The best way to minimize driving appears to be to develop in existing centers near the core of the metropolitan area, in areas of high destination accessibility where there are a whole lot of jobs near by...[O]ther factors like mixed-use and intersections and block size...fall into a second group that is less important than destination accessibility...
Density turns out as less important than land-use mix...If you’re trying to minimize vehicle miles traveled and maximize walking and transit, you’re better off emphasizing mixed-use and destination accessibility than just bumping up density. A dense development in the suburbs, far from transit and employment centers and stores, is probably not going to buy you much in the way of walking and transit use. Almost any development in the central city is going to be more efficient from a transportation standpoint. [Highlights added.]
Hm. That mostly makes sense, but the discussion of density seems a bit circular. Density doesn't matter much, but putting lots of people really close to lots of jobs and stores in the urban core is critical. Huh? How, in practice, do you do the latter without the former?
But Ewing's point is that residential density, by itself, isn't enough to relieve car dependence; while commercial density without people nearby is a wasted opportunity. And that means that complete communities -- where people live really close to jobs, stores, and services -- really do help curb car-dependence. And the best places to find that kind of "destination accessibility" are, unsurprisingly, close to major urban centers.
In short, Ewing's findings give a big analytical boost to efforts like Vancouver, BC's EcoDensity initiative and Seattle's Center City Strategy, both of which are designed to create new housing close to major urban job and commercial centers. Those kinds of efforts, according to Ewing's findings, are the best tools that cities have to curb car dependence.
Still, I have to wonder about the idea that density by itself doesn't matter much to travel patterns. Consider the research of Australian researchers Peter Newman and Jeffrey Kenworthy -- here's a recent example [PDF]. Their research also looks at car dependence -- but considers entire metro areas, not just individuals. And they find a clear and strong relationship between metro-area density and how much driving people do. Take a look, for example, at the following chart -- with metro-area density on the horizontal axis and transportation energy on the vertical axis.
(Density and energy chart - Newman and Kenworthy)
Newman and Kenworthy believe that density has such a strong effect on driving because of two fundamental travel constraints. The first is the so-called Marchetti Constant -- a long-standing and perhaps natural human desire to keep one-way commutes to about half-an-hour, a value that has held steady since Medieval times and perhaps longer. Once people have to travel more than half an hour to their workplace, they do something to get to work faster -- by changing where they live, where they work, or how they get from place to place.
The second constraint is the inherent speed of different modes of travel -- that is, how far you can travel in half an hour on foot vs. bike vs. transit vs. car. In big, sprawling cities, people just can't get from homes to jobs in under half an hour without a car. Muscles and buses just don't move fast enough! So low-density land use inherently drives up both car-dependence and energy use for transportation.
Because of these constraints, Newman and Kenworthy argue, density is a fundamental determinant of transportation patterns at the metro level. Whenever the aggregate density of a metro area falls too low, people's only option for keeping their travel time under the Marchetti Constant is to use cars for most trips.
This perspective -- that density has a fundamental impact on how people get around in a city -- is almost the inverse of Ewing's view that density is at most a side note to individual choices. How to reconcile the two? Maybe it's not all that hard. Density, as Ewing points out, is commonly associated with the features that really do affect transportation choices: big clusters of jobs and services, central locations, and short blocks and lots of intersections to make walking trips more direct. When lots of people live in or near complete, compact city and town centers, you get both density and "location diversity" in one package -- and, hence, lots of access without a lot time in cars.
This post originally appeared on Sightline Daily.
Sorry you've made it more complicated than it seemed to be. What Ewing said seemed consistent: "best way to minimized driving" 1. "develop ... where there are a whole lot of jobs near by" and "emphasizing mixed-use and destination accessibility"
It's not about creating artificial density but putting the jobs and the mixed-use where there is already density, or developing with density, jobs and mixed-use as priorities.
I could have done without the confusing intro. The rest of the article was great, thank you Mr. Williams-Derry.
I think the missing piece of your puzzle is that there just aren't that many low-density suburbs with built-in employment: low-density suburbs are generally built on the post-war "dormitory" model, deliberately a long way from anything resembling work or amenities. So the Australian study you're citing has a serious sampling bias: it deals with what exists, rather than the full range of possibilities. To make that comparison seriously, you'd need to look at explicit pairs of low-density dormitory-suburb or low-density mixed-development cities. Assuming you could find any of the latter. Or possibly they do exist, but not in "higher income" cities as graphed above.
