Moving with risk: Forced displacement and vulnerability to hazards in Colombia

https://doi.org/10.1016/j.worlddev.2021.105482 0305-750X/ 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ⇑ Corresponding author at: School of International Development, University of East Anglia, Norwich Research Park, Norwich NR4 7TJ, United Kingdom. E-mail addresses: r.few@uea.ac.uk (R. Few), T.Armijos@uea.ac.uk (M.T. Armijos), lzambrano@umanizales.edu.co (L.A.Z. Hernández), Hazel.Marsh@uea.ac.uk (H. Marsh). Roger Few a,⇑, Viviana Ramírez , Maria Teresa Armijos , Lina Andrea Zambrano Hernández , Hazel Marsh d


Introduction
The risks to livelihood and wellbeing generated by forced displacement are multiple, profound and relatively well documented (e.g. Thomas and Thomas, 2004;Christensen and Harild, 2009;Fiala, 2015;DeJesus, 2018). However, the association between displacement and subsequent exposure to hazards has received relatively little research attention, for populations that are primarily self-settling as well as those in externally-organised refugee camps (Pollock et al., 2019). Here we bring insights from diverse litera-tures together with findings from empirical research to help bridge this gap. Rooted mainly in the field of critical research on disaster risk, but drawing also on discussions of conflict, displacement, migration, participation and representation, this paper focuses on the reproduction of vulnerability through forced displacement in Colombia. It traces the social and spatial processes through which IDPs commonly become exposed to heightened risk from hazards in the places where they resettle (Siddiqi et al., 2019). The term 'hazards' can have multiple meanings, but here we refer essentially to threats emanating from the physical environment, including landslides, floods and urban fire hazards. The movement into hazardous zones is a process that is most apparent for lower-income households that have fewer livelihood assets to draw on (though by no means exclusively the poorest), and for those typically experiencing high levels of economic, gendered and cultural exclusion https://doi.org/10.1016/j.worlddev.2021.105482 0305-750X/Ó 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). (Peters, 2019). In effect, such people displaced by violence exchange one form of risk for another.
Based on data collected through an iterative combination of interviews and participatory research methods, we discuss how and why this occurred, both structurally through the constraining of options and through conscious choice, and how this ongoing vulnerability of IDPs often becomes deepened by new elements of marginalisation in the resettlement sites. Yet, we also indicate how, in part at least, there is the potential to reduce vulnerability through communal efforts and especially through efforts to bridge this marginalisation on the part of communities and authorities alike. Drawing on Ramírez Loaiza et al. (2017), we see an opportunity for displaced communities to demand social dialogues for disaster risk reduction (DRR) by sharing their remarkable stories of resilience and making visible alternative narratives that 'may uncover how people have been unfairly blamed for actions taken, not taken or still taken in disasters' (Bankoff, 2012, p. 40).
Following a brief literature review linking work on disaster risk and forced displacement, the paper introduces the case study context and research methods. We then present empirical details from the study, tracing the processes of displacement and resettlement, re-establishment, ongoing marginalisation and partial detachment from existing DRR mechanisms. A discussion section draws these points together, followed by brief concluding thoughts.

The risks downstream: Linking displacement and vulnerability to disasters
Globally much remains to be developed in the literature on migration and consequent disaster risk, and this research gap is especially evident in the case of people forcibly displaced who arguably tend to have fewer available assets and options to plan or choose where they resettle (Innes, 2015;Armijos and Few, 2017). The paper's title 'Moving with Risk' reflects this concern. The physical character of the danger people face may have changed, but the movement to hazardous rural landscapes where space for farming is available or to high-density peri-urban and urban areas with inadequate infrastructure often located on land not zoned for residence, means that multi-hazard risk commonly 'travels with' those displaced by conflict. Effectively, such people exchange one form of catastrophic risk for another, sometimes knowingly, sometimes unknowingly, but often with little real choice in the process.
The framing of risk employed in this research is broadly aligned to a political ecology conception of hazards. The basic tenet of this approach, underlined many times in the disaster literature (e.g. Hewitt, 1983;Cutter, 1996;Wisner et al., 2004) and inscribed in the international agreement, the Sendai Framework for Disaster Risk Reduction (UNISDR, 2015), is that disaster risk cannot be viewed purely in terms of the physical characteristics of the hazard. In order for a disaster to occur there must be vulnerable people who are exposed to the hazard and whose lives and livelihoods are susceptible to its effects. Disaster risk in this sense is a social condition rather than a temporally delimited event, with long-term structural antecedents and long-term implications for the trajectories of people's lives (Cutter et al., 2008;Hicks and Few, 2015). Work on climate change risks has mirrored and reinforced this approach, underlining that vulnerability is also shaped by positive capacities -capacity to adapt to change and to build resilience in the face of risks (Adger, 2006;Tanner et al., 2015).
In this light, however, it should also be noted that there is a body of critique that contests the very idea of constructing people's identities as 'vulnerable' and emphasises the need to understand people's relation to risks through their own eyes and voices (Bankoff, 2004;DeJesus, 2018;Gaviria and Zambrano, 2019). Con-ceptions of risk and the prioritisation of different forms of risk will inherently differ from person to person, group to group, and this may particularly be the case for IDPs, who know about threats only too well. It is therefore crucial for institutions working on disaster management not only to understand how people themselves view and talk about the different risks they are exposed to, but also to create spaces where knowledge about ways of dealing with these risks is produced and shared.
However, creation and occupation of such spaces can be challenging in many respects, not the least because of entrenched issues of distrust on both sides. In Colombia, many people who have been displaced by conflict have a deepened sense of distrust of authorities, a sense embedded in a context of historically strained state-society relations (Burnyeat, 2020;Weber, 2020). Conversely, IDPS themselves suffer high levels of discrimination and stigmatisation; it is common for host communities to categorise them as prone to criminality, and as unfairly benefitting from state assistance (Ángel, 2019). Arguably, the humanitarian category of 'internally displaced person' labels and stigmatises people, fixing them in particular social positions that can undermine their dignity, restrict their possibilities and continue to produce uncertain futures (Ángel, 2019;Brun, 2015). For similar reasons, in our interactions during the research we carefully avoided labelling people as 'victims' or 'displaced'.
It is also important to recognise the complex ways in which the experience of internal displacement intersects with, and is shaped by, gender dimensions, ethnicity and social inequalities, as well as by political, social, cultural, and economic discrimination. Meertens (2010) argues that surviving women have been disproportionately affected by the consequences of armed conflict in Colombia, as single heads of household experiencing forced displacement, through sexual violence by armed actors as a weapon of war, and because of the historical lack of recognition of women's rights that has allowed for the violent seizure of their land. Indigenous groups have experienced high levels of internal displacement, and, compounding their unfamiliarity with livelihood options in other regions or in cities, many indigenous people, especially women, are not fluent in Spanish, leading to further deprivation, disorientation and discrimination (Paarlberg-Kvam, 2017). Afro-Colombians' experiences of internal displacement have also been entangled with other structures of violence, such as racism and extractivism (Cárdenas, 2018). For complex reasons relating to ongoing threats, discrimination and stigmatisation, racism, remote location, socio-cultural differences, economic marginalisation and/ or informal land titles, those who have experienced forced displacement are likely to be hard to reach under conventional scales of service provision and disaster risk management. This study contributes towards a better understanding of how those who are, for cultural and material reasons, left poorly attended by formal state support, end up relocating into situations of risk, how that risk becomes compounded by exclusion, and how institutions might better work to support these populations.

