Urban vulnerability and the contribution of socio-environmental fragmentation

I Urban vulnerability in the context of climate change

Urbanization and climate change are among the major challenges of ongoing global change. The two phenomena are closely interlinked and show evidence of interrelations with other processes such as demographic, economic and land-use change. It is common knowledge that climate change has a direct effect on the living conditions of countless people, and by no means in developing countries only (IPCC, 2014). Hence the impacts of climate change, culminating for example in flooding, droughts and storms, are neither felt nor distributed equally, and human social interaction with those climate-related hazards calls for more implicit consideration (Pelling, 2003, 2011a). In this context, vulnerability has been widely accepted as the result of socio-economic systems that put people at risk (e.g. Bankoff et al., 2004; Blaikie et al., 1994; Cannon, 2008; Hewitt, 1983; Wisner et al., 2004). This touches on the question of why people are exposed differently to climate-related hazards, acknowledges their susceptibility to suffer harm and considers their coping capacities to address, manage and overcome the negative impacts of climate change.

The aim of this article is to contribute to the evolving concept of ‘urban vulnerability’. It explicitly addresses socio-environmental interaction in cities in the context of climate-related hazards. Firstly, it considers the specifics of ‘urban’ vulnerability and argues that particularly the existing urban vulnerability definitions would benefit from a more systemic perspective on the complexity of urban areas, allowing for identification of the causes of uneven exposure and susceptibility, and of the coping strategies available. When applied in urban areas, the concept of fragmentation is often seen as the opposite of compactness in spatial or morphological terms and thus as the separation of areas into fragments, pieces or parts (Irwin and Bockstael, 2007; Landman, 2011). We argue that combining the concept of fragmentation with that of vulnerability helps to gain further insights into the specific urban characteristic of vulnerability in the context of climate-related hazards. This idea goes beyond more traditional thinking on the analysis of vulnerability in cities. Up until now, research has primarily focused on analysing hazard exposure but has failed to see that the process of urbanization itself shapes susceptibility, coping capacities and local actions (cf. Garschagen and Romero-Lankao, 2013). Our approach acknowledges the twofold effect of urbanization on urban vulnerability – negative and positive – and develops a new ‘socio-environmental’ dimension of fragmentation, interpreting it as a precondition for vulnerability. Secondly, this article operationalizes the theoretical thinking by introducing an approach that combines socio-environmental fragmentation with residential vulnerability, and allows for its practical application.

Against this backdrop, we use the concepts of fragmentation and vulnerability to reveal the interface between fragmented areas and residential vulnerability to climate-related hazards. We operationalize this by focusing specifically on flood and heat hazard in an urban area of Latin America, here Santiago de Chile. Our epistemic research interest is driven by the assumption that urban fragmented areas in terms of environmental and/or socio-spatial elements impact significantly on residential vulnerability to climate-related hazards in the sense that inner-urban fragmentation patterns contribute to a specific urban form of vulnerability. One could also argue that vulnerability is a precondition for fragmentation or that there is at least interactive feedback between the two. Thus taking into account that climate change research has largely concentrated on vulnerability analysis and paid no attention to fragmentation, it seems more adequate to add this aspect to the debate than vice versa. In this respect, the article emphasizes the gap in urban vulnerability research in terms of theory, indicators and conceptualization by considering the following research questions:

To what extent are theoretical considerations of socio-environmental fragmentation and residential vulnerability interlinked and how can they help to gain a more thorough understanding of specific urban vulnerabilities to climate-related hazards?

In answering these questions, we aim to contribute further insights to the nascent concept of urban vulnerability. We follow Birkenholtz (2011), for example, who argues that understanding causality in vulnerability at local level demands an approach that concentrates on local particularity and is capable of extensifying. Once performed, extensive approaches aimed at generalizable insights can draw on local (intensive) knowledge (Birkenholtz, 2011). We adopt this approach, beginning with fragmented areas within the city as a pre-analytical step of the vulnerability assessment. In this way we produce conceptual and generalizable insights on a smaller scale (whole city) and context-specific insights on a larger scale (neighbourhoods and individuals). Theoretical premises of vulnerability are first presented, before developing a specific lens on ‘urban vulnerability’. We then present the fragmentation concept as a methodological pathway for structuring the city. In the following, an approach to address the interrelations between fragmentation and vulnerability is introduced. This is followed by a discussion of the merits achieved in terms of urban vulnerability and then a concluding section.

II Theoretical premises of vulnerability

The debates of several disciplines have made numerous attempts to conceptualize the term ‘vulnerability’ (Hufschmidt, 2011; Kasperson et al., 2005). This notwithstanding, vulnerability as a term remains fuzzy. This is largely due to the epistemological directions and theoretical insights of multiple disciplines, which have led to a range of perspectives on vulnerability and the attendant methodological practices (Eakin and Luers, 2006; Miller et al., 2010; Weichselgartner, 2001). The vulnerability concept has entered into a variety of research contexts, such as ecology, public health, poverty and development (Chambers, 1989; Chambers and Conway, 1992; Moser, 1998; Rodríguez, 2001), secure livelihoods and famine (Bohle et al., 1994; Watts and Bohle, 1993), disaster risk and hazard research (Blaikie et al., 1994; Clark et al., 1998; Cutter, 1996; Hewitt, 1997; Weichselgartner, 2001; Wisner et al., 2004), land-use change, and climate change impacts and adaptation (Downing et al., 2001; Füssel, 2007; Kelly and Adger, 2000). The concept is likewise prominent in physical and human geography and natural hazards research and has been used extensively since the 1990s (Weichselgarnter, 2001).

