Seeing like a satellite: Remote sensing and the ontological politics of environmental security

Abstract

The article furthers the debate on environmental security by highlighting the role of visual technologies such as satellite remote sensing in the construction of threats and risks. It provides a rereading of the critical literature on environmental security through the lens of Actor-Network Theory and argues for understanding environmental security as a form of ontological politics. A theoretical framework around the notion of visual assemblage is developed that accounts for the hybrid, socio-technical character of visual technologies like satellite remote sensing, and shows how these render environmental risks and threats visible, intelligible, and thereby governable. Equipped with this framework, the article traces the development of a visual assemblage of satellite remote sensing from the early days of the Cold War until today and reveals its close co-evolution with environmental security discourses and practices. Three major contemporary remote sensing projects are analyzed to reveal how this global visual assemblage enacts multiple versions of environmental security: as resilience of local populations and ecosystems, as a series of local risk factors that become manageable through market-based risk management, and through a ‘meteorology of security’ based on the collection, harmonization, and automated analysis of big (environmental) data from multiple sources.

Keywords

Assemblage big data climate change risk resilience securitization visuality

Introduction

On 3 April 2014, the European Space Agency (ESA) launched its Sentinel-1 earth observation satellite in Kourou, French Guiana, carrying sensors capable of monitoring (among other things) changes of sea ice, oil spills, land-use change, and emergencies like floods or earthquakes. Sentinel-1 is one of six earth observation satellites that are being installed as ‘Europe’s eyes on Earth’ (EU Copernicus, 2016a), through the European earth observation program, Copernicus, until 2020. Copernicus’s service component uses their data, supplemented with data from in situ sensors, to provide earth observation services in six thematic areas: land monitoring, maritime monitoring, atmosphere, climate change, emergency management, and security (EU Copernicus, 2016a). The emergency and security services, which are of particular interest for the present analysis, cover a whole range of environmental security risks – such as environmentally induced migration, land change and resource conflicts, or crises resulting from natural disasters (Danjean, 2014).

As noted by Barry (2013: 1–2), ‘in recent years geographers and social theorists have increasingly drawn attention to the critical part that materials play in political life’. In this article, I take Copernicus and a range of similar satellite remote sensing projects as a starting point to reflect upon the materiality of visual technologies such as remote sensing and their role in the construction of environmental threats. For this, I draw upon the existing literature on environmental security as governmentality (Grove, 2010; Methmann and Oels, 2015; Oels, 2013; Rothe, 2011) and on the role of images and visuals in the construction of environmental threats (Manzo, 2012; Marzec, 2014; Methmann and Rothe, 2014). This literature is reread through the lens of Actor-Network Theory (ANT) and its notions of ontological politics, immutable mobiles, and assemblage (Latour, 1986; Latour and Hermant, 2006; Mol, 1999, 2002). The resulting perspective shows that visual technologies such as remote sensing do not simply construct another discursive frame of the environment. Instead, they bring about different versions, or ‘matter-realizations’ (Moser, 2008), of the environment – as a fragile but interconnected Earth system, as a vulnerable socio-ecological system, or as a set of differentiated risk factors – each producing their own governmental effects.

Two interrelated research questions are deduced from these theoretical assumptions. First, how is the specific relation between environmental problems and security performed or enacted by visual technologies? And second, how does the specific enactment of environmental risks and threats influence their political treatment? To answer these questions, the next section outlines the theoretical framework of the article around the notions of ontological politics, immutable mobiles, and visual assemblage. The second section traces the development of a visual assemblage of satellite remote sensing from the early days of the Cold War until today and reveals its close co-evolution with environmental security thinking. The third section presents the results of an explorative analysis of three major contemporary satellite remote sensing projects and the multiple forms of environmental security these bring about.

Each of the cases examined in the third section paradigmatically represents one type of actor in the visual assemblage of environmental security. First, I take a closer look at the work of the Institute for Global Environmental Strategies (IGES), a Washington-based think-tank, as a paradigmatic example of nongovernmental organizations’ use of remote sensing technology in the environmental security field. IGES can be considered a crucial case because it is a nongovernmental organization explicitly focusing on the application of satellite data in environmental and security governance and represents a nodal point in a broader network of governmental and nongovernmental actors in the field of environmental remote sensing. Second, I use the case of DigitalGlobe to dig deeper into an emerging private environmental intelligence industry (see Thompson, 2007: 10) and the market-based management of environmental risks. DigitalGlobe is a crucial case because it represents the biggest player in the emerging private remote sensing market and is a major provider of environmental intelligence. Third, I take a closer look at the EU’s Copernicus program – as an example of a government-run earth observation program. Copernicus is crucial for this study as it represents the most ambitious contemporary governmental earth observation program and is directly motivated – and legitimized – by environmental security concerns.

Studying environmental security

Environmental security among speech acts, governmentalities, and visualities

In the aftermath of the so-called linguistic turn in the social sciences, an increasing number of scholars in International Relations studied how environmental problems have been discursively framed as security problems. Starting from the assumption that security threats are not simply objectively given, they showed how different authorized actors – including politicians, scientists, and military officials – have successfully reframed environment problems like climate change as security concerns (Buzan et al., 1998; Floyd, 2010). The merit of these works has been to demonstrate that ‘framing matters’ (Dalby and O’Lear, 2016: 5): that the way in which environmental problems are perceived and treated politically depends on how they are framed discursively. They revealed that different discourses of environmental security identify different threats (e.g. climate-induced migration, scarcity-induced conflicts, or natural disasters), multiple referent objects to be protected (e.g. the environment vs. the nation-state vs. vulnerable populations), and heterogeneous policy responses to organize this protection (Detraz and Betsill, 2009; McDonald, 2013; Trombetta, 2008).

Notwithstanding these merits, the literature on the securitization of the environment has been criticized for its narrow focus on securitizing speech acts and linguistic threat constructions and for drawing a stark boundary between the physical reality (which is not directly perceivable) and the socially constructed world (the meaningful representation of reality). By concentrating on ‘threat perceptions rather than an actual danger emanating from objective factors’ (Mayer, 2012: 167), linguistic constructivist approaches might thus unwillingly play into the hands of climate skeptics. Furthermore, the one-dimensional relation between discourses and practices implied by linguistic constructivist works has been criticized (Rothe, 2016). Asking how certain speech acts unidirectionally shape policy responses to environmental threats, they would fail to acknowledge the mutually constitutive relationship between discourses and practices.

A broader understanding of environmental security is provided by governmentality approaches (Oels, 2013; Rothe, 2016). In this perspective, environmental security is understood not so much as a discourse but as a particular field of government shaped by competing political rationalities, subjectivities, practices, and technologies. Some of these works, for example, have shown how environmental problems such as climate change have been turned into security problems through a range of expert practices including risk assessment, scenario development, and hotspot modeling (Mayer, 2012; Methmann and Rothe, 2014). Others have shown that practices of environmental security cohere with different ideal-typical governmentalities such as liberalism, biopolitics, or neoliberalism, each of which produces a different set of assumptions on the subjects, objects, and means of government (see Oels, 2013). Governmentality studies, however, have been criticized for their tendency to simply assign practices and discourses of environmental security to ideal-typical governmentalities (Corry, 2014). This would risk overstating the coherence of existing governmentalities and treating practices and political technologies as epiphenomenal to the broader political discourses that shape them.

