Eco-urbanism is a very broad topic that can be approached from many perspectives and the terminology surrounding the subject is equally diverse. Green cities or green urbanism [1
], eco-cities [3
], sustainable cities [5
], resilient cities [6
], regenerative cities [7
], biophilic cities [8
], low-carbon or carbon-neutral cities [9
], and more recently, smart cities [10
] are all terms that find their way into the general rubric of eco-urbanism.
Notwithstanding the explosion of different nomenclatures in this field, there are core factors of eco-urbanism that are common to all, some of which are:
Compact, mixed land uses with elevated densities;
Multiple sub-centres throughout the city (polycentrism or de-centralised concentration);
Use of renewable energies;
Sustainably sourced building materials;
Local food production and food sourcing nearby;
Recycling and reuse of organic and non-organic materials;
Sustainable water management systems (water harvesting, on-site water retention, etc.);
Greening of urban surfaces using biophilic architecture;
Minimisation of high capacity road construction, especially freeways;
Development of extensive and safe pedestrian and cycling facilities;
Provision of superior public transport systems, especially rail;
Generous and usable green spaces throughout;
People-oriented, protected urban spaces, such as squares, pedestrian zones, playgrounds;
Urban design to encourage walkability.
Clearly, some of these factors are related directly or indirectly to urban mobility and the need to minimise car use in cities to help foster eco-urbanism. Each major period of transport evolution has brought with it different types of city forms and a dominant transport mode. First, there were the “Walking Cities” up to about 1850, which ran entirely on non-motorised modes (walking, animal-drawn vehicles, and, in some cases, water transport). This was followed by “Public Transport Cities” between about 1850 and 1950, during which time trams, trains, and finally buses were introduced into cities, allowing cities to greatly expand in physical size due to increased speed. The current stage in the transport-led evolution of cities is “Automobile Cities” [5
Across the history of human settlements, it can be argued that the automobile, more than any other vehicular technology, has been the most profound force in reshaping cities and the most destructive in its negative effects on the ecological characteristics of cities [11
]. Some authors have devoted much time and energy to envisioning cities free of an excessive dependence on cars and how to achieve that [14
I have suggested that creating eco-cities depends upon 10 key principles for transport and urban planning in the future, which are summarised in Figure 1
The centre of this simple model suggests that sustainable urban form and transport are at the heart of developing eco-cities. This includes four key factors: Compact and mixed use, efficient urban form; nature and food growing throughout and around the city; transit and non-motorised infrastructure, not freeways; and human city centres and sub-centres for population and job growth. Achieving sustainable operation of the urban planning and transport systems of cities facilitates the achievement of so many other important goals. For example, if less land is consumed for roads and sprawl, local food production within and in the immediate hinterland of a city becomes more feasible. Denser, more clustered building forms are easier to make energy-efficient due to common walls, which result in less exposed building surfaces to dissipate the heating or cooling in individual dwellings. In addition, the heating and cooling in each dwelling are shared through internal walls and floors. Recycling systems are easier to operate due to lower waste collection costs and through residents having deposit points for glass and other materials nearer to home. If fewer roads and less parking are needed, this expands possibilities for more permeable instead of sealed surfaces for better water management.
Around these four key principles are a constellation of another four supportive principles: Environmental technologies for energy, food, and waste; economic innovation through creativity and environmental quality; a high quality public realm throughout the city; and sustainable urban design qualities for human needs.
To achieve these eight principles, it is suggested that two processes are essential: Planning is visionary, “debate and decide”, not “predict and provide”, and decision making is within an integrated sustainability framework involving, social, economic, environmental, and cultural factors.