On a related note, fitting a r-squared value to a non-linear regression is statistically extremely dubious, and doesn't give me great confidence in their analytical skills.
Interesting article. These arguments change considerably in complexity in South Africa. Densification is at present that touted darling of sustainable planning. However, we have several issues which complicate this seemingly straightforward approach. For one, Cape Town, is a city which has used planning as a means of enforcing Apartheid.
The legacy of this still remains along with calcified land-use planning management and archaic zoning schemes.
One of the main constitutional imperatives is to make cities inclusive - through densification as it so happens. The economic legacy of Apartheid, along with macro-economic policies and the effects of globalisation have driven property market prices through the roof which effectively makes developing affordable or even subsidised inclusionary housing incredibly difficult.
At the same time, jobs in the central city are generally high-skilled, and the knowledge economies do not absorb low or semi-skilled workers. Therefore the rate of labour absorption is limited. Due to globalisation, manufacturing industries have shrunk, effectively reducing the amount of jobs that could be available via this sector.
When the poor are moved into nearby areas, lack of employment and other social issues tend to result in the areas degrading. Investment in upgrading tends to boost the prices and the poor are then ejected to the periphery again. As we know, South Africa's transport system is still not brilliant and the cost of transport into the city is unaffordable to most of these people.
Essentially what this marks out for us in the South, is that the need for our professions to increasingly take on a more transdisciplinary approach in order to solve the complexities of the development issues we face.
Were that it were that simple in just choosing where to densify!
Yeah! High density dwelling doesn't make sense unless the dwelling is clustered around the dwellers' workplaces (or 'activity spaces', if you want to be more general)
To achieve that, I think you need to consider how to make those dwellings accessible to the people in those workplaces (by that I primarily mean affordable, but aesthetics plays a role as well). It's a complex equation, however: 'desirable' -> high demand -> high prices -> unaffordable -> undesirable
(What would the Sims do??)
I currently work at a place that requires about a 50 minute commute from door to desk, regardless of mode of transport (car, train, or bike). I am delighted to hear the Marchetti Constant makes me more than just a whingeing commuter!
I think there is a great option for developing outside city limits, if it were approached with the ferocity that suburban leisure communities developed around golf courses.
Linear City is a concept that can solve most dilemmas an existing city faces as it's city core vaporizes. Linear city is formed by incorporating the length of an interstate between two regular cities. Linear city expands layers of frontages upon the interstate transit system, first industrial/parking, next commercial/mixed office, next condo/high density living, then parks/schools/public works, and for the remainder of 1 mile from the interstate is filled with suburban row houses. This layered developing of undervalued land blighted by the interstate converts it into a mecca of the middle class trying to escape the city. But because it is linear, it solves the problem faced by high speed rail, the last mile dilemma. Anyone can walk a mile to the bus station, taking them less than 10 minutes to a train station. The train in turn can deliver passengers to the regular cities for greater diversity in jobs shopping or education, without congesting that city with additional car traffic. Linear City is meant for the blue collar worker, the single college student, the retiree who wants a country view without leaving the benefits of city life... 10,000 people per square mile can live comfortably in a low rise 5 story condo lifestyle. stretched 100 miles Linear City can house a million people. The family can be just as happy in a row house, and yet their yard is the countryside, less than a mile from home, guaranteed. Linear City does not provide city services to anyone past that mile. Still the rugged countryside nearby may find a home or two from the rich seeking an isolated ranch or cabin.
Ewing's observations here seem pretty straightforward and unsurprising to me.
What IS shocking to me, though, is that the overall trajectory for these factors that impact urban livability is probably monotonically downward -- at least based on my anecdotal observations in the cities I've lived in. And that, furthermore, no coherent strategies even exist to address this unfavorable trajectory -- in fact, it is not even on the radar of most urban planners. They seem to be cheerfully busying themselves with such stuff as building bike lanes to nowhere for the amusement of recreational cyclists and touting their supposedly green credentials for it. (And please note that I mean no disrespect to cyclists, seeing as I've been one myself, and completely carfree at that, for the past ten years.)