Future Urbanization Pathways: Compact, Polycentric, or Diffuse

Abstract

This chapter presents the three future urbanization scenarios created as part of the ESPON SUPER project: compact, polycentric, and diffuse. The research design is explained using a four-step procedure and discusses the distinction between exogenous (environmental) and normative (policy) scenarios. It also explores the qualitative and quantitative approaches for the analysis and communication of scenarios and, specifically, the use of the LUISETTA land-allocation model to simulate urbanization in the three scenarios. After presenting the scenario storylines, the impacts on the physical landscape are discussed as well as changes in urbanization and population density. The chapter concludes by emphasizing the potential of these scenarios to serve as a basis for public debate on preferred policy directions and as a tool for drafting a strategy to achieve sustainable urbanization.

4.1 Introduction

The future is, by definition, uncertain. At most, science can make reasoned estimations by identifying relevant trends and developments and theorizing about how particular driving forces will change over time. Hypotheses can then be made about whether it is more likely that a given trend continues, becomes amplified or diminishes. As more variables are considered, complexity increases, especially when variables interact. At some point, computer models usually take the place of human reasoning, effortlessly compiling astronomically large datasets at lightning speed. Computer models do not build themselves, but are themselves the result of human theorizing, empirical research, and testing and are continuously being calibrated and adapted. Urbanization is a good example of such a complex phenomenon with many driving forces and intervening variables.

Europe, indeed, the whole world, is facing major upheavals that will affect future urbanization. The introductory chapter identified key driving forces, many of which are being affected by current events. Geopolitical conflict, technological advancements, migration, and energy poverty can affect the economy as well as preferences regarding where individuals and businesses wish to locate. On the other hand, the transition to renewable energy adds another competitor for rural land, restricting supply. Housing costs are another uncertain factor: throughout Europe these have risen faster than economic growth, making homes an attractive investment object on the one hand, but on the other, forcing many people to economize on urban space or seek a residence further afield. Finally, the effects of Covid-19 on urbanization are highly disputed. Whereas some urban professionals feel that it will fuel suburban diffusion, others believe that it will increase demand for socially cohesive and green cities (Evers, 2020).

An appropriate method to study future urbanization is by using scenarios. Scenarios describe alternative futures when uncertainty is too high to warrant forecasting but high enough to avoid speculation (Dammers et al., 2019; Scholles, 2008). This method was first used in the context of military planning in the early postwar era as well as for large corporations and public administrations attempting to prepare for the future. Today, it is used in many different fields and has become commonplace in spatial planning (Bradfield et al., 2005; Khakee, 1991; Salewski, 2012).

This chapter discusses three policy scenarios based on the different modes of urbanization described in Chapter 1 and used throughout the book. The first section discusses the scenario method and the key choices that need to be made. The second presents the scenario design chosen for the ESPON SUPER project and outlines how the modelling was performed. The third part presents the storylines, recounting how policy orientations redirected urbanization towards different developmental pathways, resulting in changes in the magnitude and shape of urban development in 2050. The final section reflects on the differences between the scenarios and how they can be used by policymakers.

4.2 Methodological Considerations

There are many different methods available for scientists to probe the future. The most straightforward way is to extrapolate current trends to produce an estimate of a future situation. This can be done to illustrate challenges that require policy attention; well-known figures showing the exponential growth in population and CO₂ (with dotted lines for future development) fall into this category. More sophisticated analyses pay attention to the interaction between trends and driving forces. They make reasoned estimates about how the driving forces will develop and how other trends will affect the trend under investigation. Population forecasts generally rely on these kinds of methods. For broader policy concerns, a set of analyses can be performed and combined to provide an outlook; this analysis is common in the environmental field, where aspects such as air, soil, and water quality are treated separately. All these examples pertain to the production of an approximation of a probable future using the best available means; uncertainty is usually illustrated utilizing a bandwidth. However, in many cases, uncertainty is so high that forecasts become meaningless—everything seems to fall into the bandwidth of possibility, and little can be said in the way of probability.

There are various ways to deal with uncertainty in future studies (Evers & Vogelij, 2021). If uncertainty is extreme, such as long-term technological advancements and geopolitics, it can be useful to produce a number of speculations (e.g. a utopic or dystopic situation) and reflect on their origins. This can reveal institutional shortcomings and suggest policy action to stave off or anticipate such extreme situations. Similarly, speculations about unexpected extreme events—called ‘wild cards’ or ‘black swans’ in the literature—such as pandemics, war, or revolutionary technology, can provide insight into the robustness of current institutions and practices to improve preparedness (Dammers et al., 2019). Speculations need not be probable, only possible.

