Cause – effect chain modelling of sand mining using mussels

Lessons learned

In DEB models, a relatively simple description using a scaled functional response is applied. The scaled functional response describes the energy uptake rate as a function of food concentration. The effect of suspended sediment concentration can be incorporated in the functional response. The low concentrations of particulate inorganic matter have an effect on the food uptake rate. Filter feeding mussels need to invest time and energy processing the inorganic matter. Short-term increases of particulate inorganic matter have less impact on the growth performance of the shellfish compared to a continuous increase. The timing of dredging activities is also important. In the winter period, when the activity of the mussels is low, the impact of increased suspended sediment concentration is much less than during summer time. Because mussels are able to adapt to different silt concentrations in the water column, it is important that impact studies take into account the natural variability of silt concentrations the area, as well as the time spans during which these conditions may remain modified. Short term increases in suspended sediment concentration will have less impact on the food intake of shellfish than a continuous release.

This study is primarily focused on blue mussels as a model species for suspension-feeding lamellibranchiate bivalves. It is assumed that the processes will be comparable for other species and that only the values of the parameters will differ. However, it is good to check this assumption using literature data from other species.The models that were used in this case study were not directly calibrated with field observations and literature information on filtration rates and pseudofaeces production. It would be an improvement to perform an additional calibration with the appropriate data.

Case study conclusions

Under natural conditions, many factors may influence the filtration rate of bivalves. Feeding under laboratory conditions may not always accurately reflect in situ filtration where a wide spectrum of changing environmental factors and species interactions may influence the feeding behavior. The present case study is believed to reflect important basic features of mussels’ feeding behavior in nature where phytoplankton is the main source of nutrition. Among the many parameters that may affect the in situ feeding behavior, phytoplankton biomass (expressed as Chl a concentration) seems to be the most important. Yet, high concentrations of silt/seston leading to preingestive rejection/pseudofaeces production, for instance, may also affect the feeding of mussels in estuaries and exposed coastal waters. 

Although there is still no general agreement regarding physiological control of water pumping in response to (very) high concentrations of particles in the ambient water, present consensus tends to be that the filtration rate is high and constant, between a lower critical level and an upper seston concentration threshold. It remains to be clarified if the reduced filtration rate at high seston concentrations is caused by physiological regulation (supporting maximum assimilation and growth) or overloading (adversely affecting food uptake and growth). The impact of changes in silt and/or phytoplankton concentrations on the growth of an individual blue mussel (Mytilus edulis) can be modeled in a deterministic way, using the DEB-model. In cases where worst case or conservative assumptions are made to model ecological impacts deterministically, however, the use of a probabilistic instead of a deterministic approach can have several advantages (Van Kruchten 2008). For example: in a probabilistic approach worst-case assumptions can largely be prevented by incorporating the uncertainty itself in the cause-effect chain modeling. In such case, deterministic modeling may lead to an overestimation of the ecological effect, whereas the probabilistic modeling results give information on the probability of occurrence of possible effects.

Although in the DEB model uncertain parameters or variables can be identified, realistic instead of conservative assumptions are made to deal with these uncertainties. In order to quantify the uncertainty, Monte Carlo simulations with varying parameter values can be made. The added value of applying a probabilistic analysis to this case is limited and will not make a difference between a highly conservative and a realistic estimate of the impact on mussels. A probabilistic analysis might be used to quantify the uncertainty margin of the final prediction. Because a probabilistic analysis is quite laborious, however, a sensitivity analysis instead of a probabilistic analysis is recommended to give insight into the uncertainty margins of the results.