Developing population monitoring protocols for Australian sea lions
Photo: S Goldsworthy
Final Report to the Department of the Environment and Water Resources
SD Goldsworthy1, PD Shaughnessy2, B Page1, TE Dennis3, RR McIntosh4, D Hamer1, KJ Peters1, AMM Baylis1, A Lowther4, CJA Bradshaw5
1 South Australian Research and Development Institute (SARDI), 2 Hamra Avenue, West Beach SA 5024 2 South Australian Museum, North Terrace, Adelaide, SA 5000 3 5 Bell Court, Encounter Bay SA 5211 4 Zoology Department, La Trobe University, Victoria, 3086 5 School for Environmental Research, Charles Darwin University, Darwin, NT, 0909
Developing population monitoring protocols for Australian sea lions SD Goldsworthy, PD Shaughnessy, B Page, TE Dennis, RR McIntosh, D Hamer, KJ Peters, AMM Baylis, A Lowther, CJA Bradshaw South Australian Research and Development Institute SARDI Aquatic Sciences 2 Hamra Avenue West Beach SA 5024 Telephone: (08) 8207 5400 Facsimile: (08) 8207 5481 http://www.sardi.sa.gov.au/ Disclaimer Copyright Department of the Environment and Water Resources and South Australian Research and Development Institute 2007. This work is copyright. Except as permitted under the Copyright Act 1968 (Commonwealth), no part of this publication may be reproduced by any process, electronic or otherwise, without the specific written permission of the copyright owners. Neither may information be stored electronically in any form whatsoever without such permission. Printed in Adelaide, July 2007 SARDI Aquatic Sciences Publication Number F2007/000554 SARDI Research Report Series No. 219 ISBN Number: 9780730853701 Authors: SD Goldsworthy, PD Shaughnessy, B Page, TE Dennis, RR McIntosh,
D Hamer, KJ Peters, AMM Baylis, A Lowther, CJA Bradshaw Reviewers: R. McGarvey, M. Steer Approved by: T. Ward
Signed: Date: 12 July 2007 Circulation: Public Domain
Table of contents 1
TABLE OF CONTENTS
TABLE OF CONTENTS...........................................................................................................1 1 EXECUTIVE SUMMARY....................................................................................................2 2 INTRODUCTION ................................................................................................................4 3 REVIEW OF CURRENT METHODS USED TO ASSESS THE STATUS OF
AUSTRALIAN SEA LION POPULATIONS .......................................................................8 4 DEVELOPMENT OF SURVEY PROTOCOLS FOR LARGE AUSTRALIAN SEA LION
COLONIES .......................................................................................................................18 INTRODUCTION ...................................................................................................................18 METHODS...........................................................................................................................18 RESULTS............................................................................................................................22 DISCUSSION .......................................................................................................................30
5 DEVELOPMENT OF A SURVEY PROTOCOL FOR SMALL AUSTRALIAN SEA LION COLONIES .......................................................................................................................36 INTRODUCTION ...................................................................................................................36 METHODS...........................................................................................................................37 RESULTS............................................................................................................................37 DISCUSSION .......................................................................................................................38
6 DEVELOPMENT OF SUBPOPULATION SURVEY STRATEGIES FOR AUSTRALIAN SEA LIONS ......................................................................................................................40 INTRODUCTION ...................................................................................................................40 METHODS...........................................................................................................................41 RESULTS............................................................................................................................42 DISCUSSION .......................................................................................................................49
7 CONCLUSIONS AND RECOMMENDATIONS................................................................57 8 ACKOWLEDGMENTS .....................................................................................................59 9 REFERENCES .................................................................................................................60 10. APPENDICES..................................................................................................................66
APPENDIX 1........................................................................................................................66 APPENDIX 2........................................................................................................................70
Executive summary 2
1 EXECUTIVE SUMMARY
Australian sea lions (ASL) were listed as a threatened species under the EPBC Act in
February 2005. Information on the size and status of most subpopulations is poor and
significantly hampers developing appropriate management strategies for the species. Many
aspects of the species’ breeding biology and ecology are unique among otariids (fur seals
and sea lions) and make accurate assessment of pup production challenging. Unlike other
otariids where pupping seasons are short, with all pups being easily recognisable and
available for sighting at the end of the breeding season, in Australian sea lion populations the
breeding season can extend for up to eight months. As such, a significant portion of the pups
may have fully moulted, dispersed or died by the end of the pupping season.
Traditionally direct counting methods have been used to determine ASL pup production,
however, these are prone to underestimation due to sightability and availability biases, and
only provide point estimates of numbers with no confidence limits. Recently, mark-recapture
methods using the Petersen estimate have improved estimates of ASL pup production at
some large colonies, and have addressed some of the under estimation caused by
sightability biases. However, dispersal and unaccounted mortality (availability biases) may
still cause significant underestimation of actual pup production, and immigration has been
observed in some small colonies.
We developed and tested the appropriateness of two new methods for estimating pup
production in ASL subpopulations. The first utilised individual resight histories of pups and
Cormack-Jolly-Seber (CJS) models in conjunction with standard mark-recapture methods to
improve estimates of pup production for large ASL subpopulations (>40 pups). The second
developed a cumulative mark and count (CMC) method for improving estimates of pup
production in small ASL subpopulations (<40 pups).
CJS methods were trialled at Olive Island and produced pup production estimates that were
greater than those based on direct counting and on mark-recapture (Petersen estimate)
methods. Pup mortality during the study period was estimated to range from 15-52. As
recovered mortalities numbered 34 in total, ground surveys may have underestimated pup
mortality by up to 35%. There was no evidence for permanent emigration, suggesting that
the most important source of error in mark-recapture procedures at Olive Island were due to
unaccounted mortality. The best estimate of pup production for the 2006 season at Olive
Island based on CJS methods was 205 (range 193-256). This was 1.37 times the estimate
based on direct counting methods (150 pups), but was similar to the result (1.03 times larger)
Executive summary 3
obtained from the Petersen estimate (mean 197, range 191-203). However, an adjusted
Petersen estimate (adding the mortality range 34-52) produced the same estimate as the
CJS approach (206, range 191-223).
CJS analyses suggested no significant pup production occurred beyond the second session,
only three months into the breeding season. This was contrary to observations of the
presence of perinatal mothers and new born pups up to session 5 (19 June), increases in
pup abundance between sessions 2-5 based on the sum of the number of tagged pups, dead
pups and unmarked pups, and the Petersen estimate, and on evidence of a 5-7 months
breeding season elsewhere. The reasons for disparities in methods are currently unclear, but
should be addressed in the future.
The cumulative mark and count (CMC) method trialled at a small colony (Seal Slide,
Kangaroo Island), supported the observation that not all pups are available for counting
during ground surveys, and produced a consistent (repeatable) estimate on two occasions
(10 pups). The surveys would have benefited from greater numbers of pups being marked
over more sessions. The development of both CJS and CMC methods has advanced the
methods of monitoring for both large and small ASL colonies.
Distance analysis among ASL subpopulations identified 11 metapopulations in the species,
seven of them were in South Australia (SA). Among SA metapopulations, only four provided
sites where accurate, repeatable, cost effective and logistically feasible surveys could be
undertaken. Within each of these, one large (>40 pups) and one small (<40 pups) site were
selected (8 in total) as regionally representative sites to form the basis for ongoing surveys.
Ongoing surveys at these sites would provide critical data on the status and trends in
abundance within metapopulations to support management of the species across its range.
Refining of CJS and CMC methods (including trialling at WA subpopulations and
identification of regionally representative subpopulations) is recommended so that national
standards in subpopulation monitoring protocols can be developed. Resources need to be
secured and coordinated to support ongoing assessment of key representative sites across
the range of the species, to provide critical information on subpopulation status and trends,
and to measure any recovery in the species.
Introduction 4
2 INTRODUCTION
Background The Australian sea lion (ASL), Neophoca cinerea, is one of five sea lion species in the world.
Sea lions form around one-third of species in the Otariidae family of seals that includes all of
the fur seals and sea lions. Over recent decades there has been growing concern over the
status of all five sea lion species. In the North Pacific Ocean, the Steller sea lion, Eumetopias
jubatus, has been declared endangered in parts of its range and is considered threatened
with extinction in other parts (Trites et al. 2007). Although the total population of California
sea lions in California and Mexico is increasing (Caretta et al. 2004), the Mexican stock is in
decline (Szteren et al. 2006). There have also been reductions in numbers of the Galapagos
subspecies of the Californian sea lion, Zalophus californianus wollebaeki (Alava and Salazar
2006), and the Japanese subspecies, Z. c. japonicus, is possibly extinct (Mate 1982).
Numbers of South American sea lions, Otaria flavescens, have reduced considerably in
recent years (Crespo and Pedraza 1991, Reyes et al. 1999, Shiavini et al. 2004), especially
in the Falkland Islands (Thompson et al. 2005), and numbers of New Zealand sea lions,
Phocarctos hookeri (Lalas and Bradshaw 2003), and Australian sea lions (McKenzie et al.
2005) have not recovered from historic sealing, and form the smallest populations of all sea
lion species.
The ASL is Australia’s only endemic and least-abundant seal species. It is unique among
pinnipeds in being the only species that has a non-annual breeding cycle (Gales et al. 1994).
Furthermore, breeding is temporally asynchronous across its range (Gales et al. 1994, Gales
and Costa 1997). It has the longest gestation period of any pinniped, and a protracted
breeding and lactation period (Higgins and Gass 1993, Gales and Costa1997). The
evolutionary determinates of this atypical life-history remain enigmatic. Recent population
genetic studies have indicated little or no interchange of females among breeding colonies,
even those separated by short (20 km) distances (Campbell 2003). The important
management implication of extreme levels of female natal site-fidelity (philopatry) is that each
colony effectively represents a closed population.
There are 73 known breeding locations for ASLs, 47 of which occur in South Australia where
the species is most numerous (80% of pups counted), with the remainder (26 colonies)
occurring in Western Australia (McKenzie et al. 2005). The species was subject to sealing in
the late 18th, the 19th and early 20th centuries, resulting in a reduction in overall population
Introduction 5
size and extirpation of populations in Bass Strait and other localities within its current range.
Total pup production for the entire species during each breeding cycle has been estimated at
about 2,500 with an estimated overall population size based on a demographic model
developed by Goldsworthy et al (2003), of around 9,800 (McKenzie et al. 2005). A re-
analysis of this demographic model, in conjunction with improved estimates of pup
production for some sites, has increased estimates of the SA pup production to about 2,700
per breeding cycle and the size of the ASL population in SA to about10,900 individuals
(Goldsworthy et al. in review). Based on pup production estimates of 709 for WA sites
(Goldsworthy et al. 2003), the total pup production for the species is currently estimated at
about 3400 per breeding cycle, with an estimated overall population estimate of around
14,000 (Goldsworthy, unpublished data). The life tables associated with the population model
produced population estimates that were 4.08 times that of pup production (Goldsworthy et
al. in review), which is about mid-point of the range expected for pinniped populations
(Harwood and Prime 1978).
There are 39 ASL breeding sites in SA, when the criterion for classification as a breeding
colony is set at ≥ 5 pups present per breeding cycle (McKenzie et al. 2005, see Fig. 6.1). Of
these, only six (16%) produce more than 100 pups, and these account for 67 % of the State’s
pup production. The largest population is Dangerous Reef in Southern Spencer Gulf (585
pups), followed by The Pages (577 pups) in Backstairs Passage between Kangaroo Island
and mainland Australia. The next largest populations are Seal Bay (214 pups) on Kangaroo
Island, West Waldegrave (157 pups) and Olive Islands (131 pups) off the west coast of the
Eyre Peninsula, and Purdie Island (132 pups) in the Nuyts Archipelago (summarised in
Goldsworthy et al. in review). The median pup production for SA colonies is 25.5 per colony,
with 60% of breeding sites producing fewer than 30 pups per season, 42 % fewer than 20
pups, and 13% fewer than 10 pups (Goldsworthy et al. in review). These analyses do not
take into account at least another 11 breeding sites (termed ‘haul-out sites’ with occasional
pupping), where fewer than 5 pups have been recorded at some time (McKenzie et al. 2005).
Although the pre-harvested population size of the ASL is unknown, the overall population is
still believed to be in recovery. Unlike Australian fur seal, Arctocephalus pusillus doriferus
and New Zealand fur seal, Arctocephalus forsteri populations, which have been recovering
rapidly throughout southern Australia, there is a general view that the overall population
recovery of the Australian sea lion appears to be limited, and it is unclear why.
One of the most critical issues impeding effective management of ASL is the high uncertainty
in estimates of the size and status of sub-populations throughout their range. Most sub-
Introduction 6
populations are scattered on remote offshore islands and the non-annual, asynchronous and
protracted breeding seasons have made it difficult to obtain accurate estimates of pup
production.
McKenzie et al. (2005) noted that the quality of data on pup production across the range of
Australian sea lions was typically poor. Poor data are largely due to the species’ protracted
breeding season meaning that by the end of the pupping period, some pups may have died,
dispersed or moulted (and may go unrecognised). Because of this, researchers have tried to
estimate the maximum numbers of pups present from single counts, timed when maximum
pup numbers are expected in the colony, or from multiple point counts made throughout the
breeding season in order to recognise the maximum. Where possible, the accumulated
number of dead pups is added to these estimates. These methods are likely to result in
underestimates of the true number of pups produced, but to what extent is poorly understood
and is and likely to vary among sub-populations. These issues, in conjunction with the
absence of a realistic and representative population model, make it difficult to estimate the
size of the Australian sea lion population accurately.
Further, reliable estimates of pup abundance are available for few ASL colonies, and time-
series data are available for even fewer. Although the methodologies to estimate pup
numbers have advanced in recent years in conjunction with an understanding of the timing of
breeding seasons at certain colonies, the quality of time-series data is typically poor because
early records were based on limited surveys. The apparent high variability in pup numbers
recorded between breeding seasons has also made interpreting trends in population
abundance with any level of confidence difficult.
McKenzie et al. (2005) noted that these observations of major shortfalls in the quality of data
on pup production, population size and trends in the species, are significant because they
place serious limitations on our capacity to adequately manage the species. At its most basic
level, management for the recovery of the Australian sea lion will need to be underpinned by
an ability to detect changes in the status of populations and the species as a whole.
McKenzie et al. (2005), recommended that considerable effort be directed towards improving
survey methodology and precision, including the development of accurate techniques that
take into account the natural variability in pup production and mortality between seasons,
and sightability and availability biases at different locations. This will require the collection of
high-quality data, standardisation of survey techniques between seasons and sites, and
further investigation into the dispersal and mortality of pups.
Introduction 7
This project aims to review survey methods, and develop new protocols appropriate for
populations of different sizes that will enable accurate estimation of pup production with
confidence limits. We also develop a population survey strategy that identifies key colonies
within regions across the range of the species to target for on-going monitoring.
Aims & Objectives of document The aims of this report are to:
1. Review the current estimation methods used to assess the status of Australian sea
lion populations.
2. Develop a new survey protocol that will provide estimates of pup production with
confidence limits, taking account of availability and sightability biases that are
inherent in current approaches.
3. Develop a population survey strategy that identifies key and/or representative
colonies within regions across the range of the species to be targeted for ongoing
monitoring of pup production trends.
Format of the report The report essentially deals with each of the above objectives as separate chapters, followed
by a recommendations chapter. The study is focused on populations in SA, although many of
issues identified and developments in population monitoring methodology are broadly
applicable.
Review of survey methods 8
3 REVIEW OF CURRENT METHODS USED TO ASSESS THE STATUS OF AUSTRALIAN SEA LION POPULATIONS
Estimating abundance Estimates of pup production are used as an index of population abundance for
monitoring the status of pinniped populations, because pups are the only age class
that is easily recognisable and will generally remain ashore when disturbed during or
at the end of the pupping season. Although the exact relationship between pup
production and total population abundance is not known and will vary over time
depending on the age structure of the population, it is generally accepted that
estimates of pup production form a useful index of population size (Berkson and
DeMaster 1985).
McKenzie et al. (2005) provided a comprehensive review of the methods for
estimating pup abundance among pinnipeds, with particular reference to Australian
sea lions. The aims of this section are to briefly summarise those methods reviewed
by McKenzie et al. (2005), and identify and describe the biases and sources of error
associated with these methods for estimating pup production that need to be
addressed in order to achieve accurate estimates.
In most otariid species, 90 % of births occur over a 30-60 day period within each
breeding season with all pups retaining their natal lanugo (dark coat) and remaining
on or near-shore until the end of the breeding season. As a consequence, with the
exception of some pup mortality, all pups are ashore, easily recognisable and
available for survey at the end of the breeding season. In contrast, deriving accurate
estimates of pup production in Australian sea lion subpopulations is challenging,
because the extended length of the breeding season (5-7 months) means there is a
much greater spread in the age and development of pups at the end of the breeding
season, than is typical for other otariid species. For Australian sea lions, this means
that at the end of the pupping season, some pups will be:
1. fully moulted (making them difficult to distinguish from juveniles aged 1 year
and older),
2. spending increasing amounts of time in the water (playing or independently
feeding) or dispersed to nearby haul-outs, or
3. dead (as a consequence of natural mortality).
Review of survey methods 9
The two most common methods used to estimate otariid pup production are direct
counting and mark-recapture. The use of these two methods and problems
associated with their use in estimating pup production at Australian sea lion colonies
are discussed below.
Direct counting Pup production in the Australian sea lion has generally been estimated by the direct
counting of pups at colonies. Historically, direct counting of pups ranged from ad-hoc,
single counts, which might have occurred at any stage during the breeding or non-
breeding period for a subpopulation, to more rigorous multiple-survey approaches
within breeding seasons, that aimed to derive a count around the peak in pup
numbers and account for the mortality of pups born up until the survey period.
