Understanding Factors Affecting Fishers’ Wellbeing in the U.S. Virgin Islands through the Lens of Heuristic Modelling

: Ongoing efforts to improve U.S. Caribbean fisheries management include increased consideration for human dimensions data and increased stakeholder input and engagement. Given the significant pressure that the local fisheries have sustained due to environmental degradation, climate change, storms and hurricanes, and overharvesting, combined with the critical data gaps that exist in both natural and human dimensions, it becomes particularly important to understand fishers’ perceptions and aspects influencing them to promote efforts that will maximize the wellbeing of these social-ecological systems. In this study, data collected through surveys with fishers in the U.S. Virgin Islands were used to develop a correlation model to test relationships between variables using a heuristic model, the Anthropic Impact Assessment Model (AIAM) as the basis. Findings support the application of heuristic models, such as the AIAM, to develop hypotheses and test relationships to understand complex fishery social-ecological systems. The most significant findings with implications for decision making in the region include support for considering fishers’ wellbeing as an indicator of ecosystem health and for using fishers’ local ecological knowledge in the management process, particularly under data-poor conditions, for information that can be used to better target outreach and education efforts, as well as more effective recovery plans to promote resilience and adaptation to environmental change, including the impacts of natural disasters. Results of this study and future analyses using similar approaches can be used to guide the incorporation of human dimensions data into the decision-making process in the U.S. Caribbean and elsewhere.


Introduction
Current trends in fisheries management and policy call for increased consideration of human dimensions data and input from stakeholders (Stephenson et al. 2016;Karp et al. 2023;Renck et al. 2023).In the U.S. Caribbean, efforts are underway to adopt ecosystembased approaches to improve fisheries management in Puerto Rico and the U.S. Virgin Islands (USVI) (NOAA 2019).NOAA Fisheries defines ecosystem-based fisheries management (EBFM) as "a systematic approach to fisheries management in a geographically specified area that contributes to the resilience and sustainability of the ecosystem; recognizes the physical, biological, economic, and social interactions among the affected fishery-related components of the ecosystem, including humans; and seeks to optimize benefits among a diverse set of societal goals" (NOAA 2016).Understanding and incorporating stakeholder perceptions in U.S. Caribbean fisheries management is particularly important given the complexity of the region's fishery ecosystem, including the large number of species managed (over 150 stocks) (Appeldoorn 2008), and the critical data gaps that exist in both the natural and human dimensions (Tzadik et al. 2021).In addition, the region's fisheries are under significant pressure from the impacts of rapid environmental change, including anthropogenic degradation and the effects of climate change.Climate change impacts in the region are directly related to the deterioration of essential fish habitat (e.g., coral reefs) as well as to impacts on fishing infrastructure (Short et al. 2016;Alongi 2015;Baker 2014;Randall and van Woesik 2015), particularly as a result of more frequent and intense storms and hurricanes.The uncertainty generated by climate change and compounded by other stressors (e.g., pollution, overharvesting) requires that decisions in fisheries policy and management are made in an even more timely and adaptive fashion.Therefore, developing and testing strategies to incorporate stakeholder perceptions data into the decision-making process are important to aid efforts that consider stakeholder input and priorities and thus have the highest potential to maximize the ability of fishing communities to adapt to stressors.Under more holistic and systemic approaches to natural resource management, humans are considered an integral part of the ecosystem, and their wellbeing is both influenced by and influences the wellbeing of other parts of the ecosystem.
Heuristic models can be useful tools to guide the analyses and interpretation of human dimensions data to inform decision making in fisheries management.These models can be used to guide the development of data collection efforts as well as provide a framework to analyze data collected using surveys and interviews in efforts to capture stakeholder perceptions, knowledge, and experiences that can inform and be incorporated into the decision-making process (Stephenson et al. 2016).In this study, the Anthropic Impact Assessment Model (AIAM), adapted from Pollnac et al. (2008), was used as the basis for examining relationships between human variables and their impact on individual and community wellbeing in the U.S. Virgin Islands fisheries from data collected through surveys with commercial fishers, i.e., fishers' perceptions of the fishery ecosystem.The AIAM model (Figure 1) depicts interrelationships between general categories of the human ecology of a fishery.The attributes listed in the boxes for each category are examples from among the many attributes included in each category.The arrows in the model reflect interrelationships (cause-effect and cumulative impacts) between these classes of variables.For instance, external forces or stressors (e.g., population pressure, fish stock levels, etc.) influence management strategies, which, in turn, influence human activities that relate to marine resources (e.g., fishing seasons, fishing gears, crew sizes, etc.).Changes in activities impact satisfaction with them, which influences aspects of individual (e.g., fishers) characteristics and the communities in which they live, as illustrated by the individual and social attributes in the model.The AIAM model identifies wellbeing as the dependent variable.
There has been a great deal of interest in developing measures of wellbeing.This study highlights the importance of wellbeing as a desirable measure to be used in fishery development and management.The AIAM model indicates that wellbeing is a desirable outcome that should be made an essential part of fisheries' anthropic (human) impact assessment.Pollnac and his colleagues have produced several papers illustrating the utility of the model and use a measure of wellbeing (Pollnac et al. 2015(Pollnac et al. , 2019;;Seara et al. 2017).Additionally, wellbeing is affected by a large number of sociocultural and economic variables that are impacted by management decisions, making it a suitable measure in this context (Pollnac et al. 2008).According to Helliwell et al. (2020), there has been an increase in the use of wellbeing as the measure of social progress and a goal of public policy.For example, they write that The Organization for Economic Cooperation and Development (OECD with 38 member countries) evaluates wellbeing in the process of developing policy standards with the goal of facilitating sustainable economic growth.The measure of wellbeing used in this paper has been adapted from the world happiness reports which have been produced annually since 2012.Individuals in the sample are asked on a scale from 1 to 10 how happy they are with their life.

