EXPLORING SOCIAL RESILIENCE IN MARINE PROTECTED AREAS â•fi A CASE OF INDONESIAâ•ŸS CORAL TRIANGLE

The term ―resilience‖ has a very long and rich history. The term itself received a widespread attention since Holling‘s seminal paper in 1973 on system ecology. Since then, the term resilience has been widely used and defined in many academic disciplines. The examination of social resilience in the context of overall coastal community resilience has been developed during the last few years. Such studies have been important in determining factors influencing the acceptance of MPAs in resource dependent communities. The concept of social resilience has been defined mostly at the community level, and less so at the individual level. In order to fill the gap, this study is intended to measure social resilience at the individual level. The objective of this study is to explore resilience and its impact on Indonesian MPAs. It addresses the following research questions: (1) What is the degree of variability in individual resilience in Indonesia‘s Coral Triangle?, (2) Are there any relationships between degree of individual resilience and other social characteristics of a community?, (3) Are there any relationships between degree of individual resilience and a community‘ economic characteristics? (4) Are there any relationships between degree of individual resilience and community members‘ environmental attitudes beliefs and values?, and (5) How does community perception of MPA management influence their degree of individual resilience? This study has discovered some important aspects of social resiliency and it‘s relation to some aspects of MPAs. First, the social resilience of resource dependent individuals in Indonesia could be best explained by five components, which are: adaptive capacity, risk awareness, perceived social-economic status, community attachment and environmental awareness. Second, this study suggests that MPAs have some degree of influence on the level of individual social resilience. Several resource users‘ individual attributes, such as age, years of education and gender are related to their level of social resilience. Resource user‘s perceptions of some aspects of MPA planning and management processes were also found related to their social resilience. Finally, this study provides a basis for further in depth research of social resilience of resource dependent communities, specifically in the Indonesian context.


INTRODUCTION
Indonesia is the world's largest archipelagic state. It has a very complex geology, climate and ocean circulation patterns, which result in a highly diverse and dynamic marine and coastal environment (Tomascik et al. 1997). The population of Indonesia is approximately 240 million (in 2010), and nearly 60 million people live along the coast within 30 km from coral reefs (Burke et al. 2012). In order to optimize the benefits of marine and coastal resources, the government of Indonesia has rapidly expanded the extent of marine waters under protection. To date, approximately 170,000 sq. km of Indonesia's marine and coastal area has been protected with some form of marine conservation arrangement. The government of Indonesia is currently continuing to establish more conservation areas to fulfill the 200,000 sq km commitment by 2020 to the Coral Triangle Initiative -CTI (Green et al. 2012;Carter et al. 2010).
Marine Protected Areas (MPAs) have a significant role to play in the protection of ecosystems and, often, in the enhancement or restoration of coastal and marine fisheries, if they are correctly designed and effectively managed (Carter et al. 2010; IUCN-WCPA, 2008). MPAs consist of a complex combination of governing arrangements managing the interactions of humans with the natural environment (Dalton 2012). However, MPAs' implementation can cause major changes to an individual's (i.e. resource users) life and coastal communities' interaction as the result of restricting resource utilization, for protection and conservation. The coastal communities will have to be able to adapt to such changes. Their adaptation involves making adjustments to changing circumstances in order to endure the changes (Hanna, 2000).
The theory of resilience has been undergoing development for about four decades (Holling 1973;. Resilience refers to a system that maintains socialecological functions, with the ability to absorb change or perturbation and reorganize so as to maintain essentially the same function, structure, identity and feedbacks (Marshall, 2006). Resilience is the ability of a social-ecological system to cope with and adapt to external social, political, or environmental disturbances (Adger 2000, Folke et al. 2002a, Cinner et al. 2009). During, the last ten years, efforts to apply the resilience concept to marine conservation have significantly increased (Hughes et al. 2005, Cinner et al. 2009, Marshall et al. 2009, Sutton and Tobing 2012, Cinner et al. 2012).
Social resilience, as one of the essential components of resilience theory, has been developed in the context of anthropological and medical research (Vayda and McCay 1975;Rutter 1987;Abel and Stepp 2003;Bonanno, G.A. 2004). The examination of social resilience in the context of overall coastal community resilience has been developed during the last few years. Such studies have been important in determining factors influencing the acceptance of MPAs in resource dependent communities (Marshal 2007;Cinner et al. 2009;Marshall et al. 2009;McClanahan et al. 2012;Sutton and Tobing 2012).
The concept of social resilience has been defined mostly at the community level (Levin et al. 1998, Adger 2000, McClanahan et al. 2008, Cinner et al. 2009), and less so at the individual level , Marshall et al. 2009, Sutton and Tobing 2012. In order to fill the gap, this study is intended to measure social resilience at the individual level. Moreover, for the purpose of this study, general use of the term ‗resilience' refers to individual resilience-the adaptability of individual resource users to changes and perturbations in their community and ecosystem, while community resilience is the degree to which all community members are resilient.
The objective of this study is to explore resilience and its impact on Indonesian MPAs. It will address the following research questions: The next chapter provides a summary of current theory regarding resilience and social resilience, including social and economic characteristics that have been found to influence social resilience, and its potential influence on MPA management.
Chapter 3 provides a description of the methods used for data collection and analysis.
Chapter 4 presents the results. Chapter 5 discusses key findings, management implications and potential areas for improvement. Chapter 6 presents the study's conclusion.

