Read an Excerpt

Urban Transformation

Understanding City Design and Form

ISLAND PRESS

To Compare

Cities, Size, Scale, and Form

Coastlines, lakeshores, rivers, and mountain ridges are chiefly responsible for the shape of cities. Views from space suggest that even the largest of the world's great cities were shaped by their natural location. The ground plain for the city of Los Angeles originated through crustal upheaval, and the rivers that emerged on its slopes have almost disappeared, but on a satellite image the shape of the large drainage basin that the city occupies is impressive and understandably its most important form-giving element.

Only recently have new parts of Cairo started to climb away from the Nile River valley onto the rim of the Sahara desert, where they have begun to encroach on the pyramids. Likewise, Shanghai has only recently crossed the Huangpu River, even if future satellite images will almost certainly show it stretched out onto the large fertile plain of the Yangtze River Delta.

As urban population grows, it will be highly relevant to watch the cities of this world from space—and not only the expanding cities of the developing world but also the dispersing cities in the developed world. If one were to imagine for a moment that it would be possible to direct the growth of cities in the developing world—that is, not to stop the influx of rural migration but to direct the renewal and expansion at their outskirts—future satellite images would show a web of linear gaps in settlement patterns, where now continuous urbanization occurs. These gaps would coincide with the existing water drainage patterns.

For reasons that we well understand, new urbanization would stay at a distance from water: from creeks, rivers, bays, and estuaries. Of all physical measures, the preservation of land near water would provide the greatest benefits for human health, health of vegetation and animal life, quality of the air, and a more comfortable climate.

The same understanding of natural systems would direct the dispersed cities in the developed world. In both worlds, the result would lead to a better integration of cities with the forces of nature. This concern for the integration of natural processes into city design will surface again in other chapters of this book.

The second major theme of the book relates city form to human experience. The comparison of city maps in this chapter helps us reflect on the quality of life in cities. The primary urban resource, the amount of space available to each individual, is stunningly variant. The comparison between the maps should not lead us to the conclusion that more space equals a higher or better quality of life; rather, we should conclude that land is a precious resource, to be used deliberately and not to be wasted.

At first disbelief sets in when the maps of the San Francisco Bay Area are shown next to Hong Kong, and next to a map of the Randstad, the urban concentration in The Netherlands that forms a ring and includes Amsterdam, Utrecht, The Hague, and Rotterdam. The comparison suggests that all inhabitants of the San Francisco Bay Area could live together in Sausalito and the urbanized areas of a small county to the north of San Francisco under conditions Bay Area residents would find extremely uncongenial, that is, if they were to live on the same land area and at the same density as the seven million people of Hong Kong. The rest of the Bay Area would be unpopulated. When shown to Bay Area residents the comparison produces a sense of shock and disbelief. All three maps were reproduced at the same graphic scale; all three show urbanized regions that accommodate nearly the same number of inhabitants, seven million people. The comparison is made possible by computing the surface of each of the three urbanized areas and expressing it as a percentage of a fifty by fifty kilometer square.

In the year 2000 when this collection of maps was started, the population of the San Francisco Bay Area was approaching seven million people and covered 83.75 percent of such a square; Hong Kong has since exceeded seven million inhabitants, but it still covers only 6 percent of a fifty by fifty kilometer square, whereas the Randstad in The Netherlands with its 6.6 million covers 42.7 percent of the same size square. Cities in different cultures with relatively equal numbers of inhabitants accommodate people at strikingly different densities. The conditions associated with the high density of Hong Kong are quickly explained. They are chiefly related to history, the influx of refugees from mainland China, and the political status of the former British colony. But the reasons for the high density are also explained by reflecting on Hong Kong's topography. Most of the available and buildable land area is urbanized. The same is true for the San Francisco Bay Area; topography defines the extent of urbanization. Since the Bay Area has so much more buildable land, it is used at a dramatically different density. The Randstad on the other hand makes for a different comparison; the Rhine–Meuse–Scheldt Delta has much open land available, but by comparison the cities of this urbanized region have only modestly expanded into the agricultural land. The collective political will has limited the extent of the urbanization. Historically, all land above water was precious and had to be protected from flooding. As a result, residents here have lived at relatively high densities because they have also lived in settlement patterns barely at sealevel, frequently below.

It is fair to conclude from the comparison of the three urban regions that human tolerance for density varies significantly, but the metrics of density as they relate to human ecology are not very well understood. Thus the comparisons between available space in cities and numbers of people using it allow us to interpret human population as part of the earth's ecosystem.

