Shark-inflicted mortality on a population of harbour seals (Phoca vitulina) at Sable Island, Nova Scotia
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
J. Zool., Lond. (2000) 252, 405±414 # 2000 The Zoological Society of London Printed in the United Kingdom
Shark-in¯icted mortality on a population of harbour seals
(Phoca vitulina) at Sable Island, Nova Scotia
Z. Lucas1 and W. T. Stobo2*
1
P.O. Box 3504 South, Halifax, Nova Scotia, B3J 3J2, Canada
2
Bedford Institute of Oceanography, Dartmouth, Nova Scotia, B2Y 4A2, Canada
(Accepted 3 November 1999)
Abstract
Shark-in¯icted mortality on harbour seals Phoca vitulina on Sable Island, Nova Scotia, was studied from
1980 to 1997, based on carcasses washed up on shore. During this period, pup production declined
dramatically from over 600 in 1989 to 40 in 1997. Between 1980 and 1992, pup deaths only were recorded,
and only during the May±June pupping period, while deaths in all age groups were recorded year-round
between 1993 and 1997; 458 pups, 23 juveniles and 241 adults were found. Shark-in¯icted mortality in
pups, as a proportion of total production, was under 10% during 1980±93, roughly 25% in 1994±95, and
increased to 45% in 1996. Shark-in¯icted mortality occurred in all months except December, January and
February, with c. 80% of the pups killed during the pupping period, and 97% of the adults killed outside
the pupping period. The decline in pup production was not only a result of reduced recruitment owing to
pup mortality. A greater proportion of reproductive females than males was killed. We estimate that
shark-in¯icted mortality on pups and adult females reduced pup production on Sable Island by 43 to 154
pups annually between 1993 and 1997. Our results indicate that sharks are having an impact on Sable
Island harbour seals, possibly to the extent of limiting population growth, or contributing to the observed
population decline. Potential reasons for this increased mortality are discussed.
Key words: sharks, predation, mortality, Phoca vitulina
INTRODUCTION killer whales and the caloric content of sea otters. Shark
predation on pinnipeds is a well-documented phenom-
Factors in¯uencing at-sea mortality of marine mammals enon (Riedman, 1990), with early reports of shark
are dif®cult to assess. LeBoeuf & Crocker (1996) predation on harbour seals Phoca vitulina in the north-
reported that an average of 46% of young-of-the year east Paci®c being given by Scheffer & Slipp (1944),
northern elephant seals Mirounga angustirostris from based on shark stomach contents. Although shark pre-
the Ano Nuevo rookery die at sea, but were unable to dation may exert signi®cant regulation on population
determine the cause(s) of this mortality. In¯uences and growth, there are few quanti®able data available to test
assumptions on mortality, based on sampling that is the impact at either the cohort or population level.
spatially and temporally limited, and potentially biased, Shark-in¯icted mortality has been suggested as a
are often used in attempts at modelling the dynamics of signi®cant factor in the population dynamics of the
the populations of pinnipeds (McLaren & Smith, 1985) Hawaiian monk seal Monachus schauinslandi (Kenyon,
and other marine mammals. The estimation of preda- 1981; Hiruki, Gilmartin et al., 1993). Hiruki, Stirling et
tion, as a component of mortality, has been dif®cult to al. (1993) concluded that such injuries to female
achieve because of even greater limitations of data. Hawaiian monk seals adversely affected reproductive
Estes et al. (1998) have suggested that the decline in success by reducing offspring survival, but they did not
western Alaska sea otters Enhydra lutris is the result of comment on the impact on cohort size or overall
predation by killer whales Orincas orca. Their provoca- population abundance. Recent studies at French Frigate
tive analysis was heavily dependent on assumptions Shoals, Hawaii (1995±98), suggested that shark-in¯icted
relating to population size, the caloric requirements of injuries were responsible for 20% of the deaths or
disappearance of nursing and weaning monk seal pups.
*All correspondence to: Dr W. T. Stobo. Most of the predation occurred near one pupping area
E-mail: stobow@mar.dfo-mpo-gc.ca where c. 60% of the French Frigate Shoals monk seals406 Z. Lucas and W. T. Stobo have been born (M. Craig pers. comm.). The impact on during 1993±97. The long sand beaches and the lack the population is still being assessed. of terrestrial scavengers or predators on the island Sharks may have a pervasive in¯uence on the facilitate the ®nding of carcasses of dead seals and behaviour and biology of many pinniped species. Shark- provide an opportunity to report minimum mortality in¯icted mortality has been suggested as a factor in the levels by season and seal age. This, in conjunction with evolution of pinniped behaviour such as prolonged long-term monitoring of pup production, allows us to suckling in the Galapagos fur seal Arctocephalus galapa- comment on the impact of shark predation on the goensis (Bonner, 1984). The abnormal timing of the harbour seal population. breeding season in northern elephant seals on the Farallon Islands (Ainley, Henderson et al., 1985), and their diving patterns (LeBoeuf & Crocker, 1996) may METHODS be, in part, adaptations to reduce encounters with near- surface predators. Shark predation on breeding harbour This study was based on shoreline surveys of carcasses seals is known to be common at the South Farallon washed ashore on Sable Island. It is probable that some Islands, California (Ainley, Strong et al., 1981; carcasses were not found because the animals were LeBoeuf, Riedman & Keyes, 1982; Stewart & Yochem, consumed entirely, were not washed ashore, or were 1985; Long et al., 1996), but the interactions are not well buried by windblown sand soon after beaching. The understood. Ainley, Henderson et