RESILIENCE CONCEPT REVISITED
Resilience theory has been developed over the last few decades. The resilience perspectives surfaced over a theory of ecological stability resulting from studies of population interaction related to the predator--prey mechanisms in the field of ecology (Folke 2006). C.S. Holling (1973) initially utilizes the concept of ‗resilience' in ecology in his seminal paper. -The resilience approach emphasizes non-linear dynamics, thresholds, uncertainty and surprise, how periods of gradual change interplay with periods of rapid change and how such dynamics interact across temporal and spatial scales‖ (Folke 2006: 253).
The resilience perspective is constantly evolving and used in a great variety of interdisciplinary works concerned with the interaction between humans and nature , Folke 2006. The concept and associated theory began to influence other fields such as anthropology and other social sciences (Vayda and McCay 1975, McCay 1978, Thompson et al. 1990, Hanna et al. 1996, Scoones 1999, Abel and Stepp 2003, ecological economics (Perrings et al. 1992, Costanza et al. 1993, Arrow et al. 1995, community planning (Lamson 1986, King 1997, disaster and hazard (Tobin 1999), geography (Zimmerer 1994), and public health (Dyer andMcGuinness 1996, Rutter 1987).
Based on the original concept from Holling's synthesis (1973), resilience has three defining characteristics in a social-ecological system, which are; (1) the extent of change (or stress) that a system can undergo (or sustain) and still maintain the same controls on its structure and function, (2) the degree to which the system is capable of self-organization, and (3) the degree to which the system can build and increase the capacity for learning and adaptation , Folke et al. 2002a. The sequence of resilience concept development is summarized in   (Folke et al. 2010: 2). The stability dynamic of a linked systems of human and nature emerges from three complementary and interrelated attributes: (1) resilience, (2) adaptability and, (3) transformability which could determine the system's future trajectories . Henceforth, Folke et al. (2010) argue that both adaptability and transformability are the prerequisite attributes for socialecological resilience. In addition,  emphasize four crucial aspects to define resilience in the context of social-ecological systems: (1) latitude, (2) resistance, (3) precariousness and (4) panarchy.

ADAPTABILITY AND ADAPTIVE CYCLE
In the social-ecological system, adaptability refers to the extent of humans' (actors') capacity to influence resilience, intentionally or unintentionally ). The adaptability of the actors decides the level of threshold in a socialecological system (move closer/further away or more/less difficult to reach) ). Moreover,  imply that a desirable regime in the socialecological system can be created from intentional collective actions of the actors (human) to manage the resilience following a disturbance. Berkes et al. 2003(as cited in Folke et al. 2010:2) further explained, -adaptability captures the capacity of a social-ecological system to learn, combine experience and knowledge, adjust its responses to changing internal processes and external drivers, and continue to develop within the current stability domain or basin of attraction‖. Hollings et al. (1986;2001) presented a heuristic model for understanding the process of change in complex systems, called the adaptive cycle (Fig 2.1). It consists of four cyclic development phases and three characteristics, which -can be used to identify structure, patterns, and causality in a complex adaptive system,‖ (Allison and Hobbs 2004:4). Four development phases of adaptive cycles are rapid growth/ exploitation ( r ), conservation ( K ), release ( Ω ), and reorganization ( α ) (Hollings et al. 1986;2001) and the three characteristics are potential (capacity), connectedness and resilience (Allison and Hobbs 2004). The process involves an adaptive cycle triggered by a disturbance (changeevent) that breaks down the system. The cycle then moves to the next phase of growth or exploitation. During this phase, new opportunities and innovations that could shape the system arise (Marshall 2006). The cycle then continues to the conservation phase.
In this phase, any external disturbance may not significantly affect the system as the system becomes stagnant and less flexible (Marshall 2006). If an external disturbance happens that exceeded the system -threshold‖, the system would collapse and enter the release phase of the cycle. The system would then be restructured and regrown (Holling 1973;, Gunderson et al. 1995, Marshall 2006). conditions make the existing system untenable‖. It can be a deliberate or forced process by the actors . Several studies of social-ecological systems suggest that transformation attributes entail four stages; (1) preparing the socialecological systems for change, (2) a crisis that creates a window of opportunity for change occurring, (3) navigating the transition of the system and (4) charting a new direction of the social-ecological system, while building resilience for the new regime (Olsson et al. 2004a;).

THRESHOLD
One important factor in the resilience of a social-ecological system is threshold. Thresholds are used to describe the point where a regime or an alternate stable state in a system could be changed into another regime or stable state (Walker and Meyers 2004). They further explain that in theory, when a threshold level is passed, a regime shift occurs, and as a result, the nature and extent of feedback in the system changes. In a Socio-ecological system there exists thresholds (from primary components) that could determine the trajectory of the system from a desirable into an undesirable state, if it is passed (Walker and Meyers 2004). Marshall (2006:16) explained that an adequately big change event could result a switch in the system to an alternate regime if -the thresholds of coping are reached and exceeded‖. She further argued -A negative shift from ‗desirable' to ‗undesirable' states represents loss of system resilience‖ (Marshall 2006:16).
Social-ecological systems have multiple interacting thresholds that are triggered by slow and fast variables (Yorque et al. 2002, Walker et al. 2006, Renaud et al. 2010. Threshold measurement is difficult and typically has low precision; very often thresholds shift over time due to the dynamic and the complexity of the systems (Walker and Meyers 2004, Walker et al. 2006, Marshall 2006, Renaud et al. 2010).

LATITUDE
Latitude (L) refers to -the maximum amount from a system that can be changed before losing its ability to recover‖ (Walker 2005:82). It is illustrated as the width of the valley of attraction ( Fig. 2.2) . Furthermore, Walker et al. (2004:6) suggested that wide valleys -mean a greater number of system states can be experienced without crossing a threshold.‖ 2.1.6 RESISTANCE Resistance (R) suggests the level of difficulties in changing the system (Walker 2005). It is -related to the typology of the basin-deep basin of attraction; (R; or more accurately, higher ratio of R:L) which indicates that greater forces of perturbation are required to change the current state of the system away from the attractor‖ (Walker et al. 2004: 6-7). Figure 2.2 pictured Resistance as the depth of the valley. As the valley become deeper, a greater disturbance is needed in order to move a system closer to its threshold and into another alternate state or regime (Marshall 2006).

PRECARIOUSNESS
Precariousness (PR) indicates the current trajectories of a system to its thresholds , Walker 2005. It is pictured as the distance of the dot relative to the edge of the valley ( Fig. 2.2).

PANARCHY
Panarchy is the theory of the cross scale, interdisciplinary and dynamic nature of a socialecological system , Gotts 2007. It is how the latitude, resistance and precariousness are -influenced by the states and the dynamics of the systems at scales above and below the scales of interest‖ (Walker et al. 2004:7).

