DEVELOPMENT OF SURVEY METHODOLOGY FOR LONG TERM MONITORING OF HARBOR SEALS IN WRANGELL­ST. ELIAS NATIONAL PARK AND PRESERVE, ALASKA, 1994

Abstract

Coastal Wildlife Survey - Development of survey methodology for long term monitoring of harbor seals in Wrangell St. Elias National Park and Preserve, Alaska 1995

Full text

DEVELOPMENT OF SURVEY METHODOLOGY FOR LONG TERM
MONITORING OF HARBOR SEALS IN WRANGELL-ST. ELIAS
NATIONAL PARK AND PRESERVE, ALASKA
Karin Kozie, Bill Route, Resource Management Specialists,
Wrangell-St. Elias National Park and Preserve, PO Box 29,
Glennallen, AK. 99588
DATE: June 30, 1995
INTRODUCTION
Harbor seals (Phoca vitulina richardsi) inhabit coastal waters in
Alaska from southeast Alaska to the Aleutian Islands and the
Bering Sea (Hoover 1988). Although they were previously
considered abundant, surveys during the 1980's indicated declines
as great as 86% have occurred from southcentral to western Alaska
(Pitcher 1986, 1989, 1990). While the rate of decline appears to
have slowed, the numbers of seals in these areas are
substantially lower than they were in the early 1970's (Hoover-
Miller 1992). Census data for southcentral and southeast Alaska
indicate statistically significant declines have occurred in the
Prince William Sound and Ketchikan trend areas. Numbers have
declined but not significantly in the Sitka trend area (Loughlin
1992, 1994).
Limited data is available on the status of harbor seals in
Wrangell-St Elias National Park and Preserve (WRST), specifically
Icy and Disenchantment Bays. A breeding population of up to
several thousand seals was reported in Icy Bay and 700 to 900
seals were observed in Disenchantment Bay during 1992 (Kozie
1992). However, these numbers are based on incidental or casual
observations. Seal populations were surveyed by the National
Marine Fisheries Service (NMFS) during 1991-1993 in response to
the alarming declines in parts of Alaska and as directed by the
1988 amendment to the Marine Mammal Protection Act. NMFS
documented 907 seals in Icy Bay and 747 seals in Disenchantment
Bay in mid-September 1993 (Loughlin 1994). These data may under
estimate the population. A ten-fold reduction in seal
populations occurred in Glacier Bay between mid-August and mid-
September due to movements of seals outside the bay (Mathews
1993). Similar movements may have occurred in Icy and
Disenchantment Bays. Additionally, while minimum population
estimates are extremely important for baseline information, these

surveys were not designed to detect statistically significant
changes in population numbers between years.
Several factors that may affect seal populations in the Yakutat
Bay area make trend data highly desirable (Yakutaga Area
Management Plan, Alaska Dept. of Fish & Game). Local residents
have described a decline in a related marine mammal, the Steller
sea lion (Eumetopias jubatus), in Yakutat Bay. A year round sea
lion rookery/haul out area along the Malaspina forelands (WRST)
supported about 200-300 animals in the early 1980s. K.
Hundertmark, Alaska Department of Fish & Game, observed 150-250
sea lions at Sitkagi Bluffs (along the Malaspina Forelands) on
December 4, 1982 (Alaska Dept Fish & Game Memorandum, 1982).
This number gradually decreased to zero by 1991 (R. Mossman-NPS,
G. Rainy-Fishing and Flying, pers commun). Harbor seals may be
experiencing similar declines but no trend data are available for
the Yakutat area. Trend data in nearby Prince William Sound
indicate harbor seals have experienced significant declines there
(Loughlin 1992).
Proposed development of the Icy Bay area could affect unstudied
pinniped populations. Offshore oil drilling may occur west of
Icy Bay and south of Yakutat Bay in 1995 (Minerals Management
1990). Marine mammals are at risk from potential oil spills and
pollution if oil is developed in adjacent offshore areas.
Logging is occurring along the east and west sides of Icy Bay.
Expected increases in logging-related boat traffic may disturb
seals. Increases in tourism in Icy and Disenchantment Bays by
cruise ships trying to observe calving glaciers and kayakers may
also disturb seals hauled out on ice flows since they are
especially sensitive to vessel activity (Allen et al. 1984,
Calambokidis et al. 1983). Commercial fishing occurs throughout
Yakutat Bay and may affect seal populations. Because harbor
seals are distributed within coastal waters, they are commonly
caught incidental to commercial and subsistence net fishing
operations (Hoover 1988). The degree of take is unknown but
could be significant in some salmon and herring gill net and
purse seine fisheries. An unknown number are also shot by
commercial fisherman when they disrupt their nets.
The purpose of this study was to develop long term monitoring
methods for harbor seal pup and population surveys in Icy Bay.
The results of 1994 surveys and recommendations for 1995 are
presented in this paper.

