1This paper examines some long-term changes of travel patterns generated in the urbanized and semi-urbanized areas of Sweden. The problem addressed is whether certain urban sizes and densities are more mobility dependent than others. Data from the Swedish national transportation surveys 1978 and 1996/97 are used to compare the overall travel activities of inhabitants in different types of areas. Key factors, such as daily distance travelled and transportation mode used, are considered.
2During last few decades environmental problems have brought questions of travel demand management to the fore. In short, if technological measures cannot bring the polluting emissions, the traffic noise, and the emission of CO2 down to acceptable levels, then we must try to reduce the sheer volumes of traffic. Ongoing discussions concern various measures to calm the need for travel such as information and marketing, prices and taxes, telecommunications (e.g., telecommuting) and urban land-use planning, and how they might influence the way people travel and carry out their everyday activities.
3There is general agreement that the complex issue of travel behaviour - deeply integrated in most of our daily activities - is affected by a multitude of factors, for instance supply and costs of transportation alternatives, incomes and values of the population as well as urban land-use and accessibility (e.g., Vlek & Michon 1992, Gärling et al 2002). However there is often a tendency to reduce the problem by focusing on one or two factors at a time, for instance the isolated role of urban land-use. Urban land-use then refers mainly to elements such as urban form and size, urban spread, population density, mixing of activities (e.g., housing/workplace-balances). Physical distances between activities are supposed to affect the need for travel and the actual travel behaviour.
4In the perspective of a comprehensive policy, a crucial question is whether an increased accessibility within urban areas could put stop to the seemingly ever-increasing mobility of the population (e.g., Mogridge 1986, Giuliano & Small 1993, Breheny 1995, Wegener 1996). Intuitively it seems reasonable to assume that if the physical distances and geographical separation between activities could be compressed, then the need for travel in everyday life would be cut back.
5During recent years there has been an intensified research concerning matters of transportation, accessibility and the sustainability of urban environments (for a review, see van der Waals 2000). Concepts such as "compact cities" and "containment strategies" have been launched emphasizing the spatial aspects of sustainable development (Breheny 1995). Compact cities denote planning goals and measures in order to promote more densely populated cities seen as a necessary condition for sustainability. Containment strategies imply that the future expansion of urban population, activities and interactions would be kept within the existing borders of the built-up areas to hamper the urban sprawl. The theory behind this assumes close and direct links between the physical form of the city (the land use) and people's activities in time and space (Naess 1993, Newman & Kenworthy 1991). These ideas have influenced many planning documents on different levels in the European countries.
6In Sweden, at last, this planning ambition is further emphasized by a general belief, that the Swedish built-up areas continue to spread out and that the entire urban system gradually develops into a North-American situation, where extensive urban areas, sparsely populated, generate immense transportation needs, absolute car-dependency and high levels of energy consumption. The spread is thus perceived as a major driving force behind an ever-increasing mobility.
7Empirically this boils down to questions concerning people’s need for travel in different spatial structures. However, do people living in large and densely populated cities travel less, covering shorter distance, in their everyday lives than people living in smaller and more sparsely populated built-up areas? Perhaps the situation is more complex. A general access to cars and high-speed infrastructures combined with fairly homogeneous patterns of time-use among people may yield similar amounts of travel, despite different regional characteristics (Vilhelmson 1999).
8Against this background two sets of related questions will be explored regarding the actual development in Sweden of the urban or semi-urban structures, and their influence on people’s everyday mobility. The first set of questions deals with land-use and spatial growth: Do the built-up areas in Sweden expand? And are the population densities continuously thinning out in such areas? Our data concern the period 1970 - 1995.
9The second set of questions aims to investigate the main mobility trends of the population: Do people living in dense (or compact) urban areas travel less than people in small and more sparsely populated built-up areas? What shifts in general travel activity and mode use are recognized since the 1970s? This preparatory investigation aims merely of exploring some general patterns and changes in five categories of urban areas. By so doing, I hope to lay the ground for more thorough and detailed research into the geographical aspects of mobility change.
10In order to explore these questions regarding built-up area development and mobility changes, information is needed which i) exhaustively identifies and separates built-up areas from other areas (e.g., the fringe areas of population centres and the purely rural landscape); and which ii) contains a detailed description of trips and travel performed by people living within those areas. These two criteria are often hard to meet. A general impression is that most studies conducted during recent years regarding these matters are weak on the empirical side. Often the delimitations of the areas under study are vague or determined in terms of the administrative boundaries of cities or metropolitan areas. These are seldom fully compatible with respect to the built-up character of the areas. In Sweden the local authority area - the "kommun" - is often the unit of study despite the fact it normally includes wide, sparsely populated areas and of a rural type. This may lead to inaccurate categorizations and erroneous conclusions. The other problem - the lack of travel data - is often tackled by using crude substitutes such as the estimated sales of energy for transportation use within the areas under study.