U.S. Virgin Islands Fisheries
In the USVI, harvesting coral reef organisms is considered a traditional and culturally important activity and a food source that is considered a staple in Caribbean history (Kojis et al. 2017).Small scale, artisanal fishers target multiple species of coral reef fish, invertebrates, and coastal pelagic fish, with the highest-valued species including snapper (genus Lutjanus), parrotfish (family Scaridae), and spiny lobster (Panulirus argus).Approximately

U.S. Virgin Islands Fisheries
In the USVI, harvesting coral reef organisms is considered a traditional and culturally important activity and a food source that is considered a staple in Caribbean history (Kojis et al. 2017).Small scale, artisanal fishers target multiple species of coral reef fish, invertebrates, and coastal pelagic fish, with the highest-valued species including snapper (genus Lutjanus), parrotfish (family Scaridae), and spiny lobster (Panulirus argus).Approximately 200 fishers operate in the USVI (Yandle et al. 2020).The majority of these fishers are classified as commercial and divide their time between fishing, selling their catch, repairing their boats, and fixing gear (Kojis et al. 2017).
The USVI territory consists primarily of two large islands: St. Thomas and St. Croix as well as St. John and several other smaller islands (Figure 2).The islands differ in shelf geomorphology with St. Croix having a narrower and shallower shelf.These geological differences result in vast variability in both the occurrence and size of several commercially important reef fishes (Kadison et al. 2017), as well as differences in fishery attributes, including primary gear type.St. Thomas fishers predominantly use fish and lobster traps for harvest, whereas St. Croix fishers are more likely to use spear guns and SCUBA gear (Crosson and Hibbert 2017).Despite these significant differences, the fisheries had been long managed as one unit (Kadison et al. 2017).However, in 2012, fishery management authorities initiated a shift from a spatially undifferentiated management approach to an island-based approach, with separate stock reference points for Puerto Rico, St. Thomas/St.John, and St. Croix (NOAA 2019).This new approach recognized differences in the ecosystem-the economic, social, and cultural landscapes that shape fishing practices in the different U.S. Caribbean islands.This shift was also significant to establish the basis for the adoption of more ecosystem-based strategies to manage the local fisheries, and further emphasized the importance of considering human dimensions data in the decision-making process.
200 fishers operate in the USVI (Yandle et al. 2020).The majority of these fishers are classified as commercial and divide their time between fishing, selling their catch, repairing their boats, and fixing gear (Kojis et al. 2017).
The USVI territory consists primarily of two large islands: St. Thomas and St. Croix as well as St. John and several other smaller islands (Figure 2).The islands differ in shelf geomorphology with St. Croix having a narrower and shallower shelf.These geological differences result in vast variability in both the occurrence and size of several commercially important reef fishes (Kadison et al. 2017), as well as differences in fishery attributes, including primary gear type.St. Thomas fishers predominantly use fish and lobster traps for harvest, whereas St. Croix fishers are more likely to use spear guns and SCUBA gear (Crosson and Hibbert 2017).Despite these significant differences, the fisheries had been long managed as one unit (Kadison et al. 2017).However, in 2012, fishery management authorities initiated a shift from a spatially undifferentiated management approach to an island-based approach, with separate stock reference points for Puerto Rico, St. Thomas/St.John, and St. Croix (NOAA 2019).This new approach recognized differences in the