A further important extension of discourse-analytical works on environmental security has been provided by works that focus on the role of visuality in the construction of security threats and risks. As nicely expressed by Bleiker (2015: 874): ‘Images … frame what can be seen, thought and said. In doing so, they delineate what is and is not politically possible.’ Two features are said to distinguish pictures and other visuals from words in the construction of security. First, pictures and satellite images are taken as authoritative sources of knowledge, as they apparently represent reality in an objective manner (Dodge and Perkins, 2009: 498; Shim, 2014: 156). The authority of visuals is particularly evident in the realm of environmental security. Here, a whole range of visualization techniques – from colorful data visualizations to computer models or satellite images – make abstract and complex phenomena such as climate-induced migration (Methmann and Rothe, 2014) or the geopolitics of climate change (Manzo, 2012) visible and sizable. Second, images are particularly suitable for invoking certain emotive reactions, such as fear, horror, or grief, that are crucial for the creation of urgency and danger in security discourses. Again, this is mirrored in the debate on environmental security, where powerful icons – such as images of polar bears on retreating ice sheets – or colors in data visualizations are used to invoke emotive reactions and a sense of risk (see Liverman, 2009; Schneider and Nocke, 2014).

The emerging body of literature on visual security represents a promising way to broaden the analytical scope of environmental security beyond the narrow realm of language and speech acts. However, I hold that it often follows a too narrow understanding of visuality – as simply a particular form of discursive representation – that ignores the complex materiality of the visual. This materiality is to be found in the visual economies in which images are produced, circulated, and interpreted (Dodge and Perkins, 2009; Parks, 2009). Furthermore, images and pictures are not only semiotic systems of signs but also physical objects that can be printed, circulated, blocked, (re-)combined, and reproduced. Finally, looking at, seeing, touching, or interpreting visuals are material bodily practices – which are increasingly supplemented with technological practices such as algorithmic pattern detection or automated image analyses (see Brannon, 2013).

Environmental security as ontological politics

To account for the materiality of the visual and for the special role of visual technologies in the construction of environmental security, I reread the existing literature on environmental security through the lens of ANT and the notion of ontological politics. In short, ANT radicalizes the assumptions of linguistic or visual constructivism. Instead of viewing texts, images, and other semiotic systems as representations of a pre-existing single reality, it holds that the interplay of expert practices, scientific discourses, technologies, and visuals constitutes multiple realities (Hind and Lammes, 2016: 81–82). Annemarie Mol (2002), for example, shows through her long-term ethnographic research in Dutch hospitals how different medical practices and diagnostic tools produce multiple versions of atherosclerosis. To assume that phenomena are multiple implies they are ‘more than one but less than many’ (Mol, 2002: 55). According to this understanding, ontology ‘is not given in the order of things but … ontologies are brought into being, sustained, or allowed to wither away in common, day-to-day, sociomaterial practices’ (Mol, 2002: 6). As Ingunn Moser (2008: 99) puts it, natural realities become ‘matter-real’ and are ‘mattering’ only through their ‘continued enactment and re-enactment in situated practices’. Following these assumptions, environmental security can be understood as a form of ontological politics, in which the ontological status of phenomena such as climate change, deforestation, or natural disasters – as well as the ways of knowing them – becomes the subject of debate (Barry, 2013: 7; Schouten, 2014).

I argue that visual technologies such as satellite remote sensing play a crucial role in the ontological politics of environmental security. Consider Mol’s seminal work in the ontological politics of disease. A large share of the practices that render a disease such as atherosclerosis ‘matter-real’ are based on visual technologies: X-ray units, MRI scanners, microscopes, molecular imaging techniques, etc. The same holds true for the realm of environmental security. Here, visual technologies, including computer modeling, satellite remote sensing, data visualization, hotspot mapping, and graphs of future scenarios, enact different versions of ‘the environment’ as security risk (Randalls, 2014). Rather than producing mere representations of pre-existing environmental risks, visual technologies essentially bring them about: ‘The world does not exist before these ways of sensing’ (Hind and Lammes, 2016: 90). For, as Mitchell (2013: 233) crucially reminds us, ‘Nature is unable to speak for itself…. The facts of nature speak only with the help of measuring devices and tools of calculation’. To render a phenomenon such as climate change visible, intelligible, and thereby ‘matter-ing’ and ‘matter-real’, a global assemblage of visual technologies – including satellites, weather stations, computer simulations, researchers, and mechanisms of international cooperation – is required (Edwards, 2010; Wark, 2015: 166–182).

To study visual technologies as part of the ontological politics of environmental security, I first draw on the notion of immutable mobiles (Latour, 1986). Bruno Latour has coined this term to explain the power effects of ‘technologies of visualisation and inscription’ (Latour, 2013: 77), such as the printing press, cartography, photography, or the computer, that allow inscribing parts of the real world into flat and mobile artifacts (Latour, 1986: 17). Through these technologies, reality becomes flattened, mobile but immutable. Inscribed into pictures, maps, or charts, inscriptions of local realities – such as a piece of soil – can be circulated globally, reproduced, or combined with other inscriptions of reality. Pictures of different origin and different scale can be layered (Latour, 1986: 7).

Second, I draw upon the notion of visual assemblage (Bleiker, 2014: 76) to account for the fact that environmental risks and threats are never produced by a single technology – never become visible on a single image – but become enacted by a complex web of linked inscription devices, practices, and discourses (Latour and Hermant, 2006: 29). The notion of assemblage refers to a temporary form of order of heterogeneous elements ‘whose unity comes solely from the fact that these items function together, that they “work” together as a functional entity’ (Patton, 1994: 158). The visual assemblage of environmental security can be understood as exactly such a heterogeneous network of visual technologies and related infrastructures, people, laws and regulations, discourses, and practices (van Munster and Sylvest, 2016: 2). The immutable mobiles (satellite pictures, global climate models, vulnerability maps), which circulate through this assemblage, create new associative relations between actors as diverse as satellite operators, climate scientists, environmental activists, or military officials (Mayer, 2012). Together they turn ‘nature’ into a range of calculable, perceivable, and governable risks through a series of calculations, transformations, and translations (Hind and Lammes, 2016). The notion of visual assemblage makes it possible to account both for the socio-material character of the visual and for the messiness and multiplicity of practices and discourses of environmental security.

Using the framework in practice

The perspective of assemblage calls for a pluralization of methods in the study of visual politics (Bleiker, 2014). To grasp the social/technical/discursive/visual character of visual assemblages, I combine three analytical layers. The first analytical layer is a semiotic visual analysis (Rose, 2016: 106–146). A corpus of highly visible demonstration images – provided on websites, in demonstration reports, in social networks, or in demonstration videos – has been compiled in each case. Such showcase images are particularly promising for the objectives of this article because they are pre-selected by the respective actors to demonstrate the whole range of possible applications and capabilities of satellite imagery. A set of heuristic questions helps to structure the analysis: What do the images show and what is hidden? Are there common visual characteristics or patterns? How are images arranged and how are colors used? And what are the aesthetic and political effects of decisions on the composition, color, framing, or resolution of images?¹ Second, a discourse-analytical layer reveals the interdiscursive context of images through a study of the reports, homepages, flyers, or policy documents in which they are embedded (Rose, 2016: 186–219). Here, I ask how the different texts ascribe meaning to the images, and how the latter become linked to broader concepts, narratives, and storylines in environmental security discourse. A third layer focuses on the performativity of images as immutable mobiles and their function in broader visual assemblages. So, here the question is what visuals do rather than what they show: How are images produced by broader visual assemblages? How are they circulated or blocked? And how do they enact different versions of the environment? To study this third layer, I draw on a comprehensive review of the relevant secondary literature, combined with a thorough document analysis of legislation, regulations, and related policy documents in each case.

