Structure spatiale et cartographie des peuplements benthiques du bassin de Marennes-Oléron: première approche par la méthode du krigeage
Résumé
The benthic communities were sampled in the Marennes-Oléron Bay in the spring of 1984 so that their importance as competitors of the cultured mollusc populations could be studied. A stratified random sampling survey was designed to sample 370 units in a 180 km2 area. Two groups of mollusc species were distinguished on the basis of different feeding strategies: deposit-feeding and suspension-feeding bivalves. The biomass of these feeding groups was mapped after kriging estimation of the logarithmic transformed biomass. The spatial structure for each group was represented by three different models: generalized covariance, unidirectional semi-variogram, and
pluridirectional semi-variogram. Each model was used for the punctual kriging of the 693 nodes of a regular grid superimposed on the surveyed area. The variation in the kriging estimations allowed comparison of the three models. The spatial structures were different for the two groups. For both groups, estimation of the best generalized
covariance function validated the intrinsic hypothesis (intrinsic random function of order 0). Therefore the spatial structure of deposit feeders was not affected by anisotropy, and a spherical semi-variogram with a nugget representing 32% of the variance and a 5 km range was fitted to the data. Mapping of deposit feeders showed
that the maximum biomass was located at the northern and eastern parts of the bay. In contrast, the spatial structure of suspension feeders showed geometric anisotropy. Modelling the semi-variogram was achieved with the superimposition of two nested spherical models. A stretching of the north to south direction enabled us to take into
account the anisotropy. This computation resulted in two spatial scales: 4 km along the northern direction and 2km along the eastern direction. Deriving the map of the biomass from this model showed that the maximum biomass areas of suspension feeders were different from those of deposit feeders and were located in the western
part of the bay. These results were considered in relation to the abiotic and biotic factors controlling the zoobenthic communities. Some points about the spatial and temporal variability were also raised and a new study was defined on the global amount and spatial structure of the competitor’s populations. This study therefore provides
information on the importance of the stratification and on the suitable spatial scale that can be incorporated into a new sampling strategy.