The Impacts of Fishermen’s Resilience towards Climate Change on Their Well-Being

: This study aims to examine the mediating effect of climate resilience on the relationship between socio-economic, social relationship, social environment, and sense of community with ﬁshermen’s subjective well-being (life satisfaction, positive feeling, and negative feeling). This quantitative study performed a multi-stage sampling and selected 400 ﬁshermen as respondents. For analysis purpose, this study relied on Partial Least Square Structural Equation Modelling (PLS-SEM). The structural model concluded that socio-economic, social relationship, social environment, and sense of community explained 55.4% variance in resilience. The mediating analysis conﬁrmed the resiliency mediating effect on all twelve hypothesized relationships. A number of recommendations related to extending the areas of the study, to focus speciﬁcally on the small-scale ﬁshermen, and to consider the inclusion of several others’ additional mediating effects were highlighted.


Introduction
Climate change, as stated in the National Policy on Climate Change in 2009, refers to "any change in climate over time that directly and indirectly affects human and their activities as well as natural systems and processes".Several climate change impacts are detected in Malaysia [1].In term of rising temperature, for example, Ministry of Natural Resources Malaysia [1] approximately estimated a rising temperature of 0.25 • C per decade for the peninsular Malaysia, 0.20 • C per decade for Sabah, and 0.14 • C per decade for Sarawak, while Kwan et al. [2], in their studies, detected more than 51% warm day changes and more than 62% of warm night changes in areas such as Kuala Terengganu, Kuantan, Setiawan, and Bayan Lepas.With regard to sea level rise, Culver et al. [3], in their study, concluded that Setiu (a district in East Coast of Peninsular Malaysia) has recorded an increase in the rate of sea-level rise from 1.26 mm year −1 to 3.2 ± 0.6 mm year −1 , while another study by Abd Hamid et al. [4] reported trends related to rising sea levels in Malaysia ranging from 3.27 ± 0.12 mm year −1 off eastern Malaysia to 4.95 ± 0.15 mm year −1 west of Malaysia.They further concluded that, in a period of 22 years (1993-2015), the mean rising rate around Malaysia is 4.22 ± 0.12 mm year −1 , and the cumulative sea level rise is 0.05 m.
In a study by Wan Azli [5] noted that states in the West Coast of Peninsular Malaysia have experienced unstable rainfall pattern.NAHRIM [6], on the other hand, forecasted a substantial increase (11 to 43%) in mean monthly rainfall over the area of the east coast of Peninsular Malaysia, while Moyawa et al. [7], who conducted a study on 54 stations along the east coast region of Peninsular Malaysia, reported that the heavy rainfall in the studied areas increased by 1.5 days per decade over the time period from 1971 to 2010.Regarding extreme events in Malaysia, Billa et al. [8] reported an increase of rainfall intensities during

The Subjective Well-Being
Subjective views of well-being can be understood as a continuous assessment of life satisfaction in a psycho-bio, politico-socio-economic, and physical environment [21].The subjective perspective can describe well-being, which is an evaluation of the satisfaction of life that people experience.It can be in the form of life satisfaction, a positive feeling, or a negative feeling [22].Life satisfaction is a global assessment of a person's quality of life based on his/her chosen criteria, and that assessment relies on a comparison between one's circumstances and with what is thought to be an appropriate standard [23].Positive and negative feelings are people's inner guidance system.Their feelings are a feedback mechanism that informs them whether what they do is right or wrong.Positive and negative feelings are part of human life, and how much they experience these feelings provide a strong influence on their subjective well-being [24].

