Classifying fisher behaviour in the Netherlands : replicating the fishing styles method to understand social motivations and explanations for activity in the Dutch fleet

(1)

Diversity in the Dutch fishing fleet: Small, medium and large scale vessels in the same harbour (photo: Arie Mol)

Supervisor: Dr. Joeri Scholtens Second Reader: Javier Garcia-Bernado

Word Count: 17,200 excluding references and Appendix Date of Submission: August 16, 2018

Student Number: 11204001

Contact: amanda.schadeberg@gmail.com

Around 13,800 excluding references and Appendix

Date of Submission:

Student Number: 11204001

CLASSIFYING FISHER BEHAVIOUR IN THE NETHERLANDS:

Replicating the fishing styles method to understand social

motivations and explanations for activity in the Dutch fleet

By Amanda Schadeberg

Thesis submitted in partial fulfilment of the requirements of the

Research Master Social Science (MSc)

(2)

(3)

Abstract

Existing fisheries research and management policies tend to rely on the assumption that fishers make rational decisions based on utility maximisation, yet social science research has shown that this is often not the case. A solution to understanding the complex motivations of fisher behaviour in a way that balances parsimony and complexity was suggested by Boonstra and Hentati-Sundberg in their 2016 paper: ‘Classifying fishers’ behaviour: An invitation to fishing styles’. The fishing styles approach aims to achieve a better understanding of fisher behaviour by combining quantitative and qualitative methods to investigate motivations and the influence of social contexts on fisher behaviour. This thesis thus accepts Boonstra and Hentati-Sundberg’s (2016) invitation to the fishing styles concept for studying fisher behaviour, with the aim to critically replicate their classification methodology using data from the Dutch fleet. The study begins with factorial analysis of logbook data from 2001-2016 to quantitatively classify fishing activity based on species landings composition. These clusters are combined with effort data to produce a preliminary list of fishing practices. This list is then validated, reinterpreted and enriched through two phases of qualitative research: panel discussion with experts and interviews with fishers themselves. This qualitative phase of the research helps to explain the intentions of fishers and investigate the role of contextual influences as well as to enrich and transform the preliminary list of quantitatively-derived practices into a list of qualitatively-explored fishing styles. The results are 15 fishing practices based on what species fishermen bring to land and two preliminary fishing styles, based on interviews with fishermen from the demersal trawling segment of the fleet. The paper also demonstrates the value of replicating existing papers for scientific advancement and evaluates the robustness of the fishing styles concept and method.

(5)

1. Introduction

Ensuring the sustainable use of marine resources is a challenge that involves understanding the complex interactions between actors at many levels, from nation states to local communities, from individual fishermen1_{to the environment itself. While environmental management has been a component of human} societies for thousands of years (Rackham, 1980; Gammage, 2008; Pascoe, 2014), the industrial revolution brought with it an acceleration in human domination of Earth’s ecosystems, and with it environmental degradation as demand for materials and human populations simultaneously increased (Vitousek et al., 1997). This degradation was met with more complex and formalised management strategies, with the end of the 20th century seeing an explosion in environmental regulation and unprecedented international cooperation (Hahn & Richards, 1989). Maritime activity was no exception to the dramatic changes in environmental regulation. In Europe, the League of Nations began working on a convention regulating whaling in 1931 (Gambell, 1993) and the European Union (EU) has been working on cooperative fisheries policies since the Treaty of Rome in 1957 (European Parliament, 2018). The ratification of the United Nations Convention on the Law of the Sea in 1982 granted all nation states rights to an exclusive economic zone of 200 nautical miles (370km) from their coast (United Nations, 1998). This evolution in the way humans govern and interact with the marine environment at a local, national, and international level has led to complex questions about the way activities, resources, and users are classified.

Managing the sustainability of EU fishing activity has been dominated by a focus on the resource: the fish. Single-species stock management in the form of total allowable catches (TACs) has been the foundation of the EU’s Common Fisheries Policy (CFP) since the 1980s (Ulrich et al., 2012). There is, however, growing interest in using a fleet-based approach as a complement to TACs. Ulrich et al. (2012) have defined the two key concepts in this fleet-based approach: the fleet and the métier. The fleet refers to “a group of vessels with the same length class and predominant fishing gear” and a métier is “a group of fishing operations targeting a similar (assemblage of) species, using similar gear, during the same period of the year and/or within the same area and which are characterized by a similar exploitation pattern” (Ulrich et al., 2012: 39). The fleet is thus used to refer to the vessels, and métier is used to refer to their activity. This approach is a step towards a more behaviour-focused understanding of fisheries, which is necessary to measuring and understanding degrees of heterogeneity in a fishery.

However, métier analysis is conducted based on vessel and activity characteristics, and does not include information about the people who are fishing. Métier analysis “does not explain which ecological, social, economic or political factors cause fishers to maintain or change their ways of fishing” (Boonstra & Hentati-Sundberg, 2016: 81) and thus merely postulates about intention. In lieu of information about the reasons for engaging in certain activities, anyone interested in the behaviour of fishermen must work backwards from the measured outcomes of the activity and assume that fishermen are acting strategically to achieve what is represented in the data. Further, métiers are determined at the point of data processing and are thus subject to change over time under the influence of subjective interpretations of protocols. Worryingly, “no unified methods have yet been agreed upon for the standard scientific definition of fleets and métiers, despite significant activity in this field” (Ulrich et al., 2012: 41). As a result, even when minor fleets and gears are

1_{A note on terminology: While it has become common to use the gender-neutral term “fisher” in scholarly articles about this}

subject, the reality for this research was that each respondent who was currently or formerly engaged in fishing was male. While there are many gendered dimensions of fishing activity both on board and in the ancillary economy, to avoid obfuscation in this thesis I will simply use “fisherman”/“fishermen”. The exception is for the term ‘fisher behaviour’, which is a concept defined by Boonstra & Hentati-Sundberg (2016) and which I do not wish to alter.

(6)

aggregated, the International Council for the Exploration of the Seas (ICES) determined that there were 72 different fleet, gear, and métier groupings (Ulrich et al., 2012: 39), which may vary in characteristics from place to place.

To address the shortcomings of the métier approach to understanding fisher behaviour, Boonstra & Hentati-Sundberg (2016) have introduced the concept of fishing styles, which aims to study not only what people do, but how they interpret their doings. Fishing styles often run counter to the maximisation of utility, meaning that they go beyond the narrow focus of means-end motivations to incorporate non-strategic factors such as habit, affect, and values. Fishing styles builds on the current approach of métiers to add the missing dimensions of intention and non-strategic factors. These culturally embedded styles are ideal-typical ways of understanding groups of fishermen in order to satisfy scientific demands. The method tested in this thesis results in comprehensive explanation of activity via a usefully small number of categories.

