Chapter 4. Causes and consequences of urban sprawl
This chapter provides an extensive review of the main drivers of urban sprawl, as well as of its impacts on the environment, economy and society. The chapter discusses economic, geographic, technological and policy drivers of urban sprawl, including household preferences, developers’ expectations, and stringent land-use policies. It also investigates a set of potential environmental effects of urban sprawl comprising: i) increases in emissions from more extensive car use; ii) changes in energy needs and emissions associated with residential heating; iii) losses of periurban arable land and their further effect on food prices; iv) losses of biodiversity and environmental amenities; v) degradation of water resources; and vi) changes in microclimate. The potential effects of urban sprawl on land and housing prices, public finance, and on human health, are also reviewed.
4.1. Introduction
This chapter provides an extensive literature review focusing on the drivers and the effects of urban sprawl, as the latter was defined in Chapter 2. 4.2 draws from a voluminous body of literature that spans several scientific disciplines (economics, biology, hydrology, environmental and agricultural sciences), to summarise the various drivers of the phenomenon. It pays special attention to the preference-driven causes of urban sprawl, as well as to other forces (such as speculation and expectations) that contribute to a growth pattern that may appear as sprawling in the short run, but constitutes a temporary stage in the long-run development of a compact city. In addition, a series of policy interventions that may give rise to more sprawled urban environments are analysed and illustrated graphically. The chapter also examines the role of path dependence, historical drivers, the technological progress in car manufacturing and other factors that reduced the private costs of automobile use and commuting.
Similarly, 4.3 reviews the multiple impacts of urban sprawl on the environment, economy and society. The review examines the various sprawl dimensions presented in Chapter 2 in conjunction with a series of environmental variables: greenhouse gas emissions, local air quality, the relative scarcity of periurban arable land, a series of hydrological functions, biodiversity, water quality, urban heat islands and the resilience to climate change. It also examines the links between these sprawl dimensions and the social cost of providing a series of local public services such as public transport and fresh water. The discussion of consequences abstains from an emotionally-charged indictment of sprawl; instead, it investigates the internal and external validity of numerous studies, some of which provide contradictory evidence. Accounting for the various methodological limitations, the chapter highlights that, while many of the concerns raised are sensible, many others lack substantial evidence-based support. Finally, the chapter identifies concerns that may refer to effects erroneously attributed to urban sprawl; instead, these effects stem from other forms of urban development, or urbanisation per se. That distinction may be important for the development of interventions that target urban sprawl, rather than urbanisation (whose economic importance has been examined in Chapter 1). A series of such interventions will be provided in Chapter 5.
4.2. Urban sprawl drivers
Several studies have explored the nature of urban sprawl and the potential reasons for its occurrence in different contexts (Anas, Arnott and Small, 1998; Brueckner, 2000; Brueckner, 2001; Burchfield et al., 2006; Deng et al., 2008; Glaeser and Kahn, 2004; Nechyba and Walsh, 2004; Oueslati, Alvanides and Garrod, 2015; Patacchini and Zenou, 2009). Different factors are often identified as the main drivers of urban sprawl. These factors include population growth, the rise in household incomes, individual preferences favouring low-density development, complex historical processes that gradually enhanced such preferences, physical constraints on development continuity and structural density, the significant technological progress in car manufacturing and several other factors that contributed to a general decline of commuting costs. Box 4.1 discusses the determinants of urban sprawl in the context of monocentric cities. Policy-related factors include persistently low prices on car use, massive and unbalanced investment in highway capacity and a series of land-use regulations, such as maximum building height restrictions, that are known to keep population density in levels lower than those that would prevail otherwise. The above drivers are investigated in detail in the following subsections.
Box 4.1. The determinants of urban sprawl in monocentric cities: some empirical evidence
When it comes to urban expansion, the monocentric city model (Alonso, 1964; Mills, 1981; Muth, 1961; Wheaton, 1974) is the cornerstone of urban economics. The model provides a set of testable hypotheses which can be evaluated using empirical methods. First, an increase in the urban population increases demand for housing and is expected to increase the urban footprint. Second, households demand more floor space as income grows, as housing is a normal good. In turn, the increased demand for floor space leads to an extended city with a lower population density. Third, a decrease in generalised commuting costs increases the disposable income at all locations. This boosts demand for floor space and willingness to reside far from one’s workplace, and eventually leads to a city with lower population density. Changes in commuting costs can occur through changes in the pecuniary costs of transport (fuel prices, public transport fares, car ownership costs, parking fees, road taxes, etc.) or through adjustments in the required travel time (wider roads, faster transport modes, etc.). Fourth, a reduction in the opportunity cost of land, i.e. the financial return from land uses other than development, such as farming, causes cities to expand and lowers the overall population density.
Several studies have tested the empirical validity of the monocentric city model. In the context of the United States, income, population and agricultural rent have been shown to be significant determinants of the total land uptake of an urban area (Brueckner and Fansler, 1983; McGrath, 2005; Song and Zenou, 2006). On the other hand, commuting costs had an ambiguous effect, which varied with the proxies used to measure them. Income growth has also been shown to have a major influence in the People’s Republic of China’s urban expansion (Deng et al., 2008; Shanzi, Song and Ming, 2009). In the European context, Oueslati, Alvanides and Garrod (2015) consider two indices of urban sprawl that reflect changes in artificial area and urban fragmentation. They find that the fundamental conclusions of the monocentric model are valid for both indices. The monocentric city model was found to explain satisfactorily the total footprint of a city, but had limited power to predict the extent to which urban fabric will be fragmented. Urban planning policies, land availability and physical geography constraints are particularly strong predictors of fragmentation, along with other factors that discourage infill development between fragments.
Source: Alonso, 1964; Brueckner and Fansler, 1983; Deng et al., 2008; McGrath, 2005; Mills, 1981; Muth, 1961; Oueslati, Alvanides and Garrod, 2015; Shanzi, Song and Ming, 2009; Song and Zenou, 2006; Wheaton, 1974.
Preferences and pull effects
Many households are willing to pay for attributes that often characterise low-density areas. For example, proximity to open spaces and natural amenities, lower noise levels, better air quality, longer exposure to sunlight and local visibility are acting as attractors to these households, which boost the demand for housing in relatively remote areas. Gordon and Richardson (1997) claim that the preference for low density is the main driver of urban sprawl. They invoke the National Housing Surveys performed by the Federal Home Mortgage Association (1994) to argue that the observed low density in the US is preference-driven. Turner (2005) investigates the choice of residential location taking into account the willingness to pay for proximity to open spaces. He shows that, if open space is highly valued, development in remote areas with access to plenty of open space may take place before development of areas that are closer to the city centre. Therefore, part of the observed vacant land and fragmentation within urban areas may be a result of individual preferences.
The pull effect exerted by suburban natural amenities on household location decisions encourages the development of areas formerly occupied by farmers (Cavailhès et al., 2004; Coisnon, Oueslati and Salanié, 2014; Ready and Abdalla, 2005; Roe, Irwin and Morrow-Jones, 2004). Using stated preference data from the United States, Roe, Irwin and Morrow-Jones (2004) find that rural-urban fringe areas that are located within a certain distance from urban centres and are abundant in farmland could attract residential development. Therefore, efforts to preserve small parcels of farmland may induce further residential growth in areas of high accessibility. Land conservation policies aiming at preserving landscape and natural amenities have been shown to indirectly encourage the development of surrounding parcels of land (Geniaux, Ay and Napoleone, 2011; Irwin and Bockstael, 2004). Therefore, pull effects tend to increase the percentage of urban footprint hosting low-density development and the share of urban population residing in low-density areas.
Historical drivers and path dependence
Are there any historical drivers behind the formation of preferences for living in low-density areas? If housing type preferences in OECD countries have remained stable during the twentieth century, the vast suburbanisation observed during its second half has been driven by economic progress, rising incomes and s fall in the real costs of car use. If, on the other hand, preferences have changed substantially towards living in low-density areas, urban sprawl would emerge, at least to some extent, even in absence of the economic progress and population growth that followed the Second World War. Due to a large number of methodological constraints, empirical analysis has not managed to answer that important question thus far. However, historical arguments can still offer some useful insights.
Path dependence is a key driver of urban sprawl. Urban systems evolve through path dependent processes, thus history plays an important role: past policy choices and major events do not only determine the current state of a city but also the set of responses that are currently available to policy makers. For example, stringent building height restrictions implemented decades ago in many cities have led them to an urban form with very low density levels. Such land-use policies usually have effects persisting for many decades, as changing development patterns in existing urban environments is particularly challenging. If path dependence is strong, a sprawled urban area has, to a large extent, evolved in that way because the initial conditions or policy framework (e.g. stringent building height restrictions) were favouring sprawled development patterns. Through time, policy-making that failed to account for the long-run social cost of urban sprawl may have reinforced those conditions.
Expectations and speculation
Expectations for the evolution of land and housing prices may play a crucial role in the pattern of urban fabric expansion over time. Development of seemingly profitable vacant land parcels may be postponed when three contributing factors co-exist. First, the development types that maximise the present value of profits in the short and long term are different, due to e.g. expectations for population or economic growth. Second, the long-run development type is not profitable in the short run, i.e. revenues fall short of costs. Finally, conversion costs are high enough to rule out a development plan in which the short-run profit-maximising type is developed and maintained, with a view to being demolished and replaced by the other type in the longer run.
The three conditions are illustrated with an example in Figure 4.1. For example, the two development types can be considered to be a multi-story building (solid lines) and a construction of low structural density (dashed lines), such as attached family housing. The curves show that the current demand for floor space in multi-family housing types at the beginning (year 0) is too weak to generate substantial profits from development in the short run. However, developers are expecting this to change in the long run (e.g. fifteen years later); population growth and improvements in accessibility (for instance due to anticipated infrastructure investments) often underlie such expectations. On the other hand, developing single-family housing can yield immediate profits. However, these profits are substantially smaller than those yielded by the multi-family development type when considered in a longer horizon. If the short-run profits from development of single-family housing, represented by the shaded areas in Figure 4.1, fall short of the conversion costs (i.e. from single- to multi-family housing), land will stay vacant in the short run.
