Planning Practice & Research, Vol. 14, No. 2, 199± 210, 1999
ARTICLE
Spatial Development Patterns and Public Transport: The Application of an Analytical Model in the Netherlands1 L. BERTOLINI
Introduction The continuing decentralisation of cities is a widely documented phenomenon. According to many criteria, and particularly from the point of view of economic ef® ciency, it is one that appears also to make good sense (Ingram, 1998, see also Gordon & Richardson, 1997). However, the environmental and social sustainability of this pattern of urban development remain unresolved issues, as decentralisation is typically associated with high rates of consumption of non-renewable resources, high levels of spatial segregation and other social costs (an overview of these ® ndings is in Ewing, 1997). How to acknowledge the merits of urban decentralisation and cope with the challenges of its unsustainability? A promising approach is public transport-oriented development at the scale of the urban region, as for instance suggested by Breheny & Rookwood (1993), Calthorpe (1993), Owens (1992) and more recently by Hall & Ward (1998). The concept has been or is being at least partially implemented in the Portland area in the USA, in the Swiss cantons of Basel and ZuÈrich, or in the Stockholm area, in Sweden. Current plans for the development of the Thames Gateway in the UK also attempt to apply this urbanisation model (Hall & Ward, 1998). While all these and other similar ones are important efforts, analytical tools to help identify the potential for public transport-oriented urban±regional development are as yet scarce. In the industrialised world the discussion around the (im)possibilities of directing urban development towards public transportation nodes is ubiquitous, as documented by, among others, the publications cited so far. Fundamental questions remain however unasked, let alone answered. Questions as: do public transportation nodes as such have a speci® c urban (re)development potential? And what makes the task of (re)developing one node different from that of another? In order to answer these questions, in this article public transportation nodes in general, and railway station areas in particular, L. Bertolini, Faculty of Geographical Sciences, Universiteit Utrecht, PO Box 80115, 3508 TC Utrecht, The Netherlands. Email:
[email protected] .nl 0269-7459 Print/1360-0583 Online/99/020199- 12
Ó
1999 Taylor & Francis Ltd
199
L. Bertolini will be ® rst brie¯ y placed within the framework of a wider debate around the city and its future. Building upon this somewhat abstract but necessary inception an analytical tool to explore the possibilities and limits of public transport-oriented development will be introduced, and an application to the Amsterdam and Utrecht agglomerations discussed. Public Transportation Nodes and the Contemporary City Probably the most important lesson of the ongoing debate around whether in the contemporary city spatial decentralisation or concentration trends prevail is that they both are at work. Thanks to modern transport and telecommunication technologies human interactionÐ the quintessential urban activityÐ can increasingly be independent of (permanent) physical proximity. As a result, many social and economic activities have been ¯ eeing the inconveniences of high densities and have been decentralising across ever-expanding urban regions. Ingram (1998, p. 1019) ® nds evidence of this trend in ª a set of empirical ® ndings with remarkably strong regularities across countries and citiesº . However, for other activities, opposite trends are observable. Thus, while more routine economic functions as production and administration spatially decentralise, high-quality ® nancial and business services still needing face-to-face contacts concentrate (see among others Moss, 1987; Sassen, 1991; Mitchelson & Wheeler, 1994; Ascher, 1995), and the same seems also to be true for the rapidly emerging sectors of culture, entertainment and the media (Hall, 1996). In the social sphere, next to an ongoing decentralisation of homes towards suburban and exurban locations or even the emerging of virtual social networks as on the Internet, a new, if selective, popularity of dense urban neighbourhoods and of social events granting `live’ human contact can be also observed (see for instance Knox, 1993; Meulenbelt, 1997; The Economist, 1997). The present dynamics of transportation nodes is a direct product of both these spatial decentralisation and spatial concentration forces. The growing ¯ ows of people passing through transportation nodes are a direct product of the increasingly open nature of the present urban system: of people living in one place, working in a second and spending their free time in