Modelling oyster population dynamics in a Mediterranean coastal lagoon (Thau, France): sensitivity of marketable production to environmental conditions

A population dynamics model was used to simulate variations in the standing stock of the Pacific oyster, Crassostrea gigas, to assess the marketable production in Than Lagoon (France) and to evaluate the sensitivity of this production to environmental conditions. The model is based on a continuous equation of the oyster abundance as a function of individual growth rate, inter-individual growth variability and mortality rate. The growth model depends on water temperature, particulate organic matter concentration and total individual mass. Inter-individual growth variability was introduced into the general population dynamics equation by a diffusion coefficient K that was set to 0.05. The population dynamics model took into account the two culture methods (i.e. collees and pignes oysters) and rearing strategies of oyster farmers by using timetables of seeding and harvesting obtained through interviews of oyster farmers. Distributions of standing stocks were obtained through assessments conducted in the lagoon in March 2000, October 2000 and March 2001 and were used to calibrate the model. The model estimated the total marketable production at ca. 17,900 t between March 2000 and March 2001. The major part of the production (ca. 70%) was in spring. Seventy percent of the annual production came from collees oysters. Sensitivity analyses showed that the key parameters are those related to harvesting. The model was used to evaluate the effects of different environmental conditions (e.g. a decrease in the oyster growth rate, a harvesting closure due a toxic algae bloom, a massive summer mortality due to an anoxic crisis) on short- and long-term variations in the standing stock and the production for both culture methods. A decrease in the growth rate of ca. 20% resulted in losses of 18% in the first year of production for both culture methods. Long-term simulations showed that the production of pignes oysters was more affected than that of the collees oysters (reductions of 26% and 4%, respectively). Simulated scenarios included a 2-month long harvesting closure (i.e. November and December) or a massive summer mortality (i.e. 45% and 20% for collees and pignes oysters, respectively). No longterm effect was predicted for either event, although losses of ca. 10% were estimated in the first year of production for both culture techniques. The model can be a useful tool for predicting marketable production of oysters as a function of rearing strategy and environmental conditions. (C) 2004 Elsevier B.V. All rights reserved.