Note: The left panel displays the undiscounted streams of profit net of capital costs, but not land rents, for the investment in two different buildings: a constant low return option (building type A, dashed line) and an option with a high future return (building type B, solid line). The right panel shows the present value of each stream in a time horizon of hundred years using an annual discount rate of 5%. The total present value of the latter option, i.e. the area under the solid curve, outweighs the total present value of the former option, i.e. the area under the dashed curve. Therefore, in the long run (years 15-100), building B will occupy the land parcel. If the conversion costs exceed the short-run total present value of developing building type A, i.e. the area below the dashed curve for the first 15 years, the land parcel will remain undeveloped in the short run (years 0-15).
Meanwhile, the more accurately investors anticipate future profitability, the more they drive current demand up, raising the prices of those vacant land parcels even in the short run. This narrative explains a phenomenon observed in many modern urban areas: even in their interior, a substantial portion of developable land remains vacant while at the same time its price is high. A relevant stream of literature in urban economics provides various reasons why it may be efficient to postpone development of certain parcels in the interior of an urban area, so that they can be developed in a way that better suits future needs (Braid, 1988, 1991; Brueckner and von Rabenau, 1981; Clawson, 1962; Fujita, 1976; Mills, 1981; Ohls and Pines, 1975; Turnbull, 1988). This suggests that a fragmented development pattern may, to some extent, constitute a natural step in an economically efficient long-run process of urban land transformation.
Technological progress in car manufacturing
Part of the urban expansion can be explained by technological advances in car manufacturing, and the production of cheaper, faster, safer, more reliable, more comfortable and more fuel efficient cars (see also Glaeser and Kahn, 2004). Technological progress and economic growth have made cars affordable by a large share of the population in OECD countries. In turn, lower costs of car ownership and use translated into willingness to accept larger commuting distances, i.e. willingness to live further away from work. Driving distances that seemed restrictive gradually became reasonable. The outcome was that previously remote land plots became accessible to the majority of households. Simultaneously, in many places of the world the institutional framework did not account for these changes (see also the discussion below about the policy-induced drivers of sprawl). That resulted in relatively affordable periurban land prices and massive supply of accessible land readily available for residential consumption, a combination which promoted low density development. In summary, technological progress in car manufacturing contributes to lower average population densities and an increase in the share of urban footprint hosting very low density levels.
Policies encouraging car use
Car use has not only been promoted by technological developments in car manufacturing: it has also been encouraged by policies subsidising car use or failing to price its external costs. The absence of road pricing, the neglect of the external costs of on‐street parking, and persistently low taxes on gasoline and diesel have kept the pecuniary cost of vehicle use artificially low in the majority of countries. At the same time, enormous public and private investments in road – mainly highway – capacity expansion have resulted in significant reductions in travel time, as well as increased travel comfort and fuel efficiency. The consequences of improved road transport infrastructure on urban sprawl have been validated by empirical evidence (Garcia-López, Holl and Viladecans-Marsal, 2015; Su and DeSalvo, 2008). Chapter 5 discusses how policies on road transport can be transformed to control urban sprawl and alleviate its consequences.
Land-use policy drivers
Understanding the functions of land-use regulatory mechanisms is key to understanding the causes of urban sprawl. This subsection investigates the effects of three types of policy interventions that affect directly the sprawl dimensions described in Chapter 2. Building height regulations, urban containment policies and taxation of property and land are long known to influence long-run development patterns. The subsection also explains why these policy mechanisms are rigid and persistent over time.
Building height restrictions
Maximum building height restrictions contribute to the development of cities with a larger urban footprint compared to the one that would occur if they evolved without them (Bertaud and Brueckner, 2005). In general, a uniform-in-space building height restriction (e.g. 30 metres of height across the city) translates into lower average population density, although it may also (indirectly) influence the two density allocation indicators developed in the report: the percentage of urban land hosting low density levels (land-to-density allocation function) and the share of population residing in low density areas (population-to-density allocation function). Figure 4.2 illustrates the hypothetical evolution of an urban area under two different scenarios. The left panel displays the floor-to-area ratio in a city that has evolved with (dashed curves) or without (solid curves) a maximum building height regulation that is uniform across space. The right panel presents population density under the two scenarios. To keep things simple, it is assumed that the hypothetical restriction is uniform across space.
Note: Solid lines represent the urban area that has evolved without any maximum restriction on building height; dashed lines represent the urban area that has evolved with a maximum floor-to-area ratio of fifteen floors (horizontal dashed line in left panel). The vertical lines designate the boundaries of the area directly affected by the restriction.
A stringent building height restriction, such as the one depicted in Figure 4.2, will result in an urban area with a larger footprint and a different allocation of floor space and population. The environmental and economic effects of this are analysed in 4.3.
Urban containment policies and leapfrogging
Urban containment policies, such as urban growth boundaries and greenbelts, have been proposed as a remedy to the uncontrolled expansion of urban footprint. Whether such boundaries improve environmental conditions and the extent to which they affect housing prices are questions that spawned important policy debates that are summarised in Chapter 5. Instead, what is emphasised here is that part of the observed fragmentation in an urban area and a tendency towards low-density development may be attributed to the use of urban growth boundaries and other zoning policies that prevent development in designated areas. Leapfrogging is a common response to such regulations. It is characterised by scattered, low-density development beyond the buffer zones designed to restrain the footprint of an urban area. The time needed for leapfrog development to occur depends on various factors, such as the width of the buffer zone, the pace of technical progress and the growth rates of population and income.
Figure 4.3 illustrates an example of the short- and long-run effect of a ten kilometre buffer zone that is initially designed to restrain the urban area to its current footprint. The upper left panel of the picture shows that, in the short run, population growth is accommodated exclusively by densification of the existing urban area (infill development). The extent to which this policy backfires through higher housing prices depends on the ability of developers to substitute capital for land by raising higher buildings. The lower left panel of Figure 4.3 shows the developers’ willingness to pay at the moment of the urban growth boundary introduction (solid curve) and at the end of the short run (dashed curve). The increased demand for floor space causes upward pressures on land prices within the existing urban area, but these pressures may not be strong enough to generate leapfrog development: in the lower left panel the willingness to pay to use the land for development falls short of the willingness to pay for alternative land uses (e.g. agriculture) in remote areas.
Note: Solid curves represent the current state of an urban area; long-dashed curves represent the state of the urban area at the end of the short-run; short-dashed curves represent the state of the urban area in the long-run. The shadowed area represents a 10 km buffer zone that in the short-run acts as an urban growth boundary designed to restrain the urban area to its current footprint, i.e. a 20 km diameter circle (locations 20-40). Upper panels display population densities. Lower panels display land rents. The horizontal solid line represents the opportunity cost of land (i.e. agricultural rent).
However, in the long-run income grows substantially and transport costs fall. Then, the return from developing remote areas exceeds the return from alternative land uses and leapfrog development emerges. This way, an initially effective anti-sprawl policy converts into a policy that generates urban sprawl, as the latter was defined in Chapter 2.
Property and land tax
Property taxes, i.e. recurrent ad valorem taxes on immovable property, are the most widely used price instrument applied to land use. In most cities, property taxes are levied at equal rates on land and capital improvements, i.e. buildings and other structures constructed on it (Bird and Slack, 2004; Brueckner and Kim, 2003). The long-run effect of property taxation on city size has been a debated issue in the relevant literature (Arnott and MacKinnon, 1977; Brueckner, 2001; Brueckner and Kim, 2003; Carlton, 1981; LeRoy, 1976; Song and Zenou, 2006; Sullivan, 1984, 1985).
Property taxes affect urban sprawl through at least two channels. First, a property tax increases the cost of property ownership. In the long run, this provides incentives for the construction of smaller dwellings, as households will respond to the increased housing costs by substituting away from the consumption of residential floor space. In turn, smaller residences imply that the same population can, ceteris paribus, be accommodated in a city of smaller footprint. This affects negatively both the average population density and the two density allocation indicators (land-to-density allocation, population-to-density allocation) in a city. Second, the property tax will exert a downward pressure on land prices, but not on the capital costs of construction and maintenance, which are not determined within city boundaries. This may cause the development of lower buildings, a force that affects the aforementioned indicators in the opposite way. Therefore, whether property taxation impedes or fuels the spatial expansion of cities depends on which of the two effects is stronger (Brueckner and Kim, 2003).
The two possible outcomes described above are depicted in Figure 4.4. The left panel represents the case in which the first effect described in the previous paragraph dominates, i.e. the property tax eventually leads to a smaller city size. In contrast, the right panel illustrates the case where the second effect prevails, shorter buildings are constructed, and a more dispersed urban form emerges.
Note: Horizontal arrows represent the size of the urban footprint under different outcomes. Base: absence of property tax; A: property tax imposed, residence-size effect prevails; B: property tax imposed, tax-on-improvements effect prevails.
Land-use policies have long-term effects
The policies mentioned in the previous subsections are, for several reasons, rigid and persistent over time. First, it might be difficult to adjust land-use regulations since current homeowners have strong incentives to lobby for building height restrictions, as well as zoning (Fischel, 2001; Schuetz, 2009). These regulations have been shown to increase housing prices in certain contexts and are, therefore, likely to remain intact over time (see e.g. Ihlanfeldt, 2007). Another reason for which building height regulations, zoning and urban growth boundaries are persistent is that adjusting them may be impossible in an already formed urban landscape and network. For example, it is impossible to introduce an urban growth boundary in the interior of a city or a maximum building height restriction below the one observed in a specific location. The former policy reform would outlaw existing private property located outside the boundary; the latter would impose a heavy financial cost associated with demolition and reconstruction.
Similarly, immovable property taxation has been established in many OECD countries as a standard way to raise public revenue. Property taxes might in theory be easier to adjust, but in practice this is much more complicated, especially if the adjustment is justified on criteria not related to revenue-raising. Chapter 5 will highlight the importance of adjusting fiscal instruments to account for environmental objectives.
