Costa: Flood Displacement During the 2017 Coastal El Niño in Peru

Bergmann, Jonas

doi:10.1007/978-3-658-42298-1_7

Jonas Bergmann⁶

Part of the book series: Studien zur Migrations- und Integrationspolitik ((SZMI))

1303 Accesses

Abstract

The research interest in Peru’s western arid coast (Costa) was in the well-being of persons displaced short distances away from their homes by floods during the 2017 Coastal El Niño (CEN) event for two reasons. First, Peru’s coast is periodically affected by severe El Niño-driven rainfall, which climate change will increase significantly in this century. Second, related floods are the main driver of displacement on Peru’s coast. The 2017 CEN floods, specifically, were the largest push for such displacement over the past decade in Peru, with close to 300,000 cases. Thus, examining the well-being of displaced persons from villages in the Piura Region after the 2017 CEN provides a useful temporal analog for future challenges. In the first section of this chapter, I provide information on the geographical context, measured and projected climate change, exposure, vulnerabilities, local coping and adaptation, and hazard-related migration by Peru’s coast. Then, I explain the new qualitative and quantitative results of this case study of displaced persons’ well-being following the 2017 CEN. Findings underline that the displacement occurred under high structural constraints, including severe climate risks and deficient DRR/DRM; poverty and inequality; limited livelihood options; tenure insecurity; poor and hardly accessible basic services; weak governance; and limited political participation. Conversely, structural opportunities were low. As a result, the severe and abrupt CEN floods caused forced, survival migration with detrimental conditions for moving and settling. Throughout the lifecycle of displacement, people suffered extreme losses, which continued to worsen their prospects for development from a secure base and a space to live better. Conversely, social relatedness remained similar after moving. Because displacement has become prolonged without substantial improvements, people’s need fulfillment, long-term asset base, and capacities for climate adaptation have worsened. Consequently, most displaced persons evaluated their need fulfillment as negative, and only few experienced partial positive feelings or cognitive satisfaction despite their plight. Expectations for the future were mostly negative and resulted in prevalent enforced fear as well as some fragile adjustment. Based on these findings in this case, I conclude by deriving more general propositions on the well-being impacts of climate migration.

You have full access to this open access chapter, Download chapter PDF

“El agua pasaba sobre mi cabeza, era altísimo, era como para ahogarme, sí por eso ya salimos de allá. No teníamos a donde irnos, salimos a la pista por mientras pero el agua se pasaba también… No teníamos pensado, ni en sueños lo teníamos, lo que íbamos a sufrir tanto … Porque en la casa ya teníamos todo, no nos faltaba nada …”

“The water was running over my head, it was so high, it was enough to drown me, yes that’s why we left from there. We had nowhere to go, we went out on the track for a while, but the water came there too … We had never imagined, not even in our dreams, that we would suffer that much… Since in the house we already had everything, we did not lack anything …” (own translation, as in all the chapter).

Statement by a mestizo mother of six children, who was displaced after the Coastal El Niño 2017 floods in Piura. She worked in housekeeping and was in her early 40s at the time of the interview (LP-11).

For two reasons, the research interest in Peru’s western arid coast (Costa) was in the well-being of persons displaced short distances away from their homes by floods during the 2017 Coastal El Niño (CEN) event. First, Peru’s coast is periodically affected by severe El Niño-driven rainfall (Sanabria et al. 2018), which climate change will increase significantly in this century (Cai et al. 2015; IPCC 2019a; Peng et al. 2019). Second, related floods are the main driver of displacement on Peru’s coast (Bayer et al. 2014; Ferradas 2015; French & Mechler 2017; Venkateswaran et al. 2017). The 2017 CEN floods, specifically, were the largest push for such displacement (or acute, forced migration) over the past decade in Peru, with close to 300,000 cases (IDMC 2019). Thus, examining the well-being of displaced persons from villages in the Piura Region after the 2017 CEN provides a useful temporal analog for future challenges. In the first section, I provide information on the geographical context, measured and projected climate change, exposure, vulnerabilities, local coping and adaptation, and hazard-related migration by Peru’s coast. Then, I explain the new qualitative and quantitative results of this case study of displaced persons’ well-being following the 2017 CEN.

1 Context

Although Peru’s narrow coastal strip covers only about 11% of its landmass, it is home to 58% of the population, a share that has significantly increased over the past decades (INEI 2018c).^{Footnote 1} The study sites herein are situated in the lower (LP) and upper part (UP) of the northern Piura Region (Figure 7.1).^{Footnote 2} Piura is Peru’s second largest Region and was home to 1.86 million people in 2017, 79% of whom lived in urban areas. The studied sites comprise small villages of origin in two districts in LP and one district in UP. The larger districts have between 19,000 and 83,000 inhabitants (INEI 2018c). All districts consist of one larger district town each and several scattered, smaller satellite villages. On average, the sites in LP are at elevations of below 50 m.a.s.l. and those in UP at 100 m.a.s.l.

All study sites are nested in the ecological floor coined Chala (Pulgar Vidal 1972), which is characterized by a subtropical desert climate, warm to hot temperatures, and minimal, highly seasonal rainfall. The Andes separate the subtropical desert climate in the western coastal area from the tropical, humid climate in eastern Amazonia (Mächtle 2016). As a result, less than 2% of Peru’s renewable water resources accrue to the Pacific basin (ANA 2018), where 58% of Peru’s inhabitants live (INEI 2018c). About a third of Peru are drylands and desertification is a major, ongoing threat (CEPES 2015; INRENA 2006, 2011; MINAM 2016a). Most of the coastal rivers are short and seasonal. Irrigated crop farming is only possible on a narrow band west to the mountains; further downhill, shrublands extend, and the coastal plain has almost no natural vegetation cover. In certain desert areas, an intensive groundwater-irrigated and export-oriented agroindustry exists, which often overexploits local water resources (Damonte 2019; Damonte & Boelens 2019). UP has greater average rainfalls and lower temperatures than LP (Figure 7.2) because it is at a higher altitude and farther away from the coast.^{Footnote 3}

Farmers in Piura have to cope continually with “too wet and too dry conditions” (Sperling et al. 2008: 25). During recurrent El Niño events, they are threatened by abrupt events such as storms, heavy rainfall, pluvial or fluvial flooding, and huaycos^{Footnote 4} (flash floods). In addition, farmers suffer from gradual hazards such as warming, dry spells, and droughts.

Climate change has raised the frequency and intensity of these slow-onset and sudden-onset hazards. On the coast, temperatures have increased between 0.15 and 0.25 °C per decade between 1981 and 2016 (Aybar et al. 2020). An analysis of gridded data, based on satellite and station data from the PISCO dataset (Lavado-Casimiro et al. 2016), offers details for the study sites in LP. It shows higher average temperatures through most of the year in the period 1997–2016 compared to 1982–2001. Annual mean temperatures have not changed significantly, but the daily temperature range has increased. Extreme heat (95^th percentile of maximum temperature) and the number of hot days are significantly greater than before. Other studies confirm that the number of cold nights has declined while the number of tropical nights and hot days have risen (Donat et al. 2013; Skansi, María de los Milagros et al. 2013). Extremely warm days have increased at least by two in northern South America over recent decades in austral summers (Ceccherini et al. 2016; Feron et al. 2019). The analysis of PISCO data illustrates the temperature trends in LP. It shows increases in average temperatures, in maximum daily temperature range, and temperature extremes (Figure 7.3).

Average rainfall trends on the coast have not changed significantly (Haylock et al. 2006; Lavado-Casimiro et al. 2012a), but intensities may have risen in its south (Heidinger et al. 2018). Minimum annual runoff has increased (Lavado-Casimiro et al. 2012a), likely due to first stages of glacier loss in the highlands that increase meltwater until the peak flow is reached (Rau et al. 2019). The analysis of PISCO data in LP reveals that the yearly average rainfall and rainfall extremes (95th percentile) have remained similarly low, apart from El Niño years (Figure 7.4). Figure 6.1 in chapter 6 illustrates that many of Peru’s flood-exposed zones are on the coast, including in Piura (MINAGRI 2012: 38).

The El Niño Southern Oscillation (ENSO) is key for climate variability in Peru and worldwide (McPhaden et al. 2006). Every one to five years, the sea surface temperature (SST) in the Eastern (EP) or Central Pacific (CP) warms (El Niño) before it cools eventually (La Niña) (Sanabria et al. 2018). Strong EP El Niños increase rainfall by Peru’s northern coast but reduce it in other parts, whereas strong La Niña episodes have largely opposite effects (Lavado-Casimiro & Espinoza 2014). Conversely, strong CP El Niños decrease rainfall in upstream regions along Peru’s western coast (Rau et al. 2017). The latest Pacific El Niño events that severely affected Peru occurred in 1982–1983 and 1997–1998, while the strong events in 1972–1973 and 2015–16 had weaker impacts (Sanabria et al. 2018). ENSO has significantly fluctuated in the past (Cobb et al. 2013), but its variance has increased over past decades (McGregor et al. 2013). Coastal El Niño (CEN) events are rarer than CP or EP El Niños, and previous episodes occurred in Peru in 1891, 1925, and 2017 (Hu et al. 2019; Peng et al. 2019; Takahashi & Martínez 2019). The unexpected 2017 CEN—the focus here—was probably driven by anomalously warm SSTs along South America’s west coast (Hu et al. 2019; Rodríguez-Morata et al. 2019). Like strong EP El Niños, CENs can cause torrential rainfalls off Peru’s northern coast. In March 2017, the CEN generated rainfall amounts similar to the Pacific El Niño events in 1982–1983 and 1997–1998 (Rodríguez-Morata et al. 2019). Climate change made the 2017 CEN at least 1.5 times more likely to occur compared to preindustrial times (Christidis et al. 2019).

In the coming decades, climate change could affect Peru’s coast in various ways.^{Footnote 5} In a low emission scenario, average temperatures may increase between 0.75 to 1.5 °C by 2050 and between 1 to 1.75 °C by 2100 compared to 1985–2005; in a high emissions scenario, they could rise by 1 to 2 °C and 3.5 to 6 °C (Bergmann et al. 2021a).^{Footnote 6} In another high emission projection,^{Footnote 7} the 50^th percentile of mean temperature would rise from 24 °C in the 2010s to 27 °C in the 2080s (Figure 7.5).

By 2050, in a medium emissions scenario, extremely warm days and heat waves would increase 5–10 times per season. In a high emissions scenario, what is an extremely hot summer day in Lima today would become 11 times more frequent compared to 1961–1990 (Feron et al. 2019). By 2100, extremely hot summer months will be much more common, especially on the coast (Adams et al. 2014). 3-sigma heat events,^{Footnote 8} still rare today, will be the norm in Peru roughly half of the summer months in a 2 °C warmer world by 2100 and for most summer months in case of 4 °C warming. 5-sigma events, which currently do not exist in Peru, will occur in 20% (70%) of summer months for 2 °C (4 °C) global warming. In the LP area, high emissions would lead to over 350 hot days per year (Figure 7.5). Globally, the risk of heat exceeding body thresholds of temperature and humidity will increase by between 48% in a low and 74% in a high emission scenario by 2100 (Mora et al. 2017a).

This study focuses on floods, but rainfall projections come with uncertainties. Peru may see fewer rainy days but more intense rainfalls (Christensen et al. 2013; Giorgi et al. 2014). On its arid northern coast, average rainfall may increase (Sörensson et al. 2010; Vera et al. 2006). Recent rainfall models confirm this trend (Figure 7.6). A study of a central coastal basin finds that discharge is seasonal and decreases in the dry-season but rises in the wet-season for the 2050s and 2080s (Olsson et al. 2017).

Climate change will also make El Niño events, the focus of this study, more frequent, even if temperatures should stabilize over the long term (Wang et al. 2017). The IPCC (2019a) has medium confidence that extreme events will occur twice as often under both low and high emission pathways in the 21^st as compared to the 20^th century. Other projections show that strong EP El Niños could rise from six events in the 20^th to nine in the 21^st century while CP El Niños and extreme La Niñas will also occur more frequently (Cai et al. 2015; Cai et al. 2018). More El Niño events can but do not inevitably create more extreme rainfall over Peru (Sanabria et al. 2018). Extreme CEN events may also become more frequent, but the range across models is large (Peng et al. 2019).

More frequent, extreme El Niño events will also interact with permanent sea-level rise (SLR) and thus present new threats to Peru’s coastline population (Reguero et al. 2015). Globally, for 2100 compared to 1986–2005, a low (high) emission scenario could result in mean SLR between 0.29–0.59 m (0.61–1.10 m) (IPCC 2019a). For Peru specifically, the highest-emissions scenario could result in SLR of at least 0.7 m by 2100 (Church et al. 2013). After 2100, locked-in SLR will continue for long (Strauss et al. 2015) and in the highest emissions scenario, global mean SLR would reach 15 cm per decade (IPCC 2019a).^{Footnote 9} At this pace, Peru would experience 1 m SLR over the next 100 years.^{Footnote 10} Adaptation may mitigate some damages in Peru (Anthoff et al. 2006; Nicholls 2011). Permanent, local SLR impacts are less severe than in the most affected Latin American areas but large relative to the coastal area of Peru (Dasgupta et al. 2009; Gosling et al. 2011; MINAM 2010; Pearson 2009; Teves et al. 1996; USAID 2011). Moreover, coastal flooding events will add to permanent SLR inundations and may threaten tens of thousands of inhabitants in Lima alone (Reguero et al. 2015). Synergetic effects of (permanent) SLR and more frequent (episodic) strong El Niño events could also worsen periodic coastal flooding in Peru, as ENSO can create extra SLR for several months off Peru’s coast (Reguero et al. 2015). Finally, error-corrected data reveals that exposure to SLR and coastal floods may be much greater than assumed earlier (Kulp & Strauss 2019).

