More About the Grid
It is pretty well accepted in the transit industry that the grid is the most efficient form of service delivery. In a grid, each street basically has one route, which travels up and down the street for its length with very few, if any, deviations. For more information about the value of the grid, please refer to my comments on network design . Also see my review of Jarrett Walker's book and his network design class . Walker is a notable transit consultant who, in some ways, makes his living by encouraging the transit agencies who hire him to make their systems more grid-like.
We know that the grid is the most effective way of designing networks, but what is a good relationship between grid density and route frequency? Of course, in a perfect world we would have a very thick grid filled with routes that all operate very frequently, as is the case in Toronto. Unfortunately, we do not live in a perfect world and the financial constraints most of our transit agencies find themselves in require that they pick between operating a thick grid with routes that do not operate frequently versus operating a thinner grid with routes that operate more frequently. Which is better?
Before we continue, I need to define what I mean by "thick" and "thin" grid. The cities that are suitable for grid systems generally have major arterials spaced at 1/2 to 1 mile intervals. In a thick grid, routes would be placed on all arterials, or about 1/2 mile apart; in a thin grid, routes would be placed at mile intervals. If all routes cost about the same to operate, then that would mean that we could operate two routes on a thick grid at thirty minute frequency for the same cost that we could operate one route on a thin grid at fifteen minute frequency. Which would be better?
The answer depends on the market we are trying to serve. In the United States, despite constant efforts to attract green passengers and other choice riders, the vast majority of ridership in most markets is made up of the transit dependent. While some transit dependent riders could walk farther to a better bus route, many transit dependent riders are senior, disabled, and even single mothers with several children in tow - all groups that would find it difficult if not impossible to walk another 1/2 mile to a bus stop. Some of these riders would be forced onto paratransit as a result of route changes, at an extremely high cost to the transit operator.
It is because of the above riders that transit systems will continue to operate a larger number of routes with poor service, even as they know that frequencies of thirty minutes or less will never attract the kind of choice riders we need to attract to transit to improve the sustainability of our cities.
Perhaps a bigger problem than frequency in many cities that operate a grid system is that the grid has not been extended to keep pace with development. For example, in Phoenix (which has an extensive grid system of buses), the farthest west north-south bus route is at 67th Avenue, despite the fact that east-west buses go as far west as Litchfield Road, which is the equivalent of 139th Avenue. The farthest north bus route is on Union Hills Drive (18600 N), despite the fact that development now reaches Happy Valley Road (25000 N). In the southeastern portion of the metropolitan area, while there are some routes that are large holes in the grid, with up to three miles between bus routes. While most routes in Phoenix are one mile apart, there are certain areas, mostly in central Phoenix, where bus route are one-half mile apart. What if we eliminated Routes 1, 12, 15, 52, and 62 and used the resources to operate new routes in parts of Phoenix that have no service? All areas of the city that currently have bus service would continue to be an average of 1/2 mile from both east-west and north-south bus routes, while areas of the city that currently have no bus service would be accessible by transit for the first time.
In a similar vein, Las Vegas's grid network has failed to keep up with growth, with the farthest north east-west line only operating along Lamb and the farthest west north-south line only operating along Rainbow, despite significant development north and west of those streets. In addition, existing lines have for the most part have failed to be extended to areas of new growth.
Phoenix and Las Vegas both are in the process of trying to attract more choice riders, with Phoenix planning several extensions of its existing light rail line and Las Vegas in the process of deploying several BRT-light routes. However, before you can attract choice riders you need to make sure your transit dependent population is adequately served, because few people will want to purchase a product that does not work effectively even for those who require it. Before you can focus on patronage goals - designing services that will attract a large ridership - you need to meet your coverage goals by designing services that provide a basic level of service for everyone. Otherwise, you will never attract a large ridership because they cannot get to where they would like to go on your network.
Extension of the grid to currently unserved areas is also important in overcoming the spatial mismatch , which is described more here.