Density, size, dispersion: towards understanding the structural dynamics of mid-size cities.
This paper draws on a variety of evidence--distance-decay density gradients; case study of one particular mid-size metropolitan area, Kitchener, Ontario; assessment of mid-size city core areas North America-wide by planners and other professionals who are familiar with this category of cities; nation-wide place ratings--to develop a conceptual model of mid-size city structure. This model has four main components: a flat-line, low-density profile; easy automotive access between dispersed activities; depleted core areas; and a sense of place that celebrates rural-like landscapes and convenience. The paper is designed to explain how fully dispersed styles of urban form have grown to characterize the majority of mid-size urban areas across Canada and the U.S. It also exposes the policy vacuum that exists as regards mid-size planning models and provides direction for policy initiatives designed with this category of city in mind.
Keywords: depleted core area, low-density development, land use-transportation dynamic
Afin de creer un modele representant la structure urbaine des villes moyennes, cet article a recours a differents types de donnees : gradients de densite, une etude de cas de la region metropolitaine de Kitchener, les resultats d'une etude nord-americaine sur la perception des centres-villes de villes moyennes par les urbanistes et autres experts ayant une connaissance de cette categorie de villes, et classements de villes a l'echelle nationale. Ce module a quatre composantes : une faible densite; un acces facile par automobile aux differentes activites dispersees a l'interieur du perimetre urbain; des centres-villes en declin; et une perception de l'espace qui reflete la commodite et un fort attachement aux paysages ruraux. Cet article entend expliquer comment l'evolution des villes moyennes au Canada et aux Etats-Unis a resulte presque partout en une forme urbaine caracterisee par un haut niveau de dispersion. Il identifie aussi une absence de politiques urbaines adaptees aux circonstances qui confrontent cette categorie de villes.
Mots cles: secteurs centraux en declin, urbanisation de faible densite, rapports entre usages du sol et transports
Almost one quarter of the population in North America lives in urban areas that range roughly from 50,000 to 500,000 persons in size (1). Yet, urban places in this mid-size range have received little attention in either the research or policy literature. Survey of the Journal of the American Planning Association over the ten-year period, 1994-2004, revealed, for example, that 87 percent of the journal's urban-content articles concerned places with populations of 1 million or greater while a meagre seven percent considered cities in our mid-size range. In singling out mid-size urban areas, we do not mean to imply that size per se is of singular importance over other features that influence urban form. We simply believe that the size dimension should be more fully investigated.
The focus on mid-size city form is timely because issues related to urban density have been of increased policy concern in cities of all sizes. Strategies targeted at intensification, re-urbanization, new urbanism and 'smart growth' have considerably elevated the need for better understanding of interrelationships between important functional parameters such as travel and land consumption and formal arrangements of density and land use. This paper extends academic concerns with land use and transportation arrangements specifically to the mid-size category of cities. Its findings suggest that trends towards low-density profiles and dispersed land use and travel patterns are characteristic of the majority of North America's mid-size cities. This further suggests that initiatives such as 'smart growth' will probably confront circumstances that are different in mid-size communities from those in large metropolitan areas.
The research summarized here is inductive having grown out of the study of one particular mid-size Canadian metropolitan area, Kitchener CMA. This paper elaborates on the Kitchener case study by presenting subsequent, more broadly-based analysis of other Canadian and US cities. On the basis of statistical trends observed in cities across Canada and the US, this paper first identifies the absence of core area density alongside overall low-density profiles as generic features of most, though not all, mid-size urban areas. The second part of the paper examines how low-density arrangements work on the ground as observed in the Kitchener metropolitan area; it also looks more closely at core areas in mid-size cities across North America, alongside place-based ratings that highlight the socially-constructed sense of place that surrounds the mid-size city dynamic. The third section presents a conceptual model built on these observations and the final discussion relates to policy issues.
DENSITY, SIZE, DISPERSION
A considerable literature has grown around urban population density gradients (e.g. Anas 1978; Bourne 1989; Brueckner 1980; Goldberg and Mercer 1986; Newman and Kenworthy 1989; Palumbo, Sacks and Wasylenko 1990; Skaburskis 1989; Smith 1997). The groundwork (e.g. Clark 1951; Newling 1969) documented systematic relationships between residential population density and distance from the city centre which took the form of a distance-decay curve. Negative exponential statistical models usually captured the interrelationship best. However, as urban areas suburbanized and spread outward in the second half of the 20th century, traditional patterns shifted. More recent reviews of density gradient research (McDonald 1989; Papageorgiou 1989; Richardson 1988) agree about several salient trends: 1) that a centralized distance-decay gradient characterized most North American cities over most of the 19th and 20th centuries; 2) that density gradients dropped considerably in the post-WW II period; 3) that there is great variation from city to city that is related to variable features such as city size, age, municipal political structure and economic base; and 4) that mixed use nodes and higher density communities in outer parts of some larger metropolitan areas have complicated, if not altered, the fundamental parameters of city structure which, in such cases, can only be captured by more complex, non-linear models (e.g. Anderson 1985; Chen 1997; Stern 1994). A Canadian-based study that used density gradients to investigate urban form was undertaken in the late 1980s (Bourne 1987; 1989). At that time it was concluded that, despite sustained decentralization, centralized patterns remained dominant across Canadian metropolitan areas. More recent research (Bunting 2004; Bunting, Filion and Priston 2002) corroborates that most of Canada's largest metropolitan areas remain centralized and that some may even be re-centralizing. Present findings regarding mid-size metropolitan areas suggest the reverse however, calling earlier conclusions to question.
Trends towards a low-density, de-concentrating urban form, variously referred to as decentralization, sprawl and dispersion are driven by two major types of population change. The first is population loss and density decline in built-up areas, particularly in the central-most zones. The second is ongoing peripheral growth in the form of relatively low-density, 'suburban'-style development. (Reasons for decreasing density that entail changing technology, taste and mobility are well understood and not reviewed here--see, for example, Borchert 1991; Bunting and Filion 1999; Gillham 2002; Gottdiener 1994; Kling, Olin and Poster, 1991; Knox 1994; Marshall, 2000; Rowe, 1991). In the case of the mid-size city, core densities were never very high. Low density profiles accrue when there is no marked distinction between innermost and outer metropolitan parts.
