Skip to main content

Development of monotoring quidelines and modelling tools for environmental effects from Mediterranean aquaculture



In northern Europe, Canada, Tasmania and recently in Chile, considerable efforts have been made to assess interactions between aquaculture and the environment. Most models, guidelines, monitoring procedures, and environmental quality standards are, however, directed toward salmon farming. Although scientific knowledge on the impacts of salmonid farming has progressed substantially, very little is known of fish farming impacts in the Mediterranean. In the last decade aquaculture of sea bass (Dicentrarchus labrax) and gilthead sea bream (Sparus auratus) has experienced a period of exponential growth in the Mediterranean region. This process needs to be controlled in order to ensure a sustainable development of the aquaculture industry and to consider other aspects of an integrated management of the coastal zone, including tourism, fishery, and environmental protection.

Established guidelines for an initial impact assessment need to be complemented by subsequent impact monitoring.

Despite the rapid development of aquaculture in coastal areas of the eastern Mediterranean little has been published on the environmental impacts of this industry. Also, protocols, monitoring systems and techniques for the control of marine cage fish farms in Mediterranean conditions do not currently exist. In general, it has been assumed that these will, at least qualitatively, follow the pattern established in northern latitudes. Strategies developed and proven for salmon culture in northern Europe and elsewhere deliver comprehensive concepts, but need to be modified and adapted to the ecological particularities of the Mediterranean Sea. In addition to differences in the species cultured, the climate, the current regime, and the level of eutrophication, differences in the composition and diversity of fauna and flora between the North Atlantic and the Mediterranean Sea need to be addressed.

At eastern Mediterranean coastal sites, surface water temperatures are higher and water currents are generally weak, being driven by density and wind with an insignificant tidal component. Many areas are highly oligotrophic and may be phosphorus rather than nitrogen limited. Some evidence from acoustic and video surveys and diving surveys has suggested that the effects of mariculture on the benthic environment around fish cages in Greece are very much less than those typically experienced in comparatively eutrophic areas, for example in Scottish sea lochs. It has also been reported that large numbers of wild fish occur around the cages, in mid-water and near the bed and the suggestion made that the fish were responsible for consuming much of the waste food from the farm. Thus evidence of impact effects based on experience from studies of North Atlantic salmonid farms may well be unreliable if used for establishing controls for the Mediterranean mariculture industry. Environmental impact models which attempt to predict the degree of accumulation of wastes were developed and validated in strongly tidal North Atlantic areas and do not include terms relating to consumption of wastes by wild fish. Such models should be used with great caution, if at all, in non-tidal Mediterranean environments.

In order to address these concerns, the MERAMED project will develop and establish a model based control system for environmental monitoring of fish cage farms in the eastern Mediterranean. This will require the fulfilment of three major objectives:

1. to evaluate procedures used in the regulation and monitoring of marine cage fish farms in Norway, Scotland and propose an appropriate set of protocols, monitoring systems and techniques for the control of fish farms in Mediterranean conditions.

2. to carry out a field research programme to provide appropriate data on the environmental impact of marine cage fish farms in a range of conditions in the eastern Mediterranean.

3. to develop a predictive model to simulate the environmental response at Mediterranean sea cage farms to differing cage stocking levels and feeding regimes.

The protocols, monitoring systems and techniques and model will be designed as a suite of management tools for both the industry and regulatory authorities in the area.

Objective 1 will be achieved by making use of the wide experience underlying the development of the current regulations governing the Norwegian and Scottish industries and of the environmental monitoring required of cage farms operated in those countries. The aspects of these regulations and requirements, which are appropriate to marine cage farms in the Mediterranean, will be identified and adapted to the differing conditions. After testing, control protocols and monitoring techniques suitable for application in the eastern Mediterranean will be suggested. All methods, procedures, and criteria to be developed for routine monitoring will consider simplicity and feasibility.

To achieve Objective 2 surveys will be undertaken at 7 or 8 cage farms in Greece situated at different depths, at varying degrees of exposure to wave action, over different types of substratum, with different production levels, farmed species and feeding practices. Environmental conditions in the water column and sediments will be measured in the vicinity of the cages and in control areas beyond the influence of the cages. Assessments will be made of the benthic infauna and epifauna and of wild fish populations in both control and impact areas. Modem multivariate methods will be applied to identify suitable eco-indicators for Mediterranean aquaculture and if applicable, methods for routine monitoring of indicator species will be suggested. Detailed feed and production data from the farms over the years prior to the surveys will be collected. The amount of waste feed and faeces released by bass and bream farms and the proportion consumed by free-living fishes will be assessed using sedimentation traps. For quantitatively assessing fish assemblages adjacent to farms and at unimpacted locations similar in depth, exposure and sediment, non-destructive visual census by SCUBA diving and underwater video will be used. Beside the traditional belt-transect-technique, the applicability of state-of-the-art theory of distance sampling techniques and analyses will be tested. These various information sets will be analysed and collated to provide both an overview of the environmental impact of cage farms in the eastern Mediterranean and an appropriate data set to drive the predictive model, which is the subject of Objective 3. Where interactions between aquaculture and the environment are undesirable (e.g. eutrophication, loss of biodiversity) suggestions will be given as to how these effects might be mitigated. On the other hand consideration will be given to how desirable interactions (e.g. fast biodegradation of solid wastes) might be maximised. All suggestions for improvements in aquaculture husbandry will consider both ecological and economic aspects.