The Study's Objective
The advantages stated have attracted scholars to further investigate things related to a community's resiliency.A number of current studies related to community's climate resiliency have been conducted by ref [25][26][27][28][29][30][31][32], for example, looked into household subjective resilience measurement, which refers to an individual's cognitive and affective self-evaluation of their household's abilities to respond to changes or risks.Interestingly, their study tried to relate the subjective resilience with other core concepts related to perceived adaptive capacity, subjective well-being, and psychological resilience.A study by Borquez et al. [30] accentuated the importance of integrating processes of the co-production of knowledge in Chile as a means to better communicate and transform abstract concepts, such as resilience theory, into practice.Pisello et al. [31], via their study in Italy, found that people's educational background strengthens their environmentally aware behaviors, thus lessening the risks of environmental hazards such as rising temperature and heat waves.Tambo [32], on the other hand, concluded female-headed households in Ghana are weaker with regard to their resilience towards climate variability and accentuated that interventions aimed at building households' climate resilience should focus at raising household income, improving food security, and asset building.Despite several studies being produced at the international stage, a similar situation is not seen in Malaysia, and Tanggang et al. [33] stressed that more local studies should be conducted in order to formulate sound policies and offer the concerned parties with facts and figures for scientifically sound adaptation measures.Existing but few local studies have responded to Tanggang et al. [33], such as ref [13,[34][35][36][37][38].A study by Abu Samah et al. [13], for example, which attempted to compare adaptation ability between youth and older fishermen, concluded with an equal strength of both groups in terms of cognition, practice, and awareness, and discussed the possibilities of a stronger social relationship, technology usage, and money-making opportunity as factors that result in that equal adaptation strength.Another study by Shaffril et al. [39] provided descriptive explanations on fishermen's adaptation ability in climate change affected areas, and demonstrated fishermen adaptation strength related to the attachment to occupation; attachment to place; local environmental knowledge; environmental awareness, values, and attitudes; and formal and informal networks; nevertheless, they also expressed their concerns on fishermen's strong attachment to the fishing activities, and suggested alternative learning skills as one of the solutions for the issue [34]), on the other hand, examined climate change impacts on paddy production in Malaysia based on the autoregressive distributed lag (ARDL) model both with national and state level (Kedah) data.The analysis indicated that national paddy production in Malaysia is reduced by 7%, resulting from the rising temperature, while a long-term unstable rainfall pattern will cause a 0.371% reduction in paddy production.Within the case of climate change-related studies, however, most of them have not placed resiliency as their main focus, or merely focus on the direct effects of resiliency, either as an exogenous or endogenous variable.This situation denies the fact that resiliency can be used as an effective mediating factor in social science based-studies, as evidenced by Shi et al. [40], who studied the mediating factor of resilience on the relationship between big five personality and anxiety among Chinese medical students; Ferreira et al. [41], who focused on the mediating effects of resilience in the relationship between organizational support and resistance to change; Rostami et al. [42], who looked into the mediating role of resiliency in the relationship between social support and quality of life of law enforcement; and Chung-Il and Kang Yi [43], who completed a study related to the mediating effects of resilience on achievement-oriented parenting style, school adjustment, and academic achievement as perceived by children.Moreover, not much research has been done to deliver a better understanding on Malaysian fishermen resiliency towards climate change impacts.As Malaysia is formidably impacted by climate change, and the fact that fishermen are one of the groups with considerable dependence on nature stability, understanding resiliency from both the perspective of the Eastern community the fishermen are expected to contribute significant knowledge on how this group responds to the worsening impacts of the changing climate.The gaps, or lack of studies focusing on the mediating effect of resiliency on well-being and scarce number of studies representing the Eastern and fishermen views, have driven this study to its main objective, which is to examine the mediating effect of resiliency towards climate change impacts of fishermen's subjective well-being.This study focuses on resiliency as the mediating factor, and it is expected to connect the relationship between four exogenous variables, namely socio-economic status, social relationship, social environment, and sense of community with fishermen's subjective well-being.The subjective well-being is examined from three perspectives, namely life satisfaction, positive feeling, and negative feeling.

The Development of Study's Hypothesis and Research Model
As suggested by Lucas and Diener [44], continuous effort to assess all possible factors that contribute to a better subjective well-being are important.Such effort enables future theoretical work to use it to strengthen and develop a solid ground for endorsing one model over the others.To the authors' knowledge, none of the previous model attempts to examine the mediating factors of resiliency towards climate change considers the relationship of the selected exogenous variables (socio-economic, social relationship, social environment, and sense of community) with the selected endogenous variables (three components of subjective well-being-life satisfaction, positive feeling, and negative feeling).The decision to include the selected factors in the model are based on criteria-the independent variables (exogenous variables) must be proven to significantly affect the mediator, and the mediator must be proven to have a significant unique effect on the dependent variable (endogenous variable) [45].To fulfil the criteria, the researcher searched the findings of previous studies related to resiliency and subjective well-being and further discussion focusing on hypothesis development based on the selected factors.

