A statistical approach is proposed to utilize the information contained both in scientific surveys and commercial fishing logbooks, in order to determine spatial and temporal distributions of recruiting and spawning fish of exploited populations. The approach is based on multivariate descriptive methods including ordination methods and classification techniques. A typology involving a principal component analysis (PCA), followed by a hierarchical ascending classification (HAC), is applied to both scientific and commercial data. PCA that accounts for spatial and temporal contiguities were performed. These analyses allow to compare structures observed at global and local scales, with those observed without focusing on a particular scale. The method is... |