ecosystem-the economic, social, and cultural landscapes that shape fishing practices in the different U.S. Caribbean islands.This shift was also significant to establish the basis for the adoption of more ecosystem-based strategies to manage the local fisheries, and further emphasized the importance of considering human dimensions data in the decision-making process.The survival and sustainability of the USVI fishery ecosystem is threatened by different impacts and their cumulative effects.Fishing pressure, along with point and non-point source pollution, and sedimentation are considered significant chronic stresses for USVI coral reefs and other marine ecosystems as well as the overall health of the fishery resources (Kojis et al. 2017;Yandle et al. 2020;Grace-McCaskey et al. 2023).These significant impacts are compounded by the effects of climate change, particularly the loss of critical habitat such as mangrove forests and seagrass beds (Short et al. 2016;Alongi 2015), coral bleaching and disease (Baker 2014;Randall and van Woesik 2015), changes in patterns of freshwater flows (Holding and Allen 2015), ocean acidification (Rhein et al. 2013), and intensified extreme weather events (Hayward and Joseph 2018).These environmental transformations affecting fishery resources severely affect a locally significant portion of the islands' residents who depend upon fisheries for income and food supply, as well as for cultural expression and psychological wellbeing.The development of strategies to manage sustainable fishery resources and resilient communities in the USVI is crucial and time sensitive due to the socio-economic and cultural importance of fisheries for this territory and the aforementioned environmental changes compounded by climate change, disasters, and other stressors.
The objectives of this study were to test relationships displayed in the AIAM heuristic model in Figure 1 using data collected from surveys with commercial fishers in the USVI to elucidate factors affecting fishers' wellbeing and aspects of their perceived vulnerability to disturbances such as hurricanes.These relationships were tested using a correlation model between variables derived from the survey instrument to provide an understanding of direct (proximal) and indirect (distal) factors affecting USVI fishers' wellbeing.In addition, a representation of all significant relationships between variables helps us understand factors affecting fishers' perceptions of risk, environmental beliefs, climate change, natural disasters, and management.This approach can provide valuable information for fisheries management and decision making as it provides a baseline for understanding changes in the human system, as well as insights into potential impacts of future management decisions from a stakeholder viewpoint.

Data Collection
From July through to December of 2021, a total of 48 face-to-face surveys were completed with commercial fishers in the USVI.In total, 21 surveys were conducted with fishers residing in St. Thomas/St.John, and 27 in St. Croix.There are about 150 active commercial fishers in the USVI (DPNR personal communication); thus, the sample represents approximately one third of the study population.An intercept sampling method was used, which consisted of approaching fishers at informal fish markets, fish houses, marinas, and piers where they are known to land their catch or congregate.This sampling technique was considered the most effective to maximize sampling of the studied universe since no comprehensive list or directory of USVI fishers is publicly available from which to draw a random sample.The questionnaires included questions on demographics, work and fishery attributes, wellbeing and job satisfaction, perceptions of the status of fishery resources and the environment, perceptions of climate change, and views on aspects of governance and participation in fisheries management.