SOCIAL RESILIENCE
It has been understood that the resilience of the social system linked to a larger resource system is just as important as resilience of the ecological components of the system (Berkes and Folke 1998. Resilience is mostly specified within the context: ‗of what, to what' . However, researchers and managers are mostly unclear about what they have set out to measure for social resilience (Marshall 2006).
In the context of human-nature interaction, social resilience is an essential element of the conditions in which individuals and/or social groups interact and adapt to any changes in the environment (Adger 2000, Marshal 2007. The dependence of the individual and/or community on the environment through economic and livelihood activities is an example of connecting both social and ecological resilience (Adger 2000).
Researchers have attempted to define and to measure social resilience from various viewpoints. Harkes and Novaczek (2002) attempted to measure the resilience of a social system using the performance and status of a local customary institution (2010) measured social resilience in relation to community preparedness to disaster, while Force 1988, Bliss et al. (1998)  However, if an MPA improves ecosystem resilience we could expect it to improve the resiliency of resources users in the adjacent areas. Moreover, these definitions highlight several dimensions of social resilience, which thus require interdisciplinary understanding and analysis at various scales.  in their study of fishing industries in Northern Australia identified key characteristics of individual fishermen in their ability to cope and adapt to change in resource utilization policy. Such characteristics are ): 1. The perception of risk associated with change 2. The ability to plan, learn, and reorganize 3. The perception of the ability to cope, and 4. The level of interest in change.
The above-mentioned characteristics have been used in identifying and characterizing the vulnerabilities of stakeholder groups during the process of planning for prospective Marine Protected Areas in Egypt (Marshall et al. 2009) and commercial fishers response to management change in the Great Barrier Reef Marine Park, Australia (Sutton and Tobin 2012).
Social resilience is generally considered to lie at the -flip side‖ of vulnerability, (Folke et al. 2002b, Gallopin 2006. Kelly and Adger (2000:328) define vulnerability as -the ability or inability of individuals or social groupings to respond to, in the sense to cope with, recover from or adapt to, any external stress placed on their livelihoods or wellbeing‖. Resilience depends on the system's adaptive capacity to anticipate and to minimize any forthcoming harm, while vulnerability depends on the system's sensitivity to any possible harm from exposure (Folke et al. 2002b). For instance, household occupational multiplicity provides a range of options if anyone occupation within the household should suffer from a shock, e.g. the collapse of fish stock or drought impacting farming.

RESOURCE DEPENDENCY AND SOCIAL RESILIENCE
The relationship between humans and the environment is complex. The complex and reciprocal relationships that humans have with their environment have been an interesting subject that many researchers are trying to address (Dunlap andCatton 1994, Bourdeau 2004). The concept of resource dependency explains the nature of the relationship between community and the environment where they live and rely upon for fulfilling their livelihood (Machlis and Force 1988, Bailey and Pomeroy 1996, Adger 2000, Brookfield et al. 2005. The typical examples of resource dependent communities are those that are predominantly living from farming, logging, fishing, or mining (Machlis et al. 1990, Freudenburg 1992, Bailey and Pomeroy 1996, Adger 2000. The concept of resource dependency has been used to assess communities' social and economic conditions that are dependent on forest resources (Machlish and Force 1988, Little and Krannich 1988, Machlis et al. 1990) and coastal and fisheries related resources (Peluso et al. 1994, Bailey and Pomeroy 1996, Adger 2000, Brookfield et al. 2005).
Resource dependency is a description of a relationship between resource users and a resource. It -relates to communities and individuals whose social order, livelihood and stability are a direct function of their resource production and localized economy‖ (Adger 1999:254). The dependency of individuals or communities on natural resources is not always depending on a particular resource, but in most cases it depends on a whole integrated ecosystem Pomeroy 1996, Adger 2000).
Furthermore Adger (2000) implied that a community that is dependent on several natural resources is more resilient as compared to a community that depends only on one particular natural resource such as an underground mineral.
In the context of fisheries, Brookfield et al. (2005:57) defined a fishery dependent community as: -[…] a population in a specific territorial location which relies upon the fishing industry for its continued economic, social and cultural success.‖ How resource dependency and social resilience are related is well summarized in Adger's (2000:354) seminal paper: -[…] the direct dependence of communities on ecosystems is an influence on their social resilience and ability to cope with shocks, particularly in the context of food security and coping with hazards. Resilience can be undermined by high variability (or disturbance in ecological terms) in the market system or environmental system. Resilience therefore depends on the diversity of the ecosystem as well as the institutional rules which govern the social systems.‖ However, human systems adapt to high variability over time. For example in a fisheries dependent community, fishers employ multiple gears as a response to high seasonal and annual variability of fish abundance.
To observe and measure social resilience of communities or individuals, several social (e.g. demographic, attachment to place and family characteristics), economic (e.g. business size and approach, financial status and income source) and environmental (e.g. time spent on harvesting) attributes related to resource dependency of communities and individuals could be used (Adger 2000, Marshall et al. 2007). These attributes could positively and/or negatively affect the resiliency (Adger 2000).

ASSESSING SOCIAL RESILIENCE
The concept of (social) resilience, vulnerability and adaptive capacity are

Coping Ability
In the context of social systems, the coping threshold is a measure of the proximity to psychological and financial and marital terms indicators . Smith et al. (2003), in their study of commercial fishing families in Florida after the -net ban‖, found out that the policy changes had resulted in mental health impacts such as increasing level of stress, depression, anxiety and anger.
Similar results also showed in the study of job satisfaction among commercial fishermen in New England by Pollnac and Poggie (1988). Their finding indicated that management decision in various aspects of fishing could have an enormous impact on the fisher's work. They further argued that negatives changes in job satisfaction have been related to negative social impacts, such as family violence and lower worker productivity (Pollnac and Poggie 1988). Binkley's (2000) studies of families coping with the North Atlantic Fisheries' crisis in Nova Scotia's fishing-families indicated that financial well-being was an urgent problem. As a response, families engaged various short-term coping strategies to deal with financial issues such as increasing the wife's employment outside the home (Binkley 2000). This illustrates one of Marshal et al. (2010) key characteristics for measuring individual social resilience, which is livelihood diversity.

