814 
Fishery Bulletin 96(4), 1998 
Table 1 
Survey effort, line transect model parameters, density, and abundance estimates of harbor porpoise for each section ( A-E) and all 
sections combined surveyed within the northern San Juan Islands, Washington, from June to October 1992. 
Parameter Section A Section B Section C Section D Section E All sections 
Area (km 2 ) 
Effort (km) 
Transect lines 
Truncation width (m) 
Probability density /)0)/km 
Sightings of harbor porpoise 
Mean group size 
Standard error (SE) of group size 
Density (porpoise/km 2 ) 
95% confidence intervals (porpoise/km 2 ) 
SE of density 
% coefficient of variation (CV) of density 
Abundance 
95% confidence interval (abundance) 
SE of abundance 
50.26 
43.12 
50.40 
112 
120 
120 
14 
15 
15 
750 
750 
750 
2.96 
2.96 
2.96 
26 
62 
67 
2.19 
1.94 
1.81 
0.32 
0.15 
0.11 
0.60 
1.03 
1.55 
0.30-1.21 
0.60-1.75 
0.77-3.10 
0.21 
0.27 
0.52 
35.09 
26.00 
33.86 
30 
44 
77 
15-60 
26-75 
39-155 
10.53 
11.43 
26.07 
50.47 
42.43 
237 
112 
96 
560 
14 
12 
70 
750 
750 
750 
2.96 
2.96 
2.96 
80 
15 
250 
1.90 
1.80 
1.91 
0.11 
0.26 
0.07 
2.33 
0.76 
1.26 
.24-4.4 
0.35-1.66 
0.92-1.73 
0.74 
0.30 
0.20 
31.55 
39.16 
15.86 
116 
32 
299 
J— 221 
15-70 
217-406 
36.91 
12.53 
47.0 
low (37%) for categories that were not significantly 
different (Fig. 6). Given the small effect size, a power 
of 80% would require 358 locations of harbor por- 
poise in these three depth categories (there were 122 
in this study). It is, therefore, unlikely that harbor 
porpoise occur in depths within these nonsignificant 
categories in different proportions than those available. 
Mean seafloor slope for all sections combined was 
9.85% (SE=0.656, rc=584). Section B had the least 
slope and section C the greatest (Fig. 7). Harbor por- 
poise were sighted over a mean slope of 6.90% 
(SE=0.51, n=275, range: 0.37% to 45.75%). The great- 
est number of harbor porpoise sightings (79%) oc- 
curred over shallow slopes ( < 10%). There were sig- 
nificantly (P<0.05) greater numbers of harbor por- 
poise sightings than expected in category 0 to 2% 
slope, and significantly (P<0.05) fewer numbers of 
harbor porpoise than expected in categories 6 to 8%, 
18 to 20%, and >26% slope (Fig. 8). The power to 
detect a difference was fairly high (69%) for catego- 
ries that were not significantly different (Fig. 8). To 
increase power to 80%, we would need 151 samples 
within these ten categories (we had 126 samples). 
Mean sea surface temperature (SST) recorded dur- 
ing all transects was 12.6°C (SE=0.081, n=427, range: 
10.1° to 17.5°C). Little variability was found among 
the five sections. Section E had the greatest mean 
SST (mean=13.5°C, SE=0.22, n=69) and section B 
the least (mean=12.3°C, SE=0.14, n- 97). Mean SST 
recorded during harbor porpoise sightings was 12. 1 °C 
(SE=0.09, n- 267, range: 10.1° to 16.3°C). Harbor 
porpoise were sighted more frequently than expected 
(P< 0.05) in water temperatures of 11° to 12°C and 
less frequently than expected (P<0.05) in water tem- 
peratures >16° C (Fig. 9). The power to detect a dif- 
ference was moderate (50%) for categories that were 
not significantly different (Fig. 9). To increase power 
to 80%, we would need 180 samples within these ten 
categories (we had 171 samples). 
There was no significant difference between num- 
ber of harbor porpoise observed per minute during 
flood and ebb tides ( U= 315.5, n=69, P-0.076, a=0.05). 
Bootstrap estimates (resampling statistics; 10,000 
iterations) indicated an 86% chance of correctly re- 
jecting the null hypothesis that mean number of 
sightings was equal between flood and ebb tides. Fifty 
samples in each tide stage (we sampled 52 in flood 
and 17 in ebb tide) were required to reject the null 
hypothesis at a = 0.05. 
Mean Secchi reading for all harbor porpoise sightings 
was 9.3 m (SE=0.09, n- 275, range: 5.7 to 11.9 m). 
Greatest numbers of harbor porpoise were ob- 
served in mid-morning ( 1000 h) and afternoon ( 1400 
to 1500 h) throughout the study area (Fig. 10). Fewer 
harbor porpoise were observed at midday (1100 to 
1300 h; Fig. 10), although there was no significant 
difference (77=10.99, n=274, P=0.276) among mean 
number of harbor porpoise observed per minute and 
each hour of daylight surveyed (0900 to 1800 h). 
Density estimates were calculated over all four 
months of the survey period rather than by month, 
which would have yielded too low of a sample size. If 
abundance estimates of porpoise had varied greatly 
among months during our survey period, we should 
have observed a higher CV (ours was relatively low: 
0.159). 