11In an attempt to avoid such problems national transportation survey data from 1978 to 1996/97 inclusive make up the empirical basis for this study. In the 1978 survey the residence of each person studied was registered with the parish (Swedish: "församling") as the smallest spatial unit. In the 1996/97 survey this information was geocoded in terms of exact spatial coordinates. Yet, as the goal here is to investigate the long-run changes, the parish is the spatial unit chosen in our analysis. Sweden consists of 2.500 parishes making up 300 local authority areas. Further there are more than 1.900 built-up areas (or population centres) statistically defined as areas having more than 200 inhabitants and not more than an average of 200 meters separating the houses. The built-up areas, in this study, thus range from very small villages to the largest city.
12In the present study, in order to classify the degree of urbanization, information about the location, spatial extension and population of the parishes on one hand and the built-up areas on the other, have been merged by using a GIS. Each parish has been assigned to the built-up (or rural) area to which it belongs. Parishes located within a distance up to 2.5 km from a population centre have been assigned to the built-up area.
13The procedure makes it possible to determine the type of built-up (or rural) areas where the persons investigated in the transportation surveys live. However the high level of geographical precision has its price. Since the travel surveys consist of only a limited, though representative, sample of persons in the Swedish population, sometimes only a few observations may fall into a given spatial unit. So the built-up areas have to be assigned to a few broad categories in order to get a reasonably large number of observations in each group. This classification is done with respect to population size and population density of the built-up areas. The outcome of the classification procedure is presented in table 1.
Sources: Data from Statistics Sweden, data of Housing,
Population and Localities; Swedish National Transportation Survey (NTS) 1978 and 1996/97.
Table 1: Classification of Swedish built-up areas. Land area, population size and density 1995 and National transportation survey population 1978 and 1996/97.
14An example of the application, as regards the degree of urbanization of the parishes of Western Sweden, is shown in a map (see chart 1). Five types of urbanized areas are thus defined: small villages, medium-sized villages, small towns, medium sized towns and large towns. It is of major importance here to note that on an aggregate level there is (understandably) a close positive correlation between population size and population density.
15The absolute number of observations (persons 20-64 years old) of the national transportation survey (NTS) is also included in table 1. The 1978 NTS was a pre-planned personal interview including a travel diary covering a day. It was based on a representative sample of persons. The no-response rate was 11 %. The 1996/97 NTS has approximately the same design but is based on pre-planned telephone interviews. The no-response rate now increased to 23 %. Statistics Sweden conducted the surveys.
Chart 1: The classification (of the areas) of parishes according to the population size and density of built-up areas applied to Western Sweden.
16An overall picture of the expansion of urban areas, which has occurred in Sweden during recent decades, is shown in figure 1. The most important observation here is that geographical expansion has almost ceased since the period of growth during the 1970s. A period of fast suburbanisation, starting in the 1960s, culminated in the 1970s. Since then the spatial growth of the built-up areas of Sweden - at all levels - has stagnated.
Sources: Data from Statistics Sweden, concerning Housing, Population and Localities.
Figure 1: The spatial growth of built-up areas in Sweden 1970-1995
17In a corresponding way, and not surprisingly, the 1970s was a time of quick change for the urban population densities (see table 2). Especially in the towns the thinning out of population densities went fast but ebbed away during the 1980s. In the 1990s even a slight increase of density was observed in the medium-sized and large towns.
Sources: Data from Statistics Sweden, concerning Housing, Population and Localities
Table 2: The population density (and population) change within built-up areas in Sweden 1970-1995. Rate of change (%).
18The relative slowdown of population density change was partly an outcome of small population changes within the urban areas in Sweden during these years (see table 2). In the 1990s there was a tendency that the medium-sized and large towns grew somewhat faster than the other types of population centres, though further spatial expansion is not observable so far.
19However, it is not only factors of population size and population density that affect accessibility within urban areas. The locations of activities (housing, workplaces, shopping centres, health-care institutions and so on) are also important. The positive effects of large-scale and functional specialization fuel a process leading to spatial concentration. It is true that the number of shops, service-centres, hospitals and so on tend to decrease, and that the relocation often gives rise to increased geographical separation and needs for travel. But there are also examples of the reversed process. In Sweden during the last two decades a number of new universities have been established in many medium-sized towns and regional centres. Higher education accordingly has become more accessible than before, at least from a geographical point of view.
