We would like to thank all those who agreed to respond to our requests and allowed us to complete our fieldwork despite the post-disaster situation. We also thank the individuals, institutions and associations that facilitated the access to the respondents. We would like to express our deepest gratitude to the respondents for their time, trust and kindness. We are well aware that it was difficult for some interviewees to look back on this traumatic period in their lives, so we thank them for sharing their experience, memories and feelings. Last but not least, we would like to thank all the institutions that participated in the fundings of the missions, particularly the ANR which financed the TIREX project under which we conducted this study.
1In the early morning of September 6th, 2017, the “monstrous Irma” (France-Antilles 2017) devastated Saint-Martin island. The French part of the island, which is the focus for our study, then faced a major crisis situation, characterized by a high level of damage, disruptions and social disorganization. Socio-technical systems, which allow daily life and economic activities, went through, at least temporary, partial or total collapse.
2Saint-Martin is a 90km² Franco-Dutch island located in the lesser West Indies. The island is divided into a northern part of 56km² under French administration - Saint-Martin - and a southern part of 34km² under the administration of the Kingdom of the Netherlands - Sint-Maarten. The two parts of the island are therefore two very different systems, which have their own infrastructure and set their own standards (Redon 2006).
Figure 1: The Island of Saint-Martin to the North of the Lesser Antilles.
Sources: Esri, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community ©Amélie Latreille. ArcGIS, 2018.
3Irma was a hurricane classified in category 5 on the Saffir-Simpson scale with average winds estimated at 295 km/h and gusts above 360 km/h (Cangialosi et al. 2018; Nicolas et al. 2018). As on many other Caribbean islands, Hurricane Irma caused major destructions on the island of Saint-Martin: more than €1 billion insured property was lost (CCR 2018). The disaster that followed the hurricane was presented as an “extraordinary”, “exceptional catastrophe”, - quoting crisis managers we interviewed. In addition to the destruction of many buildings, Irma hit the various lifelines’ networks: roads, electricity, telecoms. Desalination plant and water distribution networks were severely damaged in the French part of the island (Gustin 2017). Yet the desalinated water production and distribution systems are officially the only source of drinking and sanitary water on the island. Indeed, the island, often described as a - “dry island”, has no permanent watercourse and the Geological and Mining Research Office (BRGM - Bureau des Recherches Géologiques et Minières) considers that its groundwater resources are very limited and brackish in nature (BRGM 2016). One of the reasons for the island's “aridity” is its low rainfall relative to other Caribbean islands, such as Guadeloupe and Martinique.
4Considering that the island didn’t have its own freshwater resources, crisis managers at national level concluded that Irma’s outcomes could lead to a severe or even total disruption of water supplies. National and local political authorities thus presented access to freshwater, especially drinking water, and the restoration of production-distribution facilities as priorities in relief operations. From September 6th, situation points alerted to the emergency affecting local communities. In addition, they mentioned numerous looting, which were also widely reported in the press and social media. The risk of freshwater shortage was therefore treated both as a threat to civil security (i. e. rescue of people) and as a threat to public security (i. e. maintenance of public order). For their part, network operators quickly communicated on their mobilisation “to respond to the urgent need to restore these essential services”. As of September 8th, UCDEM and Générale des Eaux de Guadeloupe, Veolia's two subsidiaries in charge of the production and distribution of drinking water on the island, explained that they were mobilizing all possible means to “help the affected populations and restore access to drinking water as soon as possible” (ibid.)
5In case of natural disaster, restoring access to drinking and sanitary water have long been seen as critical (Clark & Deniger, 2000). Indeed, water supply disruption or interruption, even for short periods of time, can cause serious health damage (Hardy 2009). Bross et al. recall that “drinking water supply is at the core of both, humanitarian action in times of crisis, as well as national policies for regular and emergency supply” (Bross et al. 2019). R. D'Ercole and P. Meztger classify water supply systems as “major elements at risk” considering that these territorial system components play a strategic role in the system functioning and that they present at the same time significant vulnerabilities (D'Ercole & Metzger 2009). As such, water infrastructure systems (e.g., drinking water treatment, transmission, and distribution; wastewater and stormwater collection and treatment) are considered as “critical functions of a healthy society” and “disaster resilience of these infrastructure systems during and after hazards is vital for the response and recovery to the event” (Matthews 2016).
6This paper therefore aims to address post-disaster resilience after Hurricane Irma by looking at the water supply system recovery. Resilience can be both considered as an outcome, defined by the ability to bounce back after a shock or a disruption, and as a process involving both responses, coping and adaptive capacities, self-organization and learning (Adger 2000, Folke 2006, Manyena 2006, Gaillard 2010, Weichselgartner & Kelman 2015). Our study is related to academic work that considers resilience as a post-disaster recovery process. We focus on emergency, restart, restoration, recovery, reconstruction, etc. We address the responses and coping strategies implemented at the institutional and non-institutional, public and private, collective and individual levels in order to deal with major damages and crisis (Reghezza & Rufat 2015 and 2019). We examine more specifically the drinking and sanitary water access in the aftermath of hurricane Irma in the French part of the island. Given the already abundant literature on disaster management for the 2017 hurricane season in the other Caribbean islands (Medina et al. 2019, Eakin et al. 2018, Baptiste and Devonish 2019) and the very distinct political-administrative organisation and water standards in the two parts of the island, we have chosen to exclude the Dutch territories in our study.
7First, we seek to grasp the coping strategies developed by various stakeholders in order to deal with the interruption of drinking and sanitary water services. On one hand, we consider the actions undertaken by local and national authorities and by network operators, in response to the emergency, according to the objectives associated with the water infrastructure resilience: maintain access to the service, even in degraded capacities, restart freshwater production, restore drinking and sanitary water distribution and rebuild infrastructure. We also examine NGO supply. On the other hand, we focus on the strategies implemented at the individual and collective level by local communities and inhabitants.