OBJECTIVES
1) Determine harbor seal adult/pup ratios in Icy Bay
2) Determine a minimum population estimate for harbor seals in
Icy Bay
3) Develop survey methodology for monitoring long term trends
in pup proportions and population levels of harbor seals in
Icy Bay
STUDY AREA
Icy Bay (Fig 1) is located about 85 km west of Yakutat. Seal
haul out habitat within the bay consists of glacial ice which
varies in abundance and distribution throughout the summer
breeding season. No land haul outs are present.

METHODS
Pup Proportions
Adult\pup ratios were determined using methods developed by
Mathews in 1993 (in press). Briefly, on June 15, 1994, R Mossman
took a series of non-overlapping slide photographs (200 ASA) of
seals hauled out on ice flows throughout Icy Bay (Fig 2). Seals
were photographed through an open window of a piper supercub
flying at 700-800 ft above sea level (ASL), using a Nikon 8008S
camera with a 80-200mm zoom lens, motor drive and data back. A
new roll of film was used for each aggregation of greater than
100 animals (Fig 2).
Pup proportions were determined by viewing the slides through a
dissecting microscope and classifying seals as adults, pups, or
unknown. Unknown age seals were excluded from calculations. A
mean pup proportion and standard deviation were then calculated
by averaging the percent pups from each roll of film.
POPULATION ESTIMATE
In 1993, we hired B. Mathews, Univ of Alaska - Southeast, to
develop methods for surveying harbor seal populations in Icy Bay.
B. Mathews has been working on developing standardized survey
methodology for harbor seals in Glacier Bay which consists of
counting seals hauled out on land and ice flows from elevated
observation areas along the bay and aerial photography (Mathews
1992). Methods using land based observations could not be
applied to Icy Bay because of the large size of the bay (e.g.
there are no vantage points where an observer could view the
entire bay or haul out area at a distance close enough to count
seals). Further, aerial census of harbor seals on ice flows are
complicated by the lack of land marks to distinguish which seals
have been surveyed and which have not. Consequently, no
methodology exists in the literature for surveying harbor seals
hauled out on ice flows over a large area.
We contracted with J. Kern and L. McDonald in the fall of 1993 to
analyze the data we collected during September 1993 and recommend
changes in our methodology in an effort to develop a
statistically valid sampling procedure for 1994.
Methods recommended by Kern and McDonald

Methods recommended by Kern and McDonald (in prep) consisted of
systematic aerial line transects spaced ½ mile apart throughout
the main part of the bay and a zig-zag pattern in the fiords (Fig
3). The orientation of the survey lines in the main part of the
bay was based on efficient coverage and safe flight paths. A
starting point was selected for the first transect by randomly
choosing a number between 1 and 5 from a random numbers table and
then measuring that distance (e.g. 2=0.2 miles) from a randomly
chosen area along the shoreline. For a 50% sampling intensity,

