Fishery Bulletin 93(1), 1995 



Coefficients of variation and confidence 

 intervals 



Coefficients of variation (CV) and confidence inter- 

 vals (CI) of the abundance estimates are based on 

 the bootstrap method (Efron, 1977; Buckland et al., 

 1993b). The sightings associated with consecutive 

 segments of search effort were combined to form a 

 set of subsamples of 139 km (75 nmi) of search effort 

 (corresponding to approximately one day of survey 

 effort). 5 I drew subsamples randomly with replace- 

 ment from this set of effort segments, and a pseudo- 

 population size was estimated by using the same 

 group size stratification as was used for the actual 

 abundance estimates. For each bootstrap sample, the 

 probability of detecting trackline groups, g(0), was 

 estimated as a random number between and 1 

 drawn from the probability distribution of a bino- 

 mial ratio with a mean and coefficient of variation 

 equal to the estimated values. This process was re- 

 peated 1,000 times, and the CV of the estimated 

 population size was calculated as the standard error 

 of the 1,000 pseudo-population sizes divided by the 

 estimated population size. Bootstrap 95% confidence 

 intervals on the population estimates were based on 

 the 25th and 976th ranked estimates from the boot- 

 strap samples. Log-normal 95% confidence intervals 

 were based on the method given by Buckland et al. 

 (1993b) and used the bootstrap estimate of CV. 



Results 



During the survey approximately 10,100 km of 

 searching effort were completed (Fig. 1), and 515 

 cetacean groups were seen during the sampling ef- 

 fort. Tracklines included 2,386 km in calm sea states 

 (Beaufort 2 or less) and 7,696 km in rough sea states 

 (Beaufort 3-5). During the survey, 18 cetacean spe- 

 cies were identified (as well as at least one species 

 that could only be identified to genus) (Table 1 ). More 

 detailed data summaries for this survey are pre- 

 sented by Hill and Barlow (1992), including the po- 

 sitions and school sizes of all on- and off-effort 

 sightings of cetaceans and pinnipeds, maps showing 

 the distribution of sightings for each species, distri- 

 butions of perpendicular distances for each species, 

 patterns of association in mixed-species groups, sum- 

 maries of searching effort completed under various 

 conditions, and sighting rates of individual observ- 

 ers. The fit of the probability density functions to 



5 Barlow, J. 1993. The abundance of cetaceans in California wa- 

 ters estimated from ship surveys in summer/fall 1991. Admin. 

 Rep. LJ-93-09, available from Southwest Fish. Sci. Cent., P.O. 

 Box 271, La Jolla, CA 92038, 39 p. 



the distributions of perpendicular distances are il- 

 lustrated by Barlow. 5 



Group-size estimation 



Group-size correction parameters, the slopes and 

 intercepts (in parentheses) of log 10 -transformed re- 

 gressions, were 0.922 (0.03), 1.022 (-0.03), 0.886 

 (0.07), and 0.777 (0.11) for the four observers who 

 required correction. Three of these observers appear 

 to have underestimated group size, in some cases by 

 a large amount (a group of 500 would have been, on 

 average, estimated as 328, 534, 283, and 152 by these 

 four observers, respectively). 



Probability of detecting trackline groups 



Independent observers searched a total of 8,190 km. 

 Approximately 7% of groups were detected only by 

 the independent observer; however, all groups that 

 were detected only by the independent observer were 

 groups of less than 20 individuals and accounted for 

 only 0.7% of the individuals that were seen on the 

 survey. Of all groups that had less than 20 animals 

 and were seen while the independent observer was 

 on duty, 347 were seen by the primary observers, and 

 40 were seen by the independent observer. 



Abundance estimation 



With estimated values of/(0) and ^(0) (Table 2), den- 

 sity and abundance were calculated for 19 cetacean 

 species and 9 higher taxonomic categories (Table 3). 

 Common dolphins were the most abundant cetaceans 

 by a large margin. Of the two recently recognized 

 common dolphin species (Heyning and Pen-in, 1994), 

 the short-beaked variety was much more abundant 

 than the long-beaked variety. Blue whales were the 

 most abundant species of large whale. 



Discussion 



Distribution 



The distributions of cetaceans seen during this sur- 

 vey (Figs. 2—6) are in general agreement with the 

 results of other studies in this area (Leatherwood et 

 al., 1982; Dohl et al., 1986; Smith et al., 1986; Barlow, 

 1988; Forney et al., this issue; Dohl et al. 2,3 ). How- 

 ever, the observed distribution of some species con- 

 tradicted results of previous studies. Striped dolphins 

 were seen rather commonly in mixed groups with 

 short-beaked common dolphins in southern and cen- 

 tral California between 185 and 555 km (100-300 