Social Relationship
The social relationship refers to the individual's feeling of becoming connected and accepted by others such as family, friends, colleagues, and members of community [21].A significant relationship between the social relationship and subjective well-being has been recorded by past studies.Being respected by colleagues and members of community while, at the same time, receiving motivation, for example, are significantly related to a better wellbeing [21], while receiving encouragement from a close one, on the other hand, is related to stress reduction, which eventually strengthens one's well-being [46].Shaffril et al. [20] have confirmed the influence of the social relationship on fishermen resiliency towards climate change impacts.A good social relationship with their surrounding community can provide a social support during stressful events and financial difficulties [47,48] while the good social ties are the most important sources for fishermen to gain information on money making opportunities and weather [49].Based on previous findings, resilience is evidenced to be a good mediator between social relationship and subjective well-being.In the context of communities living in an area prone to climate change impacts, for example, resilience strengthened a community's capacity to maintain or recover high well-being during life adversity, and was proven to be significantly related with life satisfaction and positive and negative feelings [50,51].The combination of past studies' findings offers a basis with which to develop the following hypotheses: Hypothesis 1 (H1).Resiliency mediates the relationship between social relationship and life satisfaction; Hypothesis 2 (H2).Resiliency mediates the relationship between social relationship and positive feeling; Hypothesis 3 (H3).Resiliency mediates the relationship between social relationship and negative feeling.

Sense of Community
Sense of community can be one of the factors that influence well-being.Having a good sense of community develops an acknowledged interdependence with community members, a willingness to maintain such a relationship by providing or doing what others expect them to do, and the feeling that one is larger dependable and stable structure [52].
A sense of community develops a stronger sense of belongingness, collectiveness, and cooperation, which in turn helps to build resilience in a community [53,54].Within the scope of areas prone to climate change impact, fishermen who are strongly attached to their community are seen to have reciprocal connections of interactions, increased levels of trust, and access to knowledge that are exchanged for mutual benefit [55].Based on empirical evidence, a sense of community is proven to be the antecedent to resilience [54], and resilience is associated with life satisfaction and positive and negative feelings [56,57].It was hypothesized that a sense of community causes a significant effect on life satisfaction and positive and negative feelings through the mediating effect of resilience.Specifically, individuals with a stronger sense of community and a greater resilience are able to strengthen their well-being.The combination of research literature provides a basis with which to develop the following hypotheses: Hypothesis 4 (H4).Resiliency mediates the relationship between sense of community and life satisfaction;

Hypothesis 5 (H5). Resiliency mediates the relationship between sense of community and positive feeling;
Hypothesis 6 (H6).Resiliency mediates the relationship between sense of community and negative feeling.

Social Environment
Social environment refers to the local social space surrounding one's life.Several past studies have concluded that social environment is where the community settles, can develop, maintain, or strengthen the existing level of community's well-being [58].Social environment influences fishermen resilience as it enables them to cooperate and strategize their response plan and, at the same time, reduces the possibility of having conflict among community members during extreme events and expediting the rescue process [20].Resilience can explain the relationship between social environment and subjective wellbeing components such as life satisfaction and positive and negative feelings.Several past studies have confirmed social environment as an antecedent to resilience [57,59,60], and resilience is associated with life satisfaction and positive and negative feelings [56,57].It is expected that social environment has a significant effect on life satisfaction and positive and negative feelings through the mediating effect of resilience.Individuals with a better socioeconomic status and greater resilience experience a healthier well-being.The combination of past studies' findings offers a basis on which to develop the following hypotheses:

Hypothesis 8 (H8). Resiliency mediates the relationship between social environment and positive feeling;
Hypothesis 9 (H9).Resiliency mediates the relationship between social environment and negative feeling.

Socio-Economic Status
Within the context of the current study, socio-economic status refers to the fishermen's income, investment, standard of living, and their access to services and accommodation.Socio-economic status was proven to be significantly associated with well-being.People with a better financial situation are, for example, able to obtain goods and services that improve their experienced quality of life, while better savings allow them to invest, which might return them with profits in the future [61].These findings were supported by Jorgensen et al. [62], and they also added that the effect of income, though small, is comparable with other determinants of well-being.Doubtlessly, socio-economic status is an important factor for developing a stronger resiliency.Within the scope of fishermen, those with a better socio-economic status such as income, savings, better standard of living, and greater access to services and facilities are more likely to report a better resiliency [20,63].Fishermen with a weaker socio-economic status, on the other hand, are lacking in terms of their flexibility, with which they can successfully absorb the costs of change and are expected to express their reluctance on taking further risks [64].The existing literature has suggested that there is a relationship between socio-economic status and well-being [61,65,66], and previous studies have confirmed that resilience may function as a variable that has a significant relationship with people's well-being [50,51].The combination of research literature provides a basis with which to develop the following hypotheses: Hypothesis 10 (H10).Resiliency mediates the relationship between socio-economic and life satisfaction; Hypothesis 11 (H11).Resiliency mediates the relationship between socio-economic and positive feeling; Hypothesis 12 (H12).Resiliency mediates the relationship between socio-economic and negative feeling.