The development of vulnerability research was shaped profoundly by human ecologists of the Chicago School, where human adjustment of physical systems to natural hazards constituted a core element (Borrows, 1923; White, 1974). Beginning in the 1970s, which saw the question of what makes the human system less vulnerable addressed (e.g. Kates, 1970), the concept became more prominent when social scientists began to question the assumption that the increased incidence of disasters is a purely natural physical phenomenon (Burton et al., 1978; Hewitt, 1983; Lewis, 1982). With the aim of advancing the risk-hazard perspective, human ecology and human geography theorized vulnerability to environmental change by relating the concept to the complexity of socio-ecological systems, interpreting human action and social structures as integral to nature (Burton et al., 1968; Cutter, 1993). As pointed out by Pelling (2002), the ‘humanistic turn’ was then introduced to vulnerability research. Today, vulnerability analysis is used to bridge disaster and development research in an attempt to understand the shaping of human exposure, susceptibility and coping capacity (e.g. Birkmann, 2006; Fuchs et al., 2011; Kuhlicke and Steinführer, 2015).

What distinguishes the use of vulnerability in the various disciplines is how they address vulnerable elements. As stated by Füssel (2007), vulnerability in natural hazard research tends to pursue a descriptive approach and is characterized as the negative outcome of biophysical risk factors combined with the potential for loss of a specific exposed population (Blaikie et al., 1994; Cannon, 1994; Cutter et al., 2003; Eakin and Luers, 2006; White, 1974). The focus lies on the distribution of hazardous conditions, the impact on different social groups living in hazard-prone areas and the degree of loss associated with the occurrence of a particular event (Adger, 2006; Weichselgartner, 2001). Key questions refer to the category of people and assets that are vulnerable, the consequences and where risk situations can occur (Cannon, 1993). Approaches that come from human ecology and political economy, in contrast, emphasize the structural causes of vulnerability in society and concentrate on the analysis of social and economic processes with interacting scales of causation and social difference (Adger, 2006; Bohle et al., 1994; Watts and Bohle, 1993). Combining the internal factors of a vulnerable system (e.g. socio-political, cultural, and economic factors) with its exposure to natural hazards is key. Together they explain differential impacts and abilities to cope with, adapt to and recuperate from past impacts and/or future threats. Important questions in this context refer to who is most vulnerable and why, and to what extent populations are vulnerable in terms of susceptibility and coping capacities (Füssel, 2007).

The Pressure and Release (PAR) Model of Blaikie et al. (1994) is a prominent example of synthesizing the human ecology and political economy perspectives into a more structural perspective on the social, economic, cultural and political context of people’s lives. The model interprets physical hazards as producing pressure on the socio-ecological system. Pressure can occur at different levels, from the global to the local, and impact on vulnerability conditions. At the global level so-called ‘root causes’ related to social, political and economic structures put pressure on the socio-ecological system in terms of, for example, demographic, land-use and economic change. The intermediate level refers to ‘dynamic pressures’ that stem from pollution, population growth, urban development and environmental degradation, while at the local level ‘unsafe conditions’ manifest themselves in social fragility, potential harm or poverty (Blaikie et al., 1994; Cardona, 2004; Weichselgartner, 2001). Although the comprehensive PAR model takes hazard and vulnerability as well as the spatial and temporal remoteness of the factors that render people vulnerable into account, the analyses fail to provide a systemic view of the mechanisms and processes of vulnerability, and operationalizing it is not an easy task (Adger, 2006).

Vulnerability in climate change research is conceived in terms of how climate change either alters the exposure of people and assets to climate-related risks or heralds new forms of risk (Hanson et al., 2011). Vulnerability to climate change cannot, however, be divorced from the social, economic and environmental challenges that, for example, cities face. This acknowledges vulnerability and the capacity to adapt as a product of the processes through which climate risks coalesce ‘with other stresses, such as scarcity of water or governance structures that are inadequate even in the absence of climate change’ (Wilbanks et al., 2007: 373). The definition of vulnerability taken from the IPCC (2007: 6) sees it as the ‘degree to which a system is susceptible to, and unable to cope with, adverse effects of climate change including climate variability and extremes’ and ‘as a function of the character, magnitude and rate of climate change and variation to which a system is exposed, its sensitivity and its adaptive capacity’. This draws attention to vulnerability as a product of the interaction between physical and social processes and to the importance of ‘adaptive capacity’ in determining it. The adaptive capacity (of systems, people, groups) is highly varied in and between urban areas, and shaped by a range of social and physical attributes (Adger et al., 2005).