Physical geography
Unsuitable physical terrain (e.g. mountains, rivers, coastlines) may prohibit contiguous development. Furthermore, geological factors posing disaster risks, such as proximity to volcanoes or exposure to earthquakes, may limit population density by imposing safety ceilings in floor-to-area ratios, lot sizes, etc. For instance, Saiz (2010) shows that physical geography is a key factor in the contemporary urban development of the United States. Land constraints, such as steep-sloped terrain and water bodies (lakes, wetlands, oceans), decrease the housing supply elasticity. This effect of geography is both direct, i.e. via reductions in the amount of land available, and indirect, i.e. via increased land values and higher incentives for land-use regulations.
4.3. Effects of urban sprawl
The primary objective of this section is to provide an objective discussion of the potential environmental and economic effects of the various sprawl characteristics (discussed in Chapter 2) while abstaining from the adoption of a prejudiced view of urban sprawl. Such an analysis should avoid confusing the environmental consequences of other forms of urbanisation with those of urban sprawl. It should also avoid adopting anti-sprawl arguments without an extensive evaluation of the methodological limitations that could condition the scientific contributions that support them.
Spatial dispersion of activity, car dependency and emissions
The association of urban sprawl with car use has been established with socioeconomic modelling and validated empirically in a series of econometric studies. This contrasts to the relationships of urban sprawl with land scarcity, pressures on water resources and biodiversity, which have not been supported equally well by theoretical or empirical evidence. These relationships are discussed in detail later in this section.
The environmental relevance of the link between urban sprawl and private vehicle use is crucial, as vehicle kilometres translate into air pollution and greenhouse gas emissions. Theoretical work highlights two main channels through which urban development may translate to more vehicle kilometres and thus emissions: i) through changes in the spatial dispersion of residences, jobs and other key locations (e.g. malls, schools); and ii) through changes in the modal split, i.e. the share of each transport mode in the total number of kilometres travelled. The former channel determines the distances to be covered in daily activities such as commuting and shopping, whereas the latter determines whether these distances will be covered by public transport or private vehicles. The two channels are interrelated, but the analysis that follows investigates them separately for expositional convenience.
Urban economic theory (Alonso, 1964; Muth, 1969) provides the basic intuition on the relationship between city size, population density and vehicle kilometres travelled. The main theoretical insights are illustrated in the left panel of Figure 4.5, which displays two cities of equal population. This population is, however, distributed very differently around a central business district (CBD), where all jobs are located. With job location fixed, the relatively less compact city will generate more vehicle kilometres, and therefore more emissions, as commuting distances grow.
Note: Solid curves display the population density around the central business districts of a monocentric (left panel) and a bicentric (right panel) urban area. Dashed curves represent job concentration around business districts. The threshold population densities determine the limits of public transport coverage in the two cases. Shaded areas provide a rough representation of the car-dependent parts of population.
The above primary effect may be reinforced or weakened when some of the determinants of modal split are taken into account. First, public transport in low-density areas will be provided at lower service frequency rates. Typically this occurs to curb the large deficits that result from the low occupancy of public transport modes in suburban areas. In turn, lower frequency of service induces more people to switch from public transport to private vehicles. Thus, less compact cities may generate more vehicle kilometres not only because distances between points of interest are larger, but also because the dependence on private modes of transport is higher. This is illustrated in Figure 4.5, in which the most car‐dependent fractions of the population are represented by the shaded areas of the residential density distributions. These groups reside in areas where population density is below a threshold level (horizontal line) that ensures frequent provision of service without large deficits.
Second, the dispersion of key points of daily activity, such as jobs, schools, malls and public services, is an important determinant of the effects of urban sprawl on private vehicle kilometres. Figure 4.5 illustrates a case in which the vehicle kilometres generated in a contiguous urban area with two distinct employment hubs (right panel) may not differ substantially from those generated in a city with a unique central business district. The figure reveals that the average land uptake per person is not necessarily a good predictor of the vehicle kilometres travelled in an urban area. The average land uptake is roughly double in the city displayed on the right panel. However, the local distribution of population around the two employment hubs in the polycentric case mimics the population distribution of a monocentric city. Furthermore, the distribution is such that only a small minority of the most car-dependent fractions of the population reside in distances that induce long commutes. As is the case with all urban sprawl dimensions, polycentricity indices are insufficient predictors of private vehicle use.
The relationship between urban sprawl and emissions has been established in empirical studies that use highly informative data sets and state-of-the-art statistical methods. Such studies (see below and Table 4.1) shed light on the observed correlation between urban form patterns and the use of private vehicles, from which emissions can be estimated with the use of conversion coefficients.
|
Contribution |
Type |
Main findings |
|---|---|---|
|
Cervero (2002) |
Case study (Montgomery county, US) |
The probability of using public transport choice increases with residential density and mixed land use at the origin of the commuting trip. |
|
Kenworthy and Laube (1996) |
International cross-country comparative study (39 cities in 15 countries) |
Non-linear (negative) relationship between annual vehicle kilometres travelled and average population density. Non-linear (positive) relationship between public transport trips per capita and average population density. |
|
Bento et al. (2003, 2005) |
Cross-city study in the US (114 urban areas) |
The probability of traveling by car increases with population centrality. |
A well-known study on the relationship between urban development characteristics and greenhouse gas emissions has been conducted by Glaeser and Kahn (2010). The authors combine data on gasoline consumption, energy use in public transport, heating expenditure of households, prices of fuel and oil, and carbon intensity of the power generation sector in different areas of the United States. The study finds a strong negative relationship between CO2 emissions and the stringency of land-use controls. The findings also indicate that the metropolitan areas with the lowest per-household emissions are those which are most restrictive towards new development.
Bento et al. (2003, 2005) investigate the effects of city shape, public transport supply, and the spatial distribution of population, jobs and residences on mode choice and distances travelled by motor vehicles in 114 US urban areas. In doing so, they control for demographic characteristics, such as age, gender and household size, weather conditions, city-specific gasoline prices, urban morphology and the characteristics of public transport supply. Their findings can be summarised as follows. First, the probability of walking or cycling to work increases as population becomes more centralised and the allocation of land between jobs and residences more balanced. Second, increasing rail route miles by 1% can increase the share of people using rail to commute by at least 2.9%. The corresponding number for bus route miles is 0.4%. Finally, increasing the amount of land allocated to highways, moving away from mixed land use, and allowing the city shape to be more elliptical rather than radial all significantly increase vehicle miles travelled.
Kahn (2000) shows that households located at the city centre drive 17-43% less than seemingly identical households (in terms of lot size and income) located in the suburbs. Furthermore, he finds that the non-linear relationship between per capita energy use in transport and population density proposed in earlier studies (Newman and Kenworthy, 1989) is robust even after controlling for unobserved city, region and household characteristics. Hankley and Marshall (2010) find that comprehensive compact development could reduce cumulative emissions in the United States by up to 3.2 Gt CO2 equivalent, or 15-20% of the projected cumulative emissions in the period between 2000 and 2020.
Urban sprawl and local public finance
Figures 4.5 and 4.6 illustrate how a lower population density may translate into car dependency and, by extension, to emissions. However, the impacts of low population density also expand to the economic sphere. Public services in low-density areas are often provided with substantial subsidies that could be significantly smaller, or even completely avoided, in compact development settings. Figure 4.6 shows this for the case of public transport provision in monocentric cities. The trade-off between quality of service and economic efficiency becomes clearer with a graphical analysis of two extreme examples of how public transport could be provided to the population group that resides in low-density areas of a highly sprawled urban area. The population distribution in that area (city A) is represented by the solid curve; the size of the affected population group is represented by the shaded area below that curve.
Note: The two curves display the population density around the central business district of a monocentric city with high (solid curve) and low (dashed curve) land uptake per person. The density threshold determines the limits of public transport coverage. Shaded areas provide a rough representation of the car-dependent parts of population.
In the first extreme case, that population group could be served with the same route frequency with which downtown areas are served, ensuring this way that waiting times are roughly equalised across the city. Car dependency is then eliminated (without this implying that car use will necessarily be reduced) with a substantial economic cost, as low-occupancy rates imply a large deficit in the public transport operator’s budget. In the other extreme, that population group is not served at all by public transport. Then, operator deficits are eliminated but the absence of public transport leaves car as the only mode to cover the long-distance trips from the urban fringe to the central business district. This policy option is environmentally detrimental and may also bear a significant distributional impact that renders it politically unpopular.
Actual policies lie between these two extreme cases, with the levels of suburban coverage and public transport subsidies ranging widely across metropolitan areas of OECD countries. Public transport subsidies show a sharp upward trend with time. Taylor et al. (2009) show that the inflation-adjusted increase in local, state and federal public transit subsidies between 1990 and 2003 in the United States was about 51%. At the same time, the per capita ridership remained at historically low levels, with public transport serving only 2.1% of all trips recorded in 2001. Savage (2004) provides a historical analysis of the Chicago Transit Authority’s deficits based on times series data from 1948 to 1997. During that period, the annual operating surplus of USD 48 million recorded in 1948 (adjusted for 1997 prices) gradually turned into an annual operating deficit of USD 543 million in 1997. The study finds that the relocation of homes and jobs – that mainly occurred in the decade following the Second World War – reduced annual revenues by USD 399 million. Together with the fact that average occupancy rates for both bus and rail in 1997 were only half of what they were in 1948, the findings indicate that urban sprawl has been one of the leading drivers of the current deficit.
Similar reasoning carries over to a wide array of public services and facilities whose provision is characterised by economies of density: road cleaning and maintenance services, garbage collection and disposal, sewerage and water provision, police and fire protection, libraries and maintenance of parks and other recreational areas. Low population density does not allow such economies of density and significantly inflates the cost of the provision of public services per user. As density may be hard to increase in the short run, time-persistent budget deficits may occur. Such deficits call for consolidation programmes that raise the price of public transport or reduce the frequency and quality of these services to undesirable levels. For example, the annual welfare loss from the increase of public transport fares in Chicago, Illinois, has been estimated to be around USD 82 million (Savage and Schupp, 1997).