This paper focuses on the core of Figure 1
–how sustainable are the urban forms and transport systems of cities? However, it also touches on key aspects of most of the other factors. Comparing cities to better understand their transport and land use patterns, to gain insights into their strengths and weaknesses, and to deliver policies about how to reduce automobile dependence and improve transport sustainability has a long history in the academic literature [5
Scandinavia is often seen as being rather progressive in many fields, with strong commitments to innovative and green city policies. For example, Copenhagen is known world-wide as being an exemplary city for cycling and trying to reduce car use [23
], as well as a focal point of Jan Gehl’s urban design work on creating people-centred cities [24
Stockholm was, in the 1950s, a magnet to planners worldwide who came to see the ground-breaking satellite town concept for places, such as Vällingby, which developed the region based around the new tunnelbana metro rail system. Cervero [26
] describes how Stockholm transformed itself into a post-war transit metropolis. More recently, Malmö, since the opening of the Øresund bridge, has expanded greatly in its bicycle orientation and quality, perhaps due partly to the closer connection with and influence of Copenhagen [28
In 2015 Sweden also created K2–Sweden’s National Centre for Research and Education on Public Transport–with the aim of increasing the contribution of public transport to Sweden’s transport task. It is the interest of this organisation in better understanding how urban transport systems work in Sweden that made the research in this paper possible.
This research therefore builds on a long tradition of city comparisons, particularly on the theme of automobile dependence and adds five of Sweden’s most populous cities to the international comparative framework on cities developed by the author and colleagues over the last 40 years. It analyses urban transport and related indicators for the year of 2015 for Stockholm (population 2,231,439), Göteborg (population 982,360), Malmö (population 695,430), Linköping (population 152,966) and Helsingborg (population 137,909). It also compares the characteristics of these Swedish cities to a large sample of other cities in the USA, Canada, Australia, Europe, and Asia. The questions it seeks to answer are:
How do Swedish cities compare in land use, wealth, private, public, and non-motorised transport, as well as other transport-related factors?
Do Swedish cities follow the patterns of European and other cities or are they different?
Are there any policy lessons that can be learned from the comparisons?
To answer these questions, the paper systematically presents the results of the investigation through a series of indicators and discusses each one in turn. It analyses and discusses the results and draws some policy implications. The next section provides the methodology for the paper, followed by the results, then the analysis, discussion, and policy implications, and finally some conclusions.
Each Swedish city is defined in Table 1
in the way Statistics Sweden defines them, while Table 2
sets out the 35 primary data variables collected for each city and a brief description of what each variable means. Readers can also refer to Kenworthy [29
] for a more detailed explanation of the history and methodology of this comparative urban research. Note that for the international comparisons, the results pertain to the years of 2005–2006, but for Swedish cities, the year is 2015. These are the latest updates on this large sample of 41 other cities, which took approximately 7 years to complete.
The validity of the comparisons is mostly not compromised by this 10-year time difference. However, it is explained that with certain variables, such as the metropolitan Gross Domestic Product (GDP), transport deaths, and transport emissions, which can change quite rapidly, the time difference is more significant.
It is important to note that for some variables, the metropolitan area definitions given in Table 1
were modified. For example, metropolitan GDP needs to be calculated on the full functional urban region or “labour market area” (in German, the Arbeitsmarktregion). Also, some public transport services, such as the regional rail, cannot be separated out into smaller areas. In these cases, the population of the larger serviced area is used to calculate per capita figures to ensure that data are not inflated.
sets out a guide of what constitutes urban land. Statistics Sweden provides detailed land use inventories for every municipality and county in Sweden on their statistics portal at [30
Those categories used for urban land area in Sweden are called “built-up land and associated land”, which consist of:
Land with one- or two-dwelling buildings;
Land with multi-dwelling buildings;
Land used for manufacturing industry;
Land used for commercial activities and services;
Land used for public services and public facilities and leisure;
Land used for transport infrastructure;
Land used for technical infrastructure;
Land with agricultural buildings and other buildings.
Careful investigations were made of these land use categories and especially the last one, but in each case, they were found to be comparable to what was used in other cities.
The methodology chosen for this paper to present the results is to compare the five Swedish cities in detail, with both tables and figures showing the values of each variable for these five cities, as well as an average for the five Swedish cities to facilitate comparisons to the other groups of cities. Only averages for the American, Australian, Canadian, and the other European cities, plus the two Asian cities are presented. This was primarily to condense the results of the analysis into a paper of an acceptable length. Readers can refer to [11
] for detailed graphs and tables showing the results for variables on all the other cities.