Level of Interest to Change
The level of interest to change corresponds to the degree of to which the system is capable of self-organization and the flexibility of an individual's financial, social, and emotional indicators (Marshall 2006, Marshall andMarshall 2007).
Individuals that have a high-level of interest to change usually have a financial, social and/or emotional flexibility (Marshall et al. 2009). These characteristics are similar to attributes of early adopters of technological innovations (Rogers 1995).
Researchers have discussed the importance of flexibility to maintain resilience (Gunderson 1999, Carpenter and Gunderson 2001, Cinner et al. 2009). Flexibility in switching livelihood strategies is important in a social-ecological system (Berkes and Sexias 2006). Loss of flexibility indicates the inability of individual or communities to exploit and benefit from other options within the industry or community ).

The Ability to Plan, Learn and Organize
This attribute suggests the ability of the individual or community to anticipate the changing future (Marshall et al. 2009). The ability to plan, learn and organize enables people to respond to disturbances by optimizing resources outside their previous experience. Understanding the perceived role of human agency in the change process can help them plan and organize for future (Cinner et al. 2009). Furthermore, the ability to reorganize after an initial change is dependent on novelty, creativity, experimentation, learning, and planning of the actors , Olsson et al. 2004b, Armitage et al. 2007).

The Perception of Risk
One of the fundamental elements in social resilience is the perception of risk .  study of commercial fishermen in Northern Australia suggested that risk perception of policy changes could influence the way the fishermen respond. The level of perceived risk by an individual determines their ability to cope and adapt to any changes and uncertainty (Marshall et al. 2010). Bradford et al. (2012) suggested that risk perception is influenced by situational (such as demographic profiles and previous experience) and cognitive factors (reflecting personal and psychological factors of the individual).

MARINE PROTECTED AREA OVERVIEW
Most of marine environments around the world are in serious decline; anthropogenic stresses and climatic related changes have caused dramatic phase or regime shifts, which are often long lasting and sometime irreversible (Huges et al. 2005). Common examples in coastal marine resources are the regime shift happening in coral reefs after habitat destruction and the collapse of many coastal and oceanic fisheries (Francis and Hare 1994, De Young et al. 2008, Huges et al. 2010).
These unwanted regime shifts are an indication that the system is losing its resilience, which has significant effects on organisms within the system and also for people who are dependent to such resources . Therefore, there has been a tremendous challenge worldwide to protect these habitats and conserve the remaining marine species that provide food, livelihood and well-being to societies (Huges et al. 2005;. A Marine Protected Area (MPA) is one of the promising tools for marine conservation and fisheries management (Tundi Agardy 1994, Dalton et al. 2012). It also serves as a link to the dynamics of social and ecological systems in the coastal waters (Pollnac et al. 2010). IUCN in Kelleher (1999: xviii) defines MPA as: -Any area of intertidal or subtidal terrain, together with its overlying water and associated flora, fauna, historical and cultural features, which has been reserved by law or other effective means to protect part or all of the enclosed environment‖.
Earlier development of MPAs drew heavily from the bio-ecological perspectives with very little attention given to social and economical aspects of the community (Christie 2004). However, researchers have shown that socio-economic factors are equally important determinants of the success or the failure of MPAs (Christie et al. 2003, Mascia 2003, Wahle et al. 2003. The management of MPAs involves some degree of restriction of human activities for resource utilization and extraction, which in most cases could create pressures and conflicts among interested stakeholders (Christie 2004). MPAs that fail to integrate the human dimension into the design and implementation processes could downplay the evolved relations of human and natural environments (Christie et al. 2003, Mascia 2003, Wahle et al. 2003. The examples of major changes brought about by MPAs are restricted resource use access, reduced fishing grounds and increased resource protection and conservation (Abesamis et al. 2006). However, in a resilient community, these changes should have the potential to generate innovation and originality among stakeholders (Folke et al. 2002b). MPA as a tool can potentially improve ecosystem resilience and therefore can be interconnected with community resilience.

SUMMARY
The resilience concept is very broad and it is indeed difficult to measure. It is a concept that incorporates all the interrelationship factors in order to understand and to assess the system. It has been used in many disciplines and has been measured in many ways. However, in order to achieve resiliency, there is a need to understand the specific context of resilience .
MPAs have been a favorable tool for managing coastal and marine areas, as it allows multiple goals at the same time. MPAs could be described as a complex system that accommodates both social and ecological goals. The management of MPAs will definitely limit some uses of resources, which could have both positive and negative impacts. This study attempts to understand one aspect of resilience, social resilience of the resource users, within the larger context of a socialecological system (MPA).

CHAPTER 3 METHODOLOGY
This chapter describes the methodology used in this study. It describes the study area, data collection methods, interviewing techniques and data analysis. There are nine priority conservation sites within the network, in which five sites have already been established as MPAs (Table 3.1).
Thirty coastal villages were selected as study sites. They are spread across four regencies within the Bali MPA network. Twenty-three study sites were associated with a managed and declared MPA, while seven villages were located in proposed sites of MPAs (Table 3.

DATA COLLECTION
Semi-structured questionnaires were used to collect the information. This study utilized three respondent categories: resource users, MPA project participants and village officials. Overlapping, but distinct survey forms were used for each category of respondent.
To facilitate interaction with the community members, local research assistants, familiar with the community and local languages conducted the in-person structured interviews (see similar methods used by Pollnac and Seara 2011 in the Philippines and Dalton et al. 2012 in the Caribbean). Local research assistants were personally trained to be familiar with the questionnaires and the interview methods.
A combination of both a systematic random and a snowball sampling methods were used to recruit respondents. At first, the head of village from each village was interviewed, to capture the general information of the village. If they were not available, another senior official was interviewed as a replacement. They were also asked to identify potential respondents for the key informant interview (MPA project participants) within their villages.
The key informants are those who are considered as local leaders. They have been involved in one or more of the MPA activities and/or functioned as the leader for local fishermen groups, operators of tourism related activities, or members of local environment and culture associations, etc.
The third category of respondent is the marine resource user. This research is focused on marine resource users as the primary respondents, as they are the ones who are most likely impacted by the MPA. For the purpose of this study, resource users are those who have their main source of income and livelihood based on coastal and marine resources utilization; e.g., fishermen, seaweed farmers, aquaculturists, boat crew and operators, dive/tourist guides, etc.
Thirty to forty resource user respondents were systematically selected from each village. The interviewers walked along the coastline in each village to identify and to recruit the respondents. All people encountered doing coastal and marine related uses along the beach during the survey, were asked their willingness to participate in the study. Interviews were only conducted with the first and the fifth persons encountered. The respondents were informed concerning the study's purpose and were asked of their availability. While most interviews were conducted on the spot, there were some interviews conducted at a different time in the same day. In this study, a very few potential respondents refused to participate, minimizing the potential for self-selection bias in the sample.