These hazards affect people in Peru in multiple ways. Exposure to El Niño events is high, with over 7 million people across the country (Figure 7.7, not counting droughts) (SINAGERD et al. 2014). They live mostly in the north, led by the Piura Region with 1.7 million people, where exposure is high due to hydro-geographical features, such as steep river gradients (French & Mechler 2017). During the 2017 CEN, rainfall deviated most strongly from the average of the previous decades in Peru’s central and northern coast (where Piura is sited) and in the central rainforest (Figure 7.8). These rainfall anomalies led to extensive floods (Dartmouth Flood Observatory 2017; Son et al. 2020). In addition, settlements in flood-zones are also growing in Peru, often involving corruption or the state’s acquiescence (Bayer et al. 2014; French & Mechler 2017). Poor dwellers in slums are most exposed, among them internal migrants who often settle on affordable but high-risk lands (Venkateswaran et al. 2017), and who tend to be less prepared for or experienced with ENSO impacts (Bayer et al. 2014).

Not only exposure but also vulnerabilities are high. People in Piura’s rural areas are mostly traditional, smallholder crop and livestock farmers, who have to irrigate their crops year-round, or day laborers (Aragón et al. 2018; Bayer et al. 2014; Oft 2009). In a large-scale survey, about 27% of the farmers on Peru’s desert strip were poor, and only 58% of the household heads had completed primary education (Aragón et al. 2018). Such exclusion from basic services is often gendered (Oft 2009, 2010). The lack of non-farming income sources and high poverty makes farmers highly vulnerable to production shocks (Aragón et al. 2021). In typical households, few members gain incomes, and their livelihoods are centered on only one ecosystem-based activity, which raises climate vulnerability (Oft 2009, 2010). Most rural farmers in Piura have almost no savings or convertible assets and can hardly cover their monthly expenses. In some exposed areas, vulnerabilities are due to the lack of land titles (Bayer et al. 2014). A lack of oversight and frail materials leads to inadequate housing especially for poor people in Piura, which increases vulnerability to hazards (French & Mechler 2017).

As a result, El Niño events can have devastating effects. Table 7.1 indicates that estimated damages for the 2017 CEN, the focus of this research, amount to USD 3.1 billion or roughly 1.6% of Peru’s GDP (French et al. 2020; Macroconsult 2017). The previous 1982–1983 and 1997–1998 events caused losses of USD 3.28 and 3.5 billion, equivalent to 12% and 6% of the annual GDPs, respectively (Sanabria et al. 2018; Vargas 2009). All these events severely damaged infrastructure, basic services, agriculture, fisheries, and livestock in Peru (Badjeck 2008; Sperling et al. 2008).

Table 7.1 Loss and damage in Peru during recent El Niño events

Full size table

Previous studies find that the adaptive capacity of subsistence farmers in Peru’s desert plains^{Footnote 12} for floods is low and depends on access to assets such as seeds, fertilizer, land and livestock (Sperling et al. 2008). For example, households that succeed in preserving livestock during El Niño events are better able to adapt and profit from their positive side effects, such as new vegetation cover. For coping with floods, the largest share of respondents in surveys across Piura asks for help from family and friends; many others either can do nothing, reduce expenses, or pursue extra work (Oft 2009, 2010). These coping strategies cover about three quarters of the incurred losses. When asked what respondents would do differently in a hypothetical future flood event, 23% would not change their farming strategies, 20% would diversify crops, and 15% would not grow anything at all. Similarly, a study in Tumbes indicates that residents find ways to live with water shortages and loss of assets and homes after El Niño events, but recovery requires extensive time (Bayer et al. 2014). For water scarcity, studies indicate a similar range of reaction strategies as for floods.^{Footnote 13} For all hazards, migration also figures among people’s coping and adaptation strategies, as is discussed next.

First, this study focuses on abrupt El Niño events, floods, and intense rainfalls, which are the largest drivers of internal displacement in Peru (IDMC 2021b). Such hazards usually force more temporary than permanent movements (Bayer et al. 2014; Espinoza-Neyra et al. 2017; French et al. 2020; Venkateswaran et al. 2017). For example, following the 1997–1998 event, 5% to 10% of people in one study left for good while a higher percentage looked for temporary work (Bayer et al. 2014). Oft’s (2009, 2010) survey finds that 7% of households in Piura’s lowlands migrate during floods, most often temporarily for day labor. When asked how they would adapt to floods over the longer term, only 1% pointed to temporary migration. For many, even temporary migration due to floods tends to be an “option of last resort” to smooth income losses through jobs in coastal cities or the Amazonian lowlands (Sperling et al. 2008: 40). Flows can be gendered: while men migrate to work on rural farms, women move to cities for domestic work (Sperling et al. 2008). Multiple hazards related to El Niño events, compounded by development challenges, have also resulted in several attempted and completed community relocations in the coastal zone (Ferradas 2015; French et al. 2020; Oft 2009, 2010; Sperling et al. 2008; Venkateswaran et al. 2017). The focus of this analysis—the 2017 CEN event—forced both temporary and permanent migration. It was the single heaviest push for displacement in the past decade in Peru, with close to 300,000 displacements (IDMC 2019). Floods, mudslides, and flash floods destroyed 63,800 houses and damaged more than 350,000 dwellings; dented thousands of schools and health posts; caused close to 140 deaths; and affected a total of roughly 1.5 million people (French et al. 2020).^{Footnote 14} In May 2017, a survey registered 13,155 displaced persons in camps in Cura Moria and in Catacaos in LP (IOM 2017b). 87% of the sites emerged spontaneously, and most were close to the villages of origin, along streets to the city of Piura (Venkateswaran et al. 2017). In September 2017, a survey in 25 camps in LP found that the displaced persons were on average at the end of their 20s, and there was almost gender parity (IOM 2017c). In a non-representative survey in April 2018, 17% male and 18% female children lived among the remaining displaced persons in 16 sites in LP. 56% of all the respondents reported prior disaster displacement experiences (IOM 2018).

Second, while not the focus of this study, slow-onset processes also contribute to migration on Peru’s coast. For example, water scarcity in Piura can damage health, livelihoods, and educational opportunities and drive migration as a result (Sperling et al. 2008). In one survey in Piura’s lowlands, 8% of households affected by water scarcity used migration as a coping strategy, and 2% would consider it for longer-term adaptation (Oft 2009, 2010). Migration after droughts can be more permanent than after floods (Sperling et al. 2008). Temperature changes can also damage agricultural livelihoods (Aragón et al. 2021; Oft 2009, 2010), but few studies examine possible links to migration.

For future movements by Peru’s coast, the climate projections debated above could have several implications. Many hazards that influence the drivers of (im)mobilities will intensify. To begin with, projected high temperatures during much of the year and periodic heat waves—combined with rising water scarcity—will likely have strong, rising impacts on farmers’ livelihoods and health, and thus affect migration and entrapment. Simultaneously, more intense rainfalls could destroy productive assets and homes and thereby compel displacement in the absence of effective disaster risk reduction and management (DRR/DRM). Especially the impacts of more frequent El Niño events could coerce more flight. More frequent, episodic strong El Niño events will also have synergetic effects with permanent SLR and could result in displacement and relocation if adaptation inaction, population growth in exposed areas, and weak governance combine (Reguero et al. 2015; Wrathall et al. 2019).

Against this background, the next section empirically evaluates how the 2017 CEN floods displaced people from the study sites and how this displacement has affected their well-being.

2 Qualitative Empirical Results

In November 2018, I interviewed 24 persons (9 m / 15f) displaced by the 2017 CEN in camp sites in LP (Figure 7.9) and in flood-affected villages in UP. Respondents’ average (median) age was 44 years (43 years), with a range of 22 to 62 years (Table 7.2). Most respondents were subsistence farmers. Few of them had supplementary but limited livestock, and various did not possess own fields but rented them or worked as day laborers for agribusinesses and other landowners. A small number of interviewees worked as civil servants, in construction, or transportation. Interviewed women usually oversaw complementary activities such as free-range animals and some vending of farmed products. All interviewees spoke Spanish and auto-identified as mestizo.

Table 7.2 Basic data on interviewees in areas affected by the 2017 CEN in LP and UP

Full size table

To gather background evidence, I also interviewed staff at the Regional government’s units for Social Development, Urban Planning, and Natural Resources, and the Authority for Reconstruction with Changes (RCC) for the 2017 CEN. Furthermore, an advisor to the mayor of Piura, two local mayors, one village alderman, and three community leaders in camps for displaced persons shared insights.

2.1 Climate Change Dimensions

Most respondents who were forced to flee after the 2017 CEN stressed that flooding and flash floods had affected them. In addition, many perceived climate-related diseases and pests that affected human beings and animals, heat, as well as strong winds and declining crop productivity (Figure 7.10). Less than one fifth of the respondents also mentioned droughts, water scarcity, and rainfall changes.

Villagers in LP and UP remembered several major El Niño events. Moreover, families living close to a drain by the Piura River in LP stated that they were flooded mildly every winter. While the 1983 and the 1998 El Niños had also displaced persons, the 2017 CEN floods were the most severe in people’s memory. In UP, respondents suffered pluvial and fluvial floods in mid-March 2017. The floods developed rapidly, blocked or destroyed access streets, isolated households, and forced many to seek shelter on hills and in communal buildings, as discussed below. Impacts were most severe for those in low-lying areas. Unmaintained drains and the nightly flood-onset exacerbated damages:

When we were sleeping, no, like at two in the morning, no, it started to rain hard ... The water started to rise, and we only saved what is necessary, clothes, it caught us off guard. When my son tells me ‘Mom, the house is already full, we can no longer do anything,’… the only thing [I could do] was to watch how I lost mattresses, clothes, shoes, everything was ruined. (UP-6)

In LP, the Piura River broke levees, overflew rapidly, and severely damaged villages twice in late March. People perceived the magnitude of the floods as unprecedented and felt overwhelmed:

The water came twice, it came on a Monday and the other came on Wednesday, and it came back even more… Thank God it was daytime, if it had been nighttime, how many would have died, creatures, old people. Quickly, the water rose and came from all sides. (LP-14)

Several respondents did not believe in an early warning they had received and refused to leave, because previous El Niño floods, although severe, had never reached the same intensity. Most villagers close to the broken levees lost their houses, fields, and other assets, and were displaced:

One could not save anything because the water rose very quickly, our concern was to get ourselves out, so we left with our clothes on, nothing else, all things stayed there, everything was lost …. (LP-5)

Beyond floods, combined gradual climatic and non-climatic stressors have affected the farmers, decreased their coping capacities, and rendered them more susceptible to periodical abrupt El Niño-hazards. Interviewees in LP and UP witnessed more heat and more or new plagues, which threatened agriculture and human health and forced some migration. For example, one farmer observed:

Ah yes, it no longer produces as before, the production is decreasing now ... Because of the climate, the sun that burns a lot ... It was not like this before then, in previous times, those that the elderly tell you about; well, they say that the plants produced like this, you just sowed and harvested. (LP-3)

Immediate coping with the 2017 CEN event was difficult for the interviewees owing to the sheer magnitude of the losses. Because they had few assets convertible to money or savings, many had to depend on the uneven humanitarian assistance. For longer-term adaptation, numerous people in LP indicated they wanted to stay in the areas they had fled to, which were protected from floods (see section 7.2.2). In UP, farmers observed a lack of knowledge, skills, and finances for DRR/DRM, and attested that, “People worry … They don’t know what to do, what they can do” (UP-10). Some attempted to build basic physical protection but lacked the required means:

Yes, a wall was made, but it does not provide guarantees, ... it is made of sand. If it [the rainfall] is slow, it does nothing, but if it comes with force, it tears the wall down as if there was nothing. (UP-5)

Soil deterioration is an additional stressor. Farmers needed to borrow money to buy fertilizers and pesticides, and to sow again after harvest losses, which might result in vicious debt circles if later harvests are also unsatisfactory. Water scarcity is also a perennial issue, especially after El Niños, and migrants said that population growth and more water-intensive crops had intensified the issue over the past decade. While some interviewees stated that they struggled to cope with water scarcity, the focus of this analysis is on the 2017 CEN, which were the main driver of displacement.