8Second, we seek to understand how institutional responses (public authorities, network operators, NGOs) and population responses, at both individual and collective levels, interact in the emergency period and in the post-disaster recovery process. We assume that responses and coping strategies implemented before and after Irma must be related to the pre-crisis situation re-examined in the light of pre-crisis inhabitants’ practices. We consider the general situation of freshwater access in Saint-Martin, the drinking and sanitary water production and distribution systems functioning, the water governance, the inhabitants’ knowledge and behavior. We also address the very notions of crisis, emergency and resilience.
9In the field of natural disaster risk reduction (DRR) and disaster risk management (DRM), water is considered as a risk, through water-related hazards and disasters. The issue of water access or water security is indirectly addressed in works on drought hazard (Luh et al. 2015) or safe drinking water scarcity (Bakker 2012, Abedin et al. 2014). Water access is also covered by a growing number of research works focusing on climate change impacts, particularly in urban systems (Hoekstra et al. 2018) and thought in terms of water systems adaptation to global changes (Scott et al. 2012). At the same time, water studies focus on water management and uses.
10Drinking water access - and to a lesser extent sanitary water - aftermath disaster is addressed in disaster studies, but paradoxically, rarely as such.
11Several works in health studies discuss the risks associated with non-drinkable water consumption: they focus on resulting infectious diseases (Fredrick et al. 2015, Basnyat et al. 2015), on the impacts of water, sanitation, and hygiene (WASH) interventions on the health of populations (Ramesh et al. 2015) or on population with specific vulnerabilities (pregnant women, children, newborns, patients, etc.) (Callaghan et al. 2007, Exum et al. 2018).
12In the literature about disaster relief, a few studies address the technical and operational aspects of water security and water sanitation within crisis management and humanitarian response. They study production and/or distribution water in the aftermath of disasters (Nolz et al. 2010), analyse technical aspects (desalination, supply chain, distribution optimisation, etc.) (Loo et al. 2012, Chew et al. 2015), or examine the social dimensions of water access, highlighting the role of inequalities, particularly those related to gender, age and income (Sultana 2010, Mirza & Mustapha 2016). Most of these research studies focus on developing countries, pointing out that the minimum standards used for humanitarian assistance do not match with those of developed countries where people mostly have permanent access to tap water (Bross et al. 2019).
13In our study, we excluded research questions that arise in developing countries because feedback from authorities, operators and associations showed that problems of drinking water contamination and hygiene problems had not emerged.
14Drinking and sanitary water issues in disaster's aftermaths are also discussed in the literature on critical networks, i.e. infrastructure of vital importance for the functioning of human society (Lhomme 2017), which has been renewed by the “resilience” concept (Curt & Tacknet 2018). Several recent studies estimate the ex-ante resilience capacity of water distribution systems and water supply networks (Christodoulou et al. 2018, Shin et al. 2018, etc.). Based on previous work on critical networks vulnerability, they identify several key aspects of drinking water system resilience (e.g. redundancy, stability, integrity) as being critical to the recovery process (Matthews et al. 2014). Some studies combine issues of vulnerability and resilience through the reliability of water systems (Gheisi et al. 2016). An abundant literature also points to the role of complex interdependencies and domino effects between different critical networks, including water facilities (Bigger et al. 2009).
15Social and biophysical aspects of freshwater resources are strongly intertwined. “Water systems" tie together biophysical systems (hydro-climatic and geological conditions), networks, production and distribution systems, users and institutions responsible for management (Ferguson et al. 2013). Moreover, water management and uses are intrinsically bound to economic and social processes, as much – if not more – than physical parameters (McCarroll & Hamann 2020).
16Water resource uses and management are deeply rooted in social and cultural structures. For instance, water knowledge and rules that support water exploitation and preservation stem from western science, but also from cultural and religious traditions, historical legacies, and more broadly from the complex relationships between communities and their environment (Ibid.).
17Sociology of water policies demonstrates that water access and intersectional inequalities are related and confirms that water crisis are socially constructed. Considering water resource governance, numerous work address freshwater distribution, use and management issues. Far from being a simple technical issue, water management is a political problem, involving a set of decisions. Scholars then identified the role played by coalitions of public and private actors, who share common representations and values and act in a coordinated manner (Poupeau et al. 2019).
18Other studies focus on the "co-constitution of water and power" (Budds & Linton 2014). Political ecology – especially critical or radical geographers – provides here a major contribution, stressing the need to "re-socialize" and "re-politicize" water, without skipping its materiality (Boyer 2020). In contrast to the hydrological cycle, which leads to consider water in an asocial and apolitical way, the concept of the hydrosocial cycle, forged in the early 2000s, encompasses both the social political and biophysical complex processes that constitute the water resource. The hydrosocial cycle has widely been used “to capture deepening entanglement of water flows and power relations” (Budds et al. 2014). The “waterscape” concept refers to the hybrid character of the water landscape, which combines the physical, social and political (Swyngedouw 1999). It reflects the concrete functioning of the hydro-social cycle.
19Given this strong social anchoring of water services, it is understandable that in the event of a malfunction, they can be a source of high vulnerability, particularly in urban and dense areas. Indeed, dependence on these networks for many activities and services (food, health care, etc.) determines their criticality. Disruptions to a water supply system have consequences for the entire territorial system, due to domino effects, sometimes causing serious crises (Hardy 2009). Identifying these vulnerabilities and accurately determining the minimum local water needs, considering the alternative sources of supply used by certain communities, often among the most precarious, can help mitigate the consequences of such crisis and implement actions to limit their severity. This is precisely what Sébastien Hardy shows in his study on La Paz, which highlights the differential vulnerability of groups who, depending on their income, mobilise alternative sources of water supply (Ibid.).