The co-evolution of remote sensing and environmental security

The roots of the visual assemblage of environmental security can be traced back to the militarization of geophysical research in the early years of the Cold War (Dalby, 2013: 41). With its successful launch of the Sputnik satellite in 1957 during the so-called International Geophysical Year, the Soviet Union fired the starting pistol for a technological and scientific race between the two superpowers. As a crucial part of this race, satellite remote sensing emerged as a dual-use technology, or ‘cyborg science – the result of an irreparable fusion of scientific and military concepts, materials and skills’ (Elam, 1999: 106). At the same time, the strategic communities’ interest in the atmospheric impact of nuclear weapons boosted advancements in numerical weather modeling, which in the 1960s were further developed into the first simulation models of global climate change. The ever more detailed global models at the same time produced ‘an insatiable thirst for data from every corner of the world’ (Edwards, 2010: 24), which since the late 1970s and 1980s was increasingly satisfied by satellite-based remote sensors. In addition, in 1972 the US administration established the Landsat program within the US National Aeronautics and Space Administration (NASA) as the first civilian satellite remote sensing program. The program’s eight satellites provided multispectral Earth imagery at moderate spatial resolution – perfectly suited for land (change) monitoring – and thereby contributed considerably to the discovery of macro-scale environmental problems, including deforestation, land degradation, and the ozone hole (DeLoughrey, 2014: 261).

NASA’s Landsat and other emerging remote sensing programs can be considered as parts of an emerging visual assemblage of heterogeneous human and non-human components (Barry, 2013: 14; Edwards, 2010). This assemblage includes hardware such as satellites, sensors, ground stations, and data centers; imagery-processing-and-analysis software; data such as images or digital maps; human actors like system operators, policymakers, scientists, and national security personnel; and legal and regulatory frameworks (Thompson, 2007: 2). The capacity to ‘see’ environmental problems, and to see the planet from a distance, is distributed across this visual assemblage. Neither the satellites, which translate the electromagnetic radiation reflected from the Earth’s surfaces into numerical grey-values, nor the operators in the control room, which process ‘wavelengths in false colours’ (Latour and Hermant, 2006: 8), nor the software tools that turn the data into realistic true-color images, nor the analysts or algorithms that interpret the resulting images would be able see anything were it not for the other elements in the assemblage.

A securitization precursor

I would argue that there was a close co-evolution between the emergence of a global visual assemblage of satellite remote sensing and the emerging discourse on environmental security in the 1980s and 1990s. The notion of co-evolution stresses the reciprocity of this development. While the satellite gaze made it possible to conceive of environmental change as a global threat in the first place (Van Munster and Sylvest, 2016: 6), environmental security also became inscribed into technical decisions, for example when Landsat’s remote sensors were optimized to monitor land change and resource degradation. The mutual imbrication of satellite technology and environmental security played out at several levels. First, the militarization of geophysical research during the Cold War created new associations between geoscientists and security actors that were crucial for the emerging imaginary of the environment as a security problem. Marzec (2014: 244) describes how this emerging actor-network led, for example, to the US defense and intelligence community’s interest in climate change in the early 1990s:

In 1992 the Central Intelligence Agency (CIA) began to establish relations with climate scientists in the program Measurements of Earth Data for Environmental Analysis, or MEDEA, when it declassified satellite imagery for ‘patriotic’ climate scientists.

Second, the described visual assemblage enabled a new ‘way of seeing’ the Earth and the human environment, which might be called a planetary gaze (DeLoughrey, 2014; Jasanoff, 2004: 44). The new possibility to look at the planet from the outside produced a sense both of the earth’s singularity and of its radical fragility. The planetary gaze produced a certain uncanniness – a feeling of being detached from planet Earth (DeLoughrey, 2014: 264). In so doing, the planetary gaze invoked deep-rooted ‘apocalyptic fears about the end of the earth’ (DeLoughrey, 2014: 266). This sense of urgency created by satellite remote sensing resonated well with the early environmental security debate in the 1990s and its alarmist tone (Marzec, 2014: 235). Yet the planetary gaze of the described visual assemblage did not simply produce another discursive frame of a pre-existing environment. Rather, it constituted a whole new reality of the Earth as a singular, interconnected, and fragile system (Van Munster and Sylvest, 2016: 4–8). This fragile Earth system became ‘matter-real’ through the immutable mobiles – such as true-color satellite images, simulation models, graphs, or numerical calculations – produced by two central inscription technologies: satellite remote sensing and computer modeling (Edwards, 2010: 2). By developing these technologies into a comprehensive global monitoring system, it was believed it would be possible to monitor, predict, control, and manage the Earth system (Jasanoff, 2004: 42; Marzec, 2014: 245). This global monitoring approach, which highlighted the role of national leaders as planetary or environmental managers, considerably influenced the early environmental security debate.

Third, the focus of the early environmental security debate in the 1990s was on large-scale global problems such as scarcity-induced conflict or migration (Rothe, 2016). The detection of these large-scale problems was closely linked to the technical capabilities of satellite remote sensing, for to detect large-scale, slow-onset changes like land degradation and resulting scarcities, an abstracted, distant view from above was required. What is more, multispectral sensors, such as those carried by Landsat satellites, made it possible to visualize processes of plant degradation in the infrared range of radiation. To be precise, these large-scale environmental changes were considered as drivers of conflict and migration in the early environmental security discourse. Remote sensors and their planetary gaze thus considerably influenced the epistemological horizon of environmental security thinking.

Opening and commercializing remote sensing

During the Cold War, access to the visual assemblage of satellite remote sensing was restricted to a closed ‘club’ of a few geoscientists, military research centers, and international bureaucrats (Elam, 1999: 98). The monopolization of space technology and security concerns imposed heavy restrictions on the circulation of the immutable mobiles of satellite sensors. Landsat is a good case in point: while the civil Landsat program was used by the USA to promote ideals of transparency and international scientific cooperation, the US government at the same time restricted the permitted spatial and spectral resolutions of Landsat images (Thompson, 2007: 8). Furthermore, the inaccessibility of satellite images and their abstract, scientific enactment of the environment further limited the number of potential users.

Shortly before the end of the Cold War, however, this picture began to change. The USA reacted to privatization efforts in France and the Soviet Union and decided to open and commercialize its remote sensing industry with the Land Remote Sensing Policy Act of 1992 (Elam, 1999: 99). The commercialization process was accompanied by the triumphal march of the personal computer and the spread of geographic information system (GIS) software to read, process, and interpret geodata, which further increased the number of users of commercial satellite imagery. As a result, satellite remote sensing grew into a massive and complex global market with a projected size of US$2.6 billion by 2020. The current global market is characterized by the dominance of a few global players, such as DigitalGlobe, and national space agencies as well as the emergence of myriad medium- and small-sized enterprises that are turning satellite data into maps, GIS products, or data visualizations.