The metropolitan regions used in this study were: The USA: Atlanta, Chicago, Denver, Houston, Los Angeles, New York, Phoenix, San Diego, San Francisco, and Washington; Canada: Calgary, Montreal, Ottawa, Toronto, and Vancouver; Australia: Brisbane, Melbourne, Perth, and Sydney; Europe: Berlin, Bern, Brussels, Copenhagen, Düsseldorf, Frankfurt, Geneva, Graz, Hamburg, Helsinki, London, Madrid, Manchester, Munich, Oslo, Prague, Stockholm, Stuttgart, Vienna, and Zurich; and Asia: Hong Kong and Singapore.
Urban density is a recurring and unifying theme throughout the paper, with the Swedish cities being compared to the highly auto-dependent metropolitan regions in North America and Australia. Some basic statistical regression analysis using a power function as the line of best fit is used in the discussion section to highlight the somewhat unique cluster of the five Swedish cities on the key two variables of urban density and car use and to use this regression as a predictor of car use for the Swedish cities compared to their actual results. This follows similar regression analyses in many of our other publications [5
The study has several limitations. For example, it only compares cities from an aggregate perspective and only for the year of 2015 in the case of the Swedish cities and 2005 or 2006 for the other cities. This impacts some variables more than others and this is explained in the text where relevant. No trends of the data are included, which would have been useful, but which were not available for the Swedish cities due to this being the first time they were included in such comparisons (except for Stockholm) as well as limitations on time and the available funding. For the other cities, 1995 or 1996 and 2005 or 2006 data were consistently available and some analyses of trends have already been made [32
The data perspectives in this paper are not the only ways that Swedish cities can be viewed in relation to each other and to other cities, and this limitation is partly addressed in the Analysis, Discussion, and Policy Implications section of the paper. Some variables that are included have limitations too, for example, the freeway length. This should ideally be the lane length to indicate capacity. However, even in an age of Geographic Information Systems (GIS) systems, this is an extremely hard, if not impossible, variable to collect consistently across such a large global sample of cities. It was the preferred variable when measuring freeways, but had to be dropped.
4. Analysis, Discussion, and Policy Implications
The first and second research questions were: How do Swedish cities compare in land use, wealth, private, public, and non-motorised transport, as well as other transport-related factors? Do Swedish cities follow the patterns of European and other cities or are they different?
The data here suggest that as a group, Swedish cities are somewhat unique and distinguish themselves in particular from other European cities in a number ways. Car use per person is almost identical in car passenger kilometres per person compared to other European cities. Use of cars in Swedish cities is much lower than in the three auto-dependent regions of the USA, Australia, and Canada. This can be seen in the analysis of urban density versus per capita car use. Figure 20
shows the urban density of the global sample of cities regressed with the annual car use per capita (passenger kilometres) using a power function as the line of best fit (see Methodology section). The five Swedish cities are shown in red. It can be clearly seen that the Swedish cities form an outlier, achieving lower car use than is typical in this global sample of wealthy cities at equivalent lower densities.
Fitting the power equation for the line of best fit (r2
= 0.75) to the average Swedish urban density of 19.8 persons per ha, we obtained a predicted annual PKT per capita of 11,127 km. However, the actual average is 6751 PKT per capita, or 39% lower than would be typical for this density. The performance of public transport, walking, and cycling in the mobility patterns of Swedish urban residents apparently keeps their car use to atypically low values for such low densities and overcomes the auto-dependence inducing effects of high road and freeway provision. The ‘transit leverage effect’ [42
]., where one passenger kilometre on public transport replaces multiple passenger kilometres by car, is likely having some effect here [5
]. This substitution of car passenger kilometres appears to be primarily related to the trip-chaining behaviour of public transport users for trip purposes that would be otherwise made in individual car journeys.