20Then, combining water studies and disasters studies to shed light on post-disaster recovery, we aim to show that the communities’ resilience, beyond purely technical responses that focus on drinking water networks restoration, is anchored in a timed and situated hydrosocial cycle, in which people have built alternative to the use of drinking water network.
21In the early 2000s, projects on lifelines network resilience introduced a territorialized approach by considering technical, organizational, social, and economic dimensions in critical networks vulnerability reduction and recovery (Bruneau et al. 2003). In a reverse perspective, some scholars focus on water supply systems' role in social systems resilience, particularly in urban context. They stress the lifeline system restoration in the aftermath of a crisis or disaster as an essential condition for the population's safety, activities restarting and rebuilding or recovery process. Water supply system disruption has indeed significant impacts, even on disaster relief operations (Berariu et al. 2015).
22Several studies also link macro-scale resilience (region, community) and water service resilience, showing that community resilience depends on water supply system resilience (Craig & Davis 2014). Based on studies that address the social dimensions of vulnerability and resilience, some researchers show that water access is a key factor of individuals and communities’ resilience. They focus more specifically on how individuals cope with water, sanitation, and hygiene (WASH) disruption after a disaster (Uprety et al. 2017).
23Some scholars use water access to assess post-disaster recovery at community scale or local scale. Based on in-depth interviews and focus groups, they seek to estimate the rate of infrastructure recovery based on people's effective access to water (Zhu et al. 2017). Others assess the evolution of consumption practices, which reveals different rates of reconstruction (Shrestha et al. 2017).
24Our study belongs to the work that seeks to understand the interactions between social systems and socio-technical systems resilience. We adopt a territorialized approach perspective (Cutter et al. 2008) that emphasizes both immediate responses to the crisis situation and mid- and long-term adaptive strategies to biophysical, social and economic environments.
25Saint-Martin's vulnerability results from a combination of biophysical and social factors (Duvat 2008). Biophysical vulnerability can be defined by physical process features (nature, frequency and intensity), degree of exposure and the socio-ecological system sensitivity. In Saint-Martin, hurricane frequency is very high (17 hurricanes in the 20th century) and since 1966, seems to have increased in number and intensity (ibid.). This exposure is aggravated by the island's topography: the narrow, sandy beaches are severely eroded by cyclonic events and are highly prone to marine flooding. The many, often illegal, constructions that occupy these spaces are therefore very exposed to risk (Duvat 2019).
26The island's social vulnerability is mainly linked to demographic and socio-economic dynamics. In 1986, the French Pons tax exemption law led to a very strong growth in the tourism and construction sectors. The new workforce needs were met by the arrival of many working poor from neighboring islands (particularly Haiti, Dominica, the Dominican Republic). In 2012, the native population represented 30.8% and the immigrant population (born abroad) more than a third of the island's total population (INSEE 2017). However, since 1988, tourism has lost its vigor. Local communities suffer from a high level of social precariousness highlighted by social indicators: 33% of Saint-Martinois are unemployed and 60% of them receive at least one social allowance (compared to 47% in mainland France) (Gustin 2017). This social precariousness is coupled with housing precariousness in the often informal and illegal settlements that have developed on the island, in areas particularly vulnerable to flooding. This is particularly the case for Quartier d'Orléans and Sandy Ground.
27The vulnerability is finally exacerbated by the partition between two national sovereignties. Although the border between the two parts of this small island is very porous - inhabitants and goods are free to move between the two sides as they wish - the infrastructure is strictly separated. This is the case of the water supply networks and the administration of water services, which are separate and uncoordinated. These coordination difficulties between the French and Dutch administrations limit global risk management strategy or coherent planning policy. In the aftermath of Irma, the lack of effective coordinate public policies was highlighted, including by national authorities (Gustin 2017).
28In the middle of the 20th century, the idea that Saint Maarten was a “dry island”, without freshwater resources of its own, gradually emerged, supported by scientific work. It explains the very limited size of the hydrographic network (Figure 2) - mainly composed of temporary gullies (active only in rainy weather) and fifteen brackish ponds – by high insolation, favorable to evaporation, and low rainfall explain (BRGM 1990, Coulon & Labadie 2015).
Figure 2: Water resources in Saint-Martin (French part).
Sources: ESRI, HERE, Garmin, OpenStreetMap contributors - BRGM - IGN. ©Amélie Latreille. ArcGIS, 2018.
29Taking this into account, in the 1960s, national authorities and the municipality of Saint-Martin wanted to anticipate the impacts of population increase on water security. They also considered new quality requirements, in order to support tourism, which was developing at that time in the Dutch part of the country. French administration then chose the desalination solution. A factory was built in 1966 in the port of Galisbay, producing 500 m3/d. Improved technologies have increased production capacity over the years: in 1975, 1000 m3 were produced daily; in 1986, the distillation system produced 5000 m3/d and in 1989, nearly 8000 m3 left the plant each day (Redon 2010; BRGM 2018). In 2006, UCDEM implemented the reverse osmosis process to enable further savings. Indeed, this production system is less expensive and more efficient as it can produce 9000 m3/d, even if it remains very energy-intensive (GDEG, 2015; Redon, 2010).
30Initially, the drinking water production, distribution and sanitation services were managed by the municipal water authority. Then in 1984, the municipality moved towards a public service delegation and concluded a 35-year concession contract with a private company, the Caribbean Union of Seawater Desalination (UCDEM), a subsidiary of Veolia, which built a thermo-compression unit (Redon, 2010). In 1998, they signed a leasing contract with a second Veolia subsidiary, Générale des Eaux de Guadeloupe (GDEG), and gave it the distribution charge. Eventually, in 2014, a six-year concession contract entrusted Générale des Eaux de Guadeloupe with the sanitation service. From then on, the desalinated water produced by UCDEM sells to Générale des Eaux de Guadeloupe. The latter is responsible for distributing the water via the supply network of the EEASM, a public water agency created in 2006 by the local authority with which the two subsidiaries have contracts.
