FISHERY BULLETIN: VOL. 84, NO. 2 



however, we felt that using the estimated proportion 

 of the diet comprised of the different prey types was 

 a more realistic representation of what was actually 

 occurring in the ecosystem. The body weight and 

 prey preference estimates were taken from Kenney 

 et al. (1985), who had based their estimates on an 

 extensive literature review. For three species not in- 

 cluded in that reference— beluga, false killer whale, 

 and rough-toothed dolphin— body weight and prey 

 preference estimates were based on Watson (1981) 

 and Nishiwaki (1972). For the categories of sightings 

 which were not completely identified, the body 

 weight and prey percentages were calculated as 

 averages for all species included in the category and 

 weighted by the number of sightings of each. (It 

 might be argued that the unidentified categories 

 should be excluded totally and that their inclusion 

 introduces too much uncertainty. However, we felt 

 that excluding them would eliminate many poten- 

 tially valuable observations and that including them 

 would provide a closer measure of habitat use Some 

 of the categories can be narrowed to only a couple 

 of species, and the number of sightings overall is a 

 valid basis for estimating the probability of an un- 

 identified sighting being a particular species.) 



The biomass data were then summed for all 

 species in each block and season, as well as for the 

 piscivorous, teuthivorous, and planktivorous subsets. 

 Values for endangered species biomass were also 

 calculated by summing the data for right, humpback, 

 blue, fin, sei, and sperm whales, as well as for the 

 estimated proportion of the unidentified categories 

 made up of these species. The biomass data for each 

 block and season were then divided by the corre- 

 sponding effort data, resulting in values of biomass 

 per unit effort (BPUE) in units of kilograms of ceta- 

 cean sighted per kilometer of track line surveyed 

 (kg/km). The final data set therefore had, for each 

 block, BPUE values for all cetaceans, for en- 

 dangered species only, and for the piscivorous, 

 teuthivorous, and planktivorous components of the 

 cetacean fauna for each season and for the entire 

 year. 



The simplest technique for looking at the pattern 

 of high-intensity habitat use by cetaceans is to plot 

 the blocks with the highest values of BPUE. Ob- 

 viously, the blocks with the highest BPUE values 

 within any of the individual data sets are those with 

 the highest intensity of habitat use The question 

 becomes one of defining the cutoff point in each 

 distribution for selecting the highest values. The fre- 

 quency distributions of each of the BPUE data sets 

 were examined for any patterns which might be 

 useful as an objective criterion to define a lower 



bound for the high-use blocks (eg, bimodal distribu- 

 tions, or 2 standard deviations above the mean of 

 a normal distribution). Log-survivorship plots (plot- 

 ting BPUE vs. log of the number of blocks with 

 higher BPUE values; see Fagen and Young 1978) 

 were also tried to look for changes in slope which 

 could serve as a means of numerically defining this 

 boundary. When these techniques failed to select any 

 specific value for the cutoff point, we opted to use 

 simple percentile rankings to classify the blocks for 

 plotting the results. 



The final step in the analysis was to develop an 

 index which would serve to define those areas which 

 are most important as cetacean habitat. By "impor- 

 tant" we include both the level of habitat use and 

 the management priority of the individual species. 

 Habitat requirements for an individual probably de- 

 pend heavily upon prey type, so each of the data sets 

 for the three feeding classes were included in this 

 process. Since management objectives concentrate 

 on the endangered species, the endangered species 

 data sets were also included. Since the endangered 

 species data are also part of the feeding type data, 

 the former are in effect being included twice This 

 gives the endangered species extra weight in the in- 

 dex, in accord with both their endangered status and 

 management focus. For each seasonal set of BPUE 

 data for the endangered species and the three feed- 

 ing classes, blocks were assigned points as follows: 

 5 if the BPUE was greater than the 99th percentile 

 value for that data set, 3 if it was between the 95th 

 and 99th percentiles, 1 if between the 90th and 95th 

 percentiles, and otherwise The value of the index 

 for a block is then the sum of these point values for 

 all data sets. Since there were four seasons and four 

 BPUE variables used, the maximum possible value 

 for the index in any block would be 80 (4 x 4 x 5). 

 For lack of a more concise term, we shall refer to 

 this as Habitat Use Index, although it does have the 

 additional dimension of focus on endangered species. 

 Since this index is based on only the top 10% of each 

 of the 16 individual data sets, it provides a simple 

 way to point out those blocks which repeatedly stand 

 out as high-use habitat in more than one season 

 and/or for more than one prey type 



RESULTS 



During the CETAP study, observers on dedicated 

 aerial or POP surveys operating within the defined 

 survey criteria made 5,304 sightings of 26 different 

 species of whales and dolphins. These include sight- 

 ing of individuals in three genera— Globicephala, 

 Mesoplodon, and Kogia— which could only be iden- 



348 