An essential part of the activity of science is categorisation (Johnson, 2006). Understanding fisher behaviour is a complex endeavour, but some simplification of reality is necessary for governance to be possible (Scott, 1998). Boonstra and Hentati-Sundberg (2016) acknowledge this, writing that a balance between parsimony and complexity is needed in policy formation and evaluation. In the current context of management and regulation, policy-oriented scientists must necessarily condense reality into manageable archetypes or examples so that policies can be designed based on the attributes, needs and behaviours of these groups. However, a challenge lies in identifying how many archetypes are necessary to adequately represent the study population. Too many and the policy response becomes impractical, too few and the policy is ineffective or has unintended negative consequences for unidentified subgroups. Further, how and on what criteria the archetypes are created is also a value-laden process. Fishing styles are an attempt to balance the inherent complexity of reality with policy’s need for simplicity.

The Netherlands is an interesting case for replicating the fishing styles classification method for several reasons. First, it is part of the EU, where data quality and availability is very good compared to the rest of the world due to strict reporting requirements of current management policies such as the CFP. Second, the Netherlands in particular is a topical case within the EU because it has undergone substantial changes in response to technological developments, market conditions and changes to regulations (Haasnoot, Kraan & Bush, 2016). Further, the Dutch fleet accounts for approximately 1% of the EU effort by number of vessels (European Commission, 2016). This

relatively low number of vessels represents 8.3% of EU fishing effort by vessel tonnage, suggesting that, unlike the case of southern Sweden selected by Boonstra & Hentati-Sundberg (2016), the Dutch fleet includes some heavily industrialised, high-yield vessels. In the Netherlands, more than 2,000 people (full time equivalent) are directly employed in fisheries, with another 3,500 employed in processing (European Commission, 2016). As shown in Figure 1, there is also great variability in the total catch per year by Dutch vessels.

(7)

This thesis is guided by the following research questions: 1. What are the fishing practices in the Dutch fishing fleet?

2. What are the main fishing styles in the Dutch demersal fleet and what distinguishes them?

3. Can the method and concepts proposed by Boonstra and Hentati-Sundberg (2016) be replicated in the Netherlands and are they useful for future fisheries research?

These questions will be answered using a critical replication of the mixed-methods methodology proposed by Boonstra & Hentati-Sundberg (2016). The aim is to replicate the fishing styles classification method while remaining critical of the qualitative and at times arbitrary methodological decisions made in the process and making necessary changes in the process. The scope of the quantitative analysis is the activity of all registered vessels in the Dutch fishing fleet in the years 2001-2016. The method begins with several factorial analyses of the logbook data to create a preliminary list of fishing practices. This list is then validated and enriched by two phases of qualitative research: consultation with experts in Dutch fisheries science, regulation and industry, and interviews with fishers themselves. The results are a set of quantitatively grounded fishing practices that are mutually exclusive and exhaustive, and a smaller set of thickly described fishing styles, which explain the motivation for fishermen using one or more of the practices.

This paper responds to a call from the leading body of marine science in Europe (ICES) for a better understanding of fisher behaviour. In doing so, it makes two main contributions to fisheries management knowledge: It is the first empirical analysis of the activity of the Dutch fleet of this scale that enriches quantitative results with an understanding of behaviour in qualitative terms. Second, it is an evaluation of a new methodology proposed as an alternative to the dominant métier approach to fisheries classification. The quantitative results shed new light on the divisions within the Dutch fleet, employing computational power to reveal behaviours and trends that are not visible or are under-examined in current approaches. The qualitative results allow fishermen to speak for themselves about their motivations for certain behaviours. Further, as a replication of a proposed new method for fisheries management social science this paper evaluates the process and concepts used and contributes to broader methodological debates about how to classify the behaviour of humans who interact with and use the environment.

The following section contains a theoretical framework and literature review. It will discuss the main theories, concepts, and variables related to this research. Section 3 is a detailed methodology which first describes the value of the replication process and then outlines the steps taken in the quantitative and qualitative phases of the project. Section 4 presents the results of each phase of the research from the fishing practices to the fishing styles. Section 5 offers the conclusions of the study before Section 6 critically discusses the substantive results by examining the value, validity, and utility of the fishing styles approach. This section also evaluates the replication process by discussing the contribution of this replication to science and the practical value of the results in the Dutch context. The Appendix contains additional results and methodological material.

(8)

2. Theoretical Framework and Literature Review

Managing Fisheries

A key assumption underlying the work of making, enforcing, and evaluating environmental policy is that humankind can manage nature. This process begins with the transformation of ‘nature’ into ‘natural resources’, where elements are categorised according to their utility for human beings (Scott, 1998: 13). For instance, plants that are valued for food or medicine become crops, while plants that are not deemed useful become weeds. In a similar way, species of marine life that are valued for consumption become catch or yield, while marine animals that are not valued become bycatch. Because these human-made decisions about value have impacts on the environment, managing the use of natural resources is not a question of managing nature but rather of managing the humans who act upon their environment to derive value (both social and economic) from their surroundings (Hilborn, 2007). This subsection touches on some key features of the current fisheries management approach, defined here as one that views a fishery as a common pool resource accessed by multiple actors who can (and should) be constrained in their activity. This constraint may come in the form of regulation and privatisation, as famously popularised by Hardin (1968) in the tragedy of the commons, or constraints may come about in the form of complex social practices and contracts drawn up by the users of resources acting as self-governing bodies (Ostrom, 1990: 17). The following subsections aim to provide a broad theoretical context for the development of fisheries management strategies by introducing and exploring key concepts and features. It is first necessary to understand marine ecosystems as a common pool resource, governed by humans. The next subsection deals with a discursive approach to understanding common pool resources known as ecosystems services. Once a common pool resource is understood as a service provider, certain institutions come about to manage the resource, such as the EU’s CFP, which is the subject of the next subsection. With these broader concepts in mind, the case is then made for a sociological mixed methods approach to understanding fisher behaviour under the CFP. This chapter concludes with an examination of the nature and consequences of classifying complex social behaviour before the concept of fishing styles, as employed by Boonstra & Hentati-Sundberg (2016), is defined.

Governing common pool resources

Common pool resources are defined as resource systems (regardless of whether they have property rights attached) characterised by difficulty of exclusion and subtractability (Ostrom et al., 1999: 278). Examples include lakes, air, forests, and seas. Difficulty of exclusion refers to the cost of excluding those who can benefit from using the resource through physical, legal, or institutional boundaries. Subtractability refers to the reduction of the resource through exploitation. Marine ecosystems are a complex case of a common pool resource because it is very difficult to force exclusion due to their size and lack of physical infrastructure (compared to lakes that have finite shores, or forests where ownership of trees can be claimed by building fences). Although it is a great challenge to govern vast common pool resources such as marine ecosystems, “sustained research coupled to an explicit view of national and international policies as experiments can yield the scientific knowledge necessary to design appropriate adaptive institutions” (Dietz et al., 2003: 1910). These adaptive institutions can take various forms such as privatisation, government regulation, or formal and informal co-management. The question of which adaptive institution is best for governing marine ecosystems as common pool resources remains an object of debate. In Europe, the current approach is manifested in the CFP.