2.2 Migration Dimensions

Next, applying the framework of section 2.2, I assess migration drivers, aspirations, capabilities, and common paths. Interviewees had to flee fast as the CEN floods threatened homes and lives:

The water was running over my head, it was so high it was enough to drown me, yes that’s why we left from there … Yes, the first time [it flooded] we stayed in the church, in the schools, but because of the downpour, the roofs were destroyed and the water was entering. The second time we already had to leave, in that humidity, in that mud that was left, it was very ugly, we were crying, the thunder was scary, we were asking God to have pity with us. (LP-11)

… [I]t started to rain hard… Then over there [by my brother’s house] they started to scream, to cry, we all went there ... the house was coming down on them… ‘No, right now there is no way we can stay here, we better get out ourselves’, … it caught us off guard. (UP-6)

Before the floods, most interviewees had had low or no aspirations to leave. One indication of these aspirations to stay is that many ignored early warnings and preferred to remain on site despite the imminent flood threat. Yet, once the water flooded their houses, they had little choice but to flee for survival and protection. Migration aspirations were thus instrumental (de Haas 2021) and the movements were forced, acute displacement. For example, a displaced woman in her early 40s said:

We had never thought about leaving before the flood. We had never imagined, not even in our dreams, that we would suffer that much… since in the house we already had everything, we did not lack anything, we had a chair to sit on, a table where the children could write, we had light, we had water. (LP-11)

Most displacement started in a similar way because the abrupt floods created a largely homogeneous, acute need to flee, which affected people simultaneously. Migration capabilities influenced people’s options in LP tangentially. Their movements did not require specific social capital, such as relatives offering shelter, since migrants fled to a nearby, uninhabited land (although conflicted land, see 5.2.2: Basic services).^{Footnote 15} Those with access to boats could escape marginally easier and faster and were able to save some belongings; others without transportation means were trapped briefly before the army evacuated them. Timing differed only slightly. Many farmers fled early on during the first flood, as the water rose rapidly to the top of their houses; others persevered for hours in the flooded area, some out of choice, others against their will. Finally, a small number waited out until the second flood hit. For most, displacement occurred over several spatial steps: they escaped first to their rooftops, nearby hills, or higher buildings; then to the close-by highway; and eventually to the nearby land where they built camps or settled in designated evacuation areas. Displaced persons were forced to stay away for long as their villages of origin remained flooded and the houses seriously damaged. When I conducted the interviews in November 2018, they had lived in prolonged displacement for close to 20 months.

In UP, in mid-March 2017, intense rainfalls damaged or destroyed several low-lying houses and forced the owners to flee over short distances. However, because the impacts were less severe in UP than in LP, the interviewees returned earlier. Displacement was also more disperse in UP than in LP, where most people ended in camps. For example, two farmers in UP had been displaced to the high school for one month but returned to their land after they had received temporary housing modules. Another female farmer and her family were displaced to relatives first and then to the high school for one month. She initially had to leave her father behind because he was unable to walk and “everything was water over there”, so they “raised his bed with bricks” until they could “get him out later” to the high school (UP-6). This case stresses the role of health status and physical ability for migration capabilities. Others stayed with relatives or in tents for several months. For example, one female farmer in her mid–20s lived with her man and their two years old child in a tent for more than a year.

Beyond short-distance and short-duration displacement, a few young farmers from UP migrated to the sea or to cities for some months to smooth income losses. For example, a man in his mid–30s worked in coastal fisheries for several months to support his family after he had lost his house. Some interviewees observed scant permanent movement, but others stated that, “Many people who lost their houses got displaced to Lima, especially if they did not receive assistance initially” (UP-10). The longer and farther away people wanted to move, the more vital were capabilities. For example, a displaced crop farmer in his early 60s recalled he could not move as far as he would have liked since “there are no resources to evacuate to another place” (UP-5). He remained in forced immobility.

Many of the respondents had been displaced multiple times in earlier El Niño events. For example, a woman in her late 30s had fled to a relative’s house in 1998 for one month, then was displaced from her house again in 2002 for one night, and finally had to leave again in 2017 (LP-12).

Among those displaced in the 2017 CEN, a few had come to UP from the highlands years ago, partially due to climate impacts in their villages of origin. These farmers can be considered “double climate migrants”, and their plight stresses the salience of accumulating shocks. For example, a crop farmer in his early 60s, originally from the rural highlands, indicated that he like many others had migrated with his whole family to UP ten years ago due to worsening agricultural production. They had built a decent life, but then lost close to everything in the 2017 CEN floods (UP7).

The camps in LP were built relatively close to the flooded home villages, which enabled varied translocal ties, such as going back to rescue belongings early on and clean premises later; sending children to unaffected schools back home (see section 7.2.3: Educational opportunities); or visiting the ruined homes for nostalgic reasons. Only few displaced persons returned completely because they wanted to be closer to their previous schools or had succeeded in restoring parts of their fields. Several farmers had dual residencies or moved between the locations for their work but lived in the camps.

During the interviews in November 2018, most displaced persons from LP had already spent 20 months in the barren desert camps and still wanted to stay. They acquiescently abandoned their old homes and accepted longer distances to fields. A central reason to stay was fear of new floods, especially among children and older adults. This fear of more inundations was built up through cumulative flood experience and boosted by radio forecasts or word of mouth:

Most have stayed here ... out of fear ... due to the danger of the river, it was not the first time… I have experienced this since I was a child. (LP-7)

These fears often combined with the experienced losses, which made a return less viable:

When [the kids] see that rain falls, they are already scared, they run inside the house, due to the rain they get really scared. They don’t want to go back anymore… why return, so that the children get scared? The houses are already destroyed, they collapsed, how are we going to stand up again? (LP-11)

Finally, besides El Niño events, the interviews confirm that droughts can drive seasonal migration. For example, the male head of one household in UP would migrate close to every year to the coast for fishing in dry seasons, when the family suffered economically.

2.3 Well-Being Dimensions

In the following, I describe the results of the qualitative text analysis of the primary data collected from interviews with displaced persons. I examine how the displacement from areas in LP and UP harmed by CEN floods have affected people’s well-being and apply the four well-being axes explained in section 2.3: development from a secure base, a space to live better, and social relatedness (objective well-being, OWB), as well as subjective well-being (SWB).

2.3.1 Development from a Secure Base

2.3.1.1 Decent Livelihoods

When I interviewed the displaced persons in LP in November 2018, 20 months after the floods, they were still struggling to make a decent living in the nearby camps.^{Footnote 16} The floods had destroyed harvests and assets, fields were not yet fully recovered, and owners of parcels “can no longer give work” (LP-11). Many men and some women switched to work as day laborers for a nearby agribusiness that had remained unaffected. The agribusiness owned the conflicted land where their camps were erected, and reportedly, the displaced could work for it as the result of negotiations on the land (see 5.2.2: Basic services). This day labor provided a key income buffer after the flood shock. Yet adults older than 50 years remained excluded, stressing how age can influence disaster recovery. The offered work was also precarious and unstable, as one of the community leaders detailed. Another leader criticized:

Let’s say that the need for work is such that you endure the work, ... the company takes advantage … [Y]our daily salary is 33 soles [Peruvian currency] and with a break of 15 minutes, not more, they exploit people. This is one of the realities that people [live], the difference of your own farm and of providing a service… Staff who do not perform, they remove after four days ... Although our people suffer at work, at least there is a source to make their family livelihood ... Imagine if these companies did not exist ... But I still say there is a mistreatment of the workers. (LP-6)

Despite these limited options for day labor, various farmers remained un- or underemployed and had entered economic hardship. Even those who could continue their pre-flood jobs (such as some moto drivers) suffered losses, albeit to a smaller extent. Regardless of the specific income source, many respondents were only earning “just enough for food, more than anything to survive” (LP-10). When asked on his income, a middle-aged farmer who was forced to switch to day labor replied:

Enough for the sustenance of the house, but not more [nervous laughter], it depends on the children you have, I have four… [Before] there was something to sell, now as the situation is, one cannot work on the farm as the water has left sands in the fields ... Yes, one works only for the sustenance of the house, for the daily expenses, it is not enough to save and buy other things [lowered voice]. (LP-3)

Before the floods, women usually contributed to household incomes by raising free-range animals. However, most could not rescue their animals. One woman in her mid–50s explained how her income had decreased significantly since they had lost all their animals and lacked jobs or savings to reinvest:

Oh yes, sometimes it is not enough and now everything is expensive. ... Sometimes it is sad because there is no work ... I am already [mid 50s], they don’t give me work anymore, that is painful to me… [I]t’s the first thing we want, that there be something to work, to buy some small things, more than anything food. (LP-4)

Some women were starting to invest in animals again with the help of NGOs and local farmers, aiming to start producing and selling food. If households combined their income activities, they could subsist but not thrive. For example, one woman in her mid–30s with six children, and highly pregnant when interviewed, had lost all animals, and was only slowly reinvesting with support from a charity. She could not produce chicha [alcoholic drink] anymore as she lacked wood and water, and her man was precariously employed in day labor, rendering their income just enough “to survive” (LP-9).

Displaced persons in UP suffered similar losses of assets, animals, and crops, due to rainfalls, floods, and subsequent pests. Those lucky enough to rescue some animals usually had to sell them after the floods to smooth the income losses. Farmers had to wait long until they could sow crop seeds and grow pastures for livestock again. Water scarcity added as an obstacle to their recovery. Most people had been poor before the floods, but afterwards, under- or unemployment aggravated their situation. The labor market provided even less possibilities to buffer losses than in LP, because UP had less agribusiness demand for day labor. Those farmers who were able to switch to day labor worked precariously and had worse income than before. A female farmer in her early 40s deplored:

What you earn is only to survive, you cannot progress, because they pay you little, that is why when the phenomenon [El Niño] came, it hit us too strong, and left us poorer. One works to survive, to support the children, they cry, they ask for food, milk. That’s the experience we have here. (UP-6)

She and her family could only take the most necessary things and lost most assets. They had nothing to eat initially, as their crops were destroyed, pastures for livestock could not grow, and new crops required time until harvesting. They received limited help from the government, pointed to corruption, and felt abandoned by the state. Many interviewees confirmed this lack of state presence and noted that NGOs and villagers’ mutual support mechanisms mitigated these gaps to some extent.

In summary, the 2017 CEN eroded livelihoods. It also increased the climate vulnerability of displaced farmers because they lacked seeds, food, and income from harvests, or savings, and no DRR/DRM were taken to protect livelihoods. A mayor rued:

Eh, well if it [a future flood] comes it would be worse, but we must prepare to prevent… Well, right now with this government there is no support, there is nothing, the municipality is worse, we are close to the end of the year and there is no support. (UP-1)

2.3.1.2 Health and Food Security

Shortly after the CEN floods, the displaced persons had large needs regarding health and nutrition. In November 2018, the interviewees confirmed that diseases remained widespread. The vulnerabilities of older adults, children, people with physical disabilities, pregnant women, and people with preexisting diseases were particularly exacerbated by poverty and the unhealthy living conditions during the displacement. In UP, insufficient water quantity and quality was a widespread issue because the displaced persons lived close to the flooded areas, sanitation had collapsed, and people lacked funds to buy water. As a result, infectious diseases spread fast, especially among children.

With their livelihoods destroyed and no savings, most respondents also suffered from hunger and depended on humanitarian assistance or mutual support. In UP, the floods had blocked access streets and exacerbated food insecurity. A farmer remarked: “There was no passing, we were isolated then. For [nearby cities] there was no access, for example in terms of food, we suffered a lot here because of food” (UP-9). Dependency on external assistance was high and when I interviewed the displaced persons 20 months after the floods, many still could not produce or purchase enough food for adequate nourishment. A highly pregnant woman stressed the difficulties for her and her six children:

Sometimes, here we suffer because there is none [food]; we endure hunger, but children do not. Sometimes, there are days when you eat, there are days when you don’t, and sometimes we help each other with a plate of food because the spouses don’t have a job. (LP-9)

People who lacked housing or land titles struggled to access social programs, and the National School Feeding Program (Qali Warma) seemed not easily portable. A community leader described widespread food insecurity, which affected children in particular:

Yes, it affects the alimentation of the children, the reason is that there is no work ... Yes, because we do not know how to get [food] ... We asked for a comedor [popular dining room] that is suitable for children, but we were not lucky. (LP-5)

Camp sector leaders also stressed the psychosocial impacts of the floods and displacement (see 5.2.2: Emotional balance and cognitive satisfaction). Several women suffered trauma due to the experienced threats to live, losses, and flight. In few areas, NGOs provided basic care, but most respondents rued that they received no emotional or psychological support. Others had to make substantial efforts to receive help. As one mother explained, “I had to take my 7-year-old daughter there to [town] to be seen by a psychologist because she was scared and did not want to stay …” (LP-11).