al. (1985), suggested numbers presented in this paper are therefore minimum increases in white shark Carcharodon carcharias sight- estimates of shark-in¯icted mortality. ings around the Farallon Islands were more probably Data on shark-in¯icted mortality of harbour seal the result of increases in the numbers of northern pups were collected during the pupping season from elephant seals in that area than increases in white sharks 1980 to 1996, and between 1993 and 1997, data for all or environmental change; more recently Long et al. ages were collected year-round. Data on pups were (1996) have reaf®rmed that suggestion. But as McLaren collected primarily during annual tagging programmes. & Smith (1985) suggest, shark predation as an agent of During the pupping season (May±June) between 1978 population regulation, needs more exploration. As and 1996, almost all harbour seal pups born on Sable support for this contention, they refer to the suggestion Island were marked by applying Jumbo Rototags of Boulva & McLaren (1979) that harbour seal popula- (Dalton Supplies Ltd, Henley-on-Thames, Oxfordshire, tion growth on Sable Island may have been constrained UK) to the interdigital web on the hind ¯ipper. The tags by shark predation and the counter suggestion by were numbered on one side, with a `reward' notice and Brodie & Beck (1983) that the reduced shark abundance mailing address printed on the other side. Different in that area at the same time allowed the Sable Island coloured tags were used each year to facilitate visual grey seal population to expand. identi®cation of cohorts in subsequent years. Harbour Grey seals Halichoerus grypus and harbour seals seal pups are mobile, frequently enter the water within a breed on Sable Island, Nova Scotia, and are year-round few hours of birth, and attempt to escape into the water residents (Mans®eld & Beck, 1977; Boulva & McLaren, when approached. Thus we used sequentially numbered 1979; Stobo & Zwanenburg, 1990). The island (448N, tags to mark each pup on initial capture to facilitate 608W) is a 40-km-long sand bar in the north-west complete monitoring of pup production and eliminate Atlantic, located 170 km south-east of mainland Nova the possibility of double-counting. Scotia, and 40 km north of the edge of the continental During the 1980±82 pupping seasons, the island's shelf. Between December and April, we have seen a few perimeter was usually surveyed every other day, and mostly immature ringed seals Phoca hispida, harp seals during the 1983±96 seasons usually once a day, using Phoca groenlandica and hooded seals Cystophora cris- all-terrain vehicles. The long uninterrupted beaches of tata resting on the beaches of the island. Although we Sable Island and the speed of these vehicles allowed and others (Boulva & McLaren, 1979; Brodie & Beck, capture of pups before they could escape into the water. 1983) have observed shark predation on all ®ve species The naive young (usually < 2 days old when tagged) in the Sable Island area, the harbour seal population were relatively easy to capture; all dead pups were has the greatest proportion of observed shark-in¯icted counted and marked with paint to avoid duplicate mortality on pups and adults. counting. As a result of the frequency of circuits over Earlier studies indicated that shark predation on the pupping beaches, probably few pups were missed. Sable Island harbour seals was only occasionally high, The occurrence of only 1 or 2 newborn during the last and primarily on pups (Boulva & McLaren, 1979; several days of the annual ®eld trips suggests few pups Brodie & Beck, 1983). Boulva & McLaren (1979) would have been born afterwards. The 1997 pup pro- reported that wounded and dead seals, severely muti- duction was estimated from visual counts made during lated by shark bites, were seen most frequently on the surveys conducted on several consecutive days at the island between late July and late October. In this study, end of the pupping seasons we report the incidence of shark attacks on harbour seal Causes of mortality for 1978 and 1979 were not pups on Sable Island during the last 17 years, and on all recorded. The data from 1980 to 1983 were obtained age classes during 1993±97. We also report on a from the onset of pupping until roughly the end of the dramatic increase in predation on adult harbour seals pupping period. Between 1983 and 1993, observations
Shark-in¯icted mortality on habour seals 407
Table 1. Shark predation on harbour seal pups on Sable Island from 1980 to 1997. Pupping period mortality was selected as
mortality occurring from the start of pupping until 11 days (inclusive) after the median pupping date for each year. In 1997 data
were not collected on non-shark-related mortality or median pupping date, thus only cohort size and total shark-in¯icted
mortality estimates are available
Total observed Pupping period Total annual shark- Pupping period shark-
Year Cohort size mortality mortality in¯icted mortality in¯icted mortality
1980a 333 77 77 25 25
1981 391 28 28 1 1
1982 403 9 9 0 0
1983 482 28 24 4 4
1984 487 55 46 23 20
1985 526 78 40 44 16
1986 531 61 59 22 20
1987 577 54 43 23 16
1988 585 78 49 40 18
1989 625 36 31 10 8
1990 591 55 41 15 5
1991 570 36 34 18 14
1992 470 45 34 27 18
1993 380 37 13 33 7
1994 270 81 35 75 32
1995 180 49 13 43 10
1996 91 45 28 41 24
1997 40 14
a
The number of deaths attributed to sharks may be biased upwards as a result of a domestic dog also killing pups.
were progressively extended into the post-pupping on 6 June would be considered as a May kill). This
period, until by 1993 the full period from birth to adjustment allows for monthly values of total numbers
completion of weaning for each pup cohort was over the 5 years of the study even though there were
monitored. several 2- to 4-week periods during each year when
The frequency of the surveys around the island surveys were not carried out.