In the climate change debate, two opposing research focuses have emerged: (1) ‘vulnerability as outcome’ and (2) ‘contextual vulnerability’ (Adger, 2006). Adger and Kelly (1999) conceptualize vulnerability as the outcome of environmental, social, cultural, institutional and economic structures and processes that generate the uneven distribution of exposure, capacities and resources. Here, vulnerability to climate change is characteristic of a specific system and a function of exposure, sensitivity and adaptive capacity (Adger, 2006). Consequently, the aim of reducing outcome vulnerability with climate change mitigation or adaptation is to diminish exposure (O’Brien et al., 2007). Contextual vulnerability, in contrast, takes the multidimensional view of climate-society interaction into account and is influenced by contextual conditions such as dynamic social, economic, political, institutional and technological structures and processes (O’Brien et al., 2007). In this context, vulnerability related to climate change adds to existing vulnerabilities that are structural in nature and often linked to poverty and inequality (e.g. Bull-Kamanga et al., 2003; Heinrichs et al., 2013; ). This ‘contextual’ vulnerability framing sets wider boundaries around the issue of climate change, bringing it into the equity context (cf. Heinrichs et al., 2013). Hence it refers to the interaction of social and environmental systems (Pelling, 2011a) and, according to Newell et al. (2005), represents a conceptual cluster for integrative human environment research. Against this backdrop, reducing vulnerability means ‘adjustments’ to the context in which climate change occurs so that individuals and groups can respond more adequately to changing contextual conditions.

All of the theoretical vulnerability approaches discussed have key parameters in common that refer to the stress to which a system (assets, people) is exposed, its sensitivity, and its coping and adaptive capacity (Miller et al., 2010). Although there appears to be some consensus on these parameters, at least in climate change and disaster risk communities (Pelling, 2011a), how they are framed and contextualized differs. The degree of vulnerability and its characteristics depend on the specific context and are a matter of perception (Weichselgartner and Kelman, 2015). Thus, a comprehensive conceptualization of vulnerability needs to take into account a variety of risks, thresholds and adaptive responses and resources, given that vulnerability manifests itself differently at different scales (Kasperson and Kasperson, 2001).

In the urban context, adaptation and resilience to climate vulnerability are emerging as responses. In the early 1990s, when the international community began to develop policy frameworks, targets and mechanisms to address climate-change mitigation, municipalities likewise sought to establish their own responses to the climate change agenda (Bulkeley and Tuts, 2013). This suggests the historical and contemporary production of urbanization rather than urban climate vulnerability as a discrete phenomenon (Bulkeley and Tuts, 2013). Understanding vulnerability under conditions of climate change in urban areas is therefore not only a matter of enhancing knowledge on the potential risks posed by a changing climate system and their impact, but of engaging with the often historically complex and politically contentious factors that structure vulnerability more broadly, and the complex trajectories of development. At the same time, the socially differentiated occurrence of vulnerability within and between cities needs to be considered (Bulkeley and Tuts, 2013; Button et al., 2013). Set against this backdrop, we explore the structural conditions of urban areas and argue for ‘the urban’ in conceptualizing vulnerability.

III The ‘urban’ in the conceptualization of vulnerability

Our interest in urban vulnerability is primarily driven by the fact that today more than 50 per cent of the world’s population lives in urban areas (UN, 2014), with a tendency to increase, and that urban systems and their individual characteristics are key to high complexity and interdependencies. Processes of climate, land-use and demographic change are interlinked and take place parallel to urban developments of varying degrees. As both culprits and victims of climate change, cities are advancing into the limelight. Climate change is viewed as additional pressure, compounding the ‘bundles’ of stress cities already face. It is also expected that climate change will affect frequency, variability and intensity of extreme weather events (e.g. extremes in temperature, precipitation, wind) and aggravate climate-related hazards, although some parameters will increase, others will decrease and some may not be affected. There is, for example, low confidence that anthropogenic climate change has affected the frequency and magnitude of fluvial floods on a global scale (IPCC, 2014). As a result, vulnerability research is now tackling climate change in cities to a greater extent and calls for innovative methodological approaches and indicators that allow for comprehensive analysis and appropriate responses. Specific research on the vulnerability of cities has therefore become a focal point (Weichselgartner, 2001). This notwithstanding, concrete conceptualization of ‘urban vulnerability’ to climate change remains vague, not least as a result of the diverse interpretations and paradigms of the core attribute at risk – the urban area (Pelling, 2006; Romero Lankao and Qin, 2011). Given the focus on the simultaneity of time and space, as well as of society and environment, questions such as ‘vulnerability of what, to what and on what scale’ are crucial, especially in the field of geography (Weichselgartner and Kelman, 2015). Bearing this in mind, we are particularly interested in the characteristic features of the ‘urban’ in the vulnerability debate. These are not still only vague but also contested in the literature.

Urban areas are seen for the most part as the aggregation of multiple components (Pelling, 2011a) dynamically connected in terms of speed, scale, scope and complexity, each generating interactive effects (Ascher, 1995; Graham and Marvin, 2001; Heinrichs et al., 2012; Moor, 2001; Wirth, 1938). This frustrates a uniform definition of the term ‘urban’ (ARL, 1994), which is the subject of a wide range of discourses, theories and concepts, and often used as a buzz word, albeit without a clear definition. Perhaps the term is assumed to be understood without further clarification or that its definition will emerge in the specific context or as the objective of research. The latter would entail a spatial-temporal contextualization. Hence the ‘urban’ is frequently used as a contrast to the ‘rural’, although the urban-rural divide no longer exists (Häußermann and Siebel, 1997; Hiddinga, 2000; Pahl, 1966). Boundaries are becoming more diffuse and debates centre on the ‘urban hinterland’ and on ‘sub-’, ‘counter-’ and ‘re-urbanization’ (Häußermann and Siebel, 1997). In general terms, urban areas have large settlements, high population densities, mostly non-agricultural activities and high mobility (Heineberg, 2001). They are characterized by physical building structures and the associated patterns in conjunction with environmental conditions (e.g. vegetation, water), climatic patterns and their specific functions.

