Temporal patterns and regional comparisons of recruitment rates of United States fish stocks

Several previous studies of marine fish stocks have demonstrated time-varying recruitment productivity and indicated that including time-varying parameters can track process variation in recruitment. Few studies have synthesized signal-to-noise ratios and underlying reasons for time-variation across stocks and regions. Using Peterman's productivity method (PPM), we provide a broad synthesis of time-varying density-independent productivity in 84 stocks across five regions of the United States. Of all stocks investigated, 50 were found to have time-varying productivity, challenging assumptions on the stationarity of recruitment parameters and dependent reference points. Our results demonstrate the power of PPM for synthesizing the form and pattern of recruitment time-variation among regions, including general summaries of directional change over time. Furthermore, our results show regional differences in time-varying patterns, particularly the signal-to-noise ratio (SNR) of low- to high-frequency variation. The SNR was lower in the California Current region than in two Atlantic regions and two Alaska regions. Generalized linear modelling used to synthesize results suggests that stocks with higher contrast in spawning stock biomass over time, standardized regardless of actual spawning stock size, were more likely to have time-varying productivity than stocks with low contrast. The likelihood of time-variation in productivity of a given stock was also found to be closely related to the autocorrelation of the recruitment time series. Such inter-regional and inter-stock comparisons of variation are vital in understanding the roles of local and global environmental change on fish productivity.