31Consequently, the water supply governance currently involves many stakeholders (Figure 3), who are responsible for the operation, maintenance, upkeep and management of the various components of the drinking and sanitary water production and distribution system. These actors are part of the public and private sectors since the public authority delegates the management of the public water services to private companies.
Figure 3: Actors and water governance in Saint-Martin between 2006 and 2018.
Source: eeasm.fr, consulted in November 2018. ©Amélie Latreille. Publisher, 2018.
32In the summer of 2017, Veolia's subsidiaries began negotiations with the local authorities to put an end to their contracts. This process was interrupted by Irma. The termination eventually happened at the end of November 2018 and in December, Veolia's two subsidiaries were replaced by a single company: Saur.
33This termination testifies to the crisis that the public water service of Saint-Martin was going through, long before the passing of Irma. In 2015, the water supply network efficiency in Saint-Martin was estimated at 56.3% compared to 80% in mainland France (Générale des Eaux de Guadeloupe 2015). For example, more than half of the water produced at the Galisbay desalination plant did not pass through the water meters. Leaks on the network are the main factor explaining this low performance. They are generally related to the equipment's obsolescence and the lack of maintenance work. In addition, "internal deterioration" due to the water composition weakens equipment and pipes. Tap water is four times more chlorinated than normal to prevent the growth of bacteria, favored by the dilapidated state of the pipes and the high temperature of the water produced (40°C at the plant's outlet) (Redon, 2010). Chlorine attacks equipment and pipes from the inside. Impacts can be seen in the field on pieces of pipe and gnawed joints reported by plumbing customers.
34Those low performances result in frequent water cuts. Consequently, many inhabitants choose independent modes of supply, adapting to the dysfunctions of the public water service. In addition to the extensive use of bottled water and rainwater recuperation tanks, private distribution systems are very common. Water infrastructure map shows that they are very important in the sector of Terres Basses, mainly in villas occupied by wealthy categories (Figure 4). According to the Territorial Chamber of Accounts of Saint-Martin, these private networks are a threat to public service, but also to the individual's health. Users install autonomous production systems that they do not declare so local authorities have no way of controlling the water quality (Chambre territoriale des comptes de Saint-Martin 2017).
Figure 4: Organisation of drinking water supply and sanitation networks and location of the building.
Sources: ESRI, HERE, Garmin, OpenStreetMap contributors - EEASM - INSEE 2014. ©Amélie Latreille. ArcGIS, 2018.
35Inhabitants have thus set up alternative ways to access freshwater in response to the crisis situation encountered by the public water service. Considering that, we must then examine whether these practices and habits played a role in crisis management and the resilience of population aftermath Irma.
36When hurricane Irma passed over Saint Martin in the early morning hours of September 6, it partly destroyed the only desalination plant of the French part: roof was washed away, some walls were down, control cabinets were flooded, etc. (feedback presented by Veolia, Chalandre 2017, Fagiani 2018). The distribution network was also affected: the underground part was weakened and some pipes were destroyed by marine submersion. In Saint-Martin, connections and counters were aerial: some of them were blown off by the wind; and when the streets were cleaned, many of them were destroyed by backhoes (interviews 2018 and 2019). According to the operator Veolia, 20% of connections and meters were out of service after Irma. In addition, there was damage to the sewer system. Treatment plants were affected to varying degrees - two of them were even destroyed. Consequently, the purification activity was stopped (Fagiani 2018). Almost all discharge stations were out of service and the sewerage networks were blocked by debris, creating a risk of stagnant wastewater pockets.
37These elements, combined with the idea that the "dry island" had no other water resources, led the authorities and the media to talking about a "water crisis". However, the testimonies of the populations, those of practitioners and crisis managers on the spot, the feedback and reports of the various NGOs, highlight the continuity of water access, the relatively rapid restoration of networks and the absence of health problems.
38Three questions arise then. First, how has the “water crisis”, or more precisely the drinking and sanitary water shortage risk, been managed by the authorities, locally and from mainland France? Secondly, how was the water service restoration, particularly in drinking water distribution, achieved? Finally, how did the population perceive drinking water supply issues during the emergency and immediate post-cyclone phase and what responses did they provide?
Table 1: Methods used for the article and their objectives.
Method
|
Objective
|
Period and place
|
Analysis of press articles, grey literature
|
1. Pre-Irma situation analysis 2. Crisis and response chronological reconstitution and analysis
|
2016-2018
|
Participating and non-participating observations
|
Crisis and response chronological reconstitution and analysis
|
Sept 2017- Sept 2018 : Continental France and Saint-Martin
|
40 semi-directive interviews with public authorities, civil and internal security forces, network operators, business owners and NGO members.
|
1. Pre-Irma situation analysis 2. Crisis and response chronological reconstitution and analysis
|
2017-2019: Continental France, Guadeloupe and Saint-Martin
|
Focus groups with teenagers (six groups of six to eight persons)
|
1. Analysis of the teenagers’ perception of the water service quality 2. Analysis of the families’ strategies for accessing drinking water in normal times and in times of crisis
|
March 2019 : Saint-Martin
|
Questionnaire survey in two secondary schools in Saint-Martin (119 answers)
|
1. Evaluation of water facilities used by families in normal times and in times of crisis 2. Analysis of the teenagers’ perception of the water service quality 3. Analysis of the teenagers’ perception of the water crisis management
|
March 2019: Saint-Martin
|
Questionnaire survey for adults (33 answers)
|
1. Analysis of the inhabitants’ strategies for accessing drinking water in normal times 2. Analysis of the inhabitants’ perception of the water service quality 3. Reconstitution of the inhabitants’ response for accessing drinking water in times of crisis 3. Analysis of the inhabitants’ perception of the water crisis management
|
March 2019: Saint-Martin
|
39We used three types of data. The first are based on grey literature analysis, e. g. press archives (particularly the daily newspaper France Antilles Guadeloupe or the Souliga Post and Le Pélican, which are local newspapers) between 2017 and 2018. We also used participating and non-participating observations in various feedback reports carried out in the mainland in the aftermath of Irma. Based on these data, we established the crisis chronology and identified the responses implemented by the authorities, network operators and NGOs.
