Modelling Hazards in Fisheries and Aquaculture Activities in the Mediterranean Sea and the Risk of their Transmission and Dispersion. Is it Feasible?

Modelling approaches in marine science is a controversial issue as no model is, or can be, a perfect representation of nature. Models can provide useful information for the dynamics of ecosystems and inform us about the likely consequences of human activities in fisheries and aquaculture. Applying a suite of dynamic models can be valuable predictive tools for modelling hazards transmission in fisheries and aquaculture activities in the Mediterranean Sea. These should include a. A low trophic level ecosystem model and sub-models to describe the ecosystem functioning of the sea for the background physical information and a biogeochemical sub-model which simulates functional groups. b. A pelagic fish individual-based model (IBM) to describe the bio-accumulation of chemical and biological hazards. c. An aquaculture integrated model, a mass balance model, to calculate the input of effluents into the environment as a result of the fish farm and feeding regimes. A dynamic energy budget for cultured bivalve species bioaccumulation heavy metals algal blooms. These dynamic models can contribute to develop and/or improve systems ensuring process efficacy and validation for hazard control by identifying “hot spot” zones and concentrations of hazard agents above certain limits, improve the effectiveness and efficiency of the controls performed by food safety Competent Authorities along the seafood chain, identify areas of hazard agents accumulation and contribute to the transparency and reliability of food safety in the Mediterranean fisheries and aquaculture production sites.


Introduction
The term "hazard" (the intrinsic potential to cause harm) is often confused with the term "risk" (the probability of harm occurring at a given exposure) and often the public do not generally differentiate between these two terms [1]. The microbiological definition of hazard in Codex Alimentarius [2] is "inherent property of an agent or situation having the potential to cause adverse effects when an organism, system, or (sub)population is exposed to that agent" whereas of risk is "the probability of an adverse effect in an organism, system, or (sub)population caused under specified circumstances by exposure to an agent". Aquaculture and fisheries are important activities in the Mediterranean Sea entailing significant socioeconomic implications [3,4]. Both activities face some potential hazards including natural, anthropic or both causes. Aquaculture hazards include diseases in both fish and shellfish whereas high stocking densities may lead to chronic stress that has important implications for fish immuno-competence [5-9], but relationships with infection levels are variable [10]. In addition, marine litter and especially micro-plastics (MPs -0.1 to 5,000 μm) and nano-plastics (NPs -0.001-0.1 μm) concerns for potential toxicity of the plastic particles towards human health are growing, as they could potentially induce physical damages Oceanography & Fisheries Open access Journal through particles themselves and biological stress through MPs/NPs alone or leaching of additives (inorganic and organic) [11,10].
Is it feasible to model hazards in fisheries and aquaculture activities in the Mediterranean Sea and the risk of their transmission and dispersion? The use of modelling approaches in marine science is a controversial issue as no model is, or can be, a perfect representation of nature [12]. However, as predictive modelling can provide a better understanding and potentially either prevent hazardous events or identify hotspot areas or describe the spread patterns and concentration of hazards (pollutants, infectious diseases etc), their transmission and dispersion needs to be integrated in dynamic models but in an environmentally realistic manner [13,7]. Therefore, a suite of dynamic models can be valuable predictive tools for modelling hazards transmission in fisheries and aquaculture activities in the Mediterranean Sea.
These should include the following:

A low trophic level (LTL) ecosystem model
The low trophic level (LTL) ecosystem model must be fully dynamic and to consist mainly of two sub-models: a hydrodynamics sub-model that describe the ecosystem functioning of the sea area and will provide the background physical information e.g. a Princeton Ocean Model -POM [14] or MICOM [15] or ROMS etc, to a second biogeochemical submodel which simulates functional groups. The LTL ecosystem model shall provide the dynamics of biological functional groups that consist of population processes (growth and mortality) and physiological processes (ingestion, respiration, excretion and egestion). The biotic system is subdivided into three functional types, producers (phytoplankton), decomposers (pelagic and benthic bacteria) and consumers (zooplankton and zoobenthos). These broad functional classifications are usually subdivided, according to their trophic level (derived according to size classes or feeding method) to create a food web. The plankton pool should have the functional groups based on size and ecological properties like diatoms, nanophytoplankton, picophytoplankton, and dinoflagellates. Bacteria, heterotrophic nano flagellates and microzooplankton represent the microbial loop. All groups in the phytoplankton and the microbial loop have dynamically varying C:N:P ratios. The chemical dynamics of nitrogen, phosphorus, silicate and oxygen are coupled with the biologically driven carbon dynamics.