The importance of population density to local economic performance was recognised long before urban sprawl became an important topic in public debate. Real Estate Research Corporation (1974) estimated that providing services to areas where high-rise apartment buildings are the predominant form of development is substantially cheaper than doing so in areas characterised by single-family housing. The report found that the per dwelling capital cost of school provision in areas with high-rise apartment buildings is just 30% of the corresponding costs in areas dominated by single-family housing. The relevant numbers for the provision of roads and utilities were 26% and 17% respectively.
Speir and Stephenson (2002) simulate the capital and energy costs of providing water and sewer services under sixty scenarios of growth of a hypothetical town of 30 000 inhabitants. The scenarios differ with respect to the lot size and fragmentation in the areas that are planned to be developed. Provision costs are found to be much more responsive to the average lot size than to the fragmentation of urban fabric. Larger lots imply that the total length of mains required for serving dwellings within newly developed urban fragments increases. In contrast, fragmentation inflates provision costs only by increasing the required length of mains so that fragments are connected to each other. As the distance between fragments diminishes, the extra cost introduced by fragmentation per se sharply decreases. Similar reasoning holds for other relevant public services such as electric power and gas supply.
Table 4.2 provides a brief, non-exhaustive, overview of the literature investigating the relationship between urban form and costs of public service provision. The econometric study by Hortas-Rico and Solé-Ollé (2010) disentangled the effect of low-density development from other determinants of municipal expenditure in Spain. The study covers six expenditure types (basic infrastructure and transport, community facilities, local police, housing and community development, culture and sports, and general administration) that represent 70% of total local spending. They show that low-density development leads to greater provision costs in all spending categories considered, apart from housing, and basic infrastructure and transport.
Another strand of literature has argued that the cost savings from economies of density are limited: densification decreases the per capita cost of service provision as long as population density lies below a threshold. For instance, Ladd (1992, 1994) analysed the spending profile of 247 large US counties that in 1985 accounted for 59% of the U.S. population. Her findings suggest that a substantial fraction of the analysed counties had a level of density lying above the implied cost-minimizing level. However, her findings should be interpreted with caution due to multiple methodological limitations.
Sprawl and consumption of energy from stationary sources
As it is the case with vehicle-kilometres travelled, consumption of energy from stationary sources can be translated into emissions if the energy efficiency of the technology used for space and water heating in different building types and the emission intensity (greenhouse gases, air pollutants) of the energy sources used for this purpose in a given geographical region are known. Suburban residences are on average larger and therefore consume more energy, and detached housing is ceteris paribus less energy efficient than multi-household buildings. Table 4.3 presents a decomposition of the total energy consumption of the residential and commercial sector by end-use in the United States, the European Union and Canada. The data show that up to 85% of the total energy consumption originates from end-uses that may be indirectly related to urban form.
Temperature in the interior of residential buildings may be sensitive to different local conditions and interactions. Such possibly determining conditions include the degree to which a building is detached from buildings in its close neighbourhood, the entire building morphology at the block and neighbourhood level, and the proximity to trees, plants and roads.
Unlike the links between urban sprawl and travel demand, however, the link between urban form and energy consumption has been poorly explored in the literature. Does lower residential density translate into significantly lower lighting needs? Are energy needs for space heating higher in the suburbs compared to high-density areas? Are buildings in areas characterised by high fragmentation harder to warm up or cool down? Does heating require significantly more energy in areas dominated by detached housing? Does a low floor-to-area ratio imply extra energy needs for space and water heating? Do denser urban forms entail higher energy needs for cooling? Filling these knowledge gaps could have important policy implications, since a significant fraction of the total GHG emissions may be attributed to residential heating and cooling. Unfortunately, the existing knowledge on these issues is limited.
Research on the impact of urban sprawl on residential energy use is at embryonic stage, with the proposed effects reflecting speculations and intuition, rather than well-established evidence. In one of the few empirical studies of this subject, Kahn (2000) fails to confirm the hypothesis that suburban dwellings consume more energy for residential purposes, such as heating and cooling. After controlling for important determinants, such as the household’s annual exposure to very high or low temperatures, its size and income, energy consumption differences between low-density suburbs and high-density urban cores are found to be negligible. However, this finding reflects the fact that important factors such as the insulation technology of the building and local development patterns are unobserved. Therefore, results may just reflect the fact that a detached development pattern in the suburbs is offset by a superior insulation technology characterising contemporary dwellings, which are more likely to be located in the suburbs. Therefore, future studies should attempt to measure important determinants of temperatures in the interior of buildings.
Earlier studies found that consumption of energy from stationary sources is significantly lower in areas of higher density. For instance, the study by Real Estate Research Corporation (1974) suggested that apartments in high-rise buildings consume only 44% of the energy single family-households do. Future research should provide a better understanding of the degree to which these differences can be attributed to local development patterns and the degree to which they simply reflect the fact that single-family housing units are on average much larger than high-rise apartments and are usually occupied by residents with higher income.
Loss of periurban arable land and environmental amenities
Agricultural land has declined, on average, by 4% in OECD countries over the past two decades and this decline is projected to continue (OECD, 2009). The effect of urban expansion on agriculture has always been at the heart of the debate on sustainable land-use patterns. There are at least three major concerns over the continuing urban sprawl and farmland loss. First, the conversion of the most fertile farmland to development reduces agricultural productivity, which decreases food supply in the short run and threatens food security in the long run. Second, urban sprawl reduces amenities and quality of life in rural communities. In many places, urban sprawl has encroached into communities to such an extent that the communities themselves have been lost (Wu, Fisher and Pascual, 2011). Third, farmland loss may have a detrimental effect on agricultural infrastructure. With farmland loss, the local agricultural support sector, such as input suppliers or output processors, may go out of business because of insufficient demand for their output or insufficient supply of input for their production (Lynch and Carpenter, 2003; Wu, Fisher and Pascual, 2011). In addition, land-use conflicts, such as urban residents’ complaints about noise and odour from agricultural production, will likely increase as more farmland is converted to development. Such land-use conflicts may lead to more stringent land-use regulation that restricts traditional farming practices (Lisansky, 1986). Consequently, agricultural economies may shrink in the short run and become unviable in the long run.
The uncontrolled expansion of urban areas has been argued to consume the most fertile parts of arable land. This concern is based on the idea that cities emerge from the evolution of smaller settlements initially hosting populations living from agriculture (Wilson and Chakraborty, 2013). Eigenbrod et al. (2011) model the change in urban land cover in Britain based on a projected population increase of 16% by 2031 under two development scenarios: i) a densification scenario, in which existing suburban housing is converted to dense urban housing, and ii) a sprawl scenario, in which housing demands in the study horizon are met by creating new dwellings in pre-defined undeveloped areas at the existing density observed in suburbs. In the latter scenario, densification of suburbs occurs only when no more space is available. In that context, they investigate the effects upon flood mitigation, agricultural production and carbon sequestration. Their simulation shows that losses of agricultural production are over three times higher in the sprawl scenario compared to the densification scenario.
There are serious counter-arguments, however, to the concerns about the loss of farmland and possible increases in food scarcity due to urban expansion. First, cities occupy only a small fraction of the total land area in a global scale, so urban expansion per se does not seem to pose a real threat for future food supply. At the world scale, for example, Alcamo et al. (2005) projected agricultural land to increase by 10-20% between 2000 and 2050. Fischel (1985) showed that even if the entire US population lived at suburban sprawl densities of 0.25 acres per person, just 3% of the total land area of the forty eight contiguous states would be utilised.
Another counter-argument neglected in some other reports on urban sprawl (e.g. EEA, 2016) is that the methods employed in many simulation studies, including the one by Eigenbrod et al. (2011), fail to account for the effects of urban sprawl on agricultural land prices. Accounting for those effects reveals that urban sprawl is not very likely to cause food scarcity. If food production declined sharply due to urban expansion, food produced in periurban areas could be substituted by imports of food produced elsewhere. If substitution was relatively cheap, cities would expand in an uncurbed way (as the upper panels of Figure 4.7 suggest) but food scarcity would not be an issue. On the other hand, if food produced in periurban areas were costly to substitute, the financial return from cultivating periurban areas would increase sharply (as the lower panels of Figure 4.7 suggest) preventing urban sprawl from being realised in the first place. Thus, uncurbed urban sprawl scenarios are inconsistent with a sharp increase in food prices (Brueckner, 2000), because market price setting mechanisms act as automatic stabilisers that prevent possible effects of urban sprawl on food scarcity. The above arguments suggest that the scenario displayed in the upper panels of Figure 4.7 is probably a better representation of reality.
Note: The figure displays the effect of a major shock in the demand for floor space on the allocation of land between residential development and agriculture when the price elasticity of food supply is high (left panels) and low (left panels). Solid curves: initial willingness to pay for a unit of land surface for development (initial bid rents); dashed curves: final willingness to pay for a unit of land surface for development (final bid rents); Solid lines: initial financial return from cultivating land (initial opportunity cost of land or agricultural rent); Dashed lines: final financial return from cultivating land (final opportunity cost of land or agricultural rent). Shaded areas: land allocated to development; non-shaded areas: land allocated to agriculture. Vertical lines: city boundaries.
The current literature lacks reliable studies on the pressures exerted by urban development on food prices. Future efforts should take into explicit account the fertility rate in periurban areas, the competition for land between the agricultural and the development sector, the import price of food substitutes from the international markets and the trade agreements between countries.
Urban expansion may be socially inefficient not only for reasons related to the negative externalities of residential development, such as congestion and pollution, but also for reasons related to the foregone positive externalities generated from alternative land uses in periurban areas, such as forestry or agriculture. Forests and cultivated areas are hubs of wildlife, perform functions related to air quality and carbon sequestration, prevent surface runoff and soil erosion, and provide amenities or other public environmental attributes. Figure 4.8 displays two different allocations of land between residential development and an alternative land use that generates positive externalities, such as periurban forestry. Under free market conditions, displayed in the left panel, city limits are determined through the equalisation of marginal private benefits of the two land use types. That is, locations in which land rents (solid curve) exceed the private return from forestry (dashed line) are conceded for development and vice versa.