differed during, and following, the pupping season, and Shark-killed seals examined between 1993 and 1997
the duration of the ®eld seasons during the pupping were divided into 3 age groups (pups, juveniles, adults),
periods varied between years. Values are provided for with age being assigned on the basis of the calendar
total mortality observed, but for a common reference year. For example, a seal born in May 1993 and killed
period for all years, the mortality values from the start before the end of December 1993 would be assigned age
of pupping inclusive of 11 days after the median 0; if killed between January and the end of December it
pupping date for each year (the duration of the shortest would be assigned age 1. Seals killed during their second
®eld season) are also provided. (age 1) and third (age 2) calendar year of life were
Collection of data on shark-in¯icted mortality for considered juveniles, and seals killed during or after
all age groups of harbour seals on Sable Island began their fourth (age 3 and older) calendar year were
in 1993 when year-round surveys were carried out in considered adult. For untagged seals, age group was
addition to the ongoing pup-tagging programme of estimated using body size. While there could have been
May±June. From January 1993 to December 1997, the some dif®culty in accurately assigning age group to
beaches were surveyed for pups, immatures, and adults small untagged individuals, all but 2 of the 23 juveniles
killed by sharks, on average once every 5 days, except were tagged, and thus of known age (see Tables 3 and
for 2±3 periods each year of 2±4 weeks, when no 4). For the untagged animals, we compared their size
surveys were carried out. Survey effort was roughly and length to known-age animals. The 2 untagged
equal in all 5 years, with little seasonal bias. Information animals classi®ed as juveniles were, in size and length,
on time of kill, location, age, sex, condition, reproduc- clearly 1 or 2 years old. All of the untagged animals that
tive status and presence/absence of tags, was collected. we classi®ed as adults had body sizes and lengths that
Carcasses were marked, or removed from the beach, to indicated they were older than age 3.
prevent duplicate records. Carcasses were found in
various states, from very fresh (bleeding and with a core
temperature warm to the bare hand), to partly decom- RESULTS
posed and scavenged at sea. For decomposed and
scavenged carcasses found during the 1993±97 period Our examination of shark predation at Sable Island is
death was assumed to have occurred in the previous based on data collected until 1997. Although beach
month (e.g. a partly decomposed carcass washed ashore surveys continued throughout 1998, only 10 shark-408 Z. Lucas and W. T. Stobo
Shark kills/total deaths – total observed
1
Proportion of deaths due to sharks
Shark kills/total deaths – reference period
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96
Year
Fig. 1. The proportion of total annual mortality of harbour seal pups caused by sharks compared to that during a common
reference period during the pupping season, from initiation of pupping up to 11 days after the annual median pupping date.
50 Percentage
Per cent of of pups
pups that died
which died
production
a per cent ofofproduction
Percentage killed
Per cent killed byby sharks
sharks
40
a percentage
30
20
Deathsasas
10
Deaths
0
80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96
Year
Fig. 2. Total deaths, and shark-in¯icted deaths, of harbour seal pups given as a percentage of total annual production.
killed harbour seals were found: nine adults (seven a peak of 625 pups (Table 1), and subsequently de-
male, one female, and one of unknown sex), all between creased to low levels in 1997. Total observed annual
April and August, and one 1998 pup, in June. mortality, as a percentage of total production, from all
Total observed annual pup production and mortality causes was 23% in 1980, then decreased to < 15% before
due to sharks and other causes are given in Table 1. 1994 (Fig. 2); subsequently it increased to 50%. Much of
During 1980±97, 458 shark-killed pups were found. The the increase in recent years was owing to sharks, from
1980 value for pupping period mortality is abnormally < 10% of total observed annual production between
high because of deaths caused by a domestic dog before 1980 and 1993 to 27.8% and 23.9% in 1994 and 1995,
its removal from the island. Of the 77 that died in this respectively (Fig. 2). In 1996, sharks killed > 45% of all
year, 15 were known dog kills, and 25 of the remainder pups born. Shark-in¯icted pup mortality, as a percent-
were attributed to sharks, but the shark-in¯icted mor- age of total mortality ranged from 27% to 60% during
tality may be biased upwards because of insuf®cient 1984±92, increasing between 1993 and 1996 to c. 90%
diligence in ascertaining the cause of mortality; the annually. Shark-in¯icted mortality was, however, not
remainder could not be assigned to a speci®c cause. proportional to pup production. Predation on pups did
Total observed mortality and the mortality observed not increase as pup production increased, but when pup
during the common reference period (from the start of production declined by almost 100 pups per year
pupping inclusive of 11 days after the median pupping between 1991 and 1996, shark predation increased for
date) show similar trends (Fig. 1), indicating that most most of that period (Table 1).
of the observed shark-in¯icted pup mortality occurred Between 1993 and 1997, 470 harbour seals of all ages
during the pupping period. We therefore decided to were killed by sharks (Table 2). The trends in mortality
present our analyses based on total observed annual over that period (Fig. 3) differed by age group. While
mortality. pup mortality during the 5 years of observation peaked
Between 1980 and 1989, pup production increased to in 1994 and declined subsequently, total adult mortalityShark-in¯icted mortality on habour seals 409
Table 2. Harbour seals killed on Sable Island by sharks between 1993 and 1997. M, males; F, females; U, animals too mutilated
for sex to be determined (unknown)
