Sociologie Românească ENVIRONMENTAL INJUSTICES IN EUROPEAN CLIMATE HOTSPOTS

: This paper starts from the observation that development projects that create various forms of environmental injustice in Europe are an integral part of the process of biospheric expulsions, that is of pushing out groups from adequate land, water or air, as described by Saskia Sassen (2014). Apart from the environmental, socio-economic and health-impacts of ecologically destructive projects, there is an added dimension of concern that has been less obvious in the past, but tends to become increasingly pronounced in a warming world. Is it possible that accumulating environmental inequalities and forms of injustice can create new and “unnatural” vulnerabilities to the projected climate change impacts? The ﬁ rst question that we tackle is whether environmental justice conﬂ icts in Europe tend to take place disproportionately in climate hotspot areas, which are geographic spaces with above-average social sensitivity, potential vulnerability, potential social impact, potential environmental impacts or response capacity (ESPON, 2011). The second question concerns the distribution of diﬀ erent characteristics of projects and of their associated conﬂ icts in climate hotspot vs. non-hotspot areas. The ﬁ nal goal is to establish, at a preliminary level, the emergence of a climate edge in Europe, a spatial conﬁ guration in which vulnerability to climate change impacts is shaped by processes of biospheric expulsion, as postulated at a general level by Sassen. For the analysis, the most current data on environmental justice conﬂ icts (444) from the Environmental Justice Atlas and ESPON climate impact projections, mapped on the Climate Adapt platform, are used. The expected result is to provide a preliminary description of the postulated climate edge.


Introduction
The Merriam Webster dictionary defi nes the edge as "a line or line segment that is the intersection of two plane faces", while the Free Dictionary also adds the meaning of "the point at which something is likely to begin".
The overall aim of this paper is to establish the climate edge as an empirically grounded concept to explore current processes of vulnerability creation to the expected impacts of climate change in Europe. The climate edge integrates existing notions of climate injustice -whereby class, race, ethnic and gender dimensions shape the distribution of climate risks -but reconstructs them through an explanation of how vulnerability emerges under the current expansion of a particular kind of political economy, that of resource-intensive development projects. The climate edge is the space where vulnerability is shaped by the complementary processes of social expulsion and resistance. In this paper, development projects refer to opencast coal mines, shale gas exploration and extraction, infrastructure expansion projects, tourism infrastructure, urban confl icts and similar interventions that create environmentally unequal outcomes. Equally importantly, development projects are engulfed in controversies, since they appear to threaten rights to health and a clean environment. Noteworthy cases are the shale gas development proposals in the UK or Poland (Lis and Stankiewicz, 2017;Nyberg, Wright and Kirk, 2018) or the continued expansion of open cast mines in Germany (Brock and Du nlap, 2018).
The climate edge is a new concept that builds on two recent research trends. The fi rst is the political economic theory of expulsions and of the systemic edge, advanced by Saskia Sassen (2015Sassen ( , 2014. The systemic edge is the point where a condition takes a form that is so extreme that it fails to be recognisable by governments and experts and hence becomes incomprehensible (Sassen, 2017). In the present context, development projects are seen to generate a host of interrelated impacts on local ecosystems, livelihoods, families and citizenship rights, for which there is no one overarching system that takes all of them into account. Their recognition is further hampered by the framing of some development projects as "strategic" for development, in terms of energy security or key natural resources (e.g. coal or gold) (Cehlár et al., 2016;Dodge, 2016), which makes their side eff ects less salient. A third reason for the invisibility of impacts is their apparent local nature, which may conceal their frequent occurrence and occasional regional concentrations. For example, the Environmental Justice Atlas shows 455 confl icts in Europe with distinct concentrations in Southern and Eastern Europe (Temper, del Bene andMartinez-Alier, 2015 but updated until 2019).
