Regime-shifts in the southern Benguela shelf and inshore region

Over the past two decades, several species have undergone distributional shifts in the southern Benguela. The commercially-important West Coast rock lobster Jasus lalandii is one of these, and its shift in distribution has had profound effects on the rest of the ecosystem along the south-west coast. Reasons for these shifts are not fully understood, but are probably linked to changes in environmental conditions. We applied a sequential t-test algorithm for analyzing regime shifts (STARS) to physical (wind and upwelling) and biological (rock-lobster catch and growth, Bank Cormorant abundance) data for the southern Benguela inshore region and performed sensitivity tests on each of the variables. Regime shifts were defined as ‘robust’ or ‘possible’ if they were detected for ⩾70% or ⩾60% of the sensitivity tests respectively. To corroborate the shifts detected by STARS, we then applied two additional methods: change point analysis and the Chow breakpoint test. The STARS method outperformed the other two methods because it could handle shorter time series and detect shifts towards the end of the time series, but most of the significant shifts detected by STARS were also detected by one or both of the other methods. A significant shift in Cape Point winter winds occurred in 1983, an El Niño year. However, measurement methodology changed during the same period and this is discussed in relation to the shift. A decline in rock-lobster growth rate occurred in the mid 1980s, and significant increases in upwelling variability and mean summer winds were detected in the early-to-mid 1990s – the same period in which rock lobster abundance underwent an eastward shift, declining on the west coast and increasing on the south-west coast. Bank Cormorants underwent respective declines and increases in the mid-to-late 1990s on the west and south-west coasts following the shift in lobsters. Further shifts in mean wind and upwelling were detected in the 2000s. These results are a step towards developing ecosystem indicators for the region, which can be monitored with the intended use of providing an early warning system for long-term, ecosystem-scale changes.

► We collate environmental and biological data for the southern Benguela inshore region.
► We apply three statistical methods to these data to detect shifts in the time series.
► Significant changes in wind and upwelling were detected in the 1980s and 1990s.
► All three methods detected similar results.
► We discuss these environmental changes in relation to ecosystem changes.