40In parallel, we conducted interviews with stakeholders involved in crisis management, including people at the top of the chain of command. Forty semi-directive interviews were conducted in mainland France, Guadeloupe and Saint-Martin. We interviewed managers at national and local level (public authorities, civil and internal security forces, network operators, NGOs). Some people were present in the field during or after Irma. Others, on the contrary, had never been to Saint Martin. We were also able to meet with various water and sanitation stakeholders, particularly agents of the Saint Martin water utilities and professionals from the water and sanitation sectors (employees working for the water service provider (the SAUR), desalination plant technicians, managers of companies specializing in boreholes and rainwater harvesting systems, etc.). The interviews lasted between one and four hours. They were recorded and the verbatim transcribed. They were cross-checked with various written and oral feedback. Some have not been disclosed to the public. We also conducted non-participating observations, such as an evaluation visit of potential beneficiaries of Red Cross assistance or the accompaniment of a Saur "listening agent" for a morning of meter reading.
- 1 Interviews conducted with secondary school pupils as part of a collaboration with several teachers (...)
41Finally, we carried out a field survey in Saint-Martin to understand the responses implemented before, during and after Irma, but also to grasp practices and behaviors related to drinking water and sanitation before the hurricane. We chose to target adolescents1, on the one hand, and adults, on the other hand, by working at the neighborhood scale.
42We therefore conducted focus groups with six “4eme” and “3eme" (8th and 9th grade) classes in the Quartier d'Orléans and Soualiga secondary schools. While the social and geographical origins of Soualiga secondary school students are relatively diverse, the population of Quartier d'Orléans secondary school is much more homogeneous, very representative of this priority district of the city, which is among the poorest on the island (Duvat 2008). We met four classes from Quartier d'Orléans and only two in Soualiga; this is why 54.6% of the students interviewed live in Quartier d'Orléans. On this occasion, we also distributed questionnaires to students, some of which focused on access to water and consumption practices. The responses to the 119 questionnaires collected were processed using the Sphinx software. The questionnaires were designed in French, which could have been a problem, as some children mainly spoke English Creole and had a poor command of French. Some questionnaires were therefore unusable. However, the questionnaires do identify a number of invariants. We also conducted participant observation days with children from the Madtowz Association (Sandy Ground Youth Association created in 2012), which confirmed and clarified the results of the focus groups and questionnaires.
43We also carried out interviews with adults: 33 people were interrogated about their experience of Hurricane Irma and the days that followed. Questionnaires were followed by informal exchanges that lasted up to 45 minutes and were not recorded but taken in notes. The sample is too small to be treated quantitatively, but these exchanges confirmed a set of conclusions on the practices and uses of drinking water and freshwater before and after Irma.
44On the morning of September 6th, 2017, people on the island discovered the devastation caused by Irma. Major destruction and debris of all kinds were hindering the progress of rescue teams and the relief efforts.
45The 72 hours following Irma were the most stressful for the populations, especially since they were warned of the arrival of Hurricane José. Victims were traumatized and many wanted to leave the island. This period corresponded to the most critical emergency phase. The insular nature of Saint-Martin and the great distance from mainland France, from where most of the emergency assistance originated, complicated and slowed down the organization of the relief effort: it took several days before the logistics chain was fully operational and the relief effort was deployed on the ground.
46National officials, who piloted the crisis management strategy from mainland France, correctly anticipated freshwater shortages risk. The interviews showed that the securing of drinking and sanitary water access was planned before the hurricane by the Inter-ministerial Crisis Management Operational Centre (COGIC), the gendarmerie and Veolia, with the pre-positioning of human and logistical means. These resources could be projected the day after the hurricane and be gradually increased. Based on their experience, emergency management NGOs (for example, the French Red Cross or the Veolia Foundation) had also anticipated the consequences of the hurricane's impact on drinking water access.
47However, when crisis managers, located in Paris, learned that the desalination plant was out of order, they interpreted the news as the disappearance of the only source of drinking water. One of them, who arrived in Saint-Martin the day after the cyclone, told us:
“Quite quickly, I was told there is no more water and there won't be any for several weeks', because there is no spring on the island. There is no drinking water or water that can be treated to become drinking water. It's only seawater or brackish water. And the desalination plant was broken by the hurricane and probably cannot be restarted, so a new one will have to be installed. (...) It's a story of several weeks, even several months. And so there, it is a very hard situation, everything is destroyed, there is no more water, and you are told that "you will have no more water” (interview February 2019).
48The distribution of drinking water and the rapid restoration of water service were identified as priorities for civil security. National and local authorities therefore sought to provide the victims with safe drinking water to meet their vital needs (20l/day/person). For example, a civil security official stated: “The issue of water distribution was a major one” (interview February 2019). This support mission was carried out by the civil and internal security forces and by NGOs (the Red Cross and the Veolia Foundation in the first place).