Depending on the concerns or development goals addressed, different factors are used to describe urban attributes (ARL, 1994). In the same vein we found that Pelling (2002) highlights social, physical, economic, ecological, political, cultural and historical factors as key determinants of vulnerability in cities. A closer look at these factors, particularly in terms of their impact on vulnerability, shows that cities are first of all seen as a social product. The diversity and composition of social structures in cities shapes the access to all kind of resources. Accordingly, lack of social power has an immediate impact on physical vulnerability (Pelling, 2002). Social factors such as age, household structure, economic means, educational and occupational opportunities, the existence of basic social utilities, and the access to property and credit determine the complex nature of risks and opportunities of the urban environment. In the context of climate change and its related hazards, these factors affect the level of support and ability to respond when it comes to hazard occurrence (Cardona, 2004; Cutter et al., 2000; Morrow, 1999).

Secondly, cities are characterized by a concentration of physical assets such as housing, technical, ecological and social infrastructure, transport and communication networks, potable water, sanitation, and drainage (Pelling, 2003), all of which are crucial to the citizens concerned. With regard to the vulnerability of urban inhabitants and urban assets, the existence, accessibility and affordability of basic technical infrastructure, the quality of housing conditions and the geography of opportunity (Galster and Killen, 1995) are vital sources of information in terms of preparedness and resilience, and possible resistance to damage during a hazardous incident (Jean-Baptiste et al., 2011; Müller, 2012).

Cities are, thirdly, seen as an engine of economic growth, representing a financial or business system or an integrated system of consumption and production (Pelling, 2002). In the current form of globalization, cities have become strategic nodes in globalized economic networks and flows of capital (Parnreiter, 2012; Vale and Campanella, 2005). As a result, local economic vulnerability becomes a global concern, since globalization itself is seen as a hazard and cities around the world are moving closer together into a new and hazardous environment (Moor, 2001).

Cities are, fourthly, centres of political power and representation, institutional practice and custom. Urban governance in particular and the role of local or municipal government rule affect the diversity and intensity of sectors, scales, and areas of responsibility. The city’s institutional and political context is therefore seen as the major determinant of the urban system (Eakin and Luers, 2006) and one of the biggest challenges in confronting vulnerability in urban areas (Pelling, 2011a). In this regard, the adaptive capacity of the city to climate change refers to a broad set of resources (skills, competences and social relations), e.g. the degree to which local authorities have integrated (or are in the process of doing so) climate change considerations into their long-term planning and development processes in order to facilitate evacuation activities. This contributes to preventive, mitigation and adaptation initiatives and consequently to the capacity to tackle and lower vulnerabilities (Smit and Wandel, 2006).

Fifthly, cities are consumers of ecological resources such as land/soil, energy, water, air, biodiversity and food. At the same time, they are producers of contaminants such as air pollution, waste water and waste, factors strongly associated with urbanization processes and for the most part the result of changes in land use and land cover, and the growth of populations in cities. The availability of services provided by ecological resources (ecosystem services) impacts on the well-being of urban residents. The loss of important ecosystem functions such as cooling and water retention through destruction of green areas is often directly linked to the increased risk of floods, heat and landslides, which is exacerbated by climate change (Depietri et al., 2012). This in turn points to the issue of vulnerability in cities, notably in the context of climate-related hazards.

The sixth determinant refers to cities as hubs of cultural diversity and resources. According to Bourdieu (1983), the level of cultural capital of urban residents indicates their educational progress and capacity, both based on the management of embodied, objectified and institutionalized cultural capital and its advancement. By making the urban environment attractive, cultural diversity in cities is perceived as stimulating new economic markets and opportunities for development (Florida, 2005), it is a place in which cultural change takes place (Redfield and Singer, 1954). In the context of climate change, cultural diversity is embedded in modes of production, consumption, lifestyles and social organization. Risk identification, response decisions and the means of implementation are mediated by urban culture (Adger et al., 2013). In any case, climate change can transform cultures and communities in ways that people may find undesirable and perceive as the loss of cultural resources (Adger et al., 2013; Kirsch, 2001), and the relationship between people’s vulnerability and the adaptation of its culture still needs further investigation (Bankoff, 2001). In this sense, cultural adaptability can play a key role in determining exposure to risk (Bankoff, 2001), thus also their susceptibility and coping capacity.

Finally, cities are the outcome of the historical aggregation of physical assets seen as a social product (Lippuner, 2005; Löw, 2001; Schroer, 2006; Soja, 2000; Werlen, 1987). Society and space (e.g. urban areas) produce each other and are interrelated. Consequently the production of space has the same societal roots as the making of history and is in fact a historical product (Parnreiter, 2007). Space, on the other hand, cannot be understood without the ‘time’ that generates it and arises from it. Set against this backdrop, cities represent a certain historic configuration of space and/or degree of spatio-temporal reality, which in turn produces a specific urban area, depending on its historical condition (Featherstone, 1999; Schmid, 2005). Cities – and their vulnerability – should therefore be understood as a dynamic entity rather than a static outcome, a container where urban processes take place.