Many of the earlier pup production estimates consisted of only one count, with little
knowledge of the timing of breeding seasons for different subpopulations. Many of
these occurred as one-off surveys. The protocol for estimating pup abundance by
direct counting was primarily developed by Dr Peter Shaughnessy throughout the
mid 1990s, and involved a series of direct ground counts of both live and dead pups
throughout the pupping season (Shaughnessy and Dennis 2000). Multiple surveys
were introduced in order to 1), determine when the first pups for the season were
born so that timing of the peak of pup production could be estimated (about 5 months
after the beginning of pupping), 2) improve estimates of pup numbers at the peak of
pup production, 3) obtain better estimates of accumulative pup mortality and 4) to
acquire information on the schedule of breeding. The last has been important in
improving the planning of surveys for a number of breeding sites.
During surveys, pups are typically classified as brown, moulted, unclassed or dead.
Brown pups are in their natal pelage or are in the process of moulting it; moulted
pups have completely moulted their natal pelage which is completed at about 5
months of age (Shaughnessy et al. 2005). If pups are not classified in these two
categories, they are referred to as ‘unclassed’ pups. Another useful category is that
of pups aged less than a month; these are recognised by their small size, loose skin
folds, and a relative lack of coordination. In addition, many pups less than 3 weeks of
age have a relatively pale crown and dark mask across their face (Ling 1992). Pups
aged less than a month seen at the beginning of a pupping season are useful for
estimating when pupping began. They are also useful when surveys are made at
monthly intervals after the maximum pup count in a colony, because they represent
Review of survey methods 10
pups born since the previous survey and can be added to the maximum count to
obtain a more accurate estimate of pup production. This has been done for several
surveys at Dangerous Reef (eg. Shaughnessy 2004).
Classifying some young Australian sea lions can be difficult because moulted pups
can be confused with small juveniles of similar size born in the previous pupping
season, which are then older than 1 year. Small juveniles can be recognised by their
cranial development, particularly their slightly longer noses. When pups moult their
lanugo coat, they replace it with a silver grey and cream pelage. When juveniles that
were born in the previous pupping season moult, their newly emerging silver grey
coat shows through their aged, ginger coloured outer hair, which gives them a
different coloration from that of pups.
Dead pups are removed from the colony or marked to ensure they are not recounted
during subsequent surveys. The number of cumulative dead pups is added to the
number of live pups recorded for a given count, to estimate the number of pups born
to that date. The maximum number of pups (live and cumulative dead) is then taken
as the index of abundance for the pupping season (Shaughnessy et al. 2006). New-
born pups sighted on subsequent surveys may be added to this total to improve
estimates.
Due to the remoteness, costs and other logistical difficulties of visiting many
Australian sea lion colonies, monthly ground counts throughout the pupping season
are not feasible at all sites. Where the timing of breeding is known for a particular
site, 3-4 surveys are typically undertaken per breeding season. This counting method
has been applied in South Australia since 2001 at a number of colonies including:
West Waldegrave, Jones, Nicolas Baudin and Olive Islands, and islands in the Nuyts
Archipelago (Franklin Islands, Breakwater Island, West Island, Shaughnessy et al.
2005, Shaughnessy un-published data).
Mark-recapture Direct counts are known to underestimate total pup production, because they may fail
to count pups that are hidden from view (sightability bias) or absent from the colony
(availability bias) at the time of the survey. The influence of these factors on
estimates of pup numbers can be reduced to some degree by undertaking a mark-
recapture procedure.
Review of survey methods 11
Petersen estimate mark-recapture methods have been used to estimate pup
production at fur seal colonies in Australia since 1988 (Shaughnessy et al. 1995a,
Shaughnessy et al. 2002, Kirkwood et al. 2005), but have only been applied recently
to estimating pup production in the Australian sea lion (McIntosh et al. 2006a). Mark-
recapture involves randomly marking a known number of pups in a population on one
occasion, then recording the proportion of marked animals ‘recaptured’ (resighted) on
one or a number of subsequent occasions. Confidence intervals can then be
calculated for mean pup abundance estimates. Because pups may die between
marking and recapture, dead pups are counted and removed during recapture
sessions and classed as marked or unmarked. The cumulative total of dead pups
recorded and marked on previous occasions is then added to the mark-recapture
estimate to provide an estimate of pup abundance and mortality.
Mark-recapture estimates of pup numbers ( ) are calculated using a variation of the
Petersen method (formula attributed to D.G. Chapman by Seber 1982), with the
formula
N̂
1)1(
)1)(1(ˆ −+
++=
mnMN ,
where M is the number of marked pups at risk of being sampled during recapture
operations, n is the number of pups examined in the recapture sample, and m is the
number of marked pups in the recapture sample.
The variance of this estimate is calculated as
)2()1(
))()(1)(1()ˆvar( 2 ++−−++
=mm
mnmMnMN ,
and the 95 % confidence limits calculated from
. )var*96.1(ˆ 5.0±N
Where several mark-recapture estimates ( ) are undertaken (one from each
recapture session), they are combined by taking the mean ( ) using formulae from
White and Garrott (1990, pp. 257 & 268):
jN̂
N̂
Review of survey methods 12
∑=
=q
j
j
qN
N1
ˆˆ
where q is the number of estimates for the colony (i.e., the number of recapture
sessions). The variance of this estimate is calculated from
)ˆ(var1)ˆvar(1
2 j
q
jN
qN ∑
=
=
and its standard deviation as
5.0)ˆvar(Nsd =
The Petersen estimates yields an accurate result as long as a number of conditions
are met. These include:
1. the probability of capturing an individual is the same for all individuals in the
population,
2. no animal is born or immigrates into the study area between marking and
recapturing,
3. marked and un-marked individuals die or leave the area at the same rate, and
4. no marks are lost (Caughley 1977).
Mark-recapture methods have been used in conjunction with direct counting methods
at four Australian sea lion sites in SA: Seal Bay, Dangerous Reef and North Page
and South Page Islands. Table 3.1 compares estimates of live pups based on direct
counts and mark-recapture methods at these locations, and indicates that at all sites,
direct counting yielded lower estimates of the number of live pups present compared
to mark-recapture methods. The degree of error varies between sites, with the
highest (mean = 1.87) recorded at Seal Bay, and moderate levels recorded for
Dangerous Reef (mean = 1.22) and The Pages (mean 1.32) (Table 3.1). The extent
of under-counting is largely due to the level of cover available to pups to evade
detection. At Seal Bay there is extensive cover from coastal shrubs and vegetation
for pups to avoid detection, with less cover available at Dangerous Reef and The
Pages Islands. Importantly, available data suggest that between sites, the extent of
undercounting may vary somewhat (ie. is not constant). Within site comparisons are
only possible at Dangerous Reef, where over three seasons, the 95% CL of the
‘difference’ overlapped, indicating the comparison was similar between seasons
(Table 3.1).
Review of survey methods 13
Table 3.1. Comparisons of direct counts and mark-recapture estimates of live pups at four Australian sea lion colonies. “Comparison” indicates the ratio of the mark-recapture estimate to the direct count. Colony Date Direct count MR Est. Comparison Source Seal Bay Jun-03 77 144 1.87 McIntosh et al. 2006a Dangerous Reef Jul-99 240 285 1.19 Shaughnessy and Dennis 1999 Jan-04 333 423 1.27 Shaughnessy 2004 Jul-05 272 326 1.20 Shaughnessy 2005a North Page Oct-05 152 177 1.16 Shaughnessy 2005b South Page Oct-05 148 219 1.48 Shaughnessy 2005b
Problems with current methods As indicated above, there are two main sources of error associated with the counting
of Australian sea lion pups at breeding colonies. The first relates to underestimating
pup numbers because some pups may be hidden from view, and we term this
sightability bias. Sightability bias is caused by 1) the level of cover afforded to pups in
colonies (eg. bushes, rock crevices, caves) that may enable them to evade detection
during ground surveys, or 2) pups becoming more aquatic and increasingly absent
from colonies as they age and develop foraging skills (ie. undertaking foraging trips).
As indicated above, sightability bias can be very significant and vary considerably
among sites. At Seal Bay, almost 50 % of pups can be hidden from view during direct
counting procedures (McIntosh et al. 2006a).
Fowler et al. (2006) instrumented 6 and 15-month old Australian sea lion pups with
dive recorders and identified increasing periods of time spent at sea as animals
aged, 10.3 % at 6 months and 39.9 % at 15 months of age. In October 2005, pups at
Lounds Island (Nuyts Archipelago) that were approximately 6 months old, were
observed undertaking foraging trips to sea, independent of their mothers (S.
Goldsworthy, unpublished obs.). Gales et al. (1992) reported pups beginning to leave
breeding colonies possibly on foraging trips from 4.5 months of age at colonies on
the west coast of Western Australia. In addition, some brown pups tagged at Seal
Bay have been observed at the seal Slide (approximately 20 km away, McIntosh
2006b). Clearly, such behaviour will lead to underestimates of the number of pups
ashore during ground surveys, especially late in the pupping season.
Review of survey methods 14
The second source of error is due to the absence of pups at the time of surveys, that
we term availability bias. There are two main sources of availability bias. 1)
unaccounted mortality - pups that die during the breeding season whose bodies are
not recovered due to disappearance through high tides, storms, natural
decomposition or scavenging by predators; and 2) dispersal - pups dispersing to
nearby haul-outs before the end of the breeding season.
Unaccounted mortality throughout the breeding season may be a significant source
of under-estimation of pup numbers, even when colonies are visited at monthly
intervals for pup surveys. The extent to which unaccounted mortality impacts on the
accuracy of mortality estimates is likely to vary among sites, with sites where pupping
is focused around near-shore areas (exposed to high tides or storms) likely to have
higher rates of unaccounted mortality than sites where pupping is focused in more
protected areas.
Dispersal could also be an important source of bias in estimating pup production at
some breeding sites. There have been numerous observations that dispersal to
nearby haul-outs occurs in some pups before the pupping season is completed. Such
movements have been noted from the Seal Bay colony on Kangaroo Island, based
on movements of tagged (or otherwise marked) pups to the Seal Slide in 1975 (Ling
and Walker 1976), in August and October of 2003 (D. Dowie in Dennis 2005), and in
June 2006 (two pups, see Chapter 5), and to Black Point in July and August 2002 (J.
McKenzie, personal communication), in March 2005 (McIntosh et al. 2006b), and to
Cape Bouger in 1978 (Ling and Walker 1979). Movements of pups from Dangerous
Reef to English Island were suspected in August and September 2002 (Shaughnessy
et al. 2005a) and in July 2005, some pups seen at English Island had been marked
at Dangerous Reef as part of the mark-recapture procedure (D. Hamer, pers.
comm.). All these observations suggest that even young pups have the capacity to
travel to nearby haul-outs and spend periods at sea or away from the natal colony,
although the extent of this behaviour is poorly understood.
In addition to sightability and availability biases, a poor understanding of the timing of
breeding can significantly affect the appropriate timing of surveys, and therefore the
quality of the results obtained. Difficulties in predicting the timing of breeding seasons
at various subpopulations occur because data for some sites are poor (see Chapter
6), because of the asynchrony in the timing of breeding among sites (Gales et al.
1994), and because the interval between breeding seasons can also be variable (14–
Review of survey methods 15
20 months, see Shaughnessy et al. 2006). Unless the pattern of breeding seasons is
known for a colony, it is difficult to ensure pup counts (or any other method of
estimation) are conducted at the appropriate stage of breeding. This highlights the
need to determine the beginning of a pupping season at a colony where abundance
is to be estimated.
Figure 3.1 demonstrates how some of these sightability and availability biases may
affect the numbers of live pups counted in a subpopulation throughout different
stages of the breeding season, relative to the actual distribution of births. It shows
that the numbers of live pups present in the colony peaks prior to the end of the
pupping season (because of mortality and dispersal). It also gives an example of a
survey that includes live pups and accumulated mortalities, factoring in a 15 %
sightability bias. In this example, the best case scenario of surveying the colony at
the peak (when all live and dead pups in the colony can be counted) results in a pup
count that is 71% of the total births. With a 15% total sightability/availability bias, the
pup count reduces to 64 % of the pup production (Figure 3.1). Importantly, the
difference (degree of error) between total births and pup counts (with or without
accumulative deaths) increases as the season progresses. Even if a mark-recapture
estimate is made at the peak of pup numbers and accumulative mortality is included,
it will still underestimate total pup production.
0
10
20
30
40
50
60
70
80
90
0 50 100 150 200
Days
Num
ber
of p
ups
Cumulative births
Cumulative dead
Cumulative dispersal
Live pups
Max Census (live+cumulativedead)
Example census (includes 15%sightability bias)
Figure 3.1. Hypothetical example of the complicating factors associated with determining accurate estimates of pup production in Australian sea lion populations. In this example, the breeding season occurs over a 6-month period where a total of 85 pups are born. As a consequence of mortality and dispersal of pups, the number of pups that can be counted at any given time is less than the total number of births, and the magnitude of the error increases as the breeding season progresses.
Review of survey methods 16
Direct counting methods are most susceptible to errors associated with sightability
and availability biases, because they only survey a subsample of live and dead pups
present in the colony at any given time. If all the assumptions of mark-recapture
methods are met during estimation procedures, then this approach should be able to
account for any potential sightability biases, as long as the probability of marked and
unmarked pups evading detection (hidden from view on land or at sea) are the same.
However, neither method is adequate to account for potential errors associated with
availability biases as a consequence of longer term dispersal or unaccounted
mortality.
Colony size The numbers of pups born and the nature of the habitat of breeding sites can affect
the suitability of different survey methods and the extent or influence of different
sources of bias in estimating pup abundance. At smaller pinniped colonies, where
densities of pups are low and pups are often widely dispersed, there may be
insufficient mixing of marked and unmarked pups to satisfy the requirements of the
mark-recapture protocol. We have arbitrarily categorised large colonies as those with
counts of more than 40 pups. Pups in colonies with fewer than 40 pups (small
colonies) tend to be dispersed and we suspect mixing in them is minimal (especially
in young pups).
The biases inherent in direct counting methods for estimating pup abundance in
small colonies are the same as those for large colonies, with the basic constraint that
all pups are unlikely to be available or visible during any one survey. Hence direct
counting methods are also likely to underestimate pup production in small colonies.
A new approach the surveying Australian sea lions The goal here is to develop new approaches to surveying pup production that takes
account of sightability and availability biases, and to provide estimates of total pup
production with confidence limits. The major limitations of current methods are that
they only enable estimation of the numbers of pups ashore at any given time, and as
indicated above, this does not take into account any pups missing due to
unaccounted mortality or dispersal. The only way to estimate the numbers of pups,
which are not available for surveying is to obtain data on the recapture and survival
probabilities of pups, using methods such as capture-mark-recapture (CMR)
(Cormack 1964, Jolly 1965, Seber 1970). However, such approaches are only likely
to be suitable for large colonies (>40 pups) with sufficient mixing. The next chapter
Review of survey methods 17
(Chapter 4) details the trialling and results of this method for estimating pup
production in large colonies using the Cormack-Jolly-Seber (CJS) method. Chapter 5
presents results of a trial for a new approach to estimating pup production in small
colonies, that we term cumulative mark and capture (CMC).
Developing survey protocols for large colonies 18
4 DEVELOPMENT OF SURVEY PROTOCOLS FOR LARGE AUSTRALIAN SEA LION COLONIES
Introduction As indicated in Chapter 3, the aim of this section of the study is to explore and
develop new methods for estimating Australian sea lion pup production in large
colonies (>40 pups), that address the issues of availability and sightability biases by
calculating re-sight and survival probabilities of pups in conjunction with Cormack-
Jolly-Seber (CJS) capture-mark-recapture (CMR) models.
Methods
Field site Field-work was undertaken at Olive Island (32º 43’ 19” S, 133º 58’ 05’ E), accessed
by charter vessel from the township of Streaky Bay, Eyre Peninsula South Australia,
(see Figure 6.1), between 7 March and 15 July 2006. Six visits (CMR ‘sessions’)
were made to the island (Table 4.1). On the first visit (7 March), the pupping season
was well underway, and based on the age structure of pups may have commenced
as early as late December 2005 (see Figure 6.3). Pups were tagged in the trailing
edge of each fore-flipper with individually numbered plastic tags (Dalton® Size 1
Supertags). Small pups were numbered with bleach (mixture of bleach powder and
peroxide solution) applied to the fur on the shoulder or the flank. During each field trip
to Olive Island, individual re-sight records were collected for marked individuals with
the aid of binoculars. A record of dead pups was obtained by placing rocks on top of
carcases to avoid repeat counting. Records of the total number of marked, unmarked
and newly recorded dead pups were noted on each field trip, and mark-recapture
was undertaken to provide information on survival, site fidelity and population closure
(see below). The number of re-sights of individually marked pups on the days prior to
recapture surveys were used as the number of ‘marked’ individuals in subsequent
recapture events using the Petersen estimate procedure (see Chapter 3 and below).
Survival We used Cormack-Jolly-Seber (CJS) capture-mark-recapture (CMR) models
(Cormack 1964, Jolly 1965, Seber 1970) implemented in program MARK (White &
Burnham 1999) to model the survival and recapture (re-sighting) probability (p) of
pups. Because our surveys identified previously tagged pups that had died during the
interval between capture and re-sighting sessions, we employed the Burnham (1993)
Developing survey protocols for large colonies 19
joint live-dead modification to the CJS model. The classic CJS model only allows for
the estimation of apparent survival (φ) given that it is confounded by permanent
emigration (Burnham 1993). By including information on the confirmed mortality of
known individuals, the processes of permanent emigration and true mortality can be
separated. As such, the joint live-dead CJS model estimates true survival (S), the
probability of identifying and reporting a dead (marked) individual (r), live capture
probability (p) and the fidelity (F) probability (i.e., the probability that a pup remains
on the study site for the duration of the CMR program and is available for live
recapture given that it is alive). As such, the probability of permanent emigration is 1
– F (Burnham 1993).
