Vanishing Automobile update #17

How Dense Is Dense Enough?

How dense is dense enough for smart growth? Smart-growth advocates in the Twin Cities (average density: 1,800 people per square mile) say the Twin Cities needs to be as dense as Portland (average density: 3,000 people per square mile).

Smart-growth advocates in Portland say that Portland needs to be as dense as Los Angeles (average density: 5,600 per square mile). Smart-growth advocates in Los Angeles say that Los Angeles needs to be as dense as Chicago (average density: 12,000 people per square mile).

Smart-growth advocates in Chicago say Chicago should be as dense as San Francisco (average density: 16,000 people per square mile). Smart-growth advocates in San Francisco want the City to be even denser still. Where will it all end?

Fortunately for a puzzled nation, the Sierra Club answered this question in a web page that supposedly calculates the environmental impacts of density. You enter your preferred density in households per acre along with your idea of average automobile fuel efficiency (miles per gallon) and the price of gasoline. The Sierra Club then projects the environmental and social impacts of your density. For comparison, it includes the environmental and social impacts of the "efficient urban density" and a "sprawl density."

When originally posted on June 18, the web site indicated that the efficient urban density is 500 households per acre. Since the U.S. has an average of 2.4 people per household, this represents 1,200 people per acre or 768,000 people per square mile. This indicates that Manhattan, at only 52,000 people per square mile, has a ways to go before it reaches smart-growth perfection.

Demographer Wendell Cox points out that this is denser that the densest parts of Mumbai (Bombay) and Hong Kong. In fact, Cox adds, at this density everyone in the United States could fit into an area a little larger than Portland, Oregon's urban-growth boundary.

Perhaps in response to Cox's comments, on June 20 the Sierra Club modified the web page to compare four different densities:
* Dense urban, which is 400 households per acre or slightly less than the "efficient urban" of the day before;
* Efficient urban, which is "only" 100 households per acre;
* Efficient suburban, which is 10 households per acre; and
* Sprawl, which the Sierra Club defines as one household per acre.

Even the efficient urban density is incredibly dense compared to what most people are used to. One hundred households per acre is 153,600 people per square mile, or three times the density of Manhattan. The "dense urban" 400 households per acre is 614,400 people per square mile, or nearly twelve times as dense as Manhattan.

Of course, 400 housing units could fit on an acre in a twenty-story building, each story containing twenty apartments averaging a little over 2,000 square feet. Add four or five more stories for shops and offices and some underground parking and you have a nice dense city of twenty-five-story buildings. But few cities have large areas of twenty-five-story apartment/mixed-use buildings. Even in Manhattan, most residences are in four- to ten-story buildings.

But the so-called efficient density of 100 households per acre is scary enough. At that density, the population of the United States could fit in the Los Angeles urbanized area -- call it "Sierra Club City." The entire population of the world would fit into the state of Virginia.

The Sierra Club assumes that all or nearly all office and retail establishments would be mixed in with the residential areas. It calculates that the efficient density would provide 48 "shopping opportunities per acre," whatever that means, as opposed to just 0.65 opportunities per acre at sprawl densities.

To be fair, some additional land would be needed for factories, warehouses, and other industrial areas. But that would still leave most of the rest of the world for farms, parks, and wilderness, which of course is the Sierra Club's goal.

How much land would Sierra Club City save? At the present time, U.S. cities, towns, and other urbanized areas occupy about 109,000 square miles of land. Roughly a third of that is industrial. If Sierra Club City replaced the other two-thirds, that would allow the restoration of about 72,000 square miles of land to farms, forests, or nature preserves. That sounds like a lot, but it amounts to just 2 percent of the land area of the United States.

Of course, if Sierra Club City stacks industry in twenty-five-story buildings too, then up to 3 percent of the nation's land could revert from urban uses to open space. Whether saving 2 or 3 percent is worthwhile depends on the costs of high-density living.

Start with congestion. The Sierra Club says that people living in sprawl densities of one household per acre would drive more than 32,000 miles per year. But at the efficient densities, says the club, they would drive only 7,600 miles per year, less than a fourth as much. Of course, with 100 times as many people per square mile, that means that total driving would be nearly 24 times more per square mile in Sierra Club City than in sprawl.

Urban Americans drive an average of 40,000 miles per square mile of urbanized land each day. As the highest density urban area, Los Angeles also has the highest density of driving: 124,000 miles per square mile of land.

But residents in the Sierra Club's efficient city would drive 1.3 million miles per day for each square mile of residential area. That's 33 times more than in the average urban area and 10 times more than in Los Angeles.

Curiously, no matter what the population density, the Sierra Club model dedicates 93 acres of land per square mile to roads and sidewalks. It is not clear whether this is included in household per acre densities or is in addition. Assuming that it is in addition, then it represents about 13 percent of the land area, which is about right for suburban areas but is far lower than the percentage of high-density urban areas that is devoted to streets.

