816 
Fishery Bulletin 96(4), 1998 
gence have a high probability of remaining undetec- 
ted during the passage of an aircraft or vessel, re- 
sulting in availability bias (Marsh and Sinclair, 
1989). Several studies (e.g. Marsh and Sinclair, 1989) 
of harbor porpoise have indicated, on the basis of 
perception bias, that the probability of detecting a 
harbor porpoise on the trackline, gtO), is less than 
one (Barlow, 1988; Palka, 1993; Calambokidis 3 ). 
Using an independent team of three observers, 
Barlow (1988) reported an estimated 22% of harbor 
porpoise that surfaced on the trackline were missed 
by a team of five observers (perception bias) travel- 
ing on a vessel at 18.5 km/h. Using three observers 
per survey, Calambokidis 3 and Palka (1993) esti- 
mated the probability of observing a group of harbor 
porpoise on the trackline, g(0), was less than 0.5. We 
assumed g(0) was one during our study because we 
were unable to determine availability or perception 
bias. It is probable that some 
porpoise did avoid the vessel 
and might have been sub- 
merged for up to five minutes 
(Raum-Suryan, 1995). It is, 
therefore, likely thatg(O) is less 
than one and harbor porpoise 
abundance is underestimated. 
The ability to estimate 
group size can vary by the 
number of animals within the 
group and by the species of 
interest. Data from land- 
based calibration studies off 
the Washington coast indi- 
cated that observers on ves- 
3 Calambokidis, J. 1991. Vessel 
surveys for harbor porpoise off the 
Washington coast. In H. Kajimura 
(ed.). Harbor porpoise interactions 
with Makah Salmon set net fishery 
in coastal Washington waters, 1988- 
89. National Marine Mammal 
Laboratory Processed Report, Na- 
tional Marine Mammal Laboratory, 
Natl. Mar. Fish. Serv., NOAA, 7600 
Sand Point Way NE, Seattle, WA 
98115. 
70 
25 50 75 100 125 150 175 200 225 250 
Depth (m) 
i | Observed number of porpoise [~| Expected number of porpoise 
Figure 6 
Depth distribution of 275 harbor porpoise sightings determined from random boat 
surveys (June to October 1992) in relation to expected distribution of harbor porpoise 
if they were distributed randomly with depth (as determined from depths at 584 ran- 
dom locations). An asterisk (*) designates a significant (P< 0.05) difference determined 
with chi-square goodness-of-fit analyses. 
Table 2 
Mean, standard error (SE), number of sightings, and number of harbor porpoise determined during fixed boat surveys in 1991 
and 1992. In 1992, the number of harbor porpoise observed (Obs.) are presented, as are values from four randomly chosen surveys 
used in analysis (Anal.), comparing mean number of harbor porpoise sighted along each transect in 1991 and 1992). n refers to 
the number of transect “sides” (bow out to 90° on port or starboard) surveyed. 
1992 
Transect 
no. 
1991 
Mean 
SE 
n 
No. of 
sightings 
No. of 
porpoise 
Probability 
Mean 
SE 
n 
No. of 
sightings 
No. of 
porpoise 
Obs. 
Anal. 
Obs. 
Anal. 
Obs. 
Anal. 
Obs. 
Anal. 
Obs. 
Anal. 
i 
2.20 
0.80 
5 
11 
18 
2.50 
2.00 
0.42 
0.41 
8 
4 
20 
8 
31 
13 
P = 0.843 
2 
1.83 
0.48 
6 
11 
16 
1.25 
1.25 
0.45 
0.63 
8 
4 
10 
5 
22 
17 
P = 0.474 
3 
1.00 
0.38 
7 
7 
13 
0.13 
0.25 
0.13 
0.25 
8 
4 
1 
1 
1 
1 
P = 0.200 
5 
2.83 
0.54 
6 
17 
29 
2.25 
3.00 
0.49 
0.58 
8 
4 
18 
12 
25 
18 
P = 0.844 
6 
0.60 
0.40 
5 
3 
6 
0.88 
0.50 
0.40 
0.29 
8 
4 
7 
2 
12 
3 
P= 0.853 
7 
1.75 
0.63 
4 
7 
10 
1.50 
1.00 
0.82 
0.41 
8 
4 
12 
9 
26 
19 
P = 0.356 