yet a third, but also of business relationships requiring face-to-face contacts of persons based in far away locations, or of extensive spatial patterns of movement generated by new forms of `urban consumption’ . In the compressed space of public transportation nodes the different populations that characterise the modern metropolisÐ residents, commuters and `city users’ of various sortsÐ are all represented. The coming together of multiple spatial scales (in the most extreme cases from the global of the High Speed Train to the local of the station neighbourhood) is mirrored by a broad range of users (again, in its extreme from the cosmopolitan manager to the drifting homeless). Public transportation nodes and the surrounding areas appear thus one of the very few places in the contemporary city where the participants in its increasingly heterogeneous communities still physically meet (Bertolini, 1996). The fact that all these people pass through public transportation nodes does not necessarily imply, of course, that people are also interacting with each other there. However, these intense and diverse ¯ ows of people do have the potential 200
Spatial Development Patterns of translating into equally intense and diverse patterns of human interaction. If the right conditions are met, social and economic activities still requiring physical proximity can thrive in these areas. And this potential can be realised in a relatively sustainable way, as it can be coupled to environmentally more ef® cient transport and land-use patterns. There appears to be, however, an important prerequisite: that the transportation node not be considered separately from its urban surroundingsÐ or the place of activities. On this principle is based the node±place model that will be discussed next. A Node± Place Model The realisation of the potential for physical human interaction at and around public transportation nodes is the essence of the strategy for public transport-oriented development envisaged here. In this respect, the enhancement and exploitation of the accessibility of the area can provide a unifying principle, on the condition that the term is not interpreted in too narrow a way. In a broader connotation accessibility is not just a feature of a transportation node (`how many destinations, within which time and with which ease can be reached from an area?’ ), but also of a place of activities (`how many, and how diverse are the activities that can be performed in an area?’ ). A third important component of accessibility is the user, or the question `by whom?’ (for theoretical underpinning of this notion of accessibility see Hansen, 1959; Breheny, 1978). In this wider connotation an accessible area is thus one where many, different people can come, but also one where many, different people can do many different things: it is an accessible node, but also an accessible place. As node and place a public transportation node and its surroundings are part of a system of both competing and complementary nodes and places. In general these are all the alternative locations within a region. More speci® cally they are all the transportation nodes belonging to a particular transportation network. Within these systems of locations hierarchies develop, that point to development potentials of different order. At the Faculty of Geographical Sciences of the Universiteit Utrecht, we have developed a model to chart these differences. The model has been initially conveyed in a simple xy diagram (Figure 1). In the diagram, the y value corresponds to the node-content of an area, or to the accessibility of the node, and thus to its potential for physical human interaction (following the reasoning: the more people can get there, the more interaction is possible). The x value corresponds to the place-content of an area, or to the intensity and diversity of activities there, and thus to the degree of actual realisation of the potential for physical human interaction (according to the idea: the more activities are there, the more interaction is actually happening). Four ideal±typical situations can be distinguished. Along the middle diagonal line are areas where the node and the place are equally strong. At the top of the line are areas `under stress’ . Here the intensity and diversity of transportation ¯ ows and urban activities is maximal. This indicates that the potential for physical human interaction is highest (strong node) and that it has been realised (strong place). However, these are also locations where the great concentrations of ¯ ows and activities mean that there is an equally great chance of con¯ icts 201
L. Bertolini Node Unsustained node
Stress
Accessibility
Dependency
Unsustained place
Place
F IGURE 1. The node±place model.
