Hobbs and Waite: Abundance of Phocoena phocoena in three Alaska regions 
263 
in our analyses better reflect conditions encountered 
during aerial surveys for porpoise in Alaska because 
they incorporate a direct measure of animals missed 
by observers during the surveys, as well as the best 
available estimate of the animals missed while out of 
view underwater. 
It is likely that the shorter sighting time for the 
observer in the belly window increased the probability 
that inexperienced observers misidentified species of 
similar size during observations. Observers in the 
belly position of the aircraft during this survey had 
approximately 2-4 seconds to perceive, identify, and 
enumerate a group of animals. This is about half of 
the time available to the side observers and leaves 
little time for the observer to double check cues to 
distinguish among species. Thus, the observers are 
left with their first impressions which may be mis- 
taken if there is little prior experience in observ- 
ing and recording individual and groups of harbor 
porpoises. Laake et al. (1997) found a difference in 
perception bias between experienced (g(0) = 0.86) and 
inexperienced (g(0) = 0.23) observers in an experiment 
where the sighted species was known. We concur with 
Laake et al. (1997) that experienced observers should 
be positioned at the belly window and a training 
period should be considered for new aerial observers 
before their data from the belly position is used to 
estimate g(0). 
This analysis was completed in 2000 with the soft- 
ware that was available (DISTANCE, vers. 3.5), which 
did not include features to use multiple resights, so 
that perception bias had to be estimated separately. 
The current software DISTANCE 5.0 can use multiple 
resight data to estimate perception bias but does not 
correct for bias in the estimation of group size or for 
errors in species identification. Although some of the 
components of the analysis presented here are now 
completed automatically within the current software, 
the analysis of observer performance would have to be 
completed separately. 
Acknowledgments 
Funding for this project was provided by Recover Pro- 
tected Species Program, National Marine Fisheries 
Service, National Oceanic and Atmospheric Adminis- 
tration. We thank A. Andriolo, R. Angliss, K. Forney, 
K. Leinbach, J. Lerczak, L. Litzky, M. Merklein, S. 
Moore, S. Norman, D. Rugh, K. Shelden, V. Vanek, P. 
Wade, and K. Wynne for participating as observers. We 
thank K. Shelden for acting as field team leader on sev- 
eral occasions. We thank the pilots of the NOAA Twin 
Otter (M. Finn, J. Hagan, P. Hall, J. Longenecker, M. 
Moran, M. Pickett, and T. Strong) for their dedication 
and excellent handling of the aircraft. D. DeMaster, 
S. Moore, D. Rugh, J. Laake, and N. Friday provided 
valuable reviews of this manuscript. This research 
was conducted under permit no. 782-1438 issued by 
the National Marine Fisheries Service. 