The Research Model
Based on the findings of past studies, the current study has come out with its own research model (as illustrated in Figure 1).The model consists of four exogenous variables, namely socio-economic, social relationship, social environment, and sense of community, and four endogenous variables, namely, resiliency towards climate change impacts, life satisfaction, positive feeling, and negative feeling.One of the endogenous variables, namely resiliency towards climate change impacts, also act as the mediating variable.

Participants
The study was conducted among 400 fishermen in climate change-affected areas.A huge majority of the respondents were male (97.3%) and the mean score for age was 46.9.
A total of 32.8% of them possessed an upper level of education (completed their secondary school at form 4 or form 5) and almost three quarters of them were married.More than 80% of the respondents were coastal fishermen, while, in terms of their household income, the recorded mean score was RM1, 850.53 a month (roughly equal to USD 450 a month).Nearly three quarters of them (73.5%) had four or more household members.A total of 38.8% of the respondents can be considered as experienced, as they had involved in the fishing industry for more than 21 years.On average, the respondents spent 19 days a month conducting fishing operations, and most of them (76.8%)relied on fish as their main catch.Only 20.5% of the respondents were equipped with a vessel sized 26 feet or longer.More than one third of the respondents used a boat engine ranging from 31 to 40 horsepower, and 71.8% of them used seines as their main catching tool.

Procedures
As the population of fishermen settling in climate change-affected areas are unknown, the sample size calculator Raosoft was used.Based on this tool, relying on the confidence interval of 95% and a margin error of 5%, the suggested number of the sample was 377.The study, however, aims to have a higher number of respondents, namely of 400.Multistage sampling was employed, whereby, at the first stage of sampling, climate change-affected areas were listed.The list of climate change-affected areas was gained based on Meteorology Department data and previous studies by Kwan et al. [3] and Awang and Abdul Hamid [67].The main criteria for the areas to be included were at least that an area was affected by either one or both of the following climate change impacts: rising temperature, sea level rise, or both.Based on the list, a total of 14 areas was listed (refer to Table 1).Then, based on a cluster sampling (a sampling technique that is based on geographic segmentation), a total of four areas were randomly selected from the list, namely Bayan Lepas, Kuala Terengganu, Kuantan, and Setiawan (please refer to Table 1 for details).Subsequently, at the second stage, based on a purposive sampling, a total of 100 fishermen from each of the selected areas were selected as a respondent.Purposive sampling is a technique that allows the researcher to rely on his or her own judgment when selecting members of population to involve in the study.Based on this technique, the researcher selected the respondents only if they confirmed that he or she is a full-time fisherman of their respective area.If not, the researcher opted to select another person, and this process was repeated until the number of respondents required for that area was fulfilled.Permission to conduct data collection at the areas was gained from either the village leader or related officers.The data collection was conducted at fishermen's places of interest, such as a waqf (small cater, usually can be found at the coastal areas), coffee stall, and jetty.The process was assisted by trained and experienced enumerators and was monitored by the research team members.The survey was the main data collection technique employed, and it was conducted in Malay language.The respondents were allowed to ask questions if there was any confusion related to the items queried to them.

Measures 2.3.1. Socio-Economic Status
Socio-economic refers to the fishermen's income, investment, standard of livings and their access to services and accommodation provided.This section consists of four items using a scale from 1 (strongly disagree) to 5 (strongly agree) on a Likert-type scale.The items were adapted from Chi et al. [52].Examples of items were 'I or my family has enough household income' and 'I or my family has a high standard of living in this society'.The Cronbach alpha value for this factor was 0.622.There are several perspectives of scholars on the accepted Cronbach alpha value, while 0.700 is stated as the best value by Nunally, Hinton et al. [68,69], on the other hand, claimed that a value between 0.500 and 0.690 is considered as a moderately reliable.

Social Relation
The social relation refers to the degree to which individuals feel connected and accepted by others.This section consists of nine items using a scale from 1 (strongly disagree) to 5 (strongly agree) on a Likert-type scale.The items were adapted from Chi et al. [52].Examples of items were 'I have a good relationship with other fishermen colleagues' and 'I have a good relationship with my neighbours'.The recorded Cronbach alpha value for this factor was 0.936.

