Better Public TransportThe first change required is a change in attitudes. This must begin with the city government and include the public transport agency. In most places today, public transport is regarded as a second-class service for second-class citizens. This is why few people in the USA use it: everyone but the poor is expected to have a car and use it routinely. In a few cities, such as New York, the middle class does use public transport and reasonably good service is offered. More positive attitudes prevail in most of Europe.
Public transport must become a first-class service. It must be safe at night, clean, operated by professionals, reliable, and on time. It should be free of all advertising. Recently, "bus-wrapping" (ads that cover an entire bus) and animated subway advertising have become common. This is one more "second-class" stamp on public transport.
It is accepted in most jurisdictions that people will have to stand during rush hour. It is thought too expensive to provide seats for all. With buses, any increase in seating capacity requires more buses and more drivers. However, rail systems can operate longer trains to provide more seats, without increasing the number of operators. A few rail systems have been fully automated, so operating labor costs fall to zero. Whatever is done, most people should be able to find a seat nearly all of the time. Sardine-packed conditions should never arise.
One advantage of driving is ease with which you can take along practically anything you wish. We will not achieve this level of convenience with public transport, but we can get close by making it easy to take along shopping carts and baby buggies. (If the system is fully accessible to wheelchairs, then it is also accessible to small carts.) This requires level loading, with the boarding platform and the vehicle floor at the same height. This also shortens station stops, because loading up and down stairs is slow. Level-loading can be provided with newer, low-floor buses and trams if boarding platforms just a foot (30cm) above the street are provided.
Walking to the halt and waiting for a vehicle to come along are usually the two largest components of a trip by public transport. We need to bring halts to within about 500m (1600 feet) of every location in the city (the new standard for Beijing). This does not entirely solve the problem because the rider may not be going where vehicle is, which necessitates a transfer.
It is not practical to establish a route network in a large city that does not require an occasional transfer. However, we can hold transfers to a minimum and ensure that the actual transfer is quick and easy. This means short walks, no stairs, and short waits for the next vehicle. The best route network is usually a hub-and-spoke arrangement, with transfers occurring mostly in the downtown area, where service is concentrated and vehicles arrive frequently. This also implies the need to concentrate the most common activities in the city center (or to situate them in every neighborhood, as with grocery stores and elementary schools).
Good public transport service is too costly unless the population is clustered around relatively few halts. There is otherwise not enough demand to justify frequent service. The Reference Design for carfree cities concentrates a million people into 100 districts in which the public transport halt is no more than 380m (1250 feet) away. The necessary density was usual in most cities less than a century ago and can still be found in many cities. High density became unbearable in the USA during the 1920s, when private cars usurped huge amounts of space for parking and movement. At the same time, cars imposed a huge burden of noise, pollution, and danger. This is why, after WW II, American cities rapidly dispersed into low-density, auto-centric suburbs, a pattern now common around the world. Good public transport is impossible in the suburbs: destinations are too spread out and density is too low for any reasonable system to work well.
Service-quality considerations dictate the choice of rail systems over bus systems whenever possible. The quality of service is much higher with rail systems, and it is now clear that people perceive the difference. As I once said, half in jest, "Nobody with a choice ever took a bus anywhere."
If surface-running rail systems are chosen, they must be installed on their own rights-of-way, where they never compete with cars for road space. When Zurich, Switzerland, decided not to build an underground metro but to fix its trams instead, the first change was to move cars out of the places where they blocked trams.
The choice between trams and metros is argued endlessly, and the subject is complex. However, in cities with populations over about one million, at least a few metro lines are almost essential, as only they can provide rapid service over the longer distances that large cities entail. Trams can provide supplementary, lower-demand service in big cities and can be the principal mode in smaller cities. The smaller trains are actually an advantage given the lower levels of demand.
Capital costs of new metros in existing cities are extremely high, whereas tram systems can be installed comparatively cheaply. However, when demand warrants the high capacity of a metro, it can still be cost-effective due to the comparatively low per-passenger operating costs. A further advantage of the metro is that it can reach very high speeds on its protected right-of-way. Trams can exceed 50 km/hr (30 MPH), but urban surface vehicles should not travel faster due to noise and safety issues.
Very large cities need limited-stop regional rail service. Consider the RER in Paris, which links the suburbs with a few major stops in central Paris. New York runs express trains that serve a similar function. Once again, it is a question of speed. Local trains making many stops achieve quite low average speeds, which makes longer trips tedious.
If we are to ask people to abandon their cars, then public transport must be available at all hours of the night. Many larger European cities have buses that run at least once an hour during the night, and New York's subway never quits. Some level of night-time service is nearly essential.
Route planning is usually a tiresome chore for passengers. It can be surprisingly time-consuming and difficult to find a route to an unfamiliar destination. The Dutch have an excellent nationwide system that will plan any trip you can make by public transport (which is nearly all of them). I found it reasonably easy to use and highly reliable. This kind of Internet service is moderately expensive to establish but cheap to operate. It is a hallmark of a first-class system.
Likewise, information should be available at tram and metro halts regarding the time until the next vehicle arrives. This is complex to implement but has been done. This allows you to run an errand when you discover that you have a few extra minutes before the train arrives.
Faster ServicePublic transport service must become much faster than it is today. Ideally, it should be faster to take transit than it is to drive, which makes it much easier to persuade people to give up their cars. A number of conditions must be met.
Transit managers need to speed journeys. Seconds count. This is not the current attitude at most systems, where service is considered acceptably good if trips run less than five minutes late. But fast service can only be achieved if every aspect of operations is considered from a time perspective. For instance, it would be common in many systems to order new trains without carefully considering how long it takes to close the doors and depart. "Only" five seconds might be saved here, but even on a journey of moderate length, that can save the passenger a minute. When millions of trips are made every day, it adds up to a lot of saved time.
Increases in top operating speed have a smaller effect on stop-to-stop time than might be supposed, but it is still important. Metro trains should routinely reach speeds of 100 km/hr (60 MPH), even though they may remain at that speed for a fraction of a minute. Most of the energy can be recovered through regenerative braking.
More important is the achievement of high acceleration, which has a pronounced effect on stop-to-stop times. Research in the 1930s showed that quite high acceleration is tolerable, even for standing passengers, if it is smooth. Peak acceleration should take about one second to be reached, without any jerks. As long as this is achieved, acceleration somewhat above 0.2 G is acceptable. Even some modern trams only reach 0.1 G, which significantly reduces average speeds, but the PCC streetcar of 1935 achieved peak accelerations of 0.22 G.
Keep the vehicle moving. Station dwell times are a large part of total time. It should be possible to open doors on both sides of the train to allow faster boarding. If one side is used for disembarking passengers and the other for boarding, times are still further reduced, and there is less jostling for the passengers.
Fewer stops make for much faster service. This point is rarely appreciated. Cutting out a halt saves not only the time that the vehicle is stopped but also the time lost to braking and acceleration. Buses are chronically affected by this problem - they'll stop almost anywhere. Good practice keeps the halts at least 400m (1300 feet) apart, and 700m (2300 feet) is better. Reducing the number of stops also makes the journey more comfortable. Each district must have only one, centralized halt for a given route.
Finally, fare-free transit is worth serious considering. We will not go into the details here, but it saves a lot of time, money, and aggravation for the passenger, which makes transit a more attractive option.