Time Series of Growth in the Genus Sebastes from the 



Northeast Pacific Ocean 



George W. Boehlert, Mary M. Yoklavich, and Dudley B. Chelton 



ABSTRACT: Marine fish populations respond to 

 their physical and biotic environment in complex 

 ways. While direct studies may discern short-term 

 responses at the individual level, time series are 

 needed to describe or predict the population level 

 response to environmental variation or cycles. The 

 ageing technique presented in this paper extracts 

 historical growth information from otoliths 

 through sectioning and careful measurement in 

 order to establish time series of growth. Otoliths of 

 two long-lived species, Sebastes pinniger and S. 

 diploproa, were collected off the west coast of North 

 America in 1977-84. Fish ages ranged from 1 to 86 

 years; corresponding birth dates were as early as 

 1896. Otolith measurements allow description of 

 growth at ages 1-6 for several decades of this cen- 

 tury. Although the technique has certain limita- 

 tions, significant interannual variability in growth 

 is obser\'ed, and its relationship to the species' en- 

 vironment is interpreted. Within species, coherence 

 in growth among age groups was not always evi- 

 dent; first year growth in S. diploproa was partic- 

 ularly different from growth in other years. For 

 both species, growth responses to environmental 

 factors were not clear; the dominant signal in the 

 time series appears to be increased growth rates 

 after about 1970. This signal is apparently related 

 to density-dependent factors (most likely prey 

 availability) associated with stocks depleted by 

 fishing pressure. 



Long-term changes in marine fish populations 

 can be caused by physical and biotic factors as 

 well as man-induced changes. An important goal 

 of fisheries research is to evaluate the effect of 

 fishing on population levels, and this task is 

 easily accomplished if natural variability is 

 understood. Determining causal relationships 

 and superimposing fishing mortality can lead to 

 predictive capability; indeed, many studies in 



George W. Boehlert, Southwest Fisheries Center Honokilu 

 Laboratory. National Marine Fisheries Service. NOAA, 

 Honolulu, HI 96822-2:396, and Joint Institute of Marine and 

 Atmospheric Research, University of Hawaii, Honolulu, HI 

 96822. 



Mary M. Yoklavich. Northwest and Alaska Fisheries Cen- 

 ter, National Marine Fisheries Service, NOAA, Seattle, WA 

 98115-(XI70. 



Dudlev B. Chelton, College of Oceanography. Oregon State 

 University, Corvallis, OR 973:31-550:3. 



Manuscript accepted .June 1989. 

 Fishery Bulletin, U.S. 87: 791-806. 



fisheries oceanography model past changes in 

 fish stocks with the objective of forecasting fu- 

 ture trends in populations for purposes of fish- 

 eries management. Long-term biological data 

 sets are also valuable for ecosystem research, 

 particularly in evaluating the range of natural 

 variability" (Wolfe et al. 1987). Historical catch 

 records have been used to assess long-term 

 changes in marine fish populations or stock sizes, 

 and such data from many decades are available 

 for some Pacific salmonid stocks (Mysak et al. 

 1982; Rogers 1984), for several North Atlantic 

 fisheries (Gushing 1982), and in the Southern 

 Hemisphere for Tasmanian fish populations 

 (Harris et al. 1988). Changes in species assem- 

 blages and biotic interactions in the California 

 Current region have been described on both the 

 decade scale (Loeb et al. 1983; Moser et al. 1987) 

 and the century scale (Soutar and Isaacs 1974). 



Long-term time series are developed either 

 from continuous data collection or by the e.xtrac- 

 tion of information stored naturally as a chrono- 

 graphic record. Continuous data collection must 

 occur over generations of biologists; starting a 

 new series may not allow achievement of objec- 

 tives for 30 or more years, so available time 

 series, which are often collected for other pur- 

 poses, are used. A classic e.xample of extracting 

 data from a chronogi-aphic record is the study of 

 fluctuations in population abundance of Eu- 

 gmnlis mordax, Sardiiiops saga.w and Mer- 

 luccius productus by Soutar and Isaacs (1974). 

 By enumerating fish scales preserved at differ- 

 ent depths in anoxic, varved sediments, they 

 defined natural cycles of abundance of these 

 species over 150 years. This approach has re- 

 cently been apphed by Shackleton (1987). Chron- 

 ological information is also stored in fish otoliths 

 (Radtke 1984; Campana and Neilson 1985). Esti- 

 mating age of fishes and using otoliths for back 

 calculation are simple examples of extraction of 

 historical information. In addition, the isotopic 

 composition of otoliths has been used to define 

 past thermal habitats occupied by individual fish 

 (Mulcahy et al. 1979; Radtke 1987) and to iden- 

 tify stocks (Mulligan et al. 1987). 



791 