(9)

The European Common Fisheries Policy

With the challenge of governing common pool resources amplified by the complexity of marine ecosystems, the contemporary response to managing fisheries in the EU is the CFP. The policy was proposed as part of the common agricultural policy at the constitutional foundation of the EU in 1957 (European Parliament, 2018). Biological conservatory management measures were introduced in 1983, comprised of TACs and quotas in an attempt to cap and then allocate the total amount of fish that could be brought to land. Regulation in 1992 introduced a licensing system to address the imbalance between fleet capacity and catch potential by incentivising the reduction of fishing effort. When this proved ineffective to prevent overfishing, a reform of the regulations was introduced in 2002. These regulations aimed to 1) improve conservation and encourage the sustainable exploitation of fish stocks, 2) give detailed rules about EU structural assistance to the sector, and 3) establish a measure that encouraged the scrapping of vessels to reduce capacity (European Parliament, 2018). The 2002 reforms also introduced a more cooperative approach to management, including fishermen and others related to the industry in the form of regional advisory councils.

The most recent iteration of the CFP was implemented on January 1, 2014. While some national jurisdiction over coastal and inland waters remains, the CFP currently governs fisheries in the Netherlands via rules to manage and conserve fish stocks, facilitating the sustainable exploitation of the natural resource through equal access and competition (European Commission, 2018). Key features of the present CFP that are relevant to this study include maximum sustainable yield limits, changes to fleet capacity that favour small-scale operations, a decentralised approach to governance that allows member states to determine implementation, more data collection and sharing to encourage scientific assessments, and the implementation of discard bans by 2019 (European Parliament, 2018). The increased stakeholder participation that began in the 2002 reforms continues today.

Scholars have identified several shortcomings of the CFP. From a conservationist’s perspective, Khalilian et al. (2010: 1179) have argued that TACs calculated by scientists are ignored in favour of socio-economic requirements. To this end, the capacity of the fleet is artificially maintained by subsidies (Khalilian et al., 2010: 1180). Further, Griffin (2009) has argued that adjudicating between competing knowledges is power-laden. Until recently, the scientists that inform the makers of the CFP have mostly neglected meaningful dialogue with local resource users in what Griffin has called the “hegemony of science” (2009: 562). This hegemonic position of natural scientists in the formulation of the CFP means that non-scientific ways of thinking are overlooked, often at the expense of understanding local- and temporal-specific practices and phenomena. This critique relates to a broader conversation about the “politics of measurement” (Scott, 1998). Scott has argued for the value of non-scientific and non-standardised ways of measuring with the example of a piece of land. When asked about local rain conditions, a scientist may give a figure of the rainfall in that area of land in millimetres per year. A non-scientific person may describe the land as ‘well-watered’. If the purpose of the land is to bear crops, knowing that it receives adequate water to sustain plants in an apparently vague phrase such as ‘well-watered’ may be more useful than an absolute figure of rainfall, which does not incorporate information about the distribution of that rain or the land’s ability to retain the water and transfer it usefully to crops.

Ecosystems services

There are thus various systems of knowledge, symbols, meanings and values both formed and embedded in social interactions with nature (van Ginkel, 2010: 21). The way marine common pool resources are governed depends on the way they are defined, valued, and classified and these choices have consequences for the natural resource and for the people who use or depend upon it. Many traditional indigenous peoples have complex knowledge-practice-belief systems that prioritise ecosystem stability via religious and moral systems

(10)

(Gadgil, Berkes & Folke, 1993). Therefore, the anthropocentric, predominantly Western concept of ecosystem services should be seen as a response to the predominantly Western activity of modern industrialised agriculture which is, according to Scott (1998: 2), “a radical reorganization and simplification [...] to suit man's goals”.

Kull et al. (2015) have argued that the framing, ontology and use of the concept of ecosystem services have historical roots and political outcomes. As one such political outcome, the CFP is a form of governance that reflects the values in the ecosystems services understanding of the environment. The ecosystem services concept gained prominence in the early 2000s (Millennium Ecosystem Assessment, 2005), alongside the growth in popularity of two important trends in environmental politics: (neo)liberal solutions to regulatory failures, also known as ‘market solutions’; and ecological modernisation, which positions technological and scientific innovation as the most viable solution to environmental crises, more so than changing consumption or production patterns. Seeing common pool resources, such as the ocean, as ecosystem services also creates ontological boundaries. Some critical questions about the ecosystems services approach to understanding common resources are,

 What is the value of the ecosystem beyond the services it renders to humans?  Who can access the ecosystem and its services and how?

 Who are the gatekeepers of access to this ecosystem and its services?

As the questions above suggest, the ecosystem services concept has consequences in its use, with the potential to benefit some groups or outcomes more than others. In the case selected for this thesis, the ecosystem service is the fishery in the North Sea (and occasionally international waters beyond), as it has been accessed by Dutch (and other) fishermen in their diverse vessels. Therefore, to adapt the definition of Kull et al. (2015), an ecosystems services understanding of the case is as follows,

The North Sea (and waters beyond) provides fish which is useful to fishermen as a commodity and to other people as food and this should be valued in monetary terms.

However, there are several problematic elements in this definition that relate to the limitations of understanding the sea as an ecosystem service. First, common pool resources are often valued for more than their monetary worth. They also provide livelihoods, dignity, purpose, pride, and cultural and spiritual connections. Second, as mentioned above, the boundaries that an ecosystem services understanding of a fishery creates mean that it is not clear what the value of the ecosystem is beyond its service to human beings. For instance, the fisheries management discourse uses stock assessments to determine how many fish are available for human exploitation, but less attention is given to the ecosystem’s role in feeding non-human animals, or even to the intrinsic value of the ecosystem. Third, the definition is insensitive to the question of which humans are entitled to benefit from the ecosystem services. Under the CFP, historical access is seen as legitimate reason to grant or deny access to the fish in European waters, but this creates boundaries for others. There are also technical and economic boundaries that exclude some humans from accessing the resource. Finally, once the common pool resource has been discursively constructed as an ecosystem service, certain adaptive institutions (such as the CFP) become viewed as ‘more appropriate’ than others, sometimes in contradiction to local relationships with the resource. Take as an example private property regimes being applied by the government to Maori fisheries in New Zealand, despite Maori culture viewing belonging and ownership as “a two-way affair” (McCormack, 2010: 21), where the sea is man as much as the man is the sea. This paper aims to move beyond an ecosystem services understanding of the marine environment by incorporating the ways fishermen value the sea beyond its ability to provide fish and financial returns.

(11)

Understanding fisher behaviour

Experimental evidence shows that humans are not good at making rational choices, especially when faced with complex information (Tversky & Kahneman 1971, 1974, 1992) or circumstantial constraints (Camerer et al., 1997). Fishermen are constrained in their decision-making by regulations and the environmental conditions in which they work. Within these boundaries lie many possible strategic decisions which, depending on the intentions and desired outcomes of the fisherman, can be taken to adapt to the resource, environment, regulations, or market (Salas & Gaertner, 2004). These strategic decisions include long-term choices such as which species to target, which gear to use, and which market to sell the fish in (i.e., free market or contract), as well as shorter-term tactical choices such as which grounds to frequent, whether to discard some of the catch, and when to return to the harbour. Van Putten et al. (2012: 5) have also argued that the motivations for fisher decision-making vary based on seen and unseen drivers and rules, which “may be based on profit maximisation or may include a number of other drivers or behavioural responses”.

