CHAPTER FIVE ANDREA I. FRANK POPULATION DYNAMICS AND URBAN GROWTH PATTERN Cities are nodes of mans greatest impact on nature, the places where he has most altered the essential resources of land, air, organisms, and water. The city is the quintessence of mans capacity to inaugurate and control changes in his habitat. Through urbanization man has created new ecosystems within which the interaction of man, his works, and nature are complex. This complexity - and the importance of our understanding it - grows as cities burgeon in the modern world. --Marcus and Detwyler Introduction The Urban Age In view of societal evolution urbanism is a very recent development. Considering that in 1800 only a meager 3% of the worlds population lived in cities over 100,000 (Detwyler & Marcus, 1972) - the twentieth century may well be called the urban age. In 1969 Davis speculated that in 1990 more than half of the globe's population would be living in cities of 100,000 or more (Cities, 1969). This prediction was remarkably good. According to World Resources Data Base (WRD), as of 1995, 45.2% of the worlds population is living in urban centers. The portion of urban population in South America (78%), Europe (75%), North and Central America (74%) and Oceania (71%) and the former USSR (68.1%) lies well above this mark; whereas Africa and Asia still show percentages under 40%. Within each continent we find a wide range of urbanization for individual countries. It can be generalized that countries with a currently low percentage of urban population experience the highest annual change rates. Urbanization is therefore more significant in those countries than in countries with an already high percentage of urban dwellers. For example, Canada, the United Kingdom, Australia, the US or Sweden had a percentage of urban population of more than 70% in 1965. This figure did increase only marginally over the past 30 years. Countries with less than 30% urban population in 1965 experienced dramatic changes (Table 1).
Urban Population as a Percentage of Total Country 1965 1995 %Increase United Kingdom 87.1 89.5 2.4 Australia 83.0 85.2 3.2 New Zealand 78.9 84.3 5.4 Sweden 77.1 84.7 7.6 Canada 72.9 78.1 5.2 Turkey 34.1 68.8 34.7 South Korea 32.4 77.6 45.2 Honduras 25.7 47.7 22.0 Kenya 8.6 27.7 19.1 Botswana 3.9 30.9 27.0 Table 1- Country Comparison of Urbanization Rate
With the earth's growing population the continuation of the urbanization trend and of urban growth is more than likely to persist. Urban form and development vary a great deal. In view of the complexity of the system and the multitude of interacting variables it seems presumptuous to think that urban dynamics and growth can be controlled and governed. It is however assumed that certain measures and policies could help influence changes toward a more favorable and "harmonious" development and shape of urban centers. With most of our earth's population living in urban areas, the creation of a livable and environmentally sensitive urban habitat is not only a moral but a survival need. Increasing the understanding of the urban-population-environment dynamics might be a first step towards addressing this demand. An improved insight in those dynamics then might help to devise policy to manage urban development in a fashion that sustains the environment, preserves valuable natural resources and biodiversity and mitigates negative effects.
Urbanization: Factors and Problems The reasons for urbanization are manifold. One reason for the increase of the urban population is total population growth. The rise of mercantilism drove the urbanization in 16th century Europe (Hartshorn, 1992). More recently, urban growth stems from the shift of labor force from agriculture to industry and service sectors; the latter two are generally located in cities or urbanized areas, thus serving as attractors for human migration and agglomeration. Other motivations to move from rural to urban settings are job opportunities and education. The benefit of the city versus the country can be summarized as the maximization of stimulation, exchange and opportunity with a minimization of travel time (Register in Aberley, 1994). This is certainly one view. During the early stages of European urban history the city represented freedom. Still, city life today is associated with a sort of independent and anonymous living. The human-imposed order within the city boundaries suggests security and safety from untamed nature or enemy forces. Modern man's fear is derived from dangers within the city. In many cases the aspect of safety and order of the urban center has lost its validity. The inner city often is perceived as unsafe and dangerous. Other benefits, such as minimization of travel time, are often offset by the increasing physical size of the urban area. In a spread out multi-million person city that covers several hundred square miles it can take considerable time to get from one destination to another. Automobile travel is slowed down due to speed limit and traffic density. With no efficient public transit system in place it might take longer to cover the same distance than if one had to commute a similar distance between rural villages. Continued (rapid) urbanization generally poses a number of problems. Housing and infrastructure has to be provided for an increasing number of city dwellers. This requires planning and the resolution of land use conflicts. If the influx of new urban residents exceeds the city planners capabilities to plan lots and streets for new neighborhoods it often results in poorly planned and uncoordinated growth development. In so-called lesser developed countries (LDCs) with high population growth rates, migrants with little and no economic resources often settle illegally on vacant areas in, or adjacent to, large cities creating pouvre barrios or squatter communities. If the city does not meet the expectations for opportunities, jobs and wealth, dissatisfaction can lead to criminal or aggressive activities. Stimulation, while good and needed for human development also has a flip-side. Life in a dense and crowded environment also can pose enormous stress upon the city dweller, possibly leading to various health and behavioral disorders such as depression, illness or aggression. Cities use large amounts of energy, for industries and to heat and light thousands of homes. The energy is often created by burning fossil fuels, which leads to air pollution. Pollution of the air and water, human and industrial wastes causes negative environmental degradation and imposes on the quality of life and well being of urban residents. Managing growth in order to mitigate environmental degradation will be of utmost importance to ensure a livable urban setting for the future (Hartshorn, 1992). In summary, problems of urbanization are especially pertinent and difficult to address in rapidly changing cities of the third World (Cadman & Payne, 1990). These periods of rapid urbanization or de-urbanization can be regarded as periods of transition. Transitions are times of vulnerability; in the case of urbanization they are a threat to the integrity and vitality of a city (Drake, 1992). Rapid urbanization is potentially unhealthy and equally harmful to urban residents, the functioning of the city and the natural environment.
