Cities are the most intricate of human creations and no two cities grow and develop in exactly the same way. It is important to understand not only how different cities grow but why they grow as they do. Why does one city grow through leapfrog development, while others grow contiguously outwards? Why are some cities denser than others? To what extent does physical geography dictate the form of a city? What is the role of policy and governance in shaping a city?

To explore some of these questions, we undertook an "apples-to-apples" comparison of three Canadian metropolitan regions: Calgary, Toronto, and Vancouver. We first used satellite imagery to define the geographical extent of each city, and census data to track changes in population and dwelling counts over time. This method of comparing cities can be used elsewhere. Previous comparisons of cities have foundered on differences in defining local jurisdictions and different ways of tracking and recording growth in different places.

We then looked at how land use plans and growth policies have developed in each region, and assessed the degree to which the observed growth patterns corresponded to these plans and policies. A study of this kind is both useful and unusual because the process of planning and the content of plans usually get more attention than their outcomes. By exploring the relationship between the two we can draw conclusions about what types of policies are most likely to be effective in different circumstances.

We focused on Toronto, Calgary, and Vancouver because the population of each has grown rapidly over the past half-century. At the same time, each is governed by a different configuration of provincial, regional, and municipal institutions for land use planning and infrastructure provision, and faces different geographical constraints on growth. The approach allows us to examine some of the effects of these differences.

We were also interested in exploring the differences in their planning cultures, which are often overlooked as being less significant than geographical differences. Vancouver is often held up as a model of sustainable development and good planning, while Calgary and Toronto have been condemned for unplanned, automobile-oriented, low-density sprawl. Rather than play into these stereotypes, we hope to foster discussion about the merits of different approaches to planning and metropolitan governance.

To study how urban growth occurred in the three regions between 1991 and 2001, we first looked at changes in the population, number of dwellings, extent of urban land, composition of the dwelling stock, and gross urban density for each region between 1991 and 2001. Second, we studied the location and characteristics of urban development on previously rural lands. Finally, we examined the location and characteristics of infill and redevelopment in existing urban areas.

Changes in population, housing, and urban land

One common definition of "sprawl" is whether the rate of increase in urban land exceeds the rate of growth in population or housing. Between 1991 and 2001, in Calgary, the urban land area grew by 43% while the population grew by 24%. A similar pattern occurred in Toronto (28% versus 19%). In Vancouver, by contrast, the urban area grew at a rate two-thirds that of the population growth rate (16% versus 24%). Put another way, for every 100 new residents, Calgary added 6.3 hectares of urban land, Toronto 4.4, and Vancouver 2.3. By this definition, Calgary and Toronto sprawled during the study period, but Vancouver did not.

To some extent this pattern can be explained by the fact that almost three-quarters of residential construction in Calgary between 1991 and 2001 was in the form of single detached houses, compared to 50% in Toronto and 16% in Vancouver. By contrast, 50% of construction in Vancouver was in the form of apartments, compared to 25% in Toronto and 12% in Calgary.

The overall density of each metropolitan region also changed during the study period. Vancouver's population and housing density increased between 1991 and 2001, while Toronto's and Calgary's declined.

Intensification vs. greenfield development

Vancouver accommodated 80% of its residential growth through intensification (that is, by adding new housing units within the already urbanized area), while in Calgary, 78% of the housing growth occurred as greenfield development in the urban fringe. Toronto's rate of intensification (44%) was in between Calgary and Vancouver's values.

Using spatial analysis (GIS) techniques, we studied patterns of development in the urban fringe, an area just outside the urban boundary and found that greenfield development was fairly contiguous with existing urbanization in all three regions, rather than in "leapfrog" form.

Finally, we studied patterns of intensification within the urban boundary. In the Toronto region, intensification tends to be clustered in certain areas, many of which are designated in planning documents as "nodes" to which intensification should be directed. In Vancouver, the downtown and eight town centres attracted 24% of all intensification and 19% of total growth. Calgary's comparatively small amount of intensification is sparsely distributed across the urbanized area.

In Toronto, regional planning institutions have come and gone, and regional governance has been weak to non-existent. However, within the region, the promotion of intensification became increasingly important over the study period, and the policy of directing growth to nodes has been fairly strong.

Vancouver, by contrast, has a long tradition of regional planning and strong regional governance. Support for intensification, nodal development, and urban containment is high.

Calgary is a single municipality that has expanded through periodic annexation of surrounding territory, avoiding the need for regional cooperation among different municipal jurisdictions. Support for intensification, nodal development, and urban containment, however, is low.

All three city-regions have policies to ensure that greenfield development occurs in an orderly way on fully serviced land at the edge of the existing urban area, rather than in a "leapfrog" manner.

Do observed regional growth patterns correspond to planning principles in each region?