Sense of Community
Sense of community refers to the perception of similarity to others, an acknowledged interdependence with others, a willingness to maintain this interdependence by giving to or doing for others what one expects from them, and the feeling that one is part of a larger, dependable, and stable structure.This part consists of seven items using a scale from 1 (strongly disagree) to 5 (strongly agree) on a Likert-type scale.The items were adapted from Chi et al. [52].Examples of items were 'I feel welcome to the people around me' and 'I will contribute to the community here'.The recorded Cronbach alpha value for this factor was 0.920.

Social Environment
Social environment refers to the local social space surrounding one's life.This section contains 11 items using a scale from 1 (strongly disagree) to 5 (strongly agree) on a Likerttype scale.The items were adapted from Chi et al. [52].Examples of items were 'the community over here respect each other' and 'my place of residence is safe and peaceful'.The recorded Cronbach alpha value for this factor was 0.937.

Resiliency
Marshal and Marshal [18] assessed levels of resilience by four key characteristics, including: their perception of risk, their ability to plan and cope, and their level of interest in change.Resiliency was measured based on 12 items developed by Marshall and Marshall [18].The original instrument is in English and was translated to Malay language using forward-translations and back-translations to ensure accurate and reliable translation.A Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree) was used.Example of items included 'I have many options available if I decide to no longer be a fisher' and 'I am confident that I could get work elsewhere if I needed to'.Originally, there are four negatively worded statements; nevertheless, these statements were reversed into positives, as Wong [70] argued that negatively worded statements might create confusion and result in a lack of understanding among the respondents.The resulted Cronbach alpha value for this factor was 0.860.

Subjective Well-Being
The present study examined the subjective well-being based on life satisfaction, positive feeling, and negative feeling.For life satisfaction, five items based on a scale developed by Diener et al. [22] was used while for positive and negative feelings.A SPAEN instrument, which consists of a total of 12 items (6 positive feelings and 6 negative feelings), was referred [24].The original instrument was in English and was translated to Malay language using forward-translations and back-translations to ensure accurate and reliable translation.For satisfaction of life, a seven-point Likert-type scale was used, ranging from strongly disagree (1) to strongly agree (7) while, for SPAEN, a scale of five ranging from 1 (very rarely or never) to 5 (very often to always) was used.For answering SPAEN, the respondents were asked to recall their affective well-being in the past four weeks.Drawing on Diener et al. [24], the period of four weeks was selected, as it offers an adequate sample of feelings compared to a shorter period that might not have been representative.An example of items include 'In most ways my life is close to my ideal' and 'The conditions of my life are excellent'.The Cronbach alpha value recorded for life satisfaction was 0.719, positive feeling was 0.951, and negative feeling was 0.875.

Data Analytic Strategy
A descriptive analysis such as frequency, percentage, and mean score was used to describe the demographic data.In order to achieve the objective determined, a Partial Least Square Structural Equation Modelling (PLS-SEM) was used.PLS-SEM was chosen for this research as it is less restrictive compared to covariance-based structural equation modeling (CB-SEM) and it suits the research's objectives.Hair et al. [71] have specified several criteria for the researchers to opt for PLS-SEM rather than CB-SEM, such as when the analysis is concerned with testing a framework from a prediction perspective and is concerned with distribution issues such as a lack of normality.PLS-SEM has two main components, namely the measurement model (also known as an outer model) and a structural model (also known as an inner model).The outer model investigates the quality of all constructs, taking into account the measurement's reliability and validity.The inner model, on the other hand, measures the relationships between the different constructs of the model [71].Raw data based on a survey of 400 respondents were analyzed by using tools such as Smart PLS and SPSS.In relation to the sampling technique, to have PLS-SEM in a study that relies on multistage sampling was not a problem, as PLS-SEM has the ability to cater to any kind of sampling, including non-probability sampling, such as purposive sampling.Nevertheless, it should be noted that data based on non-probability sampling cannot be generalized from the sample to the population with known confidence.

Common Method Variance (CMV)
The study also checked the possibility of having a common method variance (CMV)a situation where a common source could result in a correlation which produces wrong internal consistency before proceeding to hypothesis testing.A different Likert scale was used to measure the constructs as a way to reduce CMV, while, as suggested by MacKenzie and Podsakoff [72], the capability and ability of the respondents are considered before they answered the questions.Harman Single Test was performed, and the resulting value was less than 50% (26.3%) and demonstrated that CMV is not a pervasive issue.