It is possible to apply other metrics, in addition to geographic size, to a comparison between these three urban settlements; for example, car ownership, average length of journeys to work, number of trips made by public transit, air quality, per capita energy consumption, human health statistics, distance to open space—in short, all measures that attempt to quantify sustainable life in cities. Some of this work had its beginning in a metropolitan world atlas. For example, the atlas compares trips by public or private transportation. Hong Kong residents use transit for 61 percent of their work-related trips, Randstad residents for 49 percent, and Bay Area residents for only 4.7 percent. According to the atlas, the carbon monoxide (CO) emissions for the Bay Area amount to 313.3 tonne per square kilometer, double the CO emissions of Randstad's 167 tonne. But Hong Kong's emissions are 757.9 tonne. In making these comparisons we might not always find direct causality in relation to density and spatial distribution of people because of the many and complex variables that also influence such metrics, but hidden in these comparisons lies a sustainability index that is waiting for discovery. A research group in Australia has correlated automobile use to density. The group at Curtin University has collected data on energy consumption for private automobile use for the world's one hundred largest cities and compared it to population density. This global Cities Database was a challenge to compile because of the variable methods used in compiling population statistics worldwide. The work has been criticized for this reason, but the conclusions are correct. The graphs show an inverse correlation between density and private energy use for cars. The densest cities in the world also use the lowest amount of energy for private transportation. The reverse is equally true: the cities of lowest density use by far the highest amount of fuel for private cars. This conclusion might be obvious. However, the magnitude of the difference explains many social, geo-political, and environmental conditions of our time.

The urban region of Atlanta, for example, with its four million inhabitants, consumes 103,000 megajoules (MJ) of energy per person per year for private transportation. Houston ranks close behind with 95,000 MJ followed by Denver, the San Francisco Bay Area, San Diego, Phoenix, Los Angeles, Washington, Chicago, and New York in descending order. These U.S. cities are followed by Australian and Canadian cities. The European cities used one fifth of Atlanta's energy, below 20,000 MJ per capita. Asian cities follow with 15,000 MJ descending to just a few thousand megajoules like Ho Chi Minh City. One needs to add up the per capita energy use for private automobile use for thirteen of the world's very dense and largest cities from the collection of maps in this book to come close to the per capita energy consumption for private vehicles in Atlanta. In fact such a list would include, in ascending order starting with the lowest users, Mumbai, Shanghai, Cairo, Beijing, Jakarta, Manila, Hong Kong, Tehran, Seoul, Taipei, São Paulo, and Tokyo. Cumulatively the per capita energy consumption for private transportation in these thirteen cities is equivalent to 100,000 MJ. In these thirteen cities transport-related energy is used primarily for public transportation but at a much lower per capita rate.

These are 1995 statistics; they will not stay the same. Residents in Chinese cities are motorizing rapidly. For example, the city of Foshan, a city of six million people in the Pearl River Delta, was registering 5,665 cars every month in 2007. The neighboring city Guangzhou has ten million inhabitants. In 2006, it was adding 274 cars to its streets every day, a total of 8,228 a month. These are averages. There was a month in 2007 when 16,268 cars were added in Guangzhou, which equates to 542 per day. But even if the private per capita energy consumption of Chinese cities would increase one hundred-fold, it would still be well below that of Atlanta. Even if all the growing Chinese cities increase the consumption of energy for private transport, which is happening, they would still be significantly below a comparable number of cities in the United States, Canada, and Australia. However, globally these comparisons raise much concern.

Comparisons between the world's largest cities are made possible by placing them into a fifty by fifty kilometer square. Most of the world's urban agglomerations fit into such a square. Some, such as Tokyo, Los Angeles, Chicago, and the San Francisco Bay Area spill significantly out of the frame. The tristate New York metropolitan area comes surprisingly close to fitting inside; it barely spills out. Moscow fits snugly; so does Paris, if it were not for its five satellite cities. London spills out a little. Sydney almost fits into the square, but it contains only four million inhabitants. The 3.4 million inhabitants of Melbourne have spread beyond this frame in the southeast. São Paulo, one of the world's most populous cities still, but barely, fits. The more populous Mexico City no longer fits. The rest, and these cities will be the most populous in the future, like Cairo, Mumbai, Calcutta, Shanghai, and Beijing , all fit amply into the fifty-kilometer square. And it is here, of course, where the transformation of the urbanized area will be most noticeable. These five cities—relatively small in surface area—have long bypassed the ten million mark; some are heading for twenty million inhabitants by 2015.