Mid-Size Cities: Low Density Profiles
Figure 1 shows the 2001 metropolitan-wide, distance-decay density profiles for Canada's three largest CMAs, Toronto, Montreal, and Vancouver, and the three largest mid-size metropolitan areas (with the exception of Kitchener CMA, considered later), London, Halifax, and St. Catharines-Niagara. In the largest CMAs mean densities are considerably higher than in the medium-size metropolitan regions; core/inner city densities peak before dropping off; and the relatively steep slope at the centre tapers off in outer, suburban parts of the metropolitan envelope. Although usually undetected in curve-fitting exercises, (see Millward and Bunting, in press) large cities also often post sub-peaks where secondary centres have been engulfed or where densification has taken place along major transportation arterials. Either way in comparison to large metropolitan areas, mid-size metros appear as veritable flat-liners whose core parts barely approach densities found in suburban parts of larger places.
Table 1 summarizes test statistics that allow us to assess the goodness-of-fit of the negative exponential, distance-decay model across the largest Canadian CMAs, bounded here at 400 [ppkm.sup.2], at the approximate edge of the built-up urban area (2). Although goodness-of-fit ([R.sup.2]) cannot be expected to vary simply as a function of size, we observe, as expected, that there are statistically significant differences in the average densities at which people live in large versus mid-size places; core (intercept) densities are also significantly different (3). There remains, of course, a considerable degree of variability between cities like Halifax and Victoria whose traditional, centralized density gradients contrast with cities like Kitchener, Trois Rivieres, Moncton or Regina. Despite this, we can conclude that most midsize metropolitan areas in Canada are characterized by reduced densities at the core alongside overall flat, low-density profiles.
[FIGURE 1 OMITTED]
Size, Density and Travel Mode
We broaden our perspective by turning to US metropolitan statistical areas (MSAs) with a population of 100,000 or greater in the 2000 census year (N = 256). When we regress mean density against metropolitan size for this larger set of places, as in Figure 2, the exponential [R.sup.2], at 0.5117, is relatively high given the gross nature of mean density as an indicator of urban form alongside our lack of control for other important factors known to affect density profiles. The mean density for the largest places (population> 1 million) is 544 persons per sq. mi. (210 persons per [km.sup.2]); 240 persons per sq. mi. (93 persons per [km.sup.2]) for places with populations between 500,001 and 1,000,000; and 143 persons per sq. mi. (55 persons per [km.sup.2]) for those we have defined as 'mid-size'. Variance within these categories is considerable, especially so in the largest category of places. The standard deviation for the largest MSAs is 347; for places less than one million but greater than 500,000, it is 261. In contrast, variance for the mid-size category of metropolitan areas drops to 108, suggesting that there is more consistency in density profiles across mid-size metropolitan regions. The pivotal role that density plays in the evolution of urban form becomes even clearer when we turn to examine intra-urban travel.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Figure 3 presents findings about relationships found between city size and journey-to-work travel mode. The weighted average modal split for very large places registers 7.8% (4.1%, unweighted) for transit against 1.1% in mid-size urban areas, reflecting the well-known fact that only the very largest and most dense metropolitan areas are able to sustain effective rapid transit and hence enjoy relatively large numbers of choice transit users. Auto-based travel which is best suited to lower density arrangements is less susceptible to the influence of city size because of its all-pervasive popularity ([R.sup.2] for transit is 0.376; for auto it is 0.108). But here too there is a significant difference. In this instance, lower density cities in our mid-size range register greater auto use than larger places.
DISPERSION AND CORE AREA DECLINE: MID-SIZE CASE
Increasing concern has been registered about the meaningless nature of universal relationships--such nation-wide population density profiles, US or Canadian--without supplementary 'local' knowledge and detailed observation about 'real world' places (e.g. Barnes 2004). Accordingly, to better understand the mid-size case, we examine other kinds of locality-based knowledge in an attempt to flesh out the distance-decay models. This more detailed examination sharpens our focus on the critical nature of density at the core.
Kitchener CMA (pop. 2001, 414,284) is a metropolitan region, located roughly one hundred kilometers west of Toronto that has been the focus of previous research on dispersed city structure (Bunting and Filion 1999; Filion, Bunting and Curtis 1996; Filion Bunting and Warriner 1999). In Kitchener CMA accentuated rates of suburbanization in the post-WW II decades were driven by the region's heavily industrial economy, by its lack of centralized amenity landscape features and by the suburban locus of major, non-industrial employers such as the two universities in the City of Waterloo. We have purposively chosen this regionally-amalgamated metropolitan area--compromised of three originally independent but coalesced municipalities, Kitchener (190,399), Waterloo (86,543) and Cambridge (110,372) and administered by two-tier, municipal and regional government--on the grounds that its recognizably exaggerated, multi-nodal structure facilitates understanding of patterns that would be overlooked in more conventional uni-centered CMAs such as St. John, London or Regina.
Figure 4 outlines the distance-decay density profile for Kitchener CMA as a whole alongside tabulated statistical summaries for the three major constituent municipalities (tabulations are based on regression trends for the built-up portions of the urban envelope to facilitate comparison with Table 1 earlier). Whether assessed at the level of each individual, mid-size municipality or the entire amalgamated metropolitan area, the distance-decay, [R.sup.2] values are notably low. Because of the region's multi-nodal structure, higher densities are in fact registered in discrete suburban locales in the Cities of Kitchener and Waterloo than at the core. We are not, however, claiming universality for Kitchener CMA as an exemplar of mid-size cities across North America. Rather, we use patterns observed in Kitchener to help us hone in on distinctive land use features that might characterize other similarly-sized places.
