FISHERY BULLETIN: VOL. 86, NO. 3 



alone (group size of 1) and larger groups (group size 

 of 3 and greater) (P = 0.56). 



Environmental Conditions Affecting 

 Sightings 



Sighting efficiency was not significantly affected 

 by rain, fog, or sun glare. Rain/fog conditions were 

 considered "poor" if rain or fog were present within 

 5 km of the vessel and "good" if neither were pres- 

 ent. The distributions of perpendicular sighting 

 distances were not significantly different between 

 these two strata (P = 0.32, 0.44, 0.78, and 0.64, 

 respectively, for surveys 1, 2, 3, and 4), and the 

 number of porpoise per kilometer surveyed was 

 higher in the "poor" category for two of the surveys. 

 Sun glare from the water's surface was considered 

 to contribute to "poor" sighting conditions if the sun 

 was within 45° of the trackline in front of the ship. 

 Conditions were considered "good" when the sun 

 was in other positions or was obscured by clouds. 

 As with rain/fog conditions, the distributions of 

 perpendicular sighting distances were not signifi- 

 cantly different between these "good" and "poor" 

 sun glare categories (P = 0.87, 0.47, 0.30, and 0.55, 

 respectively, for surveys 1, 2, 3, and 4). The number 

 of harbor porpoise per kilometer surveyed were 

 slightly higher in the poor category for three of the 

 surveys. In paired comparisons when glare was pres- 

 ent on only one side of the bow, approximately equal 

 numbers of sightings were made on the sides with 

 and without glare (60 vs. 59, respectively). All 

 categories of rain, fog, and glare are included in 

 subsequent analyses. 



Sea state did have a significant effect on porpoise 

 sightings. Sea state was categorized as calm 

 (without white-caps, Beaufort sea states 0, 1, and 

 2) or rough (with white-caps, Beaufort sea states 3, 

 4, and 5) following the classification used by Holt 

 and Cologne (1987). Distributions of perpendicular 

 distances were not significantly different between 

 these categories for any of the surveys (P > 0.05); 

 however for all surveys combined, the number of 

 harbor porpoise detected per kilometer was much 

 lower during rough seas (0.32 km"^) than during 

 calm seas (1.22 km"^). There were insufficient 

 sightings to estimate density for rough seas separ- 

 ately; therefore, rough sea data were excluded in 

 subsequent analyses. For all three surveys, the 

 numbers of harbor porpoise detected per kilometer 

 was higher at Beaufort & 1 than at Beaufort 2, 

 and for survey 3, the distributions of perpendicular 

 sighting distance were significantly different be- 

 tween these categories {P = 0.03). Porpoise density 



is, therefore, estimated separately for Beaufort 

 & 1 and for Beaufort 2 conditions. (For comparison, 

 harbor porpoise abundance was also estimated pool- 

 ing Beaufort sea states 0, 1, and 2. Estimated abun- 

 dance was approximately the same by both methods, 

 but the variance was slightly lower using the 

 stratified sea state categories. For this reason, only 

 the stratified estimates are presented here.) 



Helicopter Observations 



Helicopter observation of the behavior of harbor 

 porpoise in response to the survey ship were made 

 on only 6 groups of animals. Plots of vessel tracks 

 and movements of the groups are given in the cruise 

 report (see footnote 5). Only in one case was a 

 distinct behavioral change noted in response to the 

 ship. In that case, when the vessel was within 800 

 m, the group moved rapidly, perpendicular to the 

 path of the vessel and then parallel to and in the op- 

 posite direction of the vessel. Observers on the ship 

 saw this harbor porpoise group as they moved rapid- 

 ly out of the path of the vessel. Observers on the 

 ship also saw 2 of the other 5 groups. Although this 

 sample of behavior is small, movement in response 

 to the survey vessel appeared limited to within 1 km 

 of the vessel and, when it occurred, animals did not 

 travel far from their original positions. 



Porpoise Density 



The probability density distributions of perpen- 

 dicular sighting distances are shown in Figure 3 for 

 surveys 1 and 3 and for Beaufort sea states & 1 

 and 2. The hazard rate model gave acceptable fits 

 for all sighting distributions {P > 0.1) when the trun- 

 cation criteria was set at 400 m (Table 1). For survey 

 1, the optimum truncation points were chosen as 400 

 m for Beaufort & 1 and 800 m for Beaufort 2; for 

 survey 3, this distance was 400 m for both Beaufort 

 sea state categories. The fits of these models are 

 shown in Figure 3. Estimates of density and stand- 

 ard errors are given in Table 2. 



Depth Distribution Model 



The model of harbor porpoise depth distribution 

 was based on the relative densities of harbor por- 

 poise at different water depths. Ship survey data 

 were pooled into five depth ranges: 18-37 m (10-20 

 fathoms), 37-55 m (20-30 fathoms), 55-73 m (30-40 

 fathoms), 73-91 m (40-50 fathoms), and 91-110 m 

 (50-60 fathoms). Ship surveys are generally not 

 practical inshore of the 18 m isobath, but estimates 



424 