Further health challenges related to decreased protection possibilities from adverse weather conditions, for example, because the displaced persons remained in inadequate shelter:

We have been in a tent for eight to nine months, there we already burned, we were very burned ... When there was sun, we could not enter the tents, it burned too much…. (LP-4)

And now children with this climate—the children get sick with the flu, the cough. At night [it’s] quite cold and here, we sleep on the floor, and it is cold. (LP-9)

Health service provision has stayed fragmented after the CEN floods. In November 2018, the interviewees felt that the health situation had slightly improved compared to the start of their displacement and the serious infrastructure losses then, but was still worse than before. Several camp sectors had only a health post in a container with intermittent service. Other areas stated that services were gradually withdrawn and that “there is no health center here, the nurses are leaving because they are hired, and we are left with nothing” (LP-11). Several respondents had to travel to unaffected health posts in surrounding villages. Health emergencies constituted severe challenges because they required costly travel to the district city for help and transportation options were low at night:

In the afternoon, in the evening, there is no one attending you. There have already been abortions here, there have been people who have come out of an emergency, children with respiratory infections late at night,… the routes are not adequate, and we do not have the mobility to be able to help people. (LP-6)

Finally, the lack of income or insurance to pay for health services exacerbated health problems:

Yes, they [the agribusinesses] don’t pay you as it should be…. You work your working day, you don’t have insurance, you don’t have anything, you fall ill, and you die since you don’t have money. (UP-6)

2.3.1.3 Educational Opportunities

When interviewed in November 2018, respondents rued the substantial losses of school infrastructure and the lack of services for the large number of displaced children. They observed that schools in the camps had been built fast and for temporary use (linked to the land conflicts, see 5.2.2: Basic services), but remained the only available option for a long time. According to one community leader, infrastructure was only slowly improving and promised support such as school feeding programs did not arrive. All respondents confirmed that educational opportunities for their children had severely deteriorated after the CEN. Various children had to walk strenuous, long distances to the few unaffected high schools in their villages of origin: “they have to go up hills and go down”, remaining “in the sun, the children burn, they get blisters, they get tired” (LP-9). The exhaustion, adding to mental health issues, made it difficult for children to concentrate in class. Parents felt that, “Sending children to schools is scary” (LP-11) as it involved a passage over the nearby highway and the risk of (at times deadly) accidents (LP-6). Yet, sending them by moto (autorickshaw) on a regular basis was too expensive for most parents. Finally, the lack of light in homes also hindered homework.

2.3.2 A Space to Live Better

2.3.2.1 Adequate Housing

The quality of housing heavily deteriorated for most displaced persons. The interviewees confirmed that many arrived with nothing but their clothes to the desolate, empty desert area. They received tents that provided basic shelter initially, in which they spent one to twelve months. 20 months after the floods, several of the interviewees still had only basic additional structures built around their tents. Only few respondents received temporary housing modules later. While they perceived these modules as an improvement over tents, the modules were of poorer quality than the houses they had had before the CEN. Inadequate housing conditions exacerbated other vulnerabilities, especially of pregnant women, older adults, children, and people with physical limitations:

Another of the great needs here is, the type of life as we have, no, the houses for example as you see are of rustic material, we have the elderly, we have children, we have the disabled [sic], and housing is not so adequate because of the cold here, the wind that makes people fall with respiratory infections. There have also been diarrheas, well the situation is critical, no ... The conditions are, let’s say, precarious, no, rustic material, kitchens of this type and well…. (LP-6)

For example, a highly pregnant woman in her mid–30s explained how the shelter failed to protect her family from the cold and how she was afraid of the health impacts on her baby. Cooking occurred mostly with collected wood, which created health risks through indoor pollution and fire risks.

All interviewees described the shelter reconstruction as arduous and “expensive” (LP-3). “Little by little” (LP-4), they made small improvements, although the materials remained “rustic” (LP-3):

The material is expensive ... we bought it suffering hunger ... [It is] expensive to make such a house, here we have invested to make this room, not more, over there I still need to do one, but I do not have money to buy the materials. Well, that is the situation [sighs]. (LP-2)

Compared to LP, displacement was more dispersed in UP and resulted in different shelter situations. The displaced persons lived in tents, public buildings, and unaffected houses of relatives. For example, two crop farmers were forced to move to the high school and stayed there for one month before returning to their land, where an NGO provided temporary housing modules. Overall, the farmers described that housing had significantly deteriorated compared to before the floods, and that damages were long lasting. Although many respondents had already suffered damages in previous El Niño events, they ignored or could not adhere to warnings by authorities not to rebuild in low-lying land and to avoid poor materials, as they lacked financial means and safe plots for protected homes.

2.3.2.2 Basic Services

Most of the interviewees had had almost complete access to basic services in their villages of origin before the 2017 CEN. They struggled with managing the ensuing lack of provision in the barren desert camps. Land conflicts posed a major obstacle to receiving investments in basic services. Interviewees had allegedly received the land to which they escaped after the 1998 El Niño; yet “because of bad leaders” (LP-6), it had been sold to two agribusinesses, which created land conflicts after the displacements in 2017. After the CEN, this land conflict stalled government investments:

… [T]he state itself did not even want to set foot here due to the very situation of the private [land]… … The minister came also, he told us that he could not invest because they were not in the capacity to invest in a property that did not belong to the state. (LP-6)

Conflicts (see 5.2.2: Security) and talks led to partial solutions for persons in some sectors, who could stay and work for the company. For these sectors, municipalities agreed to a land-use change from rural to urban to legalize settlements. Notwithstanding, people doubted that the agreement was final:

First his [the landowner’s] family came to throw them away, the fences we had; they fought, they set their dogs on us [laughs]. Then the young lady from the municipality came to resolve, they came here and had an agreement, but now the gentleman says no, and does not want to give the terrain. So, we only have certificates. (LP-12)

In other camp sectors, the company did not compromise and ended the discussions. For these displaced persons, the land issues resulted in high uncertainty. They were worried because “many told us that they were going to take the land from us” (LP-8) and longed for a solution:

What we most want is that they give us a proof of possession, we want it to be a true one, that they do not take away from us. Because here there are problems sometimes, if the businessman throws us out, where are we going to go back to? We want that they give us a document that protects us…, for the well-being of the children… The other time, they also said that they were going to evict us, and the children are scared when they hear that, they start crying. We sleep alert, too. (LP-11)

These obstacles and the fragmented humanitarian assistance resulted in an uneven availability of basic services in the camps. Various respondents voiced their frustrations with the government and often alleged corruption. Crowded living conditions added further pressure on the insufficient services.

Although water provision was a crucial need for survival and recovery, many camps remained without sufficient, safe, and reliant water for long. In November 2018, the interviewees remarked that water quantity, quality, and costs diverged across sectors. In some zones, the agribusiness owning the land provided water in tanks, in limited amounts but of sufficient quality. Other areas regularly received free water to group tanks from the state. For still others, private trucks provided water to group tanks in sufficient quantity but at high costs. In addition, water deliveries were becoming less frequent, so that some sectors now experienced scarcity. Finally, various sectors had to rely on water of insufficient quantity and quality from wells in nearby villages.

Regarding sanitation, the difficult situation with few latrines shared by many households had improved at first but then declined again. By November 2018, the interviewees rued a widespread lack of sanitation and drainage. In several areas, the temporary constructions built by the state or NGOs were decaying, after more than one and a half years of use by numerous displaced persons, or had already collapsed and were dysfunctional:

… [I]n the very school, for example their toilets, the toilets were also built thanks to the NGOs…. but these services were not, in other words, this is temporary, and because we have been around for a while, we will have spent two years here, and these hygienic services have already collapsed. (LP-6)

In contrast to their villages of origin, in November 2018, most interviewees described a persistent lack of light and electricity in the camps. Only few had access to solar lanterns or panels outside of their modules for limited hours of light. Several interviewees recalled that they had been promised access to electricity and light but expressed doubt that the promises would be fulfilled. The darkness complicated children’s schoolwork, resulted in security risks, and lead to fire risks due to the use of candles. For example, a father of four children in his late 50s pleaded that,

Hopefully the new Regional authorities ... we hope they help us and remember us, because here we are without water, without light, which is the most essential thing for children who have to do their homework. (LP-8)

Poor roads created additional obstacles for the displaced persons. Transportation providers (motos) could not enter the sandy camps as most “roads are pure sand, are not even planned”, which increased respondents’ perceived isolation and made them “suffer” (LP-6). They got “blisters” and were exhausted from walking on the sands (LP-4), especially children and older adults:

Here the only thing we want is access roads, the ladies from [nearby town] have to carry their bundle from here to there, no motorcycle enters, it is what one suffers the most without access roads. (LP-3)

2.3.2.3 Pleasant Surroundings

By the time of the interviews, the displaced persons in UP had returned to their homes, and many of them rued the destruction of their fields and missed the work there. Yet overall, they described limited changes regarding the satisfaction with their locality, natural surroundings, and related activities.

By contrast, the displaced persons in LP arrived in barren desert lands that most perceived as hostile compared to their prior home villages: “In these small terrains we have all been; pure thorns, herbs were these lands; all together, we had to help clean, men and women” (LP-1). According to a community leader, the living conditions were crowded, with more than 2000 families living in an area apt for 500, which raised stress on the already stretched infrastructure (see 5.2.2: Basic services). Additionally, the significantly smaller sizes of the new lots complicated the traditional raising of free-range animals and the storing of produce. Positive feelings about the surroundings were rare.

2.3.2.4 Safety from Hazards

In UP, a large number of displaced persons moved back to their old lots and thus into the same flood exposure as before. Consequently, many feared new floods and felt less prepared for possible impacts than before (see also 7.2.1; 5.2.2: Safety from hazards, Outlook on the future). In November 2018, the respondents in LP stated they felt safer from floods as the camps were located on sandy ground that could absorb downpours. They were slightly more elevated and farther away from the canal and drain. Especially traumatized persons reported relief to be on safer grounds. For example, a displaced woman in her mid–30s, when asked if she felt safer or not in the new destination, replied:

Yes. Although it was difficult for us to tarry here, we were crying, it was sad to be here… Remembering is sad, we didn’t eat, we only took care of the children, but little by little we got used to here, and to return there, no? Not anymore. (LP-9)

While safer from floods, the displaced persons in LP faced exposure to new hazards in the barren desert, which was exacerbated by poor housing (see 5.2.3: Adequate housing). The exposure to adverse weather conditions harmed especially the health of at-risk groups, such as children and older adults (see 5.2.2: Health and food security). By day, the desert sun caused sun burns and the heat in the tents was unbearable while overnight, they were exposed to winds and cold temperatures.

2.3.2.5 Security

In November 2018, interviewees reported insecurity, including violence due to the land conflict (see section 7.2.3: Basic services), thefts, and robberies. First, after fleeing, the land conflict with the private landowner created tensions. Violent clashes allegedly killed at least one displaced person. By the time of the interviews, some sectors had found a compromise on the land, but others continued in uncertainty (see 7.2.3: Basic services). The second insecurity issue concerned thefts and robberies, which were widespread in UP during the flood and at the start of displacement. In LP, robberies and thefts of water tanks or animals were still common. Farmers said, “When we call them [the police], they come whenever they feel like it, and the thieves have already fled” (LP-9). Finally, some sectors self-organized citizen patrols (rondas) that contributed to interviewees’ perceived security:

We make the ronda [citizen patrols] at night ... since a month ago there were robberies in a row ... as the animals are stolen and that’s because they see that there is no security, with this ronda things are improving ... Several things have been taken ... Right now, we are, as one says, relaxed, we feel that there is no theft anymore. (LP-8)

However, other sectors without rondas or police forces perceived perennial insecurity. The lack of light (see 7.2.3: Basic services) exacerbated the security risks: “More than anything, we want them to improve our town so that we too can live more peacefully, because the darkness of the night is dangerous” (LP-7). Women in the interviews did not explicitly mention gender-based violence. However, when male household members migrated to other areas for work, insecurity for women and children could rise. For example, the young mother of one child described how her husband’s migration to the coast to smooth income losses raised insecurity:

Yes, here with the children I stay ... It is not easy, and it is also dangerous since it has happened to us once that they scared us, that hooded men came here, they scared us. So sometimes I’m also afraid of being alone. (UP-3)

2.3.3 Social Relatedness

The interviewees asserted that they had formed functional, cohesive communities prior to the floods. In UP, most people fled within their villages and felt that “we are united” (UP-9). They maintained good relations and observed solidarity: “[W]e all give a hand to each other. For those affected, we tried to help, to give them a way out so they could progress, and not leave [them] alone” (UP-9).

In LP, the CEN displaced entire villages. The displaced explained that social cohesion and unity remained high because neighborhoods and social structures were replicated in the camps: “Yes, we know each other here, we live in one community, and we support each other” (LP-9). A community leader said, “Here we treat each other as one family because we stand together and we are always there, in dialogue” (LP-5). The displaced persons observed good social relations, with mutual support systems buffering the worst damages. Sharing tools or assets and mutual help in reconstruction and communal tasks was common. The displaced persons in LP also swiftly self-organized to elect community leaders who acted as local camp sector coordinators.^{Footnote 17} One of these leaders stated, “Yes, we hold meetings almost weekly, biweekly, in order to organize ourselves, to coordinate” (LP-5).