These key components shape the ‘urban’, while their interrelation defines the mode of everyday urban life. This underpins the overall complexity cities face and leads to both risks and opportunities in the urban environment (Heinrichs et al., 2012; Pantuliano et al., 2012), highlighting the view of urban vulnerability as a dynamic social process (Lewis, 1999; Weichselgartner and Kelman, 2015). Part of the challenge of understanding urban vulnerability is exploring the ways in which these components interact in space, time and scale.

Although research on urban vulnerability has gained substantial currency in recent years, it still shows evidence of inconsistency in terms of focus, definition of key terms, methods and the policy implications of its findings (Romero Lankao and Qin, 2011). According to Pelling (2011a), this may be a result of the challenges posed by the holistic view of the ‘urban’. We agree with this statement and argue that the specific urban vulnerability concept would benefit from a more systemic perspective on the complexity of urban areas, allowing for analysis of people’s uneven exposure, susceptibility and coping capacities. Identifying a specific set of vulnerable objects, their complex drivers and their heterogeneity in spatial units of the city is thus fundamental to recognizing patterns of urban vulnerability (Pelling, 2002).

Having screened several theories in human geography, we consider the concept of fragmentation as a methodological pathway to operationalizing the ‘urban’ in the vulnerability concept, according to the urban specifics stated before. We thereby acknowledge that fragmentation structures the urban environment into heterogeneous or homogeneous fragments along a set of predefined elements, in our case socio-environmental aspects that are key to urban areas. We recognize that urbanization and the numerous processes involved could have a positive effect on risk reduction and accordingly on urban vulnerability (Garschagen and Romero-Lankao, 2013). Up to now, however, the specifics of the ‘urban’ have rarely been investigated. Nor has there been an attempt to combine the concept of vulnerability with that of fragmentation.

IV Fragmentation as a methodological pathway to structuring cities

Analysing urban spatial structure as an integrative socio-ecological system first emerged as a research strand of human geography and urban sociology. It examines urban morphology and the cultural, economic, social and political processes behind its development (Kuhlke, 2007). Fragmentation has its roots in landscape ecology, where fragmentation relates to morphological and often spatial or structural properties of patches in terms of habitat composition (Inostroza et al., 2013). Here, fragmentation has the negative connotation of population decline and lower densities as a result of the ineffective distribution of individuals across habitat networks, especially when human land use is the cause (Li et al., 2010; Opdam and Wascher, 2004). A prominent example of structuring the urban environment in spatial terms is the work of the Chicago School of Social Ecology in the 1920s. Two of its representatives, Park (1915) and Burgess (1926), for example, highlighted the dynamic spatial arrangement of homogeneous urban structures, characterizing it in terms of density as well as sociological and functional features. They detected patterns of social differentiation in cities and offered explanations for the underlying processes, but were widely criticized. The Chicago approach was refined within the framework of neo-classical economy (e.g. Alonso, 1975), behavioural theory (Brown and Moore, 1970; Roseman, 1971; Wolpert, 1965), institutional analysis (Pahl, 1977; Rex and Moore, 1967), and political economy (Castells, 1977; Harvey, 1973; Marcuse, 1989).

Theoretical considerations led to various lines of approaching the structural-spatial analysis of urban areas, based on such criteria as ‘social area analysis’ (e.g. Bähr, 1978; Friedrichs, 1983; Shevky and Bell, 1955), ‘functional structuring of the city’ (e.g. Bähr and Mertins, 1995; Borsdorf, 1982) and ‘ecological structuring of the city’ (e.g. Breuste, 1989). Social area analysis, as an early concept, adopted a statistical procedure to depict social structure in terms of social rank or economic status, urbanization or family status, segregation or ethnic status. It was applied to several contexts but also contested, particularly in terms of a precise definition of social area (Hawley and Duncan, 1957). Urban morphological structures were analysed, for example, in Latin American cities. Models featuring the right-angled form of the colonial city and its structural fragmentation were designed (Borsdorf et al., 2002). Satellite images and aerial photographs entered the research arena and were used for area-wide classifications of land use and land cover, according to the spatial, spectral and temporal resolutions of the images. Biotope type mapping is slightly different, as it classifies (urban areas) into habitats or biotopes in line with their importance for the balance of nature and is used, for example, as a data base for nature conservation and sustainable urban planning (Sukopp and Weiler 1988; Wittig et al., 1998).

Despite these and other approaches, urban fragmentation is the concept that has attracted most attention in international urban studies since the 1990s (see, e.g., Graham and Marvin, 2001; Soja, 2000). The fragmentation of urban areas is perceived as an urban phenomenon that impacts heavily on contemporary cities, notably in Latin America (Jirón and Mansilla, 2014; Prévot-Shapira, 2001). The multidimensional concept seeks to address ruptures in the comprehensive structure of the city (Caprón and Gonzalez, 2006), uncovering segmented urban territories that generate spatial discontinuities (Jirón and Mansilla, 2014). It distinguishes between ‘city of fragments’ and ‘fragmented city’. The former refers to a constructive process of incorporating separate structural fragments, while the latter describes the process of disintegration of socio-spatial fragments within the city itself (Borsdorf, 2003). The concept contains a highly diverse range of values regarding the connection between different parts of the city and certain isolated fragments. Since fragmentation alludes to the structure and functioning of the city as an unequal setting, it is often used as a synonym for terms such as ‘spatial segregation’, ‘spatial separation’, ‘spatial polarization’, ‘social exclusion’ and ‘disconnected cities’ (Borsdorf, 2003; Hidalgo and Borsdorf, 2005; Link, 2008; Prévot-Shapira and Cattaneo, 2008; Trufello and Hidalgo, 2015). Given the difficulty of clearly identifying which of these features contributes to the different forms of urban fragmentation and which are the result of it (Landman, 2011), fragmentation has the reputation of being a ‘slippery concept’ (Harrison, 2003: 15). Hence, we need to ask – not least with respect to the vulnerability debate – what forms of fragmentation are important, how much they matter and why. As Landman (2011) states, analysis of urban fragmentation calls for a thorough understanding of the internal lines of influence and the attendant relationships, and therefore a systemic approach.

