Towards the development of an environmental rule-based model for predicting recruitment in Brazilian sardine Sardinella brasiliensis

A rule-based model was built in order to analyse the effects of environmental variables (wind stress, Ekman transport, turbulence, and sea surface temperature [SST]) on the recruitment of the Brazilian sardine Sardinella brasiliensis. The purpose of modelling was (1) to test the optimal environmental window hypothesis (characterised by ‘dome-shape' formats) for the variables studied and (2) to evaluate the application of the model as a tool for recruitment prediction. The model was run under different conditions and the following results were obtained: (1) when environmental variables only were used as inputs, the model response was poor; (2) when biomass was added as an input, and the dome-shape effect enhanced, the model's results were good; and (3) when the dome-shape effect was excluded, the model response was intermediate. These findings support the hypothesis of an ‘optimal environmental window' for wind and SST, and that incorporation of such relationships leads to better predictions of recruitment. Hence, although the model was calibrated with a relatively short time-series, it may prove to be a useful decision-support tool to define more flexible limits for allocation of fishing effort. Keywords: models, optimal environmental window, prediction, recruitment, rule-based, Sardinella, south-eastern BrazilAfrican Journal of Marine Science 2005, 27(3): 539–54