between multiple, extensive claims on a limited space. The property development ideal of maximal intensity of land use and the transport development ideal of maximal ¯ exibility for infrastructure adaptation and expansion have to ® nd here a dif® cult synthesis. At the bottom of the middle line is a second ideal±typical situation, represented by the `dependent’ areas. The struggle for space is here minimal, but the demand for transportation services from area residents, workers and other users and the demand for urban activities from travellers are both so low that supply can be held in place only by the intervention of factors other than accessibility. Finally, two `unbalanced’ situations can be identi® ed. On one sideÐ at the top left of the diagramÐ these are the `unsustained nodes’ , areas where transportation facilities are relatively much more developed than urban activities. On the other sideÐ at the bottom right of the diagramÐ these are the `unsustained places’ , where the opposite is true. This theoretical model has been operationalised by Zweedijk (1997) and Serlie (1998). The node and the place dimensions have been translated in a node- and a place-index, each combining different variables by means of a multicriteria analysis (MCA). The node-index is a measure of the accessibility of the node. Intensity and diversity of transport supply are here the key criteria. The index combines accessibility by train (number of directions served, daily frequency of services, amount of stations within 45 minutes of travel), by bus, tram and underground (number of directions, daily frequency), by car (distance from the closest motorway access, parking capacity) and by bicycle (number of freestanding bicycle paths, parking capacity). The place-index is a measure of the intensity and diversity of activities in the area. For this purpose, the area has been de® ned as the surface included within a `walkable radius’ of 700 metres from the main pedestrian entrance to the public transportation node. The 202
Spatial Development Patterns variables are the number of residents in the area, the number of workers per each of four economic clusters (retail/hotel and catering, education/health/culture, administration and services, industry and distribution) and the degree of functional mix (for methodological and technical details see Serlie, 1998). The node±place model represented in Figure 1 and operationalised as outlined above has been applied to railway station areas in the Amsterdam and Utrecht agglomerations (Figure 2). The two agglomerations form together what is called the ª North Wingº of the Randstad (Figure 3). The Randstad is the urban ring containing the four biggest cities of the Netherlands and other minor centres, arranged around a ª Green Hartº with mostly agricultural functions. As can be seen from Figure 2, Amsterdam Central Station (AC) and especially Utrecht Central Station (UC) are examples of station areas `under stress’ . Hollandsche Rading (HR) epitomises the `dependent’ station area. In the `unsustained node’ Amsterdam Sloterdijk (AS) the transportation node is relatively much more developed than the activity place. In `unsustained places’ as Maarssen (Ma) and Amsterdam Bijlmer (AB) happens, albeit in a less extreme way, the opposite. Most station areas lie in between these extremes.
From a Static to a Dynamic Picture Interviews with local actors and experts con® rm that Figure 2 is a convincing characterisation of station areas in the Amsterdam and Utrecht agglomerations (Serlie, 1998). However, in order to assess (re)development opportunities and threats, such a static picture is not suf® cient. A more dynamic one is needed, and that means a perspective on movements within the node±place diagram. This will be discussed next. The starting point is the assumption that in the long termÐ and provided that no `disturbing’ factors intervene (such as peculiarities in the topography of the area or in the morphology of the transportation networks, but also continuing external subsidies)Ð all locations will lie around the middle line. In other words: provided, and in the measure that demand and supply mechanisms are free to operate, the demand for transportation services from the activity place and the demand for activities from the transportation node will ® nd a (temporary) balance. This assumption of `equilibrium’ in the long term is coherent with the earlier discussed, broad de® nition of accessibility: an accessible node (a location that `can be reached’ in a certain degree) needs an equally accessible place (a location where `something can be done’ in a corresponding degree), and the other way round. The assumption is also connected with the plea made above that the potential for physical human interaction of the station area (i.e. its transportation accessibility, or node-content) be realised through a proportionally intense and diverse concentration of urban activities (or place-content). A clustering around the middle, `equilibrium’ line is indeed what the application of the node±place model to station areas in the Amsterdam and Utrecht agglomerations shows (Figure 2). There are however also a number of locations that are far away from the middle line. These are particularly interesting locations. It can be assumed that in the long term also these locations willÐ pro203
F IGURE 2. Application of the node±place model to station areas in the Utrecht (italics) and Amsterdam (bold) agglomerations.