The theory of urban fabrics possibly also helps to explain some of the results [12
]. This theory shows that cities are all made up of different compositions and proportions of walking city fabrics, public transport city fabrics, and automobile city fabrics, the first two of which have much more sustainable mobility patterns. Walking cities, public transport cities, and automobile cities are depicted graphically in Newman and Kenworthy [5
While Swedish cities do have lower overall densities than is typical of European cities, they also still have significant areas of urban fabric that are either walking or public transport in their orientation, with significantly higher urban densities and mixed land uses. In these areas, public transport is more effective and walking and cycling are more frequently used due to the shorter travel distances required.
There are other ways of investigating in more detail the differences in urban fabric typologies within individual cities, which may further explain why some areas achieve better outcomes in sustainable transport. A particularly useful approach is to use street-based metrics to characterise urban typologies [43
]. This approach can distinguish, for example, differences in vegetation in different areas of cities, which can in turn have a significant effect on the attractiveness of neighbourhoods for walking and cycling. It is also useful in depicting in a quantitative way the relationship between streets and buildings.
Two important factors in urban design regarding streets and buildings is the level of horizontal and vertical enclosure, which affects the walking and cycling attractiveness of neighbourhoods, and even the propensity or willingness of people to access public transport stops on foot or by bike. Street metrics can help to reveal this in a systematic and quantitative way. In areas where streets are too wide relative to the height of buildings (bad vertical enclosure), poor walking environments are generally evident. In areas where streets have significant holes in the urban fabric due to, for example, surface parking lots (poor horizontal enclosure), they also detract from walking and cycling [11
Therefore, urban density per se, while fundamental in determining transport patterns, has numerous mediating factors, such as building heights and volumes, that help determine the experiential quality of both higher and lower densities. In the case of this research on a broad scale comparisons of cities, it was not possible to compare cities on such a detailed basis. However, further research on street metrics might reveal additional supportive reasons why these five Swedish cities achieve an overall lower car use at relatively low densities.
Given the above matters, it is important for readers to have at least some visual appreciation of the character of Swedish urban development and transport systems, which the numbers in this paper can only partially convey. Figures 21 to 29 show key examples of the differences in the urban fabrics for the Stockholm and Malmö metropolitan areas. Figure 21
shows a dense, traditional European five to ten-storey apartment development in Malmö, but with lower density neighbourhoods in the background. It is likely that these denser kinds of areas, which exist in all the Swedish cities to greater or lesser extents, are critical in explaining the relatively high rates of walking and cycling, as well as bolstering public transport use.
reveals the high-quality walking and cycling environments that can occur in such dense areas. A common theme in Swedish urban environments is to link urban development with a relatively good walking and cycling infrastructure. In Stockholm’s satellite towns, such as Vällingby and Kista, pedestrians and cyclists are physically separated from traffic and it is possible to travel between these satellite centres by bike or foot, sometimes through areas of forest or farmland.
shows the new West Harbour development in Malmö, which is based strongly around walking and a very good cycleway system. Such redevelopment of brownfield sites, for example, former dockland or industrial areas in cities, is often important in demonstrating new, more sustainable ways of urban development [44
]. A significant problem for Malmö’s West Harbour is that it is only served by buses and therefore lacks a clear centre or anchor point based around a rail station to give it more focus and sense of place. Much of the land, which is still vacant and awaiting build-out, is used for parking. This tends to set auto-oriented mobility habits somewhat early, rather than engendering a more positive public transport-oriented mobility culture from the outset.
depicts a much more car-oriented neighbourhood in Lund, which is only served by buses. It does, however, have rather good off-road cycling facilities, especially through farmland and other green areas, which are conducive to safe and attractive cycling that in turn helps to minimise car dependence. Cycling is, however, limited to a degree by a lack of mixed land uses near housing, which is in turn linked to low density. Without a critical mass of population to frequent local amenities, such as shops, medical facilities, and so on, it is not viable to provide them. Based on the author’s own observations and cycling in this area, there is, nonetheless, significant recreational cycling, which, among other things, has health advantages.