Case study context & methods
The last 50 years have seen millions of people forced from their homes by armed conflict in Colombia between guerrillas, paramilitaries and military forces. As of February 2020, the national agency responsible for assisting those affected by conflict, Unidad para la Atención y la Reparación a las Víctimas (UARIV), had recorded just under 8 million people internally displaced by violence or the threat of violence, equivalent to 16% of the country's current population (www.unidadvictimas.gov.co). Many have self-settled in new rural or urban locations (Shultz et al., 2014). Existing research on forced displacement due to conflict in Colombia has revealed how displacement can trap people in poverty through the loss of assets from short-term economic coping mechanisms and persistent social, political and cultural barriers to economic mobility in the places where they resettle (Petesch and Gray, 2009;Ibáñez and Moya, 2010). The disproportionate impact of displacement and marginalisation on the vulnerability, security and rights of women and children has also been explored (e.g. Alzate, 2008;Lemaitre & Sandvik, 2015;Meertens, 2010).
To date, however, little research has been carried out on how the economic and social repercussions of conflict-forced displacement interact with exposure to alternative forms of disaster risk. Colombia faces a complex mix of so-called 'natural' hazards, including landslides, floods, drought, earthquakes, and volcanic hazards, together with exposure to more 'anthropogenic' threats such as fire disasters and industrial accidents. Figures compiled by the national disaster risk management agency, Unidad Nacional para la Gestion del Riesgo en Desatres (UNGRD), indicate that around 20 million people were affected by disasters in the country from 2000 to 2019 (http://portal.gestiondelriesgo.gov.co). Many hazards are sporadic and infrequent, but those associated with intense rainfall are an annual occurrence. Both rural and urban areas are affected, and landslides in particular can be especially devastating in poor urban neighbourhoods where incoming migrants commonly reside in informal settlements on steep terrain (Zeiderman, 2013). A prolonged series of flood and landslide disasters in the country in 2010/2011, for example, displaced approximately 2 million Colombians, and Shultz et al. (2014) note that many of those had been previously displaced by armed conflict. The informal urban settlements to which lower-income IDPstypically move are also places of high fire risk to homes (Carillo, 2009), owing to the close proximity of the dwellings, the flammability of wood and bamboo constructions, and the unsafe nature of electric, gas and woodfuel sources. We are not arguing that IDPs are necessarily more exposed to physical risk than non-IDPs in such locations, but we do contend in this paper that their life experience of multiple forms of risk shapes their ongoing vulnerability in complex ways.
In the research reported in this paper, work took place in 2017-2018 within four case study locations in Colombia: the city of Manizales in Caldas; the settlements of Caimalito and Galicia (official name Esperanza Galicia) in Risaralda; and the area known as Cazuca (or Comuna 4) in the municipality of Soacha in Cundinamarca. There we worked in urban/peri-urban neighbourhoods selected because they host high proportions of people displaced by conflict, and because they are classed as at risk from hazards, particularly landslides, flash flooding and fires. Located on the outskirts of Bogota, Cazuca is a major centre of resettlement for IDPs in Colombia, but each of the case studies were in key resettlement areas for regionally displaced populations. Their mix is broadly representative of the scale range of urban and peri-urban IDP sites across the country. Not all IDPs in the cities and municipalities selected for the study occupy hazardous zones, but it was this new form of risk that many face that was the focus of our specific site selection, and, as noted above, such conditions are common in many IDP resettlement locations nationwide.
Ethnically, each site had a mix of people identifying as Mestizo, Afro-Colombian and Indigenous groups. All of the study sites were originally informal settlements, and most continue to be so: technically illegal, known colloquially as 'invasiones', subject to potential eviction, and with weak or minimal access to utilities and services. Some households and neighbourhoods had become formalised in recent years, acquiring legal titles to the land, but, in general, their populations have remained economically marginalised. Households across all sites had achieved access to electricity, but the majority did not have access to gas or potable piped water. The trajectories of displacement and arrival had varied greatly for households we worked with, both across and within the sites, with a range of arrival dates spanning less than one year to 30 years prior to the time of the study. Table 1 provides further information about each of the four study contexts (note that, for reasons of sensitivity and privacy, we did not attempt to gather systematic information about where people originated from and which armed group caused their displacement; we also do not show the specific names of the neighbourhoods or 'barrios' in which we worked).
Although all of our sites were in locations of significant hazards, there were important differences in how the land was classified, between medium-and high-risk designations. Law 1523 (2012), which set in place the current disaster risk management system in Colombia, classes a 'mitigable' risk as a situation in which corrective measures could be set in place to decrease hazard exposure and vulnerability to an acceptable level. In cases where this is not considered feasible by the state authorities, the risk is classed as 'non-mitigable', and the land to which this applies becomes zoned as high-risk within the municipal land use plan, the 'Plan de Ordenamiento Territorial' (or POT). We will return to discuss the implications of this zoning mechanism later in the paper.
Through the course of the project, the research team held a series of meetings with governmental and Red Cross organisations, to introduce, discuss and authorise the research, establish links for joint community-institution engagement work, and present the findings of the research. A total of 8 national level and 23 municipal level meetings were conducted. These interactions provided an important, triangulating layer of contextual insight. However, the prime focus for our research was on the voices and perspectives of IDPs themselves within the four study sites.
Data collection in each site was based on qualitative interview and lifecourse methods, but with incorporation of creative elements in an effort to humanise the exchange between participant and researcher and create a stimulating but 'safe' space for discussion (O'Neill, 2008). The challenge to work with IDPs who are also at risk of disaster, required a sensitive approach to data collection, investment of time to build trust, and an ethic that compelled us to design methodologies that were mutually beneficial rather than merely extractive. Intensive introductory and post-research engagement work also took place with each community and institutional stakeholders, but for the purposes of this paper we focus primarily on the data collection methods (discussion and reflection on the full methodological approach in the project is provided in Marsh et al., 2020). The research interactions with community members took place in three main phases. The first we referred to as 'Conversaciones'. These interactions with individuals and household/neighbour groups took the form of semi-structured interviews, but, building on the work of Anderson (2004), O'Neill (2008) and Allett (2011), we developed an interview process that centred on the use of music/art as a medium to elicit conversation and to create trust between the researchers and the participants. Prior to addressing the interview questions, in each interaction we first invited participants to share with us a song, image or artefact that had significance for them, and to talk to us about how this particular tangible or intangible object acquired meaning for them in relation to their past or present experiences. Through a reflection on these experiences we then asked questions about the risks they have faced and their migrations, their survival strategies and needs, their livelihood transitions, their social networks and civic engagement, and their own understandings of what constitutes wellbeing for their communities. The focus was not squarely on the negative experiences of risk but also on the capacities and resources that have been deployed by people in their efforts to re-establish their lives in new locations. In all, 103 sets of interactions took place (24-30 per site) with 138 participants (84 women and 54 men). For ethical and practical reasons, it was not feasible to follow random sampling procedures in this research, and, in effect, all participants were self-selected volunteers. However, the sample was recruited through an iterative process starting with initial meetings with community leaders and invitations to all residents to a public meeting, and followed up by visits to households by the research team, designed to encourage a broad crosssectional representation of IDPs (additional details of this process are provided in the Online Appendix).
For the next phase, the research team gathered their own reflexive feedback on what they had learned as researchers and people from the conversations they had with different people, and displayed them with photographs and drawings of the territory to create a 'Museum of Re-encounters' to which community members were invited. This forum then opened up discussion space for reflections on how people faced new risks in their life trajectories. In the third phase, in each study site we selected 8 of the original participants (with a range of distinctive stories) to reflect further on the decisions made and capacities deployed through their lifecourse, using a technique we called 'HistoriAndando' (or 'Walking History'. We presented a large-format visual timeline and sequence of drawings of key moments from that person's testimony and the participant was encouraged to walk through the story and discuss those moments in greater depth (for more detail on this approach please see the Online Appendix).
The interviews and lifecourse discussions conducted through the Conversaciones and HistoriAndando methods were designed to produce 'thick descriptions': detailed accounts that offer insights into what people believe and value, how they view events and processes, how they describe the world and how they represent themselves and their capacities (Rapley, 2006). The interactions were recorded and transcribed, and content analysis consisted of reading the transcripts to identify key themes. From this reading a series of codes were developed and then applied to each of the transcripts, double-coded as a verification check, and compiled using NVivo software as a means to compare and draw commonalities across people's narratives (Coffey & Atkinson, 1996). It is to these findings that we now turn our attention.