Scenarios, which are essentially multiple narratives about the future, provide a middle-ground between speculations and forecasts. Like forecasts, they are based on an analysis of the drivers of existing trends and developments and try to make reasoned statements about the future on this basis. Like speculations, they require imagination and are open to higher levels of uncertainty that can challenge current paradigms. To be useful, do not need to be probable but should always be plausible. Plausibility is enhanced if the scenarios are transparent about why and how the different futures emerge: often this is done by varying a crucial variable or logical set of variables.

The scenario method has become increasingly common in architecture (Coleman, 2014), urban planning (Abou Jaoude et al., 2022) and is already a hallmark of modern strategic spatial planning (Vogelij, 2015) because it can address economic, environmental, and social uncertainties (Abou Jaoude et al., 2022; Wiebe et al., 2018). An advantage of the scenario method is that it invites a discussion on the desirability of alternative futures as well as a discussion on what current planning interventions could bring them about. In this way, it contributes directly to planning decision-making (Chakraborty & McMillan, 2015; Khakee, 1991) and is a common element in planning support systems (Abou Jaoude et al., 2022). The Netherlands has pioneered the use of scenarios in planning, producing countless studies over the past half-century on how the territory could and should develop, usually with the aid of urban designers (Salewski, 2012).

Depending on the purpose, stakeholder involvement, scope, orientation, data, and other considerations, different scenario methodologies can be applied (Chakraborty & McMillan, 2015; Radeljak Kaufmann, 2016). While this has resulted in a rich and varied tradition, some authors have described the state of scenario development as methodological chaos (Bradfield et al., 2005). Rather than giving a comprehensive overview or detailed chronology of this method, this section will prove some of the most important defining features of scenario design. This will help to place the SUPER scenarios in context.

Most scenario studies follow a standard process. Scholles (2008) identified four basic phases of scenario design in spatial planning. The initial ‘system analysis’ phase is about identifying key factors and deciding whether they should be considered stable or variable over the scenario period. The second phase regards the selection of which key factors, and their possible trajectories over time, should define the scenarios. The third phase determines how the key factors should vary, based on a combination of scientific insight and creativity (Kosow & Gaßner, 2007). In the final phase, the scenarios are elaborated and communicated using appropriate means (e.g. narratives, diagrams, figures) to support societal discussion or decision-making. This four-step scenario-design procedure is usually just part of a wider project. For example, guides exist for researchers on how to produce effective scenarios within ongoing environmental and spatial planning policymaking processes and ensure they impact decision-making (Dammers et al., 2019).

A key distinction in scenario design is between exogenous (environmental) and normative (policy) scenarios. Recalling the discussion on drivers in Chapter 1, environmental scenarios vary exogenous factors that cannot easily be influenced, such as global economic development, geopolitics, and climate change. These kinds of scenarios are useful for identifying robust measures that would be beneficial in all situations or drawing up contingent strategies. Here, it is important to point out that it is impossible to choose between environmental scenarios. Normative scenarios, on the other hand, vary factors that represent (exogenous) policy choices and are very useful for showing the implications of decisions. Sometimes one or more scenarios are constructed as an ideal which can be used as the basis for a later strategy. Either of these two types could be applied to urbanization. An environmental scenario study would explore the difference, for example, between high and low economic growth on the magnitude and form of urban development under stable policy conditions. A policy scenario study would hold constant as many external factors as possible to explore how urban development would occur under different policy conditions.

Another crucial decision regards the number of scenarios to make. A single but fundamental policy choice or extreme uncertainty can sometimes be expressed by two scenarios (best-case versus worst-case), although this often invites criticism of being oversimplistic. In some instances, three scenarios are created, e.g. showing the probable, the possible, and the desired future (Börjeson et al., 2006). A common method is to vary two key variables, chosen based on their significance and level of uncertainty, to create four scenarios along two axes (Kosow & Gaßner, 2007). One can also choose to create embedded scenarios, which can illustrate the implications of different policy orientations within different environmental contexts. This multiplies the number of scenarios and hence the complexity of the design. The desire for simulating a broad range of possible futures (completeness) should therefore be weighed against the time and costs of creating and using these scenarios. Most authors limit themselves to three to four main scenarios (Alcamo & Henrichs, 2008).