1993 1994 1995 1996 1997
Age
group M F U M F U M F U M F U M F U Total
Pups 3 4 26 26 24 25 22 17 4 14 13 14 2 7 5 206
Juveniles 0 0 1 7 4 2 3 0 0 4 1 0 0 1 0 23
Adults 3 6 7 5 42 5 25 52 1 31 32 0 14 17 1 241
Totals 6 10 34 38 70 32 50 69 5 49 46 14 16 25 6 470
difference occurred only in 1994 and 1995 (w2 = 29.1 and
50 Pups
9.5, respectively; d.f. = 1, P = 0.05). Overall more adult
40 females than males were killed (64.6% compared to
34.4%), and in each year adult female mortality ex-
30
ceeded that of males. Since the sex of some of the seals
20 could not be determined, it is possible that this differ-
ence could occur if males tended to be over-represented
10
in the `unknown sex' category. However, even if all
0 adults of unknown sex were assumed to be males, in
every year except 1993 a positive bias towards females
50 Juveniles would remain.
There were 264 adults and juveniles (Table 2) killed;
Numbers killed
40
of these, 33 were too damaged or decomposed to
30
determine tagging status; 36 bore no evidence of having
20 been tagged; 43 showed tag loss, demonstrated by a neat
10 round tag hole in hind ¯ipper tissue; and 152 bore tags,
parts of broken tags, or brands, that showed they had
0 been marked, either as pups or adults, on Sable Island
50 during various studies carried out by Fisheries &
Adults Males Oceans Canada. Of these animals, 58 could be identi®ed
40 Females as having been tagged or branded as pups (before
30 Unknown weaning) on Sable Island, and thus the age of this sub-
20
set of animals could be established (Table 3). Over 70%
of the mortality of animals aged 1 year or older was
10 represented by animals < 10 years of age. The oldest
0 males killed were 9 years old, and the oldest females
1993 1994 1995 1996 1997 were 16 years old. In the age 1 year and older female
Year group, most deaths (65.5%) were females that were fully
Fig. 3. Numbers of harbour seal pups (age 0), juveniles (ages recruited to the reproductive component of the popula-
1±2) and adults (ages 3 and older), by sex, killed annually by tion (age 7 years and older).
sharks between 1993 and 1997. Data from the 1993±97 survey show that harbour
seals were killed by sharks during all months except
December, January and February (Fig. 4). Most of the
increased from 1993 to 1995, then declined. Relatively shark-in¯icted pup mortality occurred during the May±
little juvenile mortality was observed, which corre- June pupping season, with a cumulative mortality of
sponds with our observations that few juvenile harbour 167 pups (81.1%) during this period over the 5 years
seals are on the beaches around Sable Island throughout (Table 4). The lack of pup mortality for March±April, is
the year. Total adult mortality from sharks was greater an artifact of our ageing criterion; pups were designated
than that of pups during the 1993±97 period (Table 2). 1-year-old (i.e. juveniles) as of 1 January following their
Of the total shark-killed seals, sex could be identi®ed birth year. Three of the ®ve juveniles killed in March or
in 80.6%. There was no selective mortality of pups by April were actually < 12 months old. Most juvenile
sex in any year between 1993 and 1997 (w2 = 0.14, 0.08, mortality occurred during late summer and autumn
0.64, 0.02 and 1.30; d.f. = 1, P = 0.05) with almost equal (73.9%); < 5% of juvenile mortality occurred during the
numbers of male and female pups killed (50.8% and pupping period. Adult mortality exhibited a pattern
49.2%, respectively). More male than female juveniles similar to that for juveniles. Almost all the adult
were killed (70.0% and 30.0%, respectively), but the mortality (97%) occurred outside of the pupping period,
numbers were too few to warrant a statistical examina- mostly during the post-pupping period (75%); among
tion of vulnerability by sex. For adults, there seemed to these adults, 93.6% of the males, and 98.0% of the
be a selection for females (Fig. 3), although a signi®cant females were killed during this period (Table 4).410 Z. Lucas and W. T. Stobo
Table 3. Known-age juvenile and adult harbour seals killed by sharks on Sable Island from 1993 to 1997. Unknown: animals
that were known age, but too mutilated to be identi®ed by sex
Numbers at age
Sex 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Total
Male 5 7 4 1 1 2 1 ± 3 ± ± ± ± ± ± ± 24
Female 3 3 3 1 ± ± 2 ± ± 1 4 3 2 1 4 2 29
Unknown ± 3 ± ± ± 1 ± ± 1 ± ± ± ± ± ± ± 5
Total 8 13 7 2 1 3 3 0 4 1 4 3 2 1 4 2 58
Table 4. Seasonal mortality of harbour seals killed by sharks estimates of shark-in¯icted mortality. Some seals may
on Sable Island from 1993 to 1997 in numbers, per cent by have been killed far enough offshore that carcasses were
season for each age group, and per cent of the total mortality not beached, and a proportion of smaller (i.e. younger)
by season. No carcasses associated with shark kills were found
between December and March seals may have been consumed entirely, but we cannot
provide estimates of numbers. Long et al. (1996) sug-
March±April May±June July±Nov. gested that northern elephant seal and harbour seal
Age Total pups may be ingested whole by great white sharks based
Group n % n % n % no. on the observations of LeBoeuf et al. (1982) that the
Pups n/aa n/aa 167 81.1 39 18.9 206 stomach contents of white sharks often included large
Juveniles 5 21.7 1 4.4 17 73.9 23 portions of pinniped bodies. But we observed many
Adults 54 22.4 8 3.3 179 74.3 241 harbour seal pup carcasses washed ashore with largely
Totals 59 12.6 176 37.4 235 50.0 470 intact vertebral columns, suggesting that many were not
a entirely, or even largely, consumed. We are still investi-
Ages based on calendar year. Since pupping commences in
May, no pups (age 0) exist between January and May. gating the species of shark(s) involved. Sable Island is a
large sand bar with very gradual sloping beaches with
an intertidal zone, which often extends for 30±40 m.