The climate edge describes the space beyond the systemic edge, namely the uncharted territory where expelled individuals and communities are made to feel the dangers of a changing climate. In contrast to Sassen's systemic edge, however, the climate edge also captures the forms of resistance that have sprung up among the disenfranchised to counter expulsions (Martinez-Alier et al., 2016). The second research trend is, thus, the remarkable growth of environmental and climate justice research worldwide, and especially in Europe, where it enjoyed little visibility until not long ago (Elvers et al., 2008). Both trends converge in defi ning a new space of climate justice confl icts which seem more striking, because they are more recent in European societies than elsewhere.
The twofold aim of this article is to use the concept of climate edge to explore these aspects of climate change. The fi rst is to explore to what extent development projects that generate resistance tend to take place in areas that are already aff ected by multiple environmental impacts that have, in turn, social repercussions. In this sense, the climate edge is the line of intersection between the accumulation of environmental injustice -as engendered by development projects in Europe -and the advancing edge of climate vulnerability, which is revealed by the maps of projected climate impacts across Europe (European Commission, 2009). The second aim is to show that the climate edge is not a deterministic, metahistorical process but one that is shaped by diff erent groups of actors resisting the encroachment of development on their lands.

Theoretical framework
The article is based on three streams of social scientifi c literature, which are detailed below. The fi rst concerns social vulnerability to climate change, the second introduces the concept of expulsions and the systemic edge, while the third discusses current environmental justice movements.

Social Vulnerability to Climate Change
The recognition of vulnerability to climate change and the importance of adaptation have undergone a spectacular increase in the scientifi c and policy arenas over the last three decades. The International Panel on Climate Change (IPCC) stated in 2001 that due to the role of vulnerability and its causes in producing impacts on society, it is as essential to understand the dynamics of vulnerability as it is to understand the climate system itself (Smit and Pilifosova, 2003). Vulnerability is unequally distributed in society, depending on the socio-economic and cultural conditions of diff erent groups exposed to climate change impacts. Where such conditions are biased against certain groups or individuals, such as ethnic minorities or indigenous groups, there is a problem of environmental or climate injustice (Pellow and Brulle, 2005). More generally, a sociological perspective on vulnerability fl eshes out the social conditions underlying vulnerability, the choices available to diff erent actors in given situations and the social decisions that shape individual or collective climate change adaptation eff orts (Carmin et al., 2015).
However, most research on adaptation employs a scientifi c framing in which the changing climate system generates impacts on society, seen as a steady-state system, and the impacts need to be assessed to devise technical and managerial adaptation measures (Bassett and Fogelman, 2013). The alternative model is much less visible in the scientifi c or policy arenas, albeit it can off er a more realistic and applicable perspective: it centres on human security and postulates that adaptation will be successful if individuals have options to mitigate or adapt to climate risks and can make use of them (O'Brien et al., 2007). The climate edge advances the human security framing by revealing where the limits of human security lie under conditions of expulsion from the societal safety nets.
Recent sociological research on vulnerability has underscored the role of human agency, in particular the ability of weaker groups to mobilize in order to reduce their vulnerability by securing access to resources and entitlements (Bassett and Fogelman, 2013). According to Wisner and his colleagues (2014), the social processes creating vulnerability are recognized and opposed by people who sometimes have the capacities to resist, avoid or mitigate the sources of their vulnerability. The paper aims at a conceptual and empirical breakthrough by showing that people do not react to some abstract processes taking place away from their everyday experiences but to concrete expulsions that threaten their immediate life spaces.

The Age of Expulsions and the Systemic Edge
Sociological work on the environmental implications of the political economy of advanced capitalist societies has a long and distinguished career (Gould, Pellow and Schnaiberg, 2004;Schnaiberg, 1980;Schnaiberg and Gould, 2000). Schnaiberg explored the distributional impacts of environmental policies, in particular their eff ects on the lower classes (1975). Leading scholars of political economy have recently taken up the topic of ecological and climate crisis in discussing the future of capitalism (Collins, 2013). This paper builds on the recent work of sociologist Saskia Sassen (2014Sassen ( , 2015Sassen ( , 2017, in particular on her related concepts of expulsions and the systemic edge. Expulsions are an accelerating and acute process of growing social inequalities which push the poor and middle classes out of their living spaces and the rich out of the responsibilities of societal membership (Sassen, 2014). The systemic edge is the space where a condition takes on a form that is so extreme that it fails to be visible to the standard ways of knowing and measuring available to governments and experts and hence becomes ungraspable (Sassen, 2015).