The interviewees did not mention changes in family relations due to the flood and displacements. They only described adverse social effects in cases when peers migrated to other areas for work. For example, a young mother and her children missed her husband who had migrated to the coast to smooth losses, and recalled how his absence raised security issues (see section 7.2.3: Security). Her husband, returned by the time of the interview, confirmed the difficulties during the family separation:

… [T]he situation is ugly because you leave your family in the situation that was here… Yes, well, it would be nice if there was work in only one site, if we both worked together, with the children. (UP-4)

2.3.4 SWB Dimensions

2.3.4.1 Emotional Balance and Cognitive Satisfaction

Disaster experiences and displacement can have a long-lasting impact on SWB. 20 months after the CEN disaster, the interviewees still experienced strong negative emotions due to the floods and the displacement. For several religious displaced persons, who considered the flood a punishment by god for previous misbehavior, the disaster evoked feelings of guilt and regret (see section 7.2.3: Outlook on the future). Yet, most prevalent were feelings of helplessness, sadness, and pain. Respondents linked these negative emotions to losses, increased poverty, poor shelter, threats to health, and stress:

… [F]or the one who no longer has a place to live, until now it is something traumatic. My sister-in-law always says, ‘I would like just one room [to live in]’, but nothing. I’m begging to God, and I really do want some help to come because it hurts. (UP-6)

These negative feelings were salient especially at the start of displacement when deprivations were greatest. However, for many, these emotions persisted despite slow advances as they worried about the arduous recovery. For example, a woman described how her family was sad about the losses, the new exposure to weather extremes, the costly and troublesome reconstruction, and the hunger (LP-2). Family separation due to longer-distance migration also created sadness (see section 7.2.2).

Many displaced persons stressed perennial psychosocial issues, including anxiety, fear, shock, and trauma. For example, one woman described how she had feared to drown in the floods, lost all her assets, arrived horrified in the camps, and still suffered 20 months later: “We were already scared, and the other year left us traumatized” (LP-11). Many intensely feared that more floods might occur:

We don’t [want to go back] anymore, because the radio stations start saying that the water is going to come again, we are already traumatized by that, we all cried when it happened. (LP-4)

Children, in particular, were traumatized after the floods and afraid of new disasters. Mothers suffered to witness the continued anxiety of their children. One of them observed that,

… [The children] no longer want to return, they are afraid. When they left, they were scared, traumatized by what happened because well, that was never seen in life before. (LP-5)

The interviewees criticized that professional psychological care was insufficient in most areas (see section 7.2.3: Health and food security). Still, a small number of displaced persons slowly lightened up because they had started rebuilding lives and felt safer from future floods. Various respondents showed notable resilience, which was at times supported by humor. For example, a respondent in UP was saddened by the loss of his house and assets. Although he worried about his wife’s health and was afraid of a new flood, he was also humorous, laughed during the interview, and made jokes (UP-5). Yet, in other cases, laughing was a mechanism of defense rather than an adaptive humorous take:

Interviewer: “We are investigating the social impacts of the phenomenon of El Niño. How did you pass that experience?” LP-4: “We don’t want to remember [laughs]”. LP-13 [adds]: “It was a sadness.”

Various children processed their experiences by playing games. One mother reported how she, saddened by her losses, felt pain when she watched her three children playfully simulate the floods. However, the children seemed to enjoy the game (LP-4). Another mother said her son had liked the time when they were forced to live in the school building, although he had suffered from a serious stomach infection shortly before, and the family lacked money to receive adequate health care.

My son used to tell me, ‘Mom, how cool, living in a school’. I have never had a house with a second floor, but my son would go up to the second floor in the school, and he liked to live in the school, he was happy, he did not want to leave from there. (UP-7)

While various interviewees expressed gratefulness toward humanitarian assistance by NGOs, they were angry, annoyed, disappointed, and frustrated about the government’s handling of the CEN and the recovery. Many felt angry or deceived due to the empty promises for improving their situation (see also e.g. section 7.2.3: Decent livelihoods). Various observed corruption. For example, a woman who had lost her house in 1998 for the first time, and then again in 2017, was angry about the “neglect of authorities” to maintain drains, because otherwise the flood “could have been avoided” (UP-6). She was also frustrated and disappointed by the flawed government responses after the flood:

To [name of an NGO] we are grateful, they helped us a lot… The state is absent… The state is taking a long time, many obstacles; what happens is that the state has a lot of corruption… Yes, the governments that we have had have been bad, in a row, bad, bad. (UP-6)

This perceived corruption also evoked feelings of injustice and shame among the displaced persons.

2.3.4.2 Outlook on the Future

When prompted to assess their views of the future, various people stressed religious faith as their anchor for consolation, hope, and patience, even in their dire circumstances and with evident trauma:

There are families that are still in need, but they also have to wait, yes God makes that it continues, otherwise they will have to wait just with patience, and someday God will touch their hearts again…. I have consolation, hope and I have faith in God that someday we are going to recover some things, not everything, because the situation does not allow anymore to buy everything we had. (LP-5)

Even so, various religious persons perceived the CEN as a divine punishment, which instilled guilt, regret, and surrender in some of them. Others hoped that virtuous behavior could improve their plight:

But since one believes in God, these are tests that little by little they are making for us, sometimes we say, ‘Why does it happen, why does it happen?’, but it happens because sometimes we also no longer obey God and God gets tired of correcting us. We do bad things and then God, when he punishes, he sends for everyone, good and bad, and thank God we are surviving here. Yes, but when you are with God everything comes to your house ... [It is] the will of God and you have to have patience. (UP-6)

Beyond faith, the solidarity in communities, the decreased exposure to floods, and trust in individual capabilities also made respondents in LP hopeful that they would be able to progress and improve the future of their children (see section 7.2.3: Safety from hazards). People hoped that progress would be possible if the land conflict was solved (see 5.2.2: Basic services), if the state invested more, and if NGO support continued. Many followed the Peruvian narrative that progressing (salir adelante) was always possible. This idea seemed to be a sign of resilience in several cases; in others, where possibilities for growth were extremely low, this narrative of progress created extra pressure or led to self-deception. For example, when asked about his view of the future, a community leader contended:

Ah well, we aim to salir adelante [to progress], keep fighting, start living a new life. In other words, start again from the bottom since certainly all the things that we had, well, we no longer have. Now, to start a new life, start buying things, start building a new home—and sometimes one does not have the resources, and the materials are expensive and cannot be bought, but anyway, one is getting up little by little ... Also, hope, clearly, [we] value everything we have, and we will recover little by little. (LP-5)

Conversely, several displaced persons expressed resignation, discouragement, and fatigue due to economic and land insecurity (see section 7.2.3: Basic services). Two respondents feared that, “Maybe we will never replace what we have lost [sad]” (LP-2) and, “Well, I think everything will remain the same [nervous laughter]” (UP-2). In UP particularly, farmers stressed that their precarious agricultural livelihoods limited their economic prospects, and to make things worse, El Niño floods, alternating dry spells, and warming have repeatedly disrupted their lives and decreased hopes for the future. The CEN losses made farmers feel “clearly less, less prepared” for future hazards, which could undo their slow recovery: “God forbid too [laughs], pucha [slur], it would be worse [if a flood happened again] … Well, we no longer have somewhere to sow and how to survive” (UP-9).

3 Quantitative Empirical Results

In this section, I complement the qualitative study of Piura with a quantitative analysis based on secondary data that cover all districts across Peru which were declared in a state of emergency due to the CEN. These data yield additional insights into people’s differential displacement risk and the impacts of displacement on their well-being one month and seven months after the event (whereas the qualitative data was collected more than one and a half years after the disaster).

As detailed before in section 3.2.2, the analyses are based on two surveys. These two surveys are (a) the CEN Survey collected around one month after the disaster, between mid- and end of April 2017, and (b) the National Census enumerated six months later, at the end of October 2017. The CEN Survey contains data on 398,148 persons in 199,938 dwellings and 2,615 public buildings in the 892 districts declared in a state of emergency (see Table 3.1). Both analyses below focus on a subsample of 186,437 affected adults, namely those whose homes suffered at least minor damages in the disaster.^{Footnote 18} 88% of this group (164,084 adults) could be tracked in the National Census for the well-being analysis, which compares well-being effects between the affected, disaster displaced adults and those respondents who were affected but could remain at home as the most suitable comparison group.

3.1 Magnitude of Displacement and Differential Displacement Risk

To begin with, the CEN Survey provides insights into the magnitude of the overall displacement and households’ differential displacement risk. However, the data does not explicitly identify displaced households.^{Footnote 19} In this dissertation, I focus on the habitability of respondents’ homes to infer if they were displaced due to the disaster.^{Footnote 20} In the CEN Survey, enumerators observed and indicated whether the visited houses were (a) unaffected; (b) habitable but in need of renovations; (c) uninhabitable; or (d) collapsed.^{Footnote 21} Because the comparison here focuses on the affected population, category (a) was excluded. The analysis operates with a new binary variable which uses category (b) for the status of affected but non-displaced and merges categories (c) and (d) as the proxy for disaster displaced. Applying these proxies, 24.1% of all affected households (or 42,410 adults) surveyed after the CEN were displaced: the homes of 15.53% of the households (or 27,583 adults) were uninhabitable while those of 8.57% (or 14,827 adults) were entirely collapsed. In some districts where rainfall was intense, more than 20% of the households were displaced; Figure 7.11 illustrates the extreme rainfall anomalies during the CEN in the 20 districts with the highest share of displaced households.

In a next step, three logistic regression models were applied to assess who among the affected 86,734 households had been most at risk of displacement (Table 7.3). The first model uses exogenous environmental factors as proxies for people’s exposure to the disaster, namely data on rainfall anomalies between January and March 2017 as well on the average distances to inland water bodies, maximum elevations, and elevation ranges in the analyzed districts.^{Footnote 22} The second model adds household composition and demographic characteristics to these environmental factors, including household size, the share of children, older adults, or people with disabilities in the households, if the household was headed by a single parent, and the level of education. The third model expands these prior parameters further with variables on livelihood characteristics and the wealth of the households, such as occupation, tenure security, as well as housing and infrastructure quality.

Regarding performance, model 3 has the highest Pseudo-R² and therefore the highest explanatory power. In this best-fit model, most parameters had a statistically significant influence on displacement risk. All results in the models are marginal effects calculated at the mean of all covariates. The risk increased greatly due to environmental factors, such as the intensity of the CEN rainfall anomaly (0.2507***) and the elevation range in the districts (0.2344**). If households were farmers (0.3178***), lived in small rural villages (0.2036**),^{Footnote 23} had young or older children (0.1965*** or 0.0769***), were headed by single parents (0.1691***), or had members with disabilities (0.1589***) also critically raised displacement risk. By contrast, dwelling in decent types of housing, such as houses or flats (–0.3122***),^{Footnote 24} and owning property with a title (–0.2373***) strongly reduced the risk. Smaller decreases in displacement risk followed for households with larger sizes (–0.1303***), who rented or lived with family (–0.1281**), and at greater distance to inland water bodies (–0.0721***). For every 10-percentage point increase of adults with at least secondary education or female members in the household, the risk of being displaced also decreased by 1.6% (**) and 1.3% (***), respectively.^{Footnote 25} Parameters without statistical significance were limited access to public infrastructure prior to the CEN, living in non-standard housing (such as hotels or asylums), the maximum elevation in the district, the number of older adults in the household, and if all adult household members were unemployed prior to the disaster.

Table 7.3 Logistic regression models analyzing marginal effects of drivers of household displacement risk

Full size table

3.2 Short- and Mid-Term Impacts of Displacement on Well-Being

The next analysis focuses on the impacts of displacement on respondents’ well-being directly after the CEN and seven months later. The two surveys provide insights into a selection of the central well-being dimensions applied in this dissertation (section 2.3 explains the full framework), mainly into a space to live better and development from a secure base. Yet, the data are limited on various key subitems of these dimensions, such as education, safety from hazards, and physical security. Additionally, the surveys lack data on social relatedness or SWB. The well-being analyses below are thus not as comprehensive as the qualitative study in section 7.2.3 but they illuminate key changes in the well-being of a much larger number of people. For context, note that the quantitative data cover people’s situation directly after the event in April and half a year later in October 2017, whereas the qualitative data sheds light on the situation in November 2018, more than a year later.

To begin with, shortly after the disaster, people displaced by the CEN reported substantially higher damage on livelihoods, health, shelter, and basic services than those residents who were affected by the disaster but able to remain in their houses within the districts declared in a state of emergency (Table 7.4). Because the sample covers close to the full affected population, summary statistics render robust results on people’s well-being outcomes. Regarding development from a secure base, the survey registered severe damages to livelihoods: the agricultural production of 92% of the displaced compared to 85% of the non-displaced, but affected households was damaged; 25% of the displaced and 17% of the non-displaced households lost work due to the disaster; while 90% of the businesses in homes of displaced households and 72% of those in the non-displaced group were damaged. Additionally, the disaster strongly affected people’s health. In Piura—the focus of the qualitative study—109 persons died, out of a total of 280 fatalities countrywide. In all of Peru, 6% of the displaced and 3% of the affected non-displaced persons were injured in the event. As many as 43% of the displaced and 41% of the non-displaced households suffered from health problems one month after the CEN, such as respiratory infections, diarrhea, dengue, and tuberculosis. Finally, food insecurity was rampant in the districts declared in a state of emergency. Approximately 80% of the displaced households compared to 69% of the non-displaced respondents suffered from moderate or severe food insecurity in the month after the event. In addition, the CEN Survey registered 81 damaged health centers. As the only indication of changes in educational opportunities in the data, the survey also recorded damages to 518 educational facilities. Other sources document an even higher damage to 934 health posts and 2,870 schools (French et al. 2020: 5). The effects on a space to live better were similarly severe: around 85% of the displaced persons noted that the disaster had severely damaged the walls or roofs of their houses, compared to less than 10% for the non-displaced group. Among the displaced households, 78% had no access to sanitation, 71% lacked access to water, and more than half of them lived without electricity, which represent substantially higher shares than for the affected, non-displaced group. The data is visualized in Figure 7.12 below.