Measures
Table 1 includes the variables that were used in the analyses and development of the model depicted in Figure 3 and their respective levels of measurement and survey questions used to collect the data.The following sections provide descriptions and methods used for developing all composite variables used in the model.

•
If the corals die it will make a difference for fishing.

•
Unless mangroves are protected we will not have any fish to catch.

•
There are so many fish in the ocean that no matter how many we catch, there will always be enough for our needs.(Reversed scale) Ordinal Scale: ( 5 Principal component analysis (PCA) with varimax rotation was used to develop scales based on interrelationships between the 7 items of fishers' environmental beliefs (See Table 1 for question and items).Two components were created and are identified as representing two important aspects of fishers' beliefs concerning causes of the beliefs evaluated: Human Causes and Natural Causes (Table 2).Component scores represent the level scored by each individual on each component.These scores are the sum of component coefficients times the sample standardized variables.Component coefficients are proportional to the component loadings.Hence, items with high positive loadings contribute more strongly to a positive component score than those with low or negative loadings.All items contribute (or subtract) from the score, and items with moderately high loadings on more than one component contribute at a moderate level, although differently, to the component scores associated with each of the components.This type of component score provides the best representation of the data.The scores are standardized with a mean of zero and a standard deviation of one.The total variance explained by the analysis is 57 percent.

Fishers' Environmental Beliefs
Principal component analysis (PCA) with varimax rotation was used to develop scales based on interrelationships between the 7 items of fishers' environmental beliefs (See Table 1 for question and items).Two components were created and are identified as representing two important aspects of fishers' beliefs concerning causes of the beliefs evaluated: Human Causes and Natural Causes (Table 2).Component scores represent the level scored by each individual on each component.These scores are the sum of component coefficients times the sample standardized variables.Component coefficients are proportional to the component loadings.Hence, items with high positive loadings contribute more strongly to a positive component score than those with low or negative loadings.All items contribute (or subtract) from the score, and items with moderately high loadings on more than one component contribute at a moderate level, although differently, to the component scores associated with each of the components.This type of component score provides the best representation of the data.The scores are standardized with a mean of zero and a standard deviation of one.The total variance explained by the analysis is 57 percent.Principal component analysis (PCA) with varimax rotation was also used to develop scales based on interrelationships between the 10 items of perceived environmental risk to fishing activity (See Table 1 for question and items).Three separate components of perceived risk were created: Risk 1: Extreme Weather, Risk 2: Human Effects, and Risk 3: Natural Effects (Table 3).The total variance explained by the analysis is 61 percent.

Job Satisfaction Measures
The 9-item scale used in the model is derived from a list of 22 indicators used by Poggie and Gersuny (1974, p. 56).This relatively long list of items was reduced in number (see Pollnac et al. 2015) by factor analyzing the items from a geographically diverse set of data and selecting items with the highest loadings on the three components, reducing the number of indicators for each component to the three manifesting the highest loadings on each component.Multiple correlations between the top three items and the factor scores for each component were high enough (R 2 = 0.79 and above) to accept them as reliable representatives of each component.Respondents were asked about their satisfaction level with each item on a scale ranging from 1 to 5 (very dissatisfied to very satisfied) (see Table 1).These values are summed for each job satisfaction component, Basic Needs, Social and Health Needs, and Self-Actualization, resulting in a scale that varies between 3 and 15, with a median of 9.