INTERVIEWS
One thousand and four face to face interviews were conducted in the study location. The questionnaires and interviews were designed to address the research questions posed in Chapter 1. The interviews were conducted in Bahasa Indonesia and usually lasted between 30 minutes and 1.5 hours, depending on the type of questionnaire used.

VILLAGE OFFICIAL AND KEY INFORMANT INTERVIEWS
The interviews with local officials were aimed to get a general profile of the community and to obtain a local permit to conduct the survey in the village. The questions for these two respondent groups were mainly focused to gather community information on: (1) community profile, (2) resource utilization activities, (3) MPA management, (4) MPA benefits, (5) community organizing and involvement, and (6) any village related problems.

RESOURCE USER
The resource users are the primary source of information for assessing social resilience. The survey form for this respondent group is focused on: (1) personal information, such as their individual, social and economic attributes, (2) environmental attitudes, beliefs and values, (3) MPA management and implementation processes, and (4) social resilience variables.

MEASUREMENT OF VARIABLES
Measurement of some variables was based on a direct response. For example the evaluation of age, education, etc. Some questions such as -have you heard of an MPA?‖ required a -yes‖ or -no‖ answer. Many questions, however, especially those evaluating attitudes, beliefs or values were measured using ordinal Likert scales. In this type of question, respondents were asked to rate how strongly they agreed with each statement using a 5-point rating scale (e.g. 1=strongly disagree, 2= disagree, 3=neutral, 4=agree, 5=strongly agree) (Likert 1932, Spector 1992).

INDIVIDUAL CHARACTERISTICS
In this study, individual characteristics measured were respondents' age, gender, years of formal education, and their primary occupations.

ENVIRONMENTAL ATTITUDES, BELIEFS AND VALUES
Respondent's environmental attitudes, beliefs and values were analyzed based on their evaluation of conservation beliefs and their subjective assessment of the degree of relationship of themselves with nature. The conservation beliefs variables were constructed of nine statements. Each of the nine statements involves some aspect of the relationships between coastal resources and human activities (see Pollnac and Crawford 2000). The following are the statements used: 1. We have to take care of the land and the sea or it will not provide for us in the future. 2. Fishing would be better if we cleared the coral where the fish hide from us.
3. If our community works together we will be able to protect our resources. 4. Farming in the hills behind the village can have an effect on the fish. 5. If we throw our garbage on the beach, the ocean takes it away and it causes no harm. 6. We do not have to worry about the air and the sea, God will take care of it for us. 7. Unless mangroves are protected we will not have any small fish to catch. 8. There are so many fish in the ocean that no matter how many we catch, there will always be enough for our needs. 9. Human activities do not influence the number of fish in the ocean.
The statements were arranged in the interviews so as to limit interference between similar statements. It will also be noticed that agreement with some would indicate an accurate belief, while agreement with others would indicate the opposite.
This was done to control for responses where the respondent either agrees or disagrees with everything. Statements were randomly arranged with respect to this type of polarity. Respondents were asked if they strongly agree, agree, disagree, strongly disagree, or neither (neutral) with respect to each statement. This resulted in a scale with a range from one to five. Polarity of the statement is accounted for in the coding process, so as a score value changes from one to five it indicates an increasingly stronger and accurate belief concerning the content of the statement (Pollnac 2013).
Responses from all nine statements were dichotomized at 3. Scores above 3 were coded -1‖ which indicates -correct‖ beliefs. Scores below 3 were coded -0‖ which indicates -incorrect‖ beliefs (Pollnac 2013). All the -correct‖ responses from respondent were summed to create conservation belief score. Conservation belief score value is hypothetically ranging between 0 -9.
Respondents were also asked to describe their subjective relationship with nature. Seven diagrams illustrating the human-nature relationship were used (adapted from Davis, Green & Reed 2009) (Fig. 3.2). The respondents were asked to choose a diagram that best describes their perceived relationship with nature (Davis et al. 2011).
Responses were coded from one to seven, respectively. As the score changes from one to seven, it indicates a closer relationship between oneself and the nature.

MPA Economic outcome
The economic outcome variable was constructed from the perceived MPA benefits to community and whether or not there was equal opportunity to receive such benefits. Respondents who perceived MPA benefits community were coded -1‖ and respondent who do not were coded -0‖. Moreover, respondents who perceived the benefits are equally distributed were coded -1‖ and respondents who perceived the opposite were coded -0‖. Only respondents who had knowledge of MPAs were asked this question.

Ecological outcome
Ecological outcome parameters were constructed from the combination of perceptions of improvement of fish abundance, coral reef condition and mangrove condition in the last five years. If respondents mentioned that there was improvement in any or all of the variables they were coded -1‖, and -0‖ if no improvements in any were mentioned.

Process quality
To measure the process quality, respondents were asked whether or not they were consulted during the planning process and whether or not the plan reflected their views. Respondents who answered -yes‖ were coded as -1‖ and -no‖ were coded as -0‖. Only respondents who had knowledge of MPAs were asked this question.

MPA Management and implementation level
For these variables, respondents were asked if there was any clear leader for the MPA, whether or not the MPA boundaries are clear, and whether or not more MPAs should be established. Respondents who answered -yes‖ were coded as -1‖ and -no‖ were coded as -0‖. Respondents were also asked their perception of MPA management committee effectiveness on a scale of from 1 to 5 where 1 = very weak and 5 = very strong. Respondents' responding above 3 were coded -1‖ and coded -0‖ for responses 3 and below. Only respondents who had knowledge of MPAs were asked this question.

