KENNEY and WINN: CETACEAN HIGH-USE HABITATS 



study, the entire data base was archived on magnetic 

 tape at the University of Rhode Island Academic 

 Computer Center. The data base is very large, com- 

 prising nearly 70,000 entries and 112 variables; it 

 includes almost 25,000 sightings of cetaceans, sea 

 turtles, or other large marine animals (eg., sharks, 

 ocean sunfish, swordfish, rays, etc.). 



For this paper, the study area was partitioned in- 

 to blocks measuring 10 minutes of latitude by 10 

 minutes of longitude The area of the blocks ranges 

 from about 243 km 2 at the northern extreme of the 

 study area to about 281 km 2 at the southern end, 

 due to the curvature of the earth's surface and 

 resulting convergence of the meridians toward the 

 north pole The data were further grouped by calen- 

 dar seasons across all the years of sampling. All 

 dedicated aerial and POP data which met defined 

 criteria were included in the analysis. These criteria 

 included observer(s) formally on watch, clear 

 visibility of at least 2 miles, and sea states of Beau- 

 fort 3 or lower. Although the dedicated aerial and 

 POP data were not directly compatible for the pur- 

 pose of absolute abundance estimation, we are 

 justified in combining them for this analysis. An ex- 

 amination of sighting effort in the 1979 CETAP data 

 (Hain et al. 1981) demonstrated a significant corre- 

 lation between numbers of sightings and length of 

 line surveyed for both aerial and POP surveys. Re- 

 analysis of these same data shows that the average 

 number of sightings per mile of track line surveyed 

 was somewhat higher for the POP surveys, but that 

 the difference is not statistically significant at the 

 5% level (paired Student's £-test). Since we are in ef- 

 fect using the number of sightings per unit length 

 of track line as a measure of relative abundance in 

 this analysis, the two data types can be combined. 



To remove any bias due to uneven allocation of 

 sighting effort among the blocks, the effort was first 

 quantified. A computer program was developed 

 which calculated the length of track line surveyed 

 each season within each of the 10-minute blocks, in- 

 cluding only line segments surveyed within the 

 criteria defined above Each line surveyed is recorded 

 in the data base as a sequence of latitude-longitude 

 positions. For any pair of successive positions, the 

 length of track line between the points (D, in km) 

 can be calculated by: 



D = 111.12 arccos [sin (X : ) sin (X 2 ) 



+ cos (X : ) cos (X 2 ) cos (Y 2 - Y x )], 



where Xj and X 2 are the latitudes of the two posi- 

 tions, and Yj and Y 2 are the corresponding longi- 

 tudes. This calculates great circle distance Flight 



or cruise tracks would actually be rhumb lines rather 

 than great circles, but the algorithm required to 

 calculate rhumb line distance is much more complex. 

 Furthermore, for two points around 10 km apart, 

 typical of track line segments in the data, great cir- 

 cle and rhumb line distance differ by <L m, an error 

 of <0.01%. 



For a pair of points within a single 10-minute 

 block, the length of the intervening line segment is 

 simply assigned to that block. The difficulty arises 

 for successive points located in separate blocks. It 

 is then necessary to find the point(s) of intersection 

 where the track line crosses any block boundary(ies). 

 The bulk of the computer program is concerned with 

 this procedure For a pair of points in separate 

 blocks, the equation describing the great circle 

 through the points is defined. The point where that 

 line crosses a boundary is then determined by insert- 

 ing the latitude or longitude value defining the 

 boundary into the great circle equation, and then 

 solving for the other coordinate The line segment 

 which originally spanned two or more blocks is there- 

 by partitioned into smaller segments, each wholly 

 contained in a single block, whose lengths are then 

 calculated as above The final step in the procedure 

 is to sum the lengths of all the line segments within 

 the block, which represents the amount of sighting 

 effort expended in the block. 



All cetacean sightings made during track 

 segments meeting the defined criteria were also ex- 

 tracted from the data base These data were sum- 

 marized to produce, for each species, the total 

 number of individual animals sighted in each block 

 and season. (This is not to say that this number 

 represents all different individuals. An individual 

 may be sighted repeatedly by different surveys, but 

 this is taken into account by the correction for ef- 

 fort.) In order to combine different species, the num- 

 ber of animals of a particular species was multiplied 

 by the species' estimated average body weight to 

 calculate biomass sighted per block and season. The 

 biomass data for each species were then partitioned 

 into three feeding classes— piscivorous, teuthivorous, 

 and planktivorous— based upon the estimated per- 

 centages of each species' diet composed offish, squid, 

 and zooplankton, respectively. In an earlier analysis 

 of prey consumption by cetaceans in the CETAP 

 study area, Scott et al. 4 classified each species into 

 a single category based on its principal prey type; 



4 Scott, G. P., R. D. Kenney, T. J. Thompson, and H. E. Winn. 

 1983. Functional roles and ecological impacts of the cetacean com- 

 munity in the waters of the northeastern U.S. continental shelf. 

 Paper presented at 1983 annual meeting, International Council for 

 the Exploration of the Sea, ICES CM. 1983/N:12. 



347 