Over the last 50 years in Kitchener CMA the foremost land use planning problem has been core area decline. All three core areas, especially the largest and most central in the City of Kitchener, have lost most of their original business (i.e. retailing and employment) and, at the time of writing, suffered high vacancy rates. The main streets in Kitchener and Cambridge especially are marked with deteriorated buildings and other visible signs of blight. Numerous locally-based research and policy studies have documented this decline (Bunting and Filion 1999; Curtis 1996; Bunting et al. 2000; City of Kitchener 1998)which has persisted despite all kinds of public initiatives aimed at remediation. Furthermore, since most of the planning initiatives targeted at Kitchener's core have been documented to work well in many other places (Robertson 1995), revitalization practices per se cannot be held directly responsible for continued decline in core parts of the Kitchener CMA.
When we ask what else, if not strategic planning, might explain pervasive core-area decline in Kitchener, we are left with the conclusion that it must be a structural by-product of the larger region's overall low-density, decentralized, auto-oriented makeup. Extensive decentralization across the Kitchener region is indeed seen clearly in Figures 5 and 6 (4) which outline major employment concentrations and travel patterns. Both indicators exhibit spatial distributions that can best be described as spread out and dispersed. Table 2 supplements the maps by outlining some of the local features found in conjunction with dispersed land use arrangements and decentralized travel. Tne only parts of the Kitchener CMA not fully auto-accommodating are the original core areas. These areas have lost ground because they cannot offer the convenience and accessibility advantages characteristic of suburban locales. Under such conditions, the well-documented body of literature cited earlier explains that most households and business establishments seek out auto-friendly locations. Developers, who will only to be interested in building in a low-density, auto-oriented style, perpetuate the dispersed land use/transportation dynamic.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
Contrary to reason, we find that policy makers' continuing concern with pervasive core area decline is not shared by the public at large, at least not in our mid-size, Kitchener exemplar. Over recent years, there have been a number of studies confirming that the vast majority of residents of Kitchener CMA believe that they live in a 'good place' (e.g. PGM 2003, Environics 2005). Focus groups conducted in the Cities of Kitchener and Waterloo have found inhabitants to be predominantly attracted to suburban neighbourhoods and nearby rural, fringe areas (Bunting and Filion 1999). Their place predispositions are bolstered by preferences for greenery, privacy, pastoral-like landscapes and convenience. On the other hand there appears also to be a strong aversion towards central parts of the urban region. A recent survey of Kitchener and Waterloo residents found that 66% of a random sample of 526 respondents claimed to avoid Downtown Kitchener for a variety of reasons that included: rundown appearances; lack of things to do; the presence of undesirable elements in the population; poor parking; and an inconvenient location (City of Kitchener 1998). Circumstances are similar, if somewhat attenuated, in the Region's two other smaller core areas where the downtowns have been dismissed by many as 'hopeless'.
The Kitchener case exposes a significant contradiction: serious decline at the core in the face of widespread public endorsement of the urban area as a 'good place to live'. We might explain away this contradiction by hypothesizing that preference for newness and open space represents a sentiment that in itself further propels core area decline and outward dispersion. However, more evidence is needed before we even tentatively propose this kind of conclusion.
Mid-size City Core Areas: North American Planners' Survey
Aside from the Kitchener case, there has been virtually no research linking core-area decline to metropolitan size. Some years ago, Broadway (1989; 1995) suggested that size might be closely associated with central-area decline in North American cities but he did not pursue the observation. 'Downtown' decline has been well documented in both larger (e.g. Abbott 1993; Beauregard 1986; Brooks and Young 1993; Carey 1988; Gillette 1985; Robertson 1995; 1997) and smaller places (Robertson 1999) but not in mid-size cities as such. To this end, in 2003 a web-based survey was sent to planners and related professionals in the 202 urban areas with populations ranging from 100,000 to 500,000 in the US and Canada (for further details, see Filion, Hoernig, Bunting, and Sands 2004). This survey asked respondents to rate 'downtowns' in their regional vicinity along a seven-point scale from very successful to very unsuccessful. Open-ended comments were also solicited. There was a 34.4% response rate to the survey.
Findings from this survey confirmed that the vast majority of the downtowns of mid-size metropolitan areas experience serious difficulties. Only nineteen metropolitan regions, less than ten per cent of all the places surveyed, were rated in the two 'successful' categories on the seven-point scale. All of these 'successful' metropolitan areas possessed at least one (and often a combination) of several distinctive attributes including: high levels of pedestrian activity; a university in, or close to, downtown; strong tourist or visitor orientation; a well-preserved historical district; attractive centralized natural features such as waterfronts; and a state capitol, provincial legislature or major hospital complex. These nineteen were the exceptions rather than the 'rule'. The remaining 90 percent of mid-size city core areas were deemed not to be so blessed. In the case of places rated six or seven, features of absolute decline were predominant. Strong terminology--"unsuccessful"; "unhealthy"; "unattractive" and "dying"--were descriptors used by respondents. The picture in between the two extremes on the scale was less distinct, but "stagnation" and "loss" were the overriding characteristics that were remarked upon. Among other things, core areas rated at levels, 3, 4, and 5, were described as: "empty"; "ignored"; "slowly dying"; and "unsafe".
Table 3 presents verbatim results (deleting specific place names or other identifiable attributes) from open-ended responses about cores rated as "unsuccessful" or "very unsuccessful". There appear to be four broad categories of responses relating to: 1) actual abandonment or disappearance of the core ("has no core area activity and no vision from political, business, or community leadership"); 2) identification of features in downtown held responsible for decline ("too many oversized buildings; no continuous streetscape; incursion of suburban style patterns; large vacant parcels"); 3) evaluations of revitalization efforts ("they have put too much effort into skyways and institutions ... and really haven't found their niche"); and finally, 4) metro-wide contextual factors perceived to be associated with decline ("suffers from a struggling economy that makes business reluctant to invest, a declining population, and there are too many urban communities which are too spread out"). It is interesting, we think, that overall little 'blame' is targeted at strategic planning efforts and that the greater body of responses describe the broader context in which decline is situated. The message seems clear. Across North America, planners are dismayed and voice negative sentiments about the state of downtowns in mid-size places. They sense that broader, contextual factors are responsible but do not entirely understand that this might represent a structural dynamic shared by most places in the mid-size category.
