Résumé
A comparative study, combining three different assessment methods (fish gillnet sampling, artisanal fisheries surveys and hydroacoustics) was conducted in Mali where two man-made reservoirs (Sélingué and Manantali) are particularly suited for investigating the impact of fishing effort on the fish assemblage. These two ecosystems have relatively similar areas, edaphic and environmental properties but are subjected to different levels of fishing exploitation (low at Manantali, high at Sélingué). The comparison is based on two indicator parameters: the abundance indices and the size spectra distributions, obtained by the three methods at two contrasting hydrological seasons (April and October). The results were compared first between the two seasons, and then between the two lakes. The present work is based on two main hypotheses: (1) that there is a higher fish abundance in October associated with smaller overall sizes, after spawning; (2) a lower abundance and smaller sizes in the Sélingué reservoir than in Manantali, because of the much higher fishing pressure in Sélingué. The relevance of each method to the selected indicators is discussed. On the one hand, the three methodologies on the whole gave similar conclusions and they also complement each other. On the other hand, some results do not match the hypotheses because of biases due to difficulties and technical limitations of each method in such ecosystems (shallow water with vegetation and stumps of former forests)