The literature discusses six complementary dimensions of fragmentation in cities: morphological, social, economic, socio-cultural, political-administrative, and environmental. Each can be related to key determinants of the ‘urban’ environment, presented in Section III. The morphological dimension of fragmentation refers to ‘urban fabric’ and ‘building morphology’ features and describes the physical dimensions of the urban. It explores, for instance, issues of physical accessibility and connectivity between territories, the presence of physical barriers and homogeneity of land use (Rodríguez and Winchester, 2004; Salinas, 2010; Salingaros, 2005; Tella, 2005). The economic dimension examines the spatial distribution of economic ‘goods’ such as economic command centres, the location of labour, the dispersion and relocation of industry and the functional specialization of districts (Janoschka, 2002; Rodríguez and Winchester, 2004; Sánchez, 2007). The socio-cultural dimension includes both physical and symbolic aspects, and considers elements such as social inequality, symbolic barriers, as well as levels of social cohesion and victimization (Dammert, 2004; Link, 2008; Prévot-Shapira, 2001; Sánchez, 2007; Schteingart, 2001; Veiga, 2004; Villela et al., 2010). The diversity of administrative units and their potential for articulation within an urban management scheme connects to the political-administrative dimension, which reflects on issues of social organization, such as community boards, as well as on variables of metropolitan urban governance (Marcuse, 1989; Michelutti, 2010; Rodríguez and Winchester, 2004; Sánchez, 2007; Salinas, 2010). It also refers to independent local jurisdictions, which are rarely coordinated and thus proceed in isolation (Harrison and Hoyler, 2014). The environmental dimension looks at the distribution of ‘ecological services’, green spaces and public spaces. This implies ‘environmental’ fragmentation of the city, understood as access to and availability of these components. It refers to the constitution of borders and the territorial limits of environmental elements within an urban area (Bizama et al., 2011). Finally, the historical dimension – a key feature of the ‘urban’ – is rarely taken into account in concepts of fragmentation, revealing their lack of temporality and giving cities the appearance of being atemporal in nature. Jirón and Mansilla (2014) claim that the absence of a temporal dimension is the principal weakness of the fragmentation concept. In the context of urban studies, authors have referred to the idea of fragmentation in at least one of these dimensions, making it difficult to formulate a general and operational definition of the concept (Borsdorf, 2003; Link, 2008; Low, 2006; Michelutti, 2010; Vidal, 1999).

V Unravelling the theoretical linkages of fragmentation and vulnerability in cities – How can they be operationalized?

This section unravels an approach to operationalize our theoretical interlinkages between spatial-structural fragmentation and residential vulnerability. As a case in hand we take a contemporary city in Latin America affected by flood and heat hazards – Santiago de Chile. The main link between the concepts we worked out in the previous section is the ‘urban’ dimension and its specifics. Given the definition of fragmentation that comprises the organization of a city’s environmental and/or socio-spatial elements by differentiating and disintegrating space into internally homogeneous or heterogeneous spatial fragments in line with selected characteristics (Link, 2010), we argue that combining it with the concept of vulnerability gives us an added tool with which to gain further insights into the urban characteristic of vulnerability in the context of climate-related hazards. We thus assume that the spatial-structural homogeneity or heterogeneity of a specific urban area is a precondition – whether positive or negative remains to be seen – for the occurrence of hazards and the vulnerability of populations to them. This idea goes beyond the more traditional vulnerability thinking, which has so far focused on analysing exposure to hazards in cities, and for the most part neglecting that the process of urbanization itself shapes susceptibility, coping capacities and local actions (cf. Garschagen and Romero-Lankao, 2013). Acknowledging that there is no one correct or best conceptualization of vulnerability, our approach (see Figure 1) follows Füssel’s (2007) four dimensions vital to the description of a vulnerable situation: (a) the system of analysis; (b) the hazard with a potentially damaging influence on the system of analysis; (c) the exposed attribute(s) of concern that is/are threatened by exposure to hazard; and finally, (d) the temporal reference alluding to the point in time or time period of interest. Furthermore, we argue that the specifics of ‘urban vulnerability’ can be enhanced by integration of the fragmentation concept, thereby acknowledging the complexity of urban processes. This allows for joint analysis and the superposing of urban characteristics to achieve a complementary interpretation of the specific aspects of ‘urban vulnerability’.

OPEN IN VIEWER

Figure 1.