The maps follow a morphological definition of metropolitan areas. Independent of administrative boundaries, they show the boundaries of what is continuously urbanized. The maps were created from satellite information. Eight bands of information are sent back to Earth from satellites like Landsat 7; the combination of three spectral bands makes possible the mapping of urbanized areas. For example, the San Francisco Bay Area is shown here mapped according to such a multispectral analysis. The information is checked against aerial imagery and placed into a fifty by fifty kilometer square.

To select a frame of fifty by fifty kilometers was an arbitrary decision, but when the frame is consistently superimposed onto all maps, comparisons of scale and size become possible. For example, Rome can be compared to cities of similar population size or to cities with similar surface areas. The surface areas are also represented graphically and are shown in the lower left-hand corner below each black and white map. This simple device became the basis for comparisons in the following discussions on city size, scale, and form.

Collection of City Maps

Making of the Maps

In a global comparison involving many cities, it was not possible to verify on the ground the urban boundary delineation. Therefore, we checked the boundaries against visual information. In the case of the San Francisco Bay Area we compared the map with the 2007 image created from satellite data shown in figure 1.4, which we purchased from the NPA Group, Edenbridge, UK. In addition to the San Francisco Bay Area, we used Landsat 7 information to map Rome, Calcutta, Cairo, and Mexico City. These maps resulted from multispectral analysis, a method that uses data bands 7, 4, and 5 of the eight bands sent by Landsat 7 back to Earth. The information contained in these three spectral bands permits the delineation of the continuously urbanized areas. For the rest of the cities in the collection we used feature extraction analysis. From the European Space Agency 2001 publication, "Mega Cities," Geospace, Salzburg , Austria, we read information from their satellite imagery. This helped to create the black and white maps of Bangkok, Beijing, Berlin, Buenos Aires, Mexico City, Cairo, Calcutta, Chicago, Delhi, Hong Kong, Istanbul, Jakarta, Lagos, Lima, London, Los Angeles Manila, Moscow, Mumbai, New York, Paris, Rio de Janeiro, Rome, São Paulo, Seoul, Shanghai, Sydney, Teheran, Tianjin, and Tokyo. We also used imagery from the NPA Group in Edenbridge, U.K., which we found in the 2003 Oxford Atlas of the World. They include Los Angeles, New York, Santiago de Chile, Cairo, Teheran, Tokyo, and Sydney. In general, we preferred to make maps from data that we could compare to multiple sources; we found images in books and on the Internet. We georeferenced the images, compared the planar projection against a consistent source, the city map portion of the 2003 Oxford Atlas of the World and used a feature extraction method in a conventional graphic application program to check the boundaries. The majority of the maps in this collection were produced in this manner. The feature extraction method involved composite viewing or layering of two, sometimes up to four, satellite images for interpretation of visible features along the urban edge. We then traced the extent of settlements in a consistent manner. The comparisons between satellite imagery from different years made us aware of the pace of change in the rapidly expanding cities in this collection. These cities include Beijing, Shanghai, and Tianjin, but also the three cities in India, Calcutta, Mumbai, and Delhi, as well as Lagos in Nigeria. For example, satellite imagery from the year 2000 shows a highly compact Shanghai still neatly framed by agricultural land use; five years later urbanization punctured that line in many places. In many Chinese cities the process of urban expansion has continued at an accelerated pace.

Frequently, we had to remind ourselves that geographic information system application software had been designed to conduct analysis at a far greater precision than necessary for a comparison where the size, shape, and form of cities would be compared within a five-inch square. For example, to produce the "footprint of a city" in a black-and-white graphic convention, we needed to establish a relatively precise line defining the edge of urban development. However, in reality this line rarely exists unless it is coincident with the edge of a water body, or the toe line of mountains. On the ground, instead of a line, the edge of a city is frequently a zone, and therefore any line is an approximation. But given the large graphic scale of the representation the decisions where to draw the lines were obvious ones. More difficult to make out were the distinctions between constructed patterns and those found in nature. An established suburb under a dense tree canopy can be difficult to discern; likewise, hillside squatter settlements, with their fine resolution of built elements and irregular shapes, can easily be misinterpreted for patterns found in nature. In general, squatter settlements rarely show on political maps of cities; a high resolution is also needed to detect them on satellite images. When in doubt, Google Earth became a much used tool to "fly" along the edges of urbanized areas to view in detail the transition zone between urban and nonurbanized areas. But it is important to remember that Google Earth imagery shows conditions with a significant delay and not simultaneously with conditions that exist on the ground when the request to visit the Google site is made.

---

Excerpted from Urban Transformation by Peter Bosselmann. Copyright © 2008 Peter Bosselmann. Excerpted by permission of ISLAND PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---