The opening and commercialization of remote sensing technology happened at the same time as a shift in environmental security discourse towards human security and resilience (Corry, 2014; Detraz and Betsill, 2009; Trombetta, 2008). Increasingly, the focus of environmental security was shifted from the international system or the nation-state towards individuals’ and local communities’ vulnerabilities and local environmental risks (McDonald, 2013: 46–47). A global managerial command-and-control approach to environmental problems made way for more regionalized approaches of risk management and other measures to enhance the adaptive capacities of vulnerable populations against environmental shocks (Oels, 2013: 24–26).

Again, I am not claiming the existence of a direct causal relation between the commercialization of remote sensing and the rethinking of environmental security. Nevertheless, the two developments are interlinked in multiple complex ways. First, the closed visual assemblage of remote sensing that was concentrated around a few global nodes became increasingly decentralized and dispersed. New actors including nongovernmental organizations (NGOs), social scientists, and businesses entered the assemblage. As the immutable mobiles of commercial remote sensing became increasingly available, these actors were provided with new means of participating in discourses and practices of environmental security. Second, calls for opening the formerly restricted satellite technology in the 1990s have been argumentatively linked to the growing ‘importance of essentially non-military targets such as environmental change, large-scale population movements and migration flows’ (Elam, 1999: 103). In other words, narratives and storylines of environmental migration and conflict became an essential part of the visual assemblage of remote sensing. It is thus not a coincidence that a large share of the services provided by private satellite businesses such as DigitalGlobe are dealing with environmental risks and threats. Third, new technologies and devices in the visual assemblage, including enhanced image processing and analysis, as well as the dramatic increase in the spatial and spectral resolutions of remote sensors, led to new possibilities of visualizing local environmental risks. This made it possible to zoom in and to scroll through the smaller components of the Earth system and challenge the planetary gaze.

Environmental security multiple

The following subsections present the findings of three explorative case studies that try to dig deeper into the contemporary visual assemblage of environmental security. Multiple versions of environmental security exist simultaneously, each unfolding in a different network of actors, practices, discourses, visualities, and technologies.

A resilience machine: NGOs and the new familiarity of remote sensing

The Institute for Global Environmental Strategies (IGES) is a Washington-based NGO that lobbies for the civil application of satellite remote sensing technology. For this purpose, IGES is, for example, organizing the Alliance of Earth Observations to bring together the private sector, researchers, NGOs/think-tanks, and the strategic community in the remote sensing sector (IGES, 2016). Furthermore, it conducts a range of awareness-raising and demonstration activities to prove the value of satellite images for the work of NGOs. One of these activities is the release of a report entitled Sanctuary: Exploring the World’s Protected Areas from Space (IGES, 2014) that seeks to demonstrate the capabilities of satellite remote sensing in the protection of vulnerable ecosystems and populations. I take IGES and its Sanctuary report as a starting point to investigate new forms of civil-societal agency in the visual assemblage of environmental security.

With the commercialization and opening of remote sensing, also the character of the technology changed. While the restrictive use of satellite remote sensing during the Cold War was justified through reference to the technology’s complexity and inaccessibility, remote sensing now ‘gained a new familiarity’ (Elam, 1999: 104). IGES’s report perfectly mirrors this logic when it states: ‘Every time we use our eyes to inspect our environment, we are using a form of remote sensing’ (IGES, 2014: 2). Remote sensing is presented as a ‘prosthetic technology’ (Elam, 1999: 104) that opens completely new ways of seeing (and engaging with) our environment: ‘Scientists and engineers have created electronic sensors that allow us to overcome our biological limitations and detect light beyond what our eyes can see’ (IGES, 2014: 2). Furthermore, the planetary gaze of early remote sensing obscured everything that could not be sensed by the macro-scaled remote sensors – such as the social and economic structures underlying global environmental problems, indigenous populations (Marzec, 2014: 245), and ‘individuals’ embodied environmental experience’ (Houser, 2014: 328). The images in the IGES report, on the contrary, highlight the embeddeddness of people within their natural environment: ‘the fates of the people and the land are intertwined’ (IGES, 2014: 22), and ‘the “environment” is where we all live’ (IGES, 2014: 24). To visualize this embeddedness, the report shows a whole range of beautiful true-color satellite images of different ecosystems around the world, including mangrove forests, mountain areas, small islands, and deserts. As humans are invisible on these satellite images, the report complements them with pictures of local/indigenous populations (see e.g. IGES, 2014: 2, 14, 16, 22, 30 ).

In the report, technical descriptions of remote sensing capabilities alternate with narrative parts in which the people’s stories are told. For example, the report describes how remote sensing as a prosthetic technology contributes to local resilience against the threat of malaria by empowering local communities to ‘manage their own habitat’ (IGES, 2014: 16). In picturesque terms, the report describes a village in Mali that successfully adapted to the threat of malaria-transmitting mosquitos by ‘using insecticide-treated mosquito netting’, and shows a picture of the modified huts (IGES, 2014: 16). The report explains that satellite remote sensing provides the information required for taking such preventive adaptation measures – for example, information on ‘temperature and rainfall, identifying areas most conducive for mosquito breeding’ (IGES, 2014: 16). But also satellite monitoring of deforestation, according to IGES, can help stop the spread of the disease: ‘Deforestation in support of agriculture creates new habitat for mosquito larvae, with farmers especially susceptible as they move into new forest clearings’ (IGES, 2014: 16).

Nongovernmental remote sensing as described in the IGES report enacts environmental security as the security of a new referent object: local communities and local ecosystems as mutually dependent, symbiotic systems. Furthermore, it promises to provide the information necessary to advance the resilience of these socio-ecological systems by rearranging their constituents. In the given example, it reassembles the relations of people, plasmodium parasites, mosquitos, huts, rainfall, temperatures, forests, etc. Things like huts, rainfall, or trees become disassembled out of their local context, become flattened and mobile, as streams of data, visual patterns, and geographic forms on satellite images, and related to other immutable mobiles such as figures of the spread of disease vectors or population statistics. Having passed a whole process of translation as immutable mobiles, the things are then reintroduced into the local context – but they are no longer the same.

The visual assemblage of remote sensing, for example, turns forests into storm defense infrastructures, because from a distance their ‘tremendous energy-absorbing and land-stabilizing effects’ become visible (IGES, 2014: 22). Through satellite images, mangroves become ‘matter-real’ as natural flood defenses that are ‘protecting coastal communities from severe storms and aiding human resiliency and adaptability’ (IGES, 2014: 22). The local populations themselves cannot see the things in their environment in the same manner because they are too close and biologically limited: for them, the mangroves are a ground for shrimp fishing but not a flood defense. From a local perspective, small-scale forest fires – to give another example – are an important means of agricultural optimization. On satellite images, owing to the view from above and the combination of thermal and multispectral sensors, however, the threat of fires to local socio-ecological systems becomes visible (IGES, 2014: 26).

The analysis of the IGES report seems to confirm the observation that discourses and practices of environmental security have recently been reconfigured around notions of human security and resilience (Boas and Rothe, 2016).² This shifts the focus of environmental security from national territories to local populations and socio-ecological systems. The downside of the spatial reassembling of satellite technology is that it also implies a multiplication and dispersion of control technologies. NGOs’ access to high-resolution imagery provides them with means to control and govern vulnerable communities at a distance. Protecting and surveilling the vulnerable here become two sides of the same coin (Marzec, 2014: 236). Moreover, as shown in the next subsection, the satellite industry is quite monopolized and high-resolution images have a high price tag. This limits the number of potential nongovernmental satellite actors in the environmental security field to financially strong (Western) NGOs³ and those with good ties to government and industry – and excludes NGOs and social movements with a more radical, counter-hegemonic agenda.