Tracing trajectories
In this main empirical section of the paper we trace the stories of how displaced people in the Colombian study sites have 'moved with risk', describing commonalities and differences in their trajectories, in the current issues they face and in their capacities to engage in disaster risk reduction in the places where they have resettled. In the accounts that follow, all statements made by the research participants are anonymised using pseudonyms.

Displacement and resettlement
First, we look at the narratives of how people came to resettle from their former homes, covering the processes and decision influences that accompanied their move of often hundreds of kilometres to places where they could find land to occupy. It is important to note that in our interactions with research participants we tried to focus not on the causes but on the outcomes of forced displacement. We took careful steps not to orient our questions on the trauma of the displacement events themselves. Nevertheless some participants openly wanted to convey this part of their stories and it is therefore appropriate for us to reflect those desires briefly here.

Underlying vulnerability
One point to emphasise is that the forcing of displacement often occurred in the context of other profound pressures on economic livelihoods and wellbeing. Many of those we worked with had experienced pre-existing vulnerability conditions that, in part, would shape subsequent trajectories. Notable among these were underlying poverty, gender-based violence and environmental hazards. Some referred to lifetimes lived in poverty that have always prevented more secure options. As José (pseudonyms used) from Cazucá noted: 'we'd have better houses, we'd live in other, better places, but as you see . . . we've been poor all our lives'. There were women in all the study communities who spoke of their exposure to gender-based violence in their original places of residence, some connecting it with poverty: 'if I ever said anything, he used to say ''your word is worth nothing" because of course, since I was poor, they believed him more' (Carmelina, Caimalito). For some, this was itself a trigger to leave a home village or at least create the idea of seeking refuge elsewhere: 'I was crying, I wanted to die, I wanted to get away' (Diana, Galicia).
It is also important to note that there were people in our study who had experienced l hazards in previous living places. More than 20 participants across the study sites spoke of flood and landslide risk in the same places where they experienced the armed conflict, and this may have shaped their subsequent view of hazards. One person observed 'I have always been on high-risk land' (Rosa, Caimalito), and another described how he and other farmworkers routinely cleared landslides with spades and pickaxes in his former rural area (Armando, Manizales). While a number of people reported that hazards had damaged their assets, this seldom represented a reason to move out from their homes. The impacts and the imminent threats associated with armed conflict generally marked a turning point in the imperative to leave. As Nelly (Caimailito) explained: 'it was difficult to cope with rivers and storms . . . but the real suffering for us was the war'.

Conflict and forced displacement
The immediate causes of forced displacement represented in the four case studies included: the murder of a member of the family; sexual violence; occupation of the home by armed groups as a place to keep weapons, sleep and eat; verbal threats; forced recruitment of children; combat between armed groups for the control of certain areas; and pressure to carry out illicit activities, such as coca cultivation. Diego from Cazucá explained how a neighbour's warnings had saved him from a large group of armed men who had come looking for him at his former home. Another research participant explained: 'I had to get out of that farm because the guerrillas wanted to take my two lads, so I resisted and instead sent them to their grandparents and that's why I had to leave, because I didn't want to cooperate' (Raul, Caimalito) Displacement became the only way to survive and seek greater security for all family members. For those people who did not receive direct verbal threats, the context of armed conflict by itself also forced the decision to move to another place. Harassment by the armed groups in their daily life led families to feel at permanent risk. Many of the participants remembered how they hid from armed confrontations, how they kept silent when combatants were close to their homes, how they had to serve drinks and food to combatants, and even how young women would pretend to be breast-feeding to avoid being raped. The need to distance themselves from the armed groups ultimately became the overriding motivation in terms of wellbeing, despite the difficulties inherent in moving away.