one line in each of two adjacent lines was randomly selected (1
or 2, 3 or 4, etc.) by tossing a coin. Further, the side of the
plane the observer counted seals from was randomly chosen by the
flip of a coin. The zig zag pattern in the fiords was
constructed by drawing a line across the bottom of the fiord and
randomly choosing a number between 1 and 10 (e.g. 1=0.1 mile) to
determine a random starting point. The route that the plane flew
in the fiord was chosen by using an angle that would allow for
safe turning of the plane and adequate coverage of the fiord.
For safety, it was recommended that a single 1/8 mile strip be
flown along the perimeter of the bay to count seals hauled out on
rocks, beaches, or ice along the shore. To correct for
visibility bias, Kern and McDonald (in prep) recommended
recording the number of seals and their distance from the
airplane (1/8, 1/4, and 1/2 mile) as determined from markers
along the wing struts.
We attempted to implement the methods they recommended in August
1994 but it became apparent when we reached the survey area that
the methods recommended by Kern and McDonald (in prep) were not
adequate due to the following factors: 1) seal distribution was
highly clumped rather than evenly dispersed throughout the bay
and the recommended 50% coverage did not adequately sample the
main aggregation; 2) we were unable to collect distance data
because of the high numbers of seals encountered in a short
period of time; and 3) high winds and limited fuel supply
decreased survey time and access into some fiords where seals
were aggregated. During the 3 survey days, we evaluated 4
different survey procedures in an effort to develop a safe and
efficient survey method that would provide adequate coverage of
seals in Icy Bay.
Methods used in population surveys, August 14-16
The best minimum population estimates are obtained during the
annual molt, which extends from late July to September in Glacier
Bay (Mathews 1992). Molt periods in Icy Bay were assumed to be
on a similar schedule and population surveys in Icy Bay were
flown on August 14, 15, and 16, 1994. B. Route was the observer
and M. Sharp the pilot in a piper supercub. Flights took place
between 1000-1400 hours during the maximum haul out period
(Calambokidis et al 1987).
1) Transects

On August 14, we attempted to follow the recommended survey
methodology as closely as possible. Ice distribution was
determined from a high altitude (3000 ft) pre-survey flight. The
bay was stratified into high, medium, and low ice density. Pre-
determined random transects were flown at 700 ft ASL in a piper
supercub based on recommendations by Kern and McDonald (in prep).
The observer recorded transect number or letter, number of seals
per observation, and distance from plane in two categories (0-1/8
mile, 1/8-1/4 mile), into a hand held tape recorder. He was not
able to clearly observe seals >1/4 mile from the plane, so seals
beyond this distance were not recorded. One of every two
adjacent transects in the main part of the bay was randomly
selected for survey, resulting in a 50% effort. All fiords, with
the exception of the north Guyot and the Taan, were surveyed in a
zig zag pattern as recommended. High winds and low fuel
prevented surveys in north Guyot and Taan fiords, respectively.
Consequently, only a small portion of one randomly selected
survey line passed through the bulk of the seals which were
aggregated mainly in north Guyot fiord (Fig. 4). No perimeter
survey was flown since no seals were observed using this area and
fuel was limited.
On August 15, seal rather than ice distribution was determined
from a high altitude (3000 ft) pre-survey flight since ice
density did not necessarily correlate with seal numbers. The bay
was stratified into high, medium, and low seal density. We
repeated the transects in the main part of the bay using the same
methodology as the previous day. All fiords were circled for
complete coverage at 1200 ft ASL (for safety reasons). We found
that circling provided complete coverage of the fiords and was
safer and easier to implement than the original zig zag pattern.
On August 16, we concentrated our efforts on counting the main
aggregation since it was clear from the previous two days that
95% of the seals in Icy Bay were there (Fig. 4). We flew a
random sample of 6 out of 12 parallel lines (50% coverage) spaced
1/4 mile apart at 1200 ft ASL through the aggregation. The strut
marker was recalibrated to 1/8 mile at the beginning of the
flight. The higher altitude allowed us to count seals at a
slower rate than on previous days which we felt allowed more
accuracy. The side of the plane seals were counted from was
randomly chosen for each line by flipping a coin.
2) Minimum Count

