Physical enrichment research for captive fish: Time to focus on the DETAILS

Growing research effort has shown that physical enrichment (PE) can improve fish welfare and research validity. However, the inclusion of PE does not always result in positive effects and conflicting findings have highlighted the many nuances involved. Effects are known to depend on species and life stage tested, but effects may also vary with differences in the specific items used as enrichment between and within studies. Reporting fine-scale characteristics of items used as enrichment in studies may help to reveal these factors. We conducted a survey of PE-focused studies published in the last 5 years to examine the current state of methodological reporting. The survey results suggest that some aspects of enrichment are not adequately detailed. For example, the amount and dimensions of objects used as enrichment were frequently omitted. Similarly, the ecological relevance, or other justification, for enrichment items was frequently not made explicit. Focusing on ecologically relevant aspects of PE and increasing the level of detail reported in studies may benefit future work and we propose a framework with the acronym DETAILS ( D imensions, E cological rationale, T iming of enrichment, A mount, I nputs, L ighting and S ocial environ-ment). We outline the potential importance of each of the elements of this framework with the hope it may aid in the level of reporting and standardization across studies, ultimately aiding the search for more beneficial types of PE and the development of our understanding and ability to improve the welfare of captive fish and promote more biologically relevant behaviour.


| INTRODUCTION
Fishes represent important laboratory animals, with increasing numbers of species used as model organisms in research (Braasch et al., 2015;Laland et al., 2011;Powers, 1989;Schartl, 2014;Utne-Palm & Smith, 2020). Increasing the diversity of model organisms can be a boon for research (Alfred & Baldwin, 2015) and fishes, as the most numerous group of vertebrates, are of interest to many researchers. Fishes encompass a range of important ecological roles and specific niches with exceptionally diverse communities, for example in coral reef fishes (Stuart-Smith et al., 2013) and rift lake cichlids (Muschick et al., 2012), and represent a considerable diversity of morphology, behaviour and reproductive biology (Fernö et al., 2020;Helfman et al., 2009;Wootton & Smith, 2014). In addition to research opportunities, a large variety of species of fish is also kept in captivity for aquaculture and ornamental pets. When keeping animals in captivity, housing conditions play an important role in the welfare of captive animals and environmental enrichment is an important component of these conditions (Mason, 2010;Newberry, 1995;Shepherdson et al., 1999;Swaisgood, 2007). However, relatively little is known about what enrichment to provide for most fishes in captive settings or the effectiveness of the different kinds of enrichment available.
Comparatively few studies on the effects of enrichment have been conducted on fish, with an analysis of the literature in 2007 revealing that fish were subjects in less than 0.5% of all enrichment studies conducted on vertebrates (de Azevedo et al., 2007). Of those studies involving fish, most are focused on relatively few species, with a focus on salmonids and zebrafish, Danio rerio (Hamilton 1822) . This relative lack of knowledge is increasingly important to address as there is a mounting drive to improve and regulate welfare for fish species kept in captivity for research, aquaculture or ornamental reasons (Browman et al., 2018;Jacobs et al., 2018;King, 2019;Saraiva et al., 2019;Sloman et al., 2019;Sneddon et al., 2017;Stevens et al., 2017).
Physical enrichment (PE), also referred to as structural enrichment, is a form of environmental enrichment that generally refers to any form physical complexity added to housing for captive animals.
Physical structure has long been known to provide potential benefits for fish and a heterogeneous environment can provide shelter from water currents, reduce aggression from other fish and act as landmarks around which to establish territories (Kalleberg, 1958). Knowledge of the importance and potential positive effects of PE has increased as research interest and effort has grown. It is now understood that adding PE to fish housing can have significant effects and may provide a range of potential benefits, both in terms of welfare for the fish, but also for research validity, where different housing and rearing environments can result in behavioural differences across studies involving the same species (Webster & Rutz, 2020). Experiments on effect of enrichment can also be relevant for the aquaculture industry as some studies of common aquaculture species have highlighted positive effects of enrichment. For example, PE has been shown to result in lower levels of cortisol in captive Atlantic salmon, Salmo salar (L. 1758), and Chinook Salmon, Oncorhynchus tshawytscha (Walbaum, 1792) (Cogliati et al., 2019b;Näslund et al., 2013;Rosengren et al., 2016), and is a commonly ascribed tool to reduce stress in captive fish and improve welfare Stevens et al., 2017). When discussing potential benefits of PE, we need to be clear what the beneficial outcomes can be for the fish in terms of welfare, better health, more stimulating environment, and also for scientists and aquaculturists where benefits may include improved survival and more 'natural' behaviour and physiological responses. Whether or not PE provides any benefit also depends on what that benefit is, how the 'benefit' is being valued and its connection to the goals of the enrichment programme. For example, a detailed study of the effect of PE on Atlantic salmon, S. salar, showed a welfare benefit for the fish from PE with reduced levels of cortisol and stress from disturbance, while mean growth was lower than in unenriched tanks (Rosengren et al., 2016). Decisions as to what PE to add to housing for fish must also take into account the pragmatic consideration of the need for usability and ease of maintenance (Lidster et al., 2017). Research into the effects of PE is the primary way to inform these decisions.