Table 7.4 Summary statistics on loss and damage experienced by households due to the CEN event in 2017

Full size table

In a next step, linear regression analyses were used to discern the impacts of displacement on the individual (not household) well-being of respondents seven months after the disaster. The analysis draws on a dataset with the affected adults of the CEN survey tracked in the National Census. A well-being index was constructed based on 35 items available in the Census. Most items existed for a space to live better, including the type of housing;^{Footnote 27} construction materials; access to water, sanitation, and electrical light; property title; number of persons per rooms; fuel used for cooking; and the availability of household assets (16 assets, such as computer, cell phone, refrigerator, and car). For development from a secure base, the Census provides three items of interest, namely indictors on people’s employment status, their health insurance, and disabilities, whereas no relevant data existed for education. All individual indicators were combined in a well-being index using principal component analysis (PCA). In PCA, items are assigned weights based on their relative importance in the sample. Well-being items that allow to differentiate more effectively between different groups are given a higher weight. The resulting index was normalized to a scale from 0 to 100 to simplify the interpretation (Table 7.5, Table 7.6).

Table 7.5 Summary statistics of variables used for constructing the well-being index

Full size table

Table 7.6 Summary statistics of the principal component analysis

Full size table

Figure 7.13 illustrates the distribution of the respondents along this normalized well-being scale. It indicates that the respondents’ well-being is skewed toward the lower end of the scale.

In a next step, five models linearly regressed peoples’ well-being score on displacement. As a baseline, the first model uses displacement as the only variable, while the subsequent models gradually add more controls. The second model includes the exogenous environmental factors from the prior analysis as controls, such as the intensity of the CEN rainfall anomaly. The third model controls for these environmental parameters as well as for household composition and demographics. In addition, the fourth model also encompasses livelihoods and wealth variables. Lastly, the fifth model adds individual characteristics to the parameters already included in the previous models. By adding further environmental, household, and individual variables to the models, the analysis attempts to control for self-selection effects that might confound the estimation of displacement impacts on well-being. Displacement is not a random process, but factors, such as the age, sex, or socioeconomic status of respondents can change the probability of displacement in systematic ways (Aksoy & Poutvaara 2021; Borjas et al. 1992; Kaestner & Malamud 2014). Because displacement is driven by a range of such factors that might simultaneously influence the well-being outcome, the observed changes in well-being in the dataset might not be due to the displacement itself, but rather due to pre-movement factors that made displacement more likely in the first place. Controlling for these factors in the models renders a more accurate representation of displacement effects, which is reflected in changes in the regression coefficients as the numbers of controls increase from model 1 to 5.

The results in Table 7.7 demonstrate that displacement led to statistically strongly significant well-being decreases in every model. The negative effect of displacement on well-being remains consistent and statistically significant across all models, while the strength of the effect decreases from model 1 to 5 as these models successively account for more confounding variables. Concerning model performance, model 5 has the lowest AIC score, which indicates the highest explanatory power. Based on the estimated Adjusted R², this best-fit model explains about 36% of the variance in the well-being outcome (whereas displacement alone in model 1 explains 1.5% of the variance). In the best-fit model 5, displacement caused statistically significant losses of 3.14 points on the well-being scale that ranges from 0 to 100 (–3.1389***, compared to –6.2189*** in model 1).

In addition, most control parameters have statistically significant effects on well-being. Strong influences in the best-fit model 5 are secondary education (14,9207***), housing type (11,9633*** and 4,9532***), living in small rural villages (–10,4514***), and owning property with a title (5,9972***). Migration between districts in the five years prior to the Census also substantially raises well-being (5,3214***). Examples of the various other control parameters with strong influence include being an agricultural household (–4,4098***) or having limited access to infrastructure (–4,2752***) and unemployment (–4,2015***) prior to the CEN. Two controls without statistically significance are unemployment prior to the disaster and having household members with disabilities.

Table 7.7 Linear regression models analyzing drivers of individual well-being impacts seven months after the CEN

Full size table

4 Discussion

In this chapter, I interpret the empirical results of the Costa case study and situate them in the broader literature. The focus is on the key mechanisms of action and structural conditions that have influenced the observed hazard-(im)mobility dynamics and resultant well-being effects.

4.1 Hazard-(Im)mobility Links and Pathways

The abrupt floods during the 2017 CEN acutely forced farmers from LP and UP who lacked previous migration aspirations to flee. The qualitative data from Piura suggest that because the floods constituted a largely uniform, rapid, and critical push to move, migration capabilities were tangential at first. Yet, the quantitative data offer a more nuanced view of differential displacement risk across Peru. Clearly, higher exposure to the disaster strongly increased displacement risk. Both pluvial and fluvial flooding raised the displacement risk, since rainfall intensities and the proximity to inland water bodies (thus the possible excess of water) as well as the elevation range in the districts (thus slopes for possible floods and mudslides) had significant effects. Other studies confirm that hydro-geographical features have contributed to people’s exposure to the CEN in Peru (French & Mechler 2017). Therefore, spatial disaster risk analysis is imperative to understand and address related displacement risk. Peru has done first steps in the direction of such spatial analyses (SINAGERD et al. 2014), but the data requires updates and should also account for the associated displacement risk. However, the regressions also reveal that not all the exposed people faced the same risk of being displaced; rather, this risk was greatly influenced by their social vulnerability to flood impacts.

First, the threat of being displaced due to CEN impacts was high for people living in poorly protected shelter. The Peruvian state recognizes that deficient housing is a major cause of disaster vulnerability in the country (SINAGERD et al. 2014). Four out of five houses in Peru are self-constructed, with limited oversight or control of building codes (Calderón et al. 2015).^{Footnote 30} As a result, precarious housing with poor materials is widespread. In cities, over one third of dwellers lived in slums in 2014 (World Bank 2019), and informal settlements have grown threefold between 1993 and 2012 (Calderón et al. 2015). The expansion of informal settlements continues and is accompanied by plot trading and challenges such as clientelist state behavior and unequal access to opportunities. Although the vulnerability of informal settlements in Peru is known, it remains unattended (French & Mechler 2017). This analysis stresses the need to improve housing quality in disaster-exposed areas to reduce displacement risk. Conversely, the regression results emphasize that owning property with a title strongly reduced displacement risk, as did renting or living with family, albeit to a lesser extent. A possible interpretation is that homeowners with a title have more income which allows them to protect their houses better from hazards. Furthermore, they might be more likely to live in legal settlements in more affluent neighborhoods with better infrastructure, DRR/DRM, governance, and access to social programs.^{Footnote 31} In the same way, people who rent or live with family may dwell in more resilient conditions than those without a title. Consequently, addressing the widespread tenure insecurity in Peru (Calderón et al. 2015; Prindex 2022) may help reduce disaster displacement risk.

Moreover, farmers were at higher risk of displacement than those in other economic sectors. One explanation could be that many smallholders and subsistence farmers have undiversified income sources and depend on ecosystem services and livelihood assets close to their homes that disasters tend to disrupt for long. Numerous poor farmers also lack buffering capacities to cope with shocks. In Peru, this lack of access to financial mechanisms for self-protection is a major cause of disaster vulnerability (SINAGERD et al. 2014) and, by extension, of displacement risk. Farmers also tend to depend on water for irrigation and might therefore have lived close to water sources that are prone to spill over during heavy rainfalls and damage people’s homes as a result. The substantial displacement risk of farmers should concern policymakers because around 30% of the Peruvian labor force works in agriculture (World Bank 2019), a share that is considerably higher in rural areas. 97% of these agricultural jobs were informal in 2014 and lacked proper social security (CEPLAN 2016), which can raise susceptibility to disasters and displacement. Related, the regressions indicate that living in small rural villages increased displacement risk. Most farmers in Peru live in the countryside. Peru’s rural areas are generally poorer and more deeply deprived, and they have worse infrastructure and DRR/DRM systems in place than its cities. An estimated 46% of rural residents were below the national poverty line in 2014 compared to 15% of the urban population, and the poverty gap for the rural poor was on average 14% below this line, but only 3% for the urban poor (World Bank 2019). As a result, the risk of displacement is considerably higher in rural than in urban areas. Even so, urban displacement risk can also be high, especially in marginalized neighborhoods (IDMC 2021c).

Finally, various intersectional social factors—such as being a household headed by a single parent, having limited education, having household members with disabilities, or having young children—made displacement more likely. Other studies have also demonstrated that intersectional factors strongly influence the climate adaptation options available to people in Peru (Erwin et al. 2021). These findings can be interpreted in two ways. First, due to systematic marginalization and resulting disadvantages in Peru (Sanborn 2012), these social factors may deprive people of basic resources required for protection against disasters, such as resilient shelter. Second, all the variables above, and particularly lower education levels, probably also correlate with having less income, buffering capacities, and resources for fleeing. Displacement is generally more associated with fast- than with slow-onset hazards (Koubi et al. 2016; Zickgraf 2021), but when hazards hit fast and hard, vulnerable households may be unable to evacuate out of harm’s way (Boas et al. 2020).^{Footnote 32} Finally, the analysis indicates that households with more male members were more likely to be displaced. Data corroborates that the probability of flight is usually higher for men than for women in other contexts as well (Aksoy & Poutvaara 2019). Some studies suggest that households with more women may be better prepared for disasters, but the literature is inconclusive on this question (Alam & Rahman 2017; Ashraf & Azad 2015; Bronfman et al. 2019; Castañeda et al. 2020; Sharma et al. 2015). An alternative explanation, but outside of this framework, is that attitudes toward risk-taking behavior, such as flight, might have differed during the disaster depending on people’s sex. Another possible explanation could be that social norms or sex-specific discrimination in learning, for example concerning swimming, could have trapped more women than men in damaged buildings (Rigaud et al. 2018). A different possibility is that caregiving norms may have forced women to stay with and assist at-risk family members (Ariyabandu 2009; Valdés 2009). Taken together, these findings on displacement risk related to intersectional social factors underscore the need to address systematic disadvantages in highly unequal societies such as Peru to lower disaster displacement risk. In addition, the findings call on humanitarian actors to prepare for the fact that people displaced by disasters can include a high share of persons in vulnerable situations due to certain social factors that do not only make displacement more likely but also increase post-disaster vulnerability. These results contradict another seminal study which finds that selectivity of migration may be limited in sudden-onset disasters; the displacement observed in that study was not predicted by age, education, sex, or socioeconomic status (Gray et al. 2014), although these factors are usually seen as influencing vulnerability to hazards (Cutter et al. 2003; Cutter et al. 2010; Cutter et al. 2014). By contrast, the statistical analysis herein indicates that certain social factors do raise displacement risk. This finding is in line with other papers which observe that (im)mobility in disasters is differentiated by various population characteristics (Black et al. 2013; Boas et al. 2020). Consequently, these factors may provide entry points for policy and planning to reduce displacement risk in future El Niño events in Peru. For example, improving education would support development and simultaneously decrease vulnerabilities to disasters and displacement. Education reduces disaster vulnerability regarding preparedness and response because it improves skills, knowledge, information, and resources to handle disasters, shapes risk perceptions, and supports wealth and health (Muttarak & Lutz 2014).

In summary, the statistical analyses underline that displacement risk was substantially driven by poor shelter and tenure insecurity; undiversified and vulnerable agricultural livelihoods located in deprived rural zones; physical disaster exposure; and intersectional social factors that reduced coping capacities and increased humanitarian needs. These detected factors could also inform recent efforts by international organizations to develop indicators for monitoring disaster displacement and ensuing risks (Guadagno & Yonetani 2022; IOM & IDMC 2022). The indicators are intended to feed into disaster-related assessment and monitoring processes. Beyond, these efforts to develop risk indicators should also consider other commonly applied disaster vulnerability metrics, such as social (e.g. social networks), institutional (e.g. local disaster training), or community capitals (e.g. community cohesion) (Berkes & Ross 2013; Cutter et al. 2014; Cutter & Finch 2008; Kim et al. 2018; Norris et al. 2008; Sherrieb et al. 2010), which this study could not explore due to a lack of available data.

