Selecting and combining survey based indices of fish stocks using their correlation in time to make diagnostic of their status
Résumé
Research fisheries monitoring surveys provide an ensemble of measurements on fish stocks and their environment. Presently fish stock diagnostics are based on the evaluation of abundance and the survey data contribution is through abundance indices only. Here we shall consider a large set of survey-based indices that represent a target stock over its time series through its spatial pattern (longitude of the center of gravity, latitude of the center of gravity, inertia, anisotropy, positive area, equivalent area, spreading area, and microstructure index), its vital traits (Lbar, L25, L75, L at 50% maturity, and Z), and its abundance (logarithm of the abundance of the recruits, logarithm of the abundance at age or pooled age, and logarithm of the total abundance). The present work aims at establishing a procedure that analyses these indices in order to diagnose and forecast the state of a stock. Because the interannual variability in survey based indices is high and because diagnostics on fish stock cannot be based on noise, the first concern is to select and combine those indices that support a reliable diagnostics, thanks to their continuity in time. First, the set of indices is reduced by selecting the ones that present the highest autocorrelation in time. Then Min/max Autocorrelation Factors (MAF) are used to combine the indices into factors that present the maximal continuity in time, and that will be used here to monitor the population status. Finally, the population status forecast to the next year is attempted based on the forecast potential of the MAF method when modelling the time continuity. The procedure will be applied to North Sea cod and the Bay of Biscay anchovy, two stocks that have experienced a decline and for which survey series exist for more than 10 years. Results are presented and discussed in view of the present ICES assessments, which demonstrates the potential of the proposed fishery independent indicator based diagnostic procedure.