L. Bertolini
204
Spatial Development Patterns
F IGURE 3. The Amsterdam and Utrecht regions, or the `North Wing’ of the Randstad. Existing situation in 1997 and situation in 2005, showing matching and mismatching of public transport and urban development. Source: author’ s own elaboration of Metz & P¯ ug (1997).
vided that no disturbing factors (continue to) interveneÐ move towards the middle line. The crucial question is: how? Schematically this can happen in one of two ways (see Figure 1). An `unsustained node’ can either move to the rightÐ namely because of strengthening of the placeÐ or fall downÐ because of 205
L. Bertolini (relative) weakening of the node. In the agglomerations under examination an example of the ® rst case (see Figures 2 and 3) would be the localisation of more and/or more diversi® ed urban activities (living, working and/or other facilities) at and around stations as Amsterdam Sloterdijk (AS) or Duivendrecht (Du). An example of the second case would be the (relative) reduction of transportation services in the same stations. It is interesting to remark that in the present policy discussions, both options are considered possible, while a continuation of the status quo seems unlikely. Alternative trajectories towards the middle line can be also imagined for the `unsustained places’ : either invest in the node and see the location move up on the diagram, or disinvest in the place and see the location move to the left. According to Figure 2, in the Amsterdam and Utrecht agglomerations such a dilemma will have to be resolved for station areas such as Amsterdam Bijlmer (AB) or Maarssen (Ma). Similar dynamics applies also to locations closer to the middle line. These could (and in a growing or declining urban region will always) move because of (dis)investments in the node and/or the place towards one of the two unbalanced situations, where the options sketched above would emerge. The combined result will be a movement of station areas upwards or downwards along the middle line. In the (hypothetical) case of perfect coordination of (dis)investments in the place and the node this movement will be slant instead of stairs-wise. From Compact Cities to Accessible Urban Regions The contrasting movements across the diagram sketched above do not only have implications for each station area separately, but also for entire systems of areas, as the two systemsÐ the Amsterdam and the Utrecht agglomerationsÐ of the application. The situation in an urban region at a given moment can be described in terms of clustering of locations: on the top, at the bottom or elsewhere along the middle line of the node±place diagram. Differences in clustering between the Amsterdam and Utrecht agglomerations can be observed in Figure 2. The Amsterdam agglomeration currently shows a clustering in the centre-top of the line, the Utrecht region a clustering at the bottom, with the sole exception of Utrecht Central Station (UC). The network of railway stations in the Amsterdam agglomeration appears thus to (potentially) provide a better support to urban development than the network in the Utrecht agglomeration. As already contended, station areas at the very top of the middle line can offer high opportunities for physical human interactions but may trigger also the most intense spatial con¯ icts. The (relative) borders of growth will be reached here before locations with lower node and place values. The withdrawal at the beginning of the 1990s of an ambitious project to redevelop the areas around Amsterdam Central StationÐ the so-called IJ-banks projectÐ and the continuous delays of a similar plan at Utrecht Central StationÐ the so-called Utrecht Centrum ProjectÐ are concrete signals of this. In both cases crucial factors of failure were the great complications and high costs of accommodating a massive new programme in already congested locations (details in Bertolini & Spit, 1998). It becomes then important to have in supply alternative station areas with adequate transport capacity, if growth is not to be diverted to areas not well connected to the public 206
Spatial Development Patterns
F IGURE 4. (a) Compact city policy: build in or next to the existing city. (b) Public transport-oriented development: build within walking/cycling distance of station. Legend: s , existing urban area; d , future urban area; Ð , railway; Ð , motorway.