Mid-size Cities and Place-based Ratings
National and international rankings of cities have been used to inform policymakers, researchers and planners alike about the overall state of different cities' health as well as about their most valued assets (Pacione 2005). Such rankings also invite comparison. We are, nonetheless, cautious in presenting place-based ratings as evidence about cities of different sizes because of the unstandardized and journalistic nature of the source material. For present purposes, this dimension of mid-size city structure remains largely conjectural.
Over the last two decades, surveys undertaken by popular magazines, such as Fortune in the US and Macleans in Canada, have helped to identify those ingredients that make cities 'good' or 'not so good' as places to live or do business; citizens' opinions also provide a snapshot of trends and sentiments that underlie public decision-making. On either score, a good deal of consensus prevails about attributes of mid-size cities. In most of the listings we examined, for example, mid-size cities predominated as 'best places' (Levine 1988; Loeb 1989; Lanbich and Gannes 1989; Martin 1994; Kay 1994; Saporito 1994; Kirkpatrick 1999; CB Media Inc., 2003; and Spilner 2003). A wide range of mid-size cities have become known 'as nice places to live' and are viewed as supporting a quality of life which larger cities cannot provide by virtue of being perceived as plagued with well-publicized problems, such as congestion, pollution, crime and poverty (Barlyn 1995; Lanbich and Gannes 1989; Saporito 1994; Wahl 2000). In a recent ranking of 331 US urban areas, the four highest-ranked urban areas belonged to our mid-size category, as did eight of the ten most highly ranked places (Sperling and Sander 2004, 33).
Overall, residents of mid-size communities are reported to regard their cities as having a good mix of 'big-city cachet' and 'small-town comfort and convenience' (Demont 1995; Kay 1994). Many mid-sized cities are believed to be attractive because they possess affordable housing and a lower cost of living than larger cities (Loeb 1989; Lanbich and Gannes 1989). The relative absence of traffic congestion and short commutes (by auto) make them easy and convenient. Cities in the midsize size range are considered ideal places to work and raise a family because of being safe, clean, having little crime activity, and a laid-back lifestyle (Kay 1994; Kirkpatrick 1999; Patterson 1996; Steinberg 1999). In addition, because of low-density arrangements and proximity to rural areas, mid-size cities are described as having an abundance of natural amenities (such as mountains, parks, riverfronts, and lakes) and recreational opportunities (Demont 1995; Kirkpatrick 1999; Lanbich and Gannes 1989; Loeb 1989; Martin 1994; Simard and Simard 2005).
Size-related aspects of sense of place as described above seem to confirm observations, however soft, of Kitchener CMA as a 'good place'. They also explain two features that may be important to full understanding of the predominantly dispersed, mid-size city dynamic. First, mid-size city residents' predisposed preferences for natural amenities, private home ownership, and auto-oriented convenience serves to perpetuate low-density, dispersed arrangements. Second, overall high levels of satisfaction produce inertia and resistance to change.
The Structural Dynamic of Mid-size Cities
We have identified several salient features that seem to be characteristic of a large proportion of mid--size cities: an overall low-density profile and a distinct absence of central density; very good auto-based accessibility and poor transit; a lack of traditional centralization, manifest as core-area stagnation and decline. There may also be a collective 'place' preference predisposed to convenience, privacy and green-space and disinclined to central, older parts of the built-up city. A sequential development path is suggested in Figure 7.
[FIGURE 7 OMITTED]
In the pre-WW II period, the majority of places we classify as 'mid size' today were much smaller, so would not host large downtowns with substantial retail or employment concentrations. Moreover, because an urban growth boom commenced in the 1950s, the spatial 'fix' (Harvey 1973) of most mid-size urban areas grew to be predominantly 'suburban' in character. Population in the City of Kitchener for example was 35,000 in 1941 whereas in 1971 it was 112,000, over three times larger, so by the early seventies over two thirds of the urban fabric had been built in the post-war style. Because densities associated with this era of urban development tended to be relatively low, motorists were able to travel almost everywhere within the metropolitan region enjoying near blanket accessibility--except at the centre (the argument made by Gordon and Richardson 1996; 1997; Gordon, Richardson and Jun 1991 in support of poly-nucleated over centralized metropolitan form). Transit, on the other hand, would not work well.
Unless there were unusual historic circumstances, most employment nodes in mid-size metropolitan areas are found in places with good vehicular access, well outside of the core. Suburban centres and malls offering attractive conditions, convenience, free parking and good access have been well poised to outpace the core as the major source of retail goods and services. In the post-1970 era for example when Kitchener CMA attracted the head office of a large insurance company, the site chosen was at an expressway junction (the 'exit ramp economy'--Taylor and Lang 2004), not in the core. Likewise, fashionable retail boutiques that might otherwise be found in core parts are today found in a small town in the rural-urban fringe.
When rates of auto ownership are high and motor vehicle accessibility to most activities is good, suburban and exurban locales reap the benefit. Such conditions pertain not just to mid-size cities but also to expansive, outer parts of most large metropolitan areas across North America (Benfeld et al., 1999; Cevero and Wu 1997; Crane 1996; Downs 1994; Gillham 2002; Kling, Olin and Poster 1991; Marshall 2000; Nelson and Dueker 1994; Soja 2000). Distinctive to the midsize city, however, is the lack of activity at the centre. As a result, most mid-size places are unable to generate a self-reinforcing, inner-city, land use/transportation dynamic based on transit and pedestrian travel alongside concentrated, high-density arrangements. We suspect that reversing these conditions may prove challenging, especially when mid-size city residents have become strongly attached to the very attributes that bring about dispersion, as place-based research tentatively suggests.
The Policy Dilemma
Mid-size cities suffer from a scarcity of planning models suited to their particular circumstances, in part because there has been little recognition of their distinctive status. Senior government policies have also been directed primarily at larger urban places since serious problems most often have their origin there. The urban policy literature reflects this bias.