illustrates how the Stockholm region has a mixture of very high density core areas around tunnelbana stations, which then taper off into much lower density, suburban-style housing areas. One of the original planning ideas for such sub-centres was to achieve a reasonable jobs-housing balance, but in practice, people who live in one sub-centre work somewhere else. The transport problems of this reality are, however, mitigated to a high degree by the excellent way these sub-centres are served by the tunnelbana to the city centre and other sub-centres along the lines, as well as the feeder buses that bring people from lower density areas to the tunnelbana stations [26
and Figure 27
show the dense central and inner areas of Stockholm with excellent public transport and walking conditions. Public transport here includes buses, trams, and tunnelbana, as well as regional rail services. The urban fabrics in these areas were created in the walking and public transport city eras and are thus ideally suited to these more sustainable modes. Stockholm has also pedestrianised significant parts of its central area, which makes it very conducive to walking (and cycling), though the winters make these modes much less attractive in colder months.
and Figure 29
show the public transport orientation and walkability of both older satellite towns in Stockholm (Vällingby) and newer neighbourhood developments, such as Hammarby Sjöstad, which is serviced by a new tram line.
In the first case in Figure 28
, the new centre is formed linearly into a transit-oriented development (TOD) due to the close spacing of the tram stops. Throughout the neighbourhood there are excellent conditions for walking and cycling, though shopping and other facilities can be some distance away for some residents of the area due to the less than ideal mixing of land uses.
shows a nodal form of TOD due to the much wider spacing of stations along the tunnelbana lines. These sub-centres are separated in some cases by open land with urban development being restricted to well-defined areas. They have a strong profile of increasing density towards the station, both for residential and non-residential development. They are also very well-served by extensive feeder buses, often having large bus interchanges for many different routes located at the stations. Within about 400 metres of these stations, the environments are generally well designed to support pedestrians.
Overall, the urban comparative data on these five Swedish cities suggest that they form an interesting and somewhat unique “cluster” of cities on a global scale, with at times some paradoxical results.
On the one hand, Swedish cities are atypically low in density, and very high in roads and freeways compared to most European cities. However, employment is still rather centralised, with 16.3% of jobs in their CBDs compared to 18.3% in Europe. Swedish cities are the second highest in this factor across the city groupings. This generally works in favour of public transport, at least for work-related travel. Additionally, for trips to the CBD, parking is comparatively limited in Swedish cities with only 246 spaces per 1000 jobs, theoretically meaning that only about one in four people working in the CBD would be able to park a car. It is about the same in other European CBDs (248 spaces per 1000 jobs). This also favours public transport, walking, and cycling access to Swedish city centres.
Other factors working in favour of sustainable transport and eco-urbanism include the fact that Swedish cities have significantly lower car ownership levels than might be expected, lower even than other European cities. Average wealth levels as measured by metropolitan GDP are below typical European levels (though comparable to Australian and Canadian cities in 2006). Despite low densities, Swedish cities have developed relatively well-performing and more extensive public transport systems than many comparable lower density cities.
There seems to be a generalised European cultural factor at work here, such that in Europe, public transport is more accepted and utilised across a wide range of incomes [11
]. Swedish cities have generous amounts of public transport lines compared to other cities and the highest amount of reserved public transport route per person in the global sample. Unfortunately, this is offset by the high per capita provision of urban freeways, which leads to a relatively low ratio of reserved public transport to freeways in Swedish cities, thus somewhat offsetting the advantages of their public transport systems. Service provision as measured by seat kilometres is second only to the other European cities, but only by a small margin (5% lower). The average operating speeds for public transport in Sweden seem to be higher than most other cities, which leads also to the best average ratio of public transport system speeds to general road traffic of all groups of cities (0.98, the next closest result being 0.88 in the other European cities).
However, despite some of the above favourable conditions for public transport, Swedish cities on average have much lower per capita public transport boardings than other European cities (roughly half), but at the same time, this is much better than in the more auto-dependent regions in the USA, Australia, and Canada, where densities are also low. This might be expected given the significantly lower density of Swedish cities compared to other European cities, but is somewhat unusual when compared to similarly lower density cities in the USA, Australia, and Canada.
