Abstract. — The results of a 

 canonical correspondence analysis 

 (CCA) of data from research ves- 

 sel surveys of the eastern tropical 

 Pacific were applied to time series 

 of estimated dolphin abundances 

 from tuna vessel sightings. The 

 research vessel survey data con- 

 sisted of daily dolphin school 

 sightings and concurrent environ- 

 mental variables for August-No- 

 vember of 1986 through 1990. Sea- 

 sonal fields of habitat quality for 

 1975-90 were calculated from his- 

 torical bathythermograph data by 

 using the CCA ordination results. 

 For spotted (Stenella attenuata) 

 and eastern spinner (S. longi- 

 rostris orientalis ) dolphins, annual 

 abundance estimates or inter- 

 annual changes in those estimates 

 are significantly correlated with 

 habitat quality. This effect is at 

 least partly due to expansion of 

 high quality habitat beyond the 

 geographic ranges assumed for the 

 abundance estimate. We discuss 

 ways that environmental data 

 could be used to reduce error in 

 dolphin abundance estimates. 



Interannual variability of dolphin 

 habitats in the eastern tropical 

 Pacific. II: Effects on abundances 

 estimated from tuna vessel 

 sightings, 1975-1990 



Paul C. Fiedler 

 Stephen B. Reilly 



Southwest Fisheries Science Center, National Marine Fisheries Service. NOAA 

 P O Box 27 I , La Jolla. CA 92038 



Manuscript accepted 18 October 1993 

 Fishery Bulletin 92:451^163 (1994) 



The eastern tropical Pacific Ocean 

 (ETP) supports a diverse and abun- 

 dant cetacean fauna. By the late 

 1960s, it had become clear that 

 large numbers of dolphins were 

 being killed in tuna purse seine 

 operations (Perrin, 1969). The dol- 

 phin species affected by the tuna 

 fishery, known as "target species," 

 are spotted dolphin (Stenella atten- 

 uata), the "whitebelly" form and 

 "eastern" subspecies of spinner dol- 

 phin (S. longirostris and S. /. orien- 

 talis, Perrin, 1990), common dol- 

 phin (Delphinus delphis), and 

 striped dolphin (S. coeruleoalba). In 

 1973, the U.S. government initi- 

 ated a formal program to place ob- 

 servers on purse seiners to monitor 

 dolphin mortality (Smith, 1983). In 

 1975, the Inter-American Tropical 

 Tuna Commission began putting 

 observers on the international 

 fleet. The tuna vessel observer data 

 from these programs includes 

 sightings of cetaceans, as well as 

 mortality data and biological 

 samples from incidental catches. 



Time series of target species 

 abundance have been estimated 

 from tuna vessel observer data 

 (Buckland et al., 1992). These are 

 yearly estimates of stocks of spot- 

 ted, spinner, and common dolphins 

 within nominal stock boundaries 

 known as the Status of Porpoise 

 Stocks or SOPS boundaries. 1 



Buckland et al. (1992) analyzed 

 smoothed time series of these esti- 

 mates and detected significant 

 trends for some stocks. 



Time series of dolphin abundance 

 estimates are subject to consider- 

 able sampling error plus the effects 

 of environmental variability on 

 abundance and distribution. The 

 effect of environmental factors on 

 abundance estimates must be 

 quantified before such a time series 

 can be properly interpreted in 

 terms of abundance changes or 

 trends. This paper analyzes rela- 

 tionships between abundance esti- 

 mates from tuna vessel observer 

 data and environmental variability. 

 Since little or no environmental 

 data are collected on these ships, 

 we base our analysis on species- 

 environment relationships derived 

 from research vessel data (Reilly 

 and Fiedler, 1994). 



National Marine Fisheries Ser- 

 vice research vessels have surveyed 

 the ETP several times since 1974 

 to collect data for abundance esti- 

 mates and supplement the tuna 

 vessel observer data (Holt et al., 

 1987). The most extensive research 

 vessel observer data were collected 



1 Au, D. W. K., W. L. Perryman, and W. F. 

 Perrin. 1979. Dolphin distribution and the 

 relationship to environmental features in 

 the eastern tropical Pacific. NOAA/SWFC 

 Admin. Rep. No. LJ-79-43, 59 p. 



451 