We would argue that the habit of planning for cities in the mid-size range with scaled-down versions of 'big city' solutions can lead to problems. For one thing, it may mean the adoption of policies that are more elaborate, costly and larger-scale than need be. Take, for example, the case of downtown revitalization strategies, originally designed for large North American cities and centered around infill malls, office towers, hotel and mixed-use developments (Frieden and Sagalyn1989; Robertson 1995). Unthinking application of generic schemes based on this type of model to mid-size places has resulted in a succession of initiatives that have faltered. Planners responding to our survey of mid-size core areas frequently mentioned vacated and/or run-down infill malls. In the City of Kitchener, there were two infill malls in the 1980s--one has subsequently converted to 'back' office premises of a suburban-based head office; the other houses a range of low-order functions--offices, a large gym, discount retailers. The lesson should be abundantly clear. Without the benefit of substantial core-oriented markets, malls and office towers in mid-size cities have proven to be no match for suburban competitors. Thus, despite decades of remedial work, downtowns in mid-size cities virtually everywhere suffer from chronic vacancies and an over-presence of low-end commercial units because the 'big-city' strategy failed.
Mid-size urban areas should not unquestioningly rely on scaled-down versions of policies formulated in large metros without careful analysis. Today for example, when major planning concerns centre on controlling urban expansion and reducing urban development costs, cities of all sizes are faced with proposals for intensification, re-urbanization, infill, growth boundaries and other sorts of initiatives associated with 'smart growth'. Mid-size cities should be wary. While all cities will confront huge obstacles in turning back outward growth to previously built-up parts of the urban envelope, the dynamic of the mid-size city may be especially inimical to intensification. Particularly in this instance, an important first step towards resolving how to plan more strategically will be to better understand the distinctive properties of mid-size urban areas.
Summary and Conclusions
In order to expose features distinctive to mid-size urban areas, this paper presented a variety of evidence ranging from 'hard' statistical models through to more subjective 'place' dimensions. We have tentatively argued that mid-size urban form creates a self-perpetuating, land use/transportation dynamic, characterized by core area depletion partnered with expansive styles of outward growth. What disturbs us most about dispersion in mid-size metropolitan areas is its entrenched nature because when all sites must be adjusted to widespread car use, the urban fabric becomes inhospitable to alternative modes of transportation and more concentrated forms of building. This is particularly concerning at present when more compact urban forms are widely held to be the path to a more sustainable future.
Abbott, C. 1993. Five downtown strategies: Policy discourse and downtown planning since 1945. Journal of Policy History 5: 5-27.
Anas, A. 1978. Dynamics of urban residential growth. Journal of Urban Economics 5: 66-87.
Anderson, J. E. 1985. The changing structure of a city: temporal changes in cubic spline urban density patterns. Journal of Regional Science 25:413-425.
Barlyn, S. 1995. Where to work: It's San Francisco! Fortune November 12: 84.
Barnes, T. 2004. A paper related to everything but more related to local things. Annals, Association of American Geographers 94: 278-83.
Beauregard, R. 1986. Urban form and the redevelopment of central business district. Journal of Architectural and Planning Research 3: 183-98.
Benfeld, F.F. et al. 1999. Once There Were Greenfields: How Urban Sprawl is Undermining America's Environment, Economy and Social Fabric. Washington: National Resources Defense Council.
Borchert, J. 1991. Futures of American cities. In Our Changing Cities, ed. J.F. Hart, 218-250. Baltimore, MD: Johns Hopkins University Press.
Bourne, L. S. 1989. Are new urban forms emerging? Empirical tests for Canadian urban areas. Canadian Geographer 33: 312-28.
--. 1987. Evaluating the aggregate spatial structure of Canadian metropolitan areas. Canadian Geographer 31: 194-220.
Broadway, M. 1995. The Canadian inner city, 1971-1991: Regeneration and decline. Canadian Journal of Urban Research 4: 1-19.
--. 1989. A comparison of patterns of deprivation in Canadian and US Cities. Social Indicators Research 21: 531-551.
Brooks, J., and A. Young. 1993. Revitalizing the central business district in the face of decline. Town Planning Review 64: 251-271.
Brueckner, J. 1980. A vintage model of urban growth. Journal of Urban Economics 8: 389-402.
Bunting, T. 2004. Decentralization or recentralization? A question of household vs population enumeration. Environment and Planning A 36: 127-47.
--, and P. Filion. 1999. Dispersed city form in Canada: A case of Kitchener CMA. Canadian Geographer 43: 268-287.
Bunting, T., and H. Millward. 1998. A tale of two CBDs: The decline and revival(?) of downtown retailing in Halifax and Kitchener. Canadian Journal of Urban Research 7:139-166.
--, P. Filion and H. Priston. 2002. Density gradients in Canadian metropolitan regions, 1971-1996: Differential patterns of central area and suburban growth and change. Urban Studies 39: 2531-52.
--, P. Filion, S. Frenette, D. Curry, and R. Mattice. 2000. Housing strategies for downtown revitalization in mid-size cities: A City of Kitchener profile. Canadian Journal of Urban Research 9: 145-175.
Carey, R.J. 1988. American downtowns: Past and present attempts at revitalization. Built Environment 14: 47-59.
CB Media, Ltd. 2003. Economic development: The location equation. Canadian Business 76: 84-89.
Cervero, R. and K-L. Wu. 1997. Polycentrism, commuting, and residential location in the San Francisco Bay Area. Environment and Planning A 29: 865-886.
--. 1998. Core Area Housing: A Feasibility Report. Kitchener: City of Kitchener, Department of Planning and Development.
Chen, H. 1997. Models of urban population and employment density. Geographical and Environmental Modelling 1: 135-151.
Clark, C. 1951. Urban population densities. Journal of the Royal Statistical Society 114: 490-96.
Crane, R. 1996. Cars and drivers in the new suburbs. Journal of the American Planning Association 62: 51-65.
