FISHERY BULLETIN: VOL. 86, NO. 3 



For 1985, both inside and outside swaths were in- 

 cluded. As in previous stratifications, apparent den- 

 sity clearly decreased with increasing sea state 

 (Table 5). This effect was significant for surveys in 

 both 1984 (P < 0.05) and 1985 (P < 0.005). Transect 

 segments surveyed during sea states of 2 or greater 

 were excluded in subsequent analyses. 



•Uncorrected harbor porpoise den- 

 ^) for the stratification based on sea 



Table 5.- 

 sities (km 



state. Data include only high-density areas sur- 

 veyed when cloud cover was <25%. Data for 

 1984 include inside swaths only; data for 1985 

 include inside and outside swaths. Numbers in 

 parentheses refer to area (km^) surveyed 

 under the given condition. 



Between Survey Differences 



We considered the 1984 survey and the 1985 

 surveys at 0.61 km and 1.85 km from shore as three 

 independent estimates of harbor porpoise density. 

 Because apparent density was shown to vary greatly 

 with sighting conditions and because sighting con- 

 ditions varied between surveys, it was necessary to 

 compare these three under similar conditions. 



The highest (and presumably least biased) den- 

 sities were obtained when sea state was Beaufort 

 & 1 and when cloud cover was <25%. Between 

 survey comparisons imder these conditions are given 

 in Table 6 for the eight geographic regions given 

 in Figure 2. For 1985, there were no transect 

 segments at 0.61 km from shore under the condi- 

 tions Beaufort & 1 and clear skies. For 1984, only 

 three regions contained more than 10 km- of 

 searching effort at 0.61 km from shore. For 1985, 

 only four areas had any searching effort at 1.85 km 

 from shore. The only direct density comparisons 

 with reasonable sample sizes are for region 1 (0.000 

 vs. 0.048 porpoise/km^) and region 3 (0.111 vs. 

 0.110 porpoise/km^) (respectively for 1984 and 

 1985). The densities for all regions pooled (0.671 and 

 0.510 porpoise/km^) are similar, but because of the 

 small sample size and geographic variation in sam- 

 pling, a statistical test of this difference is mean- 

 ingless. 



In comparing surveys, sample size and regional 

 coverage improved slightly when Beaufort 2 was 

 considered (still allowing a maximum of 25% cloud 

 cover) (Table 7). For 1984, coverage was relatively 

 complete in all regions. For 1985, coverage at 0.61 

 km from shore was limited to regions 2-4, and 

 coverage at 1.85 km was limited to regions 1-5. 

 Comparing the two surveys in 1985, distance from 

 shore made little difference in overall density for 

 all regions combined, and neither survey had con- 

 sistently higher values than the other. Comparing 

 the 1984 survey to the two 1985 surveys, the former 

 had a higher overall density for all regions combined, 

 but again this difference was not consistent among 

 regions. Sample size and regional coverage were 

 again too poor for meaningful statistical tests. 



Table 6.— Uncorrected harbor porpoise densities (km"^) in 

 eight geographic regions surveyed during Beaufort & 1 con- 

 ditions. 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. 



Ventilation Patterns 



Harbor porpoise did not appear to react to the 

 helicopter during aerial observations; they were 

 visible throughout a surfacing series and were not 

 visible during dives. Knowing this, we were able to 

 use data on surfacing series and dive times to deter- 

 mine the fraction of time harbor porpoise would be 

 visible from the air. Data on ventilation patterns 



440 



