262 
Fishery Bulletin 108(3) 
Figure 6 
Relative probability of detection of harbor porpoise ( Phocoena phocoena) by distance from 
the survey trackline (km) determined with a half-normal model with cosine adjustment for 
the 1998 and 1999 harbor porpoise sighting data. The solid line represents the probability 
of a sighting when both side observers and a belly observer were present. The dashed line 
represents the probability of a sighting when only side observers were present (without a 
belly observer). The histogram shows the distribution of sightings in 0.05-km groups. 
harbor porpoise occurrence by depth may account for 
the higher density estimated for the Bering Sea stock; 
however, the survey comprised only a portion of the 
entire stock. 
The SEA stock abundance estimate is not signifi- 
cantly different from the 1991-93 abundance estimate. 
The abundance estimates for the GOA stock (31,046) 
and the BS stock (48,215) are significantly higher 
than the 1991-93 abundance estimates (8497 and 
10,946, respectively) (t-test, natural log of means, 
PcO.Ol). It should be noted that the GOA stock abun- 
dance estimate may be biased low because it includes 
a survey of Cook Inlet which could not be corrected 
for perception bias, and the BS stock may have been 
underestimated as described in the previous para- 
graph. However, differences in survey design with the 
earlier surveys confound direct comparison between 
the abundance estimates. Overall, the area covered 
in the 1997-99 surveys was larger than that of the 
1991-93 surveys and included a wider range of pos- 
sible harbor porpoise habitat. The 1997-99 surveys 
were designed to include a sample of bays and inlets 
within the study region that the earlier surveys did 
not sample. The 1997-99 survey also included some 
larger bodies of water, such as Icy Bay and the inside 
waters of Southeast Alaska, that were not included in 
the earlier survey and gave more thorough coverage to 
some areas such as Yakutat Bay and Prince William 
Sound. The offshore extent of the 1997-99 survey 
was determined by water depth rather than distance, 
which extended it farther offshore in the Southeast 
Alaska and Gulf of Alaska stock regions. The 1999 
survey in the Bering Sea stock region covered much 
of the same area as the 1991 survey but at a higher 
density of effort. In 1999, the survey area to the south 
of the western end of the Alaska Peninsula (a survey 
area that was not completed in 1998) was surveyed. 
This area was not surveyed in the 1991-93 surveys. 
The survey design allowed for the inclusion of poten- 
tial harbor porpoise habitat that was not covered in 
the previous surveys, especially areas such as Yakutat 
Bay and Sitkalidak Strait (Kodiak Island). Another 
difference between the surveys was the use of correc- 
tion factors. A perception-bias correction was estimat- 
ed from independent observer data, and therefore only 
the Laake et ah, 1997 correction of 2.96 for availabil- 
ity bias was required to make a combined visibility 
correction factor of 4.62 for the SEA stock, 4.06 for the 
GOA stock, and 3.96 for the BS stock; these correction 
factors are 49%, 31%, and 28%, respectively, larger 
than the factor of 3.1, used in the 1991-93 surveys 
(Hill and DeMaster, 1998). The correction factors used 