SOCIAL RESILIENCE
The operationalization of social resilience, used in this study as the dependent variable, was developed by  and Marshal et al. (2010).
The key components of social resilience measured here are the individual's subjective beliefs and assessments about themselves rather than objective measures of a communities' abilities on these dimensions. Respondents were asked to self-assess their expected level of well-being in terms of their adaptability, flexibility, financial and social characteristics, and willingness to be creative and novel in their approach to adapting to the requirements of (policy) change (Marshal and Marshal 2007).
A list of statements was used to measure the respondent's response to social resilience indicators. Respondents were asked to rate their attitude to each of sixteen statements (see Table 3.3) using a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. This resulted in a scale with a range from one to five.
Polarity of the statement is accounted for in the coding process, so score value changes from one to five it indicates an increasingly stronger and accurate belief concerning the content of the statement (Pollnac 2013).
The dependent variables of social resilience measured in this study were the social resilience score and the social resilience components scores. The social resilience (SR) score was derived by summing the response scores across all sixteen questions. This resulted in a total possible score from 16 to 80.
A second measure of social resilience was derived from Principal Component Analysis (PCA). This was used to identify the underlying variables comprising the response to social resilience statements to reduce the complexity of factors to a more manageable number. Statements that are correlated with one another but are largely independent of other responses are combined into factors (Jolliffe 2005, Tabachnick andFidell 2006). In this study, the factors in the analysis were rotated using varimax, which simplifies the factor structure by maximizing the variance of a column in the pattern matrix (Abdi 2003, Jolliffe 2005. Various independent and dependent variables related to individual characteristics, perceived MPA processes and managements, environmental attitudes, and social resilience were used in this study. Table 3.4 provides the summary of variables used in this study. During the interview session, respondents were also asked to identify their primary occupation. Most of the respondents are fisherman (63%), followed by seaweed farmers 15%. While the rest of the respondents' occupations are within tourism related jobs, such as dive guide (6%), boat related jobs (5%, such as boat captain or boat crew), fish sellers (5%) and other marine related occupations (4%).
There were also respondents who reported two main occupations (2%). Three respondents did not disclose their main occupations (Figure 4.2).   Level of respondents' education also varied, ranging from 0 -18 years. Figure   4.3 provides respondents' years of education distribution. The overall mean of respondents' education found in this study is 7 years (N=934; std.dev. = 3.716).   To identify participation of respondents in MPAs, they were asked about their involvement in MPA monitoring and sea watch/patrol activities. Figure 4.5 shows that only 186 respondents (26%) mentioned that they have been involved in any of those activities (N = 717).

ECONOMIC OUTCOME
Respondents, who were aware of the MPA were asked whether or not MPAs have benefits for the communities. In total, 328 respondents said that MPAs have benefits in the community, for either themselves or others, and only 8 respondents said that MPAs do not benefit the community (Figure 4.6). Respondents were further asked whether or not the community members have the opportunity to receive equal benefits from the MPAs. In total, 294 (87.5%) respondents said that they have equal opportunity to receive benefits, and only 42 (12.5%) respondents said that they do not have the opportunity (Figure 4.6).

ECOLOGICAL OUTCOME
Respondents were asked their perception of the ecological outcome of the MPAs, which in the case of this study was the condition of fish abundance, coral reef and mangrove in the last five years. One hundred and ninety eight of the 344 respondents (58%) perceived that there is no improvement being made in terms of fish abundance. As for coral reef condition, most of the respondents (77%) perceived an improvement in its condition compared to five years ago. The majority (79%) of the 97 respondents perceived an improvement of mangrove conditions in the last five years. Respondents were asked whether or not they were consulted and the extent to which respondents' views were taken into consideration during the MPA planning process. In total 283 respondents (87%) said that they were consulted during the planning process (N = 327), and 217 respondents (89%) perceived that the MPA plans reflected their views (Figure 4.7).

DEFINING AND OPERATIONALIZING SOCIAL RESILIENCE
Principal Component Analysis was used to examine resource users' responses to 16 statements (see Table 3.3) related to social resilience indicators (adapted from Marshall et al. 2010). The social resilience of resource users in Indonesia could be best explained by five major components: (1) the adaptive capacity of the individual, (2) risk awareness, (3) perceived of socio-economic status, (4) community attachment and (5) environmental awareness (Table 4.2). These components represented 48.8% of the variance.
Individual resilience of the respondents found in this study could be best described by five components. The first component contains the statements related to respondents' ability to cope, level of interest to change, ability to learn, employability and livelihood diversity. This component of social resilience represents the adaptive capacity of individual to cope with changes and the capacity of individual to improve its condition (Smit andWandel 2006, Galoppin 2006   in the Northern Australia. They found that risk perception is one of the important denominators of social resilience of fishermen in the Northern Australia. The third component contains business characteristics, financial status and perception of equity. This component represents the socio-economic perception of respondent. In the complex of the social and ecological system, both the ecological and social economic has the same influence in the system (Perrings 1998, Levin et al. 1998). Equity issues, resilience and stewardship could be integrated in a complex system resource management (Peluso et al. 1994, Young andMcCay 1995). Forbes Environmental awareness is related to environmental knowledge (Acury 1990) and could be used to predict ecological behavior (Kaiser et al. 1999).