Note: The figure displays land-use allocation under two alternative regimes. Left panel: free-market conditions; right panel: socially optimal conditions; horizontal dashed line: private return of alternative land-use (e.g. forestry); horizontal solid line: marginal social value of alternative land-use; vertical dashed lines: city limits under free-market conditions; vertical solid lines: city limits under socially optimal conditions; solid curve: return from residential development of land. The vertical distance between the two horizontal lines represents the marginal external benefit of the alternative land use. Shaded areas: total social benefit from containing city size to its socially-optimal level.
The above allocation is not socially optimal, because it fails to take into account that every square metre of land developed at the city fringe removes a square meter of forest land that provides useful environmental functions and amenities. Such services provided by forests are valued positively by a major part of the public, but no market mechanism exists for them. Furthermore, the scarcity of periurban forests in many OECD urban areas – in over 40% of them forests account for less than 10% of their total surface – indicates that the external social benefits of them may be substantial. To obtain the socially desirable city size, policy makers may use market mechanisms, such as taxing periurban development, or regulatory instruments (e.g. greenbelts). Policy-makers should, however, ensure that the selected instruments are designed in a way that deters leapfrog development. The shaded areas in the right panel of Figure 4.8 represent the welfare gains from containing urban expansion. This is the difference between the aggregate social value of the land being protected and the aggregate willingness to pay by developers for the same land.
Hydrological effects
Soil performs important filtering and runoff-regulating functions that may be impaired by land development. Local water balance is affected by sealing surfaces for development purposes. Surface sealing leads to a rise in water run-off on impervious surfaces, which results in a reduction in groundwater recharge and a decline in effective evapotranspiration (see also Haase and Nuissl, 2007).1 Haase (2009) highlights the need to identify the exact channels through which development affects the above functions. Both contributions analyse the impact of urban land-use change and development on the urban water balance for a 130‐year period in the German city of Leipzig. Poelmans, van Rompaey and Batelaan (2010) compile land cover maps of the highly urbanised region of Flanders – Brussels to simulate twelve urban expansion scenarios for the future (2025 and 2050). These scenarios are an input to a spatially distributed water balance model that assesses the impact on surface run‐off, evapotranspiration and groundwater recharge.
Development density and lot size may have an important effect on groundwater levels in the presence of domestic wells. Rayne and Bradbury (2011) used models for groundwater flow to test the impact of suburban development on groundwater levels and discharge to streams. In their simulations they use lot sizes of ca. 4 000, 12 000 and 20 000 m2 with one domestic well per lot that pumps water from shallow aquifers. They showed that pumping had little impact on water levels and groundwater discharge to streams, provided that the developed area was of moderate size. However, domestic wells had the potential to impact local groundwater levels and baseflows in township-wide development scenarios of ca. 4 000 m2, where drawdowns beneath developed areas ranged from 0.3 to 5.5 metres, and baseflow reductions ranged from 20 to 40%.
Urban expansion has been blamed for increasing flood risk (Semadeni-Davies et al., 2008). That study simulates forward climate change scenarios characterised by different precipitation rates and various urbanisation patterns to show how urban expansion may be related to drainage. Several studies discuss or simulate the link between land-use change and flood risk (Bhaduri et al., 2001; Rose and Peters, 2001; Bronstert, Niehoff and Bürger, 2002; Hundecha and Bardossy, 2004; Tang et al. 2005; Liu et al., 2006; McColl and Aggett, 2007). All the aforementioned contributions model land-use change; however, the underlying settings do not take into explicit account the critical differences between urbanisation, urban expansion and urban sprawl.
There also appears to be a link between the fraction of land covered by impervious surfaces and the quality of drinking water. The United States Environmental Protection Agency (EPA) has highlighted that water run-off from impervious surfaces may be a threat to water quality due to washed-off contaminants carried directly or indirectly into waterways or the groundwater (United States Environmental Protection Agency, 1994). Berke et al. (2003) compared fifty matched pairs of new urban and conventional developments in five U.S. states to highlight that conventional low-density developments create larger impervious surfaces that generate more runoff than compact developments. Schueler (1994) estimated that compact development can reduce imperviousness by 10‐50%, depending on the average lot size and road network. This reduction stems from the large surfaces that have to be tied-down to roads and parking lots under low-density development patterns. Foster, Morris and Chilton (1999) discuss the impact of rapid urbanisation on water quality.
Effects on biodiversity
Suburban areas often display a relatively high diversity of flora and fauna. The natural habitats at the urban fringe are usually occupied by native species: plants, animals, such as mesopredator mammals, as well as ground-foraging, omnivorous and frugivorous birds. Such species may utilise gardens, forest fragments and other habitats available in suburban areas (McKinney, 2006).
As it is the case with hydrological effects, a part of the existing literature confuses the impact of a fragmented, low-density development pattern with the general effects of urban expansion. Alberti (2005) reviews the empirical evidence of the effects of urban expansion on biodiversity. In that review, it is shown that most of the current knowledge stems from studies that use correlations of aggregate measures of urbanisation, rather than detailed urban form metrics, with environmental performance indicators (including biodiversity). While generic urban expansion affects biodiversity in a clearer manner, the impact of fragmented, low-density development on it is more complex. What is reviewed below is a stream of literature that is more precise when it comes to the effects of specific urban development patterns on biodiversity.
McKinney (2008) reviews 105 studies on the correlation between urbanisation, suburbanisation and biodiversity, focusing on mammals, reptiles, amphibians, invertebrates and plants. For all groups, species richness tends to be reduced in areas with extreme urbanisation, i.e. urban core areas. However, the findings in moderate levels of urbanisation, i.e. suburban areas, vary significantly among groups of species. Most of the studies (65%) indicate that moderate suburbanisation is statistically associated with increased plant biodiversity. The corresponding percentage of studies showing a positive association of low density and biodiversity is substantially smaller in the case of invertebrates (30%) and very small in the case of non-avian vertebrates (12%).
Results from such correlation-based studies must be interpreted with caution for two reasons. First, the issue of reverse causality should be taken into account. People may be more willing to suburbanise areas characterised by rich initial biodiversity conditions. This is reflected in the findings of a voluminous literature, which indicates that the willingness to pay for residential space in locations closer to natural amenities may be high (see e.g. Gibbons, Mourato and Resende, 2014; Mahan, Polasky and Adams, 2000). Failing to account for the fact that specific areas are characterised by low-density development partially because of their biodiversity may result in biased conclusions. A more reliable empirical strategy would involve the repeated measurement of biodiversity indicators at fixed locations and their association with changes in urban form characteristics occurring earlier in time.
At the same time, studies should also focus on disentangling the effect of intentional or unintentional import of animals and plants by suburban households from the evolution of native biodiversity. McKinney (2006) goes as far as claiming that the import effect may be strong enough to fully offset the biodiversity loss in the surrounding landscape, but this argumentation involves questionable assumptions regarding the substitutability between native and invasive species. In fact, invasive species have been shown to have detrimental effects on the evolution of native species (see e.g. Karousakis et al., 2012).
Suburbanisation has been blamed for disrupting periurban ecological corridors that connect different complex ecosystems. A series of studies highlight the potential consequences of habitat fragmentation in different contexts. Soule (1991) conducted a case study in 37 isolated canyons in San Diego, California. That study showed a statistical relationship between habitat fragmentation and the extinction of native birds, with the author concluding that the provision of corridors linking habitat patches may be essential to maintain the balance of wildlife. The field of island biogeography established the principle of inverse size, according to which the rate of species extinction in an isolated patch of habitat is inversely related to its size (MacArthur and Wilson, 1967).
Simulation models provide further insights on how habitat fragmentation may trigger or accelerate the extinction of particular species. Fahring and Merriam (1985) constructed a model to simulate populations in a series of interconnected habitat patches. The model was applied to white-footed mice (Peromyscus Leucopus) inhabiting patches of forest in an agricultural landscape. It predicted that mouse populations in isolated woodlots would have lower growth rates and be more prone to extinction than those in connected woodlots. Fahring and Paloheimo (1988) developed a detailed simulation model of the movement behaviour of adult female cabbage butterflies (Pieris rapae) in patchy habitats. The purpose of this study was to examine the effect of the spatial arrangement of host-plant patches on the local abundance of the butterfly. The results suggested that the effect of the spatial arrangement of habitat patches on local population size also depends on the distances between patches.
Howe (1984) surveyed forest islands, i.e. small isolated forest patches of areas between 0.1 and 7.0 hectares, in eastern New South Wales (Australia) and southern Wisconsin (United States) between 1977 and 1981. Woodlands in both regions were cleared extensively during the century before the study was conducted. The study examined how fragmentation of forest habitat affected the composition and dynamics of local bird populations. Species in forest islands were compared with a control group of species residing at the edge of a large, continuous forest. The study suggested that the disruption of continuous tracts may affect not only bird populations in the interior of the forest island, but also those residing along or near the edge.
Lynch and Whigham (1984) conducted surveys to estimate the abundance and diversity of birds in relation to the size, spatial arrangement and characteristics of 270 forest patches in the state of Maryland (United States). In that study, densities of permanent resident and short-distance migrant species tended to be less affected by the site characteristics, or showed responses of the opposite direction to those of long-distance migrants. The findings indicate that different species may respond differently to factors such as patch area and isolation. Partition of the habitat into small, highly isolated patches of forest may adversely affect some species, but other factors may be more important for other species.
While the relationship between biodiversity and fragmentation of natural habitats has been roughly established, the relationship between habitat fragmentation and fragmentation of urban fabric is not as straightforward as might be expected, as Figure 4.9 suggests. An urban area can be highly fragmented with well-connected periurban natural habitats, as the left panel of the figure suggests. In such a context, trying to reduce sprawl by connecting the fragments of urban fabric may have an adverse effect on biodiversity. As the middle panel of Figure 4.9 suggests, such an option will result in high fragmentation of the natural habitats as all major ecological corridors in periurban areas will be disrupted. Finally, the right panel of the same figure suggests that low fragmentation of the urban fabric may coexist with contiguous natural habitats in periurban areas.