The process of moving (and moving again)
Broad social differences were apparent in the process of moving and attempting to find a place of refuge and ultimately resettlement. In the case of the indigenous communities that we worked with in Pereira, they had generally decided to move as a group to reach the cities or their peripheries. There they would look to appropriate a piece of land informally, each family setting up its shelter but with the group always remaining together, to help cope with new problems and challenge actions by the authorities to dispossess them of these areas.
In the case of the Mestizo and Afro-Colombian families with whom we worked, and the indigenous families that moved to Cazucá, the pattern tended to be more individualised. The family's choice of where to go after forced displacement was based on having social contacts in the area, having lived in nearby locations and knowing about the costs of living and possibilities of acquiring or appropriating land, or knowing of places through previous visits from their territories of origin (often to seek work). Esperanza explained how her nephews had previously relocated to Caimalito and then found a house that she could move to: 'so they found out about this place . . . bought this one and told me to come here'. Having friends, relatives or former neighbours in the destination site was commonly a key factor in the decision, and these contacts often contributed to supply basic needs for a while, help plan their next steps and help look for new income sources.
The complex political context of the conflict brought with it heightened tensions and mistrust when people were attempting to settle within new social settings. Many remained reticent about the circumstances of their displacement, as the impacts of the conflict left people deeply polarised in their perspectives on the different armed groups. During the research it became clear that people displaced by opposing forces were often living side by side.
However, after the forced displacement, people had seldom moved directly to the places where they were now settled. The majority of the research participants transited between two to four territories looking for economic opportunities, escaping again from armed groups or from gender-based violence, or being evicted from informal dwellings or because of rent arrears. Estela (Galicia), for example, told us of two forced displacements by the same armed group in different locations. During this transitional period, or, as some referred to it, 'wandering time', at least 12 of the research participants indicated thatthey had takensignificantsteps tore-establishtheir lives in intermediate locations, which ultimately failed: 'I arrived in Bogotá . . . with all that I'd worked [for] I bought these two really cheap plots of land, I didn't know you weren't allowed to buy this land, and I bought these two plots of land and the house. I had a little car then and I lost it, I lost everything. The Mayor came one day and gave the order to evict us' (Carmen, Cazucá).
It is important to note that none of the research participants referred to accessing/receiving governmental aid in the immediate aftermath of their forced displacement, let alone assistance in relocating. The aspiration to find a 'safe' place to live effectively overrode the insecurities posed by occupying illegal lands and/or sites at high risk from hazards. The decision to resettle was one of the most significant life moments for many participants because it reduced their sense of isolation and, notably, of vulnerability.

Re-establishment
Given their predominantly rural origins, the research participants faced major challenges in adapting to a new existence in sites where a former farming lifestyle was no longer feasible. Housing issues were central in this. The people that we worked with essentially had three choices in terms of resettling in the informal sites that have since come to be their established home: they could occupy (rent or buy) an existing dwelling; they could buy land acquired by someone else and build on it; or they could appropriate a vacant lot and build on it. For many, the economic stress of paying rent was not only difficult, but unfamiliar: 'In Pereira, when they told me that [I had to] pay rent I said, what's that? Paying rent, what's that? It's just that I'm from the mountains. . .. so I didn't know what paying rent was' (Lorena, Caimalito). People also emphasised the importance of having the opportunity to adapt their living space to permit some rural practices in their daily life, because otherwise: 'the city's like a prison. . . because the freedom you have in the countryside, you don't have it here' (Marcela, Manizales). Accordingly, many people's narratives highlighted the crucial role of self-construction as a milestone in achieving a sense of security and opening up the chance for recovery, even if in most cases the land was informally occupied.

Constructing homes to stop 'wandering'
For many, creating their own permanent home, however simple, was seen as a highly significant symbolic act in that it enabled people to break the cycle of displacements and, in their words, 'stop wandering'. Laura explained how she decided to transition from renting in Manizales and instead begin to construct her own home: ' . . ...and that's what I did, look, I put together just one room and bit by bit I built it up slowly, and look how I've got it now, just with reed mats for walls, but me and my children are proud of it because we stopped wandering' (Laura, Manizales).
During the search to find a place to settle, the participants also valued places where there was some semblance of similarity to their homeland environments: 'So I got here. . .. I liked it because you can see the river, it's much like what you used to see where I'm from' (Juan, Caimalito). Across the case studies and the different ethnic groups there was also high value placed on the chance to create a space where there could be a small garden or a yard, to adapt living space to recreate aspects of a past home environment and regenerate a sense of belonging. Sara saw the place she had made as somewhere she could still nurture plants: 'Look (I have) my plants, hanging there . . . because in the countryside I grew everything' (Sara, Cazucá). Despite living in a high-risk zone, Alirio viewed his small plot as an opportunity to grow crops: 'Here you're not going to just sit twiddling your thumbs, here there's going to be a lot of [vegetation], I've planted everything' (Alirio, Cazucá). This common theme seems to have been key in helping our research participants finally achieve the notion of a 'safe' and cherished home.

A collective goal
Even in cases where the arrival of displaced families was uncoordinated, 'land invasion' and the development of informal settlements came to be a collective project that fostered a degree of community organisation. After families built their basic shelter, the next step would be to bring water and lighting to their new homes: 'The whole community worked together to get electricity, water, gas . . . that's like a basis that really helps to strengthen relationships and empower ourselves as people' (Luisa, Cazucá). With a growing sense of solidarity, these grassroots efforts could gradually turn toward gaining access to basic public services, improving transport links and coordinating resistance to eviction by the police authorities. Although the economic conditions of many participants remained unstable and precarious, these collective goals led many to feel fully established and secure in their situation: I speak for this family and at a community level. The people who settle here, feel good here. . . We have the primary school, the high school, electricity, water, gas and everything necessary to live here' (Jairo, Galicia).
These transitions also provoked a change in gender roles, because many women had to use their knowledge and skills to bring in economic resources and to learn new work activities. Most of the indigenous women in the Pereira sites indicated that resettling in groups or extended families was a key advantage to their recovery. This allowed them to travel safely together to the city to sell items such as beaded necklaces and bracelets that they made in their new homes -one economic option that was possible in the absence of land for crop farming.