We circled the aggregation intensively at 700 ft ASL on August 14
to obtain a minimum count. A population estimate can be
calculated for the entire bay using the minimum count of the
aggregation and adding the estimate from the line transect survey
for areas of the bay with low seal density (e.g. everything
except the large aggregation).
3) Double Count
We also attempted a double sampling survey where we did a total
minimum count of the aggregation from 1200 ft. We then went back
and recounted a smaller area (about 1/4 of the total area of the
aggregation) at a higher intensity to get a correction factor
(Gasaway et al 1986).
4) Aerial Photography
We took 72 overlapping slide photographs on August 15 at 3000 ft
ASL to obtain a composite of the entire aggregation. We then
dropped to 1000 ft ASL and took 20 additional non-overlapping
photographs to get a sample of the aggregation within this
composite. By calculating the average number of seals on the low
altitude slides, and determining the approximate area of the low
altitude slides in relation to the high altitude slides, we hoped
to determine a statistically valid population estimate.
RESULTS AND DISCUSSION
Objective 1: Determine harbor seal adult/pup ratios in Icy Bay.
The proportion of harbor seal pups in Icy Bay during 1994
appeared higher than 1993 (Table 1) although sample sizes were
smaller and this may have affected pup proportions. We will
attempt to collect a larger data set in 1995.
Table 1. The proportions of harbor seal pups in Icy Bay, 1993-
1994.
_________________________________________________________________
ROLLS # OF # OF TOTAL SEALS MEAN
AREA YEAR OF FILM ADULTS PUPS PHOTOGRAPHED % PUPS SD
Icy Bay 1993 8 916 269 1185 25 7.1
Icy Bay 1994 6 231 81 231 34 4.9
_________________________________________________________________
Objective 2: Determine a minimum population estimate for harbor

seals in Icy Bay.
We estimated seal numbers using five different methods (Table 2).
Table 2. Population estimates of harbor seals in Icy Bay, August
14-16, 1994 using all methods.
_________________________________________________________________
POPULATION
DATE ESTIMATE 80% CI 90% CI METHOD
_________________________________________________________________
8/14/94 1851 -- -- Strip Transect
of Total Bay
8/16/94 3804 +/-850 +/-1092 Strip Transect of
Aggregation Only
8/14/94 1054 -- -- Min Count of Aggregat
+Strip Transct of low
8/16/94 1331 -- -- Double Count
+Strip Transct of low
8/15/94 21098a +/-5899 +/-7581 Aerial photography
_________________________________________________________________
a Suspect sampling bias or an unknown error in area calculation
resulted in an over estimate.
1) Transects
Total Bay
A minimum population estimate for August 14 was 1864 seals based
on strip transect methodology and stratification on seal density
(versus ice density)(Table 3). There is no confidence limit on
this estimate since only a small part of one survey line went
through the high density seal aggregation. High winds prevented
further sampling in the north Guyot fiord where the seals were
aggregated.
Table 3. Population estimate of harbor seals in Icy Bay on
August 14, 1994 based on line transect surveys.
_________________________________________________________________
SEAL SEALS/
DATE AREA LINE DENSITY MI2 #SEALS MI2
8/14/94 ICY BAY A LOW 0.75 1 1.33

8/14/94 ICY BAY G LOW 0.53 0 0.00
8/14/94 ICY BAY H LOW 0.47 0 0.00
8/14/94 ICY BAY I LOW 0.34 0 0.00
8/14/94 ICY BAY J LOW 0.34 0 0.00
8/14/94 ICY BAY 5 LOW 3.01 3 1.00
8/14/94 ICY BAY 6 LOW 2.42 8 3.31
8/14/94 ICY BAY 9 LOW 0.86 0 0.00
8/14/94 ICY BAY 3A LOW 0.08 0 0.00
8/14/94 ICY BAY 3C LOW 1.55 0 0.00
8/14/94 ICY BAY 3B HIGH 0.23 149 647.83
Term Low High Total
Total area 77.45 2.81 80.26
Sampled area 7.92 0.23 8.15
% sample 0.15 0.08 0.14
Total survey lines 12.00 8.00 20.00
Lines sampled 10.00 1.00 11.00
Mean 0.56 647.83
Variance 0.96 ----a
Low: 0.56 seals/mi2 x 77.45 mi2 = 43.37
High: 647.00 seals/mi2 x 2.81 mi2 = 1820.40
1863.77 = Best Estimate
_________________________________________________________________
a n = 1 for high strata provided no variance.
The tape did not work correctly on August 15, and part of the
line transect data was lost. No estimate could be calculated for
August 15 using partial data from the line transect methodology.
Aggregation Only
The total estimate of seals in the large aggregation on August 16
was 3804 (90% CI+/- 1092, 80% CI+/- 850)(Table 4). The wide
confidence intervals on our data suggest a greater sampling
effort is needed (e.g. 100% versus 50%) in areas where seals are
aggregated.
Table 4. Population estimate of harbor seals in Icy Bay on
August 16, 1994 based on transect surveys of the high
density seal aggregation.
_________________________________________________________________
# SEALS/ EST #
DATE AREA LINE MI2 SEALS MI2 SEALS
8/16/94 ICY BAY
8/16/94 ICY BAY 1 0.09 27 300 3540