49As a sign of the special atmosphere that prevailed in Saint-Martin, civil security issues were linked to public security issues. In interviews, several respondents pointed out that for populations severely affected by the cyclone, fear of running out of water increased the risk of riots and looting. A “gendarme” explained to us:
“Given the intensity and nature and extent of the damage, they said to themselves, we're going to miss everything and very quickly. They did not imagine that behind it there would be a response of support for the population. They said to themselves, there is chaos and in chaos, well I recover everything that I can recover and that can have either an immediate value or a future value” (interview November 2018).
50This diagnosis explains the intervention of internal security forces, particularly “gendarmes”, in the distribution of drinking water.
51At the same time, drinking water production and distribution system restoration had begun, with the projection of human and material means to support local practitioners and agents, who were themselves affected.
52Water distribution mobilized all territorial levels and stakeholders, including law enforcement agencies, which do not normally carry out civil security missions.
53As early as September 6th, less than 24 hours after the cyclone hit, staff in the field (local authority services, police and gendarmerie, fire brigades) improvised the distribution of bottled water packs and food. The “gendarmes” then reported looting of water stocks in the gutted stores.
54Within 48 hours of Irma's passage, water distributions began. They were organized by the civil security forces (particularly the fire brigade) and the gendarmes present on the spot. Initially, preventive stocks built up by local authorities were mobilized, those of companies requisitioned. Distributions were made by people who had been affected by the hurricane and who weren’t familiar with the procedures. It sometimes created situations that shocked residents already traumatized by what they have just experienced. A “gendarme" explains:
“They didn't know how to do it. The way they did it, technically.... They would get to the top of a truck, a semi-trailer, and throw food and water like that. So, when I first saw it, I said stop” (interview October 2018).
55Security issues were omnipresent. For example, “gendarmes” evacuated water packs stored in containers or supermarkets to avoid “crowd fixing points” and to restore “feasible technical conditions” (interview November 2019). The fear of a riot was very high (gendarmerie interviews 2018 and 2019). The “gendarmes” therefore channelled the crowd at the supply points, secured and organized the distributions.
56Water distributions were definitively structured with the arrival of the first reinforcements and of the Red Cross, from September 9th. While drinking water was still imported, the sources were diversifying. For example, the Civil Security Department distributed water from the tanks of the desalination plant in the form of bags of several litres. This water, presented as drinkable, could be stored for three days in a closed bag and must have been consumed within 24 hours once the bag was opened. After this period, it could only be used as sanitary water (Prefecture of Guadeloupe 2017). There were fixed distribution stands but also tours: a truck circulated in the streets, volunteers could even knock on doors (Prefecture of Guadeloupe 2017).
57Sanitary water supplies were also set up. For example, the Red Cross set up 13 storage points in known, easily accessible locations chosen to cover the majority of districts. The equipment was available on September 18th; the tanks were assembled on site on September 19th and 20th. Initially, some reservoirs were filled with water points identified through BRGM's hydrogeological studies. Among these water points, for example, the fountain of the Pelican well in the centre of Marigot. The Red Cross also sought the collaboration of water trucking companies. This water was treated with chlorine in the truck or directly in the tanks. Self-service distribution ramps were then installed. Chlorine consumption and concentrations were monitored daily (Red Cross interviews, 2017).
58The restoration of drinking water distribution was based on four actions: desalination plant repair and production relaunch; focus on sanitation networks to avoid stagnant water pockets formation; underground distribution network repair; connections (tap water) rehabilitation. Several actors intervened, depending on their competences. In the case of the distribution network, the Veolia Foundation was a key player. Its members were volunteering employees of the Veolia Group, trained to put their water and sanitation skills to good use in emergency missions. During an interview, a Veolia Water France manager stressed the importance of distinguishing between the group's action, then in charge of the networks in Saint-Martin, and that of its Foundation, which intervened after Irma for an emergency mission: “Since the Foundation's ethic is not to intervene to help Veolia, it intervened under the Red Cross banner”.
59Veolia's action, whether it be for the company or the Foundation, was based on a strong capacity for anticipation, which was forged over the years, even if, according to the officials interviewed, the situation in Saint Martin was unprecedented and unseen. When Irma was announced on August 30th, 2017, the Foundation pre-positioned two experts in Guadeloupe so that they could intervene as quickly as possible in Saint-Martin, report the damage and draw up a list of equipment. The Foundation had also anticipated the destruction of the desalination plant and sought to identify mobile desalination units that could be mobilized and transported to Saint-Martin. Finally, the choice was made for a unit located in Madrid. It landed on the island on September 15th, after being transported by an army plane to Guadeloupe and then to Saint-Martin by boat. A few days earlier, a team of experts from the Veolia Foundation was specially dispatched to the site to welcome and set up this mobile unit. It was first assembled and then dismantled to protect it from the arrival of Hurricane José. It was finally put into service on September 24th.
60The Foundation then worked with Veolia's local teams. At the beginning of the crisis, about 20 people worked to rehabilitate the network, then once production was restarted, reinforcements from Veolia France arrived to repair underground connections and pipes. On September 22nd, 30% of the plant's production before the disaster could be delivered (Veolia Foundation 2017). On September 28th, the production of all desalination units (including the mobile desalination unit) exceeded 3.5 million litres per day. At the beginning of November, the plant was operating at full capacity and producing nearly five million litres per day (ibid.).
61Restoring water production was essential for consumption, but it was also a prerequisite for repairing leaks sector by sector. Whenever an additional area was opened, leaks could be identified and repaired. This would reduce the volume of losses and provide enough water to open a new sector. The work of locating leaks was done at night, "by ear". It was only at the end of this process that the defective individual connections were replaced.
62The distribution network reopening allowed the gradual closing of the sanitary water distribution ramps. As of December 21th, the reservoir at the Cul-de-Sac roundabout was removed. The last tank was dismantled on February 23rd in Quartier d'Orléans (EEASM 2018).