Therefore, while rational choices within boundaries and profit-seeking can account for much of fisher behaviour, it remains the main source of uncertainty in fisheries management (Wilen et al., 2002; Fulton et al., 2011; Simons, Döring & Temming, 2015). An important concept for the study of fisher behaviour is what Boonstra and Hentati-Sundberg (2016) refer to as ‘non-strategic factors’. These are the dynamic, ‘irrational’, and thus unpredictable behaviours that are missing from utility maximisation explanations of fisher behaviour. Possible non-strategic behaviours include targeting a species based on historical practice rather than value and abundance, selling the catch to processors who may be relatives or friends despite non-competitive prices, frequenting grounds that have been fished by ancestors, or returning to harbour early on Friday not because the vessel is full but in order to spend time with family during the weekend. Finally, paying attention to non-strategic factors is particularly pertinent for understanding fishing decisions because “fishermen may not have sufficient information, or predilection, to effectively evaluate the expected value of the alternatives they face” (Holland, 2008: 328). It is thus important to ask one what basis they make decisions, and why these motivating factors are deemed important when other alternatives are available. If an understanding of fisher behaviour is based solely on the outcomes of behavioural processes (such as logbook data), the wrong conclusions may be made about the decisions that lead to those outcomes.

Many of the aforementioned non-strategic motivations for fisher behaviour are related to social norms. In his book Predictably Irrational, Ariely (2008: 88) argues that “social norms play a far greater role in society than we have been giving them credit for”. They are perhaps missing from many studies because they are difficult to see and to study. Social norms are often implicit and appear to the individual to be self-evident, yet they can have large consequences for societies and environments alike. Van Ginkel (2010: 16) has observed that fishermen “are embedded structurally in larger social configurations that act upon them”, and that understanding norms is necessary to understand social adaptations to ecological systems. As for the usefulness of understanding these social factors, Ariely (2008) has argued that, when it comes to changing human behaviour, social norms are not only cheaper than financial incentives, they are often more effective (Ariely, 2008: 86). As for understanding fisher behaviour post-hoc, inherent bias (an aversion to or preference for a particular decision based on unacknowledged influences) is often overlooked as an explanation (Ariely, 2008). Inherent bias can exist despite vast experience, and it is formed by (and therefore serves) factors beyond the outcome being measured. Inherent bias makes an individual behave in a way that may seem irrational to an observer, but which is justified by the individual. For example, a fisherman may protest or disobey rules that seem unjust to him, risking fines or exclusion from the fishing grounds. At the collective level, Gezelius (2002) found that the legitimacy of the law in a fishing group in Norway was contingent on the moral ideals of the community. In other words, the community enforced rules based on moral rather than regulatory standards.

(12)

When studying behaviour, it is important to remember that fishermen in the EU, and even within the Netherlands, cannot be seen as a homogenous, predictable group. Van Putten et al. (2012: 5) have argued that “fishers do not all respond [to incentives] in a similar manner, as they are faced with different levels of information, different information-processing abilities and different constraints (social and/or economic).” Van Ginkel (2010) has also observed that fishermen’s responses to rules and standards vary, and that fishermen are not passive recipients of regulation. Rather, they also “actively act upon their natural and social worlds to create new opportunities and new meanings” (van Ginkel, 2010: 16). Although van Putten et al. (2012) and van Ginkel (2010) work within quite different academic disciplines and methods, they agree that a policy-relevant understanding of marine ecosystems can only come about when proper research attention is paid to the resource users (i.e. the fishermen).

Categorisation

While it has been established that fisher behaviour is a complex and dynamic subject of study, policy can neither be made nor enforced at the individual level. Thus it is necessary to simplify the complex social reality of the activity of fishermen into useful categories. Johnson (2006) writes that the paradigmatic activity of science is categorisation, which can be useful as frames of reference, but which come with consequences. Categorisation creates what Scott (1998: 11) refers to as tunnel vision, which “brings into sharp focus certain limited aspects of an otherwise far more complex and unwieldy reality”. This is a useful pursuit, as it renders the “phenomenon at the centre of the field of vision more legible and hence more susceptible to careful measurement and calculation” (Scott, 1998: 11). This practice of narrowing the field of vision is useful in many empirical pursuits, such as biological taxonomy, biomedical research, and computer science. Some have even gone as far as to argue that all human cognition is for, and about, is the categorisation of the world around us so that we can “do the right thing with the right kind of thing” (Harnard, 2017).

Scott continues by asserting that “no administrative system is capable of representing any existing social community except through a heroic and greatly schematised process of abstraction and simplification” (1998: 22). It is always important to consider who or what the abstraction and simplification is useful for, how it is schematised, and what the consequences of making such divisions are. Categories have lists of variables that attempt to empirically assess the group, but they also create and respond to dominant narratives. The consequences of categorisation can be as globally significant as the effect of Orientalism (Said, 1978), where through clichés and stereotypes of the East as weak, passive, and mystic, the West came to define itself as strong, privileged, and rational. These characteristics were then deployed in narratives justifying colonialism and abuses of power. In the context of fisheries, Rita (1991) documents a dispute between fishermen and the IRS in Massachusetts in the US. Vessel owners wanted crew to be classified as contractors so they would be responsible for paying their own taxes, while the IRS preferred crew to be classified as employees, placing the burden of taxation on the vessel owners. The complications of this decision would have severe financial consequences and affect the viability of the industry in the future. Johnson’s examination of small-scale fisheries as a category reveals that “they can only be identified in relational terms, which creates a constant impression of elusiveness and categorical imprecision” (2006: 751). Thus to codify, categorise, or classify is a political process with real consequences and whose results require interpretation.

The quantitative technique used in this study, namely principal component analysis, has traditionally been used by biologists to determine the number of discrete species present in a sample, or to differentiate between genes within a species. In the first example from social science, Bourdieu (1979) used correspondence analysis to classify social behaviour as a novel and more empirical way of measuring the way ‘taste’ (for objects and entertainment) varies between classes. There are important ethical considerations

(13)

here, given the troubled history of using biological logic in social settings, such as eugenics and the pseudoscientific discipline of race biology. The motivation for choosing such an analytical approach is to bridge the gap between disciplinary perspectives, as outlined by Degnbol et al. (2006). In other words, this quantitative analysis technique is familiar to biologists, and is certainly applicable to the structure of the data. However, as discussed below, the results will not stand alone, and will be valid only by integrating concepts familiar to economists and other social scientists.

Fishing styles

As a way to address the challenges of categorisation for the sake of common pool resource governance, Boonstra and Sundberg have introduced the concept of fishing styles. Boonstra and Hentati-Sundberg (2016: 82) define fishing styles as:

“collectively shared and enacted, durable, habitual patterns of systematic and coherent actions, which aim to create congruence between normative notions about how fishing should be practiced, and fishers’ dependence on different social and ecological contexts.”