SOCIAL RESILIENCE SCORES AND COMPONENT SCORES
Social resilience (SR) scores of respondents were constructed from summing all 16 of social the resilience response values. The SR scores were ranged from 16 -80. Figure 4.9 presents the histogram of SR scores for all respondents. Descriptive statistic analysis found that mean SR score is 58.38, the minimum score was 43 and 77 as the maximum score (N = 934; std. dev. 5.471).
Differences between MPA and non-MPA sites with regard to the SR score were analyzed using the independent sample t-test. The analysis showed that there was a statistically significant, but very small difference (  Further analysis between MPA and non-MPA sites with regard to SR components scores were conducted (see Table 4.3). Significant differences (equal variance not assumed) were found for the risk awareness (p < 0.001; t = -6.846) and environmental awareness (p < 0.001; t = -8.323) components. The means of SR components scores were higher at non-MPA sites compared to MPA sites. Individual characteristics were constructed from personal attributes such as age, education and gender. In order to analyze the relationship between the SR score and individual attributes of age and years of education simple linear regression analyses were used. There is a statistically significant but very weak negative relationship between age and SR scores (R = -0.074; r 2 = 0.005) F = 5.82; p = 0.024).
Analysis of respondents' years of education and SR score indicated a somewhat stronger, statistically significant relationship between these two parameters (R = 0.206, r 2 = 0.042, F = 41.239; p = 0.001).
Further analysis between SR components and age revealed that a statistically significant, but weak negative relationship was found with the adaptive capacity component (R = -0.139; r 2 = 0.019; F = 18.472; p = 0.001), while the analysis between education and SR components found significant relationships with two components, which were the adaptive capacity (R = 0.28; r 2 = 0.078; F = 79.068; p = 0.001) and environmental awareness components (R = -0.073; r 2 = 0.005; F = 4.963; p = 0.026).
An independent sample t-test was conducted to see whether or not the means of SR scores differed between genders. The analysis found that there was a significant difference in SR score (

SOCIAL RESILIENCE (SR) AND SOCIAL CHARACTERISTICS
The relationship between the Social Resilience score and selected independent variables is examined in Table 4.4. Responses related to MPA were analyzed. A statistically significant, but small difference was found between respondents who were aware and those not aware of the MPA in terms of their SR score (p < 0.001; t = -3.975). The difference between respondents who participated and those who did not participate in MPA planning and management processes was also found to be statistically significant (p < 0.001; t = -6.99). However, there is no relationship between respondents who said that community members were consulted and not consulted during the process in terms of their SR score (p = >0.05; t = -1.666). A similar result was found between respondents' who perceived their view were considered and not considered (p >0.05; t = -1.676).

Independent sample t-tests between each of the social characteristic parameters
and SR components were employed to evaluate differences between means of component scores and the social parameters.  benefits. An independent sample t-test was conducted to investigate the difference in means of SR score between respondents who perceived there were economic related benefits and who were not. Respondents' perception of ecological status represents resource indicators.
Ecological outcome parameters were constructed from the combination of perceptions of improvement of fish abundance, coral reef condition and mangrove condition in the last five years. An independent sample t-test was conducted to investigate the difference in means of SR scores between respondents who perceived there were improvements and who were not.
In total 85% respondents perceived that the MPA has helped to improve the ecological condition, and only 15% of respondents perceived the opposite. An independent sample t-test between the two responses in related to their SR score revealed a statistically significant difference between these respondents: those who perceived positive ecological outcomes have a higher SR score than those who do not (p = 0.013; t = -2.507). 98% of the respondents perceived that the MPA has benefits to community and 88% respondents perceived that the benefits were equally distributed in the community. The independent sample t-test result found no statistically significant difference between the respondents who perceived that the MPA has benefits and those who did not with regard to their SR score (p = 0.871; t = -0.161). A similar result was also found in respondents' responses concerning equal MPA benefits in terms of their SR score (p = 0.150; t = -1.442).
A further independent sample t-test analysis of the economic characteristics and the SR components conducted to investigate the whether or not the difference in means existed. Table 4.7 presents the results found between SR components and the economic characteristics (only statistically significant results presented). The analysis indicated a statistically significant difference between the perceived ecological outcome and the risk awareness components (p = 0.002; t = -3.153). The mean component score of respondents who perceived ecological improvements is higher as compared to respondents who perceived no improvements being made. A statistically significant, but weak difference was also found between the equal MPA benefits parameter and the risk awareness component (p = 0.03; t = -2.184). Respondents who perceived that the MPA benefits are equally distributed scored higher as compare to respondents who perceived that the benefits were not equally distributed in the community.

SOCIAL RESILIENCE (SR) AND ENVIRONMENTAL ATTITUDES, BELIEFS AND VALUES
Linear regression analysis is used to examine the relationship between environmental attitudes/values and the SR score. The analysis showed that there is a statistically significant, weak relationship between the conservation score and SR score (R = 0.114; r 2 = 0.013, F = 12.290; p = 0.001). Further analysis between perceived self-nature relationship and the SC score indicated no significant relationship (R = 0.033; r 2 = 0.001, F = 1.012; p = 0.315).
The analysis of the conservation score with the SR components found a statistical significant, but weak relationship with the risk awareness component (R = An independent sample t-test was used to investigate the relationship between the SR score and respondents' perceptions of the MPA management indicators (Table 4.8). An independent sample t-test analysis found a statistically significant, but weak difference between respondents perception in relation to their SR score (p = 0.022; t = -2.311). There are no statistically significant relationships between the other MPA variables And the SR score.
Further independent sample t-tests were conducted between the MPA management and implementation parameters and the SR components. Table 4 Higher scores were found for both the risk perception and community attachment components for respondents who perceived clear MPA boundary as compare to respondents who perceived the boundary was not clear. It is the opposite for the environmental awareness components, respondents who perceived clear boundary scored lower as compare to respondents who perceived an unclear boundary.
In regards to MPA management committee parameter, statistically significant, but weak differences were found with regard to the risk component (P = 0.001; t = -3.604) and perceived social-economic status component (p = 0.032; t = -2.152). Respondents who perceived a strong MPA committee scored higher in both risk awareness and social-economic components as compare to respondents who perceived a weak committee. No significant differences were found with any of the components of social resilience with regard to establishment of more MPAs.

