Relationship between survey abundance and weight indices, survey design and wind conditions: the French groundfish survey in Bay of Biscay

No abstracts are to be cited without prior reference to the author.

Abundance indices derived from scientific trawl surveys are used to inform on population trends either directly, e.g. survey based assessment methods and the indicator approach, or indirectly for tuning stock assessment models. The implicit assumption of such use of survey indices is that they reflect true trends. Furthermore, inter-annual variability in survey indices is partly attributed to sampling variability and partly to yearly differences in recruitment or survey coverage of the stock. In certain cases indices for a given year might be qualified as outliers if they seem too variable to reflect true abundance changes. If the indices of several species appear to be outliers for the same year, the suspicion is raised that something happened during the survey of the given year. We refer to this effect as annual survey catchability effect whose main characteristics are that it affects many species simultaneously, appears on the scale of the whole survey and is not caused by technical problems. We studied annual survey catchability for the autumn Western IBTS groundfish survey in the Bay of Biscay for the period 1987 to 2003. For explaining annual survey catchability across species we derived several indicators for survey design and wind conditions. Survey starting dates had been gradually delayed by about one month which resulted in stronger westerly winds. Furthermore, gradually fewer hauls had been taken close to the coast. Using GAMs we found that a substantial part of interannual variation in abundance indices could be explained by those indicators, but not the overall time trend.