A Morphological Approach to Population Dynamics, Urban Growth and Urban Form Development of Urban form
Many studies of urban growth development and form in the past have taken a historical or a functional perspective. The historical perspective reveals a cycle of growth and decline, formation and restructuring of the city (Cadman & Payne, 1990). Towns and villages have been categorized according to physical growth pattern into cluster and linear band types or hierarchies of rank-size (Hartshorn, 1992; Christaller et. al.). The functional approach perceives cities as an organizing center, serving typically a predominant function such as banking, administration, or services. We distinguish trade, military, industrial or company towns and many more. Geographers also have categorized cities in terms of their spatial location as coastal or river cities and so forth (Hartshorn, 1992). While each of these different perspectives is interesting and valid, they fail to address the system dynamics of urban development in a more comprehensive manner.
Systems approach and Transition Theory Marcus & Detwyler et. al. (1972, 1992) suggest viewing the urban agglomeration as a dynamic system or even as an ecosystem. Systems are not static, but generally evolve over time. At times, however, systems change rapidly due to a changing systems component (i.e. population). In terms of urban form other systems components that trigger change are transportation modes and technology amongst others. From a systems perspective of view, periods of change can be viewed as temporarily bounded or transitional (Drake, 1993). Periods of rapid change will be followed by periods of relative stability. Transition theory assumes interdependence and relationship between transitions of different sectors such as agriculture, education, etc. This means the amplitude and time frame of one transition, i.e. urbanization is likely to be influenced by others, such as education. Implications for society may be enhanced or reduced, if the timing of transitions can be influenced by policy. Urbanization Transition and Urban Shape This study proposes to look at urbanization from a spatial point of view investigating the linkage of population growth, urbanization and urban shape. It will explore population changes in time and the corresponding spatial changes of urban form. The investigation is based on the premise that urban form is the physical manifestation of urbanization. Urban growth is a result of population growth, rural-urban migration and urban expansion. The author believes that the rate and time of change in those dynamics and the accompanying fringe conditions (landscape, technologies, climate etc.) will determine the urban shape (Figure 1). In a generalized way, these factors can be attributed to either the natural environment or the human society. Although the determinants for urban form of human origin, such as technology, transportation systems and culture may dominate, topography and land form are especially influential in early spatial patterns (Detwyler & Marcus, 1972). Figure 1 - Influences on Urban form
Towards a morphology of urban form The dynamics of urban-population-environment systems are complex. Transition theory at the global level and at the local level will be used to investigate urbanization trends. In conjunction to the urbanization data a number of factors that may influence urbanization, such as industrialization, transportation technology etc. will be explored. These factors are identified and linked to physical development pattern. It is assumed that different values of factor sets will result in different physical growth pattern. Furthermore, it is assumed that the timing of transitions in other sectors will influence the growth pattern of urban agglomerations as well. An attempt is made to develop a morphology of urbanization stages based on change and behavioral similarities at a national level. The categorization of urbanization stages will become a framework from which a morphology of urban development pattern at the local level can be devised. This local morphology probably will be correlated to behavioral pattern (i.e. traffic) and pattern in the environment (i.e. topography) if appropriate [Adams, ref. comment]. The author hopes that a general survey of structure and form of urban development could be helpful to derive a typology or classification of certain trends. In the long run, a morphological study may help to indicate development pattern that are preferable in respect to newly developed environmental value systems. In morphological terms a range of different classes of cities is imagined. A morphology could contain simple size distinction (small, medium, large) as well as functional-structural categories (old, new, dense, dispersed, high-tech, low-tech, etc.). In terms of development, transition theory will help to envision likely future development depending on the initial conditions of an urban system. Since technology and industrialization are thought of being major shape factors of urban growth this study starts developing an initial morphology of urban form with respect to the changes in transportation modes and urbanization. In particular individual motorized transportation is investigated for its impact on urban form. When linking urban growth and transportation technology the following scenarios for urban shape development can be constructed. A possible test for these morphologies, would be a study of urban shapes comparing urbanization and development pattern. Three initial morphologies for urban shape and change of transportation modes were developed: 1) Introduction of individual automobile transit prior to major urban growth > - increase in urban area - decrease in density - low overall urban density (except may be historic core) - transportation arterial location greatly influence urban form - depending on the size, satellite-like subcenters might form. The new urban growth is build for the automobile. 