A cross-ecosystem comparison of spatial and temporal patterns of covariation in the recruitment of functionally analogous fish stocks

Temporal and spatial patterns of recruitment (R) and spawning stock biomass (S) variability were compared among functionally analogous species and similar feeding guilds from six marine ecosystems. Data were aggregated into four regions including the Gulf of Maine/Georges Bank, the Norwegian/Barents Seas, the eastern Bering Sea, and the Gulf of Alaska. Variability was characterized by calculating coefficients of variation and anomalies for three response variables: ln(R), ln(R/S), and stock–recruit model residuals. Patterns of synchrony and asynchrony in the response variables were examined among and between ecosystems, between- and within-ocean basins and among functionally analogous species groups using pair-wise correlation analysis corrected for within-time series autocorrelation, multivariate cross-correlation analyses and regime shift detectors. Time series trends in response variables showed consistent within basin similarities and consistent and coherent differences between the Atlantic and Pacific basin ecosystems. Regime shift detection algorithms identified two broad-scale regime shift time periods for the pelagic feeding guild (1972–1976 and 1999–2002) and possibly one for the benthic feeding guild (1999–2002). No spatial patterns in response variable coefficients of variation were observed. Results from multivariate cross-correlation analysis showed similar trends. The data suggest common external factors act in synchrony on stocks within ocean basins but temporal stock patterns, often of the same species or functional group, between basins change in opposition to each other. Basin-scale results (similar within but different between) suggest that the two geographically broad areas are connected by unknown mechanisms that, depending on the year, may influence the two basins in opposite ways. This work demonstrates that commonalities and synchronies in recruitment fluctuations can be found across geographically distant ecosystems but biophysical causes of the fluctuations remain difficult to identify.