8/16/94 ICY BAY 2 0.34 204 600 7080
8/16/94 ICY BAY 3 0.38 112 295 3478
8/16/94 ICY BAY 4 0.61 204 334 3946
8/16/94 ICY BAY 5 0.61 206 338 3985
8/16/94 ICY BAY 6 0.5 59 118 1392
AVG = 331.0 3904.0
STD = 141.5 1669.2
TOTAL AREA = 11.5
TOTAL ESTIMATE = 3804
90% CI+/- 1092
80% CI+/- 850
_________________________________________________________________
2) Minimum Count of the Aggregation + Strip Transect for Low
Density Seal Areas
A minimum count of 1011 seals was obtained on August 14 by
intensively circling the large aggregation. It was difficult to
differentiate boundaries within the 2.81 mi2 ice pack which may
have resulted in some double counting or missed seals. However,
we feel we missed more seals than double counted so this should
be considered a minimum count. By adding the estimate obtained
from the low density transects (described above), a total
estimate for Icy Bay would be 1054 seals (1011 in the aggregation
+ 43 in the remainder of the bay).
3) Double Count of the Aggregation + Strip Transect for Low
Density Seal Areas
A minimum count of 1288 seals was obtained on August 16 by using
the double count technique. As with the minimum count, it was
difficult to differentiate boundaries within the aggregation. By
adding the estimate obtained from the low density transects
(described above), a total estimate for Icy Bay would be 1331
seals (1288 in the aggregation + 43 in the remainder of the bay).
4) Aerial Photography
Estimates from the aerial photography taken on August 15 were
difficult to calculate because of the unknown area of a low
altitude slide. We originally thought we could calculate the
area of the low altitude slides by identifying them on a
composite from the high altitude slides but this was not

possible. We then attempted to calculate the area of the low
altitude slides by estimating the average length of a harbor seal
(Byrden 1972) and extrapolating the number of harbor seals across
the length and width of each slide to calculate the area. This
resulted in a gross over estimate of 21,228 seals (90% CI+/-
7581, 80% CI+/- 5899)(Table 5).
Table 5. Population estimate of harbor seals in Icy Bay on
August 15, 1994 based on aerial photography.
_______________________________________________________________
MM M FT
Seal length (actual) 1550.000 1.550 5.084
Seal length (slide) 5.875 0.006 0.019
Projected slide length on wall 700.000 0.700 2.296
Projected slide width on wall 466.000 0.466 1.528
Scale (1 mm on slide = ) 263.830 0.264 0.865
Dimensions of photo: Length 184680.800 184.681 605.753
Dimensions of photo: Width 122944.600 122.945 403.259
M2 KM2 MI2
Area of photo 22705.520 0.022705 0.008766
AVG STD N
Seals per slide 25.150 28.370 27.0
KM2 MI2
Estimated density 1107.580 2868.630
Total Area 19.170 7.400
Total population = 21,228 Seals
90% CI +/- = 7,581 Seals
80% CI +/- = 5,899 Seals
_______________________________________________________________
Objective 3: Develop survey methodology for monitoring long
term trends in pup proportions and population
levels of harbor seals in Icy Bay
Pup Proportions
Surveys should be flown mid-day (1000-1400 hours) when the
greatest number of seals are hauled out (Calambokidis et al
1987). Seal distribution will be determined from a high altitude
(3000 ft) pre-survey flight. The distribution and a rough