The fishing styles classification method goes beyond simple summaries of the fleet such as vessel- or engine-size segments, and instead builds upon the philosophy behind the métier approach (which focuses on gear, species, area and seasonality). It is an attempt to also include necessary attention to the social, cultural, historical, and economical aspects of fisher behaviour that cannot be found through statistical analysis alone. As humans rarely behave in purely rational ways, these factors are useful to understand what motivates fisher behaviour and the differences between groups. At the individual level, the authors stress that these styles are not exclusively the result of conscious choice, but rather that they often come about in reaction to values, affect, habit, and history. At the collective level, culture influences the choices that appear to be available to fishermen, and thus social and cultural contexts must also be examined. Finally, as Boonstra and Hentati-Sundberg (2016: 82) have emphasised, fishing styles are:

“ideal-typical modes of human action and interpretation [...] In their conceptual purity, styles cannot be found in reality. Nonetheless, styles are not virtual realities, or reifications, because they are grounded empirically. By constructing styles as ideal-types, researchers try to find a purposeful trade-off between scientific demands for parsimonious explanations and detailed descriptions of complex empirical reality.”

In short, fishing style portraits are not to be taken literally, but rather as a way to understand the general trend within groups of fishermen and to offer fishermen’s own explanations for behaviours such as non-compliance with laws or non-cooperation with scientific research.

3. Methodology

Scientific Replication

The aim of the replication process is to test the empirical foundations upon which a claim stands, preferably before a body of work begins to build upon it (King, 1995). Replicability standards do not expect researchers to be objective themselves, but approach objectivity by encouraging rigorous and systematic empirical work that keeps results independent of the researcher who produced them (Bueno de Mesquita et al., 2003). In other words, even with strong ideological disagreements, two researchers following the same method and using the same data should be able to reach the same conclusions. Engaging in scientific replication is valuable for this project for three reasons. First, replication is a useful pedagogical tool, in which junior researchers such as myself can learn from repeating and extending existing research methods in new contexts (King, 1995). Second, replications test the value of the method used in the original study and its

(14)

applicability beyond the context in which it was formulated. This paper is therefore an evaluation of whether the method that Boonstra and Hentati-Sundberg (2016) invite the scientific community to adopt in studying fisher behaviour is a theoretically and methodologically sound way to answer research questions about fisher behaviour. Third, by applying the main principles of the method proposed by the original authors and making necessary adjustments to extend the value of the work in its new context, embarking on this replication is an opportunity to achieve a rich understanding of fisher behaviour in the Netherlands. Thus I have attempted to stay as close as possible to the fishing styles method described by the original authors, yet some sensible adjustments were necessary during the process to adapt the method to the Dutch context.

Mixed Methods Research

As Degnbol et al. (2006) have identified, single-method mono-discipline research on fisheries problems is not effective because, when approaching a complex problem from a single methodology, the potential findings are limited before the study even begins. This form of mixed-method research follows a sequential explanatory design (Ivankova et al., 2006), whereby quantitative analysis forms the foundation of the study, and qualitative data are collected depending on the results of the first quantitative phase. In other words, instead of aiming to mutually validate or triangulate findings, the results of the different research methods are primarily complementary. Some validation is pursued, but the findings mainly relate to different aspects of the phenomena of study according to the theoretical traditions of the discipline from which the methods come (Erzberger & Kelle, 2010: 484).

It is worth defining some of the terms used herein.

Landings profile: The proportion of species landed during each trip (discards not included).2 Measures activity in very short term.

Fishing practice: An ideal-typical trip created by grouping similar landings profiles and determining average vessel and trip attributes. Measures activity at annual level. Similar to métier analysis. Fishing style: A way of going about being a fisherman that is determined by personal, cultural, and structural social factors and which incorporates intention as experienced and recounted by fishermen. Describes activity at the level of many years or even lifetimes.

The method proposed by Boonstra and Hentati-Sundberg (2016) uses three phases of research to first create landing profiles, transform them into a list of fishing practices, revise it with the help of experts, and then consult with fishermen to link these practices to a number of fishing styles (Figure 2).

2_{Catch is not equal to landings. Catch refers to the amount of fish pulled from the sea by nets, hooks, or traps. Landings, however,}

refers to the amount of fish that is brought to land for sale or redistribution. The difference between catch and landings is referred to as discards, which is the amount of fish that is caught but, due to size, species or desirability, is not brought to land. None of the trips in this data set are yet subject to a discard ban (to be implemented in coming years), so all references to landings mean the amount of fish caught minus the amount discarded.

(15)

The research begins with factor analysis of landings and effort data to create a preliminary list of fishing practices. Fishing trips were algorithmically sorted into practices so as to have the most similar landings composition, i.e. the proportion of different species landed per trip. The average characteristics of each practice, such as vessel length, area most commonly fished, and most common gear used, were then added to the table. This table was presented to experts for validation and explanation. The experts were asked to help explain and understand the links between the practices and their characteristics and to suggest possible motivations or behaviours that may cause the different characteristics of the practices. A revised list of practices was then devised. This revised list, as well as the qualitative input from the focus group, was used to direct interviews with fishermen themselves to more richly explain why these practices emerged. It is at this phase that the fishing practices, which only describe attributes, lead to fishing styles, which incorporate behaviour. Note that the quantitative analysis and expert consultations continued throughout the research. Quantitative analyses could be conducted to confirm or visualise input from the focus groups and interviews, and working with Wageningen Marine Research meant that there were often fisheries science experts on hand to question or with whom to discuss findings.

Figure 3. Procedural Diagram of the Method and its Outcomes

Quantitative Analysis

Raw data

This thesis uses logbook data from 2001-2016 from the Dutch in- and off-shore fishing fleet. Under the CFP, all vessels over 12m in length are obliged to record their catch via an electronic logbook (Nederlandse Rijksoverheid, 2018). The logbook entries record the vessel number, physical characteristics of the vessel such as length and KW power, the departure and return dates and locations, the ICES rectangles fished, gear and mesh size used in each ICES rectangle, and the volume and value of every species landed by the vessel. Skippers are obliged to create a new entry for each time they change ICES rectangle, gear or mesh size and thus each trip can have multiple logbook entries to record the catch made with different combinations of location and gear. The logbook data in this study include trips occurring from January 1, 2001 to December 31, 2016. The vessels included range from some that are 1.8m in length and which fish locally for a few hours, to vessels that are more than 100m in length and go to sea for weeks at a time. Although they are not legally obliged to report their activity in as much detail as larger vessels, vessels smaller than 12m in length record their fishing activity in logbook format (Nederlandse Rijksoverheid, 2018). Therefore, logbook entries for 359 vessels under 12m in length are also included in the data. The data pertaining to the small-scale activity is probably less reliable because these vessels are not obliged to record their activity electronically under the CFP.