There is a host of potential benefits afforded by enriched environments. Shelter provided by PE can reduce metabolic costs (Chrétien et al., 2021;Finstad et al., 2007;Millidine et al., 2006). This, in conjunction with reduced levels of stress, may result in improved growth rates observed in Oncorhynchus mykiss (Walbaum, 1792) (Voorhees et al., 2020;White et al., 2019) and other species provided with shelter (Batzina & Karakatsouli, 2012;Zhang, Bai, et al., 2019). The presence of PE can result in less physical damage to fishes, for example less dorsal fin damage was observed in structurally enriched tanks (Berejikian, 2011), and reduce the frequency of potentially damaging escape-related behaviours (Zimmermann et al., 2012). Provision of PE may even increase survival in disease epidemics, for example juvenile S. salar reared in enriched environments showed greater survival in an outbreak of fish pathogen [Flavobacterium columnare (Davis 1922)] than fish raised in standard hatchery conditions (Räihä et al., 2019). A physically enriched environment can also promote the development and expression of more varied and more ecologically relevant behaviours (Braithwaite & Bergendahl, 2020;Brown et al., 2003;Sundström & Johnsson, 2001;Ullah et al., 2017), affect brain physiology and development (Arechavala-Lopez et al., 2020;DePasquale et al., 2016;Fong et al., 2019;Mes et al., 2019;Salvanes et al., 2013;Ullah et al., 2020), and promote learning and performance in cognitive tests (Carbia & Brown, 2019;Roy & Bhat, 2016;Salvanes et al., 2013;Strand et al., 2010). Such positive effects of PE can result in improved survival in programmes where fish are raised in captivity for eventual release into the wild (Hyvärinen & Rodewald, 2013;Johnsson et al., 2014;Lorenzen et al., 2010;Mes et al., 2019;Salvanes & Braithwaite, 2006). However, enrichment is not always found to increase post-release survival (Brockmark et al., 2007;Solas et al., 2019;Tatara et al., 2009). Moreover, as Näslund and Johnsson (2014) pointed out in a comprehensive review of PE there are many nuances that can impact the effects of PE on fish behaviour and welfare. The presence of PE is not always beneficial. While some studies showed improved cognition and brain growth in fish kept in structured environments, other studies found no effect (Brydges & Braithwaite, 2009;Näslund et al., 2019;Toli et al., 2017) or even negative effects (Burns et al., 2009) (DePasquale et al., 2019;Schroeder et al., 2014), heavily enriched aquaria can lead to increased aggression and lower growth in the same species (Woodward et al., 2019). Despite these many examples of positive effects found in studies of enrichment, knowledge gaps remain, highlighted by contradictory findings. There is limited knowledge on the types of enrichment that have effects and how they work (e.g., physiological and neurological processes). The general level of understanding might be summarized as some PE is better than none, some of the time.
What are the proximate reasons for fish to benefit from physical structure in captivity? In the most general sense structure can provide shelter and protection from the physical environment and protection against other animals. The availability of shelter may be enough to provide benefits. One study showed that Atlantic salmon rested outside of shelters and suggested that the mere presence of a shelter was enough to significantly reduce metabolic rates (Millidine et al., 2006). There are many studies that show that fish use structure as a form of antipredator refuge, and this may be the ultimate driver for use of and benefit from structure where protection from currents was secondary to protection from predators (Valdimarsson & Metcalfe, 1998). PE may afford refuge from many other things, including artificial lighting (McCartt et al., 1997), intraspecific competition and aggression. Physical heterogeneity can potentially reduce aggression in territoriality species by, as in the rose bitterling Rhodeus ocellatus (Kner, 1866), affording landmarks that can be used to delineate territories, leading to reduced aggression between rival individuals (Smith, 2011). Similarly, individuals within a population can show specific characteristics which can lead to individual differences in shelter use. For example, in the wild older (and larger) female sockeye salmon, Oncorhynchus nerka (Walbaum, 1792), were shown to prefer and utilize deep-water refuges (Camacho & Hendry, 2020) while fish of other ages and sex clustered in different depths. These differences were attributed to the fact that larger females are more vulnerable to predation from bears when they return to spawn. Knowing the function of shelter use for a given species could allow researchers to select the optimum types of PE.
Despite the potential importance of PE, a survey of articles published in fish biology-focused journals (between 2003 and 2013) suggested that more than 70% of studies did not use any PE . This is no doubt partly driven by one of the major challenges of using PE: adding PE to a tank can increase the difficulty of and time required for cleaning. A recent survey of husbandry practices in research laboratories highlighted this issue, with more than 60% of survey respondents considering provision of PE a challenge that required intensive labour and was thought to lead to an increased chance of disease (Lidster et al., 2017). Developing forms of PE which can provide welfare benefits while reducing maintenance costs is an important challenge that will require further research into the various factors that can impact the costs and benefits of PE. It will also require the improvement of the standardization and reproducibility of studies exploring enrichment.

