Community Resilience refers to the capacity of systems to cope and adapt after extreme events in such a way that they maintain their basic structure, function, and identity [1
]. Resilience also refers to a set of processes that link and guide networked capacities to assure adaptation after disturbance [3
]. According to Cutter [1
] the measurement and identification of standards and metrics to measure Community Resilience is a challenge for governments at all levels. This is further complicated because Community Resilience is determined by the inherent characteristics of the system and the resources it possesses. Despite this, Community Resilience “…is becoming [the] de facto framework for enhancing community level disaster preparedness, response and recovery…” [4
] (p. 65). Clearly, resilience in the human environment refers to how communities efficiently prepare, plan, and adapt in the case of extreme events; hence, achieving Community Resilience implies knowing the characteristics of the territory that affect resilience, before, during and after a disturbance and taking the necessary actions to improve resilience before another catastrophic event.
One characteristic that influences Community Resilience is Disaster Governance. Effective Disaster Governance produces resilience [5
]. Tierney [6
] (p. 344) explains that Disaster Governance consists of “the interrelated set of norms, organizational and institutional actor(s), and practices (…) that are designed to reduce the impacts and losses associated with disasters”. Although governance has been traditionally fragmented, nowadays it encourages collective actions among all stakeholders with interest in the matter of concern. Since hazards and disasters are complex events that affect communities at all dimensions (e.g., physical, social, economic, etc.), extensive cross-national and cross-institutional collaborations are required and should be incorporated in efficient Disaster Governance [6
]. Indeed, a content and cluster analysis of 1069 peer-reviewed journals [7
] indicates that the characteristics that influence governance are: stakeholders’ involvement and commitment; cooperation and collaboration at a variety of scales (e.g., distribution of government functions); and flexibility, which, for example, allows for the adjustment of policies and regulations among others in case of an emergency.
Disaster Governance is also political. There is extensive literature exploring the impact of government actions on mortality rate after earthquake events [8
]. A democratic state has been found to mitigate the effects of disasters on populations by, for example, decreasing disaster mortality rates [8
]. Expected mortality rates decrease more rapidly when democracies are older and more established and when there is a higher per capita income [9
]. Despite this, these finding may not hold for corrupt governments where, for example, a corrupt public sector can lead to the construction of buildings that collapse in major earthquakes [10
]. Additionally, inequality in per capita income can also influence the number of fatalities resulting from earthquakes [11
]. Inequalities can hamper the collective actions that can ensure the creation and enforcement of building codes, the retrofit of built structures, and the definition of quake-sensitive zones. These three aspects together can considerably reduce fatalities. Besides, a number of studies highlight the influence of government responsiveness in the reduction of mortality resulting from earthquakes. Indeed, countries with higher newspaper circulation and greater electoral accountability have been found to be more effective in calamity relief [13
]. From this, it is seen that informed citizens can influence governments’ responses, and in these types of communities citizen preference is reflected in disaster policies. Overall, well-equipped government institutions can lower fatalities from disasters by providing effective regulations and planning and by improving the quality of infrastructure and emergency areas [12
These studies indicate that Disaster Governance is particularly important in developing countries [14
] because it can particularly affect the needs of poor communities. Aspects such as disparities in income, well-being, access to services, and political empowerment [6
], as well as lack of resources, poor diversity of stakeholders, and missing links among established networks [7
] can hamper the effectiveness of Disaster Governance. These disadvantages, in turn, encourage responses to disasters to focus on coordination rather than on risk reduction [15
]. These shortcomings reduce the effectiveness of institutions and deteriorate the capacity to coordinate [15
]. Particularly for developing countries, better Disaster Governance has been found when emergency institutions and the civil society interact. The Cuban response to Hurricane Sandy in 2012 [16
] is an example of such coordination between the local community and government and non-government institutions, which resulted in a high degree of Community Resilience. The Civil Defense’s ability to act together with the inhabitants of the city and the rest of Cuba prevented greater loss of human lives. This example also sheds light on the importance of integrating extreme weather and disaster scenarios into regional planning in order to facilitate governance after disasters. These types of actions and responses are more in line with the concept of governability that has gained interest over the last several years, particularly in Latin American countries. Governability refers to a state of dynamic equilibrium with a multidimensional character and emphasis on the construction of a legitimate and effective response between the market, state/government, and the demands of the civil society [17