20We will now turn to the basic question whether people living in large and (thus) densely populated urban areas travel less compared to people living within more sparsely populated areas. Data from the national transportation surveys are employed to compare the situation in the different categories of built-up areas previously defined. Some further limitations have been undertaken:
The analysis is concentrated on the population in the age bracket 20-64 years old. This in order to reduce the effects of differences in the age composition of the population in various areas.
The analysis only includes trips shorter than 150 kilometres (one way). This reduces the random effects of a small number of very long-distance trips. This delimitation is mainly motivated by our focus on trips and travel within what is here understood as the everyday activity-space of people.
Travel activities conducted exclusively by people living within the limits of the built-up areas are considered. This is in line with the stated aim of the study to get insight into the effects of urban structure on travel behaviour. It should be observed that this is not a description of all traffic flowing within the areas under study.
21The average distance travelled by persons aged 20-64 years and living in the defined urban areas are displayed in figure 2. In 1996/97 there is a generalized U-shaped curve representing the relation between built-up area size and density on one hand, and distance travelled on the other. In general there is less distance travelled in more densely populated areas up to a certain level. The medium-sized towns - such as Karlstad, Linköping, Umeå and Uppsala - are those associated with the lowest level of revealed activity space. Among the group of large towns - Stockholm, Göteborg and Malmö - the level is higher.
Source: Data from the Swedish NTS 1978 and 1996/97. Computations by the author.
Figure 2: Average distance travelled (per person and day) by the population living in various localities 1978 and 1996/97. Population 20-64 years. All trips < 150 km. Yellow bar = 1978; Blue bar = 1996/97.
22Twenty years ago, in 1978, the pattern was similar with one important exception: People living in medium-sized towns have substantially reduced their daily average trip length - from 36 to 28 kilometres per person and day. In this type of town a potential for reduced travel has thus been realized.
23It is important to explore further the reasons behind this development - which also explains why no significant change has occurred in the other types of built-up areas. Socio-economic shifts within the population, related to an altered age-composition in the medium-sized towns, might be one answer. Further, in the age bracket 20-64 years shifts may occur that affect the general mobility level. During the recent decade an increasing proportion of young people (20-25 years old) attends higher education at the newly established universities in many of the medium-sized towns of Sweden.
24Access to a car is another obvious and important factor for regional variations in activity space and travel. The distribution of this key factor is shown in table 3.
Source: Data from the Swedish NTS 1978 and 1996/97. Computations by the author
Table 3: Individual access to car (in the household) and drivers license among the population of various built-up areas in 1978 and 1996/97. Percentage of the Swedish population 20-64 years old
25As expected, built-up area size and population density are closely associated with car availability. In the small villages of Sweden almost 9 out of 10 persons in our age bracket have a car (belonging to the household) and hold a driving license. This proportion decreases among larger and denser population centres. In the large towns (Stockholm, Göteborg, Malmö) only six persons out of ten have access to a car thus defined.
26Over the total period from 1978 to 1997 car access of households increased in Sweden. It hides the fact that a period of stagnation developed in 1990's. This stagnation was to a great extent linked to drastic reductions among young people 18-25 years old anddue to reduced income levels, increased unemployment levels and other types of socio-economic changes that took place (Vilhelmson 2003) as well an expression of shifting values and lifestyles among younger people in Sweden (Andreasson & Sjöberg 1996).
27From a geographical point of view it is important to note that the general increase in car access from 1978 to 1997 concentrates, by and large, on the smaller population centres (as well as the rural areas not considered here). The levels of car access did not change at all among people living in the medium-sized or large towns. The gap in car dependency between different types of built-up areas thus widened during the period. Since the end of the 1970s people living in the more sparsely populated areas of Sweden probably have experienced more profound changes in physical accessibility compared to people living in urban areas. It is further to be noticed that the driving-license rates do not differ to the same extent between the areas. This means that there are a substantial number of potential car-drivers in the towns that may be prone to alter their travel habits if income changes, public transportation deteriorates or common lifestyles change.
28The use of different means of transportation is another major factor directly linked to the development and distribution of travelled distance in the urban system. It is also often argued that denser, more compact, urban structures would promote the usage of public transportation, walking and cycling (Fulford 1996). Thus, the major changes in mode use, which occurred during the period of investigation, will here be briefly discussed. The relevant data are presented in table 4.