63Many inhabitants prepared for the hurricane by storing bottled water. Those who owned a tank also had it filled. An employee of a Saint Martin transport company explained that he had many deliveries in the weeks preceding Irma. Customers filled their tanks completely for the hurricane season and thus built up reserves of sanitary water in anticipation of network outages. Tanks actually played a decisive role. During the focus groups, several high school students reported that tanks provided water for their families for at least two months after Irma. If the water level decreased rapidly, the heavy rains that followed the hurricane allowed recharging.
64On the other hand, the inhabitants of the collective housing units were not able to benefit from the additional water provided by the tank. The health standards to be respected are so strict that many people see them as an outright ban. N*. explains, for example, that there is a very large tanker in his residence. “No one has ever used it, no one even knows where it is located exactly. After Irma, it would have been very useful”. Failing to find water there, N*. had to make a garbage bag tank on his balcony. It was because of it that he was able to take showers until the water system was restored. Red Cross volunteers stressed the importance of this kind of “System D”. It allowed people who did not have tanks to build up stocks of meteorite water: “Rainwater harvesting systems and water tanks had been installed. So actually, I had a bucket, I filled my shower with the bucket” (interview October 2018).
65Rainwater, whether from tanks or improvised means, was used for cooking, washing clothes and daily hygiene. It was used to clean homes and streets. Some pupils reported that their parents boiled the water to make it safe to drink: thus, they complemented the supply of bottled water from the distributions. The Red Cross also distributed leaflets to the population explaining water purification techniques.
66At the same time, seawater was used as domestic water. One respondent explained:
“I need to find water what, and I see a stream, but where does it go? This water, how could I get it? And then I realize that I am on an island and that I live 300 meters from the sea! We have huge water supplies! (he laughs)” (interview November 2018).
67These coping responses were part of a broader social context that mitigated the impacts and risks following Irma. The field survey confirmed the role of solidarity networks. Mutual aid was first provided within the family. For example, a pupil of the Soualiga College explained that he had fetched water from his aunt, who lived in the Dutch part of the country, for several weeks after the hurricane. Mutual assistance was also provided within the neighborhood, particularly in the most disadvantaged neighborhoods. Tourists also took initiatives to help the inhabitants:
“We had transformed the hotel into a food and water distribution centre, so when there were people who had nothing, we would say ‘you come to the hotel’ and we would make you a little bag, and then they would leave” (interview January 2019).
68Finally, local firms helped the island's inhabitants. The water transport company mentioned above, located in Concordia, organized free water distributions: 150 people came every day to fill their cans for almost a month.
69Most of the people interviewed explained that they did not feel suffering from thirst between the passing of the hurricane and the water supply system restoration. In this regard, the objectives set by national authorities in terms of water security and health security were achieved. A “gendarme” explains:
“The only health alert we had was when the restaurants reopened, the water started coming into the taps except it wasn't treated, and so the restaurateurs made the ice cubes... Well, I might as well tell you that I have 80% of the guys who were in the pile…” (interview October 2018).
70People seem to be generally satisfied with the organization of water distribution. They have sometimes provided opportunities for discussion, exchange and comfort. For instance, a pupil explained: “The Red Cross, they were kind, they gave us the bottles and asked us if everything was all right at home” (Soualiga secondary school, Saint-Martin, March 15th, 2019).
71On one hand, all respondents agreed that very large quantities of water were distributed and this, well after the distribution system was restored. This free water was particularly well received by the inhabitants. When the distribution ramps were dismantled (between December 2017 and February 2018), some people expressed strong dissatisfaction that they could no longer obtain free sanitary water. On the contrary, other people expressed a form of relief, a sign that the “Irma page was turning”. Others, who lived nearby, appreciated the disappearance of the disturbance that tanks could cause.
72On the other hand, it must be stressed that in the 48-72 hours following Irma, the situation was very precarious. A member of the internal security forces tells us:
“women with kids, all that, ... as soon as they see you, as soon as you meet someone, ‘give me water, give me water’. And on the 4 litres of your ration, you were giving. But I said, ‘No, I'm not giving you the bottle, I'm giving you a drink’” (October 2018).
73Until September 9th, water shortages were indeed a source of concern and constant anxiety, particularly for traumatized victims, who had lost everything, including the water stocks they had built in anticipation of the hurricane. A gendarmerie officer, present in Saint-Martin, explained: “It contributes to the high anxiety of the population. People think that if you run out of water, you'll die of thirst” (October 2018).
74During the first few weeks, the lack of water was also strongly felt by the staff involved in the relief and emergency management operations. For example, a gendarme recounted: “It is 35°, a terrible heat, debris everywhere... so people walk, there are no cars, because at the beginning everything is cut off. You have to walk, go to the contact, so you consume the water very quickly and there is no water…” (interview October 2018).
75Some also pointed to the lack of hygiene due to the lack of clean water. “They could no longer wash. The sea was totally polluted, there was nothing left” (interview November 2019). For many respondents, the first shower meant the end of the emergency phase.
76The interviews show finally that crisis management assessment varies greatly depending on the social category and the relationship to the territory. People belonging to wealthier categories, tourists or inhabitants of the island, have a rather negative judgement on the crisis management, at least for the first few days.
77There is an important gap between the discourse on water scarcity expressed at national level and taken over by the media, and the water access reality in the French part of Saint-Martin in the aftermath of Irma. It is therefore necessary to discuss the notion of a “water crisis”.
78First, there is confusion between drinking water, domestic water and freshwater. It is indeed important to assess shortage risk according to the type of water and the needs. Thus, the hospital's drinking water supply is a priority to maintain dialysis, for example. Restoring the plant's freshwater production is essential for repairing leaks, but drinking water is not required. This means that it is necessary to specify which element is vulnerable to which risk associated with a type of water. Water access issues cannot be addressed in a generic way.
