Performance of Surplus Production Models with Time-Varying Parameters for Multispecies Complexes

Single species surplus production models (SSPMs) are often used to assess multispecies assemblages in data-poor situations where catch and catch per unit effort (CPUE) information are insufficient to perform separate single species assessments.  A problem with modeling an assemblage is that if the component species differ in their productivity or vulnerability, the SSPM parameters that describe the multispecies assemblage would be expected to change over time.  We examined the ability of an SSPM with time-varying parameters to estimate the dynamics of a multispecies assemblage as well and that of its single species components.  Population dynamics were simulated for a series of three-species assemblages that differed to varying degrees in life history characteristics (growth rate and carrying capacity), catchability, magnitude and trend in exploitation rates due to fishing, and observation error in CPUE.  Several SSPMs with time-varying parameters were fit to simulated catch and CPUE data summarized across all three species.  Although in some scenarios time-varying parameters were not estimable, in other scenarios the addition of time-varying parameters improved the SSPM’s ability to estimate the dynamics of the multispecies assemblage.  However, using models with time-varying parameters does not ensure sustainability of all individual species in the assemblage.