Ninety-three acres divided into twelve-foot lanes with three-foot sidewalks represents about 50 lane-miles of roadway. To handle 1.3 million miles of vehicle travel per day, each lane-mile of road would have to carry 1,100 vehicles per hour, twenty-four hours a day, seven days a week. Non-freeway arterial lanes can handle just that number, while freeway lanes can handle twice this amount, and lesser streets (collectors and locals) do less.

Thus, Sierra Club City will be about as congested as Manhattan during rush hour -- only congestion in Sierra Club city will be 24/7. The Sierra Club's "dense urban" density of 400 households per acre, of course, will be even worse: The club predicts that people will drive more miles per lane mile than the best freeways can handle.

With this much traffic concentrated in a small area, Sierra Club City will be one of the most polluted cities in the history of humanity. The city will produce less pollution per capita than in a sprawling city, but it will be far more concentrated -- and the health effects of automotive air pollution depend largely on its concentration.

Typically, the Sierra Club model crudely assumes that pollution is directly proportional to fuel consumption. The web site asks you to enter the average fuel efficiency you imagine for your city and it calculates the pounds of hydrocarbons (volatile organic compounds), nitrogen oxides, particulates (PM10), and carbon dioxide that will be produced. The Sierra Club doesn't estimate carbon monoxide emissions, but autos tend to produce about ten times as much CO as hydrocarbons.

The Sierra Club presents pollution in terms of pounds emitted per household each year. For carbon dioxide, which is implicated with global warming, total emissions may be crucial. But for many of the other pollutants, the problem is not total emissions but the concentration of emissions. Concentrations of carbon monoxide and particulates, for example, pose extremely serious health risks, while low levels can be tolerated and ignored by most people.

Sierra Club City will produce extremely dangerous levels of these toxic pollutants. Based on the Sierra Club's numbers, automobiles in sprawl emit about 100 pounds of hydrocarbons, a half ton of carbon monoxide, and 680 tons of particulates per square mile per day. But Sierra Club City will produce 2,340 pounds of hydrocarbons, nearly 12 tons of carbon monoxide, and more than 16,000 pounds of particulates per square mile per day.

Since the Sierra Club model assumes that emissions are proportional to fuel consumption, all pollutants will be twenty-four times as concentrated in Sierra Club City than in sprawl. But in fact, cars pollute more in stop-and-go traffic, and the extra congestion in Sierra Club City will make it even more polluted than the Sierra Club numbers indicate.

The Sierra Club model also claims that high densities will produce less water pollution. "When more than 20% of the watershed is paved over and developed," says the club, "water pollution skyrockets." But a suburban neighborhood of one household per acre will have much less than 20 percent of its area paved over, while an urban jungle of 100 households per acre will be nearly all paved over. Thus, we can expect the most pollution from the urban area.

The Sierra Club's model is optimistic about the effects of density on driving. The model assumes that, at any density, doubling density reduces per capita driving by 20 percent. This is about four times greater than can be observed by looking at U.S. urban areas. However, it is about the amount generated by studies that looked at driving habits of residents of individual neighborhoods of various densities.

The problem with such studies is that they usually fail to control for family size, income level, and other factors that influence driving. A disproportionate number of people in high-density areas are either poor or have no children. They either can't afford to drive or have decided they would prefer to use transit. But this doesn't mean that forcing a middle-class family of four to live in high densities will lead them to drive significantly less.

Even without this error, Sierra Club City -- a permanently congested and dangerously polluted area -- will be far less attractive to most Americans than sprawl. But smart-growth advocates will nevertheless press for increasing densities in virtually every U.S. city.

Supplemental Note

In response to the above report, the Sierra Club modified its web site again to include the following statement:

The densities labeled 'efficient' provide transportation, living, and work choices for residents and workers but do not represent a Sierra Club endorsement of a specific density level. That should ultimately be decided by the communities themselves. The denser neighborhoods in Paris and Manhattan are shown for comparison.

What does the Sierra Club mean by "communities"? Communities don't make decisions; decisions are made by individuals. It is most likely that the Sierra Club would like to see urban planners and other appointed and elected officials made decisions that everyone else will have to live with. That is what has happened in Portland, Oregon, a city whose planning is endorsed by the Sierra Club.

Rather than let city planners make decisions for other, the Thoreau Institute's position is that densities should be decided by the individuals living in those areas. That means initially letting builders build for the market and letting individual homeowners maintain their neighborhood densities through covenants protected by small (perhaps 100 to 400 homes) neighborhood associations.

Thoreau Institute | Vanishing Automobile | Vanishing Automobile Updates