2) Introduction of individual automobile transits during the urban growth period > - a chaotic state (planners are caught in the dilemma between accommodating people or cars) - dense traffic - major pollution - urban area, that is structure in an ad hoc fashion probably with decreasing density - opportunities and dangers depending on economic situation and policy decisions 3) Introduction of individual automobile transit post urban growth period - major redesign of the city structure (retrofitting the city for the car) - relative high density - traffic congestion - pollution and noise (if not mitigated by policy) - can potentially degrade quality of the city structure Note: It is very likely that the effect of the transportation technology transition varies depending on the size of the city or town. A small town will be less affected by it than a large town. This is a matter of geometry and existing density. For these examples growth is assumed unrestricted by landscape, political borders, or policies. Doxiadis organic growth model also predicts distortions along transportation routes at advanced stages of urban development (1968). Figure 2 - Morphologies Urbanization Analysis The study uses various analysis techniques to explore the feasibility of linking urban transition and urban form to a shape development morphology. It looks at many different aspects of the problem, trying to understand it without however being exhaustive or comprehensive. Hopefully, some different views regarding urbanization can be presented.
Methodological Approach and Data The methodological approach in this study is twofold. Global and local level data will be investigated. This is to emphasize the scale dependence of the issue. First, a general investigation of urbanization at the global level is conducted. This is to establish a general notion of the linkage between population growth and urbanization While there is a global urbanization trend, we probably can observe the urbanization transition at different stages, its beginning, in full force or at its conclusion for different countries. The global urbanization trends are investigated looking at time series data from the World Resource Data for 42 Countries. Urban, rural and total population development are explored and displayed. Urbanization factors that are viewed as influencing urbanization and urban form will be investigated in relation to the urbanization transition in the hope to discover correlation and linkages. One of those factors is the level of technology available in the country. Secondly, local level data will be investigated in respect to urban development and urban shape. The urban development of 5 major US cities over the past four decades serves as starting point for this investigation. Then, a more in-depth study of urban shape is conducted comparing Boston, MA and El Paso, TX. The study is based on US census data. An attempt was made to link the results of the investigation to urbanization factors that were thought to be shape-determinants. Urbanization at the Global Level National Comparison of Urbanization Levels On the basis of availability of desired data, a total of 42 countries were selected from the WRD (see Table 2). Initially urbanization information was mapped on a world projection using ATLAS GIS for Windows. Two other factors were mapped in succeeding maps; the I_index and a correlation of the urbanization and I_index. The I_index was calculated to provide some sort of measure for industrial productivity by dividing Gross domestic Product (GDP) industrial share (%) by the workforce in the industrial sector (%). The index is adjusted through multiplication by total number of workforce and division by total GDP. Supposedly this index will give a indication of the extent of industrialization. With at high output of the industrial sector shown by a relative large share of the total GDP achieved by a small share of the workforce should result in a large index. It is assumed that when few people achieve a large proportion of GDP it is a result of automation and high technology. This would indicate a high level of technology. Map 1 presents a general overview. It displays the 42 selected countries according to the proportion of people living in cities. The distribution for the four ranges was customized for the purpose of this display. The 8 countries at the top of the list in terms of percentage of their total population living in cities are in descending order: Singapore, Belgium, Venezuela, Uruguay, United Kingdom, Netherlands, Denmark and Australia. In terms of land area, except for Australia, these countries are fairly small, suggesting that there might be a correlation of land area and degree of urbanization. Australia then would be an outlier, since despite large areas of land most people live in cities. On the other hand, Australia might be right in the ball-park, when using hospitable area instead of total land area. Due to the limited data used for this study, however, the author opted not to pursue any further investigations in this direction. The display of the urbanization transition for the 42 countries reveals three distinct pattern. Urban, rural and total population development is plotted over time (see Figures 3a, 3b, 3c).