Finally, the method of analysis and communication needs to be established (step 4). This can take the form of qualitative approaches, such as essays about future events using illustrative examples and made-up statistics to pure works of fiction (future newspaper articles) or artistic renderings. More quantitative approaches use computer models to calculate how the main variables will evolve and affect other variables using econometric modelling. In both cases, it is important to establish a plausible storyline that connects drivers to the final state.

4.3 Scenario Design

This section presents the research design of the future urbanization scenarios created as part of the ESPON SUPER project (Evers et al., 2020). Recalling the first phase identified by Scholles (2008), the first task is to understand the ‘system’ of urbanization. As discussed in Chapter 1, the amount and location of land converted to urban use is considered to be the product of a combination of factors, some exogenous to the planning system and some endogenous. Concerning the second phase (selection of key factors), the SUPER project investigated the extent to which exogenous factors explained urbanization, finding only a weak relationship with population and an even weaker relationship with economic development, the two most commonly cited drivers (Van Schie et al., 2020), suggesting that policy matters. As an illustration, Chapter 3 provided myriad examples of how planning policy and practices have influenced the amount, shape, and direction of urban development.

This insight led to the key scenario-design decision (step 3) to hold exogenous factors constant across scenarios and vary policy orientations according to the three urbanization types. Moreover, given the focus of this book is on how to make urbanization more sustainable, the normative/policy scenario approach is more appropriate than an environmental scenario approach. Table 4.1 summarizes which variables were held constant and which varied.

Table 4.1 Key elements of the scenario design

The last step identified by Scholles (2008) is communication. Here it was decided to use a combination of techniques to illustrate the scenarios, from a storyline narrative to quantitative modelling. This first necessitated the identification of societal attitudes that could help explain the adoption of relevant policies, resulting in the three urbanization types. The premise and outcome of each scenario are therefore the following.

• Compact scenario: strong urban containment policies are enacted in the early 2020s as a collective response to perceived spatial challenges; sustainability and other matters of public interest are prioritized. By 2050, urban development mainly occurs within or at the edges of the largest cities.

• Polycentric scenario: to strengthen community and local identity, policies are implemented in the early 2020s to promote the creation of a well-connected network of small and medium-sized towns. This stems from attitudes about the need for social cohesion and recognition of interdependence. By 2050, development will be clustered in urban regions.

• Diffuse scenario: policies encouraging urban diffusion were implemented in the early 2020s, allowing people to flee crowded and expensive cities and buy large homes in spacious surroundings. This stems from individualistic attitudes. By 2050, the countryside has absorbed many scattered urban functions and replaced agriculture.

The final task was to translate these notions and decisions into quantitative input for the LUISETTA model; this is explained in the text box below. Because the LUISETTA model was run up to 2050 according to a single scenario logic, this year comprised the focus of the scenarios. The intervening years are relatively uninteresting for the exercise as urbanization occurred incrementally and cumulatively without adjustment of variables in the interim. In theory, it would be possible to create more sophisticated scenarios with branching at critical junctures, but this would have made them less distinct and more difficult to compare. For the same reason, no hybrid scenarios were considered. This is because the goal within the SUPER project was not to predict, but to illustrate radically different yet plausible futures to support a policy discussion.

The LUISETTA Model

LUISETTA is an open-source version of the Land Use-based Integrated Sustainability Assessment (LUISA) modelling platform developed by the EU Joint Research Centre (Jacobs-Crisioni et al., 2022). Although LUISETTA has fewer functionalities than the full version, it was deemed sufficient to draw up policy-oriented scenarios.

Approximately 40 datasets are incorporated into the model, including information on age, population, accessibility, distance to roads and water, terrain slope, soil contamination, and high-value farmland. A high-resolution (100 m) version of the 2012 Corine land cover (CLC2012) map is used as the base map, which includes the EU member states, the UK, Iceland, Norway, Switzerland, and Liechtenstein.

For each five-year interval, LUISETTA performs three main tasks. First, the demand for different types of land use (e.g. urban development) is determined in hectares using the projections included in the model at the NUTS 2 level. The second task is to distribute these demands geographically. The result is a map of land conversion pressures. Suitability for urban use, for example, is determined by distance to roads, water, existing population, and relative accessibility. The final output of the model is a modified version of the CLC2012 map for each five-year period between 2015 and 2050, with pixels categorized as urban, commercial, agricultural, and natural.