Curtis, K. 1996. Revitalizing Kitchener's CBD: Public Planning Initiatives and Effects. In The Dynamics of the Dispersed City: Geographic and Planning Perspectives on Waterloo Region, ed. P. Filion and T. Bunting, 137-166. Waterloo: University of Waterloo, Department of Geography Publication Series No. 47.
Demont, J. 1995. The last best place: Halifax has rediscovered a lost youth to go with its sense of history. Maclean's (June 19) 22: 7.
Downs, A. 1994. New Visions for Metropolitan America. Washington, DC. Brookings Institute.
Environics. 2005. Kitchener: A View Toward the Future. Toronto: Environics.
Filion, P., T. Bunting, and K. Curtis, eds. 1996. The Dynamics of the Dispersed City: Geographic and Planning Perspectives on Waterloo Region. Waterloo: University of Waterloo, Department of Geography Publication Series No. 47.
--, T. Bunting, and K. Warriner. 1999. The changing balance between space and place: Towards an urban social mosaic. Urban Studies 36:1317-1347.
--, H. Hoernig, T. Bunting, and G. Sands. 2004. The successful few: Healthy downtowns of small metropolitan regions. Journal of the American Planning Association 70: 328-343.
Frieden, B., and L. Sagalyn. 1989. Downtown Inc. Cambridge: MIT Press.
Gillette, H. 1985. The evolution of the planned shopping center in suburb and city. Journal of the American Planning Association 51 : 449-61.
Gillham, O. 2002. The Limitless City: A Primer on the Urban Sprawl Debate. Washington, DC: Island Press.
Goldberg, M. A., and J. Mercer. 1986. The Myth of the North American City: Continentalism Challenged. Vancouver: University of British Columbia Press.
Gordon, P., and H. Richardson. 1996. Beyond polycentricity--the dispersed metropolis, Los Angeles, 1970-1990. Journal of the American Planning Association 62: 289-95.
Gordon, P., and H. Richardson. 1997. Are compact cities a desirable planning goal? Journal of the American Planning Association 63:95-106.
Gordon, P., H. Richardson and M. J. Jun. 1991. The commuting paradox: Evidence from the top twenty. Journal of the American Planning Association 57: 416-21.
Gottdiener, M. 1994. The Social Production of Urban Space. 2nd ed. Austin, TX: University of Texas Press.
Harvey, D. 1973. Social Justice and the City. London: Edward Arnold.
Kay, Linda. 1994. How women rank Canada's 25 biggest cities. Chatelaine 67(4): 49.
Kirkpatrick, David. 1989. Fortune's top ten cities. Fortune 120(9): 79.
Kling, R., S. Olin, and M. Poster, eds. 1991. Postsuburban America. Berkeley: University of California Press.
Knox, E 1994. Urbanization. Englewood Cliffs, N.J.: Prentice Hall.
Lanbich, K. and S. Gannes. 1989. The best cities for business. Fortune 129(9): 56.
Levine, R. 1988. City stress index: 25 best, 25 worst. Psychology Today 22(11): 52(7).
Loeb, Marshall. 1989. What's the best ole towns in the U.S.A.? Fortune 120(9): 8.
Marshall. A. 2000. How Cities Work: Suburbs, Sprawl and the Road not Taken. Austin: University of Texas Press.
Martin, J. 1994. The rest of the major cities. Fortune 130(10): 124.
McDonald, J. 1989. Econometric studies of urban population density: A survey. Journal of Urban Economics 26: 361-385.
Millward, H. and T. Bunting. in press. Spatial patterning in urban population densities: A temporal model compared with Toronto, 1971-2001. Environment and Planning A.
Nelson, A. and K. Dueker. 1994. The new burbs: The exurbs and their implications for planning policy. Journal of the American Planning Association 45: 459-65.
Newling, B. 1969. The spatial variation of urban population densities. Geographical Review 59: 242-252.
Newman, P., and J. Kenworthy. 1989. Cities and Automobile Dependency. Aldershot, UK: Gower.
Pacione, M. 2005. Urban Geography: A Global Perspective. London: Routledge.
Palumbo, G., S. Sacks, and M. Wasylenko. 1990. Population decentralization within metropolitan areas: 1970-80. Journal of Urban Economics 27:151-67.
Papageorgiou, Y. 1989. The exponential density function: first principles, comparative statistics, and empirical evidence. Journal of Urban Economics 26: 264-268.
Patterson, J. 1996. How clean are our cities. Chatelaine. 69 (5): 51.
PGM Consulting. 2003. Living in the City of Kitchener: Report of Findings. Kitchener: PGM Consulting, December 2003.
Richardson, H. 1988. Monocentric vs. polycentric models: the future of urban economics in regional science. Annals of Regional Science 22: 1-12.
Rowe, P. 1991. Making a Middle Landscape. Cambridge, MA: MIT Press.
Robertson, K. 1999. Can small city downtowns remain viable? Journal of the American Planning Association 65: 270-83.
--. 1997. Downtown retail revitalization: A review of American development strategies. Planning Perspectives 12:383-401.
--. 1995. Downtown redevelopment strategies in the United States: An end-of-the-century assessment. Journal of the American Planning Association 61: 429-37.
Saporito, B. 1994. The world's best cities for business. Fortune 130(10): 113-122.
Skaburskis, A. 1989. Inversions in urban density gradients: A brief look at the Vancouver metropolitan area density profile. Urban Studies 26: 397-401.
Simard, M., and C. Simard. 2005. Towards a culturalist city: A planning agenda for peripheral mid-size cities. Canadian Journal of Urban Research 14: 38-47.
Smith, B.E. 1997. A review of monocentric urban density analysis. Journal of Planning Literature 12:115-135.
Soja, E. W. 2000. Postmetropolis: Critical Studies of Cities and Regions. Oxford: Blackwell.
Sperling, B., and P. Sander. 2004. Cities Ranked and Rated. Hoboken, NJ: Wiley.
Spliner, M. 2003. Walking USA: Explore America's 10 best cities afoot. Prevention 55(4): 136.
Steinberg, S. 1999. Canada's safest cities. Chatelaine 72(4): 70.