During the last 5 years of the study, of the 34 adult Usually when harbour seal pups are observed in the
females killed during March±June (pre-pupping and water, they are in the zone between high and low tide
pupping periods), 12 carcasses were extensively levels. Consequently, most of the shark attacks on pups
damaged, and reproductive status could not be deter- would occur in these shallow water areas. Recent
mined. The remaining 22 carried foetuses. The studies (e.g. Osborne & Greenwood, 1992) suggest that
reproductive status of females killed between July and incident waves associated with shallow water areas,
November was not examined since pregnancy could not during the generally fairweather conditions of summer,
be assessed because of the potential effect of delayed tend to cause shoreward transport of sediment near the
implantation. seabed. Since pup carcasses sink to the bottom, this
phenomenon could also result in them being washed
ashore. Also, tidal currents around Sable Island are
DISCUSSION semi-diurnal ellipses rotating 360 8 in a clockwise direc-
tion over a tidal cycle (Anon, 1983), thus keeping the
All the data presented here are based on seals killed by carcasses in the vicinity of the island. These physical and
sharks and washed ashore, and thus represent minimum environmental features might cause the carcasses to
Monthly mortality by age group
140
Pups
120 Juveniles
Adults
Numbers killed
100
80
60
40
20
0
Mar Apr May Jun Jul
July Aug Sep Oct Nov
Month
Fig. 4. Seasonal distribution of shark-in¯icted mortality on harbour seal pups (age 0), juveniles (ages 1±2) and adults (ages 3
and older) between March and November. No shark-in¯icted mortality was observed during December, January and February.Shark-in¯icted mortality on habour seals 411 wash ashore to a greater extent than in many other each year since 1994. This is consistent with the decrease island rookeries. in pup production, beginning in the early 1990s. But Harbour seals killed by sharks have been observed on while predation on adult harbour seals increased in 1994 Sable Island since the 1970s (Boulva & McLaren, 1979; and remained high for 2 years, adult mortality declined Brodie & Beck, 1983). The predation reported in these in 1997, and in 1998 we observed only nine adults killed. studies was primarily on pups, but this may have been a During the same period, however, we observed consis- result of most observations being made during the tently high predation on grey seals. The diminishing harbour seal pupping period. While shark predation on numbers of shark-killed harbour seas found is therefore pups increased from < 10% in 1993 to > 45% in 1996, more probably the result of a decline in the numbers of adult mortality exceeded pup mortality overall (241 vs. harbour seals in the area than to a decline in shark 206), and annually, in 3 of the 5 years of full-year activity. It is possible that the adult harbour seals observations. These observations are in contrast with remaining are those that are more wary of sharks and those of Ainley, Henderson et al. (1985) and Long et al. thus able to avoid attacks, but Ellis (1998) reported (1996), who reported a greater vulnerability of pups. In declines in the numbers of males, juveniles and pups in agreement with these authors, our observations do, the Sable Island population between 1991 and 1996, and however, show a season effect. Most of the Sable Island in females from 1993 to 1996. These observations were harbour seal pup mortality occurred during the pupping based on weekly counts of males, females and pups and weaning period (mid-summer), with that of adults observed on, or near, a 24-km section of beach on the primarily occurring outside the pupping period. This north side of Sable Island during the pupping season. shift in prey size seasonally suggests that the decline in Although this study was restricted spatially and tempo- pup mortality outside of mid-summer may be the result rally, and the data did not include shark kill counts, the of a shift in pup distribution rather than a change in the observations of reduced numbers are consistent with abundance of sharks in the area. our study. Pup production of Sable Island harbour seals exhib- Brodie & Beck (1983) suggested that depletion of the ited a gradual and consistent increase between 1980 and shark population because of commercial ®shing may 1989, followed by a dramatic decline between 1989 and have been a factor in the increased production observed 1997. Shark-in¯icted pup mortality, however, was much for Sable Island grey seals. Cessation of those ®sheries more variable (Table 1), with a decline during the peak in the mid-1970s may have resulted in a recovery of the pup production years (1985±89) and an increase during shark populations in the area, leading to the increased the subsequent decline. This pattern suggests that shark- predation observed on harbour seals, but there are no in¯icted pup mortality was also in¯uenced by factors quantitative estimates of shark abundance, or informa- other than harbour seal pup availability (i.e. produc- tion on the species of sharks involved, to assess that tion). We do not have the data to attempt correlations possibility. The inter-annual variation and the lack of with environmental factors or shark abundance. an increasing trend in shark-in¯icted pup mortality However, pup production by the grey seal population between 1980 and 1989 suggest that a complex interac- breeding on Sable Island was increasing at over 12% tion exists between the sharks and their harbour seal annually during the same period (Stobo & Zwanenburg, prey. Boulva & McLaren (1979) reported that wounded 1990; Mohn & Bowen, 1996). Ainley, Henderson et al. and dead seals, severely mutilated by shark bites, were (1985) did not ®nd a correlation with environmental most frequent on Sable Island during the post-pupping variables. However, Ainley, Henderson