This paper applies the notion of expulsions to account for the profound reshaping of social and environmental relationships in Europe through the expansion of intrusive forms of environmental exploitation. For example, the recent boom in unconventional hydrocarbon exploration in a dozen EU member states (JRC, 2017;Stephenson, 2016) has added to earlier widescale extraction projects. For example, despite its well-known transition towards renewable energy, German companies still operate and enlarge their opencast lignite mines, obliterating villages and destroying farms and forests in their path (Weber and Cabras, 2017).
Expulsions are created through chains of political, economic and technological factors that lead to what Sassen calls brutalities, which are the sudden, intense and socially meaningless eviction of people from their life spaces in the name of shortlived and highly risky industrial activities. The expulsion concept encapsulates this whole chain, which often originates in decisions taken at transnational or intergovernmental levels (e.g. the energy strategies of the EU, such as its "Projects of Common Interest" (EJAtlas, 2018)) and make their way unabated to the lowest scale of individuals' households or bodies.
The systemic edge is the result of this process of expulsion that is a new space in which the taken-for-granted values and norms of social interaction are profoundly reshaped. In the spaces of the expelled created by development projects, for example, properties are devalued, means of livelihood are obliterated, and life projects are rendered unattainable. The systemic edge is a space of extremes -of social destitution, political disenfranchisement or cultural disorganization -that appears as a hole in the tissue of societal solidarity. Questions of what is worth recognizing or defending, for instance, in the face impeding climate impacts, become muted, as these spaces descend to the status of "sacrifi ce zones" (Harlan et al., 2015).
The loosening of societal safeguards -structural, institutional and societal (Carmin et al., 2015) -against the eff ects of a changing climate is especially problematic in those places where signifi cant impacts are expected, and this happens on the development projects frontier, as in selection of case studies presented in this article. We use such frontiers as entry points for investigating the broader ramifi cations of vulnerability-making expulsions. However, the stories from the climate edge are not solely ones of ineluctable loss. The occasionally over-deterministic tone of Sassen's argument needs to be brought into productive dialogue with the burgeoning literature on environmental justice activism.

Environmental Justice Movements
The topic of environmental injustice has seen increasing scholarly attention over the last few decades. Interestingly, this perspective has been co-constituted with the increasingly infl uential environmental justice movement (Martinez-Alier et al., 2016). The famous case of contamination in the Love Canal residential area of Niagara Falls, US, started a movement against "dumping toxins" on the poor (Bullard, 1990). Only a handful of studies on environmental justice were written throughout the 1970s. Their number increased to 999 in the decade before the new millennium and to as many as 7400 since 2011 (Google scholar). The unequal exposure to environmental harms highlighted the plight experienced by the urban poor and communities of colour (Burch, 1971). Since then, environmental justice has become an established scholarly fi eld in the United States.
The fi eld has gradually grown through geographic expansion and theoretical enrichment. The expansion has taken place in the Global South, where problems of contamination have been accompanied by those of massive resource extraction (e.g. mines or hydroelectric dams) (Agyeman, Bullard and Evans, 2003). This literature is highly relevant for this paper as it provides criteria for expulsion processes at the local or community level, for example through displacement, loss of livelihood or disenfranchisement (Cernea and Mathur, 2008;Oliver-Smith, 2011). There is also a newer strand of the environmental justice literature that this paper also links to expulsions.