A pelagic fish individual-based model (IBM)
The health benefits associated with seafood consumption are coupled with concerns about potential health risks associated with the presence of hazards (chemical and biological contaminants), both those occurring naturally and those resulting from human activities, in seafood. As the bio-accumulated hazards sooner or later will end up in our plates through the consumption of seafood, there is a need to understand this process and what is transferred from one trophic level to the other. In the Mediterranean, landings continued to increase until 1994, reaching 1,087,000 tonnes, and subsequently declined irregularly to 787,000 in 2013, with a group of 13 main species accounting for some 65 percent of landings, with anchovy (393,500 tonnes) and sardine (186,100 tonnes) being by far the dominant species [16]. Pelagic species catches are higher than the ones of demersal and 30 species contribute to 90 percent of the landings in all Mediterranean subareas.
For pelagic fish, the fish model must be on-line coupled with the LTL model (described above). Currently, it has been developed and is a full-life cycle, individual based model (IBM) that includes two species, the European anchovy (Engraulis encrasicolus) and the European sardine (Sardina pilchardus) [17]. Early larvae feed on microzooplankton, late larvae start consuming mesozooplankton and juveniles/adults interact only with the mesozooplankton compartment of the LTL model. The plankton biomass (micro-and mesozooplankton) that is consumed by the fish is removed in the LTL model, while fish bio-products from egestion, excretion and specific dynamic action are directed to the LTL particulate organic matter and dissolved inorganic nutrient pools.

The above two models (LTL & IBM) can be further developed to include potential hazards (microplastics, oil spill particles, etc)
That exist in the marine environment [10]. These hazards may be attached to phytoplankton, which subsequently may be consumed by zooplankton groups. In addition, they might be attached to zooplankton organisms and these will eventually be grazed by small pelagic fish. Also, the fish may randomly attach fibres, microplastics, oil spill particles on their gills and body. Small-scale lab experiments will be needed to quantify the above processes in order to describe some parameters and to customize the models.

An aquaculture integrated model
In order to calculate the input of effluents into the environment as a result of the farm operations and feeding regime, a mass balance model should be applied for the fish farms. Mass balance models have been developed based on nutrition (feed type) and conditions for salmonids [18,19], sea bream and sea bass among others [20][21][22][23]. Input of nitrogen and phosphorus supplied in fish feed can be used to calculate the amount harvested as fish, excreted in dissolved form (Urea, NH4, PO4) and excreted in particulate form (uneaten feed, faeces). Tsapakis et al. [23] and Lupatsch and Kissil [21] calculated that the largest portion of nitrogen supplied is excreted in the dissolved form as Urea (41%) and ammonium (26%), while phosphates losses account for 22% of phosphorus supplied. Conversely particulates released consist mainly of organic phosphorus accounting for 44% of phosphorus supplied whereas particulate nitrogen losses account for 10% of the nitrogen supplied. It is also estimated that approximately 5% of feed is settling uneaten, either being Oceanography & Fisheries Open access Journal consumed by wild fish or contributing to the organic load of the underlying sediment [24,25].
The aquaculture modelling tool could be based on the above-described LTL model and can be used for ecosystem monitoring of the dispersion of parameters, including bacteria biomass and bacterial production and can be used to model the spread of diseases from farms. It can also be adopted to predict the dispersion of parasites. This will allow the creation of zones that might get infected and therefore might trigger the use of antibiotics and other chemoprophylaxis treatments. If such treatments are not properly followed, the fish that will be harvested and directed to the value/market chain might include those chemicals and all this is a potential hazard. This mapping of the zones, can be evolved to a DSS tool that will identify critical zones of potential hazards and prioritise the controls of the Competent Authorities to identify contaminants, pathogens etc.

A dynamic energy budget (DEB) model for cultured bivalve species
The bivalve mollusc model in combination with the LTL model can be useful to predict the bioaccumulation of hazards such as heavy metals, toxins from harmful algal blooms (HABs) etc. A dynamic energy budget (DEB) model for bivalve molluscs should investigate the growth and reproduction of cultured bivalve species raised under different environmental conditions (varying phytoplankton carbon biomass, particulate organic carbon and temperature) and can be tuned against field data.

Conclusion
Models can bridge the gap of the complex links between existing environmental variation and hazards presence and risk of transmission although responses of hazards over and above natural variation might be challenging. There are tools available that can be further improved to describe the most important physical and biochemical processes that, combined together, determine the dynamics of the ecosystem. Given the complexity of these processes and their interactions, mathematical models can be regarded as unique tools to deliver integrated approaches and better understand the mechanisms of hazards transmission and risk of dispersion and inform about the likely consequences of human actions.
The above-described dynamic models can contribute to: a.
Develop and/or improve systems ensuring process efficacy and validation for hazards control by identifying "hot spot" zones and concentrations of hazard agents above certain limits.

b.
Improve the effectiveness and efficiency of the controls performed by the food safety Competent Authorities (CAs) along the seafood chain by direct and guide their sampling effort to detect contaminants (Decision Support Tool that will identify critical zones of potential hazards) and prioritise the controls of the CAs to identify contaminants, pathogens etc. c.
Identify areas of hazard agent's accumulation (e.g. microplastics and Nano plastics, HABs, areas that antibiotics and chemicals are used etc). d.
Assist activities that will develop detection and monitoring tools that will allow for the data collection, integration, validation and analysis. e.
Contribute to the transparency and reliability of food safety in the Mediterranean production sites. f.
Future research should evaluate trophic transfer of hazards with their associated risks through the marine food web for humans. Models that describe the evolution of hazards are in progress and will be published in the near future.