| Aims and scope
In this article we look at the current state of the research into PE, focusing on the specific forms of enrichment provided in studies. We  Näslund and Johnsson (2014) highlighted several important considerations for the field. Most pertinent to this article, they revealed that the effects of PE vary markedly across studies and suggested that future studies should attempt to differentiate more nuanced effects of different PE. The aim of our review is to extend this idea, specifically highlighting the need to understand what and why specific items of PE are used in studies. We provide a framework which we hope will help to guide researchers to consider and report more of the characteristics of the enrichment they use. We think more refined reporting, resulting ease of attaining higher levels of standardization and a deeper understanding of what PE to use at fine-scale levels will address the issues raised by Näslund and Johnsson (2014) and other authors (Huntingford et al., 2006;Toni et al., 2019).

| SURVEY APPROACH
To examine the current status of PE research in fish biology and the levels of reporting of PE details, we conducted a nonexhaustive survey of published research articles where PE was the focus or an important part of the aim of the study. In line with our aims this survey was used to collect a representative sample of papers currently published in the field rather than to conduct a formal meta-analysis.
To do this one author (NJ) used Web of Knowledge to search for all papers with the terms 'physical enrichment' and/or 'structural enrichment' and/or 'environmental enrichment' and 'fish' from 2015 to May 2020. The resulting research papers that described experiments which included investigation of the effect of PE on captive fish were selected and included in the survey. For each paper, the introduction and methods sections were checked for the methodological details they reported, focusing on the description of PE used. We chose to collect data from the last 5 years to get a snapshot of the more recent PE-focused work and further the comprehensive review by  which highlighted the importance of and nuances that can impact the effects of PE on captive fishes.
For each study we recorded the category of PE reported, for example whether a particular item of PE was a substrate or shelter, then recorded the general type of PE where an item of PE belonging to a substrate category could be gravel, sand or mud. For each type of PE, we recorded whether additional details were provided, such as the manufacturer and name of plastic plants, species name of live plants or some description of the shelter provided (e.g., clay pot or PVC pipe). For each of these specific types of PE we recorded whether the specifications of those PE items were provided, for example the grain size and colour of gravel, length and diameter of PVC pipe, or the number, colour and size of leaves on plastic plants.
We also recorded the amount of these PE items, for example numbers of plants added to the aquarium or the depth of gravel substrate. In addition, we recorded whether the rationale for the use of the PE was explicitly stated, essentially whether the PE used was chosen based on previous studies or in an attempt to mimic certain, normally wild, conditions. In some studies density is not reported and is included as no, but this may be due to fish being kept in isolation where density is not required.
In addition to the characteristics of each PE item reported we collected information on associated aspects of each study. We recorded the fish species tested and life stage exposed to PE and origin. We also recorded information regarding factors that are known to be or may also be important to consider in studies of the effects of PE, including whether the density of fish and the duration of the exposure to PE was reported for each study, photoperiod and lighting details, water quality parameters and evidence of whether or not pathogen checks were made.

