Forney et al.: Abundance of cetaceans in California waters: aerial surveys 



19 



set. The data were subdivided by area into effort seg- 

 ments of equal length, and the segments were then 

 drawn randomly with replacement until the total 

 number of kilometers actually surveyed in each area 

 was reached. This process was replicated 1,000 times. 

 Forney and Barlow (1993) demonstrated that the 

 choice of segment lengths between 5 km and 20 km 

 did not influence the resulting estimates of precision. 

 In this analysis we also performed bootstrap simula- 

 tions for 50 km and 100 km segments and again found 

 that segment length did not affect estimates of vari- 

 ance. For the bootstrap analysis, we chose a segment 

 length of 50 km, which roughly reflects the degree of 

 sampling variability for these surveys (i.e. the dimen- 

 sion of actual gaps in the sampling grid in Figure 2). 



Each of the 1,000 bootstrap replicates was treated 

 and analyzed as a separate survey: sightings were 

 first stratified into the three species/group-size cat- 

 egories given above. Individual values for n and s 

 were calculated, and/10) was estimated with the pro- 

 gram HAZARD. The estimated value of g(0) was 

 treated as a correction factor known without error. 

 The variance, coefficient of variation, and 95% confi- 



dence intervals were obtained from the distribution of 

 the 1,000 bootstrap abundance estimates with stan- 

 dard formulae. Because the bootstrap method (Buck- 

 land, 1984) of obtaining confidence intervals can re- 

 sult in the lower 95% confidence intervals being smaller 

 than the actual number of animals seen (or even zero) 

 we also calculated log-normal confidence intervals 

 based on the bootstrap coefficient of variation. 



Results 



Detailed results of the survey, including sighting in- 

 formation and plots of sighting locations for all spe- 

 cies sighted are presented elsewhere (Carretta and 

 Forney, 1993). Results relevant to the analyses pre- 

 sented in this paper are given below. A total of 253 

 cetacean sightings were made (Fig. 3): 213 on effort 

 (while actively searching), and an additional 40 off 

 effort (24 while in transit, 8 beyond 12° declination 

 angle, 7 while circling over another group of animals, 

 and 1 by an off-effort observer). Twenty eight on-ef- 

 fort sightings could not be positively identified to the 



42° 



38°- 



T3 



3 36° -I 



35°- 

 34° 

 33° 

 32° 



31° 

 30° 



Pacific 

 Ocean 



, 1 , 1 , 1 , 1 1 1 1 1 1 1 1 1 1 1 1 



127° 126° 125° 124° 123° 122° 121° 120° 119° 118° 117° 



Longitude 



42" 



38° 



37° 



0) 

 "O 

 2 36° 



33° 



32° 



31°- 



30 



Pacific 

 Ocean 



. 1 . 1 1 1 1 1 1 F ' 1  1 ' 1 ' I ' 



127° 126° 125° 124° 123° 122° 121° 120° 119° 118° 117° 



Longitude 



Figure 3 



Locations of all 253 cetacean sightings made during the 1991 and 1992 surveys. The 213 on-effort sightings (used in the abun- 

 dance estimation) are shown by diamonds, and the 40 off-effort sightings (e.g. made while circling or in transit) are shown with 

 plus signs. 



