Accounting for linefish dependency in management of the South African small pelagic fishery

To further implement an ecosystem approach to management of the small pelagic fishery in South Africa, we attempted to develop functional relationships between spatialised time-series of the biomass of three small pelagic fish species and the catch per unit effort (CPUE) of four linefish species (i.e. species caught on hook and line, excluding longlines) that could be used to account for the forage needs of dependent predators. Granger causality tests indicated significant effects of anchovy Engraulis encrasicolus biomass on silver kob Argyrosomus inodorus and yellowtail Seriola lalandi CPUE, significant effects of sardine Sardinops sagax biomas on geelbek Atractoscion aequidens CPUE, but no significant effects of biomass of west coast redeye round herring Etrumeus whiteheadi and east coast redeye round herring Etruemus wongratanai on any linefish CPUE. Sequential t-test analysis of regime shift (STARS) indicated temporal matching in regimes between anchovy biomass east of Cape Agulhas (EoCA) and yellowtail west coast (west of Cape Point) CPUE, and between sardine biomass EoCA and geelbek southwest coast (Cape Point to Cape Infanta) CPUE. Convergent cross mapping found no causality for anchovy biomass EoCA on yellowtail west coast CPUE but causality in the reverse direction. A high degree of causality in both directions for sardine biomass EoCA and geelbek southwest coast CPUE was observed, indicating strong predator–prey interaction, and a highly significant linear relationship between sardine biomass EoCA in yearn–1 and geelbek southwest coast CPUE in yearn was obtained. This functional relationship should be considered for inclusion as a proxy for ecosystem state in management procedures that adopt the ecosystem approach to fisheries for the South African small pelagic fishery. Keywords: convergent cross mapping, CPUE, ecosystem approach to fisheries, geelbek, Granger causality, marine resources, predator–prey interaction, sequential analysi