Heuristic Modeling
The main method used in this paper to analyze factors influencing the degree of recovery from environmental risks including ecosystem degradation, hurricanes and pandemics and their relationships with individual wellbeing is the development of a model that can stimulate further analyses.The model developed is a heuristic (discovery) causal model that simplifies the domain of interrelationships between the human and non-human environment in our analyses.There is a long history in the behavioral sciences of the use and testing of causal models using partial correlations, regression and path analysis (cf.Ullman 2007;Simon 1957;Blalock 1964Blalock , 1971;;Asher 1976;Wright 1960).Figure 1 in the introduction is a heuristic causal model which has stimulated further research where segments of the model have been tested by collecting new data and subjecting the data to a statistical testing method referred to as path analysis (cf.Pollnac et al. 2019;Seara et al. 2020a).The model was created by producing a correlation matrix of variables derived from past research on human impacts of environmental risks.The criteria for inclusion in the model are set.In this paper, the criteria are Spearman's Rho with a one-tail probability (direction of relationship predicted by the AIAM model) of less than 0.05.The ultimate dependent variable used in this study is wellbeing.Using the dependent variable as a starting point, the analysis identifies variables in the dataset meeting the criteria that are directly correlated with the selected variable.These are proximate variables.The next step identifies variables correlated with these proximate variables (distal variables).The process continues until no variables remain that meet the criteria set.

Findings and Discussion
The heuristic causal model seen in Figure 3 provides a visualization of the complex interrelationships between variables in the USVI fishery social-ecological system.The variables seen in Figure 3 can be characterized by the broad categories in the AIAM model (see Figure 1) and thus the model provides a good example of the useful application of heuristic modeling to analyze and investigate correlations and factors affecting fishers' wellbeing.In the next section, we present interpretations of select relationships displayed in Figure 3.The criteria for selection constitute all relationships directly affecting the ultimate dependent variable (wellbeing) and relationships that exemplify and illustrate the major topics covered in this study: risk factors, environmental beliefs and climate change, natural disasters, and management.

Factors Directly Affecting Fishers' Wellbeing
Proximate variables, i.e., those directly impacting the ultimate dependent variable "wellbeing" in the model (Figure 3), are the job satisfaction components Self-Actualization and Basic Needs, Months Fishing (the number of months out of a year one dedicates to fishing), and 10-Year Fishery Resource Improving, indicating fishers' perceptions of changes in fishery resources over the last 10 years.Correlations suggest that fishers with higher levels of satisfaction on the Self-Actualization and Basic Needs components of job satisfaction scored higher on the wellbeing scale.These findings support previous studies that show that aspects relating to the adventure and independence of the fishing job are important for maintaining fishers' psychological wellbeing, which is referred to in some of the literature as fulfilling the role of "therapy" (Griffith and Valdés-Pizzini 2002;Pollnac et al. 2011;Seara et al. 2017).The correlation found between wellbeing and the frequency of fishing also supports the notion of fishing as therapy, and it relates to previous findings that show that fishers manifest an expressive personality type for whom the act of fishing can fulfill aspects directly related to wellbeing (Pollnac and Poggie 2006).Earlier literature discussing the occupation of fishing as being therapeutic (e.g., Griffith and Valdés-Pizzini 2002;Pollnac et al. 2011) focused on one aspect of job satisfaction: Social and Psychological Needs.The therapeutic aspect of fishing, however, also includes the satisfaction of needs for adventure, risk and the thrill of the chase (Beauchaine and Gatzke-Kopp 2012;Pollnac andPoggie 2006, 2008;Pollnac et al. 2012), which are Self Actualization needs.In a study carried out by Griffith and Valdés-Pizzini (2002), Puerto Rico fishers mentioned that the therapy of fishing kept them away from vices such as drugs and young people away from delinquency and other anti-social behavioral attributes.Seara et al. (2022), in a study that investigated the long-term impacts of the American lobster die-off on the Long Island Sound fishing industry, also show that fishers reported negative social behaviors, including alcohol abuse and violence, as a consequence of their inability to continue fishing.It appears that without the therapy provided by fishing, fishers are likely to suffer from the consequences of job dissatisfaction which can range from psychosomatic illness, lowered longevity, and family violence, all of which are related to individual wellbeing (Pollnac et al. 2008).
Fishers with more positive perceptions of the status of fishery resources when compared to 10 years ago scored higher on the wellbeing scale, suggesting that fishers' wellbeing is directly linked to their perceptions of the health and abundance of fishery resources.This finding emphasizes the aspects discussed above since fishers' wellbeing is directly tied to their ability to continue fishing.Moreover, this provides an important perspective on the potential use of fishers' wellbeing as an indicator of fishery resource health, which can be useful in more holistic management strategies which seek to implement and evaluate decision making based on a social-ecological system perspective.