transport network, as mostly is the case. Amsterdam has those alternative station areas, Utrecht does not. In Utrecht, the situation is not likely to change in the short term: with the exception of the Central Station area, current investments appear directed either to places where public transport nodes are absent or weak (as in the planned urban expansions to the west and south-west of the city, see Figure 3), or nodes where there are as yet no clear urban development prospects (as with stations of the new `Randstadspoor’ regional rail system, see Serlie, 1998). The better starting position of the Amsterdam agglomeration does not mean that planning consequently capitalises upon it. While market demand has for several years been clearly oriented towards station areas along the peripheral railway and motorway rings, node and place development initiatives have long remained uncoordinated. There are, for instance, important public transport nodes at relatively little developed places, as Amsterdam Sloterdijk (AS) and Duivendrecht (Du), and intense and/or diverse concentrations of activities around relatively undeveloped nodes, as at Amsterdam Muiderpoort (AM) and Amsterdam Bijlmer (AB, compare Figures 2 & 3). Current plans, as those for the Zuidas (a large, mainly of® ce development centred around the southerly motorway ring and the stations of Amsterdam Zuid and RAI), improve this situation only partly, because they still try to direct growth towards a few areas and not towards the entire urban±regional network of nodes and places. Finally, it is intriguing to connect the differences between the Amsterdam and Utrecht agglomerations to the different morphologies of the two railway networks: a combination of a radial and an orbital network in Amsterdam and a strongly radial network in Utrecht (Figure 3). This is a ® nding which, if generalisable, could be an additional strong argument in support of the frequently heard plea for the development of orbital public transport lines (see for instance Hall & Ward, 1998). On a more general level, the application of the node±place model to the Amsterdam and Utrecht agglomerations shows the limits of a `traditional’ 207
L. Bertolini compact city policy as currently pursued in the Netherlands and being implemented or considered in other countries. The above-mentioned urban expansions in the Amsterdam and Utrecht agglomerations are all coherent with Dutch compact city principles, which prescribe that urban development take place either within or adjoining existing urban areas (compare implementation in Figure 3 and policy principles in Figure 4a). Yet, these developments appear often at odds with a public transport-oriented urban development. This is no minor drawback. Private motorised transport is the biggest consumer of energy and the main source of pollution in the contemporary city. Furthermore, the hegemony of the automobile enhances spatial segregation and social inequality patterns, excludes from societal life those who cannot drive and poses a threat to the viability of the public domain (see among others Holtz Kay, 1997). On the other hand, high physical mobility is a structural feature of our open urban systems. The solution of this sustainability dilemma lies in a spatial development pattern where more environmentally and socially friendly transportation modes, such as public transport, can increase their share. But if urban and public transport development are to be more coordinated, the principle of accessibilityÐ as discussed in this paperÐ has to play a much greater role, and that of proximity, or compactness, must be applied in a more differentiated way. As Verroen (1996, p. 107) concludes, drawing from an extensive study of the impact of different urbanisation patterns on mobility: ª Urbanisation ¼ as close as possible to or between existing urban areas and with a good connection to high-quality public transport lead to the most favourable effects on mobilityº (emphasis added). Figure 4b shows schematically how this correction of course could look like. Parallels with visions emerging elsewhere are intriguing (see for instance Calthorpe, 1993; Hall & Ward, 1998). Conclusions: `Deconcentrated Clustering’? The node±place model introduced in this article offers a conceptual framework for the exploration of the (re)development potential of station areas in an urban region and thus improves the chances for public transport-oriented development. The model requires further re® nement, broadening of the empirical base, and inclusion of future plans. A thorough consideration of the more qualitative aspects of station area (re)development strategies has to be added, and a complementary, in-depth analysis of each speci® c situation, including process and context factors is also necessary (see, for some steps along this way Bertolini, 1996, 1998; Bertolini & Spit, 1998). In spite of these limits the application of the node±place model to the Amsterdam and Utrecht agglomerations already provides material for re¯ ections of a more general nature. I would like to conclude with some, still exploratory and deliberately provocative, remarks in this direction. The node±place model helps identify opportunities for intensi® cation and/or differentiation of urban activities around (strengthened) public transportation nodes. These might be central stations in the historical cores, but also public transport nodes in the urban periphery (as the stations along the Amsterdam railway and motorway ring) or in secondary regional centres (in the application 208
Spatial Development Patterns these are places such as Hoofddorp and Weesp south of Amsterdam, or Woerden west of Utrecht). The city is an increasingly open system. Increasingly people live, work and spend their free time independently of municipal borders. The same applies to the activity patterns of ® rms and other organisations. A traditional `compact city’ policy is possibly still sensible in situations where this structural openness of the urban system is limited. Elsewhere an urban development strategy, geared to the accessibility of urban±regional networks of transportation nodes and activity places, appears more adequate. Some will say that this strategy reminds them of the `clustered deconcentration’ policies of the past. Maybe. I would prefer to use a term like `deconcentrated clustering’ . It might seem just playing with words, but there is a fundamental difference in the focus: in the past at stake was a particular way of deconcentrating, today urban deconcentration is a fact, the point is how to cluster! Acknowledgements With thanks to Ton Kreukels, Tejo Spit, three anonymous referees and the editorial boards of Stedebouw & Ruimtelijke Ordening and Planning Practice & Research for their comments on earlier versions. Note 1. An earlier version of this article has been published, in Dutch, in the journal Stedebouw & Ruimtelijke Ordening.