Note: Left panel: A hypothetical area with highly fragmented urban fabric and minimal habitat fragmentation (three forest islands). Middle panel: Further development that minimises fragmentation of urban fabric results in massive habitat fragmentation (more than fifteen forest islands). Right panel: An alternative pattern with minimal fragmentation of urban fabric and natural habitat. Colour keys: green: natural habitats; blue: water bodies; grey: artificial land; brown: undeveloped areas.
Urban heat islands and resilience to climate change
Prolonged human exposure to high temperatures can be a primary cause of death, for instance by triggering strokes, or a major contributing factor to life-shortening conditions, such as heart and pulmonary disorders. Such health effects are becoming a major issue, as the overall temperature rises and extreme heat events occur more often (Gaffen and Ross, 2008). Robine et al. (2008) estimated that the European heat wave of 2003 caused more than 70 000 excess deaths in the course of a few months. While human exposure to extreme heat events is an overall issue, the problem seems to be more severe in urban contexts. During persistent heat episodes, specific urban areas or entire metropolitan regions known as urban heat islands (UHI) have been observed to be significantly warmer than their surrounding rural areas. The critical temperature differential, which can exceed 10°C, can be the result of several factors (Oke, 1982).
Some of these factors that contribute to the emergence of urban heat islands are related to the material composition of built environment. Impervious surfaces such as buildings, roads and other types of infrastructure are more likely to have lower reflectivity (albedo) due to their colour. Such surfaces may have limited capacity to reradiate heat. In addition, the evapotranspiration functions performed by trees and plants in a certain area may be distorted due to the loss of vegetation. In that sense, a low-density development pattern may contribute to exacerbating the phenomenon since it requires a larger uptake of land (not only for buildings, but also for roads and other types of infrastructure) to be covered by material of higher thermal storage capacity.
However, UHIs are far from being homogeneous structures. Instead, Parry (1962) suggested that a UHI can be described as a combination of several smaller areas, each one with its own microclimate that is co-determined by its unique land-use characteristics. Stated differently, the local configuration of the built environment affects microclimate. For instance, it is well-known that the presence of tall buildings gives rise to urban canyons that hamper the dispersion of the heat generated by residences and traffic. In that sense, a low-density development pattern characterised by shorter, detached constructs may contribute to the mitigation of the phenomenon.
The various features of development patterns are likely to affect the emergence of UHI in opposing ways. For this reason, current knowledge regarding the urban development pattern that is the least likely to lead to UHIs is scarce (Stone and Norman, 2006). Understanding the mechanisms through which a development pattern may mitigate or exacerbate the phenomenon will be key in promoting urban forms that mitigate climate change and are more resilient to it.
The relationship between urban form and microclimate was recognised early on in the literature, with Clarke (1972) suggesting that the relatively high number of deaths occurring in urban areas during periods of extreme heat can be significantly reduced through appropriate urban land use. Stone and Norman (2006) find that the contribution of individual land parcels to the phenomenon could be reduced by approximately 40% through the adoption of specific land-use planning policies, such as zoning and subdivision regulations, without modifications in the size or albedo of the residential structures. Golany (1996) argues that the thermally desirable configuration of the built environment significantly varies with geographic latitude.
Obesity and other health effects
The incidence of obesity is different across regions and urban areas of different development patterns. For instance, Mokdad et al. (1999) report that during the period between 1991 and 1998 obesity prevalence grew by almost 102% in Georgia but only slightly over 11% in Delaware. The fact that obesity increases relatively faster in geographic areas containing less compact cities has led some researchers to claim that variations in the built environment may have a significant impact on obesity. The rationale is that built environment may impose certain constraints that affect exercise, diet and other lifestyle choices. For instance, as already discussed, low-density areas are costly to be covered by public transport and are harder to become walkable. On the other hand, they may provide better accessibility to open spaces, such as periurban forests, where physical exercise can take place without substantial costs. In line with that, Berrigan and Troiano (2002) find a significant correlation between urban form and physical activity. A conceptual framework that makes explicit the linkages between urban design, car dependency and public health is provided by Frank and Engelke (2001).
A number of studies investigate whether there is a positive relationship between urban sprawl and obesity, but they provide mixed evidence of the existence of such a link (Ewing, Brownson and Berrigan, 2006; Ewing et al., 2003; Frank, Andresen and Schmid, 2004; Giles-Corti et al., 2003; Lopez, 2004; McCann and Ewing, 2003; Saelens et al., 2003; Zhao and Kaestner, 2010). Sturm and Cohen (2004) found that sprawl is associated with physical conditions such as asthma, diabetes, hypertension, arthritis, rheumatism, physical disability, problems in breathing, emphysema or chronic obstructive pulmonary disease (COPD), cancer, neurological conditions, stroke or paralysis, heart failure, coronary artery disease, chronic back problems, abdominal problems (ulcer, colitis, enteritis), chronic liver disease, migraine or chronic severe headaches, chronic bladder problems or problems urinating and other chronic pain conditions. On the other hand, the same study failed to identify a relationship between urban structure and mental health conditions.
While a negative statistical relationship between sprawl and physical health has been identified in the aforementioned studies, none of them managed to establish a causal effect. The most important methodological barrier in most cases is self-selection: among others, suburban residents are more likely to have selected suburbs over downtown areas because they favour a less physically-active lifestyle with limited exercise and the vast majority of their trips undertaken by car. Low-density areas accommodate these preferences more cheaply and easily. Overcoming that methodological constraint, Eid et al. (2008) provide one of the best-conducted studies in the potential relationship between sprawl and obesity. They show that previous findings of a positive relationship between sprawl and obesity were most probably due to a failure to properly control for individual characteristics and preferences. The study indicates that trying to curb obesity by changing the built environment may be a poor policy response.
Plantinga and Bernell (2007) find a two-way relationship between sprawl and the body mass index (BMI). In particular, individuals of high BMI are expected to lose weight if they move to denser locations. However, this is not to be attributed to population density per se, but to a series of mobility constraints often ceasing when someone moves to a denser area. At the same time, the self-selection effect is confirmed. BMI is an important determinant of the choice: people of high BMI tend to prefer sprawling locations, as they prefer to move by car.
4.4. Concluding remarks
This chapter provided a comprehensive discussion of the main drivers of urban sprawl, as the latter was defined in Chapter 2. Special attention was paid to the preference-driven causes of urban sprawl, as well as to forces that contribute to a growth pattern that may appear to be sprawling in the short run, but constitutes a temporary stage in the long-run development of a compact city. The chapter presented a series of mechanisms through which some policy interventions may give rise to more sprawled urban environments. Furthermore, it provided a review of the multiple impacts of urban sprawl on the environment, economy and society. Abstaining from emotionally-charged indictments of sprawl, the discussion took into explicit account the internal and external validity of numerous studies, some of which provide contradictory evidence. Accounting for the various methodological limitations and assessing the reliability of a voluminous body of literature, it was highlighted that while many of the concerns raised are sensible, many others may lack substantial evidence-based support. Equally important, it was also highlighted that some concerns may refer to effects erroneously attributed to urban sprawl; instead, these effects stem from other forms of urban development, or urbanisation per se. Accounting for the critical differences between generic urban expansion and sprawl paves the way for interventions that target urban sprawl, rather than urbanisation. A series of such interventions will be provided in Chapter 5.
References
Alberti, M. (2005), “The effects of urban patterns on ecosystem function”, International Regional Science Review, Vol. 28/2, pp. 168-192.
Alcamo, J. et al. (2005), “Changes in nature’s balance sheet: Model-based estimates of future worldwide ecosystem services”, Ecology and Society, Vol. 10/2, 19.
Alonso, W. (1964), Location and Land Use, Harvard University Press, Cambridge, MA.
Anas, A., R. Arnott and K.A. Small (1998), “Urban spatial structure”. Journal of Economic Literature, Vol. 36/3, pp. 1426-1464.
Arnott, R.J. and J.G. MacKinnon (1977), “The effects of the property tax: A general equilibrium simulation”, Journal of Urban Economics, Vol. 4/4, pp. 389-407.
Bento, A.M. et al. (2005), “The effects of urban spatial structure on travel demand in the United States”, The Review of Economics and Statistics, Vol. 87/3, pp. 466-478.
Bento, A.M. et al. (2003), “The impact of urban spatial structure on travel demand in the U.S.”, World Bank Policy Research Working Paper No. 3007, The World Bank, Washington, DC.
Berke, P.R. et al. (2003), “Greening development to protect watersheds: Does new urbanism make a difference?”, Journal of the American Planning Association, Vol. 69/4, pp. 397-413.
Berrigan, D. and R.P. Troiano (2002), “The association between urban form and physical activity in U.S. adults”, American Journal of Preventive Medicine, Vol. 23/2, pp. 74-79.
Bertaud, A. and J.K. Brueckner (2005), “Analyzing building-height restrictions: Predicted impacts and welfare costs”, Regional Science and Urban Economics, Vol. 35/2, pp. 109-125.
Bhaduri, B. et al. (2001), “Long-term hydrologic impact of urbanization: A tale of two models”, Journal of Water Resources Planning and Management, Vol. 127/1, pp. 13-19.
Bird, R.M. and E. Slack (2004), “Land and property taxation in 25 countries: A comparative review”, in Bird, R.M. and E. Slack (eds), International Handbook of Land and Property Taxation, Edward Elgar, Cheltenham, UK, and Northampton, MA, USA.
Braid, R.M. (1991), “Residential spatial growth with perfect foresight and multiple income groups”, Journal of Urban Economics, Vol. 30/3, pp. 385-407.
Braid, R.M. (1988), “Optimal spatial growth of employment and residences”, Journal of Urban Economics, Vol. 24/2, pp. 227-240.