Measurement Model
The PLS-SEM process consists of two steps: The first one is the measurement model, which focuses on convergent and discriminant validities.The measurement model focuses on the construct, with their corresponding items in the model.All of the items in the study were reflective items.

Convergent Validity
Regarding the convergent validity, it inspects items' reliability and validity.This type of validity examines whether or not the items converge on a construct they are supposed to measure [71].To test the items' reliability and validity, the value of the factors loadings, average variance extracted (AVE), composite reliability, and Cronbach alpha value were referred.The factor loading ensures the items measures the constructs and the minimal accepted value is 708 [71], while the AVE, composite reliability, and Cronbach alpha value examine the reliability and need to pass the minimal value of 0.500, 0.700, and 0.700, respectively.
The loadings of the 60 items ranged between -0.542 and 0.966 (Table 2).As some of the items did not exceed the recommended value of 0.708 by Hair et al. [71], a total of 15 items were removed.Several items were maintained, though the items did not exceed the minimal value, as Byrne [73] claimed that loading equal to or greater than 0.500 are acceptable if the summation of loading result in high loading score resulted in an AVE score greater than 0.500.The AVE for the constructs ranges from 0.512 to 0.908, the Cronbach alpha for constructs ranges from 0.713 to 0.950, while the composite reliability ranges from 0.817 to 0.967 (Table 2).Although Hair et al. [71] stated that a value of more than 0.900 for composite reliability is deemed undesirable as it denotes that they measured the same phenomenon and is therefore unlikely to be a valid measure of the construct; nevertheless, Becker [74], in his discussion, claimed that, if the items tap into different aspects of the measured constructs but are still highly correlated, then the study simply has a good measurement model.

Discriminant Validity
To examine the discriminant validity, the Fornell Larcker criterion [75] was referred.To examine whether the discriminant validity has been met, other construct correlation values must be lower than the AVE square root.The results presented in Table 3 indicate that all the constructs square of AVE was higher than the correlation of other constructs and fulfils the requirement of discriminant validity.In another method userd to determine whether the discriminant validity is by referring to the Heterotrait-Monotrait (HTMT) ratio of correlation was referred [76].Drawing on Table 4, the discriminant validity has been ascertained, as all the values fulfil the criterion of HTMT 0.900 [77] and HTMT 85 [78].

Structural Model
The next step is to examine the structural model.However, prior to this, it is vital to check the lateral collinearity test.All of the resulting inner Variance Inflation Factor (VIF) values are less than 5.0, which confirms that lateral multicollinearity is not a concern for the study.A bootstrapping of 5000 samples was tested to identify the structural significance of the path coefficient.Analyses performed demonstrated that all seven relationships have a t value of more than >1.645, which is therefore significant at 05 level of significance.Specifically, the predictor of the sense of community (β = 0.247, p < 0.001), social environment (β = 0.276, p < 0.001), social relationship (β = 0.243, p < 0.001), and socio-economic (β = 0.218, p < 0.001) are positively related to resiliency, which explains 55.4% variance in resilience.The R 2 value of 0.554 denotes a substantial predictive accuracy.Meanwhile, resilience was found to be positively related to life satisfaction, which explains almost 9% of the model, and the R 2 value of 0.088 denotes a weak predictive accuracy.Resilience was also found to be positively related to positive feeling, which explains 46.4% of the model.The R 2 value of 0.464 denotes a substantial predictive accuracy.In terms of the relationship between resiliency and negative feeling, a negative relation was recorded, which explains only 2.4% of the model.The R 2 of 0.024 demonstrates a weak predictive accuracy (Table 5).Next, the effect sizes (f 2 ) were assessed.In reporting and interpreting studies, it is important to reveal the substantive significance (effect size) and statistical significance (p-value) [79].Based on Table 6, it can be seen that sense of community (0.083), social environment (0.102), social relationship (0.093), and socio-economic (0.088) has a small effect in producing R 2 for resiliency.Moreover, resilience was found to produce a small effect on the R 2 of life satisfaction (0.096), a medium effect on the R 2 of positive feeling (0.317), and a small effect on the R 2 of negative feeling (0.025).Additionally, a blindfolding procedure was performed to determine the predictive relevance of the model.To examine this, the Q 2 value must be larger than 0 to conclude that the model has predictive relevance for a certain endogenous construct [71].All four Q 2 values for life satisfaction (Q 2 = 0.047), negative feeling (Q 2 = 0.025), positive feeling (Q 2 = 0.195), and resiliency (Q 2 = 0.257) were more than 0, demonstrating that the model possessed an adequate predictive relevance.