We claim that the precise outcomes of expulsion are blurred by an increasingly powerful resistance movement against energy or resource extraction. Impoverishment and other deleterious eff ects do not occur automatically. This has led to the emergence of a transnational confl ict zone famously called "Blockadia" (from "blocking" pipelines or mines) (Klein, 2014). There are suffi cient examples of resistance to development projects world-wide (Temper et al., 2015), some of which are interlinked, to justify its description as a global process. The struggle is waged by coalitions of local and international environmental justice NGOs, subsistence farmers, fi shers, hunters, herders, artisanal miners or all those who, according to Sassen, are pushed on the edge. The struggle of resistance movements is framed in relation to distinct parts of the expulsion chain (e.g. policy decisions, particular technologies, new industrial practices). Activists do often develop scale frames to address this multi-scalar process of expulsion (Kurtz, 2003). In contrast to the systemic edge, the climate edge is characterized by an expulsion -response nexus, rather than by mere structural expulsion.
At the empirical level, this paper is made possible by the trailblazing contribution to map environmental justice confl icts worldwide through the EJA, an atlas developed by the EJOLT project (2011) and continued by ENVJUSTICE (2016). The Academic-Activist Co-Produced Knowledge for Environmental Justice (Acknowl-EJ, 2016) is a project-cum-network that explores the transformative potential of citizen mobilizations, and their eff orts to develop alternative knowledge bases for advancing environmental justice.
In sum, development projects are seen to change the vulnerability of various marginalised groups. This can happen in two complementary ways. The fi rst is the process of accumulating environmental injustice that progressively make selected groups vulnerable and "prepare the ground" for future climate impacts on these groups. The second is a complementary and possibly opposed process, in which local mobilizations and social movement activity can strengthen the capacity of local communities to resist development projects and also possibly their capacity to work collectively to better adapt to projected climate impacts. Sassen (2014) documents how a variety of resource extraction or manufacturing industries (mines, smelters, plantations) have created "dead land and dead water." She defi nes these as "holes in the tissue of the biosphere, [as] sites marked by the expulsion of biospheric elements from their life space". Whereas Sassen's argument is based on several textbook cases of massive environmental destruction worldwide, we need to construct a micro-level equivalent of expulsions, which can be discerned at a European level. Fortunately, during the last decade, environmental justice issues have been increasingly recognized in Europe (Laurent, 2011). In support of the argument developed here, Elvers, Gross and Heinrichs (2008, 836) point to the important role of "localized sites of sometimes rapid change". It is precisely such sites that this article uses to illustrate the piecemeal social and environmental impacts that lead to a broader puzzle of vulnerability creation.
On the one hand, development projects weaken aff ected groups by polluting and transforming their local ecosystems, by aff ecting their health, by undermining local economies and livelihoods, and creating or entrenching social insecurity. On the other hand, the grassroots mobilisation against fossil fuel projects can mediate the vulnerability of those most at risk, by enhancing their adaptive capacity. This happens through collective action: once actors can mobilize to resist development projects, they learn how to marshal individual and collective resources to advance community goals, including a superior level of adaptation.

Methods and Data Collection
The research uses secondary data to provide some preliminary empirical evidence for the climate edge concept. Providing such evidence amounts to identifying circumstances or cases in which environmentally destructive projects take place in communities that are simultaneously under threats of climate change impacts. We do not assume a causal relationship between expanding development projects and their location in climate sensitive areas, but rather highlight the overlap of the two processes in specifi c areas. Such areas are the fi rst indication of an emerging climate edge where vulnerabilities to climate change are systematically produced.
To account for these overlaps, two data sources are used: one is the Environmental Justice Atlas (henceforth EJAtlas), which is a contributor-based and publicly available platform of environmental justice confl icts worldwide. The second data source is the European Climate Adaptation Platform (Climate-ADAPT), in particular its climate map viewer facility. The overall approach is to explore the overlaps between these two maps, as shown in Figure 1. The EJAtlas consists of a zoomable map of the world on which environmental justice confl icts are displayed. The cases are organized by diff erent criteria, such as category (nuclear, fossil fuels and climate change, etc.), region or country, characteristics of the confl ict, impacts and outcomes. Moreover, the EJAtlas features a fi ltering function, which allows the selection of environmental justice confl icts by individual criteria or complex combinations of criteria, using logical connectors ("and", "or" and "not"). For the purposes of this article, we selected cases by "region", covering cases in North-Western, Southern and Eastern Europe, totalling 436 entries. The EJAtlas contains case studies that are compiled by volunteer contributors, be they environmental activists or scholars. The information from these case studies are not validated by offi cial expert systems, such as governmental agencies or ministries. For example, the EJAtlas does not contain only formal data on development projects but also documents their histories, confl ict dynamics and visible/potential impacts. Since the EJAtlas is under continuous expansion, we limited the data collection to the middle of 2018. Further refi nements of the number of cases had to be carried out due to the features of the second data source, the Climate-ADAPT platform, which is presented below.