Because LUISETTA only contains information for EU member states (EU28) in 2012, projections for non-EU ESPON countries (Iceland, Norway, Switzerland, Liechtenstein) were calculated based on national demographic projections from Eurostat, which were then distributed over the respective NUTS2 regions in these countries using the same assumptions contained in the LUISETTA model. Specifically, household size converges to 1.8 across Europe in 2050.

The built-in baseline is policy-poor (few spatial policy restrictions are included) and can be run without additional input or adaptation. After examining its output, it was deemed suitable for the ‘diffuse’ scenario, with a small adjustment to have this occur more intensely in highly populated areas (consistent with the sprawl literature). Two methods were then employed to create the ‘polycentric’ and ‘compact’ scenarios. To simulate higher densities, demand for residential and commercial development was lowered so that the model would convert fewer hectares to urban use. Second, to affect the distribution of land-use change, a cartographic layer of ‘relative attractiveness’ was inserted into the baseline map that added weights to certain locations (e.g. near rail stations) for their suitability for urbanization.

4.4 Future Urbanization Storylines

Given that the LUISETTA model provides an image of 2050 urbanization, it is tempting to concentrate the discussion solely on the endpoint of the scenario. Salewski noted that this often occurs in practice: “In the communication process the final image of an extreme future state is most powerful, while the underlying diachronic analysis is usually lost” (2012, p. 305). We seek to avoid this by paying attention to all parts of the storyline.

Given that it is extremely unlikely that all regions in Europe will follow the same path given the territorial and political diversity, the scenarios should in no way be interpreted as predictions, but instead as thought experiments regarding possible urbanization pathways and their effects. As stated, creating narratives about future development is a common spatial planning tool, particularly for strategic decision-making (Albrechts et al., 2003; Throgmorton, 1996).

This section presents the three narratives on future urbanization in Europe. Stylistically, they seek to be provocative and bold, while at the same time remaining plausible. Each scenario is presented in three parts: rationale, policy package, and impact. The rationale presents the general logic of urban development based on the prevailing social attitudes within the storyline. The policy package establishes a plausible link between the social attitudes and the tools chosen to influence urbanization that were drawn from the SUPER intervention database. Many of these were already discussed in Chapter 3. The scenarios conclude with a final image, generated by the LUISETTA model, of how several European regions could appear in 2050.

4.4.1 Compact Scenario

In this scenario, social attitudes no longer hold sacred the suburban ideal of single-family houses and private cars. To those born after the introduction of the smartphone and growing up with video conferencing, commuting long distances seemed like a colossal waste of time and resources. A better alternative was to live in a smaller apartment conveniently located near services and activities in a bustling urban environment; this was also preferred by the ageing population. Moreover, given high energy prices, this was also economically prudent. Widespread awareness about the impacts of climate change gave support to policies directed at compact urbanization.

This philosophy resulted in a coherent policy package based on existing interventions in Europe for achieving compact urbanization. A relatively straightforward and tried approach is to define physical boundaries for urban growth. Examples that were drawn on included London’s Green Belt, Turin’s Corona Verde Plan, Leipzig’s Grüner Ring, the Metropolitan Cork Green Belt, and Stockholm’s Urban Containment Strategy. Inspiration was also taken from policies curtailing urbanization, such as the zero-growth plan of Cassineta di Lugagnano in Lombardy, Germany’s 30-hectare target, Switzerland’s anti-sprawl laws, Dutch ‘red for green’ schemes, and the land take reduction scheme in Flanders. This was coupled with strategies to encourage densification through the use of vacant urban land (e.g. Royal Seaport eco-district, Stockholm), redevelop brownfields and repurpose existing underutilized urban land (e.g. Reinventing Paris, Dublin Docklands), or increase the quality of existing urban spaces (e.g. Berlin Programme on Sustainable Development).

By 2050, the impacts of the compact urbanization policy were visible in the physical landscape. Figure 4.1 shows the LUISETTA results of the compact scenario in five European regions in 2050. The darkest shade shows urban development since 2020 whereas the medium shade shows areas that were already built up at the beginning of the scenario period. In large urban regions, there was further coalescence of urban areas (Randstad and Brussels-Antwerp), while in less urbanized areas urbanization was confined to the edges of the largest cities (Bologna-Ravenna, Stockholm, Constanţa).