The qualitative data from Piura reveals more details about the displacement trajectories. Most interviewees fled fast, over short distances, and for survival under strong constraining structural conditions. Subsequent settling was dispersed in UP, where people lived in public buildings, in tents, or with relatives, which stresses the need to address urban and non-camp displacement in Peru and worldwide (IDMC 2021c; UNHCR 2014). These migrants fled for a short duration and had mostly returned to their original land by the time of the interviews, which underlines the need to investigate micro-mobilities, an issue for which data is still limited and that remains unattended by policy and donors (Safra de Campos et al. 2017; Safra de Campos et al. 2020). Conversely, in LP, most people were displaced to nearby camps in the barren desert for a long period because their villages of origin remained devastated, which echoes global findings that displacement often implies prolonged challenges (Crawford et al. 2015; Devictor 2019). Most displaced persons in the camps intended to stay permanently out of fear of future floods. Nonetheless, they have kept translocal connections and occasionally dual residencies for educational and livelihood needs, which provides evidence that translocality approaches are needed in this field of study (Peth & Sakdapolrak 2020). Furthermore, the findings reinforce global calls to assess secondary displacement risk, cumulative shocks, and how more frequent climate impacts may influence patterns of recurrent displacement (Blocher et al. 2021b): in both LP and UP, many displaced persons had been previously displaced by earlier El Niño events, while several others had originally migrated to Piura due to climate impacts in the highlands, but then been forced to leave again by the CEN impacts. In addition to the large-scale, short-distance displacement analyzed in the qualitative data, other studies and anecdotal interview evidence point to limited, larger-distance moves driven by the CEN floods, which validates the global finding that vulnerability and mobility are often inversely related (Adger et al. 2014). Beyond abrupt El Niños, gradual processes such as warming and droughts have also shaped migration in Piura through complex effects, which points to the understudied intricacies of compounding effects of slow- and sudden onset hazards (IOM 2020).

4.2 Well-Being Effects, Structural Conditions, and Mechanisms

The statistical analyses demonstrate consistently that displacement negatively affected people’s well-being one and seven months after the disaster. Displacement greatly reduced well-being in statistically significant ways in all models, including in those controlling for environmental and socioeconomic parameters at the household and individual level. In the best-fit model with a sample of 162,331 affected adults, displacement led to an average 3.14-point loss on the well-being index that ranges from 0 to 100. The data reveals that displacement strongly deteriorated people’s space to live (including housing quality, tenure security, and basic services) and development from a secure base (based on the applied employment and health indicators) around half a year after the disaster.

While the quantitative analysis thus substantiates the short- and medium-term losses that the displaced persons experienced,^{Footnote 33} longer-term effects cannot be deduced from the data. The broader literature is also limited on this question but suggests that intersectional social factors influence the outcomes. In one of the few available studies, education was not predictive of post-traumatic stress one year after a disaster, but higher educated individuals had better housing, smaller declines in spending and consumption, and better psycho-social health four years later (Frankenberg et al. 2013). Similarly, several empirical studies in a special issue (Muttarak & Lutz 2014) confirm that people with higher education suffered lower loss and damage and could recover more swiftly after disasters than others. Future work could expand the analysis here and discern the determinants of recovery for displaced persons after the CEN in Peru to verify if intersectional factors also influenced their results over the longer term. For the development of the well-being effects up to one and a half years after the disaster (around one year later than the survey data), the primary qualitative data collected for this dissertation provided additional insights. These insights, including into a range of well-being metrics and mechanisms that were unavailable in the used datasets, are contextualized in the literature below.

4.2.1 Development from a Secure Base

The qualitative data demonstrates that the respondents often remained impoverished after they had lost their agricultural livelihoods, as seen in other studies of the 2017 CEN (IOM 2017c, 2018) and previous El Niños (Oft 2010; Sperling et al. 2008). Several mechanisms contributed to the severe economic losses of the displaced persons First, previous levels of livelihood diversification had been low, and as most farmers had depended entirely on farming or day labor, flood damage to their small farms was difficult to buffer. Second, displaced persons had often engaged in livelihood activities that were difficult to recover. It took long before they could plant and harvest crops or raise animals again for consumable or marketable outputs. Third, many respondents lacked savings or lost livestock in the floods, which hindered reinvestments. Fourth, few of the subsistence farmers disposed of skills transferable to job markets less affected by the floods, such as those in cities. Fifth, the effects of livelihood erosion were significantly moderated by the demand for day labor by agribusinesses in areas surrounding LP that were spared flood damage. However, this buffer was inaccessible for older adults in LP and not available to the same extent in UP. Sixth, household composition shaped the economic effects; the more healthy members at working age a household was composed of, the better it could pool income options for recovery. Finally, economic consequences were aggravated by alleged corruption and lack of humanitarian assistance by the state, and basic service gaps have impeded livelihood recovery. Conversely, help by NGOs or private actors, and mutual community support, slightly mediated economic damages.

Other studies also confirm the result that the 2017 CEN severely harmed people’s health. It damaged infrastructure (French et al. 2020); on-site health services in camps remained limited; mental health issues were widespread; infectious diseases abounded; and many of the displaced depended on gradually more irregular food distributions (Espinoza-Neyra et al. 2017, 2018; IOM 2017a, 2017b, 2017c, 2018). This study provides detailed insights into the mechanisms underlying these impacts. First, the floods destroyed health facilities and created lasting risks from new diseases. Second, in the camps, displaced persons suffered from more hazard exposure while living in substandard shelter with poorer protection. Third, flood damages and livelihood erosion, combined with a lack of savings or credit, hindered access to medicine or health services and raised food insecurity. The displaced also lacked psychosocial care options. The health risks were most salient for those with vulnerabilities relating to age, sex, and physical ability. Fourth, the lack of land titles in the camps further impeded access to social programs that could have enhanced health. Fifth, NGO support could only partially buffer the gaps of state presence, and their services dwindled over time with rising donor fatigue.

Similarly, prior work corroborates the finding that El Niño-induced damages and hardship tend to worsen educational opportunities (French et al. 2020; Sperling et al. 2008). In the 2017 CEN specifically, a large number of children were displaced and faced educational challenges (IOM 2018). The analysis here demonstrated that several mechanisms worsened their educational opportunities. First, flood damages damaged educational infrastructure and external assistance was insufficient for people’s needs during the prolonged displacement. Second, flood-impoverished families struggled to pay school fees or had to withdrew children for work. Third, physically accessing educational facilities resulted difficult due to the larger distances, strenuous trajectories, and costly passages. Finally, pupils struggled to study and learn due to the exhausting physical access, trauma, stress, anxiety, food insecurity, and the lack of light for homework.

Altogether, the findings therefore re-emphasize the concerns from the review in chapter 4 that pre-movement vulnerabilities set the limits for people’s development from a secure base, and that those most affected by severe hazards, and moving under survival conditions, have the highest risk to suffer losses (e.g. Afifi et al. 2016). This study also reinforces the result of the review that livelihood, health, and educational risks can persist for long after displacement, especially if external assistance is insufficient (e.g. Bruijn 2009; Cazabat 2020; Crawford et al. 2015; UNESCO 2020).

4.2.2 A Space to Live Better

Many of the displaced had consciously lived in flood-exposed areas because of livelihood and land needs or choices. In a vicious circle, they have struggled to repair or rebuild flooded homes at high costs close to every 15 years, which has often left them less prepared for future floods, as witnessed in the CEN. After the CEN, their prospects for recovery were poor for various reasons. First, few displaced persons had resources to reconstruct in a safe and adequate way. In LP, they settled in barren desert spaces too small for their needs and felt dissatisfied with their new setting. In UP, they returned to their familiar but damaged homes, and rued the destruction of their fields and missed the work there. Second, external assistance for shelter was deficiently funded, unevenly provided, and centered on temporary help instead of long-term development needs of prolonged displacement. Third, similar reasons (including the lack of state presence, corruption, uneven or short-term, unsustainable humanitarian assistance, and donor fatigue or withdrawal) also led to greatly worsening basic services for most of the displaced. In LP, land conflicts were an additional key hindrance for service investments by the state. The deprived displaced persons also lacked money to buy or repair decaying services, and crowded living conditions raised pressure on the weak infrastructure. This work thus echoes prior studies on the 2017 CEN in Peru which identified extensive challenges in housing, infrastructure, and basic services that persisted long after displacement (French et al. 2020; IOM 2017a, 2017b, 2017c, 2018). This analysis also reinforces the finding from the broader literature that the definitive losses suffered by displaced persons, their disenfranchised societal positions, and the lack of external help often impede recovery (e.g. World Bank 2017b).

Changes in the exposure to hazards depended on the availability of unaffected land after the displacement: in LP, people fled to an area safer from floods but exposed to other minor hazards, whereas in UP, they mostly stayed close to or returned to the same exposure. Few displaced persons received support for or could afford to build adequate shelter to protect themselves from future risk. Nonetheless, the cases in LP highlight that even displacement may yield limited gains when it helps people to move out of harm’s way, as seen in other areas of Peru (Jarman 2020). Conversely, returning may be desired for many reasons but can also mean a return to the same hazard exposure, as seen in UP. These findings echo global concerns that climate migration can bring safety from initial hazards, but also create new exposure (e.g. Adger et al. 2014), especially because migrants often settle in high-risk zones and are unaware of hazards (for Peru, see e.g. Rubiños & Anderies 2020).

Finally, two factors substantially contributed to physical insecurity. First, the absence of security personnel created conditions conducive to violence and crime during the floods and throughout the prolonged displacement. Patrols self-organized by the displaced buffered this insecurity moderately. Second, land conflicts with the private landowner in LP also raised violent tensions, which negotiations could partly offset with time. The review in chapter 4 confirms that insecurity is common in climate migration (Melde et al. 2017), especially for women (Fleury 2016), after disasters, and during displacement (e.g. McMichael et al. 2012). Prior studies specifically on the CEN 2017 also echo the insecurity reported here, and indicate that the lack of security provisions, secure lighting, or safe spaces for children and women in camps amplified risks (IOM 2017a, 2017b, 2017c, 2018). Contrary to the women interviewed in this study, IOM (2017b) also observed gender-based violence in the camps, especially where infrastructure such as public light and shelter was deficient. However, the answers in the study here may represent sample bias or relate to the interview situation, social taboos, fear of over-disclosure (Reczek 2014), or a normalization of violence.

4.2.3 Social Relatedness

Social relatedness remained good due to several mechanisms. First, communities had had strong social ties before displacement and because they preserved a similar spatial dispersion after settling in camps (in LP), or stayed in their villages (UP), they were able to maintain these bonds. Second, pre-existing mutual support systems as part of local traditions, such as communal tasks and food sharing, contributed to good relations. Third, people’s capacity to self-organize and the availability of capable community leaders were essential to channel social life and structure support. After displacement due to the CEN, this social capital was a key asset for coping and recovery, as witnessed by other studies in Peru, especially where people felt abandoned by authorities (Moncada et al. 2018; Venkateswaran et al. 2017). Although prolonged absences of family members did put pressure on social structures in some cases, this study thus contradicts the global finding that displacement often threatens social relatedness (e.g. Schwerdtle et al. 2020; World Bank 2017b).

4.2.4 SWB

How displaced persons appraised their new lives emotionally and cognitively depended on various factors. A first key variable was the severity of experienced flood losses and perceived recovery options. Many suffered greatly; sadness, pain, helplessness, anxiety, fear, and trauma were salient especially at the beginning of displacement but persisted or were (re-)produced over the arduous recovery. Second, respondents also reported subjective ill-being because they felt inadequately cared for or supported by external actors, which evoked feelings of anger, disappointment, frustration, injustice, and disgrace. Third, the lack of professional support for coping with negative experiences also worsened SWB. Fourth, personal coping mechanisms, including humor and religious faith, provided limited buffer. Fifth, social networks offered support and care that helped many households, but such networks were heavily missed when migration separated household members. The study thus adds to and echoes the limited evidence of present SWB after forced migration. Prior studies also find that SWB losses in displacement result from the lack of preparedness and the forced conditions of moving; scarring effects before and after moving; and intersectional vulnerabilities to emotional, social, and health risks of moving (e.g. Bartram 2015; Luhmann et al. 2012). The analysis here focuses on the internal standards of the displaced respondents (pre- vs. post-movement). Relative deprivation related to stayers or locals—another key strain on SWB identified in the literature (e.g. Chen et al. 2019)—is unlikely in Piura, since most respondents were similarly affected, almost no one stayed in the source areas, and destination zones lacked pre-defined host communities. Finally, the cross-sectional design makes it difficult to assess if hedonic adaptation has occurred. However, the data collected 20 months after the floods emphasized that people still suffered from a large SWB decline. The wider literature also suggests that scarring can lastingly reduce SWB (e.g. Kettlewell et al. 2020; Luhmann et al. 2012; Mousteri et al. 2018), which seems likely here due to the unemployment, financial losses, and health shocks caused by the CEN. Other qualitative studies on the CEN also report persistent suffering and despair after the disaster (Moncada et al. 2018).

Almost no displaced persons held neutral views of the future, and many expressed hopelessness as they suffered from combined, large well-being stressors. Such despair may further deepen perceptions of vulnerability and block action or engagement (e.g. Schueller & Seligman 2011). Most but not all of the mechanisms identified in the reviewed literature reduced hope (Edey & Jevne 2003; Snyder 2002). Especially experiences of trauma and loss as well as lack of control and progress (such as related to weak governance, economic hardship, and tenure insecurities) raised despair in Piura. Conversely, alienation or a lack of social connections were not observable. Religious faith was one key contributor to hope and consolation, as known from other contexts (e.g. Lim et al. 2019), but faith also evoked feelings of guilt, regret, and surrender in some cases, an effect still understudied. Other people nourished hope thanks to trust in their individual capabilities, solidarity in their communities, and a belief that external actors would support them. The social narrative that progressing was possible (salir adelante) created optimism for many migrants, and such optimism can be a key motivator of further productive action (Forgeard & Seligman 2012). However, given the factual hurdles, such optimism seemed unrealistic and self-deceiving in some cases, and such forms of illusionary hope can be detrimental (Turner 2017). Finally, many of the displaced persons held mixed sentiments toward the future, confirming the need to consider outlooks on the future as layered, as the review in chapter 4 suggested. For example, faith in god and patience regarding recovery would mix with pessimism and anxiety regarding future floods in some cases, similar as in a study of climate migration in the Pacific (Yates et al. 2021). Overall, because SWB is key for effective functioning, displaced persons’ decline of present SWB and their challenged views of the future can pose cascading threats for other OWB dimensions (e.g. Carver & Scheier 2014; Diener et al. 2018b).