This climate data platform includes several layers of climate data, stemming from diff erent research projects (e.g. ENSEMBLES, ClimWatAdapt, etc.). For our purposes, we used information from the ESPON Climate project, which collected data at the EU-level in 2009. One limitation of this data is that it excludes from the analysis the countries of former Yugoslavia. This meant that the EJAtlas data had to be tailored to the same countries for which climate impact data are available. For this reason, the number of usable cases was reduced to 296.
Our analysis focuses on place-specifi c instances of environmental confl icts (cf. Elvers et al., 2008), hence the cases of environmental confl icts that occur at the national level in each country, had also to be removed. As a result, the fi nal database consisted of 279 cases.
A specifi c group of indicators was used from each database to illustrate the dimensions of accumulating environmental injustices and of resistance-cumadaptation respectively. For the accumulating environmental injustices, the data from the EJAtlas were selected based on several key dimensions discussed by Sassen in her book on Expulsions (2014). These include the eff ects of environmentally destructive projects on the ecosystem, or what Sassen (2014, 150) refers to as biospheric expulsions. Further considered were the impacts of such projects on human health (Sassen, 2014, 210), on livelihoods and local economies (Sassen, 2010, 28;2014, 29), and on the fresh creation or entrenching of social insecurity, as theorized by Wacquant (2010). Each of these dimensions has a number of corresponding indicators in the EJAtlas, as shown in Table 1. For each of these indicators we used a dummy variable, with 1 representing the presence of every characteristic (i.e. each particular impact) and 0 for its absence. These values are attributed by the volunteer contributors to the EJAtlas. Each contributor had three options: to ascertain that a given impact took/takes place, and hence is "deemed" visible, that it is "potential" or that it is absent. We coded with 1 only those cases where the impact was deemed "visible" and used 0 for the other two possibilities.
On climate impacts, we used information from the ESPON Climate project, which includes about 20 indexes of climate impacts and their corresponding maps. Five of these were selected for the analysis, due to their relevance for assessing the social aspects of climate change impacts. These are potential environmental impact, potential social impact, social sensitivity, potential vulnerability and response capacity. Each of these is explained in the last column in Table 1. Similar to the treatment of the data from the EJAtlas, we used dummy variables for each of these indexes. More specifi cally, we used 1 when social sensitivity was high or very high, otherwise we entered 0. For potential social impact and potential environmental impact we used 1 for medium or high negative impact and 0 otherwise. For potential vulnerability, 1 was attributed to medium or high negative vulnerability and 0 in the other cases. Finally, for response capacity we used 1 for low adaptive, low mitigative and for low adaptive, high mitigative, and 0 for the remaining values.
In order to collect both environmental justice data and climate at community -(or small area) -level, we used the names listed in the EJAtlas as sites of confl ict. These same names were the included in the search function of Climate-ADAPT to locate these places on the maps for each of the fi ve indicators. Source: see the heading of each column for specifi c references.
