Bengtson et al Comparison of survey methods for estimating abundance of Phoca vitulino 



349 



study in Johns Hopkins Inlet, Glacier Bay, and 

 determined that it was feasible to count harbor 

 seals on glacial ice from medium-format aerial 

 photographs, and comparisons of these counts 

 were made to simultaneous shore counts. In our 

 study, we employed higher resolution (large- 

 format) film, conducted three simultaneous 

 aerial and shore surveys (vs. one in the pilot 

 study), and used data from different altitudes 

 to assess sources of counting error within aer- 

 ial photographs. 



Aerial surveys of harbor seals are most often 

 conducted when peak numbers are hauled out, 

 which usually occurs during the annual molt in 

 late summer. During such surveys, low altitude 

 (100-300 m) photographs of harbor seal groups 

 are obtained, from which seal counts are made 

 (Olesiuk et al., 1990; Boveng et al,, 2003). Sur- 

 veying seals in glacial fjords is more difficult 

 than on terrestrial sites because the moving, 

 large expanses of scattered ice on which seals 

 haul out offer little spatial reference to aid in 

 counting seals. Furthermore, there is often in- 

 sufficient maneuvering room for low-altitude 

 aerial surveys in the fjords. The main objec- 

 tive of our study was to compare the relative 

 effectiveness of shore-based and aerial survey 

 methods to estimate harbor seal abundance. 



Materials and methods 

 Study area 



Figure 1 



Map of Glacier Bay, Alaska, showing Johns Hopkins Inlet (shaded). 



Johns Hopkins Inlet is located in the north- 

 west arm of Glacier Bay (58=N, 138°30W) in 

 southeastern Alaska (Fig. 1). At the head of the inlet 

 is an active tidewater glacier that is currently advanc- 

 ing. The fjord walls are steep, rising to 640 m and 884 

 m within one km of the western and eastern shores, 

 respectively. Harbor seals rest, nurse, and molt on ice 

 calved from the glacier. Pieces of ice occupied by seals 

 are typically around 8-24 m^, although they vary widely 

 in size and can be much larger (up to -890 m-. /!=105) 

 (E. A. Mathews, unpubl. data). Ice that has above-water 

 features taller than about 2 m is rarely occupied. The 

 percent of ice cover in the inlet was estimated by shore 

 observers before seal counts during August for the years 

 1995-2001; for most surveys, 6-25% or 26-50% of the 

 inlet was covered by ice (?! = 83 shore surveys). Approxi- 

 mately 60-70% of harbor seals in Glacier Bay use glacial 

 ice in Johns Hopkins Inlet during the pupping, breeding, 

 and molting periods from spring to early fall (Mathews, 

 1995). 



Shore-based surveys 



Observers counted seals from an elevated (-35 m) site 

 located along the western shore of the inlet about 2.5 km 

 from the terminus of Johns Hopkins Glacier (Fig. 2A). 



From this site, the observers' field of view comprised 

 approximately 9 km (from the glacier to Jaw Point). The 

 northwestern edge of the inlet and a small area near the 

 northwest edge of the glacier face were not visible to the 

 shore observers because of obstruction by headlands and 

 other geographic features (i.e., "blind spots," Fig. 2A). 

 Ice in the inlet more commonly drifted closer to the east 

 than the west shore because of current and wind pat- 

 terns; therefore, the location of the observation site along 

 the western shore was selected as the observation site. 

 Two of the four observers involved in this study simul- 

 taneously counted seals 2 or 3 times each day between 

 12 and 23 August 2001 and 9 and 26 August 2002. Seals 

 on ice and in the water were tallied, but only seals on 

 ice were included in this analysis. Observers conducted 

 shore-based counts of harbor seals in Johns Hopkins 

 Inlet during, or within an hour of, the aerial surveys 

 conducted on 15 and 16 August 2001, and 15 August 

 2002. Surveys targeted the period around solar noon 

 when the largest number of seals hauls out on glacial 

 ice (Hoover, 1983; Calambokidis et al., 1987). 



Observers counted seals using 20x60 binoculars 

 mounted on tripods. After each tripod was leveled, ob- 

 servers locked the vertical orientation of the tripod 

 head and counted all seals in the field of view as the 