Factors Influencing Fishers' Perceptions of Risk and Environmental Beliefs
Heuristic models can also be used to explore complex multi-component concepts such as Perceived Environmental Risk to Fishing Activity and Fishers' Environmental Beliefs (see Sections 2.2.1 and 2.2.2) by examining each of the items reduced by PCA in terms of their relationships with other important variables in the model.
Variables related to the item Extreme Weather are concerned with climate change, level of perceived recovery from the hurricane, and self-actualization.Fishers presenting higher levels of concern over the Extreme Weather component also scored higher on the climate change concern scale, suggesting that fishers connect the threats of climate change to the environmental changes included in the risk component Extreme Weather (droughts, increase air and sea temperature, sea level rise, and storm frequency and severity).Those who reported lower levels of recovery from the most recent hurricanes also presented higher levels of concern over the Extreme Weather risk component, emphasizing previous observations that exposure to and experience of natural disasters affect people's perceived vulnerability to risk and change (Grothmann and Patt 2005;Smith and Clay 2010;Seara et al. 2016Seara et al. , 2020b)).This is also evidenced by the relationship in the model between the risk factor Human Effects and whether a fisher had had their fishing activity impeded or impacted by the hurricanes.Fishers with higher levels of satisfaction with Self-Actualization were also more concerned with risks to the fishery posed by the Extreme Weather risk factor.Previous research on job satisfaction among fishers stresses the importance of components related to adventure, challenge, and independence resulting in high attachment to the occupation (see Pollnac and Poggie 2006;Seara et al. 2017).These results may suggest that fishers with strong ties to fishing and who derive higher levels of self-actualization from it may be more aware and concerned about the impacts that extreme weather may have on the activity.Fishing in the US Caribbean is mostly artisanal, employing small vessels that lack the adequate technology to cope with storms while at sea.Therefore, an increase in the frequency and severity of hazardous weather prevents fishers from going out, affecting their opportunity for both monetary gain and self-actualization.
Fishers who have observed changes in the fishery resources and environment that they associate with climatic changes, as well as those who perceived resources to be in bad shape and worse today than they were 10 years ago, were more concerned about the Human Effects component of risk, which includes overfishing, pollution, and coral bleaching.This suggests that fishers' observations of changes in the environment and fish populations lead to increased concern over significant negative human impacts.
St. Thomas/St.John fishers were also more concerned about the Human Effects component of risk when compared to St. Croix.Previous research has shown significant differences in fish stock status between the islands, suggesting different degrees of overexploitation and environmental degradation (Kadison et al. 2017).In addition, St. Thomas has a large marine tourism industry, and perceived impacts of competing interests is often brought up by fishers during interviews, which may also affect their perceptions of the risks associated with human uses of the local marine environment.This is further evidenced by the higher levels of concern among St. Thomas/St.John fishers with the Human Causes component of the Fishers' Environmental Beliefs scale, which includes aspects of perceived vulnerability of fisheries to the effects of tourism, agriculture, houses, and industry.On the other hand, fishers in St. Croix scored higher on the item Natural Causes, relating to the importance of coral reefs and mangroves for fishing as well as the perception that fishing resources are subjected to overexploitation, suggesting that fishers in St. Croix are more concerned about the effects of habitat destruction and overfishing on their fishing resources.
Other factors affecting the perception of Human Effects on the marine environment included fishing frequency, household size and formal education.Fishers who were younger presented higher dependence on fishing, as well as those who scored higher on the climate change concern scale, who were more concerned about the Natural Effects risk component.Observation of these changes and the understanding that they pose a risk for fishing may lead to increased concern among fishers about environmental and climatic changes in general, which could result in further impacts on fishery.Fishers with higher level of dependence on the occupation are likely to have a closer relationship with the natural environment and the fishery resources and are potentially more aware of environmental changes and their direct and indirect impacts on fishery.Highly dependent fishers are also likely to express more concern about current environmental impacts or the potential effect on fishery, as these impacts represent a threat to their primary source of income and subsistence.While more data are needed to better understand these relationships, they point to the complex links between knowledge and resource dependency and fishers' perceived vulnerability to natural and anthropogenic impacts affecting their fishing activity.