The threat of eviction
Being free from the burden of paying rent enabled poorer households in informal settlements to focus not just on other primary needs but to envision a situation where they could meet future aspirations for their families. For many, the principal concerns were therefore not about the often palpable evidence of hazard risk in the vicinity, but in the threat of being evicted by the government authorities: The Colombian state has the power to remove people from settlements classed as illegally occupied, and most occupants of all the sites in which we worked either were still, or had in the past, been subject to the threat of eviction. In practice, attempts at eviction are commonly ineffective, but the threat of repeated displacement hangs over those attempting to create a new home. Prior to 2017, such eviction could be undertaken without provision of alternative land, but legal changes mean that it is now incumbent on the state to provide a planned relocation site before eviction can take place.
Across the different study sites, there were cases where some households or groups had managed recently to formalise their settlement and gain legal title to the land on which they had constructed their homes. Law 1848 (2017) establishes criteria for the formalisation, titling, and recognition of properties. In cases that meet these criteria, an applicant must request their interest by letter to the municipal planning department and obtain an authorisation that permits residential use. The processes are complex and difficult and, even if the people we worked with could negotiate these hurdles, there were a number of conditions that prevented entitlement being granted. One of these applies if the location is an area designated as of high-risk in the land-use planning tool, the POT. Indeed, in areas zoned as of high risk there is a legal expectation on municipal authorities to remove dwellings. For residents in such settlements, a permanent threat of eviction is therefore reinforced by land-use planning regulations.

Ongoing marginalisation
Despite the gains they feel they have made in establishing homes and communities, however, the research participants also emphasise aspects of ongoing economic, social and political marginalisation that constrain their options and leave them on the fringes of public decision-making. Though common generally for low-income migrants and informal settlement dwellers, these issues seem especially entrenched for people who have been displaced, given the prejudices against them and their own distrust of others born from their life experiences.
Only a few of the people we worked with had gained a formal occupation, with access to health insurance and pension contributions. Most of those who work, across both urban and peri-urban sites, have found informal work such as providing street food, cleaning houses, or casual jobs in construction. Attempts to claim government welfare support for which they were eligible in principle were commonly unsuccessful, or delayed for long periods, and some seemed not to be aware of their rights. Under Law 1448 (2011), the government is mandated to provide aid and reparation for those registered as 'victims' of the armed conflict, including a set of measures intended to provide compensation for losses and assistance in terms of housing, livelihood rehabilitation, healthcare and psychosocial support. However, the bureaucratic hurdles to securing aid had proven difficult for many. The first hurdle was obtaining official proof and certification of IDP status. One research participant expressed how her mother had come up against this repeated barrier when she tried to claim, with the only suggestion offered being for her to return to her original town: 'They'd say go back to where you used to live and ask the Inspector for the letter and my mum would say 'but which inspector?' The Inspector was the first person to leave the town when the violence started' (Sonia, Galicia).
Yeny from Cazucá had obtained the official certification, but complained that 'even though I have the letter. . . they've never given me any help'. Those who did receive aid including cash, food, cloth-ing and employment training were generally critical of the quantity and process of disbursement. But the state itself was not wholly responsible for such issues, as complaints of corruption were levelled at some cases where community-level structures had taken on the organisation of claims and disbursement. As Ricardo stated: 'sometimes the aid doesn't arrive, just the paper for you to sign. . . and you sign it and then nothing arrives' (Ricardo, Caimalito).
For some people at least, a disconnection from external support was actually a conscious choice. There were cases in which people decided not to disclose their forced displacement status because they were afraid of being found again by the armed group that victimised them originally. Some people also wanted to avoid what they perceived as the public stigma of being classed as an IDP: 'the truth is that displaced people are discriminated against badly, they're always seen as thieves . . . people are only taught about the war through the news' (Julián, Manizales). Not applying for support as 'victims' could also be conveyed as a demonstration of family dignity in terms of recovering independently: 'When we arrived, we didn't pass ourselves as displaced people because none of my children wanted to, they said 'no mum, let's not beg'. . . my son. . . he said 'no mum, we haven't lost our hands or our lives', so we started working and providing' (Marcela, Manizales).
As noted above, some communal initiatives have been taken to strengthen access to services and support, particularly within the indigenous communities that arguably face the greatest barriers in terms of language, culture and historical recognition of rights.
In an effort to retain social cohesion, members of the indigenous communities that moved in and around Galicia and Caimalito set up a community-based authority structure that functions as the main channel with which to communicate with other institutions and implement projects. Though not immune from complaints, its existence seems to have facilitated the receipt of assistance and reparation measures from UARIV. In addition to that, an NGO has supported social programmes and educational projects for other ethnic groups in Galicia.
In contrast, in Manizales and Cazucá few participants considered that there had been any significant external role in supporting their recovery, and most argued that the initiatives that had been offered were mismanaged. Experiences of broken promises made during political campaigns, 'extractive' research undertaken by universities, and a generally perceived lack of support by the government all solidified a generalised distrust of external agencies. Consequently, many participants were not willing to engage in governmental projects unless they provided clear-cut economic benefits.

Detachment from disaster risk reduction (DRR)
Colombia has a relatively sophisticated institutional structure established for managing disaster risk, extending through national, state and local level agencies. The UNGRD and the sub-national branches that it oversees through the country-wide system for disaster management (SNGRD), have a mandate for overseeing and coordinating DRR, including pro-active prevention to minimise the impacts of hazard events. The Red Cross (Cruz Roja de Colombia), together with Civil Defence and the emergency services, take key operational responsibility for managing the onset of disaster events, in emergency response and efforts to assist recovery. As in most countries, both DRR and crisis response functions in Colombia face many administrative and political hurdles that impede their effectiveness on the ground (Quintero & Thomas, 2018;World Bank, 2012). However, this service gap is most profound in informal settlements, where a combination of regulatory and non-regulatory barriers exists to the provision of riskrelated assistance.