| Reporting details of the PE used
Across the 65 studies surveyed we collected reporting details for 159 types of PE used (

| At the study level
Of the species of fish tested in the studies, the majority (56%) were salmonids (30%) or zebrafish (26%). This reflects the major use of   (Greenway et al., 2016). The importance of substrate colour to reduce predation risk is also crucial where fish that match the colour of the background substrate are less likely to be predated (Browman & Marcotte, 1987;Ostrowski, 1989;Sumner, 1934). This has led to recommendations for salmonid aquaculture to acclimate fish raised in hatcheries to backgrounds with similar coloration to the gravel in the habitat where they will eventually be released (Donnelly & Whoriskey Jr, 1991). Besides colour, brightness of a substrate can also be important factor in fish preferences (Wu et al., 2020), most likely for similar reasons. Substrate preferences can be driven by other aspects of an animal's ecology, for example three-spined sticklebacks, Gasterosteus aculeatus L. 1758, prefer complex substrates (with heterogeneous topography and colour) over simple substrates (homogeneously coloured and textured), but only when they are not satiated, suggesting preferences are linked to foraging preferences (Webster & Hart, 2004).

| TOWARDS A MORE REFINED UNDERSTANDING OF PE
Many forms of PE that have welfare benefits may not have direct connection to the ecology of the fish, for example novel objects, behavioural engineering and stimulation. Water flow, or regulated changes in water flow, is one such form of enrichment that can promote exercise for fish with many associated benefits to welfare and growth (Huntingford & Kadri, 2013). Provision of water flow or other forms of enrichment which promote swimming and exercise will still benefit from considering ecological aspects, for example the maximum swimming and flow speeds may be based on natural water conditions.  (Valdimarsson & Metcalfe, 1998).

| Timing of enrichment
The duration of exposure to enrichment (or lack therefore) is the other synergistic aspect to consider when designing and reporting studies of enrichment. A study of rainbow trout, O. mykiss, showed that fish kept in enriched conditions for longer durations performed better in cognition assays (Bergendahl et al., 2016). However, even short periods of exposure to PE can have effects as long as the fish were exposed to the enriched environments very recently prior to testing. For example, swimming agility and performance in behavioural assays improved even after relatively short exposure to enrichment (Bergendahl et al., 2016(Bergendahl et al., , 2017dos Santos et al., 2020). Duration of studies can also be con-

| Inputs
Inputs such as food and water are, of course, fundamental for maintaining fish in captivity. The physical and chemical properties of the water are also crucial, where water quality parameters such as dissolved oxygen and pH can impact stress and affect the underlying physiology, behaviour and ultimately welfare of fishes (Huntingford et al., 2006;MacIntyre et al., 2008;Stevens et al., 2017;Williams et al., 2009). Fishes exposed to even relatively short fluctuations in water quality in confinement showed behavioural changes with associated welfare implications (Vanderzwalmen et al., 2021). As such, water quality parameters are frequently well reported in studies of enrichment, but not to the same level across studies. Differences in reporting of water quality likely reflect differences in fish species requirement but also enclosure type, for example studies using naturally fed flow-through tanks frequently report fewer water parameters. However, measuring and reporting the details of water chemistry are important in determining the effects or benefits of PE and affording more comparative or reproducible studies. Diet and differences in food types can also be important. Levels of dietary nitrogen can have impacts on welfare (Conceição et al., 2012) and sources of nitrogen from specific diet formulations can affect growth and other welfare parameters (Bonaldo et al., 2015). Live prey may be worth considering as a form of enrichment and can be fundamental to survival for many species (Ruyet et al., 1993). Larval Pseudochromis flavivertex Rüppell, 1835, for example, do not survive without diets enriched with live prey (Olivotto et al., 2006). The provision of live prey can be especially important for species reared in captivity for later release into the wild (Brown et al., 2003). Temperature is an example of an easily measured and commonly included water parameter that has both direct and indirect effects on welfare. The provision of a varied thermal environment improved measures of welfare and growth in Atlantic salmon (Sanhueza et al., 2018).
Temperature can also impact preferences for physical complexity, for example minnows, Phoxinus phoxinus (L. 1758) significantly increased the time spent in refuges when the temperature dropped (Greenwood & Metcalfe, 1998). Temperature differences within a tank afforded by physical complexity in aquaria can impact use of particular areas or ref- uges for fishes exhibiting behavioural fever (Boltaña et al., 2013;Huntingford et al., 2020). Thermal preferences can also depend on the level of structural complexity or micro habitat, as shown for coral reef damselfish, Chromis atripectoralis Welander & Schultz, 1951 (Nay et al., 2020). Various water parameters can also have interactive effects.
For example, while high nitrate levels can reduce swimming speed and duration in juvenile silver perch, Bidyanus bidyanu (Mitchell, 1838), the effects can depend on or be masked by changes in temperature (Isaza et al., 2020). Other environmental toxins can impact the preferred temperatures of fish (Petersen & Steffensen, 2003;Skandalis et al., 2020).
Not all inputs are intentional, and parasites and other pathogens are an important welfare concern (Barber, 2007;Bui et al., 2019).
Pathogens can have many effects on fish, including effects on their behaviour and measures of welfare and potentially use of shelter (Gabagambi et al., 2019;Martins et al., 2012). They can also impact other measures of behaviour, for example parasites can impact cognitive performance (Barber et al., 2017) and shoaling preferences.