Dimensions of expulsion (Sassen, 2014) EJAtlas indicators (Temper et al., 2015)  For the resistance-cum-adaptation dimension, we also used indicators from the EJAtlas, this time pertaining to the dynamics of confl ict. These pertain to confl ict dynamics (the highest intensity reached by each confl ict), to the extent of local and transnational mobilizations and to the confl icts' empowering outcomes. The intensity of confl icts was measured on a four-point scale, from 1 (the lowest) to 4 (the highest value). Local mobilization was computed by summing up the number of mobilizing groups. Local empowerment was similarly calculated as the sum of each of the four confl ict outcomes listed in Table 2. The relevant climate impact index was considered potential vulnerability, because it includes information on the regional populations' knowledge and awareness of climate change.

Findings and Interpretation
This section provides an overview of the environmental justice confl icts that occur to diff erent degrees in climate vulnerable areas throughout Europe. Each of the two dimensions -accumulating injustice and resistance-cum-adaptation -are discussed in turn.

Accumulating Environmental Injustices and their Occurrence in Climate-Sensitive Areas
The most signifi cant fi nding is that most of the environmentally unjust projects in Europe (67%) do take place in areas that have at least one above-average climate vulnerability, potential social or environmental impact or social sensitivity. The fi nding does not imply any past causality, in the sense that development projects target climate-sensitive areas. Rather, the implication is that many areas that suff er underdevelopment impacts of the kind listed in Table 1 are the ones that will simultaneously see some of these impacts magnifi ed by climate change. This shows that in a climate policy jurisdiction generally seen as progressive (Rayner and Jordan, 2016), there are growing vulnerabilities resulting from extractive and development activities. The climate edge theory explains these interactions between development project impacts and upcoming risks as unrecognizable by governments and expert systems.
Interestingly, most cases (38%) cluster in the category with two impacts, rather than in the one impact group (20%). For this reason, we will choose this modal category for further analysis. In terms of response capacity to climate change, 73% cases fall in the "low adaptive" category, which means that they struggle to adapt, regardless whether their mitigative capacity (i.e. the ability to reduce emissions) is low or high). This complements the fi ndings in Table 3, namely that environmentally deleterious projects take place in areas that are at risk of climate change impacts, and that also have a low capacity to adapt to such impacts.
To refi ne the analysis, we turn to the main categories of projects covered in the EJAtlas. In order to obtain an accurate picture of where real projects are located, we select only projects that are either under construction, under operation or have been stopped after a certain time in which they were in operation (N=174). In

Number of cases Percent
No impact 90 32.3 One above average impact 57 20.4 Two above average impacts 105 37.6 Three above average impacts 22 7.9 Four above average impacts 5 1.8 Total 279 100.0 these cases, it can be assumed with some degree of confi dence that such projects have already created environmental and other impacts. Not considered here are situations in which projects are proposed or planned, both instances in which it is not yet clear that they have had the time to make their (negative) impacts visible to observers. Our analysis shows some noticeable variability among the nine categories by climate vulnerability/sensitivity score. This means that the modal group for each category of project is placed under diff erent climate scores (see the highlighted cells in Table 4). Five of the nine categories of projects occur mostly in areas with two above-average climate impacts or vulnerabilities and two more categories in areas with one above-average impact.
Tourism recreation projects appear to be the most common in the modal group, with 9 cases (82% of all tourism recreation projects). This is followed by water management projects (6 cases or 55%), industrial and utilities projects (11 cases or 52%), waste management (8 projects or 50%), and biodiversity conservation / biomass confl icts (9 cases, representing 45%). Surprisingly, fossil fuel and climate justice / energy projects tend to occur most often in regions that have no aboveaverage climate vulnerability/ impact score (20 cases or 65%). The same applies to nuclear projects, although to a lesser extent (11 projects or 46%). In order to test the association between project category and climate vulnerability, we transform the latter in a dichotomous variable, which takes a value of 0 if projects are developed in areas with no above average impact or 1 if they occur in areas with one to four impacts. Using the chi square test, we fi nd that the null hypothesis that the two variables are independent can be rejected at the level of 0.001 2 (see Table 5). This means that while most categories of projects do occur in climate-vulnerable areas, there are some noteworthy exceptions to this trend. One are fossil fuel projects which often occur in areas that are not vulnerable (65%). A preliminary analysis, that is not reported here, showed that coal extraction projects taking place in Germany (overall low in climate vulnerability) are largely responsible for this surprising fi nding. Another exception are nuclear projects, which are almost evenly split between climate vulnerable and non-vulnerable areas. In what follows, the focus will be on the distribution of these cases in terms of location (country), as well as environmental, health and socio-economic impacts. The picture suggested by these numbers will be complemented by drawing on several concrete examples.