A business machine: DigitalGlobe and the environmental intelligence industry

DigitalGlobe, founded immediately after the US Land Remote Sensing Policy Act of 1992 under the name WorldView Imaging Corporation, is the biggest commercial provider of high-resolution satellite imagery today (following the fusion with its former competitor GeoEye in 2012). DigitalGlobe offers not only ‘raw’ satellite images, but also a whole range of analytical tools to predict and prevent calamities resulting from man-made and environmental crises. DigitalGlobe thus represents the spearhead of an emerging intelligence industry that promises solutions to the world’s complex security problems. In 2014, DigitalGlobe conducted a social media contest in which users could vote for their favorite satellite images (see DigitalGlobe, 2014). The images in the DigitalGlobe contest are showcases for the capabilities of remote sensing and cover the whole range of applications of satellite data in the emerging environmental intelligence industry. They are hence perfect examples through which to study the forms of visuality underlying this field and to answer the question of how it enacts environmental security.

The DigitalGlobe pictures are as globalist in nature as the above-described planetary gaze: they decidedly show phenomena and problems from all over the world (globalism is even mirrored in the company’s name). The pictures show, for example, melting glaciers in the Antarctic (Image 23), offshore oil drilling in the Arctic and resulting pollution (Image 5), wildfires around San Diego, California (Image 2), flooding in Gunja, Croatia (Image 18), the environmental impact of waste disposal in Sulaibiya, Kuwait (Image 8), aquacultures in Huayuanli, China (Image 14), coal plants in China (Image 19), mining in Canada (Image 21), dams and their impact on water security in Paraguay (Image 24), and forest management in Alberta, Canada (Picture 25). However, the DigitalGlobe images do not ‘connect the dots’ (Houser, 2014: 326) – that is, establish the connections between the different places, activities, and environmental problems shown on the single images. The broader global environmental crisis that is behind many of the monitored problems is invisible (Mirzoeff, 2014: 217). The abstracted view from above even makes the human exploitation and disruption of nature appear beautiful. The winning image, for example, uses colors to visualize mineral stocks in the Rainbow Range, Canada, leading to a colorful and beautiful picture of the mountain area (Image 21). A Facebook user responded to another image, which shows receding glaciers (Image 23) due to climate change, with the comment ‘fantastique’. The traces and artifacts of a global fossil-fuels-based capitalist system – such as global ports, harbors, coal plants, oil drilling platforms, megacities, dams, aquacultures, and mining sites – appear as beautiful geometric forms and figures.

I would argue that DigitalGlobe’s services enact the environment as a global circulation of goods (fossil fuels, minerals, water, ecosystem services) and ‘bads’ (excessive greenhouse gas emissions, natural disasters, viruses, or waste). As part of a market-based risk-management approach to the environmental crisis (Grove, 2010; Oels, 2013: 18–20), the immutable mobiles provided by commercial satellite intelligence allow visualizing this circulation and thereby keeping it at an optimal level. A good case in point is the circulation of waste (Image 8). Since ‘improper disposal practices can adversely affect public health’, local civil governments can use DigitalGlobe’s services to keep levels of waste disposal at a tolerable level (DigitalGlobe, 2014). Mining and oil drilling businesses can use DigitalGlobe’s service to find an optimal balance between increasing extraction rates and avoiding pollution (DigitalGlobe, 2014). The monitoring of ports and airports (Images 3 and 16) identifies deviant, potentially dangerous patterns of global transport (DigitalGlobe, 2014). Through DigitalGlobe’s imagery, the planetary crisis becomes ‘matter-real’ as myriad local, digitized risks (and risk factors) that are subjected to a technocratic risk management. There is no insurance against global warming, war, or mass migration at a global level, but local floods, storms, crop failures, or health impacts can be calculated, financialized, and insured. Or, put in the language of DigitalGlobe, ‘You can’t control Mother Nature, but you can control your response’ (DigitalGlobe, 2016a). For this, a considerable amount of DigitalGlobe’s services are focused on damage detection. Using satellite image time series of cities before and after a natural disaster, DigitalGlobe’s analysts – assisted by automatic object detection algorithms – provide damage assessments of major natural disasters. The swift and reliable information on loss and damage represents a key prerequisite for disaster insurance (DigitalGlobe, 2016b). Multispectral images can help in calculating exposure – such as agricultural damage caused by drought or disease – that would otherwise remain incalculable, and thus create new insurance opportunities.

Commercial satellite intelligence enacts environmental security on the basis of digitized and commoditized immutable mobiles circulated in the global economy of remote sensing. This global economy builds upon a ‘highly capital-intensive space infrastructure [that] is monopolized by a few national space agencies as well as a few multinational corporations’ (Parks, 2009: 541). So, while the commercialization of satellite remote sensing led to an expansion of the visual assemblage and an increased circulation of satellite images, access to the latter continues to be restricted. Instead of military secrecy, property-rights regimes and the heavy price tag on satellite images are now regulating access (Brannon, 2013: 289). At the same time, ‘“shutter control” remains firmly in the hands of powerful government institutions and unaccountable corporations’ (Dodge and Perkins, 2009: 498). In sum, following Dalby and others, I would argue that the visual commodities of this expanding assemblage represent a crucial new resource in a market-based risk management of environmental problems (Oels, 2013), seeking to secure the global fossil-fuels-based capitalist system from its toxic local excesses (Dalby, 2015: 431–433).

A data machine: EU Copernicus and the meteorology of risk and security

The European Union’s earth observation program, Copernicus, is the most ambitious existing governmental satellite earth observation project existing today. Once fully operational, it will become a new nodal point in the visual assemblage of environmental security. Environmental security rationales have been guiding the Copernicus project from the very beginning (see, for example, GMES Support Team, 2002; GMES Working Group on Security, 2003). When the idea of a genuine EU earth observation program was first expressed in the so-called Baveno Manifesto by the European Space Agency (ESA) and the European Commission in 1998, it was called ‘Global Monitoring for Environmental Security’ (Danjean, 2014: 15). Although the program was soon renamed ‘Global Monitoring for Environment and Security’ (GMES) – and later in 2014 ‘Copernicus’ – it continues to be informed by environmental security rationales. Practically, this focus finds its expression in Copernicus’s Emergency Management Service (EU Copernicus, 2016b), which went operational in 2015, and its security services, which at the time of writing are yet to become fully operational and build upon a series of pre-operational pilot services (G-MOSAIC, 2012; G-Next, 2012). These services provide satellite imagery and digital maps on issues including environmentally induced migration, land degradation and resource scarcity, land use and scarcity-induced conflicts, and the risks of flooding, wildfires, and other natural disasters – in short, the whole bandwidth of issues in environmental security discourse.

The visual products and artifacts provided by Copernicus’s security and emergency services differ from the satellite images discussed in the previous two subsections. Whereas the latter drew upon an aesthetic of naturalistic beauty, Copernicus’s images follow a techno-scientific aesthetic of data visualization. As Diamond (2010) notes, ‘Data Visualization allows representations to be mapped onto each other, to compare and overlay vastly different data sets’. Every visual technology in these layered digital images would entail its own aesthetic rules and conventions.

