SCIENTIFIC COMMITTEE FOURTH REGULAR SESSION 11-22 August 2008 Port Moresby, Papua New Guinea SENSITIVITY OF BIGEYE STOCK ASSESSMENT TO ALTERNATIVE BIOLOGICAL AND REPRODUCTIVE ASSUMPTIONS. WCPFC-SC4-2008/ ME-WP-1

Executive Summary The outputs from mathematical and statistical models can be influenced by uncertainty in the estimates of the parameters used (known as parameter uncertainty) and to the methods and assumptions used to construct and link parameters in a model (known as structural uncertainty). Sensitivity analysis was applied to the current bigeye stock assessment to ascertain the influence of the structural assumptions on the reference point outputs of the model. The analysis examined the influence of alternative estimates of natural mortality, fecundity at length, spawning fraction at length and alternative maturity schedules. The effect of an alternative growth curve, and an alternative steepness assumption, were also assessed. The current stock assessment indicates that the fishing mortality exceeds F MSY , and that the biomass is approaching MSY (Langley et al. 2008). The model is more strongly influenced by precision in CPUE and length frequency data than by the reproductive and growth parameters directly. This sensitivity analysis demonstrates that the model is also sensitive to the structural assumptions associated with estimation of the reproductive and growth parameters. Alternative estimates for all reproductive and growth parameters and natural mortality influenced the spawning biomass reference points (SB current /SB MSY and SB current /SB 0) typically by more than 10% and influenced biomass (B/B MSY) and the F multiplier reference points by between 1 % and 5 %. The results support the need for further investment in knowledge acquisition to reduce the current level of uncertainty.