BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE 



were available from the helicopter study in Califor- 

 nia (13 groups were observed, mean group size was 

 2.7), the Oregon shore observations (11 groups, 

 mean group size was 4.2), and a previous shore study 

 in Glacier Bay, AK (28 solitary individuals). The 

 mean times spent in surfacing series were 34.2, 24.6, 

 and 30.0 seconds (respectively for the three studies). 

 The corresponding mean dive times did not differ 

 significantly between study sites (P > 0.05 using 

 pairwise ^tests). Using the pooled data set (n = 52), 

 the mean time spent in a surfacing series was 30.02 

 seconds (SE = 1.95) and the mean time spent in a 

 dive was 95.81 seconds (SE = 5.32). The percent- 

 age of time spent at the surface is 23.9%. 



Average and Adjusted 

 Density Estimates 



Harbor porpoise densities under optimal condi- 

 tions (Beaufort & 1 and <25% cloud cover) were 

 averaged for the two surveys, weighting by transect 

 length (Table 8). Given an average survey speed of 

 160 km/h and assuming that the window for harbor 

 porpoise observation is 400 m long, the time win- 

 dow during which a point would be visible is 9.0 



Table 7— Uncorrected harbor porpoise densities (km"^) in 

 eight geographic regions surveyed during Beaufort 2 condi- 

 tions. Data for 1984 are based on inside swaths of transects 

 flown 0.61 km from the coast. Data for 1985 are based on in- 

 side and outside swaths of transects flown at 0.61 and 1.85 

 km from the coast. Only those segments surveyed when cloud 

 cover was <25% are included. Numbers in parentheses refer 

 to area (km^) surveyed under the given condition. 



Geographic 

 region 



1984 survey 

 density 

 0.61 km 



1985 survey 

 density 

 0.61 km 



1985 survey 

 density 

 1.85 km 



1 

 2 

 3 

 4 

 5 

 6 

 7 

 8 

 All regions 



0.000 

 (107) 



0.092 

 (109) 



0.196 

 (117) 



0.562 

 (180) 



0.489 

 (119) 



0.439 

 (57) 



0.114 

 (70) 



0.268 

 (37) 



0.295 

 (796) 



(0) 



0.102 

 (59) 



0.062 

 (48) 



0.282 

 (110) 



0.459 

 (6) 



(0) 



(0) 



(0) 



0.192 

 (224) 



0.184 

 (103) 



0.103 

 (68) 



0.287 

 (52) 



0.070 

 (57) 



0.234 

 (171) 



(0) 

 (0) 



(0) 



0.188 

 (452) 



seconds. Using this estimate and the surface and 

 dive times estimated above, the probability that a 

 porpoise will be seen is estimated as 0.310 from 

 Equation (2). An instantaneous count would there- 

 fore underestimate porpoise abundance by a factor 

 of 3.2. Average values were therefore multiplied by 

 this factor (Table 8). 



DISCUSSION 



Results indicate that sighting conditions must be 

 very good in order to estimate harbor porpoise abun- 

 dance from aerial strip transects. Both sea state and 

 cloud cover had very large and significant effects 

 on apparent density. Limiting observations to the 

 best categories of sea state (Beaufort & 1) and 

 cloud cover (<25%) can be used to minimize the bias 

 due to missed animals. These conditions are, how- 

 ever, rare and only occurred during 5.3% and 10.3% 

 of the transects in 1984 and 1985. The actual occur- 

 rence of these conditions is even more rare if one 

 considers flights that were cancelled due to bad 

 weather. 



The effects on sea state and cloud cover on sight- 

 ing conditions were predicted by observers before 

 analysis of survey data was begun. Most harbor por- 

 poise were first seen when submerged a small 

 distance below the surface. Surfacings were rela- 

 tively inconspicuous to aerial observers and were not 

 an important cue in sighting porpoise. Both sea state 

 and cloud cover affect the ability of observers to see 

 through the water's surface and to spot submerged 

 animals. Increasing sea state causes more refrac- 

 tion of light at the water's surface, increases glit- 



Table 8.— Harbor porpoise densities (km"^) in eight geographic 

 regions from a) a weighted average of uncorrected estimates from 



1984 and 1985 aerial surveys, b) the same average adjusted by 

 a factor of 3.2 to account for submerged porpoise that were missed 

 by aerial observers, and c) a weighted average for the 1 984 and 



1985 ship surveys. Aerial estimates are based on observations 

 made under conditions of Beaufort & 1 and with <25% cloud 

 cover. Ship estimates include Beaufort sea states of 0, 1, and 2. 



Geographic 

 region 



Area 



surveyed 



(km*) 



Aerial estimates 



Uncorrected 

 a 



Research 

 vessel 

 Corrected estimates 

 b c 



441 