(16)

Data selection

The raw data contains three categories of information: vessel attributes, fishing trip attributes, and activity attributes. Vessel attributes are useful for this study to associate vessel size and power, gears, and mesh sizes with the fishing practices and styles in the fleet. Fishing trip attributes were included so as to record the seasonality, number of days at sea, and the departure and landing harbours of the trips. The activity data records the date, gear used, areas fished, and the landing composition. The dataset comprises a total of 445 unique species caught over the study period, with a total landings volume of 6.42 million tons. However, the Dutch fleet specialises in only a few target species, with one species (herring) comprising 21.5% of all landings by volume. Further, only 25 species represent 98.4% of total landings by volume (6.31 million tons). A list of these top 25 species and their codes used in the data can be found in Appendix 1. To simplify the computations, only these top 25 species by volume were selected for analysis, with the other 420 species seen as anomalous. Any observation that did not include at least one of the top 25 species was deleted. A study of the small-scale fleet would require more careful consideration of this exclusion, but given that the aim of this study is to assess the activity of all vessels, the trade-off between completeness and computational speed and simplicity was deemed worthy. The data was then aggregated at the trip level so that each observation included vessel attributes, fishing trip attributes, and the total landings per species. Because multiple gears, ICES rectangles, and mesh sizes could be recorded in a single trip, this aggregation kept only the most frequent per trip. I added new variables that record the number of unique ICES rectangles, gears, and mesh sizes used in each trip to reduce the amount of information lost by aggregating the data in this way. Therefore, after data selection, the dataset contained 320,298 fishing trips with 46 variables, of which 21 recorded vessel and trip attributes and 25 recorded the volume of landings for each relevant species. A table of summary statistics can be found in Appendix 2.

Reflecting on value or volume

Although other studies concerned with fisher behaviour use the market value of the landings as their input (Pelletier & Ferraris, 2000), I have chosen to take landings volume data for two reasons. First, Boonstra and Hentati-Sundberg (2016) use volume in their study, and replicating their work is one of the main goals of this paper. Second, in a comparison between the top 25 species by volume and the top 25 species by value, the species included in each list differs only slightly. The volume data creates the scientific representation of fishing activity that grounds the qualitative investigation, during which the experts and fishermen are consulted to explain the motivations and intentions behind their behaviour. The value of the landings is one known motivation for behaviour, but beginning the quantitative analysis with volume does not incorporate this assumption in the first phase, allowing it instead to emerge during the qualitative phases.

Analysis process

The quantitative analysis followed the method reported in Boonstra & Hentati-Sundberg (2016), which itself closely follows that of Pelletier & Ferraris (2000). First, a new data frame was created that contained only the landings volume in kilograms for the 25 species for each observation using the statistical analysis program R. Any missing values were converted to 0, to represent no landings of the species during the trip. This data frame of absolute landings was then transformed into a relative species composition proportion by dividing the value of each species by the total landings during the trip. Doing so controlled the dramatic differences in total landings volume during trips. With each observation now containing the proportion each of the species contributed to the total landings, the data frame was log transformed to symmetrise the distribution. Next, a principle component analysis was performed to account for differences in abundances between rare and dominant species. Principal component analysis is an exploratory method to reveal underlying structures in data. The next step was to use the object CLARA (clustering for large datasets) to perform a cluster analysis on this data frame to create groups of practices based on the species composition of each trip’s landings.

(17)

At this point, an important decision had to be made about how many clusters to retain. An R function (NBClust; Charrad et al., 2015) computes many different methods of determining the optimal number of clusters in a group and reports the most commonly suggested number of clusters. This was attempted with the data frame, however the computational power required to perform the function grows exponentially and it could only be performed with a sample of less than 20% of the data, even with the help of large servers. Computing the appropriate number of clusters with a sample of the data may be suitable for a normally distributed and relatively homogenous data set, but the diversity in the Dutch fishing fleet, with many boats landing smaller quantities of some species and a few boats landing enormous quantities of other species, means that performing the cluster test on a selection of only 20% risked misrepresenting reality.

A solution was devised in the form of a looped test that performed the CLARA analysis 15 times (specifying 10 to 25 clusters) and recording the results of a silhouette test for each possibility. A silhouette test measures the homogeneity of the clusters by testing how well each observation fits with its assigned group. A silhouette width of 1 represents perfect homogeneity within the group, while -1 represents most observations being incorrectly placed. An average silhouette width (ASW) across all the clusters measures how appropriately all the observations have been divided into groups. The ASW peaked at 15 clusters. Moving forward with this number of clusters was further validated by the absence of any groups with a negative silhouette width and an above-average percentage of groups with a strong silhouette width of above 0.5 when the dataset was divided into 15 clusters (see Appendix 3).

Once the optimal number of clusters was determined, the CLARA analysis sorted every observation into a cluster based on the landings composition. This allocation was then reassigned as a categorical variable to the original data frame (i.e. before it was transformed to relative landings composition). In other words, the composition of the landings per trip was the leading ordering principle for the quantitative analysis. It was then possible to create a table presenting the fishing practices. The landings composition for each cluster conveys the species attributes of the fishing practices. The average characteristics of the vessel and trip attributes for each cluster were then calculated and added to the table to create a detailed representation of the practices.

Expert Focus Groups

The first expert focus group was held with several researchers at Wageningen Marine Research who are predominantly busy with managing and analysing the logbook data. Once I had completed the data selection and processing phase, I presented the group with a number of descriptive visualisations of the data and the first cluster analysis. The purpose of this consultation was to ask for some methodological feedback and to perform some ‘sanity checks’ on the attributes of the dataset.

The second expert focus group was held at Wageningen Economic Research in The Hague with a group of six individuals involved in the fishing sector. 3_{The panellists included a former fisherman and vessel owner who} is currently a board member in an organisation representing fishers and producers; a pelagic fisheries scientist; a member of a fishermen’s representative group; a process manager of the Ministry of Agriculture, Nature and Food Quality; a member of an association representing fish wholesalers and processors; and an economic researcher in sustainable fisheries. I first briefly presented the aim of the research and the

3_{It should be noted here that a solution had not yet been found to the dilemma of selecting the right number of clusters, and}

some data-cleaning issues remained in the data selection stage. As a result, the table presented during the focus group differs slightly, though not substantively from the final quantitative analysis.

(18)

preliminary results of the quantitative analysis. The 90-minute discussion was semi-structured around the preliminary table containing the fishing practices, with additional detailed figures given for each of the characteristics contained in the table of practices, such as landings composition, area fished, and gear used. The discussion was structured by several questions asking the experts about the validity of the practices. The meeting was conducted in a mixture of English and Dutch, and was recorded and later transcribed in English. The preliminary quantitative results and a brief summary of the outcomes of this discussion are attached in Appendix 5 and 6.

Fisherman Interviews

Focus on demersal beam trawl fleet

Due to time restrictions, a choice had to be made as to how to conduct and incorporate the interview data in light of the quantitative results. With 15 valid practices, there seemed to be two options: conduct short interviews with one or two respondents from each practice to ask basic questions about their fishing style and their reaction to the quantitative analysis; or to conduct longer in-depth interviews with fewer respondents. Boonstra & Hentati-Sundberg (2016), in the spirit of the replication process, had shared their interview guide, which was quite extensive. Given that one of the aims of this paper is to replicate their work, the choice was made to focus on one segment of the practices and conduct the in-depth interviews in a similar manner to the original authors.