Because we determined that F was approximately equal to 1 (i.e., no permanent
emigration, see Results), we used the simpler CJS model with live captures only to
estimate true survival (φ is equivalent to S when F = 1). Models were compared using
an information-theoretic measure of model parsimony, Akaike’s Information Criterion
(AICc) (Akaike 1973, Burnham & Anderson 2002) and goodness-of-fit was assessed
using the simulation procedures provided in program MARK (White & Burnham
1999). A second model was constructed to incorporate the effects of sex on φ and p.
In all, 16 models were considered.
Pup production We used various mark-recapture models assuming either demographic closure (no
net immigration or emigration, including births or deaths) or those assuming
demographic openness to estimate the size of the pup population ( ). We first used
the program CAPTURE (Otis et al. 1978) to construct closed population models
based on variants of the Petersen abundance estimator. Program CAPTURE allows
the user to test the null hypothesis that the live capture probability of an individual i
(p
N̂
i) is invariant (population closure), with the alternative hypothesis being that for
some individuals p = 0 at the beginning or end of the study period (Pollock et al.
1974, Otis et al. 1978, Cerchio 1998).
The assumption of population closure was violated due to continued pup births
during the beginning of the sampling interval and pup mortality (see Results). To
assist in the estimation of the number of new pups added to the surveyed population
over the investigation interval, we applied the Pradel recruitment model (Pradel 1996)
to estimate the parameter λ , which equates to the probability that an individual
Developing survey protocols for large colonies 20
observed alive and available to be captured at time i was also alive and available for
capture at time i – 1 (‘seniority’). From 1+λ it is possible to estimate λi, the growth
rate of the population expected between time i and i + 1 ( 1ˆˆˆ
+= iii λφλ ; Pradel 1996).
Thus, an estimate of λ gives some idea of the number of pups born during an
interval between capture sessions. In addition, differences between the Petersen
estimates of pup abundance were used to estimate the number of births between
recapture sessions.
As a further verification of the fidelity assumption used to estimate S and to coalesce
the closed and open-population estimators, we incorporated the closed-capture data
(Petersen estimators) with the longer intervals between sightings into Pollock’s
robust design in program MARK (Kendall et al. 1995). This design analysis uses data
from sessions within a period using closed-capture models and Jolly-Seber methods
for data from multiple periods to allow for animals to be ‘unavailable’ for capture at a
given time (i.e., either a temporary emigrant or immigrant). The model estimates the
following parameters (1) tφ = the probability that a member of the population in
period t survives and is still a member of the population in period t + 1; (2) γ ′′ and γ ′
= the probability that a member of the population in period t is unavailable for
detection (i.e., outside the study area) given that it was available or unavailable,
respectively; (3) and = the probability that an animal that is available for
detection during period t and has not or has, respectively, previously been detected
during period t, is detected in sample s. The total population size estimator, N
stp , stc ,
t, is not
included specifically; instead, an estimate of Nt is derived from the total number of
individuals detected during period t (nt) and the estimated pooled detection
probability for period i, , where = the number of sessions in
period t. Total estimated abundance is then:
∏=
−−=ts
ssti pp
1
11 )(ˆ ,*
ts
*ˆˆ
t
tt pnN = . The closed-capture models
implemented using program CAPTURE indicated full heterogeneity in capture
probabilities, so we incorporated the full heterogeneity estimator option in MARK
(Huggins 1989, 1991).
With new pups being added, some pups dying, and an unknown number of pups
born prior to the first sampling session, we combined the parameters estimated in the
survival and recruitment models in open-population Jolly-Seber models (Schwarz &
Developing survey protocols for large colonies 21
Arnason 1996) using the POPAN option in the program MARK (White & Burnham
1999) to estimate total pup production. Our logic proceeded as follows:
1) An examination of the Pradel recruitment model indicated that λ was larger than 1
(i.e., the population was ‘growing’ due to births) between sessions 1 and 2 only (see
Results). As such, we assumed that an open-population estimate of N pups between
sessions 1 and 6 would provide a good estimate of peak pup numbers, less pups
born between session 1 to 2, and pups dying between sessions 1 to 6. In the POPAN
model, t capture occasions were modelled to provide t – 1 estimates of φ (apparent
survival), t estimates of p, t – 1 estimates of β (probability of entry into the population
per occasion), and N (the super-population size). We fitted all models using the logit
link function for and , the identity link function for , and the multinomial logit
link function to constrain the set of parameters to ≤ 1 (White & Burnham 1999).
We used AIC
φ̂ p̂ N̂
β̂ β̂
c to compare models, but we did not fully develop the φ, p and β
parameters within the mark-capture framework given our focus on estimating N.
2) To the super-population N (the larger population of individuals that are associated
with a particular area, in this case Olive Island) estimated for sessions 1 to 6, we
applied the mortality rate of 1 – 0.971 (in reality, 1 – the upper and lower confidence
intervals of – see Results) to estimate the number of pups that would have died
during each interval. This number, added the confidence interval to N
φ̂
1-6 and was
taken as the total number of pups to have been present during that interval.
3) To estimate the number of pups born between intervals 1 and 2 ( - ), we
applied a simple Petersen estimator:
2N̂ 1N̂
( )( ) 1
111ˆ
1 −+
++=
mnMN ,
where M = initial number of marked animals, n = subsequent number of individuals
from the same population and m = the number of n animals containing marks in
mark-resight surveys (formulae for variance given in Chapter 3). Differences between
the and , and their confidence limits, were taken as the range of pups born
between sessions 1 and 2 (mean: ; upper range:
; lower: 0). This was also used to estimate
2N̂ 1N̂
12ˆˆ NN −
)(ˆ)(ˆ12 lowerCLNupperCLN −
Developing survey protocols for large colonies 22
independently λ1 and was compared to λ estimated using the Pradel recruitment
model.
(4) To estimate the number of pups that would have died between the onset of
pupping (assumed to be 1 January 2006) and session 1, we assumed again a
constant mortality rate of 1 – 0.971 and applied it to determined above. The
range of expected dead pups was compared to the number of dead (unmarked) pups
found during surveys in session 1.
1N̂
Test for equal catchability The key assumption of mark-recapture studies is that the probability of capture is the
same for all individuals in the population. This was tested within the tagged
population by examining the number of times individual pups were resighted within
each capture session. We used the Leslie’s test for equal catchability, following
methods detailed in Caughley (1977), and for each of the six capture sessions,
examined the number of times known-to-be-alive individuals were resighted. We
used the Leslie’s test in favour of the zero truncated Poisson test because it enabled
us to use data on zero recaptures, animals known to be alive from subsequent
recapture session, but not sighted. This could be achieved for all but the final
recapture session. The assumption in Leslie’s test is that if catchability is constant
the recapture frequencies will form a binomial distribution. This assumption can be
tested as a Chi-square with ( ) 1−∑ f degrees of freedom, by comparing the
observed variance to the expected binomial variance, where
( )
( )2
2
22
2
fn
ffi
ffifi
∑∑
−∑∑
∑∑
−∑=χ ,
and n is the number of individually tagged pups resighted during each recapture, i is
the number of times individual pups were resighted during recapture sessions and f
is the number of individual resighted i times (Caughley 1977).
Results Marking and absolute counts A total of 142 pups were marked (tagged 136, bleached 6) over the first three visits
(7 March, 23 March and 12 April 2006) to Olive Island (Table 4.1). On each visit, the
maximum number of unmarked pups counted during surveys of the colony and
Developing survey protocols for large colonies 23
cumulative mortalities (unmarked and marked) were recorded (Table 4.1). This
enabled minimum estimates to be calculated for each visit (session) based on: total
counted (live), maximum count (total live count plus cumulative dead), and minimum
pups (cumulative marked + dead [unmarked] + maximum unmarked counted) (Table
4.2, Figure 4.1). Minimum estimates of pups based on these approaches were: 126,
150 and 183, respectively (Table 4.1). Counts based on total live pups, and live pups
plus cumulative dead pups peaked in session 3 (12 April), and declined thereafter
(Table 4.1, Figure 4.1).
Petersen estimates Results from Petersen estimates of pup abundance undertaken over the first five
sessions at Olive Island are presented in Table 4.2. Estimates suggest that the
numbers of pups at Olive Island increased steadily from about 115 (95% CL, 95-134)
during the first session (7 March) to 165 (159-170) during the penultimate visit
(session 5) (Figure 4.1). Adding the minimum accumulated mortalities to these
estimates suggests that the mean minimum number of pups in the population
between session 1 to 5 ranged from 123 to 199 (Table 4.1, Figure 4.1).
Test for equal catchability Results from Leslie’s test of equal catchability are presented in Table 4.3. Results
from all recapture sessions were non-significant, indicating that the assumption that
the distribution of recaptures is binomial, and therefore that catchability is constant is
supported.
Survival A total of six ‘capture’ sessions with 142 marked individuals (588 total resightings and
five ‘marked’ dead returns), provided estimates of survival (S), live capture probability
(p), dead return probability (r) and fidelity (F). The F parameter in the Burnham
(1993) joint live-dead model indicated strong support (combined AICc model weights
for the two top models with F(.) = 0.93) for a constant F over the course of the
program equal to 1 (i.e., no permanent emigration). The best supported models
according to AICc demonstrated time-variant r, but most were inestimable due to the
low number of returns of dead marked pups(5). However, when dead recoveries
were made (in sessions 2 to 5), the estimable parameters indicated r ~ 1.0 for
intervals 2 and 4 (r3 was 0.18 ± 0.16, indicating some dead pups may have been
missed between sessions 2 and 3).
Developing survey protocols for large colonies 24
We therefore chose to model the simpler CJS live-captures only model estimating
apparent survival (φ) and live capture probability (p) because when F = 1, φ = S. The
best supported model had strong support (AICc weight [w] = 0.711) for a time-
invariant monthly (30-day) = 0.971 ± 0.010 and time-variant p (range: 0.680 –
0.943; Table 4.4). There was, however, moderate evidence for time-variant survival
(w = 0.287; Table 4.4), but the range (0.956 – 1.000 = 30-day survival probability)
was effectively modelled by the error associated with the time-variant φ in the first
model. The model set including the combined effects of pup sex and time on φ and p
indicated only weak support (combined w of models with a sex effect on φ = 0.137)
for differences in survival between males and females (time-invariant = 0.973 ±
0.009, = 0.967 ± 0.008 over 30 days).
φ̂
malφ̂
femφ̂
The Pollock robust models considered were unable to estimate all parameters, so
they were adjusted accordingly for the number of estimable parameters (e.g.,
estimated Nt could not be resolved). The best supported model did, however, have
γ ′′ = γ ′ = 0 (w = 0.73), indicating strong support for a closed population with no
temporary immigration or emigration. There was also weak support (w = 0.12) for
time-invariant γ ′′ and γ ′ (0.06 and 0.07, respectively), suggesting that there was a
possibility that some marked pups were not always available for recapture at each
recapture phase. However, given the low probabilities and weak model support, the
assumption of fidelity F = 1 as estimated in the Pradel model appears to be robust.
Additionally, the most parsimonious models accounting for over 96 % of the model
weights estimated a time-invariant survival probability of ~ 0.97, corresponding
exactly with the simpler CJS live-capture model results.
Developing survey protocols for large colonies 25
Table 4.1. Summary of details of Australian sea lion pup marking, counts, recovered mortalities and various direct counting abundance and Petersen estimates during six visits (sessions) to Olive Island between March and July 2006.
Session 1 2 3 4 5 6
Date 7 Mar 24 Mar 13-Apr 17 May 19 Jun 15 Jul
Cumulative marked 38 84 142 142 142 142
Maximum unmarked counted 39 26 3 6 12 7
Maximum count (live) 86 57 126 84 70 64
Cumulative dead (unmarked) 8 14 22 23 29 29
Cumulative dead (marked) 0 1 2 3 5 5
Total accumulative dead 8 15 24 26 34 34 Maximum count (live) + cumulative dead 94 72 150 110 104 98 Cumulative marked + dead (unmarked) + max unmarked 85 124 167 171 183 178
Petersen Estimate (live) 115 137 145 149 163 Petersen Estimate Lower – Upper CL (No. recapture estimates)
95-134 (3)
123-150 (4)
143-147 (6)
145-152 (3)
157-169 (6)
Petersen Estimate (live) + cumulative dead Lower – Upper CL
123 (103-142)
152 (138-165)
169 (167-171)
175 (171-178)
197 (191-203
0
50
100
150
200
250
28-Feb 23-Mar 16-Apr 9-May 1-Jun 24-Jun 17-Jul
Max count (live)Cumulative deadMax count + cumulative deadCumulative tagged + dead (clear) + max clearPetersen estimate Petersen estimate + dead
Num
ber o
f pup
s
Date
Figure 4.1. Trends in accumulative dead Australian sea lion pups and various pup abundance estimates between March and July 2006, at Olive Island.
Developing survey protocols for large colonies 26
Table 4.2. Details of Petersen mark-recapture procedures undertaken at Olive Island between March and June 2006. M = number of marked pups in the population, n = the total number of pups sampled and m = the number of marked pups in each recapture sample. N = the estimated pup population size, sd = standard deviation and V = variance. % = the percentage of marked pups in each sample, CV = the coefficient of variance, and Nup and Nlo are the upper and lower 95%confidence limits of each estimate, respectively.
Date Recapture Marked Examined M-R
No. M n m N sd V % CV Nlo Nup
Session 1
7 Mar 1 38 64 25 97 8 70 39% 7 Mar 2 38 53 14 139 23 548 26%
7 Mar 3 38 44 15 109 16 269 34%
Mean 115 9.9 33% 8.6% 95 134
Session 2
24 Mar 1 83 39 22 145 17 274 56% 24 Mar 2 83 57 40 118 7 50 70%
24 Mar 3 83 50 24 170 20 404 48%
24 Mar 4 83 51 37 114 7 50 73%
Mean 137 7.0 64% 5.1% 123 150
Session 3
13-Apr 1 141 75 72 147 2 6 96% 13-Apr 2 140 90 87 145 2 3 97%
13-Apr 3 140 70 67 146 3 7 96%
13-Apr 4 140 86 84 143 1 2 98%
13-Apr 5 140 88 86 143 1 2 98%
13-Apr 6 140 68 65 146 3 8 96%
Mean 145 0.9 97% 0.6% 143 147
Session 4
17 May 1 139 83 79 146 2 5 95% 17 May 2 139 78 72 151 3 11 92%
17 May 3 139 82 76 150 3 9 93%
Mean 149 1.7 93% 1.1% 145 152
Session 5
19 Jun 1 137 58 46 172 9 84 79% 19 Jun 2 137 69 57 166 7 47 83%
19 Jun 3 137 66 58 156 5 28 88%
19 Jun 4 137 58 48 165 8 60 83%
19 Jun 5 137 58 49 162 7 51 84%
19 Jun 6 137 49 42 159 8 56 86%
Mean 163 3.0 84% 1.8% 157 169
Developing survey protocols for large colonies 27
Table 4.3 Leslie’s test for equal catchability across each recapture session at Olive Island. n is the number of individually tagged pups resighted during each recapture, i is the number of times individual pups were resighted during recapture session and f is the number of
individuals resighted i times. Chi-squared ( ) and degrees of freedom (df) values are also given. Non-significant P values indicate equal catchability.
2χ
Session No.
Recapture No. n n2 i f fi fi2
2χ df P 1 1 25 625 0 11 0 0 2 14 196 1 9 9 9 3 13 169 2 11 22 44 3 7 21 63
∑ 52 990 38 52 116 0.045 37 >0.05
2 1 16 256 0 38 0 0 2 33 1089 1 13 13 13 3 20 400 2 16 32 64 4 29 841 3 15 45 135 4 2 8 32
∑ 98 2586 84 98 244 0.050 83 >0.05
3 1 52 2704 0 23 0 0 2 67 4489 1 17 17 17 3 49 2401 2 20 40 80 4 68 4624 3 22 66 198 5 63 3969 4 19 76 304 6 44 1936 5 18 90 450 6 9 54 324
∑ 343 20123 128 343 1373 0.023 127 >0.05
4 1 53 2809 0 7 0 0 2 46 2116 1 23 23 23 3 48 2304 2 29 58 116 3 22 66 198
∑ 147 7229 81 147 337 0.010 80 >0.05
5 1 26 676 1 19 19 19 2 31 961 2 16 32 64 3 32 1024 3 11 33 99 4 32 1024 4 13 52 208 5 28 784 5 5 25 125 6 29 841 6 2 12 72
∑ 178 5310 66 173 587 0.025 65 >0.05
Developing survey protocols for large colonies 28
Table 4.4. Model selection parameters for apparent survival (φ) and capture probability (p) based on 142 tagged Australian sea lion pups at Olive Island, South Australia marked and resighted between 7 March and 15 July 2006. Shown are the four models considered (the ‘.’ indicates time invariance, while ‘(t)’ indicates time variance), the difference between Akaike’s Information Criterion scores corrected for small sample sizes and the most parsimonious model (ΔAICc), AICc weight (w) and the number of estimable parameters (k). Model ΔAICc w k
φ(.) p(t) 0.00 0.711 6
φ(t) p(t) 1.81 0.287 9
φ(t) p(.) 11.8 0.002 6
φ(.) p(.) 48.94 < 0.001 2
Pup production The closed population models examined using CAPTURE indicated a violation of the
assumption of closure (z = -7.092; P < 0.0001), and the best supported model
indicated time-based, behavioural and unrecognized sources of heterogeneity in
capture probability (model Mtbh); however, the models could not converge adequately
to provide estimates of N. Given that pups were not migrating permanently from the
population (fidelity, F = 1), disappearance of individuals was due to mortality.