Approach to combining fragmentation and vulnerability in the urban area. Source: authors, following Füssel (2007). (a) represents the system of analysis – cities and their level of fragmentation, (b) indicates the hazard that is external to the system – flood and heat, (c) relates to the exposed attribute of concern – vulnerability assessment of socio-environmentally fragmented areas, and the arrows point to the temporal reference, alluding to the point in time or time period of interest.

3 The exposed attribute of concern – Vulnerability assessment of socio-environmentally fragmented areas

Coming back to the overall purpose of assessing urban vulnerability and establishing who is vulnerable to climate change in a city and why, the exposed attributes of concern need to be defined. We therefore focus on assessing the many reasons why certain individuals or communities within a socio-environmentally (non-)fragmented urban area are vulnerable and argue that this demands a thorough evaluation of the key dimensions – exposure, susceptibility and coping capacity (see Figure 1). Following Wisner et al. (2004: 11), we understand vulnerability as the ‘characteristics of a person or group and their situation that influence their capacity to anticipate, cope with, resist and recover from the impact of a natural hazard’. This refers to a contextual perspective on vulnerability.

We follow an inductive approach to operationalizing vulnerability. Field data derived from field mapping, household surveys and expert interviews plays a major role, since data availability can be a constraint. We agree in this context with Pelling (2006), who argues that social surveys are highly useful in understanding society-wide vulnerabilities to vast threats and particularly vital where the respective hazard has not yet been experienced by the community. Furthermore, this approach acknowledges that vulnerability cannot simply be reduced to a single quantifiable metric and that translating a complex set of parameters into a quantitative metric in many ways reduces its impact and hides its complexity (Alwang et al., 2001). The assessment itself is based on a wide range of vulnerability indicators that describe the three key dimensions. It takes into account that working with indicators is common practice in vulnerability assessments but remains challenging because it calls for unambiguous definition of the vulnerability ‘entity’, transparency of the methodologies used to develop vulnerability indicators, and appropriateness at the local scale, where the entity can be clearly defined (cf. Hinkel, 2011). In line with these challenges, our vulnerability entity consists of households located in hotspots predefined by fragmentation and hazard exposure. For application in these hotspots, we suggest overall indices that include the weighting of variables in each vulnerability indicator along an axis of low to high exposure, susceptibility and coping capacity. The indicators applied are presented in Figure 2. Variable weighting is based on a review of the literature and expert knowledge, and reflects assumptions made. Normalization of each indicator according to the respective accumulated frequencies (ordinal scale) or equal points (nominal scale) represents the degrees of exposure, susceptibility and coping capacity. Applying this approach assumes that each indicator is of equal value for the overall vulnerability assessment. The result is a single vulnerability score for each household, which can be mapped spatially and used as a basis for discussion of the necessary interventions to reduce vulnerability (cf. Krellenberg and Welz, 2016), e.g. in terms of adaptation measures.

OPEN IN VIEWER

Figure 2.

Vulnerability assessment – components, indicators and methodology. Source: authors.

VI Discussion

The overall objective of this paper was to detect theoretical and methodological pathways of ‘urban vulnerability’, still a recent field in the literature. Guided by our first research question we emphasized the theoretical context of vulnerability and fragmentation by focusing on one common dimension – the ‘urban’.

Applying the concept of socio-environmental fragmentation to a city allows for a better understanding of how separate urban fragments are characterized in terms of socio-environmental conditions, a crucial aspect when considering human interaction with the environment in the context of climate-related hazards. This is particularly important in the context of hazards such as flood and heat and likely to be a precondition for the uneven occurrence and distribution of hazards in a city. The analysis of socio-environmentally fragmented city areas alone, however, does not allow for conclusions as to whether the respective urban conditions affect vulnerability to the impacts of climate change positively or negatively. We therefore suggest an in-depth vulnerability assessment in these urban areas with pre-analysed socio-environmental conditions. The joint fragmentation and vulnerability assessments will contribute substantially to the understanding of urban vulnerability. The presented approach underpins the multidimensional character of cities and should be seen as a step forward in clarifying the specific meaning of the ‘urban’ in vulnerability research. This aspect distinguishes it from other approaches that, in our understanding, look at ‘vulnerabilities in cities’ rather than at the specifics of ‘urban vulnerability’. Hence we argue with a new and more precise definition of urban vulnerability in the context of climate-related hazards.

For us, urban vulnerability in the context of climate-related hazards thought to be aggravated by climate change, continuous land-use change and other ongoing pressures, is predefined by the socio-environmental conditions of a fragmented urban area, since this is considered a precondition for the degree of exposure and susceptibility of populations to a specific natural hazard at a certain point of time and their ability to cope with the consequences. We understand urban vulnerability as the interrelation between the spatial-structural (socio-environmental) conditions of an urban area exposed to a discrete and identifiable event in nature (or in society) with the underlying susceptibility and response profile of its inhabitants.

This definition includes two main aspects inherent to both the fragmentation and the vulnerability concept: 1) spatial connectedness and 2) the accessibility of urban assets. The spatial discontinuity of an urban area and its disintegration are characteristic features when it comes to determining an urban fragment. At the same time, this spatial disconnectedness may impact on the individual access (in terms of distance, time or financial means) to goods, services and resources crucial to coping before, during and after a hazardous event. These aspects are particularly evident when defining a specific socio-environmental dimension that goes beyond traditional fragmentation characteristics. This new dimension acknowledges the interlinkage between climate change, land cover altered by land use, and socio-economic characteristics such as density, accessibility to basic services and physical housing conditions. In contrast, the residential vulnerability analysis along the dimensions of exposure, susceptibility and coping capacities that considers specific information on housing and people does not allow for conclusions on the structure of a city. Insights into both aspects calls for a more thorough observation of the urban structure and an assessment of residential vulnerabilities, a procedure that enables the two concepts to complement each other. The latter considers that vulnerability and fragmentation both have specific urban implications, although so far the fragmentation concept has not been addressed by vulnerability studies in cities or vice versa.