The gap in assistance
Previous emergencies that had occurred in the study sites, including floods, landslides and fires, indicated that emergency response in these communities is highly challenging. The informal development of most of the settlements in which we worked, with houses commonly densely built on difficult terrain with narrow walkways, makes them difficult for emergency vehicles to access. However, 'access' issues are more than physical.
Officials from some of those institutions mandated to protect Colombian citizens from disaster described a conception of IDPs in our study sites as people who are 'hard to work with' and expressed distrust around some people's claims for needing assistance. Under Law 1523 (2012), a national protocol has been established for identifying needs and distributing relief for disasteraffected people (UNGRD, 2013), including food aid, temporary shelter and housing subsidies. However, securing the housing subsidies is an administratively demanding process for households. Moreover, the regulations state that this is to be distributed only to legal owners of property or those who legally rent homes, denying support for repair of damaged properties or for rehousing to those who occupy land informally.
Evidence from the research participants and from meetings with government officials also indicates that pro-active DRR to protect the communities from hazard impacts has seldom occurred. Under SNGRD, the emphasis to date has been on much more on engineered mitigation structures than on social DRR programmes to engage at-risk communities (although some steps are being taken now through a new community-oriented programme, COMUNGERD). Even the IDP households in our studies whose occupation has now been legalised still felt neglected by DRR institutions, despite the existence of hazard risk in their locality and the palpable need for support in mitigation and preparedness. For those in informal settlements, denied access to most services, externally-provided DRR measures were even less likely to exist. This was most starkly so in zones designated as of non-mitigable, high risk from hazards. There, instead, the only risk-related government policy formally in operation was to try to enforce evictions of informal settlements.

People's response
Most of those we worked with in this study had no immediate plans to move. Through tracing life stories we came to a detailed understanding of how they had arrived at the sites and how they were using their capacities to try to re-establish themselves there. In most cases it was not that people were unaware of the potential physical hazards, nor that they would not relocate if they had a clear and workable option to do so. The problem was that very few have ever been offered a viable resettlement plan.
As is clear from the foregoing accounts, moves to organise evictions in each of the case study sites exacerbated feelings of distrust toward authority, and provoked again the sense of vulnerability that residents experienced through armed conflict. The participants felt in danger of being 'displaced again' but in this case by the government: 'They told me 'you have to leave, because look at the risk for your children', and I said 'where am I going to go?', and I said 'sorry but I'm not going to go'. . . they gave me four months to get out and I didn't want to leave. . . because imagine, I'd move out and some quick-witted person would move in and make a good home, and I'd be out there suffering with my daughters' (Rita, Galicia).
It was not necessarily the case that all people were against the idea of relocation from the hazardous sites, but they were against the idea of a forced eviction with no suitable alternative site provided. In response, eviction threats had generally been resisted, even when homes were destroyed. As Rita went on to explain: 'they'd demolish where we lived, and we'd stay there and build it again' (Rita, Galicia). These actions appear to have severely weakened already-strained relationships between people and government institutions. On the one hand, the government disaster management agencies were prevented from acting according to policy. On the other, many people were caught in a double jeopardy of unfulfilled entitlement to support as IDPs and minimal recourse to support in reducing the disaster risk they faced -other than 'the solution' of another forced relocation.
Although the risk from environmental hazards had not been a priority for the participants in deciding on a place to settle, most of them recognised that they face these risks and that they needed to figure out for themselves how to reduce the threat to their houses and their lives. Several participants had acquired private loans to increase the physical resilience of their houses, to strengthen the supporting structures for dwellings built on slopes or to make the materials less flammable. Yet, many participants could not access loans as they had no credit history or sufficient assets as a guarantee for the debt. Notably, some participants who did manage to receive economic support from the government or the Red Cross as people affected by conflict or by a hazard event used this money to strengthen their houses, even if that was not the stated purpose of this assistance: '. . . so I repaired the back of my house, I plastered the wall, organised the ditch. . . so that when it rains there aren't any leaks, all the water goes into the gutter, so I invested all the money in this' (Marta, Manizales).
Mostof thecommunities weworkedwithhad been affectedby one or more recent hazard events, and the experience had prompted some limited changes in preventive and preparedness behaviours. After a fire in Galicia, one participant had learned the following: 'on the subject of prevention, always turn off the gas, be careful with electrical things, don't leave anything turned on, watch the children, because the incident might happen again' (Isabel, Galicia). Collectively, however, more proactive DRR approaches might have potential. There were signs of this in Manizales, for instance, where groups of neighbours had used legal tools to demand structural interventions in their territories including retaining walls in order to mitigate the risk of landslides. They did so first by resisting attempts to evict them, using the argument that in the municipal land-use plan their zone were not designated as high-risk: 'The municipality couldn't get us out, because the POT in the city is in favour of us, [and so] it will be very difficult to evict anyone . . . Some councillors came here and they helped us a lot' (Leonardo, Manizales). As a result of this collective initiative, the community also improved its level of social organization in relation to communitybased waste management and clearing the local drainage and water systems -activities that help reduce flood risk. Indeed, following the research reported in this paper, a range of activities have been facilitated to try to continue a process of DRR empowerment in several of the study sites, through joint work by community activists, researchers, artists and disaster management organisations. Key in this work has been effort to strengthen the dialogue and cooperation between IDPs living in risk zones and government agencies.