The distribution of these cases by country reveals a noteworthy pattern. It is mostly Southern European countries where environmentally unjust projects take place namely Spain, Italy, Bulgaria, Portugal and Greece. The share of these projects in the total number of projects in each country is also very high, ranging from 100% in Bulgaria to 77% in Italy. At the other end of the spectrum are mostly western and northern European countries, which off er a mixed picture. Germany or Sweden, for example, have low shares of projects in climate vulnerable areas (10% and 13%, respectively), despite the fact that each has a relatively high number of projects (20 and 15).
Most of these projects, more exactly 61% are accompanied by environmental impacts, which can range from one to thirteen. Moreover, 79 projects (45% of the 174 cases) occur in areas with at one least above-average climate vulnerability/ sensitivity, as opposed to only 27 cases. The most common environmental impacts occurring in climate vulnerable areas are loss of landscape/ aesthetic degradation (28% of 49 cases), biodiversity loss (24% or 41 cases) and surface water pollution or decreasing water quality (17% or 30 cases), thus confi rming to some extent the main hypothesis of this paper. This means that accumulating environmental degradation have occurred in places that are under adaptive stress in the face of expected climate impacts. The relationship is statistically signifi cant at the level  Table 7). These fi ndings also provide prima facie for Sassen's argument on "dead land and dead water" even within Europe, and not only in the Global South (Sassen, 2014). In terms of health impacts, they are less prevalent than the environmental ones, with 63 cases, out of which 45 (71%) display one or several (up to four) health impacts in climate-sensitive areas. However, they do not occur systematically in such areas, as indicated by the chi-square test ( Table 6). The most common environmental impact is the exposure to unknown or uncertain complex risks (16% or 27 cases). Less often reported, but still noteworthy are reports of accidents (for 12 projects) and even deaths (11 cases). Environmental impacts coupled with health impacts indicate that development projects aff ect the lives of local residents in mediated ways, but an analysis of co-presence will not be attempted here.
Finally, the socio-economic impacts of projects under construction, in operation or stopped occur in 93 cases, 62 of which (67%) occur in climate-sensitive areas. The statistical association is nevertheless not signifi cant at the 0.05 level (chi square). The most common such impacts concern the loss of landscape or sense of place (39 cases or 42%), loss of livelihood (27%), increase in corruption or co-optation of diff erent actors (24%) and land dispossession (18%).
A useful case that illustrates how environmental impacts accumulate in an area of high climate vulnerability is located in South Eastern Spain. The largest aquifer in the Province of Almeria, the Sorbas-Tabernas Basin, is located in the province of Almeria in Andalucía (Dene, 2015). The aquifer supplies water to 15 towns, to numerous small villages, which would otherwise have no access to water, and also to subsistence farmers. However, Spain's Department of the Environment has allowed in 2014 that 3,600 hectares of land above the aquifer be planted with an estimated two million olive trees. The spring of El Rio De Aguas is the largest and most important spring carrying water from the aquifer and it is also the only access to water for villagers and subsistence farmers down-stream from the spring. With the spring captured, the latter see "the total loss of their livelihoods" (Dene, 2015). According to the author of this article, some irrigation lines previously used to water fruit trees and vegetable gardens are running dry, the trees other than olives are dying, and the land has become diffi cult to use. These villagers and farmers are said to be under pressure to be relocated but have nowhere to go. Compounding this anthropic crisis, rainfall is insuffi cient to re-charge the aquifer so that underground water levels are continuing to drop. Indeed, the area is located in a region that has above-average potential vulnerability and environmental impact, and is also characterized as having a low response capacity to the challenge of adaptation. All these show that recent "development" decisions intersect with long-term trends of climate vulnerability and push places such as Sorbas-Tabernas Basin towards a point of no return. Adaptation at the local level becomes thus highly impossible, and the situation can only be solved through migration, which puts stress on neighbouring regions.