References Ascher, F. (1995) Me tapolis, ou l’ avenir des villes (Paris, Editions Odille Jacob). Bertolini, L. (1996) Nodes and places: complexities of railway station redevelopment, European Planning Studies, 4, pp. 331±346. Bertolini, L. (1998) Station area redevelopment in ® ve European countries. An international perspective on a complex planning challenge, International Planning Studies, 3, pp. 163±184. Bertolini, L. & Spit, T. (1998) Cities on Rails. The Redevelopment of Railway Station Areas (London, Spon). Breheny, M. J. (1978) The measurement of spatial opportunity in strategic planning, Regional Studies, 12, pp. 463±479. Breheny, M. & Rookwood, R. (1993) Planning the sustainable city region, in: A. Blowers (Ed.) Planning for a Sustainable Environment, pp. 150±189 (London, Earthscan). Calthorpe, P. (1993) The Next American Metropolis (Princeton, Princeton Architectural Press). Economist, The (1997) The music of the metropolis, 2 August, pp. 25±26. Ewing, R. (1997) Is Los Angeles-style sprawl desirable?, Journal of the American Planning Association, 63, pp. 107±126. Gordon, P. & Richardson, H. W. (1997) Are compact cities a desirable planning goal?, Journal of the American Planning Association, 63, pp. 95±106. Hall, P. (1996) Revisiting the nonplace urban realm: have we come full circle?, International Planning Studies, 1, pp. 7±16. Hall, P. & Ward, C. (1998) Sociable Cities: The Legacy of Ebenezer Howard (Chichester, John Wiley & Sons). Hansen, W. G. (1959) How accessibility shapes land use, Journal of the American Institute of Planners, 25, pp. 73±76. Holtz Kay, J. (1997) Asphalt Nation. How the Automobile Took Over America and How We Can Take It Back (Berkeley, University of California Press).
209
L. Bertolini Ingram, G. K. (1998) Patterns of metropolitan development: what have we learned?, Urban Studies, 35, pp. 1019±1035. Knox, P. L. (1993) The postmodern urban matrix, in: P. L. Knox (Ed.) The Restless Urban Landscape, pp. 207±236 (Englewood Cliffs, NJ, Prentice Hall). Metz, T. & P¯ ug, M. (1997) De Nieuwe Kaart. Atlas van Nederland in 2005 (Rotterdam, Nai Uitgevers). Meulenbelt, K. (1997) Van e e n-naar een twee-toppige woningmarkt. De opkomst en neergang van woonbuurten in het Amsterdamse en Rotterdamse stadsgewest (Amsterdam, AME). Mitchelson, R. & Wheeler, J. (1994) The ¯ ow of information in a global economy: the role of the American urban system in 1990, Annals of the Association of American Geographers, 89, pp. 87±107. Moss, M. L. (1987) Telecommunications, world cities and urban policies, Urban Studies, 24, pp. 534±546. Owens, S. E. (1992) Energy, environmental sustainability and land-use planning, in: M. J. Breheny (Ed.) Sustainable Development and Urban Form (London, Pion). Sassen, S. (1991) The Global City: New York, London, Tokyo (Princeton, Princeton University Press). Serlie, Z. (1998) Stationslocaties in vergelijkend perspectief. Master thesis, Universiteit Utrecht. Verroen, E. J. (1996) Mobiliteitsvriendelijk e verstedelijkingsvor men na 2005, in: F. Dieleman & H. Priemus (Eds) De inrichting van stedelijke regio’ s. Randstad, Brabantse stedenrij, Ruhrgebied, pp. 100±108 (Assen, Van Gorcum). Zweedijk, A. (1997) Knoop of Plaats? Naar een operationalisering van het begrip stationslocatie . Master thesis, Universiteit Utrecht.
210
