Interim Report November 2002
Scoping Study of the carrying capacity
for bivalve cultivation in the coastal waters
of Great Britain
MJ Kaiser and HA Beadman
Contents
1.Outline of the project 2
2.Definitions 2
3.Review of approaches used to determine carrying capacity 3
of coastal waters for bivalve cultivation
4.Assessment of the present ability to predict:
4.1.Carrying capacity of bivalve production 7
4.1.1.Dynamic Energy Budget Models 9
4.1.2.Models relevant to the management of 14
mussel production
4.1.3.Conclusions16
4.2.Environmental impacts16
4.2.1.Seed collection17
4.2.2.On-growing19
4.2.3.Harvesting 24
4.2.4.Management Considerations24
4.2.5.Conclusions26
5.Recommendations of field and laboratory investigations
required to provide indicators of carrying capacity of UK
coastal waters for bivalve cultivation
5.2.Aim and Rationale27
5.3.Approach29
6.References31
1.Outline of the scope of the report
The main objectives of the project are:
1.To prepare a review of the approaches used to determine the carrying capacity of coastal waters for bivalve cultivation.
2.To assess the present ability to predict carrying capacity in terms of both shellfish production and ecological capacity (i.e. environmental impacts), and the specific methods that are employed.
3.To produce recommendations of field and laboratory investigations required to provide indicators of the carrying capacity of UK coastal waters for shellfish cultivation.
2.Definitions
In accordance with the Scottish Parliament Environment and Transport Committee in its 5th Report 2002 (Report of Phase 1 of the Inquiry into Aquaculture) the following definitions have been adopted:
Carrying Capacity of a defined area refers to the potential maximum production a species or population can maintain in relation to available food resources within an area.
Assimilative Capacity is the ability of an area to maintain a ‘healthy’ environment and ‘accommodate’ wastes.
Environmental Capacity is the ability of the environment to accommodate a particular activity or rate of activity without unacceptable impact.
3.Review of approaches used to determine carrying capacity of coastal waters for bivalve cultivation
The main driving force behind the development of models to determine the carrying capacity of waters for bivalve cultivation has generally been for commercial interests. As a result the models to date have focused upon determining the stocking density at which production levels are maximised without negatively affecting growth rates (Carver and Mallet 1990). This has therefore entailed study upon the bivalve population and the populations (e.g. phytoplankton) and processes (e.g. tidal currents, flushing) that may effect bivalve productivity rather than the effect that the bivalve population may have on other parts of the ecosystem.
The approaches towards determining carrying capacities of bivalve populations can be characterised into three categories: 1) empirical studies, 2) calculation of budgets and 3) simulation modelling (Grant et al.1993). Empirical studies generally correlate some aspects of shellfish growth to the food supply and/or environmental variables. For example Smaal and Van Stralen (1990) related carrying capacity to primary production, Grizzle and Lutz (1989) related growth to tidal currents, seston and bottom sediments and there are many other examples in the literature (e.g. Wildish and Kristmanson 1979, Officer et al. 1982, Smaal et al. 1986, Grant et al. 1990, Heip et al.1995). However, these models are limited in use due to their restricted temporal and spatial scales. The spatial variability of both the biological demand and the physical characteristics of the system are not detailed (Raillard and Mengesguen 1994). Feedbacks within the systems of the impact of shellfish culture on food sources and regeneration of food within the system are also not included. Nonetheless, Dame and Prins (1998) used an empirical approach to compare the carrying capacity of 11 ecosystems with a dominant bivalve population, in which carrying capacity was defined in terms of water mass residence time, primary production time and bivalve clearance time. While recognising the limitations of this method it did allow estimations to be made of the requirements of massive and successful bivalve populations in terms of water residence time and primary production time. This method therefore provided a way in which a prospective system could be compared to known and extensively studied systems to determine whether a pilot cultivation study would be worthwhile (Dame and Prims 1998).
The determination of carrying capacity through calculation of budgets estimates the balance between production of phytoplankton and ingestion of the bivalve populations (e.g. Carver and Mallet 1990). More complex ecosystem budgets that include more components, such as the benthos and detritus, have been developed providing a better representation of the ecosystem by acknowledging the complex interactions between the various components within the system (e.g. Rosenberg and Loo 1983, Rodhouse and Roden 1987, Grant et al.1998). Rosenberg and Loo (1983) produced an energy-flow diagram for mussel long-line culture in a Swedish outlet over 571 days of cultivation from seeding to harvest. From this study they were able to predict that mussel culture numbers could be nearly doubled. Rodhousen and Roden (1987) created a more comprehensive carbon budget, for the Killary Harbour System, with 14 carbon components and 5 spatial sectors. This budget was then used to estimate the carrying capacity of the system for mussel cultivation. Grant et al. (1998) employed a similar method to determine a carbon budget for mussel floating raft culture in a bay in South Africa, and were able to identify the potential for further intensification of the present mussel production. However, the extent of the use of these models is limited, as they do not directly allow for feedback mechanisms since the components are averaged over time. Again there is no detailed spatial variability, and no allowance for interaction between sectors where an area has been divided into smaller spatial units. This lack of spatial variability presents a particular problem in predicting where to locate new cultivation if it has been established that the system can tolerate increased bivalve production. Ecosystem budgets do, however, potentially provide a way of estimating the impact of increased mussel production on other components of the ecosystem that are included in the budget such as the benthos.
