway.
- Cost of construction: Any sort of grade separation like a tunnel will cost a fortune
- Cost of operation: I imagine the number of passengers will be high enough that the cost of operation will be lower than running buses along the Transitway.
- Speed: As far as I can tell, there's no reason to believe that LRTs will be faster than buses along the Transitway. Any speed-ups will be solely due to the tunnel through downtown. This tunnel will probably save about 5-10 minutes per trip that goes through downtown.
- Risk minimization through diversification: the LRT means Ottawa is actually increasing its risks of unusual disruptions. With buses, you can reroute buses around problems. Although a tunnel will mean fewer disruptions due to traffic, the use of trains means that every time someone throws themselves onto the tracks (I imagine it'll happen once every month or two), the whole system will grind to a halt.
- Every phase of the project must provide an incremental benefit: This is the big weakness of the plan. The proposed new LRT system only works if the whole thing gets built. If you only build the first phase, you actually end up with a worse system than you started out with. Let's say you build the LRT tunnel and then you run out of money. You've then spent a billion dollars building a train that doesn't go anywhere. Basically, you'll end up with a train that crosses downtown. Ottawa is not Toronto. There isn't enough density in downtown Ottawa to support a train that only goes from one of the downtown to the other. For people in the suburbs, you have to catch a bus for one part of your journey anyway, so it simply makes sense to stay on the bus and ride all the way to or across downtown. The only way to get them to ride the train would be to force them--bus journeys would purposely have to halt at the edge of downtown and everyone would be forced to transfer to a train. People going from the west end to the east end would have to make two transfers if not more. It's just a disaster. The LRT only works if it's long enough to stretch across the city, but I seriously doubt the funding will last that long.
So in my opinion, the proposed LRT plan involves spending two billion dollars to build a train that doesn't go anywhere new but simply cuts travel times by 5-10 minutes. Of course, if a problem develops in the plan and it isn't completed, Ottawa will instead end up with a train that increases everyone's travel times by 5-10 minutes due to increased transfers. This plan just seems like a huge risk and a huge cost with only a small possible benefit. If the risks or the costs were small, then it might be worthwhile. If the possible benefits were huge, it would also be worth it. But this plan has too many negatives. The old mayor used to claim that making transit faster would get more people to use it. This is one of those delusions by car drivers that they would switch to using transit if it could be made faster than driving. This is, of course, totally ridiculous. It's pretty much impossible to make transit faster than driving, and most people still wouldn't switch.
Of course, I have to admit that solving this transit problem is tricky. The most pragmatic plans aren't necessarily the ones that are politically feasible. I also don't have the numbers about costs and ridership, so I can't really make well-informed proposals. But in the interests of being constructive, I'll throw out some completely uninformed suggestions.
The primary transit problem that needs to be addressed by Ottawa is that the city has reached the capacity for buses in downtown. Any transit plan needs to address this problem. It would be nice if a proposed transit plan also provides other benefits, but the current system will run fine for another decade or so as long as the downtown problem is resolved. Although everyone wants to get rid of the Transitway through downtown, this is simply not an option in the next few decades. BRT is the fundamental backbone of the Ottawa transit system. If you mess with it, you might break the entire system, and that's too risky. So all planning decisions must start with the assumption that the Transitway will continue to go through downtown for the foreseeable future. Starting from that assumption quickly narrows down the number of feasible plans for solving the downtown capacity problem:
- A grade separated busway through downtown (it doesn't have to be a tunnel--a trench or elevated line might be cheaper and less risky) should probably be enough to solve the capacity problems because you can have really long bus platforms without having to be limited by the size of city blocks and traffic lights. Travel times through downtown would also probably decrease by 5-10 minutes. This plan would be expensive though and not very glamorous, so I'm not sure that it's politically feasible to go this route.
- A supplemental bus transit line through downtown to pull traffic off of the Transitway. Expanding the downtown portion of the Transitway into a four-lane bus highway would probably work but would never get political support. Having a second bus rapid transit line would probably also work (it would be something like buses that eventually go north or south will run along Laurier while east-west buses will continue to use Albert and Slater), but I don't think it would work politically either.
- Any supplemental downtown transit line would have to be light rail to be politically viable. One advantage of LRT is that it's easier to drum up support for them. Car drivers would rise up in arms if you converted part of a busy road into dedicated bus lanes. But if you convert part of a busy road into a dedicated LRT, people will often grudgingly agree (though they'll mutter about Toronto streetcars the whole time). A surface light rail line would probably be sufficient because it's cheaper and the LRT only needs to serve enough capacity to bring down the Transitway bus traffic to reasonable levels. Unfortunately, a short LRT will hardly have any traffic at all--though riders might save 3-5 minutes by taking the LRT through downtown instead of riding the bus, they'll lose all that time by needing to transfer to a bus later for the rest of their journey. You need to have a long LRT to have any hope of getting reasonable ridership. There's a problem with placement of the line though. Running an LRT along the Transitway probably makes the most sense, but is probably politically infeasible because Ottawa was originally going to do this back in 2006 but it was canceled out by the old mayor. It would look bad to simply revive the old plan. One possibility is to run a LRT along Carling and then through south downtown or centretown, and then somehow possibly down Montreal road. This LRT configuration would serve underserved neighbourhoods that are good candidates for densification, which will help drum up political support for the plan. It'll be comparatively cheap as a surface rail line. It would hopefully pull traffic from the Transitway though I admit that there's no guarantee about that. It also provides a nice diversity of transportation options for people in downtown. It wouldn't be very fast though.
- Some percentage of the traffic on the Transitway through downtown is caused by people traveling through downtown but not actually stopping there. It might be possible to pull some of the traffic off of the Transitway by developing the O-train into a proper diesel commuter train. The O-train can be extended both south past the greenbelt and north into Hull (not that anyone expects much ridership to/from Gatineau with only a single station, but until there's an actual train station on the other side of the river, no one in Gatineau is going to make appropriate transit plans that incorporate that possibility. Plus, responsibility for the O-Train can then be dumped on the NCC and outsourced). There's a bit of a risk that a commuter train will lead to more urban sprawl, making transit problems worse, but these sorts of risks are impossible to predict. It might also be worthwhile to build an experimental commuter train line that runs from Orleans, to the Via train station, to Billings Bridge, and then up to Bayview. I'm not sure if the O-train line has the capacity and signaling to handle the traffic though.
- It might be possible to gamble that the future development of the city will lead to a change in traffic patterns as jobs and housing start distributing themselves more evenly around the city instead of having the jobs mostly be in downtown and housing mostly be in the suburbs. In that case, it makes sense to ignore the problem in downtown and focus on making sure that there is a sufficient lattice of BRT lines crisscrossing the city to allow for the efficient movement of people from anywhere in the city to anywhere. If that is the case, it might make sense to create a Transitway ring around the city (possibly two--one inside the Greenbelt and one outside).