Factors Influencing Fishers' Perceptions of Climate Change and Natural Disasters
Fishers who said climate change is affecting their fishing activity as well as those who perceived the fishery resources to be in worse shape when compared to ten years ago also scored higher on the Human Causes component of the Fisher's Environmental Beliefs scale, suggesting an important connection between fishers' awareness of climate change, resource degradation, and anthropogenic factors that affect fishery.This is also evidenced by the relationship found between concern about climate change and the Human Effects component of risk perception.Fishers who said climate change was affecting their fishery were also less satisfied with the Basic Needs component of job satisfaction, which includes aspects of satisfaction with income, suggesting an interesting relationship between perception of risk and vulnerability and financial security.This may also, in part, explain why fishers who did not receive hurricane assistance were also more likely to say climate change was affecting the fishery since the most significant form of assistance mentioned by fishers during the surveys was financial.
Other factors affecting fishers' perceptions of climate change affecting their fishing activity include the perception that resources are in worse shape when compared to ten years ago and more years of formal education, suggesting that both awareness of resource degradation and information exposure through education may increase awareness and perceived vulnerability to climate change.These results are in line with findings by Seara et al. (2020b), showing that fishers in Puerto Rico who either have more years of formal education or fishing experience tended to perceive fishery resources to be in worse shape more often.In addition, in the present study, a negative correlation was found between the climate change concern scale and the resource status variable, showing that fishers who are more concerned about climate change also tended to say fishing resources are in bad shape.Fishers who fish more frequently (more months/year) were also more concerned about climate change, further suggesting that exposure and dependency likely play a role in fishers' perceived vulnerability to climate and environmental change.
Fishers who received both hurricane and COVID-19 assistance were more satisfied with the Social and Health component of job satisfaction, suggesting a relationship between receiving aid and improved sense of physical and mental health.Fishers who received hurricane assistance also scored higher on the Basic Needs component of job satisfaction.As stated previously, the most important source of hurricane assistance mentioned by fishers was government financial aid, and the fact that these fishers were more satisfied with Basic Needs show the importance of disaster financial assistance.Those who said they received COVID-19 assistance were also more likely to say that fishers and the community work well together, indicating that social networks and capital likely played a role in recovery from the pandemic.

Factors Influencing Fisher's Perceptions on Governance/Management
Fishers who fished more often (more months/year) scored higher on levels of cooperation among fishers and tended to say that fishers' compliance with rules and regulations was higher, indicating a potential heightened sense of community and commitment by those who fish more frequently and likely present higher levels of dependency and attachment to the occupation.Fishers who expressed more concern over the Natural Causes component of the Fisher's Environmental Beliefs scale scored lower on the compliance variable, suggesting that higher awareness of the effects that the health of coral reefs and mangroves and overexploitation can have on the fisheries is linked to perceptions of low compliance.Those who perceived compliance to be low shared the opinion that more enforcement of rules and regulations is needed.Perceptions that there is a need for more enforcement were also correlated with a more negative view of the health of fishery resources.These results indicate that fishers who express concerns over the health of the fishery ecosystem, including the status of fishery resources, are also concerned about compliance and support more enforcement of rules and regulations.
Results regarding fishers' perceptions of fisheries governance and management showed some significant differences between the different island groups.Fishers in St Croix tended to perceive management as being fair and tended to say fishers and the government work well together more often than their counterparts in St Thomas/St John.It is interesting to note that fishers in St Croix presented higher levels of dependence on fishing when compared to St Thomas/St John.It is difficult to pinpoint specific reasons behind the differences found between the island groups; however, these findings further support distinctions between islands and island groups for fisheries management purposes (e.g., island-based management), which better account for environmental, social, and cultural differences between locations.