Discussion
Tracing the life trajectories of people in the study communities in Colombia enables us to see how constraints on options for those displaced by violence brings them, often knowingly, to a situation of ongoing threat from hazards in their new environment. It also demonstrates how structures of marginalisation for these groups can become heightened still further in their new risk context, distancing people from external support. However, and demonstrated in part through the research process itself, we also see signs that suggest this form of exclusion can diminish. Drawing on the capacities inherent to self-settled communities, some groups are challenging the collective imagination of IDPs and disaster risk management institutions to envisage that they can work more effectively together to reduce the risks people face.
The life stories of the people we worked with were intertwined with different elements of risk, and the decisions people made to move, settle and resettle were shaped by a combination of structural violence that stripped them of assets and rights and conscious agency in how they responded to, and tried to circumvent, ongoing forms of exclusion. Importantly, we need to recognise that for most the context of risk began well before the violent events that triggered forced displacement (McIlwaine & Moser, 2003;Tovar-Restrepo & Irazábal, 2014;Celestina, 2016). Previous experiences of hazards, poverty and, particularly, gender-based violence influenced post-displacement decisions about which places to move to and how to establish a sense of greater security. Many spoke of repeated displacements, a 'restless' path, of being a 'wanderer', moving between multiple sites until they found some form of haven in the place they have now settled. Yet the haven is a relative one, and not just for the reasons for which we selected these hazardous sites. As in-migrants with few assets in a context of heightened social tensions, they were both aware of and subject to ongoing discrimination and economic, political and, in many cases, cultural marginalisation (Ibáñez & Moya, 2010;Cárdenas, 2018). Their struggle to build a life and home has continued, with most continuing to experience the instability of informal sector livelihoods and/or informal housing (Albuja & Ceballos, 2010;Ferreira, 2016).
Though deeply disturbing to hear, the struggles of vulnerability, forced displacement and re-establishment reported here are familiar -they echo the testimonies of people from many countries, already well described in the wider literature (e.g Eastmond, 2007;Brun, 2015;DeJesus, 2018). Instead, it is the much less well-documented connection between these struggles and the replication of risk through subsequent exposure to environmental hazards that we wish to focus on here, including the implications for people's engagement in disaster risk reduction.
As Table 1 indicates, all of our research locations had been sites of hazard events in the preceding years, and most study participants had witnessed the destruction of homes and property, injury to occupants, and consequent disruptions to livelihoods and wellbeing of those affected. First, then, we need to understand why the Colombian people we worked with had ended up resettling in places where they would be at heightened risk from hazards, often with knowledge of that fact. Certainly, there was an inherent reason why such sites were potentially 'available', especially in and around urban areas -across the world many informal settlements have been built on land considered unsuitable for planned development because of the threats from hazards on steep slopes, rivercourses and coastal lowlands (Doberstein & Stager, 2013;Abunyewah et al., 2018). Exposure to a wide range of environmental risks affects many low-income rural-urban migrants (Tacoli et al., 2015). In much of Colombia, hazardous terrain either confines formally urbanised areas or has necessitated investment in engineered structures to mitigate risk. Many available resettlement options were therefore in spaces at risk from landslides and flash floods. .
Yet this narrowing of choices is clearly only part of the reason. Forced displacement is a prolonged trauma (Shultz et al., 2014). We have to understand both the extraordinary difficulties people had experienced at their original homes and the 'restless' nature of their subsequent movements to see why such a place can be viewed as a suitable haven, despite it seeming to expose them once again to mortal peril. Time and again, research participants expressed to us the need to distance themselves from previous displacement threats, physically and emotionally, to find a 'safe' place in that context, and to be able, therefore to bring some stability to their lives by stopping moving and attempting to re-establish livelihoods out of a state of poverty. It was not that they were unaware of the dangers of building on precarious riverbanks, on slopes that required makeshift stilts under their dwellings, or in such close proximity that fires could spread easily between them. There was a conscious ranking of relative risks going on, that, however flawed in the eyes of city planners, made the existence of possible physical hazards a side-issue to be worked around rather than avoided. Self-construction of homes in these sites was a landmark in people's hope.
The sentiments expressed above echo those reported by Armijos and Few (2017) in research with populations exposed to volcanic hazards in Colombia. Despite the notoriety of an eruption in living memory that killed more than 25,000 people, the valleys and low-lying sites around the volcano Nevado del Ruiz that were devastated by lahars in 1985 have since seen re-occupation, principally by outsiders among whom many who moved to the newly vacant land did so because of conflict-induced displacement. Around 70% of survey respondents in that study confirmed that they knew the area they lived in could be affected by a volcanic eruption: the risk was therefore a trade-off in order to meet their livelihood needs and be somewhere at low risk from political violence (Armijos & Few, 2017).
The research reported in this paper shows that the selfconstruction of housing became a symbolic representation of recovery for many participants, and a demonstration of their dignity as residents. The resistance that people felt to the possibility of being evicted once again, this time on the grounds of protection from hazards, was therefore strongly evident. But it also contributed to an ongoing experience of marginalisation from governmental support (Siddiqi et al. 2019). These groups faced a paradox in that external recognition could have eased their problems of reestablishment, but there were also very strong reasons to remain as disengaged and 'invisible' as possible. One key constraint in Colombia is the persistence of fear and trauma in people who have been displaced by violence: this can impinge on everyday life and act as a barrier to participation in local civil and political processes (Riaño-Alcalá, 2008). Moreover, such tension has not dissipated following the 2016 Peace Accord because of the continuation and resurgence of violence (between 2017 and 2019 an additional 423,000 people were displaced in the country, according to IDMC, 2020). In our study not only did people moving into informal settlements fear that they might be identified and targeted again by armed groups, but they also wanted to keep a low profile to try to reduce the threat of eviction.
Over time, however, as they became more established, learned more about their rights, their environment and methods to resist eviction attempts, so they and their living spaces became more visible. Emerging from the shadows of attempted obscurity may have helped in the struggle to exercise their rights as 'victims' of conflict, but it also may have reinforced stereotypical narratives of the displaced populations (Cárdenas, 2018;Siddiqi et al., 2019). During meetings held to discuss the research plans and outputs for this study, some governmental personnel expressed views not only that such groups can be difficult to work with and 'hard to reach', but also manipulative in terms of seeking aid and misusing assistance. The communication gap between people and institutions directly affected how they accessed their rights as 'victims' of con-flict and the help they received to cope with the risks from hazards (as also reported by Fraser, 2016). Given their resistance to eviction from unlawfully occupied zones of hazard risk, the residents were commonly regarded as obstacles to disaster risk reduction: they should not be living in perilous places and if they continued to do so the authorities stated they could do little to protect them. This marginalisation from disaster protection is not unusual for the urban poor in many informal settlements (Castro et al., 2015;Tacoli et al., 2015), but, we contend, is a particularly acute issue for Colombians displaced by violence, given both their experience of trauma and their ongoing struggle on multiple fronts to rebuild stability in their lives.
Part of this impasse around rights and responsibilities relates to shortfalls in coordination structures between institutions managing different forms of crisis in Colombia, despite their evident overlaps in causes and consequences (Castro & Culma, 2018;Levy, 2019). UARIV, for example, is noted globally for its systematic registration of IDPs (IDMC, 2020), but the mechanisms of informationsharing between UARIV and UNGRD are presently weak. Owing to the sectorality of governmental mandates, it is ironic but therefore unsurprising to trace how those who have undergone forced displacement can come to be seen as obstacles to, rather than beneficiaries of, disaster risk reduction interventions (Siddiqi et al., 2019). Yet there does seem to be room for manoeuvre for both sides in challenging the apparent impasse posed by risk zoning regulations and/or the illegal status of informal settlements. As work elsewhere has underlined, given the complexity of the conflict-displacement-hazard situation that people in Colombia have faced, and still face, there may be a need for alternative, perhaps unorthodox, approaches to be applied to further DRR outcomes for such groups (Peters, 2019).
Within the communities, adversity has also forged capacities to take individual and collective action, even if in certain periods those capacities have been held latent. All the people we spoke with have taken the decision to leave their home, make their way to a distant place and restart their lives anew, often constructing and servicing their own dwelling. Amid the political sensitivities and generalised mistrust of authorities during the reestablishment process, many deliberately keep a low public profile. But as the process of creating a new living space progressed, materially and emotionally, people started truly 'inhabiting' their new homes, improving quality of life, and as they did so this progressively shaped individual, family and community goals -as similarly described by Osorio Pérez (2008) and Domínguez (2018). For many this included reinforcing the safety of the locale from threatsfrom landslides and flood hazards, from continuing external and internal threats of violence, and from the possibility of eviction.
For those in disaster management institutions that are keen to engage more productively with such communities, a key step in the dialogue is to recognise the strength of people's resettlement priorities and the attachment that many feel to their places they now regard as sites of safety and security, despite apparent risks from hazards (Opitz-Stapleton et al., 2019). Another key step is to shift perspectives to view people who have resettled after forced displacement as active self-determining agents, recognising that they have key capacities alongside their needs and demands (Siddiqi et al., 2019).
It seems there could be room for a justifiable flexibility. Different arms of government already operate in the space of informal settlements with contrasting agendas and competing imperatives, with the result that regulatory actions undertaken on the ground often end up as pragmatic, flexible outcomes of negotiation between institutional expectations and held at the interface between institutions and citizens (Fraser, 2017;Zeiderman, 2012). Designation of zones and attendant regulations is neither definitively 'accurate' nor immutable in practice. There are cases in the country where groups of settlers have succeeded with legal action to be allowed to remain in zones of high risk on humanitarian grounds (Fraser, 2017).
Recognising these points, should greater flexibility be enabled to review or update hazard zonation in cases where there is clearly reason to question the non-mitigable characterisation of land? This could free up the possibility to formalise titles and to work with communities on mitigation efforts to bring risk to acceptably lower levels. Some sites will indeed be too hazardous for this option to be reasonable, and that may well be recognised by the residents themselves. But even in those cases there is room for active risk management. A sustainable approach to planned relocation and the negotiations that entails inevitably requires time (Osorio Álvarez, 2017). Extending DRR to such communities in the intervening period therefore also implies strengthening disaster preparedness in situ through enhancing understanding of potential exposure to hazards, mechanisms of alert and emergency communication, and how to plan for evacuation.
Recent research on capacity development in DRR indicates that much remains to be achieved in bringing the most vulnerable people into both local decision-making and negotiation with authorities (Hagelsteen & Becker, 2013;Few et al., 2016). For reasons of physical access, socio-cultural differences, economic marginalisation and/or informal land titles, people who have undergone forced displacement are likely to be hard to reach under conventional scales of disaster risk management. Hence they are specific groups for which efforts to foster community-based participation and action are likely to be most valuable, if often challenging (IDMC, 2020). Ongoing work by the research team in Colombia has highlighted the value for both sides of approaching such interaction creatively. In this, the team has played a facilitative role in encouraging community members to come together to tell their stories as a group to a wider set of stakeholders, using the creative arts as a medium of expression. Some groups used drama and dance performances, others created murals, accompanied by narration and explanation. Fostering cultural engagement with communities through the media of music, artworks, storytelling, theatre and film has been one promising way to help build relations of trust, galvanise interest in participation and generate the mutual respect required for constructive approaches to DRR (Marsh et al., 2020).