DEGREE OF VARIABILITY IN INDIVIDUAL RESILIENCE
The social resilience of resource users in Indonesia could be best explained by five major components: (1) the adaptive capacity of the individual, (2) risk awareness, (3) perceived of socio-economic status, (4) community attachment and (5) environmental awareness.
This study also found that the social resilience (SR) scores of people who lived within MPA and non-MPA areas are statistically significantly different. The mean score of SR is slightly higher for respondents in the non-MPA area as compared to respondents living within the MPA area. Detailed analysis of SR components between MPA and non-MPA sites found statistically significant differences in the risk awareness and environmental awareness components, where respondents from non-MPA areas scored slightly higher than those from MPA sites.
These results indicate that MPAs have a weak negative impact on the level of resource users' social resiliency. As Abesamis et al. (2006) noted, MPAs could bring a major change to coastal communities such as restricted resource use access, reduced fishing grounds and increased natural resource protection and conservation. Thus, it is going to be a challenge for the MPA managers concerning how to improve the resiliency of resource users within the MPA. Lebel et al. (2006) suggested that there are at least three attributes of governance that the manager should focus on to improve the resilience of a social-ecological system: (1) stakeholder participation; (2) polycentric or multilayered governing institutions and (3)

INDIVIDUAL RESILIENCE AND SOCIAL CHARACTERISTICS
Resource dependent people are typically less flexible as they only have limited transferable skills . They argued that, young resource users typically leave formal education early for securing an apprenticeship, while older resource users typically have become too attached to their job and became less flexible for any new employment opportunities within their area. As a result, they are -locked‖ into their occupation , which ultimately could negatively affect their resilience. Age, education level and attitude to working elsewhere are some of indicators of individuals' employability ).
This study found that age and education, have a significant relationship with the SR score. Interestingly, a negative correlation between age and SR score was found. This indicates that individual resiliency decreases as age increases. An analysis of the SR components also found a negative but significant relationship between age and the adaptive capacity components. Sutton and Tobing (2012) study of fishers in the Great Barrier Reef found a similar result, where age had a significant but negative correlation with the fishermen's SR. These facts suggest that age might likely be used to predict the direction (either high or low) of individual's social resilience levels.
Although the relationship is very weak, as expected, years of education have a positive relationship with the SR score. This is somewhat similar to the Adger et al.
study in 2002 that found that education is a factor that enhances social resilience of coastal communities in Vietnam. People who are educated will have access to information, which in turn could result in more options for jobs. Education also contributes to the adaptive capacity and environmental awareness components of social resilience in Indonesia;Fulan (1970) argued that education is positively linked to individual adaptive capacity. In addition, a higher education level will increase employability (Graham and Paul 2010). A well-designed environmental education program could be used to increase environmental awareness, which in turn could change ones behaviors towards the environment (Hungerford & Volk 1990).
Although the roles of woman in the resource dependent communities have been acknowledged, the hierarchy of gender is still happening (Bennett 2005). In this study, gender was found to have relationship with the level of individual social resilience. Male resource users tend to have higher SR score compared to female. To improve the level of social resilience of female resource users, they have to be actively engaged in the MPA planning and management processes. A study of forest communities in India and Nepal found that the presence of females in community institutions for forest governance were significantly improved the forest condition (Agarwal 2009).
Social characteristics have been related to the level of either individual or community social resilience (Adger et al. 2002, Marshall 2007, Sutton and Tobing 2012. Social characteristics such as awareness and participation in MPA activities, which were statistically significantly related to resilience, could help to enhance their ability to cope and adapt to any sudden change brought by the MPA. The analysis of relationships between SR components and the social characteristic parameters indicates that two of the most important components of social resilience--adaptive capacity and risk awareness-are related to these social variables. In order to increase the resiliency, the MPA managers should have to understand the social characteristics of both the individuals and communities. Programs to compensate for the short-term impacts of MPA establishment should be designed in line with the needs and characteristics of the involved community to avoid the failure of program implementation. In order for the MPA program to be successful, the community has to be actively involved from the earliest stages of MPA planning and management processes. Mascia (2004)  The study also found that the current ecological status of the marine resource has a statistically significant positive relationship with the SR score, while the perceived benefit of MPA and whether or not the MPA benefit was equal were not related to the score. This result explains the interrelationship between the social and ecological factors in a complex social-ecological system, such as MPAs , Pollnac et al. 2010. Maintaining the ecological performance of MPAs in the long-term could positively contribute to resiliency, as healthy marine resources could potentially diversify the source of income for resource users.

INDIVIDUAL RESILIENCE AND ENVIRONMENTAL ATTITUDES AND BELIEFS
Environmental attitudes of an individual heavily influence their ecological behavior (Kaiser et al. 1999). In this study, environmental attitudes and beliefs characteristics are related to the risk awareness, perceived social-economic status and environmental awareness components of respondents' social resilience. The analysis To manage a complex social-ecological interaction system such as a protected area, an effective governance mechanism is needed. Adaptive co-management has been used and proven to be useful in many contexts and situations (Wollenberg et al. 2000, Olsson et al. 2004a. Armitage et al. (2008:95) presented four important aspect of co-management: -…innovative institutional arrangements and incentives across spatiotemporal scales and levels, learning through complexity and change, monitoring and assessment of interventions, the role of power, and opportunity to link science and policy‖.

5.6
STUDY LIMITATIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH.
The author acknowledges a numbers of limitations in this study. To build an operational definition and concepts of social resilience, an in depth interview with resource users is necessary to get descriptive information concerning social/individual resilience components to compliment the quantitative responses. Limited sets of questions were used to explain the potential social resilience indicators, which might not be best to capture the essence of such indicators in defining social resilience.
Despite some of its limitations, this study has shown that some personal and social attributes associated with an MPA could potentially have an impact on the level of individual resource users' social resilience. However, a more detailed study of demographics and socio-economic indicators to compliment the information found in this study is needed. Strategies that the resource dependent communities employed in order to cope with the changes brought by the establishment of MPA also need to be further investigated. Finally, building baseline information of people's perceptions of social resilience indicators could help to assess the potential impacts of MPAs on resource dependent people. This study has discovered some important aspects of social resiliency and its relation to some aspects of MPAs. The social resilience of resource dependent people in Indonesia could be best explained in five components, which are: adaptive capacity, risk awareness, perceived social-economic status, community attachment and environmental awareness. In order to fine-tune the finding, these components of social resilience should be tested in future studies in various locations and settings. A summary of statistically significant findings between SR score and SR components score can be found in Table 6.1.