On the positive side, public transport use measured by per capita passenger kilometres is closer to European levels due to the longer distances travelled by public transport in Swedish cities, which again probably relates to low densities. Using total motorised passenger kilometres as a measure, these five Swedish cities on average have a healthy 20.4% of total motorised passenger kilometres on public transport, beaten only by their European neighbours (24.5%) and of course the Asian cities (62.9%). The modal split of daily trips is also just under 50% for public transport, walking, and cycling combined, meaning that the modal share in these five larger Swedish urban regions is pivoted rather equitably between the more sustainable and less sustainable modes. As shown in Figure 21
, Figure 22
, Figure 23
, Figure 24
, Figure 25
, Figure 26
, Figure 27
, Figure 28
and Figure 29
, Swedish cities have significant areas of urban fabric that are supportive of non-motorised modes and where walking and cycling is high, leading to over 27.1% of daily trips in Swedish cities being completed by these modes, despite a very cold climate. Only the other European cities have more with 34.5%. There are numerous areas in Swedish cities where bicycling is supported with reasonable infrastructure and walking is well catered for too in many areas.
Three key areas of concern for eco-urbanism also perform well in these Swedish cities. Energy use in private passenger transport is commensurate with the other European cities and much lower than in the auto-cities of North America and Australia. The Swedish cities excel in their extremely low transport emissions per capita and per hectare compared to every other region in the world and even the worst Swedish cities are better than the best of the others. Of course, the other data are from 2005–6 so it is expected that the other cities would have been closer to the Swedish cities in 2015, given the large advances in automotive technology and tougher air pollution regulations. Likewise, in transport fatalities per 100,000 persons, Swedish cities are the lowest in the world, and possibly would remain so due to Sweden’s Vision Zero transport deaths policy [45
]., even if the other cities were to be updated to 2015.
The third asked research question was: Are there any policy lessons that can be learned from the comparisons? Based on the data here, urban and transport policy for enhanced transport sustainability and better eco-urbanism in Swedish cities could focus on three key areas.
The cities would do well to focus on targeted increases in higher density, mixed use development that is especially linked to expanded and improved public transport. Stockholm is by far the best of the Swedish cities in sustainable transport and although it is still a low-density region, it is bound together by strong and now diverse urban rail networks (light rail, metro, and regional rail), around which very high density, mixed use development has occurred in strong centres throughout the region. Such an approach could be emulated in other Swedish cities.
There is also a need to restrict further development of already abundant freeway and road systems in all Swedish cities and perhaps to consider some focused removal of freeways, especially to help green the cities and improve public spaces. The thought of removing freeways is a radical one and is usually met with claims that the existing traffic will simply flow over everywhere else and cause chaos. However, it has been shown repeatedly that traffic behaves more like a gas than a liquid, adjusting its volume according to the size of the “container” provided (in this case, road capacity) [11
]. It is thus important in such endeavours to eliminate freeways to provide demonstrative projects of “disappearing traffic” so that other cities can learn from these experiences and feel more confident that such an approach can work. There are numerous examples of freeways that have been successfully removed. The Seattle Urban Mobility Plan in 2008 [47
] provides a catalogue of case studies of these projects, such as the Embarcadero and Central Freeways in San Francisco, California and the Harbor Drive Freeway in Portland, Oregon.
However, one of the best examples to date (also included in [47
].) is the Cheonggyecheon restoration project in Seoul (Figure 30
), which removed 6 km of elevated freeway and surface road carrying collectively 170,000 vehicles per day and created a green river boulevard in their place. The average traffic speed improved in Seoul after the removal of the freeway [11
]. This project stands as a landmark of both how traffic adjusts to a reduction in high capacity roads and the sheer attractiveness and beauty of the green environments that can replace such traffic corridors.
Swedish cities have an over-reliance on bus systems and a need for more extensive urban rail networks, which are critical in developing competitive public transport and achieving a wide range of other environmental, economic, and social goals [48
]. A major difference between Swedish and European cities generally is that European cities have three times higher rail use and this is a critical distinguishing feature in the lower public transport use in Swedish cities compared to other European cities.