et al. (1985) and period, between late July and late October. The differ- Long et al. (1996) reported an increase in shark attacks ence between their observations and ours suggests some with increased abundance of pinnipeds, the overall shift in the timing of shark presence around the island. increase being due to one of several cohabiting species. The greater proportion of pups killed during the But although Long et al. (1996) concluded that an May±June pupping season in our study suggests that increase in white sharks had occurred, Ainley, pups are more vulnerable during the mid-summer Henderson et al. (1985) suggested that only a few sharks period. The small number of adult and juvenile kills may have been responsible and that the increase in during the pupping period suggests that sharks may attacks was related to availability of prey and an select pups during the time when the pups are likely to increase in the size of the individual predators over time. be concentrated in the waters adjacent to the island They suggested that a few young white sharks may have pupping areas. found the growing food source and continued to use it The decline in harbour seal pup production cannot be as they grew larger. In a somewhat similar manner, explained entirely by a reduction in per capita recruit- Estes et al. (1999) suggested that a few apex predators ment resulting from shark-in¯icted mortality. If the (killer whales) might have been the cause of a large problem was limited to new recruitment, one would decline in the abundance of sea otters in Alaska. expect to have observed a stabilizing of production near It is probable that the continued presence of sharks in the peak reached in 1989 with a gradual decline as a the Sable Island area is not the result of harbour seal result of adult female mortality. All adult females for abundance. Predation on the combined age groups of which pregnancy status could be determined, killed harbour seals was highest in 1994, but pups killed as a during March±June of the 5 years of our study, were proportion of total mortality of all age groups declined pregnant. Boulva & McLaren (1979) found pregnancy
412 Z. Lucas and W. T. Stobo
rates for mature harbour seal females (6 years of age Table 5. Estimated number of adult females killed by sharks
and older) of 95%, similar to our observations. Thus, between 1993 and 1997, in total and by age
the observed decline has to be at least partly related to a Age composition
reduction in the numbers of adult females in the Total females
population. Year killed 3 4 5 6+
A greater proportion of reproductive females than
1993 11 1 0 0 10
males was killed. Among mature seals of known age
1994 47 6 2 2 37
and sex, six (25.0%) males and 19 (65.5%) females were 1995 53 7 2 2 42
at least 6 years old when killed. We do not have an 1996 32 4 1 1 26
explanation for this differential adult mortality for age 1997 18 2 1 1 14
within sex, or by sex. Mature males may be less vulner-
able to shark attack, because of their slightly larger size
or their behaviour, than young males or females. Since animals was assumed to be the same as for age 4; (4) the
mature male harbour seals (5+ years) generally experi- rates for Sable Island female harbour seals reaching
ence higher mortality rates than females (see Heide- reproductive age, calculated by Boulva & McLaren
Jorgenson & Harkonen, 1988), the greater proportion (1979) for ages 3, 4, 5 and `6 and older' of 0.27, 0.55,
of older females represented in the shark-kills could also 0.79 and 0.95, still apply; (5) an annual mortality rate of
re¯ect a greater proportion of older females in the 10% was chosen to account for non-shark-related
population. If the differential mortality were related to deaths. We chose this estimate of non-shark-related
differing abundances, then one would expect the pro- mortality based on a total annual mortality estimate
portional mortalities to have stayed roughly the same made by Boulva & McLaren (1979), as we had no direct
throughout the study. Our data show that 67±90% of measure of overall mortality, and the literature (see
the adult mortality was of females in the early years of Heide-Jorgensen & Harkonen, 1988) provides only
the study (1993±95), but only 50±54% of the adult broad ranges without distinguishing between shark-
deaths were of females in 1996±97. This change in the related and other causes. Boulva & McLaren (1979)
mortalities from predominantly adult female could be provided an annual mortality rate estimate of 17.5% for
interpreted as indicating that mortality was proportional a harbour seal population in eastern Nova Scotia that
to abundance, with the 1 : 1 ratio being approached as they felt was kept at a near stationary stable-age dis-
abundances were equalized. Alternatively, if sharks tribution by an active bounty hunt; again this estimate
select for adult females, the ratio observed in 1996±97 did not distinguish between sources of natural mortality.
could indicate that the abundance of adult females had Using the ®rst three assumptions, the proportions of
been reduced below that of adult males. adult females killed at ages 3, 4, 5 and `6 and older' were
Our observations during the pupping and immediate estimated to be 0.13, 0.043, 0.043 and 0.784, respectively
post-pupping period suggest a high incidence of shark- (Table 5). Summing progressively along the diagonals
in¯icted pup mortality since the mid-1980s, which (i.e. following cohorts) to obtain the incremental loss of
would have limited the number of females recruiting to reproductive potential, and applying the adjustments
the reproductive group. Ellis (1998) found that, concur- for mortality and partial recruitment, we estimated a
rent with the steady decline in the number of harbour loss in potential pup production for the study period
seal pups born on Sable Island after 1991, the age increasing from 10 pups in 1994 to 127 pups in 1997.