Assuming a constant mortality rate over the 4.13 months elapsed between sessions
1 and 6, approximately 11 % (1 – 0.9714.13) of the pup population should have died
over that interval.
The steps outlined in the Methods to estimate total pup production from the onset of
pupping to the end of the survey period (i.e. accounting for new births and a constant
mortality rate) resulted in the following: the Pradel recruitment model identified strong
evidence for time-variant λ, and when combined with time-variant estimates of φ, λ1
was estimated to range from 1.50 to 2.57, with λ2 approximately equivalent to 1 (i.e.,
population stability, or no further pup births beyond session 2). This justified the
approach we took to estimate of N1-6 using the CJS open-population model. The
range was also confirmed by comparing the Petersen estimates of N
1λ̂
1 and N2 (plus
mortalities) to estimate λ1 independently – this provided a range of λ1 = 1.03 to 1.64.
Thus, was in the range 144 to 160 (61−N 61−N = 149), the number of births between
sessions 1-2 was between 0 and 55 (mean = 20, based on Petersen estimates, or 3
to 86 [mean = 38] based on the Pradel estimates of λ1 and the Petersen estimates of
N1), the number of deaths between sessions 1 and 6 was 11 – 41 (mean = 20), and
Developing survey protocols for large colonies 29
the number of deaths prior to session 1 was 4 – 11 (mean = 7; this latter range
agreed well with the 8 unmarked dead pups found at session 1) (Table 4.5). This
gave a total pup production estimate of between 159 and 267 (N = 199, Table 4.5).
The coefficient of variation for this estimate was 17.2 % assuming a standard error =
0.5 the mean difference between the mean and upper and lower confidence limits
of N .
The above estimates are based on data obtained just from the resight histories, and
do not include data on minimum estimates of mortality in the population.
Furthermore, the wide range in the confidence interval is a result of combining the
uncertainty in survival rate, the uncertainty in the estimation of N and the uncertainty
associated with the number of estimated deaths between intervals. As such,
estimates on the lower range of mortality in the population based on observed
cumulative mortality can be used to improve the precision of the above estimates. A
minimum of 8 pup deaths were recorded prior to session 1, and 26 deaths between
sessions 1 to 6, making 34 in total. This figure is greater than the mean and lower
estimates of mortality in the CJS open model (15 and 27, respectively, Table 4.5).
Substituting the mean and lower estimates of mortality in the CJS model with that
actually observed, adjusts the pup production estimate to between 178 and 256 ( N
= 205, Table 4.5), reducing the coefficient of variation to 15.1%.
However, based on marking, absolute count and Petersen estimates we know that
the minimum estimate of pup production is at least 183 (based on total marked +
unmarked dead and maximum unmarked pups), and potentially 191 based on the
lower CL of Petersen estimates (session 5 plus minimum accumulative dead) (Table
4.1). The latter estimate is close to the adjusted mean of the CJS open model ( N =
205, Table 4.5). Further as the range of mortality based on recovered dead and the
upper CL of the open model CJS is between 34-52 (Table 4.5), we can adjust the CL
of the highest Petersen estimate (session 5) to given an adjusted estimate of 191-
203 ( N = 206, Table 4.5). This estimate takes includes unaccounted for mortality,
and provides a similar estimate to the CJS method.
Developing survey protocols for large colonies 30
Table 4.5. Summary of estimates of Australian sea lion pup production at Olive Island based on a combination of CJS-CMR methods, minimum estimates of mortality and Petersen estimates of pup production. Numbers rounded to integers.
Pup production estimate method Estimate Low Mean High Open CJS N - Session 1-6 144 149 160 Births between Sessions 1-2 0 22 55 Deaths between Session 1-6 11 20 41 Deaths prior to Session 1 4 7 11 N 159 199 267 Adjusted minimum mortality Session 0-6 34 34 52 Adjusted N 178 205 256 Minimum estimate – total marked + dead (unmarked) + unmarked 183 Adjusted Petersen estimate (session 5) with mortality range 34-52 191 206 223
Discussion Direct count methods Results from Olive Island are consistent with previous data that demonstrate how
direct counting under-estimates total abundance (see chapter 3). Petersen estimates
consistently gave greater estimates of abundance averaging 1.8 (SD = 0.6) times
larger than direct counts (similar to results from Seal Bay, Dangerous Reef and The
Pages, see Chapter 3). Based on direct counts, pup numbers peaked at session 3
(13 April) possibly at around four months into the breeding season, although the
actual date when breeding began is unknown. Minimum numbers of pups based on
the cumulative marked plus maximum unmarked and cumulative unmarked dead,
and Petersen estimates suggest that the total number of pups in the population kept
increasing between session 3-4 and 4-5, providing further support that absolute
counts of pups ashore peak well before the end of the pupping season. Based on
these data, the season of births may have continued until between session 4-5 (19
June), 5-6 months from the predicted commencement of pupping (December 2005).
This season of births is consistent with estimates of the duration of the pupping
season at Seal Bay where the season of births typically lasts between 5-7 months,
with 90% of births occurring over a 4.7 month period (SD = 0.7, 19 seasons,
Shaughnessy et al. 2006).
Developing survey protocols for large colonies 31
Marking and Petersen estimates Although the marking of pups was undertaken to provide individual resight histories
for the CJS models, it also provided another means of estimating minimum pup
abundance across 5 of the 6 recapture sessions. Depending on which final estimate
of pup abundance is used, the 142 marked pups represented 1.27 times the
maximum pup count; 0.77 of the maximum marked, dead and unmarked pups; 0.71
of the maximum Petersen estimates; 0.69 of the mean CJS estimate and 0.53 of the
upper confidence limit of the CJS estimate. Although estimates of pup numbers using
the maximum marked, dead and unmarked were 0.69 and 0.82 of the Petersen
estimates (plus dead) in session 1 and 2, the similarity between these approaches
was high in sessions 3, 4 and 5 (0.99, 0.99 and 0.95, respectively), averaging 98 %.
This result confirms that both methods can provide accurate estimates of the total
numbers of pups ashore at any given time; however, both methods are susceptible to
missing unaccounted mortality and emigration. The minimum number of pups based
on maximum marked, dead and unmarked suggests that the pupping season had
finished around session 5 (19 June), as the minimum number had decreased slightly
by the next session (15 July) (Table 4.1).
Petersen estimates made during the first two sessions had relatively high CVs (8.6%
and 5.1%, respectively, Table 4.2), compared to those undertaken in session 3, 4,
and 5 (0.6, 1.2, and 1.8%, respectively). During the first two sessions, lower
proportions of pup estimates were marked (33 and 64%, respectively), and fewer
recapture sessions were undertaken (only 3 and 4, respectively). Both these factors
tend to increase the variance of estimates. Bayesian modelling of Petersen estimates
has suggest that the variance in estimates can be reduced by increasing the number
of recapture sessions to six, but there is little decrease in variance beyond six
recapture sessions (Shaughnessy et al. 1995b). Consistent with the minimum
estimates based on marked, dead and unmarked pups, Petersen estimates indicate
that the number of pups ashore were increasing (new births were occurring) up until
at least session 5 (19 June). This agrees with observations during surveys when new
born pups and mate-guarded (perinatal) females were seen up until the June
(session 5) surveys.
By having individually identifiable marks on pups, we were able to use individual
resights of pups in the 2-3 days immediately prior to Petersen estimate recaptures to
determine the number of live marked pups available for recapture (but without the
need to capture and mark pups again). Potential sources of bias that remain with this
Developing survey protocols for large colonies 32
approach relate to the age distribution of pups that are marked, in that youngest pups
are likely to be under-represented and oldest pups are likely to be over-represented.
Younger pups are more likely to be hidden under rocks and in crevices than older
pups, and hence less likely to be seen during recapture sessions. This could lead to
a sightability bias if there were a difference in this behaviour between very young
marked and very young unmarked pups. Similarly, older pups are more likely to
disperse into the water than younger pups, and this could also lead to a sightability
bias if there were a difference in this behaviour between the older marked and older
unmarked pups. The extent to which these behaviours of younger and older pups
affect the accuracy of the Petersen estimates in assessing pup numbers is unknown,
but could violate the mark-recapture assumption of equal catchability. The
significance of this source of bias could be examined in future, by measuring the
length of pups tagged over a single tagging period, and using length ‘bin’ as a proxy
of the relative age of pups. Difference in the subsequent recapture probabilities of
different length groups could then be tested for. In the current study, we were able to
test the assumption of equal catchability within the tagged part of the pup population,
and found no support for unequal catchability, indicating that the practice of
resighting pups did not affect the probability of subsequent resighting.
The main source of error that may have influenced the Petersen estimates in later
sessions is the difficulty in distinguishing fully moulted pups from those of the
previous seasons pups (i.e., juveniles >12 months of age). Errors in assigning
unmarked pup status to juveniles would increase the number of unmarked pups in
recapture samples, and hence inflate estimates of pup abundance. This type of error
could also inflate the minimum estimates of unmarked pups used to calculate
minimum estimates from the sum of total marked, recovered dead and unmarked
pups seen. However, the maximum number of unmarked pups sighted in sessions 5
and 6 was relatively low (12 and 7, respectively), hence the relative magnitude of
such error is likely to be small.
CJS method The new methodological approach developed in this study provides many insights
into the issues surrounding the estimation of pup production in ASL colonies.
Importantly, it provides a new context from which other methods (direct counting and
Petersen mark-recapture estimation) can be evaluated, and as such has confirmed
some previously identified sources or error or bias associated with the methods for
Developing survey protocols for large colonies 33
estimating pup abundance. It has also enabled for the first time, a means to estimate
overall pup survival, live capture probability and fidelity rates.
Unaccounted mortality was anticipated to be the biggest source of error associated
with direct counting and Petersen estimate methods at Olive Island. The mean
estimate of mortality estimated by the CJS method was 27, which was actually lower
than the minimum based on 34 recovered mortalities. However, the upper CL of
mortality estimated by the CJS approach was 52 pups, 18 (53%) more than those
recovered. Based on the range of methods, mortality rate over the entire study
interval was estimated to be 19.1% (session 3, 24/126) for maximum counted, 18.5%
(session 5, 34/183) for maximum marked, dead and unmarked; 17.1% (session 5,
34/199) for maximum Petersen estimate; 16.6% (34/205) adjusted mean CJS
method and 19.5% (52/267) upper CL CJS method. These values are very similar
and indicate that indirect methods for estimating mortality (direct counts and
Petersen estimates) all fall within the range of estimates using the CJS method (i.e.
16.6 -19.5%).
With respect to immigration, results from the Pollock robust models indicated strong
support for a closed population, suggesting no temporary immigration or emigration
in the pup population during the survey period. Although there was weak support
suggesting that some marked pups may not always have been available for re-
sighting at each session, the assumption of the Pradel model of fidelity F =1
appeared to be robust, providing little support for emigration during the study period.
Immigration of pups from neighbouring colonies is unlikely during the pupping
season, given that the nearest colony (Nicolas Baudin Island) is 36 km away. Nicolas
Baudin and Olive Islands have similar breeding schedules, so if some immigration
were to occur, it would only be possible late in the breeding seasons when pups are
fully moulted. Average distances travelled by 15 and 23 month old pups at Seal Bay
are 16 and 33km, respectively, compared to an average of 57km travelled by adult
females (Fowler 2005). East and West Franklin Reefs (Lilliput and Blefuscu Islands),
are 41 and 42 km from Olive Island, respectively, however, breeding begins there
about four months after Olive Island. As such, by the time the breeding season has
ended at Olive Island, the oldest pups at the East and West Franklin Reefs would be
aged four months and unlikely to be able to travel such distances. Immigration and/or
emigration may be more likely between colonies that are closer together or where
high density colonies are involved.
Developing survey protocols for large colonies 34
The high coefficient of variation in pup production developed through the CJS
procedure was a surprise, especially given the high proportion of pups marked in the
Olive Island population and the high level of re-sighting effort undertaken. It is
important to distinguish that estimates of the super population (N), are contingent on
estimates of survival (i.e., unaccounted mortality), re-sight probabilities and fidelity,
and are not providing a point estimate for any specific time-periods, as is the case
when estimating through Petersen estimates. The latter approach provides much
tighter CVs because it is providing a point estimate of pups present at the time of
recapture. However, as indicated in this study, CVs can be reduced by combining
methodologies, in this case, by adjusting the lower range of mortality estimates using
visual surveys, and adjusting for recruitment (births) using Petersen estimates of pup
abundance at different sessions. Ultimately, factors that include the extended season
of births and the wide range in ages of pups (and hence probabilities of survival
throughout the season) at the end of the breeding season lead to high CVs.
Additional information from further CJS exercises may help refine methods to reduce
CVs to some extent, but wide CVs may be inherent in assessments of ASL pup
production because of the unusual breeding biology of the species.
N̂
The CJS approach to estimating pup production also identified some inconsistencies
among methods. Perhaps the most important in terms of estimating abundance was
the output of the Pradel recruitment model that suggested no additional births
occurred following session 2 (24 March), about three months into the breeding
season. This result was at odds with minimum estimates based on the sum of
maximum marked, recovered dead and unmarked pups, and on Petersen estimates
that both suggested that pupping continued until at least June, and from studies
elsewhere that indicate the pupping season typically extends from between 5-7
months (Shaughnessy et al. 2006). As such, the similarity between the CJS (205,
range 178-267) an adjusted Petersen estimate (206, range191-223) may be
misleading because the former method assumed pup production ended at session 2,
while the latter incorporated additional births up to session 5. Based on differences in
Petersen estimates and minimum estimates of pups based on the sum of maximum
marked, recovered dead and unmarked pups between session between session 2
and 5, the number of additional pups born may have been between 45 and 59 (Table
4.1). At this stage in the development of survey methods, it is unclear which method
is the most appropriate, but the reason for inconsistencies is currently unclear. We
recommend further trials of CJS methods in conjunction with Petersen estimates, and
Developing survey protocols for large colonies 35
examination of additional analytical methods that may help estimate pup production
between survey periods.
Comparison with previous surveys Olive Island was recorded as a breeding colony in November 1977 when 52 pups
were seen (Dennis 2005). Pups were also seen there in April 1979 (49 unclassed,
Ling and Walker 1979) and in November 1990 (27 moulted and one dead, Gales et
al. 1994, Dennis 2005). Based on three ground counts undertaken between February
and July 2003, 121 pups were estimated to have been born (117 pups were seen in
July plus 4 dead in May 2003, Shaughnessy et al. 2005). Ground counts undertaken
in September 2004 and January 2005 estimated pup production as 131 pups
(Shaughnessy 2005). During the 2006 season, the highest ground count was 126 on
13 April with 24 dead recorded to that date (ie. 150 total). Estimates from the latest
breeding season utilising improved survey methods have substantially increased the
estimates of pup production at Olive Island to between 191-267 (this study), 1.5-2.0
times the 2004/05 estimate.
Developing survey protocols for small colonies 36
5 DEVELOPMENT OF A SURVEY PROTOCOL FOR SMALL AUSTRALIAN SEA LION COLONIES
Introduction As indicated in Chapter 3, mark-recapture methods for counting Australian sea lion
pup production are unlikely to be suitable for small (<40 pups) colonies where
densities of pups are low and pups are often widely dispersed causing insufficient
mixing of marked and unmarked pups to satisfy the requirements of mark-recapture
(eg. equal capture probabilities of marked and unmarked pups). However, regardless
of the size of Australian sea lion colonies, the basic problem remains that not all pups
born there are likely to be present or visible during any given survey, so point
estimates will invariably underestimate pup abundance. On the other hand, pup
numbers can be overestimated as a result of immigration from larger colonies during
the breeding season. Such immigration has been recorded on several occasions at,
for example the Seal Slide and at Cape Bouger on Kangaroo Island (reviewed by
Shaughnessy et al. submitted).
To determine the status of eight Australian sea lion sites on Kangaroo Island where
small numbers of pups have been observed (such as the Seal Slide and Cape
Bouger), Shaughnessy et al. (in prep.) used direct counting. They avoided the
possibility of including pups that had immigrated from the large colony at Seal Bay by
disregarding counts made more than 4 months after the pupping season began at
Seal Bay, the most likely source of immigrants. The age of 4 months was chosen
because pups were recorded moving away from colonies on the west coast of
Western Australia from about 4 ½ months of age by Gales et al. (1992). Pups of that
age can be recognised on the basis of their overall size, shape of their head and the
state of moulting of their pelage. This rule was applied because visits to most of the
small sites on Kangaroo Island during pupping seasons were irregular. Application of
the rule may have excluded some pups born at sites late in the pupping season and
hence should result in conservative estimates.
A variant of the direct counting method was used at the Seal Slide in the 2002-03
season. The number of pups was recorded there at monthly intervals from the
beginning of the pupping season by an experienced observer (D. Dowie). At each
visit, he recorded the number of pups estimated to be aged less than one month,
based on their small size, dark pelage and lack of coordination. Each of these was
Developing survey protocols for small colonies 37
assumed to have been born at the site and was added to the number recorded on
previous visits to estimate the number of pups born there in that season.
This chapter details a simple modification to direct counting procedures we term
cumulative mark and count (CMC) that should reduce the influence of sightability and
availability bias on pup abundance estimates for small Australian sea lion colonies.