| Lighting
Light, the intensity, wavelength and amount (photoperiod), is an important factor that is often overlooked in studies of PE. Light levels can have a large impact on fish behaviour, playing a crucial role in their behavioural ecology (Cerri, 1983;Keep et al., 2020;McCartt et al., 1997;Santon et al., 2020), and can drive the use of shelter and shade. Light levels can impact levels of aggression (Valdimarsson & Metcalfe, 2001), which in turn can have significant impacts on fish welfare (da Silva et al., 2020).
Light can also impact growth (Boeuf & Le Bail, 1999). A study on blunt snout bream, Megalobrama amblycephala P. L. Yih, 1955, showed that high light intensity caused stress, elevated oxidative rate and immunosuppression, but low light intensity led to depressed growth, antioxidant capability and immunity (Tian et al., 2015). Shade in aquaculture settings has also been shown to reduce levels of sea lice in pen-reared Atlantic salmon (Huse et al., 1990).
Larval development and growth rates of zebrafish can be significantly impacted by light conditions (Villamizar et al., 2014), and growth rate and aggression between individuals have been shown to improve with lower light intensity in other species too (Almaz an- Rueda et al., 2004;Arambam et al., 2020;Boeuf & Le Bail, 1999;Rahman et al., 2020;Tian et al., 2015). Recent studies have focused on the effects of other aspects of light conditions on captive fish, such as the effects of acute bursts of light from photography (camera flashlight) (Knopf et al., 2018), but despite the importance of light conditions, and the relationship with PE which may provide shelter and shade from lights, most studies exploring PE tend to ignore light and provide very few details outside of photoperiod.