distribution of live, known-age, sexually mature females These estimates of potential reproductive loss re-
shifted to older ages She suggested that the trend sulting from adult female mortality, combined with
re¯ected a lack of recruitment of young females to the observed pup mortality between 1994 and 1997 (see
breeding population, possibly as a result of higher Table 1), suggest that sharks reduced production on
mortality or emigration of young females. Sable Island by 43, 124, 134, 154 and 141 pups, respec-
We examined the potential impact of the deaths of tively, for the years 1993±97. These observations and
adult females on pup production by estimating the estimates of pup loss account for < 50% of the decline
potential reproductive output of those adult females that was observed annually during 1993±97 from the
killed by sharks. Juveniles were excluded from this peak production in 1989. However, we have not at-
procedure because of the small number involved and the tempted to estimate the numbers of pups that would
likelihood of the higher, but unknown, mortality of have been completely consumed by sharks and thus not
juveniles compared to adults. Five assumptions were observed by us, nor have we tried to account for the
made in estimating the reproductive output: (1) the ratio potential reproductive loss to the population of the
of adult females to males in Table 2 was representative shark-killed female pups that never reached reproductive
of the ratio in the `unknown sex' category; (2) the age. It is generally accepted that juveniles experience
proportions of females at ages 3, 4, 5 and `6 or older' higher mortality rates than adults, but we have been
from Table 3 were representative of the age structure of unable to ®nd estimates of juvenile mortality that we
the shark-killed animals and was used to derive could use in this treatment. Therefore no attempt was
numbers at age for each year; (3) because of the small made to account for losses of juveniles, although inclu-
sample size and the tendency for more females than sion of such potential losses would more closely
males to be taken over time, the proportion for age 5 approach the observed reductions in pup numbers.Shark-in¯icted mortality on habour seals 413
Several other factors may have played a part in both I. A. McLaren, and C. McGowan provided constructive
the reduced total numbers and the diminishing numbers comments on an earlier version of the manuscript. ZL
of pups born on Sable Island. Harbour seals may be also received logistical support from Nova Scotia Natural
shifting their distribution southward, associated with Resources, the Nova Scotia Petroleum Directorate,
changing environmental conditions. Drinkwater, PanCanadian Resources and Mobil Oil Canada.
Colburne & Gilbert (1995) have shown a reduction in
water temperatures throughout the east coast of Canada
during the early 1990s. Frank, Carscadden & Simon REFERENCES
(1996) have documented the recent appearance and
Ainley, D. G., Henderson, R. P., Huber, H. R., Boekelheide,
increase in the abundance of a cold-water species, R. J., Allen, S. G. & McElroy, T. L. (1985). Dynamics of white
capelin, in the area; and we have observed an increase in shark/pinniped interactions in the Gulf of the Farallones. Mem.
the occurrence of harp and hooded seals on Sable Island South. Calif. Acad. Sci. 9: 109±122.
in recent years. The grey seal population on Sable Ainley, D. G., Strong, C. S., Huber, H. R., Lewis, T. J. & Morrell,
Island has been increasing by > 12% annually since the S. H. (1981). Predation by sharks on pinnipeds at the Farallon
early 1960s (Stobo & Zwanenburg, 1990) to an estimated Islands. Fish Bull. 78: 941±945.
Anon (1983). Venture development project environmental impact
40 000+ animals by the late 1980s (Zwanenburg &
statement IIa. Biophysical assessment. Mobil Oil Canada Ltd.
Bowen, 1990), and > 85 000 by 1994 (Mohn & Bowen, Submitted to Candada Oil and Gas Lands Administration,
1996). Coincident with this population increase, we Ottawa.
have observed increased use of the beaches throughout Boness, D. J., Bowen, W. D. & Oftedal, O. T. (1992). In¯uence of
the year by grey seals, now extending into areas pre- storms and maternal size on mother±pup separations and
viously used only by harbour seals for pupping, fostering in the harbor seal. Can. J. Zool. 70: 1640±1644.
moulting and haul-out. The increased grey seal numbers Boness, D. J., Bowen, W. D. & Oftedal, O. T. (1994). Evidence of
a maternal foraging cycle resembling that of otariid seals in a
may have contributed to the reduced abundance of small phocid, the harbor seal. Behav. Ecol. Sociobiol. 34:
harbour seals by passive interference or resource compe- 95±104.
tition. Scienti®c investigations of harbour seals on Sable Bonner, W. N. (1984). Seals of the Galapagos Islands. Biol. J.
Island also increased dramatically in the late 1980s, and Linn. Soc. 21: 177±184.
that research continued into the mid-1990s (e.g. Ross, Boulva, J. & McLaren, I. A. (1979). Biology of the harbour seal,
1990; Muelbert, 1991; Boness, Bowen & Oftedal, 1992, Phoca vitulina, in eastern Canada. Bull. Fish. Res Bd Can. 200:
1994; Bowen, Oftedal & Boness, 1992; Walker, 1992; 1±24.
Bowen, W. D., Boness, D. J. & Iverson, S. J. (1998). Estimation of
Bowen, Oftedal, Boness & Iverson, 1994; Coltman, total body water in harbour seals: how useful is bioelectrical
1997; Bowen, Boness & Iverson, 1998; Ellis, 1998). impedance analysis? Mar. Mamm. Sci. 14: 767±777.
Harbour seal pups may have been suf®ciently disturbed Bowen, W. D., Oftedal, O. T. & Boness, D. J. (1992). Mass and
by the combined impact of the long-term tagging energy transfer during lactation in a small phocid, the harbour
programme and the increased intensity of research seal. Physiol. Zool. 65: 844±866.
activities, beginning in the late 1980s, to spend more Bowen, W. D., Oftedal, O. T., Boness, D. J., & Iverson, S. J.
(1994). The effect of maternal age and other factors on birth
time in the water, increasing their vulnerability to preda-
mass of harbour seals. Can. J. Zool. 72: 8±14.
tion. Adult females may have responded by abandoning Brodie, P. & Beck, B. (1983). Predation by sharks on the grey seal
Sable Island as a reproductive area. Unfortunately, (Halichoerus grypus) in eastern Canada. Can. J. Fish. Aquat
there are few data on harbour seal abundance elsewhere Sci. 40: 267±271.
in Atlantic Canada, and thus if they are shifting to other Coltman, D. W. (1997). Phenotype and mating success of male
areas, these movements are not documented. harbour seals, Phoca vitulina, at Sable Island, Nova Scotia. PhD
This study shows that sharks are having an impact on thesis, Dalhousie University, Halifax, Nova Scotia, Canada.