Disaster Governance in Urban Environments
Although Disaster Governance includes a broad spectrum of variables in its definition, the focus of this article is the study Disaster Governance through the spatial analysis of different human settlements structures in the event of a tsunami, to further our understanding of Community Resilience. In this respect, specific studies suggest that Disaster Governance in human environments can be characterized by three attributes: redundancy, overlap, and diversity [2
]. Firstly, the redundancy of the governance structure [2
] indicates that if one organization fails to provide emergency support to a struggling community, another organization similar in nature may take its role. For instance, if hospitals are damaged after a disaster, institutions such as the Red Cross and local clinics can continue providing health services to the community. Secondly, overlap in governance refers to the set of linked capacities that allow institutions to be efficient and flexible [18
]. Overlap in governance develops when different types of institutions (e.g., police, firefighters, and health services among others) have the capacity to perform the same function at different scales when disaster arises [19
]. For example, if hospitals are damaged, other institutions, such as firefighters, should be prepared to provide health assistance until the primary health support has recovered. Thirdly, diversity in governance refers to the diversity of institutions prepared to perform in the same areas, regardless of their role. For instance, a community with overlap and redundancy in governance may have a more diverse response to disaster [21
]; thus, disadvantaged and threatened societies can have improved Community Resilience.
Various studies support the idea that specific aspects of urban structure and space affect the attributes of Disaster Governance. Communities with polycentric urban structures show more overlap and redundancy in governance [19
]. Specifically, urban areas with various central areas often have similar neighborhoods, mixed-use areas, and modular systems [19
]. For instance, in polycentric urban areas, several fire departments are distributed within different neighborhoods; therefore, if one fails, another can fill its role. Additionally, more intense overlap has been found in the central area of cities, and this overlap diminishes toward the periphery and rural areas [22
]. As a result of being centrally located or due to infrastructure that facilitates mobility (e.g., properly marked paved streets), central urban areas are usually better equipped than rural and peripheral areas thus facilitating rapid and effective emergency responses. In contrast, peripheral and rural areas, which are sometimes harder to reach, often receive less assistance during an emergency. Furthermore, diversity of governance can be improved by attributes of the urban space. A diverse set of green areas with different sizes and spatial arrangements provides more opportunities for different emergency institutions to fulfill their roles after a disaster (e.g., for allocating temporary hospitals, refuges, etc.) [23
]. The multifunctional use that a diversity of green areas provides contributes to the diversity of governance by providing flexibility in the urban space [24
]. Diversity has also been evaluated by measuring urban density, a variable that at the same time influences the abundance/scarcity of useful open space [25
]. In this case, increased urban density is associated with less diversity because there is less open space available for emergency purposes.
This study is aimed at furthering our understanding of Community Resilience through the lens of Disaster Governance in coastal environments. To do so, we have explored how different human settlement structures influence the performance of emergency institutions after disaster. This study was undertaken in four coastal towns in southern Chile that are threatened by tsunamis. The four towns selected have different spatial and planning arrangements. The objective was to evaluate, by means of different spatial indexes, the extent that spatial aspects of towns impact the attributes of Disaster Governance, namely redundancy, overlap, and diversity. The outcomes of this study can provide guidelines for improving Community Resilience through planning.
2. Materials and Methods
This study employed a comparative approach including three steps. First, it was necessary to identify where emergency institutions go and how they use urban spaces in the case of a tsunami emergency. For this purpose, experts in emergency procedures from different institutions were interviewed using the Projective Mapping Technique. From interviews, we identified sites used for emergency purposes. Secondly, the data were subjected to content and GIS analysis to explore how institutions are distributed and concentrated. From this, we identified the redundancy, overlap, and diversity of the institutions. Finally, the results were projected on maps and were contrasted with aspects of the spatial and planning structure of the towns, i.e., density, sector types (urban, rural or indigenous), level of government within the commune, and emergency plans, to explore their influence on disaster governance.