Hypothesis Testing
The performed bootstrapping analysis confirmed that all twelve indirect effects (β ranging from −0.043 to 0.155) were significant, with t values ranging from 2.662 to 5.083.
The indirect effects, including 95% boot CI Bias-correction, showed that all of the resulting values were not straddling a 0 in between, and confirms that there was mediation [80].Consequently, the study concluded that the mediation effects are statistically significant (refer to Table 7).

Discussion
The structural model concluded that all predictors have a significant relationship and accounted for 55.4% of variance in resiliency.Resiliency, on the other hand, explained 2% variance in negative feeling and 9% in life satisfaction, while explaining more than 44% variance in positive feeling.This study confirmed that all of the twelve hypotheses were supported.Mediation analysis confirmed the mediating impacts of resiliency, as it is able to explain the relationship between socio-economic aspects, sense of community, social relationship, and social-environmental relationship with life satisfaction, positive feeling, and negative feeling.
The analysis confirmed the ability of socio-economic status to predict the fishermen resiliency, and this is in line with several previous findings [20,[61][62][63].Within the scope of this study, fishermen with good socio-economic status are expected to have greater access to resources, which allows them to plan ahead for the climate change impacts.Such advantages allow the fishermen to have more adaptation options, diversify their livelihood skills, and may readily be able to switch between occupations, be competitive, and expedite their recovery process after extreme events [20,63].Sense of community is another important antecedent for resiliency.To have this finding is not surprising, as [39] have clarified fishermen strong attachment to their community and place of settlement.In fishermen's community, collectivist cultures such as gotong-royong (mutual help), rewang (mutual cooperation during wedding ceremony), and those involving in social activities (e.g., playing checkers, repairing nets, chatting) at small shelters called a waqf are commonly practiced [81].All these create strong sense of belongingness, collectiveness, mutual understanding and cooperation, which later, according to [54], develop their responsibility to help others in dealing with the risks and damages, and for keeping the members safe and healthy, eventually strengthen their resiliency.
The analysis confirmed a significant relationship between social relationship and fishermen resiliency towards climate change impacts.Shaffril et al. [20] have explained that a good social relationship among the fishermen enforces the resilience that buffers the impacts of climate change.Fishermen with a good social relationship with their family members, fishermen colleagues, leaders, and others surrounding the community are expected to receive a higher motivation to learn and master several alternative livelihood skills, while, at the same time, are encouraged to share their knowledge and experience, especially on how to deal with the climate change impacts.Moreover, the fishermen's social environment is significantly related to resiliency, and this is in line with the previous findings of ref [57,59], and ref [60].In a community that is collectivist, maintaining a harmonious interpersonal relationship for the fishermen is important.Kito et al. [82] explained that these relationships are so long-lasting and extremely difficult to change that to not keep peace can mean unhappiness for everyone involved.This result does not contradict a study done by Shaffril et al. [20], who stated that a harmonious environment does exist within the fishermen community, where good relationships and mutual respect are highly emphasized.Such healthy social environments influence their resiliency, as it enables mutual efforts among them to plan reactive and proactive responsive strategies towards the severe impacts of the climate change.Furthermore, creating a good social environment reduces the possibility of having conflicts among fishermen during extreme events and expediting the rescue process.
Resiliency was confirmed to have a significant relationship with positive feeling and has a substantial predictive accuracy.Such a finding is not surprising, as highly resilient individuals turn out to achieve higher levels of positive feeling in a stressful situation [83].Despite the facts that climate change impacts are expected to worsen, fishermen with a strong resiliency are predicted to stay positive and happy, as they know the things that they should do to lessen the impacts.According to ref [81], although fishermen are strongly attached to their fishing job, they are still open to learn new livelihood skills and have great occupational mobility.During extreme weather, resilient fishermen do several other things to absorb the impacts.They might go to other places and seek for available opportunities, while those with alternative livelihood skills might turn out to conduct other non-fishing, money-making activities, rather than waiting for weather to return to normal.
Resiliency was found to be negatively related to negative feeling-their stronger resiliency lessens their negative feeling.A fisherman with strong resiliency is less likely to experience negative feelings as they are able to anticipate the changes and they know how to reduce the impacts [81].Furthermore, resiliency was also found to be positively related to life satisfaction-the stronger their resiliency, the more satisfied the fishermen are with their lives.Climate change impacts are among the major obstacles faced by fishermen, and being resilient to this major problem resolves many difficulties and is expected to drive them towards better life satisfaction [18].Nevertheless, despite the significant relationship, it should be noted that the predictive accuracy of resiliency on these factors was very low.It demonstrates a little influence of resilience towards climate change and concludes that resilience is not the major cause for fishermen's negative feelings and life satisfaction.