Fig. 4.1

Compact scenario output for five European regions in 2050

An analysis of the LUISETTA output revealed that the largest increase in urban area (urbanization/land take) occurred in NUTS2 regions with large cities (which we realize is directly related to the model input). Interestingly, given the scenario storyline, population density not only increased in regions with the biggest capital cities but also in urban regions in southern Germany, Italy, and Spain.

4.4.2 Polycentric Scenario

In this scenario, social attitudes favoured a return to community. Rejecting both American-style individualism associated with sprawling development as well as the forced urbanity of the compact city, life in small and medium-sized towns was held up as an ideal middle ground. Such a community could be largely self-sufficient, containing both jobs and facilities as well as homes (Handy, 2005). Food and energy production could occur in a decentralized manner according to a ‘buy and produce local’ philosophy. This urban form was especially attractive for the increasingly ageing population, due to its recognizable traditional structure, giving a sense of belonging and inclusion. For this same reason, reliable and accessible public transport both within and between towns was seen as vital.

This philosophy resulted in a coherent policy package that harkens back to the polycentric urbanization policies of yesteryear. The most prominent example is the Garden City promoted by Ebenezer Howard (Howard, 1902), which inspired planned communities in Europe and abroad. There are various historical examples to point to, such as the postwar new towns and growth centres in the UK, Sweden, and the Netherlands. More recently, the city plan of Stara Zagora in Bulgaria established settlements as secondary urban centres, with available public services and quality housing opportunities (Cotella et al., 2020).

The concept of transit-oriented development (TOD) was embraced as a way to create interconnected and walkable communities without big-city densities (Papa & Bertolini, 2015). A well-cited example is Ørestad, a district of Copenhagen built on the backbone of a light-rail system. It contains high-quality urban functions and residential densities near the stops and high-quality nature in the vicinity (Knowles, 2012). Similar developments can be found throughout Europe, such as Paris, Rotterdam, Vienna, and Stockholm (Paulsson, 2020; Pojani & Stead, 2015).

What polycentric urbanization will look like in the physical landscape in 2050 is shown in Fig. 4.2 according to the LUISETTA model results. The dark shades in the picture represent a new development of urban tissue, and light shades built urban areas that existed at the time of the beginning of the scenario period. It is evident that new urban development is not concentrated in the outskirts of major cities, but in smaller towns in their immediate vicinity or in towns located in broader gravitational areas of the major city in a row along main transportation routes.

Fig. 4.2

Polycentric scenario output for five European regions in 2050

The statistical output shows that the increase in urban land use occurred mainly in populous regions, leading to significant differences in larger countries (France, Germany, and Italy). Population density increased mainly in urban regions of major cities, but also in other developed regions (e.g. along the Mediterranean coast).

4.4.3 Diffuse Scenario

In this scenario, social attitudes are more individualistic. A contemporary Broadacre City is envisioned, where dispersed services and facilities are accessed by private transport modes, often powered by self-generated electricity. In this scenario, people wish to live in single-family homes on large plots of land. The Covid-19 pandemic was a trigger in this direction, with unpleasant memories of quarantines in small apartments. Especially the elderly population longed for homes with spacious gardens after decades of working in urban centres. The urban heat island effect was also a reason to vacate cities for a greener environment. Digital technologies that enabled work, shopping, education, and other activities at a distance reduced the necessity of proximity: a shift from physical to the virtual.

This philosophy resulted in a coherent policy package to promote diffuse urbanization. The budgets of planning departments, which were often bulwarks of outdated notions of compact development (Pagliarin, 2018), were slashed and restaffed so that existing restrictions could be swiftly abolished. A lean, flexible planning procedure was introduced, making it easier to buy land and build one’s own house in green surroundings. Various European policies inspired this paradigm shift. The area of Oosterwold near Amsterdam in the Netherlands had experimented with a very hands-off regime, with no zoning plan or coordination by the government for services and infrastructure—everything is arranged and financed by the landowners themselves (Cozzolino et al., 2017). Flanders’ rule allowing ‘fill-in’ housing construction along existing roads was also a source of inspiration, as was subsidizing suburban and exurban construction (such as in Lithuania to repopulate shrinking regions) as well as the Italian and Croatian policies of providing amnesty to illegally constructed buildings.