As for selecting which fishermen to focus on, the quantitative analysis reported some statistically robust groups for which most observations in the group were similar, indicated by the silhouette width. The most statistically strong groups were shrimp fishers and razor clam fishers, meaning that fishing trips that targeted shrimp or razor clams tended to only catch those species during their trip. Other groups from the analysis returned low silhouette width scores and therefore had large variation between the characteristics of the observations in the group. These groups were mostly from the demersal sector of the Dutch fishing fleet. Demersal fish are those that live on or very close to the sea bottom, such as flounder, sole, plaice, dab, cod, gurnard, and turbot (species codes FLE, SOL, PLE, DAB, COD, GUU, and TUR in Figure 4).

Interviewees who were actively engaged in the demersal trawler fishery were purposefully selected because the low silhouette width scores suggested large potential for qualitative exploration of the differences within this segment of the fleet.4_{Selecting only a segment of the fleet achieves two aims: to best replicate Boonstra} & Hentati-Sundberg’s (2016) method in a time-restricted context, and to add the most understanding possible after the insight from the quantitative analysis. Further, a correlation matrix (attached in Appendix 4) showed that there are many overlaps in the demersal practices. In other words, vessels tend to specialise in razor clams only and thus the practice does not correlate with others. In contrast, trawling vessels that fish for sole, plaice, flounder, or dab are likely to use more than one fishing practice throughout the year, and interviewing these fishermen about their intentions will help understand how and why they combine practices, which can lead to insights into fisher behaviour over the longer term than the existing métier approach. Finally, the demersal practices all used mostly beam trawling gear except for two practices, small scale sole fishermen (practice 4) and small scale dab and cod fishermen (practice 9), who predominantly use nets. These practices were therefore excluded from the demersal selection and can be addressed in later interviews.

4_{There is also an element of timeliness in that some vessels in this group are currently subject to discussion about the use of}

controversial ‘pulse’ gear in the European Parliament. Instead of employing the traditional chains to disturb fish from their position on the sea floor and encourage them into the net, pulse fishing gear uses electrical impulses to stun fish. Pulse fishing gear is under review in the European Parliament and may be banned for most vessels in early 2019.

(19)

Figure 4. Distribution of Landings by Species, With Demersal Species Highlighted

Respondent selection and contact

With the decision to focus on the demersal sector of the Dutch fishing fleet, purposeful sampling identified the names of fishermen who are engaging in this activity. I met with four demersal trawler fishermen, three at their vessels as they returned to the harbour from their fishing trips and one in his home. I made contact with these fishermen through colleagues at Wageningen Marine Research and through contacting people in secretarial roles in fishermen’s organisations such as VisNed or Nederlandse Vissersbond. In an attempt to create snowball sampling, fishermen were asked if they could put me in contact with other fishermen they thought I should speak to in order to better understand the fleet. This proved to be rather ineffective. One reason may a sense of secrecy or privacy about cooperating with fisheries scientists. Sending me to other fishermen for interviews would necessitate revealing that they had spent time speaking with me, which may not have been looked upon favourably in the fisher community. The most likely reason, however, is that the topics covered in the interview are so broad that they would be applicable to nearly any fisherman in the sector, and therefore respondents did not deem it necessary to name any individuals.

Several factors influenced the small sample size of this part of the study, the most important of which was time. The in-depth interviews were conducted in Dutch, a language I am only somewhat proficient in. While most fishermen in the Netherlands understand English, my respondents expressed a preference for speaking in their native language so they could best express themselves. The structured interview guide, my passive understanding of Dutch, and the option of switching to English if something was not understood were sufficient to ensure a valid interview, but my novice status in Dutch meant that transcription and analysis were exceptionally time consuming. In addition, the fishermen spend most of the working week at sea, and were therefore there were a limited number of opportunities for us to meet. As a consequence, the qualitative results for this thesis should be taken as indicative only. This thesis serves as a pilot study for a larger project, which has funding to gather a total of 20 interviews.

(20)

Interview guide

The interviews closely followed the instrument provided by Boonstra & Hentati-Sundberg (2016). I also introduced some additional questions to fit the Dutch context and to more extensively discuss the quantitative results, which, importantly, the original authors did not do. The interviews began with some short-answer questions about the respondent, such as their age, when they began fishing, what species they target, and who they work with. The conversation then moved to economic choices, such as where and how the fishermen sell their catch, how the market price of fish influences their choices at sea, and whether they are satisfied with their income as a fisherman. In the motivation and career section of the interview, respondents were then asked to tell the story of how they became fishermen, how things have changed since they began, what they enjoy most and least about their career, whether they have ever done any other work, and whether they ever considered stopping and why. To get more directly at the fishing styles, fishermen were then asked about their fishing calendar, and to describe their most recent fishing trip. They were also asked about how to build a viable fishing enterprise, the qualities of a good fisherman, choices at sea, external factors, long term decisions, measures of success, and aspirations for the future. After this part, the fishermen were asked to review the quantitative results of the first phase of the study. This was also the beginning of a discussion about their own behaviour as fishermen. They were presented with the demersal trawler practices and their characteristics, and asked if they recognise their own activity in any of the groups. There was a moment for free discussion of any issues they have with the practice characteristics, and to explain to me why they suppose such groups may have come about. There were also questions about cooperation with others in the fishery, their motivation to remain in their career and their self-assessed adaptability. At the end of the interview there was a chance for fishermen to raise any issues that we had not discussed. This open conversation was allowed to continue as long as the fishermen saw fit, and allowed them to identify what the issues are that face them in their careers at present without any topic prompting from the interviewer. Coding and analysis of the interview data was performed using the R package RQDA (R-based Qualitative Data Analysis, Huang, 2016).

Life diagrams

During the motivation and career section of the interviews, respondents were asked to make a visual representation of their life in a diagram. This technique was also employed by Boonstra & Hentati-Sundberg (2016) and has been used in life histories research (such as Wengraf & Chamberlayne, 2006). The life diagrams provided a way for the fishermen to reflect on the progression of their life, to identify points of change (positive and negative), and to consider what direction their future may take. This exercise proved to be a useful way to encourage fishermen to openly discuss how they experience their career and to reflect on the consequences of decisions that they have made, the consequences of policy or environmental changes, and the influence of both personal and contextual circumstances.