Resistance-cum-Adaptation in Climate-Sensitive Areas
To measure the resistance to development projects in Europe, we use several measures from the EJAtlas. These are the intensity of the confl ict (latent, low, medium, high), the number of mobilizing, as well as the outcomes that suggest some forms of local empowerment (victory for environmental justice through court decisions), fostering a culture of peace, strengthening of participation and negotiated alternative solution).
In contrast to environmental or health impacts, the intensity of confl icts can be measured even for projects that are not yet in operation. This is because groups can mobilize in anticipation of possible impacts, being aware that preventing development projects before they even reach the operational phase is the most eff ective way to protect their communities, their land and water. However, in order to ensure comparability with the previous tables, we still refer in the following two tables to projects under construction, in operation or stopped (N=174).
Interestingly, the EJAtlas data suggest that the confl icts having the highest intensity do occur in the areas with some climate vulnerability (55%) (Table 8). However, in climate vulnerable areas, medium and low intensity of confl icts are more prevalent (up to 73%). No statistically signifi cant association was found between confl ict intensity and climate vulnerability, using the chi-square test.
The pattern of high intensity confl icts can be explained more appropriately by the political culture of diff erent countries, rather than as a reaction to projected climate change. For example, most high-intensity confl icts take place in Germany, France, the United Kingdom or Sweden, all in areas that are less vulnerable to climate change. In contrast, among the numerous confl icts taking place in Spain (76 overall), in areas of climate vulnerability, only 10 where high intensity. When the number of mobilizing groups in each confl ict is considered, the pattern conforms to the theoretical expectations, to some extent. Regardless of the number of mobilizing groups, they are more active in climate sensitive areas. However, there is no observable trend for higher numbers of groups to be mobilized in the climate sensitive areas. For example, nine or more groups mobilize most often in climate sensitive areas (90%), but these are followed by two groups or less (78%). (see Table 9). The measures that capture local empowerment, namely victory for environmental justice through court decisions, fostering a culture of peace, strengthening of participation and negotiated alternative solutions, occur with varying frequency in climate sensitive areas. Overall, they occur in one third of cases, at most, in both climate sensitive areas and in areas that are not sensitive. The strengthening of participation occurs most frequently in vulnerable areas, but still does not exceed 35% of projects in such areas. This is followed by victories for environmental justice, through court decision, which take place in 23% of projects occurring in climate vulnerable areas. At the opposite end is the fostering of a culture of peace, occurring in just one case. Overall, it may be said that the evidence for an adaptative eff ect of resistance cannot be ascertained by the available data.

Conclusion
The article has sought to articulate the climate edge as a dynamic concept with which to capture climate change as a process that deepens already occurring environmentally unjust practices at local level. We have found preliminary evidence that diff erent categories of projects occur in areas with one or several above-average impacts or vulnerabilities. This shows that Sassen's claim that dead land and dead water are an increasingly common condition is borne out by the available data. In this sense, the article provides a useful contribution to the environmental justice literature, as it reveals patterns that were little visible in the case of Europe. In this way, it provides some empirical evidence that Sassen's argument about the systemic character of extreme conditions (environmental, health and socio-economic) is born out even in the case of the European Union. This is signifi cant since the EU is generally seen as a progressive force in climate and environmental justice.
On the other hand, we have noticed that the climate impact categories do not make a clear-cut diff erence among the projects included in the EJAtlas. This suggests that the infl uence of projected climate changes needs to be better specifi ed, rather than assuming that the climate vulnerable areas themselves have intrinsic characteristics that are noticeable in relation to the characteristics of development projects. A processual model seems to be needed to more systematically hypothesize how climate vulnerability is enhanced by environmentally unjust projects. This, however, shall be the focus of future research.