Limitation and Direction for Future Studies
This study has several limitations.Firstly, with regard to the geographical limitation of this study, as it was only conducted in Peninsular Malaysia, future studies should consider involving fishermen from Sabah and Sarawak.This is expected to enrich the data, as the socio-culture of Sabahan and Sarawakian are different compared to their peninsular counterparts.Secondly, this study focuses on fishermen in general, which means, regardless their category, either deep sea or small-scale fishermen can be chosen as the respondents.Future studies should specifically focus on small-scale fishermen, as Shaffril et al. [20] confirmed that, due to their characteristics (small vessel, small engine power, catching tools, etc.), this group is at great risk of being exposed to the severe climate change impacts.Thirdly, although resilience was proven to mediate all the developed hypotheses, it should be noted that resilience recorded substantial effects on positive feeling but merely recorded a weak predictive accuracy on life satisfaction and negative feeling.Hence, future scholars should consider several other additional mediating effects.The fishermen anxiety caused by bottom trawling activities and foreign fishermen intrusion should be one of them, as both can potentially affect their well-being.Bottom trawling is a fishing technique that causes excessive fish landing and the extinction of ocean resources.According to Mohd Ariff [84], fishermen expressed their concern on illegal trawlers and the intrusion by foreign fishermen, as these prohibited activities destroy marine flora and fauna, degrade the productivity and income, while at the same time increase the landing of 'trash fish'.In addition to these two, another mediating factor that might offer valuable information is the effects of migration of fisheries' resources associated with changes in temperature.For instance, Shaffril et al. [39] explained that a rising temperature can disrupt marine habitat and forces certain species to seek new areas.This might force fishermen to consume more time, fuel, and money to search for new catching zones, as well as reducing their catches and productivity.It might be interesting for future scholars to examine the impact of such situations on the fishermen's well-being.

Conclusions
Several climate change 'symptoms' have been detected in Malaysia.Rising temperature, sea level rise, unstable rain patterns, and the frequent occurrence of extreme events are recorded in several areas in Malaysia.Realizing this, the Malaysian government has initiated several mitigation strategies that aim to reduce greenhouse gas emissions and lessen the impacts of climate change, which include the establishment of the National Policy on Climate Change and The Renewable Act Energy and National Green Policy.Another initiative that can help to reduce the impacts of climate change, especially on the community, is by strengthening their resiliency.Resiliency can be understood as the ability of socio-ecological systems to cope and adapt with change, and having a stronger resiliency will assist the community to better prepare and enable them to design resourceprotection strategies that lessen current socio-economic impacts.Understandably, most of the existing climate change-related studies have not placed resiliency as their main focus, or simply focus on the direct effects of resiliency, either as an exogenous variable or endogenous variable, and this has geared this study to its main aim, which is to examine the mediating effect of resiliency towards climate change impacts of fishermen's subjective well-being.All twelve hypotheses were supported, and the mediation analysis confirmed the ability of resiliency to mediate the relationship between socio-economic aspects, sense of community, social relationship, and social environment relationship with life satisfaction, positive feeling, and negative feeling.Noticeably, resilience recorded a substantial effect on positive feeling.Although climate change impacts are expected to worsen in the future, climate-resilient fishermen are expected to stay positive and happy, as they know the things that they should do to lessen the impacts.The findings have also demonstrated that resilience recorded a weak predictive accuracy on life satisfaction and negative feeling.Such results indicate the little influence of resilience towards climate change, and determine that resilience is not the key cause for fishermen's negative feelings and life satisfaction.A number of recommendations are suggested for future scholars, including the involvement of fishermen from Sabah and Sarawak, taking into consideration the different of socio-culture of Sabahan and Sarawakian.Furthermore, more research focus on small-scale fishermen should be placed, as this group is at great risk of being exposed to the severe climate change impacts.Thirdly, to consider factors related to bottom trawling activities, foreign fishermen, and the migration of fisheries' resources, as these factors can potentially affect fishermen's well-being.

Table 1 .
Listed areas and the climate change symptoms at selected areas.

Table 6 .
The Structural Model.

Table 7 .
Hypothesis testing on mediation.