Methods Trialling of the cumulative mark and count (CMC) method was undertaken at the Seal
Slide (-36°.028 S, 137°.539 E) Australian sea lion colony in the Cape Gantheaume
Conservation Park, south-east Kangaroo Island. The colony was visited on four
occasions during the breeding season (7 February, 22 March, 19 May and 27 May
2006). During each visit, attempts were made to mark a number of pups, by clipping
a small patch of fur on the back using scissors, and inserting microchips (23mm
glass TIRIS RFID microchips) under the skin in the rump. Microchipping was
undertaken to ensure that marked pups born at the Seal Slide were not confused
with those micro-chipped at Seal Bay (~24km away), as dispersing pups from Seal
Bay have been reported at the Seal Slide previously.
The number of marked, unmarked and dead pups sighted were recorded on each
visit to the Seal Slide, and where possible, additional pups marked. Marked pups
were scanned for a microchip with an RFID antenna to ensure they were born at the
Seal Slide and not Seal Bay. Dead pups were covered with rocks to avoid repeat
counting on subsequent surveys. Pup numbers were estimated for each visit from the
numbers of marked pups and accumulated dead pups, plus the number of live
unmarked pups. The last item was estimated in several ways, and the maximum
number was used to estimate number of pups born to date. For the first visit, it was
simply the number of unmarked live pups seen. For the latter surveys it was the
maximum number of unmarked pups seen in one of the previous surveys, less pups
marked since then.
Results A total of 7 pups were marked over the four visits to the colony, one of which was
marked on the final visit. Details on the number of unmarked, marked and dead pups
sighted on each survey are presented in Table 5.1. The minimum number of marked,
dead and unmarked pups present in the population, based on the re-sight and
marking history are also presented. Based on these data, the maximum number of
Developing survey protocols for small colonies 38
pups estimated to have been born in the subpopulations was 10, from surveys
undertaken on 19 and 27 May (Table 5.1).
The colony was also visited in August, when a total of eight pups were sighted, four
marked (one dead) and four unmarked. Two of the marked pups were micro-chipped
pups from Seal Bay, the other two were pups born at the Seal Slide. Given that at
least 25% of the pups seen on this survey were from an adjacent colony, it is
possible that one or more of the unmarked pups seen were also from Seal Bay. As
the pupping season for the Seal Slide (and Seal Bay) was expected to finish in May
2006 (see chapter 6), this survey was undertaken some three months after the end of
breeding, and is clearly confounded by the addition of dispersing pups from Seal
Bay. As such these data were not used in the estimate for pup production at the Seal
Slide.
Table 5.1. Details of pup surveys undertaken at the Australian sea lion colony at the Seal Slide (Kangaroo Island) between February and May 2006. The number of unmarked, marked, dead and total pups seen on each survey is indicated, in additional to the number of new marks applied. The number of marked pups available to be re-sighted at each survey is presented, along with the cumulative number of dead pups recorded. The maximum number of pups at each visit is estimated by summing the number of pups marked, maximum number of unmarked pups and cumulative dead pups. Observers include: Rebecca McIntosh (RM), Mel Berris (MB), Simon Goldsworthy(SG), Rachael Gray (RG), Clarence Kennedy (CK), Andy Lowther (AL), Peter Shaughnessy (PS) and Albert Zepf (AZ).
Date Survey details Pup categories at time of survey Observers
Unmarked pups seen
Marked pups seen
New dead pups
Total counted
New marks applied
Marked pups
available
Cumulative dead
Max Unmarked
1
Minimum pup
estimate
7-Feb 5 0 0 5 1 0 0 5 5 MB, PS
22-Mar 2 1 1 4 0 1 1 4 6 AZ, CK
19-May 8 1 0 9 5 1 1 8 10 RM, CK, AL, RG
27-May 2 4 0 6 1 6 1 3 10 SG, RM, PDS, CK
1Maximum unmarked is the larger of the number of unmarked pups sighted on the survey, or from the larger previous survey, minus the number of new pups marked then.
Discussion The accumulative mark-recapture method developed here for small Australian sea
lion populations has demonstrated that not all pups were present or visible during
any of the surveys at the Seal Slide. Our results indicate that on the second survey, 4
of a minimum of 6 pups were seen, on the third and fourth surveys, 9 and 6 of a
minimum of 10 pups were seen. Based on these results between 10-40% of pups
were not sighted on each survey.
Developing survey protocols for small colonies 39
Our methodology could have been improved by applying more marks during earlier
surveys, but this was not possible due to the presence of mate-guarding males and
aggressive females, and because additional unmarked pups may not have been
available. The level of confidence in the final estimates would increase
commensurately with the proportion of pups marked, because the level of
errors/biases associated with unaccounted mortality/dispersal are reduced as the
number of marked pups increases. Additional surveys would also have helped
provide greater confidence around the estimates derived. Results demonstrate that
there can be a large variance in the number of pups sited on each visit, and that it is
not straightforward to mark more pups (especially earlier in the breeding season)
when they are available. Clearly the purpose of the CMC method is to mark as many
pups as possible in order to reduce the number of unmarked pups, and therefore the
degree of variance in the total estimate. The best means to maximise the number of
marked pups, and reduce the variance in the number of pups re-sighted, will need to
be examined in follow up surveys.
An estimate of 10 pups born at the Seal Slide for the 2006 season compares well
with estimates from the preceding two seasons based on variants of the direct
counting method outlined in the Introduction to this section. For the 2004/05 pupping
season, the estimate was 11 pups, all of which were judged to be less than 4 months
of age when recorded; for the 2002/03 pupping season, the estimate was 9 pups,
each of which was recorded when less than a month of age (Shaughnessy et al.
submitted).
Subpopulation survey strategies 40
6 DEVELOPMENT OF SUBPOPULATION SURVEY STRATEGIES FOR AUSTRALIAN SEA LIONS
Introduction A recent report to Commonwealth DEH detailing the impediments to growth in
Australian sea lion populations (McKenzie et al. 2005) noted that because of the
large number of Australian sea lion breeding sites and their asynchronous breeding
patterns, achieving high quality trend data across all breeding sites over time is
unlikely to be achievable, especially considering the difficulty and expense in
reaching many of the sites. The report recommended that focusing efforts on
obtaining high-quality pup count data from consecutive breeding seasons from a sub-
set of key and/or regionally representative colonies was the best strategy for
obtaining trend data across the range of the species.
In this chapter we develop an approach to identify regionally representative colonies
that are significant (based on size and historic data), logistically accessible and
practical to survey. Where known, the breeding chronology for these representative
sites is detailed to inform the appropriate timing of future surveys, and we suggest
the appropriate monitoring approach for ongoing assessment of these sites.
Our approach was to identify representative Australian sea lion colonies within
regions or metapopulations (a group of spatially separated subpopulations/breeding
colonies or sites of the same species which interact at some level). To achieve this
we first undertook a distance matrix (as a proxy of genetic distance) to identify
regionally discrete metapopulations; then identified a minimum of one large (>40
pups) and one small (<40 pups) breeding site within each metapopulation that were
the most logistically feasible, cost effective, practical and safe to survey. We finally
present (where known) the breeding chronology of breeding sites and develop a
timetable for surveying representative sites into the near future.
Importantly we aim to identify a minimum subset of representative breeding colonies
to form part of an ongoing monitoring program. This program would estimate pup
production at each of these colonies in consecutive breeding seasons in order to
provide information on the status and trends of Australian sea lion metapopulations
Subpopulation survey strategies 41
across their range, over the shortest time-series. Only subpopulations producing 5 or
more pups in South Australia are considered here.
Methods
Identification of metapopulations Population genetic studies by Campbell (2003) have indicated that Australian sea
lions have one of the highest levels of population subdivision of any pinniped
species, with very high levels of mtDNA haplotype fixation among colonies. Female
Australia sea lions display extreme levels of natal site fidelity (i.e. there is little or no
interchange of females among breeding colonies), with some population divisions
occurring over as little as 20 km (Campbell 2003). This suggests that although
females may forage out to and beyond 100km from their breeding colony
(Goldsworthy et al. unpublished data), dispersal and genetic exchange between
adjacent colonies within this range may be extremely limited. Although reproductive
isolation of subpopulations may be due in part to the asynchrony of breeding
seasons between colonies, genetic subdivision also occurs between adjacent
colonies for which the timing of breeding is synchronised (e.g. North Fisherman and
Buller Island in Western Australia) (Campbell 2003). Such population subdivision has
significant implications for management of the species, suggesting that breeding
colonies should be managed as separate subpopulations (Campbell 2003). Campbell
(2003) also identified that measures of genetic differentiation (such as ΦST and Nei’s
corrected D) were significantly positively correlated with the geographic distance
between colonies. Based on these findings, and because the genetic relatedness of
many Australian sea lion subpopulations is unknown, we developed a distance matrix
among 65 Australian sea lion subpopulations (Figure 6.1) as a proxy for genetic
distance. In order to identify metapopulations among these subpopulations, we
analysed the distance matrix using a Bray Curtis similarity and clustering approach
using the multivariate statistic software Primer (V5.2, Plymouth Marine Laboratory,
UK). Distances (km) were measured as the shortest straight line distance
constrained by geographical boundaries (i.e. coastlines, headlands and islands)
using the GIS software MapInfo, in conjunction with an algorithm that calculates the
distance between two latitude/longitude positions.
Within the metapopulations identified, subpopulations were considered suitable for
ongoing monitoring on the basis of:
1. logistics, particularly ease and safety of access,
Subpopulation survey strategies 42
2. relative cost of reaching the colony, including the likely method of travel
(helicopter, boat, vehicle) and travel distance,
3. physical attributes of the site, including suitability for camping for colonies with
relatively large pup numbers (where estimation procedures may take several
days on each visit), and ease of moving around the whole colony, and
4. previous knowledge of colony size.
Where possible we identified one large (>40 pups) and one small (<40 pups)
subpopulation within each metapopulation that would be most suitable for ongoing
monitoring.
Breeding chronology of subpopulations The breeding chronology of each of subpopulation was estimated based on data
available on the timing of breeding for Australian sea lion populations in South
Australia (Dennis 2005). Timing of the beginning of each breeding season was
predicted at some colonies by estimating the age of the oldest pups on the basis of
several factors, including: their size, coordination and the stage of moult (see
Shaughnessy et al. 2005). After eight weeks of age, the natal coat deteriorates and
by 12 weeks the new coat begins to appear in most pups (Shaughnessy et al. 2005).
Information on the timing of breeding at some subpopulations is presented in Gales
et al. (1994) and Shaughnessy et al. (2005) was used as well as an interpretation
from a compilation of counts of Australian sea lion pups (Dennis 2005).
Chronology data are summarised in tabular form for each month between 2002 -
2012. As the interval between successive breeding seasons is approximately 17.5
months, results were presented by alternating the timing of the beginning of
successive breeding seasons by 17 and 18 months. The length of the pupping
season was taken as 6 months in small colonies, 7 months in large colonies and
longer where supported by data.
Results
Identification of metapopulations The Bray Curtis dendrogram of distance among Australian sea lion subpopulations
indicates clear separation between Western Australian and South Australian
populations (Figure 6.2, raw data are presented in the Appendices). Within WA,
subpopulations fall into three major groups, including the West Coast, South Coast,
Subpopulation survey strategies 43
and Recherché Archipelago Regions. A fourth, minor group consists of the single
isolated subpopulation, Bunda Cliffs 10 (Figure 6.2).
Within South Australia, subpopulations broadly formed three major groups, the
Bunda Cliffs, West Coast and the Central Coast Region. These could be further
subdivided into seven main metapopulations detailed below (Figure 6.2).
The Bunda Cliffs Region forms its own metapopulation:
1. Bunda Cliffs (seven known subpopulations).
The West Coast Region is subdivided into three main metapopulations:
2. Nuyts Reef (one subpopulation),
3. Nuyts Archipelago (9 subpopulations), and
4. Chain of Bays (6 subpopulations) (Figure 6.2).
The Central Coast Region is subdivided into three main metapopulations that
include:
5. South-west Eyre (3 subpopulations)
6. Southern Spencer Gulf and nearby waters (9 subpopulations), and
7. Kangaroo Island (4 subpopulations) (Figure 6.2).
The seven SA metapopulations are considered below, beginning in the west and
moving in a south-easterly direction. Summary data indicating the size, access, and
logistics for each site are summarised in Table 6.1.
1. Bunda Cliffs
Small numbers of pups have been recorded at each of seven sites (B1-9) (Dennis
and Shaughnessy 1996). All are difficult to access, being at the bottom of 90m high
cliffs. Access would have to be by boat (which would be weather dependent and
landings would be very difficult for some sites), or by abseiling (which requires
arranging for skilled operators in cliff rescue, eg. State Emergency Service). Both
options have severe drawbacks in terms of logistical difficulties and safety. Another
option is to count pups from the cliff-top (as has been done to date), but large
boulders prevalent at these sites make that difficult and under-estimation is the likely
result. Consequently, accurate repeatable surveys are unlikely to be achieved, and
as such we do not recommend including any of these sites in the suite of colonies for
ongoing monitoring.
Subpopulation survey strategies 44
2. Nuyts Reef
The breeding colony at Western Nuyts Reef is considerably isolated, with the nearest
boat ramp at Fowlers Bay. Charter operators are scarce and the reefs are difficult to
land upon with swell and conditions usually preventing safe landing with a small craft.
If a helicopter were available at Ceduna, the distance by air to Nuyts Reef would be
~160km. Access by helicopter would be expensive. Regular repeatable surveys are
unlikely to be achieved at this site, and we do not recommend including it for ongoing
monitoring.
3. Nuyts Archipelago
Nine islands are included in this group: Purdie, Lounds, West, Fenelon, Masillon, and
‘Breakwater1’ Islands, Gliddon, West Franklin (‘Blefuscu’ Island1) and East Franklin
(‘Lilliput’ Island1) Reefs. Two islands with large numbers of pups are suitable, West
Franklin Reef (‘Blefuscu’ Island1) and East Franklin Reef (‘Lilliput’ Island1). In the
2004-05 pupping season, direct counts of pups at these sites were 84 and 67,
respectively. Each island is about 1 hour by boat from Ceduna or Smoky Bay, and
camping is feasible on each. Of the two, West Franklin Island (‘Blefuscu’ Island1) is
preferred because landing by small craft is easier and pup numbers are larger.
Other large colonies include Purdie and West Islands. Purdie Island is surrounded by
deep water and sites for safe landing by small craft are limited and hazardous.
Helicopter access would be most suitable for this site. West Island (in St Francis
Group) is about 60 km from Ceduna by boat, but has a small embayment providing
some protection for landings via small craft, and good camping and mobility around
the island. Neither of these is more suitable than West Franklin (‘Blefuscu’ Island1)
and East Franklin Reefs (‘Lilliput’ Island1).
Subpopulations with small numbers of pups include Masillon, Fenelon, Lounds and
Breakwater Island1 and Gliddon Reef. Masillon and Fenelon Islands are located in
the south eastern part of the Nuyts Archipelago (60-70 km from Ceduna) in the St
Francis Group. They are both difficult to access by small craft and safe access is
highly dependent on sea conditions. Helicopter access to these islands would be
preferable. Lounds Island is surrounded by steep drop-offs making boat landings
very difficult and unsafe. Breakwater Island1 and Gliddon Reef are the most
1 These are unofficial placenames that have been submitted to the South Australian Geographic Names Unit (pending approval).
Subpopulation survey strategies 45
accessible small populations in the region. Each island is about 30 minutes by boat
from Ceduna and boat landings are relatively straightforward. Of the two, Breakwater
Island is preferred because it is easier for boat landings and it has more pups, but the
two islands could effectively be surveyed together as they are only 2.8 km apart, and
may in fact be one subpopulation. Movement of pups between islands is likely (even
at a young age), hence we recommend both be surveyed together. In the 2005
pupping season, direct counts of pups were 17 and 7, respectively.
4. Chain of Bays
There are six islands in the Chain of Bays group: Olive, Nicolas Baudin, Jones,
Ward, Pearson and West Waldegrave Islands. The most suitable with a large number
of pups is Olive Island, which is about 32 km (45 min) by boat from Streaky Bay. It
has a protected embayment and beach on its north-eastern side that is excellent for
landings (tender typically not needed), there is a suitable camping at its northern end,
and access around the island is good. In 2006 the pup production was estimated to
be 205 (178-267, Chapter 4).
Nicolas Baudin Island is easily accessible by small craft from Sceale Bay, however
tides play a significant role in access and the timing of surveys. Camping would be
difficult but as the island is so close to shore, day trips from the township of Sceale
Bay would be feasible. Ward and Pearson are offshore islands 50-60km from Elliston
making them more logistically challenging and expensive. Access to Ward via small
craft is difficult due to fringing reefs and swell. Pearson Island has good anchorages
and landings can usually be made on the more sheltered east coast. West
Waldegrave Island is a large colony, with good access from the Elliston boat ramp.
There is a developing abalone aquaculture site adjacent to the island with daily boat
traffic that might assist with logistics. Landings with the aid of small craft on the
northern shore can be undertaken safely in good conditions (but can be difficult), and
camping on the island is feasible.
The most accessible small colony in the Chain of Bays group is Jones Island. It is
about 15 min by boat from the township of Baird Bay and landing there is
straightforward. In the three pupping seasons from 2001-02 to 2004-05, from 6 to 15
pups were counted there.
Subpopulation survey strategies 46
5. South-west Eyre
Three small breeding sites are known in this region, all on islands: Rocky (North),
Four Hummocks (northern islet) and Price Island. Because of their inaccessibility,
these colonies have only been visited by seal biologists once, in January 1996 by
helicopter when 16, 12 and 25 pups were counted, respectively. Travel to these
islands by boat would take several hours from Port Lincoln or Coffin Bay (shorter
time from Avoid Bay) and landings could be difficult because they are small islands
with steep drop-offs and the swells tends to wrap around them. Consequently, we do
not recommend any of them for inclusion as representative colonies for monitoring
pup abundance.