Pattern A This pattern shows a declining rural population and an increasing urban population. Around 1990 to 2000 the total population growth seems to level or decline. Pattern A countries are reaching or have reached the end of a fairly typical urbanization transition. The urban population percentage stabilizes at a fairly high level of 70% of the total population or higher. Over the transition period the rural population dropped to a fairly low overall level. These countries typically have a high I_Index (Figure 3a). Pattern B Pattern B countries are at the beginning or in the midst of their urbanization transition. A high growth rate for the urban population is accompanied by a (soon) declining rural population. The total population growth rate is quite high (3%++). The end of the transition in this pattern is projected beyond the year 2010 or later (all data and projections WRD). These countries typically have a low I_Index (Figures 3b).
Pattern C Pattern C countries have a very low stable rural population. The urban population is however growing in linear fashion. There seems to be no leveling or change in that trend for the next two to three decades (WRD prediction). All population increase seems to occur in the cities. The countries with this urbanization pattern are Australia, Canada and the USA, three major immigration countries (Figures 3c). This suggests that a large proportion of the urban growth is due to inmigrating foreigners. Figure 3a - Pattern A Figure 3b - Pattern B Figure 3c - Pattern C Urbanization and Technology Technology and industrialization seemed to be major factors that drove urbanization in the past (Hartshorn, 1992). These factors had a great influence in changing land use pattern (Sinclair, 1967). A further investigation of the link of urbanization to industrial production/technology appears promising Map 2 displays the calculated I_Index (Industrial GDP/workforce industrial share). The high ranking countries are what is often designated as the Industrial Nations. The order is somewhat surprising with Norway at the top and the US ranked sixth. The listing might appear skewed since strong industrial nations such as Switzerland and Germany could not be included due to the lack of comparable data. Map 3 shows the correlation between the I_index and the percent urbanization for each country. The findings are interesting. There is a group of three countries Norway, Finland and Japan with a correlation factor of 1.5 or less, meaning that the percentage of urban population and the numerical value of the I_index are almost the same. A high correlation factor indicates a discrepancy of nominator and denominator. This means that despite a low level of technology and industrialization there is a high percentage of people living in urban areas. This means that urbanization preceded industrial development and must have been triggered by different factors. It also means that urbanization will occur under different premises than in the countries with a low correlation factor. Countries with a high factor are in descending order Pakistan, Sri Lanka, Honduras, Philippines, Romania, Egypt, Panama, El Salvador, Costa Rica, Turkey, et. al. Most of those countries are at the beginning of their urbanization transition (Figures 3a-c, 4a, 4b, 5a, 5b).