79Second, we must distinguish shortage from reality and its perception. In mainland France, managers and the media perceptions were biased by distance, jet lag and the collapse of communication networks. On the spot, people's feelings were biased by the violence of the event they had just experienced and by the very special atmosphere that prevailed in the immediate post-hurricane period. In the interviews, respondents acknowledged that it is difficult to give precise dates and that the time immediately after Irma is distorted from reality (“hours that seemed to last for days”, “it seems like an eternity to me, but it is certainly much shorter”). A “gendarme” present on the island during Irma's passing explains:
“Frankly, the Red Cross, the fire brigade, at D+5, I didn't see any. Yeah, honestly. It seemed like an eternity to me.... The first gendarmes to come in as reinforcements were very late. (he stopped suddenly). Indeed, there are dates... It's true, the gendarmes, they arrive at D + 2... They're all here, aren't they? … But hey, you don't reorganize 700 gendarmes like that. There is time going on, and we, in the meantime, we're alone, so for us, it's true that it seems excessive”.
80So we are led to tighten up the temporality of the emergency period over the first 72 hours following Irma, which effectively corresponds to the time when the situation was very precarious and when the risk of a shortage of drinking water was real and high.
81But even within this period, there are significant differences according to population categories. According to testimonies, the first 72 hours were particularly critical for non-native wealthy residents or for tourists. However, over the days, they were able to mobilise their social, cultural and economic capital to implement effective coping strategies (including leaving the island).
82Contrarily, populations belonging to the least well-off social categories were relatively more resilient during the first days of the crisis. They seemed to be more proactive in clearing debris, cleaning streets, searching for food and water.
83To explain this difference, two additional explanations can be put forward. First, water shortage risk seems to be biased by the differences in standards for water access. A tourist explains: “You really count the water... And counting water, you're not used to it” (interview May 2019). On the contrary, people from the least well-off social categories are accustomed to lower standards.
84Secondly, inhabitants from disadvantaged social backgrounds had already adapted their consumption practices and habits to their difficulties in accessing public water services. They therefore felt its interruption much less strongly.
85Post-Irma analysis in Saint-Martin shows that local communities implemented proactive coping strategies to deal with the crisis situation. The coping responses are partly reactive adjustments: “system D”, family or neighborhood solidarity allowing them to take a shower, or wash clothes.
86But coping responses also reflect the local waterscape. Resilience is anchored in daily practices, local knowledge and solidarity, which were developed long before Irma and correspond to medium- and long-term adaptations to the water service crisis. This finding is consistent with the results already available in disaster studies. Individuals' resilience to major events is rooted in their resilience to everyday vulnerability. Poor, excluded, marginalized people constantly struggle to satisfy their elementary needs and must deal with the vulnerability in daily life or “ordinary vulnerability” (Mika 2019). It also confirms the role of historical, socio-economic and cultural legacies, as well as political decisions and coalitions of local actors in the access to and management of water resources, emphasized by water studies and political ecology.
87Alternative water access methods were therefore important factors of individuals and communities’ resilience, which greatly mitigated the consequences of the desalination plant shutdown and the distribution network collapse. To some extent, the so-called crisis was even an opportunity for some inhabitants, who were able to benefit from free water access.
88However, even these adaptation strategies are conditioned by economic and social capital. Feedbacks show, for example, the inequalities between those who own a vehicle, a well or a private tank (Red Cross 2017). The wealthier were even able to rely on private desalination units or on their swimming pools.
89Our study shows finally that the vulnerability of the inhabitants to water shortage risk was greatly reduced by an adaptation built well upstream of Irma, rooted in the local hydro-social cycle. It confirms the idea that networks are socio-technical systems, based not only on technical devices, but also on stakeholder relationships, regulation, practices, values and trade-offs between different interests (Jaglin and Zérah 2010). These socio-technical systems are themselves “embedded” in a set of production and consumption practices, just as they are anchored in a given geographical location (Jaglin 2014). In this sense, water systems, and beyond them, social resilience, are anchored in overarching contexts and must be considered through place-based and community-based approaches.
90First, our study shows that the idea of a water crisis in Saint Martin needs to be reconsidered. While the threat was real, the crisis was did not occur through the joint action of governments, local authorities, network managers and NGOs. The hazard - here, water scarcity - was greatly mitigated by the response implemented by the various actors, including inhabitants and local communities. They were able to implement effective and efficient coping strategies. The combination of institutional and inhabitants’ responses prevented a humanitarian disaster.
91Secondly, acknowledging that drinking water shortage risk was overestimated does not mean that individuals have not been in critical situations and that they cannot describe their own experience as a crisis. However, the individual perception must be contextualized relatively to the trauma and to the situation of disorganization that prevails on the island immediately after Irma. There is a need to consider the water access norms and standards according to socio-economic categories and lifestyles. Our study therefore stresses the relevance of discussing the notion of crisis in terms of people's perceptions and experiences, and of the relationship to norms and the temporalities of the post-disaster.
92Third, we emphasize the importance of living practices related to the island hydro-social system and the failures of the public water service. It is impossible to understand the resilience of local populations without examining the latent crisis situation of the public water service in Saint-Martin that existed before Irma. This finding challenges the normative use of resilience and encourages going beyond discourses that only highlight its positive dimensions.
93We are aware that the methodology employed is not intended to cover the entire population of Saint-Martin and that not-responding or not-included people due to the social fragmentation on the island are missed by the study. Moreover, the results obtained are only valid for the French part of Saint-Martin and their generalization to other Caribbean islands would require many additional studies. However, the results confirm the conclusions of other studies that highlight the need to examine what is behind resilience, its drivers and root causes. They also challenge the preconceived ideas of some decision makers and managers that overvalue technical response and underestimate both people's capacities and the role of overarching contexts.