(21)

4. Results

Quantitative Analysis

Fishing Practices

The principal component analysis of the logbook data sorted the more than 320,000 fishing trips into 15 fishing practices based on landings composition. There were no practices with negative silhouette widths, meaning that all groups had homogeneity greater than a chance assemblage of observations. The other characteristics of the practices were calculated by the mean, median or mode depending on what was appropriate for the variable. The data was not sorted into equal-sized groups, but rather into groups with the most homogeneity in landings composition. The smallest group contains just over 1,500 trips, while the largest has more than 135,000 trips. For instance, Practice 3, which contains the pelagic species, was smaller than average by the number of trips, but contained the largest share of total landings volume. Pelagic species are low-value high-abundance fish that are caught by Dutch vessels in local waters and internationally. The other pelagic practice (Practice 7) contained the largest vessels by length and engine power, the longest trips, and the greatest capacity (indicated by the average landings per trip of more than 1 million kilograms). As for the use of the fishing practices by the vessels, almost all practices were combined with another at least once during the year on average, with the exception of razor clam fishers. A correlation matrix (attached in Appendix 4 and reported in the penultimate column of the table) shows which practices were used together to a significant degree. In other words, aside from the very large ships that catch pelagic fish and those specialised in fishing for razor clams and shrimp, the vessels included in the study employed more than one practice during the course of a year. The preliminary version of the practices table (presented to the expert focus group) is attached in Appendix 6. An abbreviated version of the final table of fishing practices is presented below, with only key characteristics included. A more extensive version can be found in Appendix 7.

(22)

Practice N Trips Species* N Species per trip (mean) Annual Landings (mean, 1,000 tons) Landings per trip (mean kg) Calendar month (mode) Area (mode, name) Gear (mode, name) Mesh Size (mean mm) Vessel length (mean m) Trip length (mean whole days) N Practices (mean per vessel per year) Which other practices?** ASW ***

1 9,112 Razor Clams (100/0) 1 3 4,629 12 Zeeland Boat

Dredge 73 37 Same day 1 None 1.00

2 137,667 Shrimp (99/1) 1 16.2 1,886 10 Data Missing Beam

Trawl 23 21 2 2 None 0.97

3 10,111

Blue Whiting, Horse Mackerel, Sardinella, Atlantic Mackerel (25/25/19/11/20) 6 215.2 345,140 5 The Channel (Normandy) Bottom Otter Trawl 68 36 5 6 6, 9, 11 0.21

4 18,140 Sole (85/15) 2 0.3 315 7 Katwijk Nets 90 12 Overnight 5 9, 11, 15 0.70

5 27,791 Plaice, Sole (40/35/25) 9 11.5 7,195 9 Offshore Belgium

Beam

Trawl 82 39 4 5 13, 14 0.13

6 6,897 Gurnard, Striped Red

Mullet (33/32/35) 9 2 5,867 6

The Channel (Normandy)

Scottish

Seine 87 29 2 6 3 0.28

7 1,533 Herring, Horse Mackerel

(78/12/10) 2 107.7 1,126,256 11

The Channel (Normandy)

Midwater

Otter Trawl 41 68 11 4 None 1.00

8 6,795 Nephrops, Plaice

(50/35/15) 7 1.6 4,008 8 Dogger Bank

Bottom

Otter Trawl 80 25 5 4 10 0.38

9 7,187 Dab, Cod (66/8/26) 3 0.5 1,224 11 Katwijk Nets 90 16 Overnight 7 3, 4, 11, 12, 15 0.34 10 11,209 Plaice (85/15) 8 9.7 14,501 6 Skagerrak &

Kattegat

Beam

Trawl 91 35 5 5 8, 14 0.14

11 14,753 Cod (77/23) 3 0.8 1,030 1 Zeeland Hooks and

Lines 112 14 Overnight 5 3, 4, 9, 15 0.61

12 7,643 Flounder (76/24) 3 0.3 573 5 Zeeland Beam

Trawl 85 17 Overnight 4 9 0.75

13 23,572

Plaice, Sole, Dab, Flounder (35/25/16/11/13)

8 9.3 6,588 3 Zeeland Beam

Trawl 82 35 4 6 5, 14, 15 0.09

14 24,630 Plaice, Sole (68/16/16) 9 15.1 10,153 1 Southern North Sea

Beam

Trawl 80 40 4 5 5, 10, 13 0.31

15 13,258 Flounder, Sole, Plaice

(41/32/10/17) 5 1.9 2,328 7 Katwijk

Beam

Trawl 85 22 2 8

4, 9, 11,

13 0.25

(23)

The two charts above indicate a trend towards consolidation and specialisation in the Dutch fleet. The number of vessels active each year has been in near steady decline since the beginning of the study period, with roughly 25% fewer vessels active in 2016 as in 2001. The total catch in the same period shows that changes in landing volumes may be a cause (Figure 1), however catch remained relatively stable in the ten years to 2016, despite the decline in the number of active boats. Another reason for this trend may be the decommissioning of boats due to increasing competition from consolidated companies, and the introduction of ‘continuous fishing’, a recent development that sees a vessel operate with two rotating crews, rather than a single crew that takes rest days between trips. The number of practices used by a vessel in a year also shows a steady decline, suggesting that fishermen may be becoming more specialised in their activity.

Figure 5. Number of Vessels Active per Year in the Dutch Fleet, 2001-2016

(24)

Some practices produce significantly higher volumes of landings than others. Of the whole fleet, practices 3 and 7 stand out as exceptionally effective in terms of volume. Both are pelagic practices, which means that they target round fish species such as herring, mackerel, sardines, and whiting, which are found in abundant schools in open water. The trip with the greatest single landing in the dataset is found in practice 3. This trip lasted nearly six weeks and went to the coast of Chile and back, returning to the Netherlands with more than 13,000 tons of fish, mostly Chilean Jack Mackerel. The flounder beam trawl practice (practice 12) returned less than this during the entire study period. The demersal practices all returned less than 250,000 tons during the study period, with the practices that have plaice as their main catch (practices 5, 10, 13, and 14) returning the highest volumes of fish to land.

Figure 7. Total Landings per Practice, 2001-2016

(25)

The average composition of the landings per trip in each practice are reported above as composition graphs. In practice 1, where boats fished for shrimp, only 1% of the landed volume was a common species other than shrimp. In the demersal practices, some contained the same species (such as sole, plaice, and flounder), but with variations in the composition of the landings (shown in Figure 10). Compare, for example, practices 5, 13, 14, and 15, which all contain sole and plaice. Practices 5 and 14 appear to target sole and plaice alone, while practices 13 and 15 also have significant amounts of other species. It is unknown whether this is due to gear selectivity or heavy discarding practices in those practices. The trawling demersal practices also had the most diverse landings compositions, with averages of up to nine species landed per trip.

Figure 9. Landings Composition per Practice, 2001-2016

Classifying fisher behaviour in the Netherlands : replicating the fishing styles method to understand social motivations and explanations for activity in the Dutch fleet

CLASSIFYING FISHER BEHAVIOUR IN THE NETHERLANDS:

Replicating the fishing styles method to understand social

motivations and explanations for activity in the Dutch fleet

By Amanda Schadeberg

Thesis submitted in partial fulfilment of the requirements of the

Research Master Social Science (MSc)

Table of Contents

Abstract

1. Introduction

2. Theoretical Framework and Literature Review

Managing Fisheries

Categorisation

Fishing styles

3. Methodology

Scientific Replication

Mixed Methods Research

Quantitative Analysis

Expert Focus Groups

Fisherman Interviews

4. Results

Quantitative Analysis