A complete list of countries with their ranking of urban population, I_index etc. is shown below. Rank # Country Urban Popin % I_Index % Urban/I_Index 1 Norway 77.002525 64.9283 1.1861141 2 Japan 77.943899 55.4794 1.4051541 3 Canada 78.137155 51.4949 1.517521 4 Belgium 96.650384 47.3575 2.0411908 5 Finland 60.285375 46.0019 1.3105516 6 United States 76.237184 45.9387 1.6598559 7 Sweden 84.737262 45.6385 1.8570515 8 France 72.788174 45.0832 1.6146445 9 Netherlands 88.915414 44.6254 1.9927255 10 Denmark 85.477658 41.7588 2.0473691 11 Italy 70.540494 40.9309 1.7247065 12 Austria 60.628419 38.8964 1.5589719 13 Australia 85.167412 37.2962 2.2839209 14 United Kingdom 89.461725 34.704 2.5781477 15 Spain 80.652307 27.9247 2.8886929 16 New Zealand 84.290541 27.6005 3.0190022 17 Singapore 100 19.9111 5.0226017 18 Mexico 75.298388 18.8261 3.9996807 19 Venezuela 92.87809 16.3011 5.6980423 20 Uruguay 90.301318 13.6092 6.6398028 21 Trinidad&Tobago 66.50038 13.3537 4.9880179 22 Greece 65.034624 12.9655 5.0181037 23 Korea, Rep 77.628259 12.3138 6.3061137 24 Guatemala 41.465022 11.3321 3.6597548 25 Ecuador 60.615801 11.316 5.3594873 26 Colombia 72.721575 10.9699 6.6351802 27 Ireland 58.402998 10.6801 5.4684455 28 Portugal 36.361797 9.92898 3.6655037 29 Paraguay 50.684652 7.95855 6.3754279 30 Poland 63.868236 7.62871 8.3816583 31 Jamaica 55.359246 6.92128 7.9998911 32 Turkey 68.767733 6.86745 10.024451 33 Panama 54.870252 5.02267 10.930329 34 Costa Rica 49.707944 4.94479 10.144478 35 Romania 56.183791 4.44702 12.625571 36 El Salvador 46.67129 4.41119 10.583059 37 Egypt 44.763239 3.81796 11.748882 38 Indonesia 32.529271 3.63699 8.936613 39 Philippines 45.677693 3.61282 12.653101 40 Honduras 47.654155 3.47418 13.733186 41 Pakistan 34.692607 1.80723 19.273671 42 Sri Lanka 22.3809 1.58441 14.165127 Table 2: Industrial GDP/Industrial workforce Index
Different Urbanization Catalysts As a consequence of the map evaluations, the countries with the extreme values of the urban - I_index correlation appear to be interesting in terms of transitions. The charts below show the urbanization transition for Norway, Finland (%Urban/I_Index factor of 1.18 and 1.31) and Pakistan, Turkey (%Urban/I_index 19.27 and 10.02). The most remarkable difference between the two pairs seems to be that the energy consumption seemed to raise parallel to the urbanization in Norway and Finland. Both countries show a similarly high per person consumption of energy of about .2 Terajoules in 1991. For Pakistan and Turkey the per person energy consumption in 1990 is about 1/10 of the energy consumption observed in the two industrialized nations. Turkeys urbanization level (68%) is higher than Finland's (60%). The climatic differences may account only for some of the difference. It is assumed that this difference in energy consumption will lead to a dramatically different urban shape. Since data for city density, population and shape is difficult to obtain in the short period of time available for this project the impact of technology, i.e. the availability of the automobile (expressed also through an increased energy consumption per person) will be investigated in a modified way; at the local level (for US cities). Figure 4 a - Urbanization and Energy Consumption Trend in Norway Figure 4 b - Urbanization and Energy Consumption Trend in Finland Figure 5 a - Low Energy Urbanization in Turkey Figure 5 b - Low Energy Urbanization in Pakistan
Local Urbanization The urbanization at the local level is examined for five US cities using census data. The urbanization development of three cities with a population of about 1/2 million in 1990 was traced, as well as that of two multi-million person cities. As assumed earlier, despite an increase of urban population at the national level, some cities decline (Figure 6). Some of the cities that had major growth before the advent of individual traffic and others grew in the automobile area. Density seems to be one factor that impacts urban form. In comparing the growth rate - density correlation of 5 major cities we find a strong correlation of rapid growth and high spread. However there are many more factors and additional research is needed. Figure 6a - Urban population development Figure 6b - Urban population development
Figure 7 - Population Density Development The generalized data however may be misleading. Therefore a comparison of spatial density distribution was conducted for two of the five cities investigated.