| Social environment
Density of individuals in aquaria has long been known to impact behaviour (Ellis et al., 2002), for example a recent study showed that killifish, Nothobranchius furzeri R. A. Jubb, 1971, exhibit differences in body length, activity, aggressiveness and feeding behaviour across different densities (Thoré et al., 2020). Beyond 'simple' fish density, social factors can impact the effects of or requirements for PE. Being a member of a group can provide many benefits (Ward & Webster, 2016), but most species of fish do not blindly form groups.
Fish are able to differentiate between individuals and actively moderate the composition of their groups  and habitat complexity can, in turn, have a strong effect on social behaviour (Rodriguez-Pinto et al., 2020). The benefit of social enrichment and preferred group sizes is species-specific (Saxby et al., 2010) and can depend on the level of sociality of a species. Highly social species, obligate shoalers, can show strong preference for forming groups with conspecifics, for example zebrafish (Spence et al., 2008), and can show well-developed discrimination between groups, preferring the larger group, for example Lamprologus callipterus Boulenger, 1906 (Durrer et al., 2020). Specific aspects of social context, including group size, makeup and dynamics, can be important and may impact any observed 'benefit' of PE. Social context may even reduce the requirements for any PE at all. For example, solitarily housed zebrafish may use and gain benefits from structure (Collymore et al., 2015), but in groups zebrafish showed no preference for PE (Jones et al., 2019;Kistler et al., 2011). Certain visual signs of activity of other fish can impact shoaling by zebrafish (Pritchard et al., 2001) and large amounts of structure in a tank can lead to increased aggression in zebrafish (Woodward et al., 2019) and other species (Boerrigter et al., 2016).
Aggression is particularly important in captive environments, with effects on body condition and growth, and can be closely tied to social context. Levels of aggression between conspecifics can increase in captivity where movement is restricted (Doran et al., 2019;Kelley et al., 2006), especially in territorial species (Perrone et al., 2019).
Levels of aggression can depend on social dynamics, species composition, sex and stage of the individuals in a group (Bayani et al., 2017;Desjardins et al., 2012;McRobert et al., 2012;Sloman et al., 2011). Other structure Descriptive measurements, including size or volume of the whole object. Also, the number, position and dimensions of each of the openings/ refuges. a Ecological rationale b Substrate State the general relevance of substrate for study species, i.e., benthic species or commonly associated with specific substrates. Is the substrate used in the study to mimic natural conditions for the species or for some specific reason(s)?
Plants Are the live plant species associated with the natural habitat of the species or do the live/plastic plants have similar physical characteristics and dimensions to habitat specific species? State the reason(s) for using the chosen plants.
Other structure State the reasoning for the use of the structure, relative to the goals of the study and ecology of the species. Explain why the number of items and their dimensions was selected.

Stage
Life history stage of species tested, and the season fish are collected and tested.

Duration
Duration of time individual fish spend in both housing and experimental condition(s). Include the acclimation time prior to specific assays.

Amount
Substrate Surface area covered within the tank and depth of sediment layer.

Plants
Some indication of surface area of the tank covered by plants and/or the number of individual plants and density.
Other structure The number of items per tank would be a minimum to report but we would recommend including a ratio of shelters to fish number per tank.

Chemistry
Water chemistry parameters and temperature as standard, but include some measure of variation over time, especially for fish housed for longer durations.

Pathogens
Explicitly state whether checks for parasites or diseases were made.

Flow
Some measure of water movement, e.g., filter turnover rate.

Lighting
Photoperiod as standard, but include details of the lighting source, e.g., manufacturer, wattage, hue, luminance and intensity.
Social environment Density of fish and sex ratio as standard, but also size (and size range) and relatedness. Include some measure of the level of sociality of the species, level of familiarity of fish housed and tested together, the social dynamics of the species, and how these may impact aggression and resource partitioning. a For individual items of physical shelter, e.g., clay pots or tubes, dimensions should focus on the diameter of any openings fish may be expected to move through or rest within, but also the overall dimensions and size of the object, colour, and for less commonly used shelter types the texture of the material. b Decisions as to which objects to be provide as enrichment can be relatively simple when there is direct ecological relevance, e.g., providing empty snail shells for shell-dwelling cichlids or plant species commonly found in the native habitat of the fish or artificial versions that mimic these plants. However, this information should be reported. Moreover, the specific details of particular items of enrichment may be informed by the ecology of the fish and help to connect with the details provided in other parts of the framework. For example, why are a set number of shelters used per tank, or per density of individuals? Why was one colour of substrate used?
We highlight the need for developing more nuanced understanding of the factors that drive the use of PE and contribute to welfare improvements for captive fishes. There is a growing awareness and demand for more empirical studies and quantitative measurements of fish welfare (Brydges & Braithwaite, 2009;Huntingford, 2004;Huntingford et al., 2006;Johnsson et al., 2014;King, 2019;Sloman et al., 2019;Sneddon et al., 2016;Turnbull & Huntingford, 2012). Answering fundamental questions around PE is a crucial aspect of that: knowing what type, how much and when to provide structure to maximize the benefits associated with PE will help to improve the welfare outcomes for fishes. The proposed DETAILS framework may help to achieve these goals by focusing researcher attention on these questions and providing a memorable guide to reporting of PE used across studies. In conjunction with recent preparatory guidelines (Smith et al., 2018), our reporting framework may also benefit the reproducibility of empirical studies exploring the effects of enrichment. Ultimately, a more detailed knowledge of PE may allow the identification and implementation of design changes which can afford the benefits of PE while minimizing the costs associated with them, leading to benefits for the fish themselves and researchers, aquaculture industries and manufacturers of housing systems.