29Between 1978 and 1997 the use of the bicycle became more important in the Swedish towns. Especially in the medium-sized towns a marked redistribution in mode use occurred in favour of the bicycle. Here the share of the bicycle of all trips increased from 10 % to nearly 25 %. This growth helps to explain the simultaneous decrease in total travelled distance that occurred.
30Car driving is of course the dominating mode of transport in all types of urban areas here considered. The small and sparsely populated areas show the highest shares - 70 % of the total distance travelled - while the share decreases with increased urban density. This discrepancy extended between 1978 and 1997, consistent with the changes in car-access noted above. Among people living in the large towns car-use remained at same level as in 1978, while it even decreased slightly in the medium-sized towns. So, the observed long-run increase in the use of bicycles must have replaced some car travel, not only as measured in terms of frequency but also in terms of distance travelled. From an environmental perspective this is an important observation. It may partly be explained by systematic efforts to make the Swedish towns more bicycle-friendly during those years.
31Public transportation here includes trips done by bus, tramway and underground. (Travel by train is not included due to an altered categorization between the surveys). Among the population 20-64 years old - which excludes a large portion of the regular users of public transportation - no significant differences was observed between 1978 and 1996/97. As expected, the share of public transportation peaks among people living in the largest towns, where the supply of networks and systems is appropriate. A slight decrease of the use of public transportation in medium-sized towns indicates that the expanding use of bicycle also impacts this system.
32Considering trips done by foot to perform out-of-home activities - shopping, bringing children to and from day-care centres and so on - there was evidence of a long run decline of importance in all types of built-up areas. This general decline was similar to a more general trend found in most industrialized countries. Note however that the change of data collection method, from home interview to telephone interview, has affected the reporting of those types of movements.
33Finally, the share of trips done as a car-passenger also decreased between 1978 and 1997. This is associated with the fact that women to a greater extent use the car as a driver and that two-car-households grows faster than car-access in general (Krantz & Vilhelmson 1996). The resulting tendency, mostly affecting travelled distance, was evident in all types of areas.
Data from the Swedish NTS 1978 and 1996/97. Computations by the author
Table 4: Means of transportation use and built-up area 1978 and 1996/97. Population 20-64 years. All trips < 150 km included
34To sum up, the general impression is that between 1978 and 1997 no radical change in the use of different means of transportation among people living in different types of urban structures occurred. The rates of change were rather slow compared to the fast expansion of car-use that developed during 1950 to 1975. The car is the dominating mode of transport and car driving still increases in importance among people living in the built-up areas under study - with one important exception. Among people living in the Swedish medium-sized towns a decrease occurred. Here the importance of the bicycle in the daily travel activities increased considerably. Movements done by foot and as car-passenger have decreased a bit while the share of the public transportation has remained stable. This means that the process of toward individualization in using means of transportation went on, albeit at a slow rate.
35The aim of this paper has been to capture the main tendencies in travel patterns for people living in different urban environments in Sweden. The actual activity space and transportation modes used by inhabitants in built-up areas with different densities have been explored.
36It has been shown that the expansion of urban areas and the thinning out of population densities culminated in Sweden during the 1970s. Since then only minor expansions have occurred. During recent years the densities have even increased within medium-sized and large towns. Hence the change of urban areas and densities cannot be a main force of the changes in mobility identified during recent decades.
37There is an observed connection between population density and travelled distance. This relation is not linear but takes a U-shaped form. In 1997 people living in medium-sized towns (containing 50.000 - 200.000 inhabitants) of Sweden travel the least distance, on average.
38Our observations may possibly question the common idea that increases in population densities within towns - the strategies of compact cities and urban containment - automatically reduce the amount of travel, at least in the Swedish context of urban structures and activities. Instead, we could claim that increased accessibility harbours a potential - hitherto hardly realized - for travel reduction that could conceivably occur under specific conditions and within certain types of urban areas. Such conditions might include, for instance, changing costs of energy/fossil fuels, changing values and preferences of the population, and the shaping of attractive alternative modes of transportation. The observed increase in the use of bicycles in the Swedish medium-sized towns during the last two decades can be seen as a good example of this.
39Thus it may still be maintained that denser urban areas are more robust towards changes in the transportation system for example caused by such factors as energy-shortages or problems of air-pollution. But mere physical access, even when combined with distance-minimization behaviour, is not the only factor affecting the levels of everyday mobility (Salomon & Mokhtarian 1998, Schwanen, Dijst & Dieleman 2004). The decision to travel to participate in activities - where, when and how - to fulfil individual needs and commitments is obviously a highly complex issue.