2.1. Towns of Study
The study areas included four towns distributed within three communes and two regions in southern Chile (Figure 1
). While Mehuín, Queule and La Barra represent the lowest units of government, Puerto Saavedra represents a higher unit, as it is the capital of the Saavedra Commune. All towns have a monocentric structure with urban, rural, and indigenous sectors. Additionally, all towns have tsunami emergency plans. Mehuín is a small coastal town with a population of 1504 inhabitants, and it is located in the Mariquina Commune in the Los Ríos Region. Mehuín faces the sea, lies at the mouth of the Lingue River, and includes the urban sectors of M. Caleta and Balneario, the rural sectors of Pichicuyin and Mississippi, and the indigenous sector of M. Bajo. Queule is also a small coastal town with a population of 1541 inhabitants located in the Toltén Commune in the La Araucanía Region. It is located along both sides and at the mouth of the Boldo River and includes the urban sectors of Q. Corvi, Q. Caleta, and Q. Portal and the indigenous sector of Los Pinos. The town of La Barra is a small fishing town of 165 inhabitants located at the mouth of the Toltén River in the Toltén Commune in the La Araucanía Region. La Barra includes the urban sector of LB. Caleta by the coast and the indigenous sector of Mapuches located inland. Finally, Puerto Saavedra is a large coastal town of 5011 inhabitants located in the Saavedra Commune. It is the capital city of the commune and is separated from the sea by the Imperial lagoon. Puerto Saavedra includes the urban sectors of PS. Bajo and PS. Corvi, the rural sector of La Playa, and the indigenous sector of Budi.
2.2. Data Collection and Participants
Participants from government and non-governmental institutions who are involved in tsunami emergency procedures were invited to participate in an interview. The participants contacted were identified based on the information provided by the local Emergency Operation Committees (COE) of each commune; the COE is in charge of the management of disaster situations. Institutions are distributed in each town as can be observed in Figure 2
. In Mehuín, a total of 25 professionals from four government institutions and three non-government institutions were interviewed. In Queule, a total of 23 participants from four government institutions and five non-governmental institutions were interviewed. In La Barra, a total of 15 participants from two government institutions and four non-government institutions were interviewed. Lastly, in Puerto Saavedra, a total of 35 participants from five government institutions and five non-governmental institutions were interviewed. Overall, a total of 32 institutions and 98 emergency professionals (75.96% of the population) were interviewed. The role of each participant during an emergency is described in Table 1
, including the coordination of the emergency response, evaluation of damages, evacuation, medical care, and the distribution of supplies, among others. Using the Projective Mapping Technique, participants were interviewed in their own workspaces. The interviews consisted of showing participants a map of the town under study and asking them to identify the places they would use during the emergency stage of a tsunami to provide professional help to the community. Participants were also asked to mention the utility of the places (e.g., for refuge, for water collection, for temporary health service). With this information it was possible to identify the frequency of mention of sites and the site uses for emergency purposes in the event of a tsunami.
2.3. Content Analysis
Content analysis [27
] was used to reduce the number of uses assigned by participants to the places used for emergency purposes to a manageable number. The analysis involved the following steps to assure validity. First, uses with the same meaning were grouped together (e.g., the terms “for water collection” and “for water extraction” were grouped under “water supplies”). Second, the outcome of the first step was group together by theme (e.g., “water supplies” and “food supplies” were grouped under “supplies”). Accordingly, the themes represent the final number of uses assigned by participants to the sites used for emergency purposes. This procedure was performed twice and by two researchers separately to assure that the analysis was reliable and trustworthy [28
2.4. GIS Analysis and Spatial Index
Frequency data were organized in different matrices (m × n) prepared to code the mention of sites (ID × n) and frequency of mention of uses (ID × Use). In the first case, the number 1 was used when a site was mentioned by each organization; otherwise, the number 0 was assigned. In this manner, four binary matrices were created, one for each town, and were subjected to the Directional Distribution Index. The Directional Distribution Index was used to evaluate the territory that is under the jurisdiction of each institution; this index was used to determine the redundancy and overlap of each institution in the event of a tsunami emergency. This technique was originally developed for the ecological study of animal habitats and consists of obtaining a bivariate confidence interval for X and Y coordinates. Once the confidence intervals were determined, the standard distances (Euclidian) between the X and Y locations of the sites were calculated. These distances were then mapped to show the ellipsoid area of the sites mentioned by each emergency institution. The X and Y coordinates define the major and minor axes of the ellipse with the smallest possible area. Each standard deviational ellipse is a summary of the data dispersion of a point structure. The size of the ellipses depends on the standard deviation. In this case, a standard deviation of 1 covered 68% of the places mentioned. Results can be interpreted as follows: increased redundancy is shown by more ellipses representing institutions with similar roles and crossing in the same sector. Likewise, there is more overlap in governance if there are more ellipses representing different institutions prepared to assume similar roles in the same sectors.