Bronstert, A., D. Niehoff and G. Bürger (2002), “Effects of climate and land-use change on storm runoff generation: Present knowledge and modelling capabilities”, Hydrological Processes, Vol. 16/2, pp. 509-529.
Brueckner, J.K. (2001), “Property taxation and urban sprawl”, in Oates, W.E. (ed.), Property taxation and local government finance, Lincoln Institute of Land Policy, Cambridge, MA.
Brueckner, J.K. (2000), “Urban sprawl: Diagnosis and remedies”, International Regional Science Review, Vol. 23/2, pp. 160-171.
Brueckner, J.K. and D. Fansler (1983), “The economics of urban sprawl: Theory and evidence on the spatial sizes of cities”, The Review of Economics and Statistics, Vol. 65/3, pp. 479-482.
Brueckner, J.K. and H.A. Kim (2003), “Urban sprawl and the property tax”, International Tax and Public Finance, Vol. 10/1, pp. 5-23.
Brueckner, J.K. and B. von Rabenau (1981), “Dynamics of land-use for a closed city”, Regional Science and Urban Economics, Vol. 11/1, pp. 1-17.
Burchfield, M. et al. (2006), “Causes of sprawl: A portrait from space”, The Quarterly Journal of Economics, Vol. 121/2, pp. 587-633.
Carlton, D.W. (1981), “The Spatial Effects of a Tax on Housing and Land”, Regional Science and Urban Economics, Vol. 11/4: 509-527.
Carruthers, J.I. and G.F. Ulfarsson (2008), “Does smart growth matter to public finance? Evidence from the United States”, Urban Studies, Vol. 45/9, pp. 1791-1823.
Carruthers, J.I. and G.F. Ulfarsson (2003), “Urban sprawl and the cost of public services”, Environment and Planning B: Planning and Design, Vol. 30/4, pp. 503-522.
Cavailhès, J. et al. (2004), “The periurban city: Why to live between the suburbs and the countryside”, Regional Science and Urban Economics, Vol. 34/6, pp. 681-703.
Cervero, R. (2002), “Built environments and mode choice: Toward a normative framework”, Transportation Research Part D: Transport and Environment, Vol 7/4, pp. 265-284.
Clarke, J.F. (1972), “Some effects of the urban structure on heat mortality”, Environmental Research, Vol. 5/1, pp. 93-104.
Clawson, M. (1962), “Urban sprawl and speculation in suburban land”, Land Economics, Vol. 38/2, pp. 99-111.
Coisnon, T., W. Oueslati and J. Salanié (2014), “Urban sprawl occurrence under spatially varying agricultural amenities”, Regional Science and Urban Economics, Vol. 44, pp. 38-49.
Deng, X. et al. (2008), “Growth, population and industrialization, and urban land expansion of China”, Journal of Urban Economics, Vol. 63/1, pp. 96-115.
Duncan, J. and Associates (1989), The search for efficient urban growth patterns: A study of the fiscal impacts of development in Florida, Florida Department of Community Affairs, Tallahassee, FL.
EEA (2016), Urban sprawl in Europe – joint EEA-FOEN report, EEA Report No. 11/2016, Publications Office of the European Union, Luxembourg.
Eid, J. et al. (2008), “Fat city: Questioning the relationship between urban sprawl and obesity”, Journal of Urban Economics, Vol. 63/2, pp. 385-404.
Eigenbrod, F. et al. (2011), “The impact of projected increases in urbanization on ecosystem services”, Proceedings of the Royal Society of London B: Biological Sciences, Vol. 278, pp. 3201-3208.
Ewing, R., R.C. Brownson and D. Berrigan (2006), “Relationship between urban sprawl and weight of United States youth”, American Journal of Preventive Medicine, Vol. 31/6, pp. 464-474.
Ewing, R. et al. (2003), “Relationship between urban sprawl and physical activity, obesity, and morbidity”, American Journal of Health Promotion, Vol. 18/1, pp. 47-57.
Federal Home Mortgage Association (1994), National Housing Survey. Federal Home Mortgage Association, Washington, DC.
Fischel, W.A. (2001), “Why are there NIMBYs?”, Land Economics, Vol. 77/1, pp. 144-152.
Fischel, W.A. (1985), The Economics of Zoning Laws: A Property Rights Approach to American Land Use Controls, The Johns Hopkins University Press, Baltimore, MD.
Foster, S.S.D., B.L. Morris and P.J. Chilton (1999), “Groundwater in urban development – a review of linkages and concerns”, in Ellis, B.J. (Ed.), Impacts of Urban Growth on Surface Water and Groundwater Quality. IAHS Publications, No. 259, Oxfordshire, UK.
Frank, J.E. (1989), The Costs of Alternative Development Patterns: A Review of the Literature, Urban Land Institute, Washington, DC.
Frank, L.D., M.A. Andresen and T.L. Schmid (2004), “Obesity relationships with community design, physical activity, and time spent in cars”, American Journal of Preventive Medicine, Vol. 27/2, pp. 87-96.
Frank, L.D. and P.O. Engelke (2001), “The built environment and human activity patterns: Exploring the impacts of urban form on public health”, Journal of Planning Literature, Vol. 16/2, pp. 202-218.
Fujita, M. (1976), “Spatial patterns of urban-growth: Optimum and market”, Journal of Urban Economics, Vol. 3/3, pp. 209-241.
Gaffen, D.J., and R. Ross (1998), “Increased summertime heat stress in the United States”, Nature, Vol. 396/6711, pp. 529-530.
Garcia-López, M.-Á., A. Holl and E. Viladecans-Marsal (2015), “Suburbanization and highways in Spain when the Romans and the Bourbons still shape its cities”, Journal of Urban Economics, Vol. 85, pp. 52-67.
Geniaux, G., J.S. Ay and C. Napoléone (2011), “A spatial hedonic approach on land-use change anticipations”, Journal of Regional Science, Vol. 51/5, pp. 967-986.
Gibbons, S., S. Mourato and G.M. Resende (2014), “The amenity value of English nature: A hedonic price approach”, Environmental and Resource Economics, Vol. 57/2, pp. 175-196.
Giles-Corti, B. et al. (2003), “Environmental and lifestyle factors associated with overweight and obesity in Perth, Australia”, American Journal of Health Promotion, Vol. 18/1, pp. 93-102.
Glaeser, E.L. and M.E. Kahn (2010), “The greenness of cities: Carbon dioxide emissions and urban development”, Journal of Urban Economics, Vol. 67/3, pp. 404-418.
Glaeser, E.L. and M.E. Kahn (2004), “Sprawl and urban growth”, in J.-F. Henderson and J. Vernon Thisse (eds.), Handbook of Regional and Urban Economics, Vol. 4, Cities and Geography. Elsevier, Amsterdam.
Golany, G. (1996), “Urban design morphology and thermal performance”, Atmospheric Environment, Vol. 30/3, pp. 455-465.
Gordon, P. and H.W. Richardson (1997), “Are Compact Cities a Desirable Planning Goal?”, Journal of the American Planning Association, Vol. 63/1, pp. 89-106.
Haase, D. (2009), “Effects of urbanisation on the water balance – A long-term trajectory”, Environmental Impact Assessment Review, Vol. 29/4, pp. 211-219.
Haase, D. and H. Nuissl (2007), “Does urban sprawl drive changes in the water balance and policy? The case of Leipzig (Germany) 1870-2003”, Landscape and Urban Planning, Vol. 80/1, pp. 1-13.
Hankley, S. and J.D. Marshall (2010), “Impacts of urban form on future US passenger-vehicle greenhouse gas emissions”, Energy Policy, Vol. 38/9, pp. 4880-4888.
Hortas-Rico, M. and A. Albert Solé-Ollé (2010), “Does urban sprawl increase the costs of providing local public services? Evidence from Spanish municipalities”, Urban Studies, Vol. 47/7, pp. 1513-1540.
Howe, R.W. (1984), “Local dynamics of bird assemblages in small forest habitat islands in Australia and North America”, Ecology, Vol. 65/5, pp. 1585-1601.
Hundecha, Y. and A. Bardossy (2004), “Modeling of the effect of land use changes on the runoff generation of a river basin through parameter regionalization of a watershed model”, Journal of Hydrology, Vol. 292/1-4, pp. 281-295.
Ihlanfeldt, K.R. (2007), “The effect of land use regulation on housing and land prices”, Journal of Urban Economics, Vol. 61/3, pp. 420-435.
Irwin, E.G. and N.E. Bockstael (2004), “Land use externalities, open space preservation, and urban sprawl”, Regional Science and Urban Economics, Vol. 34/6, pp. 705-725.
Kahn, M. (2000), “The environmental impact of suburbanization”. Journal of Policy Analysis and Management, Vol. 19/ 4, pp. 569-586.
Karousakis, K. et al. (2012), “Biodiversity”, in OECD Environmental Outlook to 2050: The Consequences of Inaction, OECD Publishing, Paris, http://dx.doi.org/10.1787/env_outlook-2012-7-en.
Kenworthy, J.R. and F.B. Laube (1996), “Automobile dependence in cities: An international comparison of urban transport and land use patterns with implications for sustainability”, Environmental Impact Assessment Review, Vol. 16/4, pp. 279-308.
Koomey, J.G. et al. (2000), “Buildings sector”, in Interlaboratory Working Group, Scenarios for a Clean Energy Future, ORNL/CON-476, LBNL-44029, and NREL/TP-620-29379, Oak Ridge National Laboratory, Oak Ridge, TN, Lawrence Berkeley National Laboratory, Berkeley, CA and National Renewable Energy Laboratory, Golden, CO.
Ladd, H.F. (1994), “Fiscal impacts of local population growth: A conceptual and empirical analysis”, Regional Science and Urban Economics, Vol. 24/6, pp. 661-686.
Ladd, H.F. (1992), “Population growth, density and the costs of providing public services”, Urban Studies, Vol. 29/2, pp. 273-295.
LeRoy, S.F. (1976), “Urban land rent and the incidence of property taxes”, Journal of Urban Economics, Vol. 3/2, pp. 167-179.
