The Relationship between Red Cod, Pseudophycis Bachus, Recruitment and Environmental Variables in New Zealand

Abstract

This paper examines the relationship between red cod, Pseudophycis bachus, recruitment and selected environmental variables. Commercial catch data of red cod were used as indices of relative recruitment strength for New Zealand fish stocks (east coast South Island, RCO 3; west coast South Island, RCO 7). Analysis of catch per unit effort (CPUE) supports the assumption that fluctuations in commercial catches reflect changes in abundance resulting from variable recruitment. Catch was assessed against 34 environmental variables that included the Southern Oscillation index, sea surface temperature, weather types, wind speed and direction, and river flows. Lagged correlations were calculated between annual commercial catch and averaged environmental data. Variables most strongly correlated with commercial catch were included in regression analyses. Regression models were calculated using all available data in its raw form, and after application of a logarithmic transformation. Robustness, and forecast skill was tested using a repeated leave-one-out procedure. The strongest correlations between commercial catch and explanatory variables for both RCO 3 and RCO 7 were consistently with sea surface temperature over a wide range of time lags followed by frequencies of southwest and northeast flow patterns. The correlations generally operated in the sense that ‘colder is better’: negative correlations (∼−0.75) with sea surface temperature, positive with the occurrence of cool southwest weather types, negative with the occurrence of warm northeast weather types. The maximum magnitude of the correlation occurs at 14-month lag for both fisheries, corresponding to the late spring/summer of the year prior to the catch year. There were often secondary maxima in the magnitude of the correlations two or three years prior to the catch year. The commercial fishery is made up largely of two and three-year-old fish, implying that juvenile fish (age ∼ one year) are most sensitive to environmental conditions, but egg, embryo and larva periods may also be affected. Regression analyses suggest that two-predictor regressions were unstable, with often large decreases in explained variance when applied to the independent years, however, at the lags when relationships are strongest (around 14 months), the explained variance on the left-out years remains high. Fluctuations in red cod landings are dependent on sea surface temperature, controlled largely by climatic variability, including El Niño and La Niña events. For short-lived species such as red cod, fluctuations in annual landings are commonplace and analyses of environmental data may be useful in predicting catches in subsequent years.