The frequency of mention of sites was recorded for each use associated to each place mentioned by the participants. Accordingly, four frequency matrices were obtained for each town. These were then subjected to the Kernel Density Index. The Kernel Density Index was used to identify the concentration of site uses and the most used sites in the towns. This was used, hence, to shed light on the multifunctionality, or diversity, of the sites. Kernel Density refers to how a phenomenon is spread across a landscape based on the quantity of a variable, for example evacuation, which is measured at each point. This analysis is based on the quadratic kernel function and can be used to explore spatial variation in event intensity [29
]. Specifically, this index was used to calculate the uses associated with sites and their surrounding area. Thus, the analysis showed the density of 1 km2
sites mentioned by institutions; these were represented by five ranges for the four towns. The pixel size used was 25 × 25 m, including a spherical surface and a radius of 200 to 500 m. The results are illustrated in maps, which indicate the urban nodes with more intensity of use. The Kernel Density values are differentiated according to the shade and intensity of a particular color. Accordingly, the resulting kernel raster values are greater when the color is more intense, indicating that more points are located in the area. Overall, higher raster values reflect more diversity in governance, suggesting that more institutions are using the space for emergency purposes.
The study of Disaster Governance undertaken here shows that Disaster Governance is highly influenced by decisions taken during regional, urban, and emergency planning. Here it was found that towns with higher hierarchies within communal systems have better Disaster Governance. One reason for this is that most of the emergency institutions studied here are located in central and urban areas, which, in turn, have more redundancy, overlap, and diversity in governance should a tsunami occur. In contrast, emergency plans add to the limitation of governance in the towns studied. Although the emergency plans include both governmental and non-governmental institutions which altogether contribute to the diversity and redundancy of an emergency response, the roles allocated to emergency institutions are not diverse enough, neither is the distribution of safety and evacuation zones within the towns. Accordingly, and since the emergency institutions have to follow the emergency procedures established in emergency plans, flexibility-known as one of the characteristics that influences Disaster Governance [7
]—cannot be achieved during an emergency.
From our findings, it is seen that Disaster Governance is improved when governmental and non-governmental institutions interact. Although the government was found to be an important institution during an emergency, it was also found that the government is insufficient by itself. Hence, governability can improve Disaster Governance by encouraging governmental and non-government institutions as well as society to interact. In support of this idea, other studies have highlighted the need for an approach that includes the interrelationships of norms, institutions, actors, and practices during an emergency (e.g., MSP approach discussed in the discussion section) [4
]. In other words, governability calls for collective action not only among actors from different institutions, but also among actors and planning instruments. Further studies should include the role of the local community in responding to a disaster and explore the extent to which redundancy, overlap, and diversity in governance are improved.
Overall, the spatial relationships found between different town structures and the performance of government and non-government emergency institutions in the event of a disaster show that urban areas have better Disaster Governance than rural and indigenous areas. However, the availability of resources remaining after disturbance in the study areas, influence the spatial findings obtained in this study. Figure 1
shows that urban areas, unlike rural and indigenous areas, are better equipped with emergency facilities and infrastructure that host economic activities, yet most urban areas are located in tsunami inundation zones. This suggests that the effect of a tsunami would be high, disturbing the response by emergency institutions and consequently Disaster Governance. For example, in Puerto Saavedra, the emergency infrastructure including schools, fire stations, police departments, and the main area of commerce are below the inundation level (Figure 1
). In contrast, rural and indigenous areas have less emergency infrastructure, yet the community headquarters and natural resources in these areas are located above the tsunami inundation zone. Thus, this could facilitate the community’s ability to adapt from tsunamis and, hence, Community Resilience. Additionally, in the indigenous area of M. Bajo, there are social headquarters, a school, water, forest and scrubland above the inundation area. Thus, where people could congregate obtain water, food, and firewood in the event of a tsunami. As such, in certain situations, rural and indigenous areas, versus urban areas, show better Disaster Governance and could be more likely to adapt from a tsunami.