City Shape: Boston, MA versus El Paso, TX The theory, just to recap, is that the time when the urbanization transition occurs will have a significant impact on the urban shape that is formed. It is in fact the context or conditions under which urbanization occurs that will impact the form. Technology is viewed as one of the important shape factors. With Boston, MA and El Paso, TX we have two cities that have approximately the same size by population; with Boston having a population of 551675 in 1992 and El Paso having 543813. Both cities are constrained on how freely they can grow in some sense, with Boston having the Atlantic Ocean restricting growth to the south and east and El Paso having the Mexican border. In these terms the two cities are comparable. However, Bostons growth rate over the past 40 years was moderate or even declining, while El Paso went through a major post World War II growth spur. In 1940 El Paso had a population of about 100,000 (D'Antonio & Form, 1965). Boston's population exceeded 100,000 between 1840 and 1850 (Kennedy, 1992)1. The existing urban shape of the two cities is very different. The diagrams below map the shape components of population density and spatial distribution for Boston and El Paso. They were created using 1990 US census data for the city proper boundaries. At the block group level, population data, area, longitude and latitude for the centroids of each block were collected by extracting the appropriate files in electronic form and saved as text files. At this level 100% population count data is available. Block groups have an absolute population count between 700 and 2000 persons. The granularity of this data level ensures enough data points to get fairly accurate results for a density mapping. 700 + data points (blocks) were identified for Boston and 600 + for El Paso. However the variation of population within each block was felt to be too large to merely map the population. Thus, the delimited text files were read into a spreadsheet program (EXCEL) and reformatted for import in a statistics package (SYSTAT). Within the statistics package a number of computations were performed to calculate the population density per square mile within each block. A simple no-frills 2 dimensional plot was chosen to display the data spatially. The latitude variable was chosen for the y-Axis and the longitude variable was plotted on the x-Axis. The level of density is distinguished by color (Figures 8a,b) and size (Figures 9a,b) of the symbol plotted on the centroid of the each block within the citys jurisdiction. For representation purposes it was necessary to multiply the population density per square mile with a factor of 1/1000 to ensure the variable was in a suitable range. The data for both cities was forced onto the same scale in order to make the graphs more easily comparable. Although the plot where density symbols are sized by value evokes a more shape like image this might be deceptive. This is not the real city shape. The density is so high that the symbols overlap or touch each other creating a black seemingly solid city. The colored plot may be more honest however it is somewhat hard to translate it into a meaningful interpretation. The Boston shape plot below shows a very dense, and compact city shape with a high population density. Most dense areas are located in the older parts of town. Some less dense areas are found in southern parts of the town where the city extends into more rolling land. In contrast, El Pasos population density on average is 1/5 or less. The urbanized area is spread out along the major arterials. There are some areas of higher density, curiously enough one at the eastern fringe of the city. Figure 8a - City of Boston Shape in 1990 (Density by Color) Figure 8b - City of Boston Shape in 1990 (Density by size of symbol)
Figure 9a - City of El Paso Shape in 1990 (Density by color) Figure 9b - City of El Paso Shape in 1990 (Density by size of symbol)
Conclusions This paper is but one little puzzle piece in gaining insight into the urban-population-environment dynamics and especially the linkage of urban development and shape. Despite the rudimentary results, some thoughts on policy implications are presented. Many more issues became obvious for future research.
Discussion of Results While globally there is a definitive trend towards urbanization, the urbanization transition is already concluded in some of the countries investigated, such as Japan, Denmark, the UK. Other countries are at the beginning or in the middle of the urbanization transition, such as Pakistan, Indonesia, and many more. The further course of the urbanization transition in these countries in terms of amplitude and speed will depend on the course of their demographic transition. The sooner overall population growth can be slowed the sooner the urbanization transition will level. It is likely that the urbanization transition will lag behind the demographic transition for a few decades while rural to urban migration persists. A general time frame for this lag could be derived from looking at industrialized nations which have gone through both transitions successfully (Japan, Denmark, Norway). At least three countries could be identified that go through a somewhat distorted transition, which are Australia, Canada and the US. These are major immigration countries. Urban population increases in a linear fashion, while the rural population remained constant for the past 45 years. Besides rural to urban migration of younger members of the population, most immigrants seem thus to move to cities. Ports of entry and coastal cities are likely to grow more than cities in the interior of a country, since these are the places where immigrants naturally arrive and mostly settle. Depending on the specific characteristics of a country it can be assumed that between 80% and 90% of a country's total population will live in cities and urbanized areas. Some places with stringent land constraints will reach virtually 100% urbanization such as Singapore and Hongkong. Country specific characteristics influencing how high exactly the percentage of future urban population will be could possibly be derived from a conglomerate of factors such as landscape, level of technology, historical patterns of settlement, culture, available land area, total population density, policy and others. The author can only speculate at this point about the importance of each of those factors, but future research might be able to show patterns of development and correlation, particularly between land area, density and the total urbanization quota. In comparing the urbanization of industrialized nations with those of lesser developed countries it becomes quickly obvious that the urbanization happens under different conditions. The study looked at just one factor, which was energy consumption. In Turkey for example, the overall urbanization is at a higher level than in Finland - however with only a 1/10 of the energy consumption/per person. Urbanization in Turkey is based not necessarily on industrialization and improved technology (Turkey's relative low I_index also points to that direction). Urban shapes and urban living thus will look different in those countries. As Sinclair (1967) pointed out J. von Thunens land use model does not