Conclusions
Analysis and findings of this study support the application of heuristic models, such as the AIAM, to develop hypotheses and test relationships to understand complex fishery social-ecological systems.In this paper, the USVI fishery ecosystem was investigated and correlations between important factors affecting fishers' wellbeing were explored.While the relationships between variables' categories in the USVI causal model did not precisely match the AIAM, its overall structure and flow is similar.Importantly, direct and indirect relationships between perceptions of environmental change, risks, management and wellbeing are found in the USVI correlation model and provide support for the AIAM.
The findings of this study elucidate important relationships between factors affecting fishers and fishing communities.Particularly in the U.S. Caribbean, where EBFM approaches are currently in their development stage (NOAA 2019), further understanding of the human dimensions system from a stakeholder viewpoint will contribute to a more holistic conceptualization of the ecosystem (see Seara et al. 2024).Additionally, the model constructed here can help identify potential indicators of ecosystem health and risk that take into consideration the wellbeing of fishing communities.
The relationships depicted in the USVI correlations model are complex and reveal the extent of information that can be explored in the context of this fishery social-ecological system in future studies.While not an exhaustive list, the points below summarize some of the most significant findings, particularly those that are relevant for decision making in the context of the U.S. Caribbean fisheries, considering current efforts to adopt ecosystembased strategies: • The relationships found between fishers' perceptions of the health of fishery resources and their wellbeing and job satisfaction further support the importance of considering fishers' wellbeing as an important indicator of ecosystem health.Further understanding these relationships can help to support EBFM development in the region by including monitoring and incorporation of human dimension variables that align with the needs and interests of fishers, thus increasing stakeholder buy in (Karp et al. 2023).

•
Fishers' knowledge and awareness of the impacts of environmental and climate change on fishery resources, including that of storms and hurricanes, as clearly evidenced by the correlations found in the model, provide evidence for the importance of local ecological knowledge (LEK) that can be used to complement management and decision making.Under EBFM, the need to understand and conceptualize ecosystem-level relationships further unmasks scientific data limitations; thus, LEK can be an important source of information for decision making, particularly in data-poor conditions such as that of the U.S. Caribbean region (Appeldoorn 2008).

•
Higher fishing frequency, used here as a proxy for dependency, and higher levels of education were found to be associated with heightened environmental and climate awareness, suggesting that different aspects of experience and exposure to information affect perceptions of ecosystem change.In addition, fishers who expressed more concern over environmental changes were also more likely to support the need for more enforcement and compliance with rules and regulations to manage fishery resources.This information can be used to support and develop outreach and education strategies in collaboration with fishers and fishing communities.This can be particularly useful at the initial stages of EBFM development to increase understanding and support for new management approaches.

•
Relationships found between job satisfaction variables and hurricane assistance, as well as community cohesiveness perceptions and COVID-19 assistance, point to the importance of developing effective, timely, and collaborative assistance programs to aid fishers and fishing communities in the aftermath of disaster events, as well as to promote long-term resilience and adaptation.
Findings of this research and future analysis using similar approaches can provide invaluable information, including a baseline, to guide the incorporation of human dimensions data into the decision-making process in the U.S. Caribbean and in other regions.

Figure 3 .
Figure 3. Causal model for the U.S. Virgin Islands fisheries.

Figure 3 .
Figure 3. Causal model for the U.S. Virgin Islands fisheries.

Table 1 .
Variables used in the analyses and their respective levels of measurement and survey questions.

Table 2 .
Results of Principal Component Analysis (varimax rotation) with the 7 items of fishers' environmental beliefs.Bolded font indicates components with highest factor scores for each item.

Table 3 .
Results of Principal Component Analysis (varimax rotation) with the 10 items of perceived environmental risk to fishing activity.Bolded font indicates components with highest factor scores for each item.