estimate of the percent of the population in each aggregation (in
increments of 10) should be recorded on a base map Appendix 1.
The aggregations may not necessarily total 100% if some seals are
spread out in the bay. For example, if there are 3 seal
aggregations, group 1 may comprise about 20%, group 2 about 40%,
and group 3 about 30%, for a total of 90% in the 3 groups. The
remaining 10% are singles or small groups less than 100 scattered
throughout the bay.
Each aggregation that contains more than 100 seals should be
sampled for pup proportions. Up to ten rolls (24 exposures per
roll) of non-overlapping slide photographs (100 or 200 ASA
Kodachrome film depending on light conditions) of seals hauled
out on ice flows will be taken in proportion to the distribution
of seals throughout the bay. Each increment of ten represents
one roll of film so if 40% of the population is in one
aggregation, the observer would take 4 rolls of film. Ideally,
each roll will be flown along a corresponding transect. If the
configuration of the aggregation limits the number of transects
to less than the required rolls of film, the observer may need to
decrease the increment of time between shots (e.g. every 10
seconds to every 3 seconds) so 2 rolls can be taken along one
transect. Since film needs to be changed mid-transect in this
situation, the pilot should note the latitude and longitude on
the GPS and a compass bering so the plane can return to that
location and continue along the same transect. The observer may
feel it is difficult to be sure he/she can maintain non-
overlapping photographs in some aggregations but as long as the
observer and pilot make a reasonable effort to keep the transects
as separate as possible, the probability of over-lapping slides,
especially in large aggregations, is minimal, since the area of
each slide is very small (e.g. there are usually only 1-3 seals
per slide).
The location of transects, the roll number, the beginning time
for photographing each aggregation, and a rough estimate of the
number of seals in each aggregation using the following
categories (e.g. 100-500, 500-1000, 1000-2000, 2000-4000, 4000+)
should also be recorded on the base map (Appendix 1). Each roll
of film and mailing envelope (which is returned with the slides)
should be labeled with the date and the roll number.
Seals are photographed through an open window of the plane flying
at 1000 ft above sea level (ASL), using a Nikon 8008S camera with
an 80-200mm zoom lens, motor drive and data back. All

photographs should be shot as close to vertical as possible so
the entire slide is in focus and at the fastest speed (1/1000-
1/250 second). If the camera is tilted, a portion of the slide
becomes blurred. E Mathews has written directions for camera use
during the survey. A copy is stored with the camera in Yakutat
and an additional copy is in the harbor seal files in the
Resource Management Office in Glennallen.
In groups with sparse seal distribution, a photograph should be
taken each time a seal is observed through the view finder. The
observer should wait until the seal moves across the field of
view before taking the photograph. For example, if the plane is
traveling north, the observer will first view a seal at the north
end of the view finder but should wait to take a photograph until
the seal has reached the south end of the view finder. This
helps decrease the potential bias of photographing adults with
pups, which are easier to see than lone adults (E Mathews, pers
commun).
In areas of dense seal distribution, the observer should space
photographs every 3-10 seconds depending on the size of the
aggregation. The number of seconds between frames should be
maintained on each transect but may be changed between transects
if it appears the interval is too short or long for the size of
the aggregation. The observer should take the first photograph
at the edge of the aggregation and then begin counting "one-one
thousand, two-one thousand, etc...". When the desired interval
is reached, the observer should locate the first seal along the
north edge of the frame, wait to photograph it until it reaches
the south edge of the frame, and then begin counting again (E
Mathews, pers commun).
Pup proportions will be determined by viewing the slides through
a dissecting microscope for ice haul outs (and by projecting
slides onto a white background for land haul outs) and
classifying seals as adults, pups, or unknown (Mathews 1992).
Unknown age seals are excluded from calculations. A mean pup
proportion is then calculated by averaging the percent pups from
each aggregation and reporting the standard deviation between
aggregations.
Due to large increases in cruise ships in Disenchantment Bay, pup
proportion and population surveys will also be conducted there in
1995. Methodology will be the same as for Icy Bay with the
exception of surveying land haul outs. Land haul outs should be