The analysis shows substantial correspondence between policies and development patterns. Vancouver, with its physical boundaries and long-term adherence to policies restricting the urbanization of rural land, experienced a high rate of intensification during the 1990s (80%), while Calgary, largely unconstrained, had a low intensification rate (22%). The degree of urban containment grew stronger over the study period in Toronto as the Ontario government enacted stronger policies and comprehensive rural land protections; its level of intensification lies between those of Vancouver and Calgary (44%).

The urban fringe analysis suggests that policies aimed at contiguous urban expansion (as opposed to leap-frog development) have succeeded in all three regions. The vast majority of urban land added in each region over the study period was adjacent to (within 2 km of) the existing urban area.

Finally, the designation of downtowns, nodes, centres, and other already urbanized zones as preferred destinations for growth has met with modest success in Toronto and greater success in Vancouver. In Calgary, which has not pursued a similar policy, few concentrations of intensification occurred.

Explaining the correspondence between policies and outcomes

It appears that the stability and continuity of institutions in Vancouver and Calgary have led to the consistent implementation of planning policies over the long term, while Toronto's inconsistent governance may explain why some policies have been only weakly applied.

Moreover, Vancouver's and Calgary's regional institutions have persisted because they have been able to adapt to changing needs. Vancouver's regional district has acquired new responsibilities over time and expanded its territorial jurisdiction as the urban area grew. Similarly, by following a long-term policy of periodic annexation of surrounding territory, the City of Calgary has also proved to be a flexible structure of regional government. The Toronto region's structures have proven to be less flexible. The creation of the regional municipalities in the early 1970s made it difficult to address the effects of regional growth, except through provincial intervention.

In all three regions, provincial governments have largely set the rules for planning by municipalities and regional authorities. In Vancouver and Calgary, the provincial governments created systems that made it easier for local authorities to identify and pursue regional planning objectives. The opposite has been true in Toronto, where provincial intervention has generally superseded local authority.

Finally, each region has a distinctive planning culture shaped by geography and the perception of physical limits to growth. Calgary, which is relatively unconstrained by geography, exhibits a high degree of consistency in planning ideas. Since the 1950s, its planning policies have been primarily and consistently directed toward the efficient production and servicing of low-cost housing on greenfields, rather than intensification. By contrast, since the end of the Second World War, Vancouver's constrained setting has impressed on planners and citizens alike the need to regulate urbanization, even though it would be many decades before potentially developable land in the region would be fully exploited.

In Toronto there is no obvious physical limit to outward urban expansion north of Lake Ontario and no early consensus emerged on long-term limits to growth. The city grew rapidly in the postwar period and only with the rise of environmentalism and the perception of the costs of growth in the 1970s and 1980s did planning practice shift its emphasis from enabling expansion to managing growth.

What have we learned?

Calgary, Vancouver, and Toronto have very different patterns of growth.Although these three cities grew at similar rates during the 1990s, each accommodated growth in distinct ways. Calgary's growth occurred mainly through greenfield development (78%) while Vancouver directed the vast majority of its growth (80%) to already-urbanized areas. Urban growth in the Toronto region represents a combination of greenfield development (56%) and intensification (44%).

Geography matters, but cannot explain all differences. It is easy to assume that Vancouver's relatively high density can be attributed to constrictions on growth caused by the presence of mountains and sea, and that Calgary's low density is the result of the lack of geographical restrictions on its expansion, but the reality is not so simple. As much as Vancouver's mountains and ocean and Calgary's open prairie function as long-term physical determinants of growth patterns, their role is largely symbolic. The Lower Mainland has plenty of developable land onto which urban areas could expand, but this land is protected from development by strong policies (the Agricultural Land Reserve and the Green Zone), and these policies account for the high rate of intensification in Metro Vancouver. Calgary could have enacted similar policies, but has chosen to accommodate new residents largely through outward expansion and greenfield development.

Planning makes a difference and different plans produce different outcomes. Contrary to the common characterization of typical suburban development as "unplanned," the policy review shows that the urbanizing fringe of all three regions has long been highly regulated. Some researchers have argued that because cities compete for economic growth, they all tend to follow similar policies to attract new businesses and residents. This study has found that different cities have the autonomy and capacity to take distinctly different approaches to planning urban growth, and that these different approaches have shaped and channelled that growth in distinctive ways.

Continuity of policy and institutions are important. The two preceding findings must be qualified by a third. Planning policies to organize and direct urban growth are more likely to be effective if they are pursued over the long term and buttressed by supportive institutions and a sense of shared objectives. The Vancouver region has a long tradition of intermunicipal cooperation, and decades-old policies for protecting agricultural land, and its growth pattern reflects this consistency. Calgary has managed regional growth through annexation, avoiding the need for intermunicipal cooperation, and its growth pattern is equally consistent. In the Toronto region, by contrast, changes in policies and governance structures have had mixed and uneven growth outcomes, such as a few successful and many more unsuccessful nodes, and a wide variation in intensification rates among municipalities. Municipalities cannot create regional institutions and planning policies on their own. Senior governments -- Canadian provinces or U.S. states -- play a central role in shaping the institutional environments within which regional planning and politics operate.