Copernicus’s security and emergency services are, for example, using datasets that combine remote sensing data on regional ecosystems, global and regional climate data, population data, and georeferenced information on critical infrastructure (G-MOSAIC, 2012; G-Next, 2012). Figure 1 shows a flood risk assessment map for Rio Mamoré in North Bolivia that was produced by the Risk and Recovery Mapping of Copernicus’s Emergency Management Service (EU Copernicus, 2016b). It first comprises a layer of satellite data presented as naturalistic true-color image – an apparently objective representation of the Rio Mamoré area. This image can be combined with a second layer that adds geodata on infrastructures or roads and follows widely shared aesthetic conventions of cartography (the use of icons, legends, etc.). A third layer shows the varying vulnerabilities of populations and makes use of aesthetic conventions of data visualization, such as the use of colors (red) and forms (squares) to express varying degrees of risk (Liverman, 2009).

Figure 1.

Rio Mamoré – Flood Risk Assessment.

Another example is presented in Figure 2: a spatial assessment of land-use change and conflict in North Kivu in the period 2003–2008, developed by Copernicus’s pre-operational security service G-MOSAIC (2012). The digital map, first, comprises a political base-map. On top of this base-map, a second layer visualizes land-cover changes in different colors (green for deforestation, red for crop and grassland degeneration, black for urbanization), derived from the semi-automatic classification of multispectral high-resolution satellite images. Finally, a third layer maps data on human conflict in the region obtained by the Armed Conflict Location and Event Data Project (ACLED, 2016) dataset and visualizes this data in the form of grey-shaded circles of different sizes. This layer follows aesthetic conventions of data visualization rather than those of mapping: the size of the circles symbolizes a numerical value (the number of conflicts) instead of the actual spatial distribution of conflicts.

Figure 2.

G-MOSAIC map on land-use change and conflict.

In both examples the naturalistic beauty of the true-color satellite images is supplemented with the beauty of data visualizations and false-color images that can visualize land-cover changes or the distribution of risk factors in beautiful colors and forms. Copernicus’s visual products seek to harness the vast possibilities of digital satellite imagery by combining and mapping such imagery with a range of other data layers to reveal undetected patterns that point to potential risks and vulnerabilities (see Brannon, 2013: 274). Human infrastructure, population density, land use, and vegetation types become ‘matter-real’ as digital data layers, which can be combined or subtracted from each other with the help of GIS software. Hence, Copernicus’s visual products are not only mobile – like analogue maps – they are mutable and manipulable digital maps (Hind and Lammes, 2016: 85).

Copernicus’s layered digital maps – or ‘mutable mobiles’ – enact environmental security in a way that differs considerably from the remote sensing projects discussed in the previous subsections. I want to call this form of enactment big environmental data (see also Brannon, 2013). The term ‘big environmental data’ refers to the attempt to collect, distribute, harmonize, combine and visualize huge amounts of environmental data to develop predictive tools in the security field. Such a rationale has guided Copernicus from its early days onwards (see e.g. GMES Support Team, 2002: 8; GMES Working Group on Security, 2003: 6). In an online interview, Josef Aschbacher, head of the Programme Planning and Coordination division at ESA, explains how Copernicus follows the example of meteorology to develop predictive capabilities (see GeoBuiz, 2015). He elaborates how meteorology draws on the combination of satellite information with a range of other data sources – measurements of temperature, humidity, pressure, etc. – to create a foresight picture, which is then presented to stakeholders in an accessible manner. Aschbacher argues: ‘In Copernicus we are doing exactly the same, but beyond meteorology in environmental monitoring, disaster management, security services’. To make sense of Copernicus, it is thus helpful to take a closer look at the history of meteorology (see Edwards, 2010: 32).

A prerequisite for the development of predictive capabilities through big data in the security field is the creation of interoperability between databases (De Goede et al., 2014: 415). Long before the creation of interoperability between databases became a main challenge for intelligence and security actors, meteorology was facing a very similar challenge (Edwards, 2010: 24). When modern meteorology began to make use of the power of digital computers after the end of World War II, its models required an ever-bigger amount of data from a whole variety of different sources, each having its own standards and protocols of measurement. Hence, a whole subfield of meteorology has been dedicated to the harmonization of climate data through data models and other devices of standardization (Edwards, 2010: 109). GMES/Copernicus took meteorology’s approach to interoperability as an example and adopted a whole range of acts and devices of standardization (see e.g. GMES Support Team, 2002: 20). These include, for example, the development of data standards, common data-classification levels, or the distribution of ‘technical reference architectures’ for public services. A good example is the so called ‘Common Information Sharing Environment’ (CISE) in the field of maritime security (European Commission, 2010). CISE is supposed to integrate Copernicus’s remote sensing data with the databases of actors as diverse as FRONTEX, the European External Action Service, or national customs into an ‘Integrated Maritime Surveillance System’. CISE uses a ‘common informatics language’ that allows decentralized user communities to maintain autonomy over the collection of data, which is then translated into a commonly agreed format. This approach – the EU hopes – will lead to ‘a cost effective, decentralised interconnection of different information layers’ (European Commission, 2010).

Figure 3 illustrates the infrastructural network of CISE. It demonstrates how different data sources – airplanes, satellites, customs databases, fishing boats, marine patrols – become linked together into a complex visual assemblage. Each data source is linked to a decentralized CISE data center functioning as a center of translation to render the data interoperable. Now consider Figure 4, which is the World Meteorological Organization’s (WMO) illustration of its Global Observing System. The similarities are striking: not only did the EU adopt the WMO’s model of a Global Observing System, it even reproduced (almost copied) the latter’s visual representation (in terms of perspective, image composition, and positioning of elements such as the ship vessel, the airplane, and the satellite in the upper-left part of both illustrations). This shows, as I would argue, how much the enactment of environmental risks through ‘big environmental data’ is modeled after weather prediction and the underlying visual assemblage of meteorology.

Figure 3.

Common information sharing environment (CISE). © Alain Biltereyst.

Figure 4.

Global Meteorological Observing System.

While the authority of classical satellite imagery stems from an ‘overview effect’ – that is, the ability to oversee large areas at a distance – the authority of big environmental data stems from an ‘overlap effect’ – that is, the capacity to (re)combine multiple visual layers derived from different inscription devices. Environmental security risks here simply ‘matter-realize’ out of the data itself – as surprising associations and previously undetected patterns (Aradau and Blanke, 2016: 3). ‘Big environmental data’ does not provide us with causal mechanisms to explain the relations between the environment and the security that it creates (Chandler, 2015: 837). Consider the map on land-use change and conflict in North Kivu discussed earlier (Figure 2): evidence of environmental conflict here is not derived from any conflict theory. Instead, it emerges out of surprising visual patterns that appear when different visual layers are experimentally combined – here, the co-appearance of land-cover changes and a high number of human conflicts. The only way to challenge the enactment of environmental security through big data would hence be to challenge the data itself – by questioning the quality of classifications, algorithms, or databases. The problem then is that the whole assemblage behind each data layer, and all the acts of translation, calculation, and interpretation that it includes, remain hidden. This tendency of foreclosing public debate is further reinforced by the fact that, unlike with Copernicus’s other services, access to the security services is reserved for a few actors in European security governance, such as FRONTEX or the EU’s External Action Services (see Danjean, 2014; G-Next, 2012). How and when these organizations will draw upon Copernicus’s satellite products in future border control or crisis prevention missions will hardly be shared publicly. Thus, while the core idea of ontological politics is that the material characteristics of objects and the possibilities of knowing them are increasingly becoming subject to public disputes (see Barry, 2013: 13; Mitchell, 2013: 239), the opposite seems to be the case here: neither the production nor the application of ‘big environmental data’ is open to scrutiny by the broader public.