surveyed by taking slide photographs (200-400 ASA film) of the
entire haul out. In many cases, 1 or 2 photographs can cover the
entire haul out. If several photographs are required, slides
should be overlapping. If more than 1 aerial pass must be made
to photograph a haul out, the pilot should make a wide turn
rather than a tight circle, to limit disturbance. Seals hauled
out on land can be difficult to distinguish in slides. The
observer should adjust the camera so there is a slight
overexposure of 1/2-1 F stop (as much as possible without losing
film speed)(E Mathews, pers commun). Survey dates for
Disenchantment Bay should be scheduled for days when low tide
occurs between 1000-1400 hours so the greatest number of seals
will also be present on land haul outs.
Population Surveys
Neither the minimum count or double count of the aggregations
provided methodology that was precise enough to detect trends in
seal populations. We recommend these methods are discontinued in
1995.
General recommendations for all population survey methods
Seal distribution is an important factor in determining survey
methodology. Distribution was highly clumped during the August
14-16, 1994 surveys but may vary within the season or between
years. Sampling effort should be based on seal distribution
since ice density did not correlate well with seal numbers in
1994. Seal distribution should be determined from a high
altitude (3000 ft) pre-survey flight. Seals should be stratified
into high, medium, and low density areas on a map each day. If
seals are tightly aggregated similar to 1994, high and low
categories will be sufficient. To obtain a more accurate map in
areas of the bay where landmarks are limited, the pilot should
track the outline of the aggregation using GPS locations. These
locations can later be plotted and the area measured using a
digital planimeter.
Surveys should be flown mid-day (1000-1400 hours) when the
greatest number of seals are hauled out on ice flows
(Calambokidis et al 1987). Survey dates for Disenchantment Bay
should be scheduled for days when low tide occurs between 1000-
1400 hours so the greatest number of seals will be present on
land haul outs as well as ice flows.

Specific recommendations for specific survey methods
1) Transects
If seals are aggregated in one to several groups of >100 animals,
survey lines, spaced ¼ mile apart, should be flown at 1200 ft ASL
through the aggregation(s). Coverage should be 100%. A
transparency with lines marked 1/4 mile apart should be randomly
placed over the aggregation. Direction of survey lines should be
based on safety considerations and the most efficient coverage of
the aggregation. Only the portion of the survey lines going
through the aggregation are flown. The observer should
consistently count from the same side of the plane for each
aggregation. The remainder of the Bay (that is covered with ice)
should be flown according to recommendations by Kern and McDonald
(in prep) with the following changes. A laminated map with
survey lines spaced 1/2 mile apart should be used to survey the
main part of the bay at 50% coverage (every other survey line
with the first line, 1 or 2, being randomly chosen). The 1/8
mile strip along the perimeter of Icy Bay should not be flown
since seals were rarely seen in this area and fuel and time are
limited. All fiords should be circled for complete coverage at
1200 ft ASL. Circling provides more complete coverage and is
safer and easier to implement than the original zig zag pattern.
The observer should only count out to 1/8 mile since it is
difficult to categorize seals beyond this distance in high
density areas.
If seals are dispersed throughout Icy Bay, methods should follow
the recommendations of Kern and McDonald (in prep) with the
changes noted above. In Disenchantment Bay, we did not draw
the 1/2 mile transects on the base (laminated) map since we were
unsure what the safest flight direction might be on any
particular day. A transparency with transects 1/2 mile apart
should be placed over the main part of Disenchantment Bay if
seals are dispersed. It should be taped onto the base map and
photocopied when the observer returns to the office. Only areas
with ice need to be surveyed at 50% coverage.
Surveys (and analysis) of land haul outs in Disenchantment Bay
should follow the same methodology (aerial photography) as
described for pup proportion surveys. Surveys (and analysis) of
ice haul outs should follow the same methodology as Icy Bay
population surveys.

Surveys occurring over a short period of time (e.g. 3 days) may
miss cohorts of the population which haul out at different times
during the season (E Mathews, pers commun). This may have
occurred in 1994. Surveys in 1995 should span over a longer
period. There should be a minimum of 3 surveys between August 10
- 25, with at least 2 days between surveys if possible.
Additional surveys should be flown if funding and weather allow.
2) Aerial Photography
Aerial photography will also be used and should consist of at
least 30 non-overlapping slides of each seal aggregation. Prior
to the survey, the observer should take slides of a marked area
using the same flight altitude (1000 ft ASL) and camera focal
length (200 mm) that will be used in the actual survey. Three
markers should be placed 200 ft apart. This will allow for later
interpretation of the area of the slides. The average number of
seals per slide (e.g. per known area) can then be extrapolated to
the total area of the aggregation to obtain a population
estimate. Despite problems with this method in 1994, we feel it
holds good potential for assessing seal numbers as well as
leaving a permanent data record.
LITERATURE CITED
Allen, S.G., D.G. Ainley, G.W. Page and C.A. Ribic. 1984. The
effect of disturbance on harbor seal haul out patterns at
Bolinas Lagoon, California. Fish Bull 28:493-500.
Byrden MM. 1972. Growth and development of marine mammals. Pages
2-79n IN: RJ Harrison, ed. Functional anatomy of marine
mammals. Vol 1. Academic Press, New York. 451pp.
Calambokidis J, BL Taylor, SD Carter, GH Steiger, PK Dawson, LD
Antrim. 1987. Distribution and haul-out behavior of harbor
seals in Glacier Bay, Alaska. Can J Zool 65(6): 1391-1396.
Calambokidis J, LE Healey, and GH Steiger. 1983. Behavior of
harbor seals and their reactions to vessels in Glacier Bay,
Alaska. Abstract IN: Proceedings of the Fifth Biennial
Conference on the Biology of Marine Mammals. 27 Nov - 1 Dec.
1983. Boston, MA.
Gasaway WC, SD DuBois, RD Boetjie, DJ Reed, and SJ Harbo. 1986.
Estimating moose population parameters from aerial surveys.