Stern, D. 1994. Historical path dependence of the urban population density gradient. Annals of Regional Science 28:197-222.
Taylor, P. J., and R.E. Lang. 2004. The shock of the new. Environment and Planning A 36:951-8.
Wahl, A. 2000. A nice place to visit. Canadian Business 73 (5): 96.
Department of Geography and School of Planning
University of Waterloo
School of Planning
University of Waterloo
(1) We are not overly concerned about putting limits on what constitutes 'mid-size'. The range 50-500,000 is put forward as a general one that seems to us to capture the main body of places we would classify as falling within the mid-size range but the limits suggested here are broad ones, not meant to be interpreted with exactitude. Our empirical or operational definition of a mid-size urban area derives from Statistics Canada's definitions of urbanized regions, i.e. either metropolitan areas or census agglomerations. We thus exclude any lower tier municipalities that are a constituent part of a larger urban region.
(2) Figure 1, in contrast, encompasses the entire territory of the CMA, the more conventional (but less reliable because of the variability of the size of the rural-urban fringe from one metro to another) basis for distance-decay modeling.
(3) The [R.sup.2] and slope exponent formulations are slightly higher among medium-size than large CMAs because the negative exponential does not capture the more complicated, poly-centric relationships manifest in the largest CMAs.
(4) The 1996 data used to construct Figure 6 derive from the most recent source that provides origin-destination travel for all trips undertaken within the Region. Plots shown are restricted to results from a surveyed sample (5% of total population). Only flows that aggregate to 15 trips or more are plotted. This probably explains why there is a paucity of travel between Kitchener-Waterloo and Cambridge. It is anticipated that the current (2006) survey will show more travel connections to Cambridge.
Table 1: Regression Statistics, Population Density vs. Distance from CBD Negative Exponential Model, 2001 Canadian CMAs (Built up areas, 400 pp/km2) CMA Total CMA Mean Exponential Population Population Regression Density R2 Large CMAs Toronto 4,682,219 3,143 0.1899 Montreal 3,426,150 2,907 0.3437 Vancouver 1,986,965 2,509 0.1948 Ottawa-Hull 1,063,644 2,136 0.0549 Calgary 951,395 2,054 0.0184 Edmonton 937,691 2,187 0.0005 Quebec 682,757 1,514 0.3162 Winnipeg 671,274 1,995 0.1222 Hamilton 662,285 1,909 0.1072 Unweighted Mean 1,673,820 2,262 0.1497 Weighted Mean n/a 2,628 0.1924 Mid-sized CMAs London 432,451 1,904 0.0685 Kitchener 414,279 1,779 0.0394 St. Catharines- 377,009 1,558 0.0723 Niagara Halifax 359,168 1,494 0.1498 Victoria 311,902 1,547 0.6555 Windsor 307,877 1,664 0.0177 Oshawa 296,298 1,896 0.1015 Saskatoon 225,927 1,470 -0.0121 Regina 192,800 1,923 -0.0132 St. John's 172,918 1,394 0.0823 Sudbury 155,601 1,134 0.2811 St. John 122,663 1,078 0.2152 Thunder Bay 120,342 1,412 0.0078 Unweighted Mean 268,403 1,558 0.1281 Weighted Mean n/a 1,623 0.1281 All 22 CMAs Unweighted Mean 843,346 1,846 0.1370 Weighted Mean n/a 2,439 0.1803 Student's t n/a 0.0015 0.3709 probability * CMA Intercept Slope Exponent Large CMAs Toronto 7,604 -0.0365 Montreal 9,952 -0.0673 Vancouver 5,959 -0.0439 Ottawa-Hull 3,537 -0.0332 Calgary 2,927 -0.0225 Edmonton 2,657 -0.0147 Quebec 5,152 -0.1154 Winnipeg 3,812 -0.0696 Hamilton 4,027 -0.0672 Unweighted Mean 5,069 -0.0523 Weighted Mean 6,593 -0.0484 Mid-sized CMAs London 3,173 -0.0827 Kitchener 2,489 -0.0226 St. Catharines- 2,112 -0.0211 Niagara Halifax 2,836 -0.0666 Victoria 4,665 -0.1634 Windsor 2,406 -0.0466 Oshawa 3,206 -0.0895 Saskatoon 2,149 -0.0368 Regina 2,320 -0.0317 St. John's 2,602 -0.0812 Sudbury 1,774 -0.0876 St. John 2,923 -0.2840 Thunder Bay 2,048 -0.0502 Unweighted Mean 2,670 -0.0818 Weighted Mean 2,760 -0.0722 All 22 CMAs Unweighted Mean 3,652 -0.0697 Weighted Mean 5,872 -0.0529 Student's t 0.0089 0.1024 probability * * Test of difference of means of large vs. mid-sized CMAs Source: Statistics Canada, 2001 Table 2: Salient Features of Dispersion, Kitchener CMA In terms of Size and Period of Growth, in 1941, in the largest municipality, the City of Kitchener, population registered at 35,657; today, at 190,399, it contains more than five-times its pre-WW II population, with virtually all of the post-war growth being suburban in form. As early as 1971, most of the municipality's territorial jurisdiction, over two-thirds of the urban envelope, was 'suburban' in character. Core Area Employment though never high has continued to drop--in the City of Kitchener, it dropped from 18,394 in 1980 to 12,000 in 2003 Retail Vacancy Rates in the Core are high and vacant lots and empty storefronts predominate on the main street (Bunting and Millward 1998). In a little more than a decade after their much-heralded opening, downtown Kitchener's two large infill malls were closed to retail activity. Densities--over the 30-year period, 1971-2001, there has been an overall increment in mean density, but, this is produced by islands of increased densities in the inner suburbs (Figure 4) and at the edge of the urban envelope rather than at the core. Core density per se, never high, dropped off over the last 25 years. Figures for the City of Kitchener, are: 1971, 3,787; 1996, 3,011; 2001, 2,990 persons per [km.sup.2]. Densities in outer parts increased somewhat as a result of higher density development at select strategic locations (e.g. expressway interchanges) as well as planners' encouragement of increased suburban densities. In terms of travel patterns, journey-to-work commutes are on average short--5.3 km (average commuting distances in Toronto, Montreal, Vancouver and Ottawa are respectively 9.3, 8.2, 7.7 and 7.8 km); 62% of all auto trips are less than five kilometers (Statistics Canada, n.d.) Modal split in 1996, when the last major travel survey of Kitchener CMA was undertaken, registered eighty-seven percent of journeys done by car; walking, 6.6%; 3.5% for public transit. For