Conclusion

In this article, I have sought to highlight the role of visual technologies in the process of constructing environmental security and to show that environmental security is ontologically multiple. To do so, I developed an analytical framework around the concepts of ontological politics, immutable mobiles, and visual assemblage. This framework made it possible to account for the role of visual technologies in environmental security without introducing an artificial divide between the technological, the discursive, and the visual. Equipped with this framework, I first traced the roots of the visual assemblage of satellite remote sensing back to the early days of the Cold War and showed how it co-evolved with knowledge controversies on environmental security. The subsequent section took a closer look at the current visual assemblage of satellite remote sensing to answer the question of how it enacts environmental security in multiple ways. First, the analysis showed how NGOs are using remote sensing imagery to protect vulnerable local socio-ecological systems and how this novel ‘local gaze’ resonates with a recent discourse of resilience in environmental security thinking. Second, it was revealed how private companies such as DigitalGlobe produce environmental security as a private good. Immutable mobiles circulated as commodified goods in an emerging private intelligence economy allow users to assess and manage localized environmental risks. Third, the case of Copernicus demonstrated how environmental security threats and risks are increasingly becoming enacted through ‘big environmental data’: the collection, harmonization, and automated analysis of data from multiple sources.

Regarding my second research question, several political implications of each of the different versions of environmental security have been discussed. The enactment of environmental security as local resilience might promote problematic forms of paternalistic control, providing scientists or NGOs with powerful means of controlling populations at a distance. Also, the enactment of environmental security as private risk management bears some normative problems. While the attempt to prevent dangerous environmental changes in the first place has been dismissed, environmental intelligence products allow those who can afford them to adapt to the inevitable disasters resulting from environmental changes. Finally, ‘big environmental data’ can be criticized for silencing discursive struggles on environmental security. When environmental security risks simply emerge out of the data, the only way to challenge them is on the basis of the data itself. The visual products of Copernicus or DigitalGlobe are presented as apolitical services for everyone’s benefit. This apolitical vision only works because all the flows of money, subjective and/or algorithmic decisions and acts of interpretation, scientific assumptions, legal norms, technical constraints, etc. behind every single image or data layer become black-boxed.

The explorative analysis presented in this article was a first step to open these black boxes, but it could only scratch at the surface of the underlying assemblages. Hence further research is required to dig deeper. This could be done by conducting more detailed ethnographic studies to further trace the different enactments of environmental security in concrete practices. Another avenue would be to further inquire into the political implications of the different enactments of environmental security, to search for possibilities to repoliticize them, and to appropriate the means of enacting environmental security to use them in more progressive ways.

Footnotes

Acknowledgements

I would like to thank the three anonymous reviewers of this article for challenging but very productive comments.

Funding

The research for this article has been funded by the Center for a Sustainable University, Hamburg.

Notes

Delf Rothe is a researcher and lecturer at the Institute for Peace Research and Security Policy Hamburg, Germany. He received his PhD from the department of International Relations of the University of Hamburg in October 2014. Rothe has published on the securitization of climate change, visuality, energy security, and resilience in journals including International Relations, Journal of International Relations and Development, Security Dialogue and Environmental Politics. He is the author of Securitizing Global Warming: A Climate of Complexity (Routledge, 2016). Email: Rothe@ifsh.de.

References

Aradau

Blanke

(2016) Politics of prediction: Security and the time/space of governmentality in the age of big data. European Journal of Social Theory. Epub ahead of print 14 September 2016. DOI: 10.1177/1368431016667623.

Armed Conflict Location and Event Data Project (ACLED) (2016) Armed Conflict Location and Event Data Project: About ACLED. Available at: http://www.acleddata.com/about-acled/ (accessed 26 April 2017).

Barry

(2013) Material Politics: Disputes Along the Pipeline. Chichester: John Wiley & Sons.

Bleiker

(2014) Visual assemblages: From causality to conditions of possibility. In: Acuto

Curtis

(eds) Reassembling International Theory: Assemblage Thinking and International Relations. Basingstoke: Palgrave, 75–81.

Bleiker

(2015) Pluralist methods for visual global politics. Millennium 43(3): 872–890.

Boas

Rothe

(2016) From conflict to resilience? Explaining recent changes in climate security discourse and practice. Environmental Politics 25(4): 613–632.

Brannon

(2013) Standardized spaces: Satellite imagery in the age of big data. Configurations 21(3): 271–299.

Buzan

Wæver

De Wilde

(1998) Security: A New Framework for Analysis. Boulder, CO: Lynne Rienner.

Chandler

(2015) A world without causation: Big data and the coming of age of posthumanism. Millennium 43(3): 833–851.

10.

Corry

(2014) From defense to resilience: Environmental security beyond neoliberalism. International Political Sociology 8(3): 256–274.

11.

Dalby

(2013) The geopolitics of climate change. Political Geography 37: 38–47.

12.

Dalby

(2015) Climate geopolitics: Securing the global economy. International Politics 52(4): 426–444.

13.

Dalby

O’Lear

(2016) Reframing the climate change discussion. In: Dalby

O’Lear

(eds) Reframing Climate Change: Constructing Ecological Geopolitics. Abingdon & New York: Routledge, 1–13.

14.

Danjean

(2014) The security dimension of Copernicus. Window on Copernicus. Available at: http://www.gmes-bridges.eu/content/window-gmescopernicus (accessed 12 April 2017).

15.

De Goede

Simon

Hoijtink

(2014) Performing preemption. Security Dialogue 45(5): 411–422.

16.

DeLoughrey

(2014) Satellite planetarity and the ends of the earth. Public Culture 26(2): 257–280.

17.

Detraz

Betsill

(2009) Climate change and environmental security: For whom the discourse shifts. International Studies Perspectives 10(3): 303–320.

18.

Diamond

(2010) Lenticular galaxies: The polyvalent aesthetics of data visualization. CTheory. Available at: http://www.ctheory.net/articles.aspx?id=651 (accessed 5 April 2017).

19.

DigitalGlobe (2014) Top Image Contest: 2014 (Top 25 Album). Available at: https://www.facebook.com/media/set/?set=a.10152960283921289.1073741830.132015741288&type=3 (accessed 26 April 2017).

20.

DigitalGlobe (2016a) FirstLook: Accelerate your crisis response. Available at: https://www.digitalglobe.com/products/firstlook (accessed 12 October 2016).

21.

DigitalGlobe (2016b) Satellite imagery for insurers. Available at: https://www.digitalglobe.com/industries/insurers (accessed 25 April 2017).

22.

Dodge

Perkins

(2009) The ‘view from nowhere’? Spatial politics and cultural significance of high-resolution satellite imagery. Geoforum 40(4): 497–501.

23.

Edwards

(2010) A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming. Cambridge, MA: MIT Press.

24.

Elam

(1999) Spyglass on the world: The culture and politics of remote sensing. In: Jelusic

Selby

(eds) Defence Restructuring and Conversion: Sociocultural Aspects. Brussels: European Commission, Directorate-General Research, 96–121.

25.

EU Copernicus (2016a) Application domains. Available at: http://www.copernicus.eu/main/application-domains (accessed 26 April 2017).

26.