apply any more for the US or other industrialized nations but may be still valid for lesser developed countries. With urban expansion due to rapid urbanization it will be interesting to see, if new and what values will govern the development. Comparative investigations of city shapes and form in countries with high respectively low energy consumption could be an interesting future research topic. In terms of all transitions the issue of scale is important. While in general, the urban population in the US is increasing, locally decline is also observable as can be seen for the development of Chicago and Boston. Since both cities are nuclei of larger metropolitan areas the decline of center areas also could be viewed as a decentralization and dissemination trend. While studies in the past were successful in providing empirical data on the "flattening density function" and post World War II suburbanisation (Mills, 1969; Newling 1966), the causes for the trend are not agreed upon. The local level research conducted for US cities and the shape investigation for Boston and El Paso in particular suggest that it matters when and under what conditions urbanization and urban growth occurs. The much older city of Boston experienced major growth and land expansion in the past century. Due to the relative limited and slow modes of transportation at the time, Boston's density is high, even today. In comparison, El Paso's major growth spur occurred post World War II where a change in technology brings about cost efficient personal transportation. At a comparable population size in 1990 (ca. 500,000) urban densities of El Paso are about 1/5 of Boston's densities. Figures 8 and 9 show the differences clearly. Doxiadis organic urban growth model (1968) anticipates a distortion of urban growth away from a compact form along radially outbranching transportation arterials, however not to the extent observed for El Paso. The shape development analysis conducted for this study has several backdrops. While for El Paso the city limits and the urbanized area are fairly congruent, it is not for Boston. The area of Boston proper did not expand since 1912 (Kennedy, 1990). So, in order to examine the shape of urban growth development it probably would be better to look at the shape of the metropolitan urbanized area rather than city limits. Data availability is a problem, since census tracts and jurisdictional boundaries do not necessarily reflect urbanized land and its boundaries. Remote sensing and satellite imagery might overcome this problem, however since this data has been only recently released to the public time series over decades are a problem. Policy Recommendations There are two types of policy recommendations: Country and Local Level. The suggested policies are very general at this point. They aim at improving the quality of urban agglomerations. This is defined very roughly as lowering the negative impact on the environment and decreasing the energy/resource exchange flows between the urban ecosystem and the surrounding ecosystems, mainly by looking at shape and density of the city. The recommendations are based on the broad (and maybe at times invalid) assumption that a denser and more compact city is likely to be more sustainable (Dantzig & Saaty, 1973)
Country Level Recommendations It seems sensible to categorize the policy recommendations according to the stage of urbanization a country or nation is at (Pattern A, B, C)
Pattern A (Concluding Urbanization Transition) => Stabilize status quo (you do not want to have cities disintegrate after booms, marching right into the next transition) Improve efficiency of urban city (lower emission, water use, public transit) Mitigate environmental effects in surrounding areas prepare for downsizing of the city or industry consolidate metropolitan areas and provide for coordinated strategies (jurisdictional problems)
Pattern B (Starting Urbanization Transition) => slow urbanization down, or at least try to influence development Improve opportunities in the country side make vision plans (although there seems to be no time) provide public transit promote density and neighborhood clustering, decentralized avoid depletion of environment
Pattern C (Immigrant Countries) => try to influence urban development provide public transit promote density and neighborhood clustering, decentralized avoid depletion of environment Local Level Recommendations Local recommendations must be much more specific to the individual situation of the city.
Future Research Much more work in terms of developing urban shape morphologies has to be done. The author envisions that harmonious correlation can exist between urban form and environmental features. These harmonious correlation will be perceived as beautiful and particularly pleasant. It also could be characterized as being stable and sustainable. Or yet in other words, a preferable city shape is achieved by optimizing the urban metabolism2 for the surrounding landscape. During rapid changes of either the environment or population (characteristics) the harmony of this correlation will be disturbed. The author further expects to find different patterns and correlation at different scales. So the harmonious urban-environment pattern will be different for a population of 10,000 or 100,000. A harmonious relationship between urban agglomeration and environment can be described as follows: the urban form complements or assimilates the landscape, i.e. in color, form, character and scale3. the size of the city and its energy and resource demands do not cause drastic changes in the landscape. Beauty, harmony and pleasantness are difficult values to measure. It is may be possible to find indicators such as crime rate, rating of quality of live, migration rate, traffic congestion to evaluate the level of compatibility of the urban system at its current stage with its surrounding environment. For the environment measures of biodiversity, decline or increase of renewable resources, fertility of the land, need to fertilization in agriculture etc. Since both factors are interrelated a depletion of the natural resources in the vicinity of an urban agglomeration may as well further decrease the quality of life in a city. As described earlier, urban form can be conceived as a result of the different factors such as culture, technology etc. (Figure 2). Cities in the mountainous countries like Austria or Nepal and Tibet will have a different shape than a city in a desert landscape. Growth pattern and urbanization pattern will be different as well. It is much easier to build a city in a valley on flat ground than on the slope of the hills. But flooding of rivers in this valley might be a threat and the valley ground must be kept free from development for agricultural usage. Diverse and rugged topography might favor many smaller town over one or two very large ones. In contrast, the desert landscape might favor larger agglomerations along a river or at some other water source. It will be easier to distribute the water within a large city than build a pipeline to supply many smaller towns dispersed in the landscape. Technology is an important modulator of urban form. However, it tends to exert an increase in energy consumption that might be counterproductive in developing sustainable cities.