Objective 3 will be realised by developing a model designed to predict the environmental impacts arising from fish cage farms in non-tidal waters. Special emphasis will be given to the effects on the environment of solid wastes from cage farms of sea bass and sea bream (uneaten feed and faeces) and on the interactions between farmed and free-living fishes, focusing on farm effluent flux and biodiversity. This will build on the experience gained in modelling impacts at salmonid cage farms in Scotland where model simulations of environmental impacts are now a key component of fish farm monitoring. Current models configured for tidal waters include modules to estimate sedimentary deposition areas and rates and the response of benthic infauna under varying feed inputs and production levels. In order to simulate impacts in non-tidal Mediterranean conditions a similar modular approach will be taken to allow for the easy inclusion of terms which might be significant such as consumption of waste food particles by wild fish around the cages. However considerable reconfiguration and development of the particle deposition and distribution aspect of existing models will be necessary in order to ensure that model outputs truly reflect the complexity of Mediterranean conditions. Emphasis will be given to ensuring that model output is sufficiently sensitive to simulate worst-case scenarios in order to safeguard the interests and needs of both the industry and the regulatory authorities.
1. Guidelines for monitoring: Guidelines and protocols detailing appropriate and cost effective suite of techniques for routine surveillance of the environmental impact of Mediterranean cage farms were developed. The scale and complexity of these are linked to the production levels of each farm, the purpose of monitoring and to the need for appropriate data to drive the predictive model. Sugested field procedures were provided as an annex to the guidelines. Farm discharges have their greatest impact in the vicinity of the fish cages and decline with increasing distance from the cages, creating a strong gradient of effects in both chemical conditions and faunal populations in the sediments. Concepts for defining effect zones and an allowable zone of effect (AZE) are discussed in relation to site monitoring and the predictive possibilities given by the MERAMOD model;
2. Impact of aquaculture on the environment: - Changes in redox potential and other geochemical parameters as a function of distance from the cages and as a function of expected deposition rates show the same trends at all sites but are not fully comparable. No seasonal differences were encountered. - Bottom photography and sediment profile imaging techniques can be used to discriminate between heavily affected site and unaffected one, but are less useful for monitoring in the intermediate zones. - Spatial variability of benthic macrofauna is a very significant indicator of environment conditions. A number of indicator species and families for enriched and non-enriched conditions were identified. - The nematode/copepod ratio is not a reliable indicator for enrichment - Cost-benefit analyses showed that macro-benthos samples using small corers, sieved at 1 mm and identified to family level provides significant information for monitoring purposes, but at much lower cost than (traditional) full quantitative benthos surveys - A number of wild fish species may be used as indicators for environmental quality at farming sites. The abundance/biomass of Mugilids and Sparids may be used as an indicator for the organic input of the fishfarms to the ecosystem - There are indications of mass escapes from fish farms. - Observations and experiments showed that practically all feed waste from the cages is consumed by wild fish;
3. Modelling: - Particle-settling experiments were performed for different fish feeds and faeces from sea bass and sea bream. Settling velocities vary significantly according to fish size and species. Fish species, fish size and biomass can be specified in the model for each individual cage. - A wild-fish module was build into the model in order to take care of the removal of wasted fish pellets. - In order to be able to use the model for short time flux studies, the ability to specify feeding times and rates was build into the model. - Model validation resulted in acceptable agreement between observed and modelled values. The detail of husbandry data used in the model effects predictions significantly. - The validation of the benthic response model sjowed that the number of species, abundance, abundance/species index, and the Shannon-Wiener index show the best correlations while biomass, Simpson index and Redox provided a higher degree of scatter. Useful relationships were established between modelled flux and the indicator species that were identified based on the quantitative benthic surveys. However, indicator families showed less useful. Species indicative of unpolluted sites are expected to be absent where deposition is greater than 1500 g m2 yr-1. The majority of indictor species for enriched conditions are expected to be absent where deposition is less than 1500 g m2 yr-1. - MERAMOD was used to simulate various management practices and predict their effect on sediment deposition. These scenarios show deeper, dispersive sites result in less severe impact over a larger area. In addition, spacing out of cages reduces predicted deposition markedly especially where a large spacing is used. The modelling also suggests bass potentially have more impact than bream as a result of faster faecal settling velocities, despite the slightly lower feed input used for bass. The effect of inefficient feeding and high stocking density is clear. A more severe impact over a larger area will result, with a higher probability of problems with sediment and fish health. Little difference was found between the scenarios where feeding method was tested. However, where a strong diurnal pattern of wind and circulation exists at a site, the effect of feeding larger portions by hand in two feeding events may result in periods of higher deposition. - MERAMOD is capable of predicting the depositional foot print from cage installations and will enable the regulators to agree with a farmer a practicable Allowable Zone of Effect based on survey data of key hydrographic and sedimentary chemical and biological parameters.

Funding Scheme

CSC - Cost-sharing contracts


Polar Environment Centre

Participants (5)

Main Port Heraklion
Institut für Meereskunde an der Universität Kiel, Anstalt des Öffentlichen Rechts
Düsternbrooker Weg 20
24105 Kiel
United Kingdom
Prospect Place The Hoe
United Kingdom
Dunstaffnage Marine Laboratory
United Kingdom
Argyll Square
Oban, Argyll