Fiedler and Reilly: Interannual variability of dolphin habitats and abundances 



457 



Common 



100 90 80 70 



Figure 3 (Continued) 



Common and whitebelly spinner dolphins did not 

 show significant linear relationships between N and 

 H (Fig. 7). For all four species, log-transformation 

 of the abundance estimates or lagging N ( by up to 

 four seasons did not change the significance levels 

 of the linear relationships. However, common dol- 

 phin abundance appeared to be maximum at H near 

 1.0 and to decrease at lower or higher H values, 

 except in 1983. Scatterplots of N t vs. H t for spotted 

 and whitebelly spinner dolphins suggest similar 

 nonlinear relationships. For common dolphins at low 

 values of//, as in early 1983, very little high-qual- 

 ity or favorable habitat was available in the ETP 

 (Fig. 6C). The only favorable habitat with H>1 was 

 in equatorial water west of the Galapagos. Half of 

 this favorable habitat was outside the SOPS bound- 

 ary. At high values of H, as in early 1985, favorable 

 habitat for the central and southern stocks (along 

 10°N and the equator, respectively) expanded. Fa- 

 vorable habitat along the equator extended beyond 

 the SOPS boundary. At the same time, favorable 

 habitat for the more abundant northern stock, off 

 Baja California, was reduced. 



Discussion 



Estimated abundances of spotted and eastern spin- 

 ner dolphins in the eastern tropical Pacific were 

 correlated with changes in the environment, as de- 

 scribed by the CCA-derived habitat quality index H. 

 The time scale of the changes and the patterns of 

 favorable habitat relative to nominal population 

 boundaries (Fig. 6) suggest that the correlations 

 resulted from a sampling effect, rather than a popu- 

 lation effect. Thus, we have explained biases in an- 

 nual population estimates that result in apparent 

 population changes. For example, Anganuzzi and 

 Buckland (1989) suggested that their low 1983 

 abundance estimate for spotted dolphins might be 

 explained by dispersal of local concentrations of 

 schools during the strong El Nino. Our results 

 showed that spotted dolphin schools may have 

 moved outside the nominal species range when the 

 "coastal tropical" habitat expanded into equatorial 

 water west of 100°W during this unusual event. 

 Similarly, estimated eastern spinner dolphin abun- 

 dance decreased slightly with increasing H because 