The first palpable effect was seen in the immediate vicinity of cities, which spread outwards into rural and natural hinterlands at low densities. This is illustrated by the LUISETTA model output (see Fig. 4.3), which shows that urban development is becoming more amorphous and dispersed (Brussels-Antwerp-Constanţa) and encroaching on natural spaces (Bologna-Ravenna, Stockholm). The Netherlands finally received the ribbon development it had so long sought to prevent (dark shades indicate new urban fabric, light shades existing urban areas).

Fig. 4.3

Diffuse scenario output for five European regions in 2050

Finally, the statistical analysis of the diffuse scenario model output revealed a large increase in urban land use (120–140%) in most regions, the highest proportion at the regional level of all scenarios. Although diffuse urbanization occurs at low densities in terms of morphology at the local level, densities still increased in many areas at the regional (NUTS 2) level.

4.4.4 Cross-scenario Comparison

A synthetic interpretation of results derived from the LUISETTA model is a complex and challenging undertaking given the territorial diversity of Europe. One interesting finding is that the changes in demand (used as input in the scenario design) did not produce an equivalent change in supply (model output). This is likely due to intervening variables such as site suitability and inbuilt transaction costs from the previous land use. For this reason, the 50% decrease in demand introduced in the compact scenario did not produce a commensurate 50% reduction in urbanization. Instead, the differences between scenarios were far less than expected. Using the diffuse scenario as a baseline, the model indicates an 4% average reduction in urbanization across Europe for both compact and polycentric scenarios. This finding highlights a shortcoming of the model for realistically simulating urbanization scenarios, unless another method can be found than adjusting the demand module and spatial attractiveness factors. It should also be noted that since this research was carried out, there have been additional advances in land-use modelling, for example, the 2UP model which employs highly detailed datasets on population and soil sealing (although unfortunately little land-use differentiation) at the global level (Koomen et al., 2023).

Bearing these shortcomings in mind, some interesting patterns do emerge when individual countries are compared. For example, there are some cases where urbanization remains relatively constant across all three scenarios, indicating that societal preferences may have a modest effect on outcomes. In Lithuania, the scenarios produced identical urbanization levels, perhaps due to the modest demographic development predicted up to 2050. More surprisingly, in some countries, urbanization was higher in the compact scenario than in the diffuse scenario (Bulgaria, Croatia, Germany, and Latvia), albeit very slightly. In general, the compact scenario produced a reduction in urbanization, the most substantial difference being seen in Iceland (17%), the UK (14%), Malta (12%), Belgium and Luxembourg (11%), and Sweden (11%). In some countries, the polycentric scenario yielded a greater decrease than the compact scenario (Estonia, Latvia, and Denmark). The model also produced results on changes in urban population densities, which were directly derived from the output on urbanization. This usually varied little between scenarios. As expected given the model input, densities (people/ha urban land) tended to be slightly higher in the compact scenario, less so in the polycentric scenario and least in the diffuse scenario. Exceptions include Austria and Estonia. Malta showed the exact opposite tendency, with the diffuse scenario producing significantly higher densities than polycentric and compact.

Given the fact that changing demand in the model had so little effect on the results, and the surprising outcomes of individual countries, it would be premature to reflect on potential policy implications. There are simply too many unanswered technical questions. More in-depth analysis and a deeper understanding of the working functions and processes of the LUISETTA model would be essential to explain these differences with increased confidence and precision. At present, the model seems more suited to illustrating possible urbanization patterns than producing reliable statistical output.

4.5 Conclusion

The future urbanization scenarios presented here attempted to extrapolate the logical result of societal attitudes on urbanization, producing maps that display the amount and location of new development in 2050 at a high level of resolution. The aim of these scenarios is not to show which form of urbanization is the most or least favourable but to give an indication of what could happen if all European countries were to opt for a distinct form of urbanization. As such, it can serve as a basis for public debate on preferred policy directions. If one type of urbanization is deemed desirable, one can then discuss the desirability and feasibility of the interventions aligned with this scenario. Again, this analysis should be seen as a source of inspiration rather than an outlook. The choice to hold many variables constant can be questioned in light of recent events: Covid-19, the European Green Deal, the energy crisis, climate change, and geopolitical turmoil. All these may affect societal attitudes and with it the demand for urban space and locational preferences. Nevertheless, at a time when the European Union is trying to provide a common response to crisis and uncertainty, these scenarios can be of great benefit as they predict the direction of possible spatial changes in an immediately understandable way. As such, this can signify the first steps towards drafting a strategy to achieve sustainable urbanization.