The third approach towards modelling carrying capacity of bivalve cultivation is the use of simulation models. A simulation model is described as a model in which the culture ecosystem is viewed as distinct compartments of state variables (e.g. shellfish, phytoplankton), between which flows of energy or material are quantified based on internal biological fluxes mediated by external forcing functions (Grant et al. 1993). Models have been developed for both oyster (Raillard and Menesguen 1994, Bacher et al.1998, Ferreira et al.1998) and mussel cultivation (Grant et al.1993, Dowd 1997, Campbell and Newell 1998).
Oyster cultivation models have been developed for the Marennes-Orleron Bay, the most important shellfish culture site in France (Bacher et al. 1998). Raillard and Menesguen (1994) produced a box model of the site incorporating both physical and biological processes: horizontal transport of suspended matter, feeding and growth of the oyster Crassostrea gigas and primary production. Some of the biological features and physical features of the bay were accurately reproduced and the hydrodynamic regime, due to the short resident time, was found to be the determining factor that controlled carrying capacity. The model was able to predict a density dependent effect of oyster growth, with a reduced maximal dry weight with increased standing stock. The validity of the model was limited mainly by the description of the physical transport of suspended and deposited matter (Raillard and Menesguen 1994). The model of Bacher et al.(1998) has further developed the model of Raillard and Menesguen (1994), and is again based on a box model approach including nutrient inputs, mixing and transport by the currents, the turbidity level and the ecophysiology of the oysters. The model has been modified to include other parts of the bay, and each oyster age class within each box has been given it’s own dynamics. The model still predicts that the system relies on water exchange for phytoplankton and is still sensitive to the standing stock. Further refinement of the model would require greater knowledge on both oyster physiology, particularly gametogenesis, and population effects such as density dependence and mortality (Bacher et al. 1998).
An ecological model has also been developed to estimate the carrying capacity for oyster cultivation of Carlingford Lough, Ireland (Ferreira et al.1998). In this box model Carlingford Lough has been divided into 3 boxes with transport of particulate and dissolved substances between each box. The ecosystem is divided into objects that represent the different functional compartment in the model of: forcing functions (river flow, temperature, light), advection-dispersion, suspended particulate matter, phytoplankton, oysters (at the physiological and population level) and man (different management strategies). The model was able to predict that oyster cultivation in Carlingford Lough is below the level where growth is inhibited by stock density and that a five-fold increase in seeding would maximise oyster production. A limitation of the model is that of the small number of boxes in the model that may cause some bias of the result due to the positioning of the present cultivation close to the boundary of boxes 2 and 3. The methodology of this model and the Bacher et al.(1998) model are very similar as both are based on the coupling between physical and biological processes at the spatial scale of several kilometres in length (Bacher et al.1998). The major difference between them is in the population dynamics where 40 weight classes are used in the Carlingford Lough model compared to the 10 age classes in the Marennes-Orleron model. The difference in the carrying capacity of the two systems can be explained by the difference in biological and physical flows derived form the model outputs (Bacher et al. 1998).
The modelling of carrying capacity for mussel cultivation has been undertaken for both longline culture (Grant et al.1993, Dowd 1997) and a bottom culture site (Campbell and Newell 1998). Longline culture is examined by Grant et al.(1993) and Dowd (1997) who describe the same model developed for a bay in Nova Scotia, Canada. A box model approach was used to represent the system. Particle exchange was allowed between each section, which contained the state variables: seston, zooplankton, phytoplankton and mussels. The model was reasonably successful and highlighted the relative importance of the internal ecology of the cove versus exchange during different times of the year Dowd (1997). It was also able to estimate the different carrying capacities for different parts of the cove, estimating that the outer cove would have a carrying capacity an order of magnitude greater than the inner cover due to flushing time. Unlike the model of Raillard and Menesguen (1994) in which model uncertainties were placed with the description of particle exchange and mixing processes, sensitivity in this model is as a result of uncertain physiological parameters. This is discussed in more detail in the next section. The model also only includes population interactions as a function of a reduced food supply with increasing standing stock and does not include other density-dependant effects.
Campbell and Newell (1998) developed a carrying capacity model for a bottom culture site in Main, USA. This model was developed to be as simple as possible regarding both mussel physiology and physical parameters. Mussel production was simulated using site conditions (water depth, current speed and mussel mortality), forcing functions (temperature and food supply), and initial conditions (seeding date, seed weight and length and seeding density). A sensitivity analysis of the effects of seeding density was conducted and used to determine the carrying capacity of the three test sites. The development of the model also demonstrated the importance of food quality and quantity in explaining mussel growth. However, the model was not able to accurately represent mussel growth over the entire range that mussel are cultured, and this is possibly as a result of the ecophysiological simplifications.
