Is it true that while everyone loves Portland’s regular 200-foot street grid, urbanists are turning away from it as something to emulate?
Daniel Nairn, who just wanted to make a nice nerdy poster about street grids, points me to a fascinating Planetizen article by Fanis Grammenos and Douglas Pollard. It argues that the standard street grid, an easily repeated pattern where most intersections are four-way, is and should be history. The future, they argue, lies in more complex grids where there are a lot of street connections but where 3-way “T” intersections are the rule. It’s an excellent article. Read the whole thing.
Grammenos and Pollard indict the standard Portland grid on the following counts:
- Inefficient use of land. Portland’s closely-spaced grid puts 42% of all land in the street right of way. That leaves not much land for development.
- Aesthetics. Any number of artists and theorists can agree that overly regular grids are boring.
- Safety. To optimise safety of traffic flow, the authors claim, a grid
made of three-way intersections is ideal, as in this ITE diagram, which will look familiar to anyone who’s studied greenfield Transit-Oriented Development.
Now, I’m not an urban design specialist, but I did grow up in Portland, and I’ve also tried to design transit networks for a huge range of cities. From those perspectives, something’s wrong here.
To take the three points:
- Inefficient use of land. Portland’s grid uses 42% of all land for street right of way, but that’s a feature of the width of the streets, not the shape of the grid.
- Aesthetics. Humans “do not like endless vistas” in an urban context, says Andres Duany. True, but the endlessness of the vista is also a function of the width of the street. A grid where non-arterial streets are suitable narrow and visually interrupted won’t have this problem.
- Safety. If we’re praising the grid of 3-way intersections for its ability to calm private car traffic, are we perhaps still giving cars too much authority in determining this most fundamental aspect of our urban structure?
The crucial point missing from the Grammenos/Pollard analysis is legibility. Grids are easy to remember. They fit our brains. Anyone who’s navigated Manhattan knows that the shredded street patterns of Lower Manhattan, below 14th Street, take up far more of your brain than the uniform grid that stretches from 14th Street to beyond 125th Street. In the latter area, which is about half of Manhattan, almost any location is easily described in terms that look like co-ordinates: Second Avenue at 73rd Street, etc. You can hold a huge part of the map of the city in your mind with very little effort. With those simple co-ordinates, your mind can jump to whatever location you choose, and often plot a rational course to get there. In short, the regular gridded area feels conceptually available in a way that the labyrinth of Lower Manhattan is not.
Sydney, where I live now, is more like Lower Manhattan. There are grid fragments here and there but overall the street pattern is sheer chaos:
Three-way intersections are routine, and where these happen on the arterial network, they are always the worst bottlenecks. No mode of transport can expect a reasonably direct path from origin to destination. As a pedestrian I can almost do it, but it’s still a complex calculation, requiring me to assess the unique angles and slopes presented by each street or path I might use.
Grids designed for cars inevitably concentrate all the worst features of cars, but I remain unclear on why this means we should abandon the standard grid. We just need to tame it. Vancouver’s West End, for example, is a highly obstructed labyrinth for cars, but remains entirely legible to peds and cyclists, who can still follow, and remember, its simple pattern:
It’s not all that clear in the Google Earth shot, but most of the streets on this grid don’t offer Duany’s dreaded “endless vistas.” If you look up, you can see the mountains from any north-south street in the West End, but you can’t see continuously down the street, unless it’s a major arterial. That’s because the various obstructions that interrupt car traffic generally include some landscaping, so if you look a few blocks down a street, the ground-level vista is likely to end in some trees. It’s very pleasant.
The long side of the Vancouver block (145 m or 474 feet) (generally 500-700 feet) is too long for my Portland-trained eye; with the wrong architecture, it becomes an oppressively long wall. But the point here is that the legibility of the grid is available to the cyclist and pedestrian. It’s easy for either of these modes to select a travel path based on just a few considerations — far less work than the advanced calculus required to plot the most direct path through Sydney.
Finally, standard grids are very, very good for transit. Straight, continuous arterials mean that transit doesn’t have to turn. (Every time a transit line turns, it’s a little less safe, takes a little more intersection capacity, and increases the likelihood that it’s taking someone out of direction.) In this post, for example, I tried to lay a grid route network over Sydney, in which the routes continue flowing in the same general direction to the extent that the tortured street network allows. Obviously, it was a struggle:
I would never claim that a fundamental decision about urban structure should be made purely for the convenience of transit. But I’ve encountered, and lived in, a range of emotionally powerful grids — from the intimacy of Portland’s to the empowering simplicity of Manhattan’s. To abandon the four-way grid for a more choked lattice of three-way intersections seems to be me not just bad for transit, but bad for peds and bikes too. Fundamentally, it’s bad for our ability to grasp our city as an entirety, and thus to experience it as available, and as ours.