6. Southern Spencer Gulf and nearby waters
Nine islands are included in this group: English, Lewis and Albatross Islands and
Dangerous Reef on the eastern side of southern Spencer Gulf, North Island and
Peaked Rocks in the Gambier Islands in south-western Spencer Gulf, ‘East’ Island of
the North Neptune Group, South Neptune and Liguanea Islands to the south and
south-east of the lower Eyre Peninsula. The most suitable large subpopulations are
Dangerous Reef and Lewis Island. Dangerous Reef is about 33km from Port Lincoln
and landings can usually be made on its northern side. Camping on top of the island
is satisfactory (although extremely exposed) and research groups have camped
there previously. Pup numbers have been surveyed there by direct counting in the
last five pupping seasons (1999 to 2005) and have ranged from between 383 and
585. Lewis Island has reasonable landing sites and good camping. Count data for
pups are only available for one season (78 in 2004-05). The island is less exposed
than Dangerous Reef, and access during the height of the breeding season is likely
to be less of a challenge, with less movement restrictions due to lower densities of
seals, and survey procedures would take less time (smaller population size)
compared to Dangerous Reef. However, given that the Dangerous Reef
subpopulation has some of the best time-series data for the species, continuation of
monitoring at this site is preferable.
Albatross Island is situated at the south-eastern end of Thistle Island, more than
50km from Port Lincoln. It is exposed to southerly swells that wrap around the island.
Landing by small craft would be unsafe in most conditions. Helicopter access would
be the safest option.
Subpopulation survey strategies 47
The only suitable colony in the group with a small number of pups is at English
Island. It is situated about 33km NE of Port Lincoln and landing there is usually
straightforward. It is only 20 minutes from Dangerous Reef, and visits to the two
islands can often be combined. In the four pupping seasons from 1999 to 2002, from
2 to 15 pups were counted there, although there were indications that the colony was
being disturbed in some of those seasons (Dennis 2005). In the 2005 pupping
season, 27 pups were counted there.
In the Gambier Islands, North Island can be accessed in good conditions by boat and
tender, and camping is possible with relative low densities of animals affording good
access. Peaked Rocks off the south-west coast of Wedge Island are unsafe for small
boat access.
‘East’ Island of the North Neptune Group, South Neptune and Liguanea Islands are
all accessible by boat from Port Lincoln, but distances (80+ km) are significant.
Liguanea Island can be difficult to land people via small craft, but is fine in good
conditions. Helicopter access would be more reliable and the distance is relative
short from Port Lincoln airport (50km), although a helicopter is not usually based
there. ‘East’ Island (off North Neptune Island) has a large gulch on its north-western
side that can afford good protection from southerly and south-easterly conditions, if
the swell is not too high. This colony was discovered in 2004 when 14 pups were
counted. People can be landed safely at South Neptune (Main Island) via small craft
on its eastern shores, or within the large gulch of Eagle Bay. Brown pups were
recorded on the island in 1969 and 1970 (Stirling 1972), October 1991 (Dennis 2005)
and early 1993 (Shaughnessy et al. 2005). But no pups were seen during our visits to
the island in February 2000, September 2004 and February 2005 and February
2006.
7. Kangaroo Island
There are four colonies in this group: North and South Page Island (in the Backstairs
Passage) and Seal Bay and the Seal Slide on the southern and south-eastern coast
of Kangaroo Island. The most suitable with a large number of pups is at Seal Bay,
which is accessible by vehicle and accommodation is available nearby. Pup numbers
have been difficult to assess at Seal Bay because of access to parts of the colony
(Pup Cove and the Eastern Prohibited Area), although solutions are possible. Direct
counting is difficult in the inland parts of the colony where pups hide under bushes.
That problem is readily addressed by the mark-recapture approach we propose.
Subpopulation survey strategies 48
Direct maximum count of pups in 13 pupping seasons between 1985 and 2002-03
averaged 144, and ranged between 122 and 166. An important reason for including
Seal Bay in the suite of colonies to be monitored is because that series of counts
showed a decreasing trend (Shaughnessy et al. 2006). In addition, the colony is a
major tourist destination and a key attraction that supports a significant regional
tourism industry.
The only suitable small colony in the Kangaroo Island group is the Seal Slide, which
is accessible by vehicle. In three of the four pupping seasons from 2000 to 2004,
between 7 and 11 pups were counted there. In the 2005-06 pupping season, the
estimate was 10 pups using the procedure we plan for monitoring pup numbers in
small colonies.
Both North and South Page are offshore islands that are best accessed by
helicopter. The regional office of SA DEH at Kingscote (Kangaroo Island) has
undertaken regular surveying of these sites by helicopter over the last decade.
In summary, suitable large and small reference colonies were identified from four of
the seven identified metapopulations in SA, as detailed below:
1. Bunda Cliffs – no colonies selected
2. Nuyts Reef – no colonies selected
3. Nuyts Archipelago
• West Franklin Reef (‘Blefuscu’ Island) – large colony
• ‘Breakwater’ Island (including Gliddon Reef) – small colony
4. Chain of Bays
• Olive Island – large colony
• Jones Island – small colony
5. Southwest Eyre – no colonies selected
6. Southern Spencer Gulf and nearby waters
• Dangerous Reef – large colony
• English Island – small colony
7. Kangaroo Island
• Seal Bay – large colony
• Seal Slide – small colony
Subpopulation survey strategies 49
Breeding chronology of key subpopulations
The timing of the breeding season for most SA subpopulations between 2002 and
2006 is presented in Figure 6.3, and includes predictions for subsequent pupping
seasons until 2012. The extent that the timing of breeding is known for a particular
site is indicated by the number of ground surveys that have been undertaken
between 2002 and 2006 (Figure 6.3). As can be seen, the extent of data for many
sites is poor, and as such, the predicted timing of breeding should be seen more as a
guide. For example, only one ground survey is available for Gliddon Reef, and based
on this the timing of breeding is predicted to occur three months later than
Breakwater Island (Figure 6.3). However, these sites are only separated by 2.5 km,
hence the timing of breeding at Gliddon is likely to be the same as that for
Breakwater Island. For other sites including North Island, Peaked Rocks, Lounds
Island and Nuyts Reef, no ground surveys have been undertaken during breeding
seasons in recent decades if at all (Figure 6.3). For some sites, including Albatross,
South Neptune, Price, Four Hummocks, Rocky (North) and Masillon Islands, data are
so sparse and/or from many years ago that timing of breeding is purely speculative.
Table 6.2 proposes a subpopulation survey plan for three years (2007-2009) for the
eight representative SA Australian sea lion subpopulations, based on their expected
schedule of breeding and on a minimum of three surveys per breeding season for
small subpopulations (<40 pups) and four surveys per breeding season for large
subpopulations (>40 pups). Where possible the timing of surveys has been
synchronised among sites. Even though this is a modest number of sites to survey
(only 15% of 39 SA sites where 5 or more pups are born), it still requires surveys to
be undertaken on 22 (61%) of the 36 months presented (2007-2009).
Discussion Based upon a distance analyses, breeding colonies of the Australian sea lion in
South Australia were divided into seven regional groups or metapopulations, forming
a basis from which to select appropriate, representative sites where accurate
estimates of pup production can readily be derived. Monitoring of these sites will
facilitate the determination of trends on Australian sea lion populations across their
range over the shortest possible time-series. Suitable survey sites were selected
from four of seven metapopulations recognised including Nuyts Archipelago, Chain of
Bays, Southern Spencer Gulf and Kangaroo Island. We did not select sites in the
Subpopulation survey strategies 50
Bunda Cliffs, Nuyts Reef or south-west Eyre regions because none satisfied our
criteria of accessibility or the likelihood of being able to undertake accurate,
repeatable surveys.
One small and one large subpopulation within each metapopulation was selected as
the minimum required to adequately assess the status of subpopulations within these
regions for the species. The eight sites represent about 20% of South Australia’s
breeding sites, where five or more pups are produced each breeding season. The
small sites were considered especially important because almost no trend data are
currently available for them, and they form more than 60% of all breeding sites in SA
and are most vulnerable to extinction (Goldsworthy et al. in review). Selection of
monitoring sites within metapopulations was a pragmatic exercise, based on
logistics, practicalities, the capacity to undertake accurate repeatable surveys and
existence of prior data. The selection process did not identify sites on the basis of
need of information, either because of our poor knowledge on status, or because of
management imperatives/conservation concerns eg. within Marine Parks/Aquatic
Reserves, adjacent to aquaculture sites, commercial fisheries, tourism or extinction
risk (see Goldsworthy et al. in review).
Caution should be placed on the significance of the metapopulation groupings
presented, as they may not accurately reflect the extent or level of subpopulation
subdivision in the species. For example, within the Nuyts Archipelago
metapopulation, satellite tracking studies undertaken at six breeding colonies indicate
marked difference between the foraging behaviour of adult females among sites.
Females demonstrated either inshore (shallow) or offshore (deep) foraging
behaviours, that also appeared colony specific, with female body mass differing by
25% between sites (Goldsworthy et al. unpublished data). Even adjacent colonies
such as East and West Franklin Reefs (‘Lilliput’ and ‘Blefuscu’ Islands, <5km apart),
demonstrated marked difference in foraging behaviour (Goldsworthy et al.
unpublished data). These findings suggest that population subdivisions are likely to
occur at a finer scale than has been suggested by this metapopulation analyses,
which has been confirmed within the Recherché Archipelago in WA (Campbell 2003).
Genetic analyses of the degree of relatedness among subpopulations (especially
those in SA) needs to be undertaken to refine the management units for the species.
Subpopulation survey strategies 51
If a commitment to resource the ongoing assessment of pup production at the eight
sites proposed here were achievable, it should not preclude surveys being
undertaken at other sites. For most ASL breeding sites, data on pup production is
extremely poor, and no baseline information is available. Clearly, in addition to
committing to ongoing monitoring of pup production trends at some key
representative sites, there should also be a commitment to acquire baseline data on
the status of other sites where data are currently poor. In the absence of any other
historic data, such information would improve our understanding on the size of ASL
populations, and provide critical baseline data against which future surveys could be
compared. Other sites could be prioritised according to risk. For example, a better
understanding of the status of sites with very low pup production is critical, given the
high number of sites that produce fewer than 30 pups for which data are currently
poor. These are the sites at greatest risk of extinction in the near future (Goldsworthy
et al. in review).
Subpopulation survey strategies 52
Figure 6.1. Distribution of Australian sea lion breeding sites in South Australia. Seven metapopulations are also identified in bold, based on the distance analysis.
Subpopulation survey strategies 53
Figure 6.2. Dendrogram of subpopulation distance similarity of 64 South Australian and Western Australian breeding sites. Eleven metapopulations are identified.
100
80
60
Similarity
40
WA
SA
West C
oast
South C
oast
Recherche
Bunda 10
Bunda C
liffs
Nuyts R
eef
Nuyts A
rchipelago
Chain of B
ays
SW
Eyre
S S
pencer Gulf
Kangaroo Island
Abrolhos Is -E
astern Group
Abrolhos Is -S
outhern Group
Buller Is
Beagle Is
North Fisherm
an Is H
aul Off R
ock
Middle D
oubtful Is R
ed Islet
West Is
Rocky Is
Little Is M
acKenzie Is
Kim
berley Is
Kerm
adec (Wedge) Is
SW R
ock (Twin Peaks Is)
Taylor Is G
lennie Is
Stanley (W
ickham) Is
Poison C
reek Is
Salisbury Is
Cooper Is
Round Is
Six M
ile Is Ford (H
alfway) Is
Spindle
B10
B9
B8
B6
B5
B3
B2
B1
Western N
uyts Reef
West Is
Fenelon Is
Masillon Is
Purdie Is
Lounds Is B
reakwater Is
Gliddon R
eef W
Franklin Reef (B
lefuscu Is) E
Franklin Reef (Lilliput Is)
Olive Is
Nicolas Baudin Is
Jones Is W
ard Is P
earson Is W
est Waldegrave Is
Rocky Is (N
orth) Four H
umm
ocks (North) Is
Price Is
Liguanea Is Lew
is Is N
orth Neptune (E
ast) Is A
lbatross Is S
outh Neptune (M
ain) Is E
nglish Is D
angerous Reef
Peaked R
ocks N
orth Is S
eal Bay (K
angaroo Is) S
eal Slide
South P
age N
orth Page
Subpopulation survey strategies 54
Table 6.1. Assessment of suitability of Australian sea lion colonies in South Australia for representative monitoring site on the basis of accessibility, safety of access, and ability to camp and move around sites.
Metapopulation Subpopulation Colony type Logistics Physical Approx pup
Likely mode Access from Travel dist
(km) Safe access Camping production
Kangaroo Island North Page L Helicopter Kangaroo Is/Cape Willoughby 18 Yes No South Page L Helicopter Kangaroo Is/Cape Willoughby 16 Yes Yes Seal Slide S Vehicle Kangaroo Is Yes Yes 11 Seal Bay (Kangaroo Is) L Vehicle Kangaroo Is Yes Yes 214 South Spencer Gulf Peaked Rocks S Boat Port Lincoln 77 No 24 North Island S Boat Port Lincoln 73 Yes Yes 28 Dangerous Reef L Boat Port Lincoln 33 Yes Yes 585 English Island S Boat Port Lincoln 33 Yes Yes 27 Albatross Island S Helicopter Port Lincoln 54 No 15 Lewis Island L Boat Port Lincoln 39 Yes Yes 73 South Neptune (Main) Is. S Boat Port Lincoln 80 Yes Yes 6 North Neptune (East) Is. S Boat Port Lincoln 70 Yes Yes 14 Liguanea Island L B/H Port Lincoln 80/50 Marginal/Yes Yes 43 South West Eyre Price Island S Helicopter Port Lincoln 55 Boat No/Heli Yes 25 Four Hummocks (North) Is. S Helicopter Port Lincoln 78 Boat No/Heli Yes 12 Rocky Island (North) S Helicopter Port Lincoln 68 Boat No/Heli Yes 16 Chain of Bays West Waldegrave Island L Boat Elliston 9 Yes Yes 157 Ward Island S Boat Elliston 53 Boat No/Heli Yes 8 Pearson Island S Boat Elliston 66 Yes Yes 27 Jones Island S Boat Baird Bay 7 Yes No 15 Nicolas Baudin Island L Boat Sceale Bay 6 Yes No 72 Olive Island L Boat Streaky Bay 32 Yes Yes 131 Nuyts Archipelago East Franklin Reef L Boat Ceduna 39 Yes Yes 67 West Franklin Reef L Boat Ceduna 45 Yes Yes 84 Breakwater Island S Boat Ceduna 27 Yes No 17 Gliddon Reef S Boat Ceduna 28 Yes No 7 Lounds Island S Helicopter Ceduna 41 Boat No/Heli Yes Yes 26 Fenelon Island S Helicopter Ceduna 68 Boat No/Heli Yes Yes 21 Masillon Island S Helicopter Ceduna 67 Boat No/Heli Yes Yes 9 West Island L Boat Ceduna 60 Yes Yes 56 Purdie Island L Helicopter Ceduna 53 Boat No/Heli Yes Yes 132 Nuyts Reef Western Nuyts Reef S B/H Ceduna 40/160 Boat No/Heli Yes 14 Bunda cliffs Bunda Cliffs B1 S Vehicle/Cliff abseil Ceduna 15 Bunda Cliffs B2 S Vehicle/Cliff abseil Ceduna 5 Bunda Cliffs B3 S Vehicle/Cliff abseil Ceduna 31 Bunda Cliffs B5 L Vehicle/Cliff abseil Ceduna 43 Bunda Cliffs B6 S Vehicle/Cliff abseil Ceduna 12 Bunda Cliffs B8 S Vehicle/Cliff abseil Ceduna 38 Bunda Cliffs B9 S Vehicle/Cliff abseil Ceduna 17
Subpopulation survey strategies 55
Table 6.2. Estimated timing of breeding at selected representative Australian sea lion subpopulations in South Australia between 2007 to 2009. Crosses indicate approximate timing of surveys, three surveys for small breeding subpopulations and fours surveys of large breeding subpopulations. Shading indicates predicted pupping seasons.
2007 2008 2009 J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J Seal Slide x x x x x x Seal Bay x x x x x x x x English Island x x x x x x Dangerous Reef x x x x x x x x Jones Island x x x x x x Olive Island x x x x x x x x Breakwater/Gliddon x x x x x x West Franklin (Blefuscu) Is x x x x x x x x
Subpopulation survey strategies 56
Figure 6.3. Diagrammatic representation of breeding season commencement and duration at Australian sea lion colonies in South Australia between 2002 and 2006, with predicted seasons to 2012.
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F
The Pages Is. G G G G G G G G G G G G G G G G G G G
Kangaroo Is. G G G G G G G G G G G G G G G G G G G G G G G
Peaked Rocks* B
North Is.* G
North Neptune Is. (E). G G
English Is. G G G G G
Dangerous Reef G G G G G G G G G G G G G G G G G
Lewis Is. B G G G G
Liguanea Is. G G
Ward Is. G
Pearson Is. G G G G G G
West Waldegrave IsG G G G G G G
Jones Is. G G G G G G G
Nicolas Baudin Is. G G G G G G G G
Olive Is. G G G G G G G G
Franklin Is(s). G G G G
Gliddon Reef G
Fenelon Is. G G G
Breakwater Is. G G G G G G
West Is. G G G G
Lounds Is. G G
Purdie Is. G G
Nuyts Reef * G
Bunda Cliffs G G G G G
J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F
G Ground surveyB Boat based survey
Actual or predicted breeding seasons span six months for small colonies, 7–8 months at larger colonies except at The Pages Islands, where it appears to be longer or more variable.