Lisansky, J. (1986), “Farming in an urbanizing environment: Agricultural land use conflicts and right to farm”, Human Organization, Vol. 45/4, pp. 363-371.
Liu, Y.B. et al. (2006), “Predicting storm runoff from different land-use classes using a geographical information system based distributed model”, Hydrological Processes, 20/3, pp. 533-548.
Lopez, R. (2004), “Urban sprawl and risk for being overweight or obese”, American Journal of Public Health, Vol. 94/9, pp. 1574-1579.
Lynch, L. and J. Carpenter (2003), “Is there evidence of a critical mass in the mid-Atlantic agriculture sector between 1949 and 1997?”, Agricultural and Resource Economics Review, Vol. 32/1, pp. 116-128.
Lynch, J.F. and D.F. Whigham (1984), “Effects of forest fragmentation on breeding bird communities in Maryland, USA” Biological Conservation, Vol. 28/4, pp. 287-324.
MacArthur, R.H. and E.O. Wilson (1967), The Theory of Island Biogeography, Princeton University Press, Princeton, NJ.
Mahan, B.L., S. Polasky and R.M. Adams (2000), “Valuing urban wetlands: A property price approach”, Land Economics, Vol. 76/1, pp. 100-113.
McCann, B.A. and R. Ewing (2003), Measuring the health effects of sprawl: A national analysis of physical activity, obesity and chronic disease, Smart Growth America, Surface Transportation Policy Project, Washington, DC.
McColl, C. and G. Aggett (2007), “Land-use forecasting and hydrologic model integration for improved land-use decision support”, Journal of Environmental Management, Vol. 84/4, pp. 494-512.
McGrath, D.T. (2005), “More evidence on the spatial scale of cities” Journal of Urban Economics, Vol. 58/1, pp. 1-10.
McKinney, M.L. (2008), “Effects of urbanization on species richness: A review of plants and animals” Urban Ecosystems, Vol. 11/2, pp. 161-176.
McKinney, M.L. (2006), “Urbanization as a major cause of biotic homogenization”, Biological Conservation, Vol. 127/3, pp. 247-260.
Mills, D.E. (1981), “Growth, speculation and sprawl in a monocentric city”, Journal of Urban Economics, Vol. 10/2, pp. 201-226.
Mokdad, A.H. et al. (1999), “The spread of the obesity epidemic in the United States, 1991-1998”, Journal of the American Medical Association, Vol. 282/16, pp. 1519-1522.
Muth, R.F. (1969), Cities and Housing. University of Chicago Press, Chicago.
Nechyba, T.J. and R.P. Walsh (2004), “Urban sprawl”, The Journal of Economic Perspectives, Vol. 18/4, pp. 177-200.
Newman, P. and J. Kenworthy (1989), “Gasoline consumption and cities: A comparison of U.S. cities with a global survey”, Journal of the American Planning Association, Vol. 55/1, pp. 24-37.
OECD (2009), Farmland Conversion: The Spatial Dimension of Agricultural and Land Use Policies, OECD Publishing, Paris, www.oecd.org/agriculture/44535648.pdf.
Ohls, J.C. and D. Pines 1(975), “Discontinuous urban development and economic efficiency”, Land Economics, Vol. 51/3, pp. 224-234.
Oke, T.R. (1982), “The energetic bias of the urban heat island”, Quarterly Journal of the Royal Meteorological Society, Vol. 108/455, pp. 1-24.
Oueslati, W., S. Alvanides and G. Garrod (2015), “Determinants of urban sprawl in European cities”, Urban Studies, Vol. 52/9, pp. 1594-1614.
Patacchini, E. and Y. Zenou (2009), “Urban sprawl in Europe”, Brookings-Wharton Papers on Urban Affairs, Vol. 1, pp. 125-149.
Plantinga, A.J. and S. Bernell (2007), “The association between urban sprawl and obesity: Is it a two-way street?” Journal of Regional Science, Vol. 45/3, pp. 473-492.
Poelmans, L., A. van Rompaey and O. Batelaan (2010), “Coupling urban expansion models and hydrological models: How important are spatial patterns?” Land Use Policy, Vol. 27/3, pp. 965-975.
Rayne, T.W. and K.R. Bradbury (2011), “Evaluating impacts of subdivision density on shallow groundwater in Southeastern Wisconsin, USA”, Journal of Environmental Planning and Management, Vol. 54/5, pp. 559-575.
Ready, R.C. and C.W. Abdalla (2005), “The amenity and disamenity impacts of agriculture: Estimates from a hedonic pricing model”, American Journal of Agricultural Economics, Vol. 87/2, pp. 314-326.
Real Estate Research Corporation (1974), The Costs of Sprawl: Environmental and Economic Costs of Alternative Residential Development Patterns at the Urban Fringe, U.S. Government Printing Office, Washington, DC.
Robine, J.M. et al. (2008), “Death toll exceeded 70,000 in Europe during the summer of 2003” Comptes Rendus Biologies, Vol. 331/2, pp. 171-178.
Roe, B., E.G. Irwin and H.A. Morrow-Jones (2004), “The effects of farmland, farmland preservation, and other neighborhood amenities on housing values and residential growth”, Land Economics, Vol. 80/1, pp. 55-75.
Rose, S. and N. Peters (2001), “Effects of urbanization on streamflow in the Atlanta area (Georgia, USA): A comparative hydrological approach”, Hydrological Processes, Vol. 15/8, pp. 1441-1457.
Saelens, B.E. et al. (2003), “Neighborhood based differences in physical activity: An environment scale evaluation”, American Journal of Public Health, Vol. 93/9, pp. 1552-1558.
Saiz, A. (2010), “The geographic determinants of housing supply”, The Quarterly Journal of Economics, Vol. 125/3, pp. 1253-1296.
Savage, I. (2004), “Management objectives and the causes of mass transit deficits”, Transportation Research Part A: Policy and Practice, 38/3, pp. 181-199.
Savage, I. and A. Schupp (1997), “Evaluating transit subsidies in Chicago”, Journal of Public Transportation, Vol. 1, pp. 93-117.
Schueler, T. (1994), “Use of cluster development to protect watersheds”, Watershed Protection Techniques, Vol. 1/1, pp. 137-140.
Schuetz, J. (2009), “No renters in my suburban backyard: Land use regulation and rental housing”, Journal of Policy Analysis and Management, Vol. 28/2, pp. 296-320.
Semadeni-Davies, A. et al. (2008), “The impacts of climate change and urbanisation on drainage in Helsingborg, Sweden: Suburban stormwater”, Journal of Hydrology, Vol. 350/1, pp. 114-125.
Shanzi, K., Y. Song and H. Ming (2009), “Determinants of urban spatial scale: Chinese cities in transition”, Urban Studies, Vol. 46/13, pp. 1-19.
Song, Y. and Y. Zenou (2006), “Property tax and urban sprawl: Theory and implications for US cities”, Journal of Urban Economics, Vol. 60/3, pp. 519-534.
Soule, M.E. (1991), “Land use planning and wildlife maintenance: Guidelines for conserving wildlife in an urban landscape”, Journal of the American Planning Association, Vol. 57/3, pp. 313-323.
Speir, C. and K. Stephenson (2002), “Does sprawl cost us all? Isolating the effects of housing patterns on public water and sewer costs”, Journal of the American Planning Association, Vol. 68/1, pp. 56-70.
Stone, B. and J.M. Norman (2006), “Land use planning and surface heat island formation: A parcel-based radiation flux approach”, Atmospheric Environment, Vol. 40/19, pp. 3561-3573.
Sturm, R. and D.A. Cohen (2004), “Suburban sprawl and physical and mental health”, Public Health, Vol. 118/7, pp. 488-496.
Su, Q. and J.S. DeSalvo (2008), “The effect of transportation subsidies on urban sprawl”, Journal of Regional Science, Vol. 48/3, pp. 567-594.
Sullivan, A.M. (1985), “The general equilibrium effects of the residential property tax: Incidence and excess burden”, Journal of Urban Economics, Vol. 18/2, pp. 235-250.
Sullivan, A.M. (1984), “The general equilibrium effects of the industrial property tax: Incidence and excess burden”, Regional Science and Urban Economics, Vol. 14/4, pp. 547-563.
Tang, Z. et al. (2005), “Forecasting land use change and its environmental impact at a watershed scale”, Journal of Environmental Management, Vol. 76, pp. 35-45.
Taylor, B.D. et al. (2009), “Nature and/or nurture? Analyzing the determinants of transit ridership across U.S. urbanized areas” Transportation Research Part A: Policy and Practice, Vol. 43/1, pp. 60-77.
Turnbull, G.K. (1988), “Residential development in an open city”, Regional Science and Urban Economics, Vol. 18/2, pp. 307-320.
United States Environmental Protection Agency (EPA) (1994), The Quality of our Nation’s Water: 1992, Report #EPA-841-S-94-002, US EPA Office of Water Quality, Washington, DC.
Ürge-Vorsatz, D. et al. (2007), “Mitigating CO2 emissions from energy use in the world’s buildings”, Building Research and Information, Vol. 35/4, pp. 379-398.
Wheaton, W.C. (1974), “A comparative static analysis of urban spatial structure”, Journal of Economic Theory, Vol. 9/2, pp. 223-237.
Wilson, B. and A. Chakraborty (2013), “The environmental impacts of sprawl: Emergent themes from the past decade of planning research”, Sustainability, Vol. 5/8, pp. 3302-3327.
Wu, J.J., M. Fisher and U. Pascual (2011), “Urbanization and the viability of local agricultural economies”, Land Economics, Vol. 87/1, pp. 109-125.
Zhao, Z. and R. Kaestner (2010), “Effects of urban sprawl on obesity”, Journal of Health Economics, 29/6, pp. 779-787.
Note
← 1. Groundwater recharge is the hydrological process of water moving from the surface downward to soil pore spaces and fractures of underground rock formations. Evapotranspiration is the process of water moving to the air directly from the soil or through plants.