Based on the theoretical pathway developed in this article, we provided an approach for operationalizing our understanding of ‘urban vulnerability’. Here we were guided by the second research question of how to operationalize fragmentation and vulnerability in the interests of a more systemic perspective on the complexity of urban areas in a holistic approach. This assessment approach includes in a first step a socio-environmental fragmentation analysis of a city with a deductive approach that differentiates the city in urban fragments along specific socio-environmental conditions. This approach is innovative per se as it uses a set of variables that combines social, economic, morphological and environmental aspects, rather than working with a single dimension or, as was often the case in fragmentation analyses of urban areas, solely with traditional socio-economic variables. The second step in this methodological approach includes a definition of the hazard in place. It allows for the selection of hotspot areas that are a) characterized by certain socio-environmental conditions, and b) potentially exposed in the case of a hazardous flood and/or heat event. Hence, unlike many other vulnerability analyses, the starting point of this methodological approach is not the unequal access to resources and distribution of risks, nor does it take as read that poorer people are more vulnerable (Garschagen and Romero-Lankao, 2013). Instead, we take socio-environmental fragmentation and spatial hazard occurrence as the point of departure for our analysis. The hotspot areas are analysed further in light of the residential susceptibility to suffer harm when exposed to hazard and the capacity to anticipate, cope with, resist and recover from the impact of the natural hazard along a place-based set of indicators. Data analysis includes the weighting and combining of selected indicators into composed indices with empirical data obtained by techniques such as field mapping or household surveys. Application of this methodology allows for evaluation of the prevailing degree of vulnerability to climate-related hazards in selected areas of socio-environmental fragmentation.

The proposed approach poses nonetheless a number of challenges. First of all, working with different data sources such as general statistics, field mapping and household surveys is challenged by the issue of scale when methodologically combined in a comprehensive manner. The approach considers three different spatial scales: household, neighbourhood and municipal. Whereas (deductive) fragmentation analysis is linked to the neighbourhood scale across the city, a comprehensive (inductive) vulnerability assessment is only feasible on block scale across selected neighbourhoods due to time and resource constraints. As a result, the inductive approach to vulnerability shows a first approximation of the complex nature of the risks, opportunities and capacities of individuals and households affected by natural hazards.

In order to operationalize both concepts – fragmentation and vulnerability – a context-specific set of indicators was proposed and tested. The clustering, aggregating and weighting of indicators was based on a review of the literature and expert knowledge, and reflects assumptions made. Fragmentation and vulnerability were translated into quantitative metrics and obtained generalized results by reducing impact and concealing complexity (see also Alwang et al., 2001). Consequently, indicator verification proved to be a highly useful pathway to a better understanding of the suitability of the selected indicators (Eriksen and Kelly, 2007; Weiland et al., 2011).

To sum up, the approach presented here is both innovative and comprehensive, and its practical application, feasible. Given the wide range of methodologies and analytical steps involved, it will be crucial to test to what extent it can be transferred to other cities. We argue that while its potential transferability is high, much depends on the availability of existing data bases. If more up-to-date census data were available in other cities, the analysis itself would be less time-consuming and field mapping and household surveys less important.

VII Conclusions

This article considered theoretical considerations of fragmentation and vulnerability and worked out their interlinkages with a view to advancing our understanding of specific urban vulnerabilities to climate-related hazards. It presented a conceptual framework and discussed benefits, challenges and limitations of the proposed methodological approach. We acknowledged the twofold effect of urbanization on vulnerability – negative and positive – by introducing a ‘socio-environmental’ dimension to fragmentation. Here we argued that the fragmentation concept allows us to analyse an entire city and structure it into spatial socio-environmental fragments. This wealth of overarching information must, however, be reduced again, since the vulnerability assessment, with its in-depth information on people’s perceptions, can only be undertaken on a large scale as a result of data assessment constraints in terms of time and resources.

Analysis of the temporal dynamics involved is only partly included in the approach in the form of dynamic vulnerability indicators. In this sense, the approach has its limits, primarily in assessing the dynamics of fragmentation and vulnerability. Enhancement of the approach in this respect would also contribute to a more thorough understanding of how adaptation to climate change can help to reduce vulnerability over time. Looking to a greater extent at less visible roots of vulnerability, such as social, cultural, economic and political factors that often overlap and interact (Pelling, 2011b) can further strengthen the urban vulnerability approach. It is particularly the temporal dimension of vulnerability that is crucial in order to understand the underlying processes of fundamental changes in urban areas, being itself an important driver of change (see Kabisch and Kuhlicke, 2014). This follows O’Brien (2012), who argues that a broader and more holistic approach to adaptation involves viewing the vulnerability context from different spatial and temporal perspectives. The next vital step and possibly the most challenging would be to enlarge the approach to the temporal dimension and test its applicability or transferability to other cities.