Further SCAA/ASPM assessments of Gulf of Maine Cod including data for 2007 and exploring the impact of age-dependence in natural mortality

Abstract

The ASPM (SCAA) Gulf of Maine cod assessments presented at previous GARM meetings are extended to take account of data for 2007. In response to concerns about the reliability of earlier annual catch data, the assessments commence in 1964 rather than 1893; this necessitates careful consideration of the parameters that determine the starting numbers-at-age vector. Further, in response to recommendations from the Panel, scenarios are investigated which maintain asymptotically flat survey selectivity but increase the value of an age-independent M or admit an estimable increase in M with age for older ages. Either of these mechanisms is able to achieve model fits that do reflect the relative paucity of older cod in particularly the proportionsat-age data for the NEFSC surveys. From 16 assessments presented, a selection is made based on AIC, but factoring in concerns about the lack of direct evidence for the cryptic older fish implied by domed survey selectivity, and the biological realism of the higher values of age-independent natural mortality required to account for the relative paucity of older cod in the surveys. Specifically assessment “Ricker G”, which has asymptotically flat NEFSC survey selectivity and M increasing above 0.2 for ages above 4, is put forward as the basis for providing management advice. This option has an AIC-weight of some 1010 relative to the current default of flat selectivity and an age-independent M of 0.2, which is unable to fit the proportion-atage data for older ages adequately. The analysis also provides strong statistical evidence that selectivity for the commercial fishery is domed even if that for the NEFSC surveys is asymptotically flat. A number of reasons for preferring ASPM to VPA as the basis for providing management advice for the stock are put forward.