The models developed to determine bivalve carrying capacity range from simple correlations between forcing functions and mussel growth, to ecosystem carbon budgets to simulation models. Carbon budgets have provided the only means by which the effect of bivalve cultivation on other components in the ecosystem can presently be modelled. However, through allowing for no inclusion of spatial or temporal variability, or feedbacks within the system they provide limited use towards the comprehensive calculation of bivalve carrying capacities. Simulation models do allow for more temporal and spatial variability and have provided reasonable estimates of carrying capacity for bivalve cultivation. It would be possible to develop these models to address the impacts on other components within the ecosystem, but this would require a large amount of research in order to produce a full ecosystem model for every potential bivalve cultivation site. Therefore, while carrying capacity can be established to a certain degree of accuracy there is no specific method for the determination of assimilative or environmental capacity through modelling. There are still limitations to the prediction of carrying capacities by simulation models concerning spatial variability. The model of Campbell and Newell (1998) does not allow for spatial variability over the mussel bed, and in box models (e.g. Raillard and Menesguen 1994, Bacher et al.1998, Ferreira et al.1998) spatial variability is restricted to box size, which can be kilometres in length. Smaller scale variability within the mussel bed, or cultivation area, is not accounted for by any of the present models and this will have repercussions in predicting the variability in flow and hence food supply and resultant growth of bivalve populations.
4.Assessment of the present ability to predict:
4.1.Carrying capacity of bivalve production
The production of cultivated mussels in the UK far exceeds cultivation of other bivalves (Figure 1). For this reason mussels have been focused upon in this report, although the same approaches are equally applicable to other bivalve production. The previous section has highlighted that simulation models present the most probable approach towards modelling carrying capacity. Minimum requirements for such carrying capacity models can be detailed as transport processes, sediment dynamics and submodels for organism and population level processes (Smaal et al.1998). These models will permit production (growth and reproduction) to be forecast as a function of food supply and other environmental factors. Complicated interactions at the population level (mortality - self-thinning and predation) requires that models of individual production are integrated (e.g. Dynamic Energy Budget (DEB) models) with models that describe the consequences of these processes on the production of mussels at the population-level. To better describe the levels of model complexity Figure 2 illustrates a hierarchy of modelling, demonstrating how with a need to represent important processes at higher levels in the hierarchy (e.g. population level), the potential complexity of the modelling task increases. As a consequence there is a need to consider the appropriate level of detail required of the physiological DEB model, while meeting the objectives of a useful and ecologically relevant management tool. It has therefore been considered most relevant to assess the current ability to predict mussel production using models based upon a Dynamic Energy Budget approach.
Figure 1.Farmed shellfish production for the UK in 2000 (Shellfish Association of Great Britain)
Figure 2. Hierarchy of Mussel Modelling, showing components at each level and interactions between levels. The strength of interactions is indicated by the thickness of the arrows.
Collectively the physiological components determine the size and reproductive capacity of a mussel. The individual mussels in turn interact at the population level influencing both the size structure and recruitment to the population. The mussel population will have a limited effect on the ecosystem through providing a food source for predator populations and altering the local topography. The ecosystem provides the largest influence from population level to component level as the driving force providing food and the environmental conditions in which the mussel population is situated.
4.1.1.Dynamic Energy Budget Models
There are a number of DEB models specifically designed to represent mussel growth. Each of these seeks to represent mussel growth as the balance between components of feeding, respiration and reproductive output. Within each model this is achieved with differing levels of complexity of mussel physiology and by the inclusion of various physical and biological factors (Table 1). The differences between each of the models occur as a direct result of the approach taken during the development of the model and are to some extent dependent on the specific aim/s of the model.
The most sophisticated model, in terms of physiological complexity, is that of Scholten and Smaal (1998). This model was developed to simulate the growth and reproduction of a subtidal mussel incorporating all the available ecophysiological knowledge. Specifics in the model are detailed from filtration through to ingestion, absorption (incorporating the optimal feeding model of Willows 1992), respiration and excretion. Energy flow is represented by carbon and nitrogen fluxes between the five main compartments in the model: blood, body tissue, storage products, the organic component of the shell, reproductive tissues and activity (gametes and spawning). Growth and reproduction are ascertained from the rates and efficiency of the physiological processes that vary with seasonal variation in temperature, food quality and quantity, and metabolic demands. The incorporation of all the available knowledge on the ecophysiology of mussels resulted in a highly complex and over parameterised model, which is difficult to calibrate (Scholten and Smaal 1998). The complexity has also made the model unidentifiable i.e. there are redundant or ambiguous hypotheses within the model (Scholten and Smaal 1998), and this must be addressed before further meaningful development of the model can occur. However, Scholten and Smaal (1998) state that at present there is insufficient knowledge of mussel ecophysiology to rectify the situation. Nonetheless, the model predicted growth well for the site for which it had been calibrated and moderately well for another site with a high seston level. However, it was not successful in predicting growth at an alternative site that had a low seston and food input. This may be as a result of the adaptation of the mussels to their environment of low total particulate matter (TPM). To overcome this problem would require either a separate calibration of the model with adapted mussels for use in low TPM environments, or further complexity added into the model to account for mussel functions altered by the adaptation to low TPM.