journey-to-work in 2001, Statistics Canada reported that 3.9% of Kitchener CMA workers used transit (across all Canadian CMAs an average of 8.7% of metropolitan workers are transit users.) Table 3: Responses About 'Unsuccessful' Core Areas Overall Assessment/Characteristics of the Core "Dreadful.... very bleak. No sense of place"; "Very unsuccessful"; "dump, with abandoned buildings"; "... is one of the least healthy downtowns I have seen. Why I'm not sure", "Used to live there--dying downtown and no signs of willingness to do anything about it."; "I'm not impressed"; "Very unsuccessful"; "Having lived there, not impressed"; "An unmitigated disaster"; "snore"; "Very sad--great architecture, but the life has been sucked out of it."; "Down at the heels." "Has not kept pace with ... to say the least"; "Scary"; "... has 3 distinct and separate downtowns ... these are all very unsuccessful"; "..., despite being the capital and a larger city, is widely considered much less "liveable" than.... Downtown has very little to recommend it. After business hours, it is essentially shut down."; "... has declined catastrophically"; "Downtown has a concentration of bars, sex related businesses and everything else that goes along with that development model."; "poor pedestrian environment, lack of overall plan, fragmented and disjointed land use pattern"; "the downtown fails to gather its share of tourism traffic, overnight guests, and secondary spending. Too many oversized buildings; no continuous streetscape; incursion of suburban style patterns; large vacant parcels"; " Brutalism and one-way streets" Abandoned Core / Absence of Core "No real core area"; "... has no core area activity and no vision from the political, business, nor community leadership"; "Not well-defined"; "Abandoned main street"; "There is absolutely no core to these post WWII communities"; "Where is the ... city core ...?", "Can't buy a pair of socks ... except maybe at the buck store"; "... downtown retail virtually nonexistent"; "... Retail is mostly gone, leaving only a few restaurants", "No retail, too many vacant buildings, too many gaps in the streetscape"; "Not much left of what once was a very viable downtown", "Many vacant lots, and businesses. Lack of cultural activities. Lack of waterfront access and facilities"; "... derelict buildings; urban blight" Evaluation of Revitalization Efforts "... they have put too much effort into skyways and institutions ... and really haven't found their niche"; "Worse than.... same problems magnified by little effort to fight sprawl or redevelop downtown"; "No life in downtown, city has non motivated staff, no creativity"; "Obscured beachfront with commercial development; too uncontrolled"; "failed covered main street attempt." "They 'Paved over paradise and put in a Parking Lot.' If you ever want to see a town that incorporated every mistake in the book this is it. I use it as a classic example of 'What not to do!'" Metro-wide Contextual Factors "Depressing city, sprawl and big box stores with huge parking lots make for a horrible urban landscape"; "Too many shopping mails on outskirts of downtown"; "No sense of place, uncontrolled development"; "Retail presence is negatively affected by big boxes ... at ... Center", "Major regional suburban centers. Not sure where core is"; "Downtown ... was badly damaged, a generally declining economic profile.... but there is very little of cultural or recreational interest"; "... suffers from a struggling economy that makes business reluctant to invest, a declining population, and there are too many urban communities which are too spread out (e.g. ...)"; "... but declining region, aging population"; "Three cities, three little downtowns, big suburban malls"; "Total disaster, despite new retail center adjoining the city core and a new state office building, q-his city has not been able to define a new economic base and reason for existence since the ... industry collapsed 20 years ago. The city is currently in bankruptcy under the financial supervision of the state, and the mayor is in jail on morals charges"; "Downtown is dying as a result of malls and sprawl"; "An unattractive place that even the summer tourists avoid.., has little in the way of historic character, and is pretty devoid of much commercial/retail activity. It has turned itself largely away from its principal natural attraction, ..., and the few tourist-oriented activities in the city are scattered around the area rather than focused in downtown core.... It is not a walkable downtown, and tends to fraught with both pedestrian and vehicular hazards"; "The core area appears to have been discarded for suburban development purposes"; "The core area has been ignored, and discarded in favour of suburban development"; "Very strong suburban pressures, mobile homes and cheap apartments"; "long long main street, not pedestrian friendly"; "quite a few vacancies"; "Much more rural than cosmopolitan". Figure 4: Negative Exponential Regression Model for Population Density, Kitchener CMA, 1971 and 2001 Negative Exponential Model, 2001, Built up areas only (400 persons/k[m.sup.2]) Mean Exponential Total Population Regression Municipality Population Density [R.sup.2] Kitchener 176,897 1,908 0.1936 Waterloo 80,173 1,985 -0.0714 Cambridge 104,509 1,489 0.0423 Unweighted Mean 120,526 1,794 0.0548 Weighted Mean n/a 1,804 0.0911 Kitchener CMA 414,279 1,779 0.0394 Slope Municipality Intercept Exponent Kitchener 3,850 -0.1529 Waterloo 2,239 -0.0018 Cambridge 2,446 0.0830 Unweighted Mean 2,845 -0.0239 Weighted Mean 3,087 -0.0512 Kitchener CMA 2,489 -0.0226
|Printer friendly Cite/link Email Feedback|
|Author:||Bunting, Trudi; Filion, Pierre; Hoernig, Heidi; Seasons, Mark; Lederer, Jeff|
|Publication:||Canadian Journal of Urban Research|
|Date:||Dec 22, 2007|
|Previous Article:||Placing social policy? Reflections on Canada's New Deal for Cities and Communities.|
|Next Article:||Local urban form measures related to land-use and development period: a case-study for Halifax, Nova Scotia.|