Biol Paper No. 22, Univ of Alaska, Fairbanks.
Hoover AA. 1988. Harbor Seal (Phoca vitulina). IN: Lentfer JW,
ed. Selected Marine mammals of Alaska: Species accounts with
research and management recommendations. Marine Mammal
Commission, Wash. D.C. pp.125-157.
Hoover-Miller AA. 1992. The Harbor Seal in Alaska. IN: LentferJW,
ed. Revised Marine mammals of Alaska: Species accounts with
research and management recommendations. Marine Mammal
Commission, Wash. D.C. 78 pp.
Kozie KD. 1993. Coastal wildlife survey - seabirds and marine
mammals along the Malaspina Forelands 1992. Rep to Natl Park
Serv., Glennallen, AK
Loughlin TR. 1994. Abundance and distribution of harbor seals
(Phoca vitulina richardsi) southeastern Alaska during 1993.
Rep to NMFS, Seattle, WA. 42 pp.
Loughlin TR. 1992. Abundance and distribution of harbor seals
(Phoca vitulina richardsi) in Bristol Bay, Prince William
Sound, and Copper River Delta during 1991. Annual Report for
1991, NMFS, MMPA Popul. Assessment Program. Office of
Protected Resources, NMFS, Silver Spring, MD. 27 pp.
Mathews EA. 1992. Harbor seal censuses in Glacier Bay National
Park and Preserve: a comparison of aerial and land-based and
aerial censusing. NPS Report. Glacier Bay National Park
and Preserve, Gustavus, AK 99826.
Mathews EA. 1995 (in press). Long term trends in the abundance of
harbor seals (Phoca vitulina richardsi) and development of
monitoring methods in Glacier Bay National Park, southeast
Alaska. NPS Report. IN: D. Engstrom (ed) Proc from 3rd
Glacier Bay Sci Symposium, Sept., 1993.
Mathews EA and BP Kelly. (in press). Extreme temporal variation
in harbor seal (Phoca vitulina richardsi) numbers in Glacier
Bay, a glacial fiord in Southeast Alaska. Mar Mamm Sci.
Mineral Management Service. 1990. Outer continental shelf natural
gas and oil resource management comprehensive program 1992
to 1997. US Dept Inter, Mineral Management Serv, Washington,
D.C.

Pitcher KW. 1986. Assessment of Marine mammal-fishery
interactions in the western Gulf of Alaska and Bering Sea:
population status and trend of harbor seals in the
southeastern Bering Sea. Final report to NOAA, National
Marine Mammal Lab, Seattle, WA. (Contract No.
NA-85-ABH-00029, 12pp.
Pitcher KW. 1989. Harbor seal trend count surveys in southern
Alaska, 1988. Final report to Marine mammal commission,
Washington, D.C. Contract MM4465853-1,
Pitcher KW. 1990. Major decline in number of harbor seals
(Phoca vitulina richardsi) on Tugidak Island, Gulf of
Alaska, Mar Mamm Sci, 6(2):121-134.

Data Analysis
Multiple surveys (a minimum of 3 separated by at least 2 days) were
recommended during the molting season to estimate the total number of "seal
days". A seal day is defined as 1 seal hauled out on 1 day during the
survey period that could be potentially counted if a survey were conducted.
The average number of days a seal is hauled out must be estimated to use
this method. This information is currently unknown and would likely
require radio-tagging animals. However, data can be collected annually and
calculations corrected in the future when this data does become available.