5 Summary and Induction of Propositions

In this section, I first briefly recap the observed well-being dynamics, mechanisms of action, and structural conditions in this case of displacement (forced, acute migration) under survival conditions. After that, I induce broader propositions on the potential well-being implications of climate migration.

Figure 7.14 provides a visual summary of the identified well-being effects, relevant structural conditions, and mechanisms of action. It illustrates that displacement occurred under high structural constraints, including severe climate risks and deficient DRR/DRM; poverty and inequality; limited livelihood options; tenure insecurity; poor and hardly accessible basic services; weak governance; and limited political participation. Various of the adverse structural conditions identified through the qualitative analysis as influential drivers of well-being were confirmed by the complementary statistical analyses. Conversely, structural opportunities were low, such as the influence of non-state actors. As a result, the severe and abrupt CEN floods caused forced, survival migration with detrimental conditions for moving and settling. Throughout the lifecycle of displacement, people suffered extreme losses, which continued to worsen their prospects for development from a secure base and a space to live better. Conversely, social relatedness remained similar after moving. Because displacement has become prolonged without substantial improvements, people’s need fulfillment, long-term asset base, and capacities for climate adaptation have worsened. Consequently, most displaced persons evaluated their need fulfillment as negative (deprivation), and only few experienced partial positive feelings or cognitive satisfaction despite their plight (adjustment).^{Footnote 34} Expectations for the future were mostly negative and resulted in prevalent enforced fear as well as some fragile adjustment. The light gray boxes in the figure summarize key conditions and well-being mechanisms which may also be influential in other contexts.

Based on the findings in this case of survival displacement (forced, acute migration) by Peru’s coast, one can derive more general propositions on the well-being impacts of climate migration.

(1)
First, the results stress that in the constraining structures of societies with weak governance, ineffective DRR/DRM, insecure tenure, and undiversified subsistence farming with minimal margins, sudden-onset hazard displacement can greatly reduce people’s agency and prospects for development from a secure base. When effective DRR/DRM is unavailable, repeated displacement may be the only way to safe one’s life. When such forced, acute migration becomes prolonged or protracted without an avenue for self-reliance, it is likely to worsen people’s long-term asset base and capacities for climate adaptation in dangerous areas. Women, minors, older adults, and those with health limitations are at especially high risk of losing resources.
(2)
Second, climate change also increases the risk that poor farmers who need to carve out a precarious living in areas exposed to recurring, abrupt hazards can suffer from cumulative shocks, whose destructions diminish their chances to obtain a space to live better. After displacement, farmers’ losses, insecure asset base, the lack of land and housing titles, along with neglect by authorities, can force them into vicious circles of inadequate housing, deficient access to basic infrastructure, and perennial insecurity.
(3)
Third, strong community bonds and traditions, self-organization capacities, and neighborhood continuity can create conditions that shield displaced groups from some of the possible harm. Where these factors exist, social relatedness may be preserved.
(4)
Finally, the results suggest that the SWB of forced, survival climate migrants is at risk. Profound losses, arduous recovery, and negative governance experiences make it likely that OWB and SWB losses converge into a present subjective state of deprivation. Extreme stress results as the displaced lose resources, or resources (such as farming skills) become devalued or obsolete, and as their goals are obliterated or obstructed after displacement. Subjective adjustment may occur exceptionally when strong social networks combine with personal coping mechanisms such as faith or humor. Whether deprivation and adjustment are accompanied by optimism (precarious hope or high adjustment) or by pessimism regarding the future (enforced fear or fragile adjustment) depends on the balance between persistent, large stressors and setting-, community-, or person-related resilience factors. The greater the stressors, the more likely more fear than hope for the future. Because despair is likely to be salient in displacement under survival conditions, negative ripple effects on other well-being dimensions are to be expected.

Notes

1.
This section partially draws on my previous work published in Bergmann et al. (2021a).
2.
For privacy protection, neither the names of the study sites nor of interviewees are specified.
3.
The station name is concealed for privacy protection. I would like to thank again my colleague Stephanie Gleixner who kindly produced graphs for the results chapters on the Sierra and Costa with PISCO data, and for the Selva chapter with station rainfall data and PISCO temperature data.
4.
Huaycos are flash floods formed in the highlands after extreme rainfall, which carry mud, rock, and debris flows through ravines and valleys.
5.
Climate models in the IPCC’s Fifth Assessment Report used four representative concentration pathways (RCPs) to cover possible future emissions, see van Vuuren et al. (2011). RCP8.5 (high) assumes the highest CO2 concentration, leading to a global surface temperature increase of 2.6 to 4.8 °C for the end of the 21st relative to the end of the 20th century. RCP2.6 (low), the most optimistic scenario that requires strong mitigation suggests an increase of 0.3 to 1.7 °C. RCP6.0 and RCP4.5 (medium) fall in between with 1.1 to 2.6 °C and 1.4 to 3.1 °C warming. See IPCC (2014).
6.
The results are averages of four ISIMIP models.
7.
SSP 3.70, leading to above 4 °C mean global temperature increase by 2100.
8.
3-sigma or 5-sigma events are monthly and seasonal temperatures >3 or >5 standard deviations (SD) warmer than mean temperature.
9.
The collapse of the West Antarctic Ice Sheet could add several meters over the long term and a loss of the Greenland ice sheet would add about seven meters on a timescale of centuries to millennia. See IPCC (2019a).
10.
In 2010, Peru’s Ministry of Environment expected SLR of 0.60 to 0.81 m along the northern coast of Peru over the next 100 years (MINAM (2010)), but recent studies suggest more severe impacts, see Pörtner et al. (2019).
11.
I would like to thank my colleague Roman Hoffmann, with whom I collaborated for the quantitative analysis in this chapter, for producing this figure. It uses MERRA-2 rainfall data by NASA, which is based on GPM satellite data. Riccardo Biella helped with the data extraction and resampling to a 1 km resolution applying bilinear interpolation.
12.
This section does not cover strategies of fishers and scallop farmers living directly by the sea, who employ different coping and adaptation techniques to deal with ENSO hazards. Their strategies include prey- or gear-switching, changing harvesting areas, or exiting fisheries or scallop farming. See Badjeck (2008); Badjeck et al. (2009); Kluger et al. (2019).
13.
To deal with water scarcity, most coastal farmers ask for help from peers or reduce expenses; others cannot do anything, work additional time, ask for credit, use different crops, extend the used land, migrate temporarily, sell livestock, attempt to produce charcoal, and resort to child labor, see Aragón et al. (2018); Oft (2009, 2010); Sperling et al. (2008).
14.
While not the focus of this study, El Niño events also shape the availability of marine resources, which, in turn, affects fisher’s reliance on seasonal or long-term migration, see Badjeck (2008); Kluger et al. (2019); Kluger et al. (2020).
15.
Interviewees had allegedly received this land after the 1998 El Niño; it was, however, eventually sold to two agribusinesses, resulting in land conflicts after the CEN caused displacement in 2017.
16.
Similarly, the displaced persons perceived that the few longer-distance migrants who had gone to work temporarily in Lima or Piura city often ended in precarious work, without social security, unstable income, and irregular housing.
17.
A non-representative survey in May 2017 showed that in 80% of camps, displaced persons had formed communal committees to help with the site management. 63% of respondents in camps in Piura received their information primarily from community leaders and 77% identified them as the most trustworthy sources. See IOM (2017b).
18.
The analysis excludes children below 18 years to avoid double-counting and because the survey did not contain relevant data for them on key well-being items, such as employment.
19.
The CEN survey identifies “damaged” persons, but this designation is only an indication of possible displacement. One cannot take for granted that all “damaged” persons were displaced because INEI applies the label “damaged” to persons who either had uninhabitable houses, or suffered severe economic losses, including unemployment or losses of agricultural assets, or whose health was affected due to the CEN. See INEI (2017c: 8).
20.
As explained in section 3.2.2, this assumption that the habitability of homes is a proxy for displacement is a plausible basis for the analysis but comes with certain limitations.
21.
For 1,550 datapoints, enumerators noted that there was no infrastructure or land in the designated lot; they were excluded from the analysis. Likewise, households that consisted only of children were excluded from the analysis.
22.
The data sources for these regression parameters included GTOPO30 and MERRA-2, as explained in footnote 28 in Chapter 3.
23.
INEI considers settlements with less than 2,000 inhabitants as rural. See INEI (2018c: 15).
24.
The variable is coded as follows. Decent types of housing refers to houses and apartments or rooms in tenements. Non-standard types include collective housing (hotel, hostel, lodging, asylum) and other types. The reference for both is deficient types of housing, such as huts, cabins, improvised housing, or premises not intended for human habitation.
25.
All variables referring to a percentage share of household members are scaled from 0 to 1. For example, 0.1 represents 10% of female household members. Thus, the effect on the dependent variable gauges an increase from 0% to 100% of women. A 10-percentage point increase of women in the household would lead to an effect of −0.1283/10 = 0.0128.
26.
Footnote 25 applies to all parameters that gauge percentage shares of household members.
27.
Decent types of housing include houses and apartments or rooms in tenements, as opposed to huts or cabins, improvised housing, and others. Hotels, hostels, lodging, boarding houses, health facilities, penitentiary, residential care centers, army bases, and houselessness are excluded.
28.
The reference here are all other types of housing, including non-standard and deficient housing.
29.
Footnote 6 applies to all parameters that gauge percentage shares of household members.
30.
The insights from other studies referenced in this and the following paragraphs draw in part on my previous work published in Bergmann et al. (2021a).
31.
Another hypothesis which cannot be assessed with the applied data is that homeowners might also have decided to stay longer due to higher place attachment, for example, to protect their assets.
32.
A third hypothesis, which cannot be examined through the data, is that households with at-risk members might have had to move fast to seek assistance and protection and, as a result, were less likely to remain in their damaged homes.
33.
As discussed in the limitations in section 3.2.2, the data does not allow to distinguish if the displaced people remained in displacement or were able to return with the analyzed period. Thus, it is unclear if displacement itself has an impact regardless of the time spent away from home or if short periods of displacement affect well-being even after months.
34.
For terminology, see section 2.3. Deprivation refers to adverse objective conditions accompanied by low cognitive satisfaction and negative emotional balance. Adjustment is subjective well-being despite adverse objective conditions.

Author information

Authors and Affiliations

Berlin, Germany
Jonas Bergmann

Authors

Jonas Bergmann
View author publications
You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Jonas Bergmann .

Rights and permissions

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Reprints and permissions

Copyright information

About this chapter

Cite this chapter

Bergmann, J. (2024). Costa: Flood Displacement During the 2017 Coastal El Niño in Peru. In: At Risk of Deprivation. Studien zur Migrations- und Integrationspolitik. Springer VS, Wiesbaden. https://doi.org/10.1007/978-3-658-42298-1_7

Download citation

DOI: https://doi.org/10.1007/978-3-658-42298-1_7
Published: 20 October 2023
Publisher Name: Springer VS, Wiesbaden
Print ISBN: 978-3-658-42297-4
Online ISBN: 978-3-658-42298-1
eBook Packages: Life Science and Basic Disciplines (German Language)

Publish with us

Policies and ethics

Costa: Flood Displacement During the 2017 Coastal El Niño in Peru

Abstract

1 Context

2 Qualitative Empirical Results

2.1 Climate Change Dimensions

2.2 Migration Dimensions

2.3 Well-Being Dimensions

2.3.1 Development from a Secure Base

2.3.1.1 Decent Livelihoods

2.3.1.2 Health and Food Security

2.3.1.3 Educational Opportunities

2.3.2 A Space to Live Better

2.3.2.1 Adequate Housing

2.3.2.2 Basic Services

2.3.2.3 Pleasant Surroundings

2.3.2.4 Safety from Hazards

2.3.2.5 Security

2.3.3 Social Relatedness

2.3.4 SWB Dimensions

2.3.4.1 Emotional Balance and Cognitive Satisfaction

2.3.4.2 Outlook on the Future

3 Quantitative Empirical Results

3.1 Magnitude of Displacement and Differential Displacement Risk

3.2 Short- and Mid-Term Impacts of Displacement on Well-Being

4 Discussion

4.1 Hazard-(Im)mobility Links and Pathways

4.2 Well-Being Effects, Structural Conditions, and Mechanisms

4.2.1 Development from a Secure Base

4.2.2 A Space to Live Better

4.2.3 Social Relatedness

4.2.4 SWB

5 Summary and Induction of Propositions

Notes

Author information

Authors and Affiliations

Corresponding author

Rights and permissions

Copyright information

About this chapter

Cite this chapter

Download citation

Share this chapter

Publish with us

Search

Navigation