EU Copernicus (2016b) Copernicus Emergency Management Service: Risk and recovery mapping products. Available at: http://emergency.copernicus.eu/mapping/ems-product-component/EMSN014_01RIOMAMORE_04RISK01POPVULNERABILITY_01DETAIL/1 (accessed 25 April 2017).

27.

European Commission (2010) Integrating Maritime Surveillance: Communication from the Commission. COM(2010) 584 final. Available at: https://ec.europa.eu/maritimeaffairs/sites/maritimeaffairs/files/docs/body/integrating_maritime_surveillance_en.pdf (accessed 11 May 2017).

28.

Floyd

(2010) Security and the Environment: Securitisation Theory and US Environmental Security Policy. Cambridge: Cambridge University Press.

29.

G-MOSAIC (2012) Product portfolio: Space-based products and services in support of EU External Action. Available at: http://www.gmes-gmosaic.eu/sites/gmes-gmosaic.eu/files/G-MOSAIC%20Product%20Portfolio_final_LOW.pdf (accessed 13 April 2017).

30.

G-Next (2012) Product portfolio: Space-based products and services in support of EU External Action. Available at: http://www.d-copernicus.de/sites/default/files/dokumente/G-MOSAIC%20Product%20Portfolio_Mar2012.pdf (accessed 12 April 2017).

31.

GeoBuiz (2015) Earth observation: From space to surface – online panel. Available at: http://geobuiz.com/join-the-panel-earth-observation-from-space-to-surface/ (accessed 12 October 2016).

32.

Gisat (2012) Map linking locations of armed conflict with main clusters of land cover changes 2003–2008 North Kivu-DRC. Available at: http://www.gisat.cz/content/en/news/gisat-news/813-new-land-cover-map-for-kivu (accessed 25 April 2017).

33.

GMES Support Team (2002) GMES Issues and Need for Action: Report of the 1st GMES Forum. Brussels: GMES Forum.

34.

GMES Working Group on Security (2003) The security dimension of GMES. Brussels: European Council & European Space Agency.

35.

Grove

(2010) Insuring ‘our common future’? Dangerous climate change and the biopolitics of environmental security. Geopolitics 15(3): 536–563.

36.

Hind

Lammes

(2016) Digital mapping as double-tap: Cartographic modes, calculations and failures. Global Discourse 6(1–2): 79–97.

37.

Houser

(2014) The aesthetics of environmental visualizations: More than information ecstasy? Public Culture 26(2): 319–337.

38.

Institute for Global Environmental Strategies (IGES) (2014) Sanctuary: Exploring the World’s Protected Areas from Space. Arlington, VA: Institute for Global Environmental Strategies.

39.

Institute for Global Environmental Strategies (IGES) (2016) Alliance for Earth Observations. Available at: https://www.strategies.org/environmental-information/alliance-for-earth-observations/ (accessed 12 April 2017).

40.

Jasanoff

(2004) Heaven and Earth: The politics of environmental images. In: Jasanoff

Martello

(eds) Earthly Politics: Local and Global in Environmental Governance. Boston, MA: MIT Press, 31–52.

41.

Latour

(1986) Visualisation and cognition: Drawing things together. In: Kuklick

(ed.) Knowledge and Society Studies in the Sociology of Culture Past and Present. Greenwich: Jai Press, 1–44.

42.

Latour

(2013) An Inquiry into Modes of Existence: An Anthropology of the Moderns. Cambridge, MA: Harvard University Press.

43.

Latour

Hermant

(2006) Paris: Invisible city. Available at: http://www.bruno-latour.fr/sites/default/files/downloads/viii_paris-city-gb.pdf (accessed 26 April 2017).

44.

Liverman

(2009) Conventions of climate change: Constructions of danger and the dispossession of the atmosphere. Journal of Historical Geography 35(2): 279–296.

45.

McDonald

(2013) Discourses of climate security. Political Geography 33: 42–51.

46.

Manzo

(2012) Earthworks: The geopolitical visions of climate change cartoons. Political Geography 31(8): 481–494.

47.

Marzec

(2014) Militarized ecologies: Visualizations of environmental struggle in the Brazilian Amazon. Public Culture 26(2): 233–255.

48.

Mayer

(2012) Chaotic climate change and security. International Political Sociology 6(2): 165–185.

49.

Methmann

Oels

(2015) From ‘fearing’ to ‘empowering’ climate refugees: Governing climate-induced migration in the name of resilience. Security Dialogue 46(1): 51–68.

50.

Methmann

Rothe

(2014) Tracing the spectre that haunts Europe: The visual construction of climate-induced migration in the MENA region. Critical Studies on Security 2(2): 162–179.

51.

Mirzoeff

(2014) Visualizing the Anthropocene. Public Culture 26(2): 213–232.

52.

Mitchell

(2013) Carbon Democracy: Political Power in the Age of Oil. London & New York: Verso.

53.

Mol

(1999) Ontological politics: A word and some questions. The Sociological Review 47(S1): 74–89.

54.

Mol

(2002) The Body Multiple: Ontology in Medical Practice. Durham, NC: Duke University Press.

55.

Moser

(2008) Making Alzheimer’s disease matter: Enacting, interfering and doing politics of nature. Geoforum 39(1): 98–110.

56.

Oels

(2013) Rendering climate change governable by risk: From probability to contingency. Geoforum 45: 17–29.

57.

Parks

(2009) Digging into Google Earth: An analysis of ‘Crisis in Darfur’. Geoforum 40(4): 535–545.

58.

Patton

(1994) Metamorpho-logic: Bodies and powers in A Thousand Plateaus. Journal of the British Society for Phenomenology 25(2): 157–169.

59.

Randalls

(2014) Climate change multiple. In: Stripple

Bulkeley

(eds) Governing the Climate: New Approaches to Rationality, Power and Politics. Cambridge: Cambridge University Press, 235–242.

60.

Rose

(2016) Visual Methodologies: An Introduction to Researching with Visual Materials. London: SAGE.

61.

Rothe

(2011) Managing climate risks or risking a managerial climate: State, security and governance in the international climate regime. International Relations 25(3): 330–345.

62.

Rothe

(2016) Securitizing Global Warming: A Climate of Complexity. London & New York: Routledge.

63.

Schneider

Nocke

(eds) (2014) Image Politics of Climate Change: Visualizations, Imaginations, Documentations. Bielefeld: Transcript.

64.

Schouten

(2014) Security as controversy: Reassembling security at Amsterdam Airport. Security Dialogue 45(1): 23–42.

65.

Shim

(2014) Remote sensing place: Satellite images as visual spatial imaginaries. Geoforum 51: 152–160.

66.

Thompson

(2007) A political history of U.S. commercial remote sensing, 1984–2007. PhD thesis. Virginia Polytechnic Institute, Virginia State University, Virginia, VA.

67.

Trombetta

(2008) Environmental security and climate change: Analysing the discourse. Cambridge Review of International Affairs 21(4): 585–602.

68.

Van Munster

Sylvest

(2016) Introduction. In: Van Munster

Sylvest

(eds) The Politics of Globality Since 1945: Assembling the Planet. London & New York: Routledge, 1–19.

69.

Wark

(2015) Molecular Red: Theory for the Anthropocene. London & New York: Verso.

70.

World Meteorological Organization (WMO) (2016) Global Observing System. Available at: http://www.wmo.int/pages/prog/www/OSY/GOS.html (accessed 25 April 2017).