HOLT: DENSITY OF DOLPHIN SCHOOLS 



1982 



6 5^ 10 IS 2b 2S 3b 3S 40 <S SO 55 60 65 70 75 80 BS 30 



SIGHT InG ANGLE IM DEGREES 



study: 1) schools directly on the trackline are 

 never missed, 2) schools do not move in response 

 to the approaching ship or plane; and 3) no sys- 

 tematic measurement errors occur. All three as- 

 sumptions have been made in analyzing previous 

 aerial survey data (Holt and Powers 1982); how- 

 ever, field studies have subsequently been con- 

 ducted to investigate the ability of observers to 

 detect trackline schools (Holt 1983a), and 

 whether or not dolphins avoid approaching ships 

 (Au and Ferryman 1982; Hewitt 1985). In addi- 

 tion, assumption 3 was not accepted because an 

 inordinately large number of schools detected 

 from the ships was recorded on the trackline. 



i 



0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 S.S 6.0 6.5 7.0 7.5 8.0 



RHDIflL DISTflnCE in nflUTICFlL niLES 



Dl 



ama 



0.5 1.0 1.5 2.0 2.5 3.0 3.5 



?ERPEnDICULRR DISTflnCE in nflUTlCflL niLES 



Figure 2.— Continued— and 1982 ship data. 



where n is the number of schools sighted, D is the 

 density of dolphin schools per km^, L is the total 

 linear distance searched (km), and fiO) is a proba- 

 bility density function ipdf} evaluated at perpen- 

 dicular distance, x = 0. The Fourier series (FS) 

 model (Grain et al. 1979) was used to estimate /(O) 

 based upon criteria developed by Burnham et al. 

 (1979). Burnham et al. (1980) is recommended for 

 a full presentation of the FS model and for vari- 

 ance estimation. 



Several assumptions must be met for valid use 

 of LT theory. I investigated three of them for this 



Data Treatment 



All species of dolphins encountered in the study 

 area were included in the analyses. Of these, only 

 schools with a mean minimum or mean best esti- 

 mate of more than 14 animals were used because 

 my field experience indicated that the probability 

 that all animals in a school of at least this size 

 would be submerged at one time, and hence unde- 

 tectable, was very small. In addition, species af- 

 fected by the fishery generally occur in schools 

 with more than 14 animals. 



During the first 18 of 20 flights of the 1979 

 aerial survey, two independent teams of three ob- 

 servers each searched for dolphin schools. Mem- 

 bers of each team always searched for dolphins 

 during the same time, alternating with the other 

 team. 



For aerial and 1979-83 ship data, observers 

 recorded sea state conditions according to individ- 

 ual Beaufort, but during analyses, I grouped the 

 data into Da "calm" sea state category: seas 

 without whitecaps (Beaufort conditions 0-2) or 

 2) a "rough" sea category: seas with whitecaps 

 (Beaufort conditions 3-5). Data for Beaufort con- 

 ditions >5 were omitted from the analyses. The 

 presence of whitecaps was important because an- 

 imal splashes were used as sighting cues during 

 calm conditions but could not be easily distin- 

 guished from whitecaps during rough conditions. 



For aerial data and 1982-83 ship data, sun 

 glare effects were investigated by classifying ef- 

 fort at various sun positions into "good" and 

 "poor" categories depending on the amount of sun 

 glare on the trackline (see Holt* for method used 



*^Holt, R. 1984. Testing the validity of line transect theory 

 to estimate density of dolphin schools. Southwest Fish. Cent. 

 Adm. Rep. No. LJ-84-31, 56 p. National Marine Fisheries Ser- 

 vice, NOAA, P.O. Box 271, La Jolla, CA 92038. 



423 