NOTES 1. Boston and El Paso are different in age. The settlement of Boston began around 1630. Boston reached its peak in terms of spatial extend of city limits in 1912. The first permanent settlement on the present site of El Paso was in 1827. It was known by its present name by 1859 and formally chartered as town in 1873 (Kennedy, 1992; et. al). 2. Urban metabolism views the city as an organism, an open system, with inputs and outputs from and to the environment (H. Girardet in Cadman & Payne, 1990). 3. The conceptual idea of complementation of man-made world and nature is discussed in Norber-Schulz (1984). The structure of the natural place is complemented by either adaptation of the same structure in the man-made artifacts or contrasted; i.e. diverse landscape with multiple features is "harmoniously" complemented by a small scale architecture with lots of detailing like we find it in Norwegian wood carvings. The large scale desert landscape is complemented with geometrical simple artifacts, straight lines, square features, monumental. etc. REFERENCES Aberley, Doug ed. Futures by Design. Philadelphia, PA. New Society Publishers. 1994. Batty, Michael and Paul Longley. Fractal Cities. Academic Press. 1994. Cadman, David and Geoffrey Payne (eds.). The living City - towards a sustainable future. New York, NY. Routledge, Chapman and Hall Inc. 1990. Cities. Scientific American. New York, NY. Alfred Knopf. 1969. County and City Data Book 1956. A statistical Abstract Supplement. Department of Commerce. Buereau of the Census. United States Printing office D.C. 1957. County and City Data Book 1962. A statistical Abstract Supplement. Department of Commerce. Bureau of the Census. United States Printing office D.C. 1962. County and City Data Book 1972. A statistical Abstract Supplement. Department of Commerce. Bureau of the Census. United States Printing office D.C. 1973. County and City Data Book 1994. A statistical Abstract Supplement. Department of Commerce. Bureau of the Census. United States Printing office D.C. 1995. Detwyler, Thomas R. and Melvin G. Marcus. Urbanization and Environment: The physical Geography of the City. Belmont, CA., Duxbury Press. 1972. Dantzig, G. B and Thomas L. Saaty. Compact City. San Francisco, CA. Freeman and Company Inc.. 1973. Doxiadis, C. A. Ekistics: An introduction to the science of Human Settlements. London, Hutchinson. 1968. Drake, William D. Towards Building a Theory of Population Environment Dynamics: A Family of Transitions. In: Population - Environment Dynamics. Ann Arbor. University of Michigan Press. 1993. Frank, Andrea I. In Search for Pattern: An analysis of socio-economic Data: Genesee County, Michigan. Unpublished term project. Engineering 503, University of Michigan, Ann Arbor, MI. Fall 1993. Frankhauser & Sadler. Fraktales Stadtwachstum. in ARCH+ 109/110, Dec 1991. Hartshorn, Truman A. Interpreting the City: An Urban Geography. New York, NY. John Wiley. 1980, 1992 2nd ed. Johnston, R.J.. Spatial Structures: Introducing the study of spatial systems in human geography. Methuen & Co. ltd. 1973. Kennedy, Lawrence W. Planning the City upon a Hill - Boston since 1630. Amherst, MA. The University of Massachusetts Press.1992. Mills, Edwin S. "Urban Density Functions". Urban Studies. Vol 7. pp 5-20. 1969. Newling, Bruce. "Urban Growth and Spatial Structure: mathematical models and empirical evidence. Geographical Review. Vol 56. pp 213-225. 1966. Norberg-Schulz, Christian. Genius Loci: Towards a Phenomenology of Architecture. New York, NY. Rizzoli International Publications, Inc. 1984. Sinclair, Robert. "von Th|nen and Urban Sprawl." Association of American Geographers Annals. Vol. 57. pp 72 -87. 1967. State and Metropolitan Area Data Book 1986. v. D'Antonio W. and William H. Form. Influential in two Border Cities: A study in Community Decision - Making. University of Notre Dame Press. 1965.