Conclusion
By the end of 2019, there were more internally displaced people (IDPs) living around the world than ever recorded before: an estimated total of 50.8 million (IDMC, 2020). Ninety per cent of those that make up this figure had been forced from their former homes by conflict and violence, with the remainder displaced by disaster events. The issue of how displaced people can find secure resettlement is a global problem, and there will be strong parallels between the stories that have emerged from Colombia and those from many other countries.
In tracing how different groups of displaced people from four case study sites in Colombia have interacted with multiple forms of risk, this paper underlines why they may seek to create permanent homes in sites associated with hazards. Moving away from the sources of violence created by the context of armed conflict was the overwhelming focus for families and communities who experienced the forced displacement. A perception of risk dominated by the experience of violence meant that avoiding other types of risks associated with hazards was seldom factored into the decision on where to resettle. Although there are currently some actions undertaken by the communities to mitigate the risk associated with hazards, many people do not prioritise moving to another place as an option. Creating a home in the houses they inhabit has symbolic meaning reflecting both a personal struggle and a collective effort, building a stronger sense of belonging to the place where they now live.
Understanding how and why people 'move with risk' in this way, and its logicality even in the face of genuine threats from landslides, floods and fires, should make us question the oversimplification of risk avoidance approaches that automatically demand people undergo another episode of eviction. In some cases, yes, the gravity of the hazard threat may make occupation of a zone untenable. But, in others, where the threat is less clear, there would seem to be case for reconsideration. In the context of their extreme life experiences and their aspirations to create a living space of relative safety when options to live elsewhere remain structurally limited, how does society judge what hazard risk is acceptable? And how can disaster risk reduction be extended to support people to remain in areas that a narrowly-conceived hazard avoidance approach would not sanction? In Colombia, and in places elsewhere where people have faced similar risk trajectories, we advocate for a more thoughtful discussion around the 'moral hazards' of hazards-forced relocation.
We also see great potential for such communities to become more engaged in DRR in their places of resettlement. Though we show how an ongoing marginalisation of displaced people may become further entrenched in the process of resettlement, leaving them 'detached' from DRR activity, we also point to the capacities for agency that they evidently possess and the inherent understanding that they have about coping with risks. New forms of engagement are always emerging, and with creativity and commitment, bridges can be built from both sides to support people to remain safe in their places of refuge.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
ing transcripts. Finally, we are grateful for the enthusiastic partnership and support received from the National Unit for Disaster Risk Management (UNGRD) of the government of Colombia and the Colombian Red Cross. The project, 'Moving with Risk'/'Identifi cArte', was funded by the UK's Economic and Social Research Council (ES/P004326/1) and the Universidad de Manizales.