Another aspect that influences our findings is that the distribution and allocation of resources during an emergency is not balanced. In the case studies, communal governments are the main bodies responsible for making decisions and making resources available to the community during a disaster. Specifically, the Emergency Operations Committee (COE) at the communal level makes most of the decisions during an emergency. The Municipality is part of the COE and mainly focuses its actions in urban areas through the Land Use Plan. As the Land Use Plan is only relevant to urban areas, the Disaster Governance of rural and indigenous areas is negatively impacted. As a result, rural and indigenous areas are disadvantaged compared to urban areas. Giving priority to intercommunal planning instruments, which have the capacity to regulate urban, rural and indigenous areas as a whole, could improve Disaster Governance in the study areas.
The political aspects of Disaster Governance [8
] also influence our findings. A recent study has evaluated the physical and social causes of mortality resulting from a tsunami, finding that social aspects can be negatively affected by political decision taken in the territory [33
]. The study highlights the fact that knowing the stock of social capital can influence mortality rates during a tsunami. In particular, mortality can be affected by the merging or creation of new localities. This information is relevant for the case studies because the Chilean coast is a highly dynamic environment where rural areas are often absorbed by urban zones. Indeed, in Mehuín and Puerto Saavedra future land use plans are proposed to include rural as well as urban areas. According to Aldrich and Sawada [33
], the merging of rural and urban areas is a political condition that can diminish social capital because it causes people to lack identity with their new locality and representation in the political arena. This is why the authors recommend allocating funds to build the social ties among people and not only to recover physical infrastructure after a disaster. Such a recommendation can be of great value for improving Disaster Governance in the case studies presented here.
The findings of this study provide insight regarding the Disaster Management cycle, particularly with respect to the preparedness, response, and recovery stages. First, during the preparedness stage a more flexible evacuation plan should be developed for the case studies that includes all units of government (urban, rural, and indigenous). Particularly in the rural and indigenous areas, the emergency plan should diversify the emergency institutions involved during a disaster. This would assure the action of emergency institutions at lower governmental levels, which would provide efficient responses to all territorial scales. Training during the preparedness stage could be used to evaluate the equity in the quality of the response of existing emergency plans. Second, during the response stage, more fluidity in the assistance of institutions to all units of government could be achieved by diversifying the role of institutions. Specifically, members of institutions should be trained so that they can take on different roles in the event that other institutions fail. In this way, the redundancy of governance will be improved. Additionally, social institutions should be prepared in order to enhance social capital. This is particularly important in rural and indigenous areas where the action of emergency institutions is limited. Finally, during the recovery stage the allocation of funding and applications of grants should be adequately guided as to enhance Disaster Governance. Here we have found that Disaster Governance of urban, rural and indigenous areas differs. Hence, funding should be used to improve Disaster Governance in the areas and aspects that is particularly needed. This may be hard to achieve due to the political influence which has historically affected the distribution of funding [11
]; however, it is necessary if Disaster Governance and hence, Community Resilience, are to be improved. In this sense, the division of the territory into urban, rural, and indigenous areas can hinder Governance. The increased population densities of urban areas greatly influence political will and thus the allocation of funding and the creation of laws. Regardless, the implementation of transitions zones in disaster prone areas can be rapidly implemented to improve disaster response. From this, physical and human infrastructure can be used to provide an equal, flexible and fluid response regardless of disparities in income, access to services, and political power.
Today, the relationship between government and society is not determined by the government alone. Disaster Governance is two-way where society and citizenship assume a higher level of responsibility and interaction. Disaster Governance is also political. The effectiveness of emergency actions for the adequate recovery and protection of communities depends on the bonds of society as well as on the state’s capacity and willingness to provide aid. This is particularly relevant in disaster scenarios where rapid as well as efficient responses are required to assure that communities are resilient: this can only be achieved by including all visions in a dynamic and flexible manner.