* Sites of uncertain status and season commencement; requiring further survey
Sites where breeding has been reported but data is lacking include: South Neptune, Albatross, Four Hummocks (N), Price, North Rocky, Greenly and Masillon Islands
2011 20122002 2003 2004 2005 2006 2007 2008 2009 2010
Conclusions and recommendations 57
7 CONCLUSIONS AND RECOMMENDATIONS
Australian sea lions present unique challenges in obtaining accurate information about
the size and trends in their populations. This stems from difficulties in obtaining accurate
estimates of pup production at individual subpopulations, due to the extended breeding
seasons and sightability and availability biases, and from difficulties in obtaining status
and trend data across the species range due to the large number of subpopulations, and
their asynchronous breeding cycles. Given these challenges, there is little surprise that
data on the size and trends in numbers at specific colonies is poor and scarce, and that
data on trends in abundance are largely non-existent. For conservation and
management purposes, there is a critical need to develop better methods for estimating
abundance and trends over the shortest time series possible. Given the difficulties and
costs of surveying at many sites, it is important that the aim of obtaining quality time
series data it is focused at regionally representative sites where ease and reliability of
access enable repeatable, accurate and cost effective surveys to be undertaken.
We developed and tested the appropriateness of two new methods for estimating pup
production in Australian sea lion subpopulations. The first method utilised individual
resight histories of pups with Cormack-Jolly-Seber (CJS) models in conjunction with
standard mark-recapture methods to improve estimates of pup production for large ASL
subpopulations. The second approach developed a cumulative mark and count (CMC)
method for improving estimates of pup production in small ASL subpopulations.
CJS methods trialled at Olive Island produced pup production estimates that were
greater than those based on direct counting and the Petersen estimate. Pup mortality
was estimated to range from 15-52, with recovered mortalities (34 in total) potentially
accounting for only 65% of maximum mortalities. There was no evidence for permanent
emigration, suggesting that the most important source of error in mark-recapture
procedures at this colony is due to unaccounted mortality. The best estimate of pup
production for the 2006 season at Olive Island based on CJS methods was 205 (range
193-256). This was 1.37 times the estimate based on direct counting methods (150
pups), but was very similar to the result (1.03 times) obtained from the Petersen estimate
(mean 197, range 191-203). The adjusted Petersen estimate (adding the mortality range
34-52) produced essentially the same estimate as the CJS approach (mean 206, range
191-223).
Conclusions and recommendations 58
The CJS approach, although providing significant improvements to previous methods,
was not without its problems. CJS analyses suggested no significant pup production
occurred beyond the second session (24 March), only three months into the breeding
season. This was contrary to observations of the presence of perinatal mothers and new
born pups up to session 5 (19 June), increases in pup abundance between sessions 2-5
based on absolute counts and Petersen estimates, on evidence of a 5-7 month breeding
season elsewhere and on time series from direct counting in previous seasons. The
reasons for disparities in methods are currently unclear, but should be addressed in
future.
One of the clear challenges in estimating pup production in ASL populations is
integrating and utilising often disparate estimates and methodologies. One potential
methodological approach that allows for utilising of different information sets is the
unified likelihood method. This methods provides a single set of confidence intervals
which both 1) account for all the uncertainty together of all data sets and combined sub-
models such as pup mortality, emigration, resight probabilities and 2) does so around
mean estimates that are more reliably accurate, providing more appropriate maximum
likelihood estimates rather than more ad hoc combinations of separate maximum
likelihood estimates from separate models, all seeking to describe the same population.
Such an approach was recommended by Lebreton et al. (1992) for mark-recapture data
and specifically for extending CJS.
Cumulative mark and count methods trialled at a small colony (Seal Slide), supported the
observation that not all pups are available for survey during each visit, and produced a
consistent (repeatable) estimate on two occasions. The surveys would have benefited
from greater numbers of pups being marked in earlier sessions, but that was difficult
because of the aggressive behaviour of adult seals. The development of both methods
has demonstrated improvements in the survey of both large and small Australian sea lion
colonies.
Distance analysis among Australian sea lion subpopulations identified 11 main
metapopulations in the species, seven of them were in South Australia. Among the SA
metapopulations, only four provided sites where accurate, repeatable, cost effective and
logistically feasible surveys could be undertaken. Within each of these one large (>40
pups) and one small (<40 pups) site were selected as potential regionally representative
sites to form the basis of ongoing surveys to provide critical data on the status and
Conclusions and recommendations 59
trends in abundance within metapopulations to support management of the species
across its range.
We recommend to following:
• refine CJS and CMC methods by repeating surveys at Olive Island and the Seal
Slide, as well as other sites and develop a unified likelihood method for analysis
of these datasets.
• Assess the suitability of the survey methods developed here for WA
subpopulations and develop national standards in survey protocols.
• Secure Commonwealth and State resources to enable nation-wide ongoing
assessment of representative subpopulations across the range of the species, to
provide critical information for on-going assessment of ASL populations.
• Improve estimates of pup production at other sites (not part of representative
subpopulations), where data are poor in order to develop a baseline for as many
sites as possible to provide a basis from which future changes may be assessed.
• Undertaking broad genetic subpopulation analysis to improve understanding of
the genetic relatedness of subpopulation, and improve metapopulation analyses.
8 ACKOWLEDGMENTS
We thank the NHT/MSRP for funding this project, and project managers Melissa Giese,
Zoe Cozens and Abigail Beeson for their support. A number of SA DEH staff provided
assistance including Brett Dalzell, Robbie Sleep and Andy Causebrook (Ceduna), Meg
Goecker (Port Lincoln), Bill Hadrill, Clarence Kennedy and Albert Zepf (Kangaroo
Island). Field support was also provided by Mel Berris (Kangaroo Island) and Rachael
Gray (University of Sydney). We thank the PIRSA Animal Ethics Committee for
approving the study, and Kate Lloyd and Peter Canty (SA Department for the
Environment and Heritage) for provision of SA NPW Permits. We thank Matt Guidera
(Streaky Bay Boat Charters) for support with logistics to Olive Island. We also thank
James Bushell, Dennis O’Malley and Tim Ward (SARDI Aquatic Sciences) for project
management support. We thank Rick McGarvey and Mike Steer for providing critical
feedback on the draft report.
References 60
9 REFERENCES
Akaike H (1973) Information theory as an extension of the maximum likelihood
principle. In: Petrov BN, Csaki F (eds) Proceedings of the Second
International Symposium on Information Theory, Budapest, Hungary, p 267-
281.
Alava, J.J., and Salazar, S. (2006) Status and Conservation of Otariids in Ecuador
and the Galápagos Islands. pp 495-519. In ‘Sea Lions of the World’. (Eds A.
W. Trites, S. K. Atkinson, D. P. DeMaster, L. W. Fritz, T. S. Gelatt, L. D. Rea
and K. M. Wynne). Alaska Sea Grant College Program, University of Alaska:
Fairbanks, Alaska.
Berkson, J. M. and DeMaster, D. P. (1985). Use of pup counts in indexing population
changes in pinnipeds. Canadian Journal of Fisheries and Aquatic Sciences
42, 873-879.
Burnham KP (1993) A theory for combined analysis of ring recovery and recapture
data. In: Lebreton JD, North PM (eds) Marked individuals in the study of bird
populations. Birkhauser Verlag, Basel, Switzerland, p 199-213.
Burnham KP, Anderson DR (2002) Model Selection and Multimodal Inference: A
Practical Information-Theoretic Approach, Vol. Springer-Verlag, New York,
USA.
Campbell, R. (2003) Demography and population genetic structure of the Australian
sea lion, Neophoca cinerea. PhD thesis. University of Western Australia,
Western Australia. 153pp (including appendices).
Carretta, J.V., K.A. Forney, M.M. Muto, J. Barlow, J. Baker, and M. Lowry. (2004)
U.S. Pacific marine mammal stock assessments. pp 5-11. NOAA-TM-NMFS-
SWFSC-358 Tech. Memo. 295 pp.
Caughley G (1977) Analysis of vertebrate populations. John Wiley & Sons. 234 pp.
Cerchio S (1998) Estimates of humpback whale abundance off Kauai, 1989-1993:
evaluating biases associated with sampling the Hawaiian Islands breeding
assemblage. Mar Ecol Prog Ser 175:23-24.
Cormack RM (1964) Models for capture-recapture. Biometrika 51:429-438.
References 61
Crespo, E.A., and Pedraza, S.N. (1991) Estado actual y tendencia de la poblacion
de lobos marinos de un pelo (Otaria flavescens) en el litoral norpatagonico
(Spanish). Ecologia Austral 1: 87–95.
Dennis, T. E. (2005). Australian sea lion survey (and historical) records for South
Australia. Report to the Wildlife Conservation Fund, Department for
Environment and Heritage, South Australia.
Dennis, T.E. and Shaughnessy, P.D. (1996). Status of the Australian sea lion,
Neophoca cinerea, in the Great Australian Bight Wildlife Research. 23: 741-
754.
Dennis, T.E. and Shaughnessy, P.D. (1999). Seal survey in the Great Australian
Bight region of Western Australia. Wildlife Research. 26: 383-388.
Fowler, S.L. (2005) Ontongeny of diving in the Australian sea lion. PhD Thesis.
University of California, Santa Cruz.
Fowler, S.L., Costa, D. P., Arnould, J. P. Y. , Gales, N. J. and Kuhn, C. E. (2006).
Ontogeny of diving behaviour in the Australian sea lion: trials of adolescence
in a late bloomer. Journal of Animal Ecology 75: 358-367.
Gales, N.J., Costa, D.P. (1997) The Australian sea lion, a review of an unusual life
history. In: Hindell, M., Kemper, C. (eds.). Marine mammal research in the
Southern Hemisphere, Vol. 1 Status, ecology and medicine. Surrey Beatty
and Sons, Chipping Norton. pp. 78-87.
Gales, N. J., Shaughnessy, P. D., and Dennis, T. E. (1994). Distribution, abundance
and breeding cycle of the Australian sea lion Neophoca cinerea (Mammalia:
Pinnipedia). Journal of Zoology, London 234: 353-370.
Goldsworthy SD, Bulman C, He X, Larcombe J, and Littnan C (2003) Trophic
interactions between marine mammals and Australian fisheries: an
ecosystem approach. In: Gales N, Hindell M, and Kirkwood R. (eds) Marine
Mammals and Humans: Fisheries, tourism and management. CSIRO
Publications. pp 65-99.
Goldsworthy SD., Hamer D, Page BC (in review) A Risk-Assessment Approach to
Evaluating the Significance of Seal Bycatch in two Australian Fisheries.
Biological Conservation.
Harwood J, and Prime JH (1978) Some factors affecting the size of British grey seal
populations. Journal of Applied Ecology 15:401-11.
References 62
Higgins, LV., and Gass, L. (1993) Birth to weaning: parturition, duration of lactation,
and attendance cycles of Australian sea lions (Neophoca cinerea). Canadian
Journal of Zoology. 71:2047-2055.
Huggins, R. M. 1989. On the statistical analysis of capture-recapture experiments.
Biometrika 76:133-140.
Huggins, R. M. 1991. Some practical aspects of a conditional likelihood approach to
capture experiments. Biometrics 47:725-732.
Jolly GM (1965) Explicit estimates from mark-recapture data with both death and
immigration-stochastic models. Biometrika 52:225-247.
Kendall, W. L., K. H. Pollock and C. Brownie. 1995. A likelihood-based approach to
capture–recapture estimation of demographic parameters under the robust
design. Biometrics 51:293–308.
Kirkwood, R., Gales, R., Terauds, A., Arnould, J. P. Y., Pemberton, D.,
Shaughnessy, P. D., Mitchell, A. T., and Gibbens, J. (2005) Pup production
and population trends of the Australian fur seal (Arctocephalus pusillus
doriferus). Marine Mammal Science 21: 260-282.
Lalas, C. and Bradshaw, C.J.A. (2003) Expectations for population growth at new
breeding locations for the vulnerable New Zealand sea lion (Phocarctos
hookeri) using a simulation model. Biological Conservation 114: 67-78.
Ling, J. K. (1992). Neophoca cinerea. Mammalian Species 392: 1-7.
Ling, J. K., and Walker, G. E. (1976). Seal studies in South Australia: progress
report for the year 1975. South Australian Naturalist 50: 59-68, 72.
Ling, J. K., and Walker, G. E. (1979). Seal studies in South Australia: progress
report for the period April 1977 to July 1979. South Australian Naturalist 54:
68-78.
Lebreton, J-D., Burnham, K.P., Clobert, J., and Anderson, D.R. (1992). Modelling
survival and testing biological hypotheses using marked animals: a unified
approach with case studies. Ecological Monographs, 62: 67-118.
Mate, B.R. (1982) History and present status of the California sea lion, Zalophus
californianus. FAO Fisheries Series 4: 303-9.
McIntosh, R. R., Shaughnessy, P. D. and Goldsworthy, S. D. (2006a). Mark-
recapture estimates of pup production for the Australian sea lion (Neophoca
cinerea), at Seal Bay Conservation Park, South Australia. In ‘Sea Lions of the
World’. (Eds A. W. Trites, S. K. Atkinson, D. P. DeMaster, L. W. Fritz, T. S.
References 63
Gelatt, L. D. Rea and K. M. Wynne). pp 353-367. (Alaska Sea Grant College
Program, University of Alaska: Fairbanks, Alaska).
McIntosh, R. R., Shaughnessy, P. D. and Goldsworthy, S. D. and Kennedy, C.
(2006b). Developing appropriate monitoring strategies for the declining
Australian sea lion population at Seal Bay. Report to the Wildlife Conservation
Fund, Department for Environment and Heritage, South Australia.
McKenzie, J., Goldsworthy, S. D., Shaughnessy, P. D., and McIntosh, R. (2005).
Understanding the impediments to the growth of Australian sea lion
populations. Final report to Department of the Environment and Heritage,
Migratory and Marine Species Section. South Australian Research and
Development Institute (Aquatic Sciences), Adelaide, 107 pp. SARDI
Publication number RD4/0171.
Otis DL, Burnham KP, White GC (1978) Statistical inference for capture data on
closed animal populations. Wildlife Monographs 62:1-135.
Pollock KH, Solomon DL, Robson DS (1974) Tests for mortality and recruitment in a
K-sample tag-recapture experiment. Biometrics 30:77-87.
Pradel R (1996) Utilization of capture-mark-recapture for the study of recruitment
and population growth rate. Biometrics 52:703-709.
Reyes, L.M., Crespo, E.A., and Szapkievich, V. (1999) Distribution and population
size of the southern sea lion (Otaria flavescens) in central and southern
Chubut, Patagonia, Argentina. Marine Mammal Science 15: 478–493.
Schwarz CJ, Arnason AN (1996) A general methodology for the analysis of open-
model capture recapture experiments. Biometrics 52:860-873.
Seber GAF (1970) Estimating time-specific survival and reporting rates for adult
birds from band returns. Biometrika 57:313-318.
Seber, GAF (1982). The Estimation of Animal Abundance and Related Parameters.
MacMillan, New York.
Shaughnessy, P. (2004). Population assessment of New Zealand fur seals and
Australian sea lions in some South Australian breeding colonies and haul-out
sites, 2003-2004. Report to Department for Environment and Heritage (South
Australia). 49 + ix pp.
Shaughnessy, P. (2005a). Australian sea lions at some colonies on the coast of Eyre
Peninsula, South Australia: abundance in 2004 and 2005. Report to
Department of the Environment and Heritage, 33 pp. October 2005.
References 64
Shaughnessy, P. (2005b). Population assessment of New Zealand fur seals and
Australian sea lions at some colonies in South Australia, 2004-05. Report to
Department for Environment and Heritage (South Australia). 48 + xi pp,
December 2005.
Shaughnessy, P. and Dennis, T. (1999). Seal research in South Australia,
1998/1999: abundance of New Zealand fur seal pups on Kangaroo Island and
Australian sea lion pups on Dangerous Reef. Report to South Australian
National Parks and Wildlife, Department of Environment, Heritage and
Aboriginal Affairs, October 1999, 39 pp.
Shaughnessy, P. and Dennis, T. (2000). Establishing monitoring guidelines and
assessing abundance of Australian sea lions at key breeding localities in
South Australia. Consultancy report to Marine Species Protection Program of
Coasts and Clean Seas, Environment Australia, 25 pp.
Shaughnessy P D, Goldsworthy S D, and Libke JA (1995a) Changes in the
abundance of New Zealand fur seals Arctocephalus forsteri on Kangaroo
Island, South Australia. Wildlife Research 22:201-15 .
Shaughnessy, P.D., Testa J.W., and Warneke, R.M. (1995b) Abundance of
Australian fur seal pups, Arctocephalus pusillus doriferus, at Seal Rocks,
Victoria, 1991-92 from Petersen and Bayesian estimators. Wildlife Research
22:625-32.
Shaughnessy PD, Kirkwood R, and Warneke RM (2002) Australian fur seals,
Arctocephalus pusillus doriferus: pup numbers at Lady Julia Percy Island,
Victoria, and a synthesis of the species population status. Wildlife Research
29: 185-192.
Shaughnessy, P. D., Dennis, T. E. and Seager, P. G. (2005). Status of Australian
sea lions, Neophoca cinerea, and New Zealand fur seals, Arctocephalus
forsteri, on Eyre Peninsula and the far west coast of South Australia. Wildlife
Research 32: 85-101.