Drinkwater, K. F., Colburne, E. & Gilbert, D. (1995). Overview
Sable Island harbour seals, possibly to the extent of of environmental conditions in the northwest Atlantic in 1994.
limiting population growth, or causing population North. Atl. Fish. Org. SCR Doc. No. 95/43.
decline. The abundance and species composition of Ellis, S. L. (1998). Maternal effects on offspring traits from birth
sharks in the area are not known, although examination through weaning in the harbour seal, Phoca vitulina. PhD thesis,
of wounds on harbour seals suggests that only a few Dalhousie Universty, Halifax, Nova Scotia.
species are involved. Additional research is required on Estes, J. A., Tinker, M. T., Williams, T. M. & Doak, D. F. (1998).
factors affecting the vulnerability of seal populations to Killer whale predation on sea otters linking oceanic and
nearshore ecosystems. Science 282: 473±476.
shark attacks and on the abundance of sharks necessary Frank, K. T., Carscadden, J. E. & Simon, J. E. (1996). Recent
to affect such population control. excursions of capelin (Mallotus villosus) to the Scotian Shelf
and Flemish Cap during anomalous hydrographic conditions.
Can. J. Fish. Aq. Sci. 53: 1473±1486.
Acknowledgements Heide-Jorgensen, M.-P. & Harkonen, T. J. (1988). Rebuilding seal
stocks in the Kattegat-Skagerrak, Mar. Mamm. Sci. 4:
231±246.
We thank G. Forbes and the Sable Island Atmospheric
Hiruki, L. M., Gilmartin, W. G., Becker, B. L. & Stirling, I. A.
Research Station staff (Atmospheric Environment (1993). Wounding in Hawaiian monk seals (Monachus schauin-
Service) for assistance throughout the study, and slandi) Can. J. Zool. 7: 458±468.
B. Beck, J. McMillan and G. M. Fowler for their help Hiruki, L. M., Stirling, I. A., Gilmartin, W. G., Johanos, T. C. &
in the harbour seal pup tagging work. Drs K. Kovacs, Becker, B. L. (1993). Signi®cance of wounding to female414 Z. Lucas and W. T. Stobo reproductive success in Hawaiian monk seals (Monachus concolor) on Sable Island. MSc thesis, Dalhousie University, schauinslandi) at Laysan Island. Can. J. Zool. 7: 469±474. Halifax, Nova Scotia, Canada. Kenyon, W. K. (1981). The Hawaiian monk seal: an endangered Osborne, P. D. & Greenwood, B. (1992). Frequency dependent marine mammal. Water Spectrum 13: 1±6. cross-shore suspended sediment transport, 1. A non-barred LeBoeuf, B. J. & Crocker, D. E. (1996). Diving behavior of shoreface. Mar. Geol. 106: 1±24. elephant seals: implications for predator avoidance. In Great Riedman, M. (1990). The pinnipeds: seals, sea lions, and walruses. white sharks, the biology of Carcharodon carcharias: 193±205. Los Angeles: University of California Press. Klimley, A. D. & Ainley, D. G. (Eds). London: Academic Ross, P. S. (1990). Immunocompetence of free-ranging harbour Press. seals (Phoca vitulina) mothers and their pups over the course of LeBoeuf, B. J., Riedman, M. & Keyes, R. S. (1982). White shark lactation MSc thesis, Dalhousie University, Halifax, Nova predation on pinnipeds in California coastal waters. Fish. Bull. Scotia, Canada. 80: 891±895. Scheffer, V. B. & Slipp. J. W. (1944). The harbour seal in Long, D. J., Hanni, K. D., Pyle, P., Roletto, J., Jones, R. E. & Washington State. Am. Midl. Nat. 32: 373±416. Bandar, R. (1996). White shark predation on four pinniped Stewart, B. S. & Yochem, P. K. (1985). Radio-tagged harbour species in central California waters: geographic and temporal seal, Phoca vitulina richardsi, eaten by white shark, Carchar- patterns inferred from wounded carcases. In Great white sharks, odon carcharias, in the Southern California Bight. Calif. Fish the biology of Carcharodon carcharias: 263±274. Klimley, A. D. Game. 71: 107±128. & Ainley, D. G. (Eds). London: Academic Press. Stobo, W. T. & Zwanenburg, K. C. T. (1990). Grey seal Mans®eld, A. W. & Beck, B. (1977). The grey seal in eastern (Halichoerus grypus) pup production of Sable Island and Canada. Fish. Mar. Serv. Tech. Rep. 704, 1±81. estimates of recent production in the northwest Atlantic. Can. McLaren, I. A. & Smith, T. G. (1985). Population ecology of Bull. Fish. Aquat. Sci. 222: 171±184. seals: retrospective and prospective views. Mar. Mamm. Sci. 1: Walker, B. G. (1992). Mass changes and feeding behaviour of sub- 54±83. adult and adult male harbour seals (Phoca vitulina) during the Mohn, R. & Bowen, W. D. (1996). Grey seal predation on the breeding season. MSc thesis, Dalhousie University, Halifax, eastern Scotian Shelf: modelling the impact on Atlantic cod. Nova Scotia, Canada. Can. J. Fish. Aquat. Sci. 53: 2722±2738. Zwanenburg, K. C. T. & Bowen, W. D. (1990). Population trends Muelbert, M. M. C. (1991). Weaning and post-weaning changes in in the grey seal (Halichoerus grypus) in eastern Canada. Can. body mass and composition of harbour seal pups (Phoca vitulina Bull. Fish. Aquat. Sci. 222: 185±197.
You can also read