Shaughnessy, P. D., Dennis, T.E., Dowie D., McKenzie, J. and McIntosh, R. (in
prep.) Status of small colonies of the Australian sea lions, Neophoca cinerea,
on Kangaroo Island, South Australia. Australian Mammalogy
Shaughnessy, P. D., McIntosh, R. R., Goldsworthy, S. D., Dennis, T. E., and Berris,
M. (2006). Trends in abundance of Australian sea lions, Neophoca cinerea, at
Seal Bay, Kangaroo Island, South Australia. In ‘Sea Lions of the World’. (Eds
References 65
A. W. Trites, S. K. Atkinson, D. P. DeMaster, L. W. Fritz, T. S. Gelatt, L. D.
Rea and K. M. Wynne ). pp 325-351. (Alaska Sea Grant College Program,
University of Alaska: Fairbanks, Alaska).
Shiavini, A.C.M., Crespo, E.A., and Szapkievich, V. (2004) Status of the population
of South American sea lion (Otaria flavescens Shaw, 1800) in southern
Argentina. Mammalian Biology 69: 108-118
Szteren, D., Aurioles, D., and Gerber, L.R. (2006) Population Status and Trends of
the California Sea Lion (Zalophus californianus californianus) in the Gulf of
California, Mexico. In ‘Sea Lions of the World’. (Eds A. W. Trites, S. K.
Atkinson, D. P. DeMaster, L. W. Fritz, T. S. Gelatt, L. D. Rea and K. M.
Wynne). Alaska Sea Grant College Program, University of Alaska: Fairbanks,
Alaska.
Thompson, D., Strange, I., Riddy, M. and Duck CD. (2005) The size and status of
the population of southern sea lions Otaria flavescens in the Falkland Islands.
Biological Conservation 121:357-367.
Trites, A.W., Miller, A.J., Maschner, H.D.G. et al. (2007) Bottom-up forcing and the
decline of Steller sea lions (Eumetopias jubatus) in Alaska: assessing the
ocean climate hypothesis. Fisheries Oceanography. 16: 46-67.
White GC, Burnham KP (1999) Program MARK: survival estimation from populations
of marked animals. Bird Stud 46 (Supplement):120-138.
White, G. C., and Garrott, R. A. (1990). 'Analysis of Wildlife Radio-tracking Data'.
(Academic Press: San Diego).
Appendices 66
10. APPENDICES
Appendix 1
Capture- mark-recapture data for 144 Australian sea lion pups at Olive Island during the 2006
breeding season, including live and dead recaptures (resighting) over six session (details in
methods, Chapter 4).
LIVE DEAD 16 20 32 33 24
ID SEX S1 S2 S3 S4 S5 S6 S1 S2 S3 S4 S5 S6
PUP033 F 1 1 1 0 1 0 0 0 0 0 0 0
PUP042 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP043 F 0 1 1 0 1 0 0 0 0 0 0 0
PUP045 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP035 F 1 1 1 1 1 0 0 0 0 0 0 0
PUP046 F 0 1 1 1 0 1 0 0 0 0 0 0
PUP034 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP047 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP048 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP027 F 1 1 1 1 1 0 0 0 0 0 0 0
PUP050 M 0 1 1 1 0 1 0 0 0 0 0 0
PUP032 F 1 1 1 1 1 0 0 0 0 0 0 0
PUP037 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP029 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP051 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP052 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP053 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP055 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP056 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP058 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP057 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP017 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP060 M 0 1 1 1 0 1 0 0 0 0 0 0
PUP061 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP062 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP064 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP065 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP066 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP024 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP067 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP010 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP069 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP070 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP072 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP006 M 1 1 1 1 1 0 0 0 0 0 0 0
PUP073 M 0 1 1 1 1 1 0 0 0 0 0 0
Appendices 67
Appendix 1. Cont.
LIVE DEAD
16 20 32 33 24
ID SEX S1 S2 S3 S4 S5 S6 S1 S2 S3 S4 S5 S6
PUP074 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP007 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP075 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP005 F 1 1 1 1 1 0 0 0 0 0 0 0
PUP077 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP078 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP080 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP081 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP022 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP019 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP023 F 1 1 1 0 1 1 0 0 0 0 0 0
PUP082 M 0 1 1 1 1 0 0 0 0 0 0 0
PUP083 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP084 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP021 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP014 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP085 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP086 F 0 0 1 1 0 0 0 0 0 0 0 0
PUP087 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP088 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP020 M 1 1 1 1 0 0 0 0 0 0 0 0
PUP089 M 0 0 1 1 0 1 0 0 0 0 0 0
PUP090 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP091 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP092 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP093 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP012 M 1 1 1 1 1 0 0 0 0 0 0 0
PUP040 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP108 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP041 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP109 F 0 0 1 1 0 0 0 0 0 0 1 0
PUP112 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP110 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP028 F 1 1 1 1 1 1 0 0 0 0 0 0
PUP115 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP026 F 1 1 1 1 1 1 0 0 0 0 0 0
PUP114 F 0 0 1 0 0 0 0 0 0 0 0 0
PUP011 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP030 F 1 1 1 1 1 1 0 0 0 0 0 0
PUP044 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP094 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP095 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP097 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP049 F 0 1 1 1 1 0 0 0 0 0 0 0
Appendices 68
Appendix 1. Cont.
LIVE DEAD
16 20 32 33 24
ID SEX S1 S2 S3 S4 S5 S6 S1 S2 S3 S4 S5 S6
PUP025 F 1 1 1 1 1 1 0 0 0 0 0 0
PUP068 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP039 F 0 1 1 0 0 0 0 0 1 0 0 0
PUP079 M 0 1 1 1 1 1 0 0 0 0 0 0
PUP096 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP098 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP054 F 0 1 1 1 1 1 0 0 0 0 0 0
PUP099 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP100 M 0 0 1 0 0 0 0 0 0 0 0 0
PUP101 M 0 0 1 1 1 0 0 0 0 0 0 0
PUP103 F 0 0 1 0 1 0 0 0 0 0 0 0
PUP102 U 0 0 1 1 1 1 0 0 0 0 0 0
PUP018 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP111 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP059 F 0 1 1 0 0 0 0 0 0 1 0 0
PUP104 F 0 0 1 0 0 1 0 0 0 0 0 0
PUP105 F 0 0 1 1 0 1 0 0 0 0 0 0
PUP009 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP107 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP106 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP116 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP015 M 1 1 1 1 0 0 0 0 0 0 1 0
PUP016 M 1 0 1 1 1 1 0 0 0 0 0 0
PUP119 U 0 0 1 1 0 0 0 0 0 0 0 0
PUP120 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP013 F 1 0 1 1 0 1 0 0 0 0 0 0
PUP121 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP122 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP123 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP124 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP125 F 0 0 1 1 0 1 0 0 0 0 0 0
PUP126 F 0 0 1 0 0 0 0 0 0 0 0 0
PUP127 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP071 F 0 1 1 1 1 0 0 0 0 0 0 0
PUP002 M 1 1 1 1 1 0 0 0 0 0 0 0
PUP128 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP129 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP130 M 0 0 1 1 1 0 0 0 0 0 0 0
PUP003 M 1 0 1 1 1 0 0 0 0 0 0 0
PUP131 F 0 0 1 1 0 0 0 0 0 0 0 0
PUP132 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP036 M 1 1 1 1 1 1 0 0 0 0 0 0
PUP038 M 1 1 1 1 1 0 0 0 0 0 0 0
PUP134 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP135 F 0 0 1 1 1 1 0 0 0 0 0 0
PUP136 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP137 M 0 0 1 1 1 0 0 0 0 0 0 0
Appendices 69
Appendix 1. Cont. LIVE DEAD
16 20 32 33 24
ID SEX S1 S2 S3 S4 S5 S6 S1 S2 S3 S4 S5 S6
PUP004 M 1 0 1 1 1 1 0 0 0 0 0 0
PUP138 M 0 0 1 1 1 1 0 0 0 0 0 0
PUP008 M 1 0 1 1 1 1 0 0 0 0 0 0
PUP140 F 0 0 1 0 1 1 0 0 0 0 0 0
PUP139 F 0 0 1 1 1 0 0 0 0 0 0 0
PUP001 M 1 0 1 1 1 1 0 0 0 0 0 0 PUP054 F 0 1 1 1 1 1 0 0 0 0 0 0 PUP141 M 0 0 1 1 1 1 0 0 0 0 0 0 PUP142 M 0 0 1 0 0 0 0 0 0 0 0 0 PUP143 M 0 0 1 1 1 1 0 0 0 0 0 0 PUP031 F 1 0 0 0 0 0 0 1 0 0 0 0 PUP113 M 0 0 1 0 1 1 0 0 0 0 0 0 PUP117 F 0 0 1 1 1 0 0 0 0 0 0 0 PUP118 F 0 0 1 1 1 1 0 0 0 0 0 0 PUP133 M 0 0 1 1 1 0 0 0 0 0 0 0 PUP144 F 0 0 1 0 0 0 0 0 0 0 0 0
Appendices 70
Appendix 2
Distance matrix detailing the distance (km) between Australian sea lion colonies in South and Western Australian.
Nor
th P
age
Sou
th P
age
Sea
l Slid
e
Sea
l Bay
(Kan
garo
o Is
)
Pea
ked
Roc
ks
Nor
th Is
land
Dan
gero
us R
eef
Eng
lish
Isla
nd
Alb
atro
ss Is
land
Lew
is Is
land
Sou
th N
eptu
ne (M
ain)
Is.
Nor
th N
eptu
ne (E
ast)
Is.
Ligu
anea
Isla
nd
Pric
e Is
land
Four
Hum
moc
ks (N
orth
) Is.
Roc
ky Is
land
(Nor
th)
Wes
t Wal
degr
ave
Isla
nd
Jone
s Is
land
War
d Is
land
Pea
rson
Isla
nd
Nic
olas
Bau
din
Isla
nd
North Page -35.76 138.30 0 2 75 99 240 240 240 240 240 240 239 239 239 302 319 323 438 494 446 430 520 South Page -35.78 138.29 2 0 73 97 241 241 241 241 241 241 240 240 240 303 320 324 439 495 447 431 521 Seal Slide -36.03 137.54 75 73 0 24 182 189 226 246 200 214 176 188 232 276 288 337 412 468 418 403 494 Seal Bay (Kangaroo Is) -36.00 137.33 99 97 24 0 118 169 206 226 180 194 156 168 212 256 268 317 393 448 398 383 474 Peaked Rocks -35.19 136.49 240 241 182 118 0 8 48 67 31 47 37 38 82 126 141 200 272 326 274 256 352 North Island -35.13 136.45 240 241 189 169 8 0 40 59 25 41 38 36 77 121 136 195 267 321 268 251 347 Dangerous Reef -34.82 136.22 240 241 226 206 48 40 0 20 28 25 58 48 63 107 130 189 262 316 263 245 341 English Island -34.64 136.20 240 241 246 226 67 59 20 0 48 41 77 67 79 123 146 205 278 332 279 261 357 Albatross Island -35.07 136.18 240 241 200 180 31 25 28 48 0 305 30 21 52 96 119 178 251 305 252 234 330 Lewis Island -34.98 136.03 240 241 214 194 47 41 25 41 17 0 39 28 38 82 105 164 237 291 238 220 316 South Neptune (Main) Is. -35.33 136.11 239 240 176 156 37 38 58 77 30 39 0 12 58 102 116 175 248 302 249 231 327 North Neptune (East) Is. -35.23 136.07 239 240 188 168 38 36 48 67 21 28 12 0 48 92 107 166 239 293 240 222 318 Liguanea Island -35.00 135.62 239 240 232 212 82 77 63 79 52 38 58 48 0 44 59 118 191 245 192 174 270 Price Island -34.71 135.29 302 303 276 256 126 121 107 123 96 82 102 92 44 0 23 82 155 209 156 138 234 Four Hummocks (North) Is. -34.76 135.04 319 320 288 268 141 136 130 146 119 105 116 107 59 23 0 59 132 186 133 115 211 Rocky Island (North) -34.26 135.26 323 324 337 317 200 195 189 205 178 164 175 166 118 82 59 0 87 145 107 98 173 West Waldegrave Island -33.60 134.76 438 439 412 393 272 267 262 278 251 237 248 239 191 155 132 87 0 59 47 61 87 Jones Island -33.19 134.37 494 495 468 448 326 321 316 332 305 291 302 293 245 209 186 145 59 0 62 85 29 Ward Island -33.74 134.29 446 447 418 398 274 268 263 279 252 238 249 240 192 156 133 107 47 62 0 23 82 Pearson Island -33.95 134.26 430 431 403 383 256 251 245 261 234 220 231 222 174 138 115 98 61 85 23 0 104 Nicolas Baudin Island -33.02 134.13 520 521 494 474 352 347 341 357 330 316 327 318 270 234 211 173 87 29 82 104 0 Olive Island -32.72 133.97 556 557 530 511 388 383 378 394 367 353 364 355 307 271 247 209 122 64 117 139 36 East Franklin Reef -32.45 133.67 596 597 570 550 427 422 417 433 406 392 403 394 346 310 287 250 163 105 155 176 77 West Franklin Reef -32.46 133.64 596 597 570 550 427 422 417 433 406 392 403 394 346 310 286 250 164 105 154 175 77 Breakwater Island -32.32 133.56 613 614 587 567 444 439 434 450 423 409 420 411 363 327 304 267 181 122 172 192 94 Gliddon Reef -32.32 133.56 613 614 587 567 444 439 434 450 423 409 420 411 363 327 303 267 180 122 171 192 94 Lounds Island -32.27 133.37 627 628 601 581 458 453 447 463 436 422 433 424 376 340 317 282 197 138 184 204 109 Fenelon Island -32.58 133.28 603 604 577 557 433 427 422 438 411 397 408 399 351 315 292 262 178 122 159 177 93 Masillon Island -32.56 133.28 605 606 579 559 435 430 424 440 413 399 410 401 353 317 294 264 180 123 161 179 94 West Island -32.51 133.25 611 612 585 565 441 436 430 446 419 405 416 407 359 323 300 269 185 128 167 185 100 Purdie Island -32.27 133.23 635 636 609 589 464 459 454 470 443 429 440 431 383 347 324 291 205 147 191 210 119 Western Nuyts Reef -32.12 132.13 713 714 687 667 540 535 529 545 518 504 515 506 458 422 399 376 296 241 270 284 212
Appendices 71
Appendix 2. Cont.
Nor
th P
age
Sou
th P
age
Sea
l Slid
e
Sea
l Bay
(Kan
garo
o Is
)
Pea
ked
Roc
ks
Nor
th Is
land
Dan
gero
us R
eef
Eng
lish
Isla
nd
Alb
atro
ss Is
land
Lew
is Is
land
Sou
th N
eptu
ne (M
ain)
Is.
Nor
th N
eptu
ne (E
ast)
Is.
Ligu
anea
Isla
nd
Pric
e Is
land
Four
Hum
moc
ks (N
orth
) Is.
Roc
ky Is
land
(Nor
th)
Wes
t Wal
degr
ave
Isla
nd
Jone
s Is
land
War
d Is
land
Pea
rson
Isla
nd
Nic
olas
Bau
din
Isla
nd
Bunda Cliffs B1 -31.49 131.07 833 834 807 788 659 654 649 665 638 624 635 626 578 542 519 498 418 363 392 405 334 Bunda Cliffs B2 -31.59 130.58 860 861 834 814 685 680 674 690 663 649 660 651 603 567 544 528 450 397 421 432 369 Bunda Cliffs B3 -31.58 130.15 893 894 867 847 717 712 706 723 696 682 693 684 635 600 576 562 487 434 456 466 407 Bunda Cliffs B5 -31.59 130.05 900 901 874 854 724 719 713 729 702 688 699 690 642 606 583 570 495 443 463 473 415 Bunda Cliffs B6 -31.61 129.77 921 922 895 875 744 739 733 749 722 708 719 710 662 626 603 591 517 466 485 494 439 Bunda Cliffs B8 -31.64 129.38 949 950 923 903 771 766 760 777 749 735 747 737 689 654 630 621 549 498 515 523 472 Bunda Cliffs B9 -31.65 129.30 955 956 929 909 777 772 767 783 756 742 753 744 696 660 636 627 555 505 521 530 479 Bunda Cliffs B10 -32.28 126.01 1205 1206 1179 1159 1022 1017 1011 1028 1001 987 998 989 940 905 881 888 830 788 788 790 765 Spindle Island -33.76 124.16 1334 1335 1308 1288 1147 1142 1136 1152 1125 1111 1122 1113 1065 1029 1006 1025 981 949 936 933 930 Ford (Halfway) Island -33.77 124.04 1345 1346 1319 1299 1158 1153 1147 1163 1136 1122 1133 1124 1076 1040 1017 1036 992 960 947 944 941 Six Mile Island -33.64 123.97 1354 1355 1327 1308 1167 1162 1156 1172 1145 1131 1142 1133 1085 1049 1026 1044 999 967 955 952 947 Round Island -34.11 123.89 1354 1355 1328 1308 1166 1161 1156 1172 1145 1131 1142 1133 1085 1049 1026 1046 1006 975 960 956 957 Cooper Island -34.23 123.61 1379 1380 1352 1332 1190 1185 1180 1196 1169 1155 1166 1157 1109 1073 1050 1071 1032 1002 986 982 984 Salisbury Island -34.36 123.55 1382 1383 1356 1336 1194 1189 1183 1199 1172 1158 1169 1160 1112 1076 1053 1075 1037 1008 991 987 990 Poison Creek Island -33.92 123.33 1408 1409 1382 1362 1220 1215 1210 1226 1199 1185 1196 1187 1139 1103 1079 1099 1057 1026 1012 1009 1007 Stanley (Wickham) Is. -34.02 123.29 1410 1411 1384 1364 1222 1217 1212 1228 1201 1187 1198 1189 1141 1105 1081 1102 1061 1030 1015 1012 1011 Glennie Island -34.10 123.11 1426 1427 1400 1380 1238 1233 1227 1244 1217 1203 1214 1205 1156 1121 1097 1118 1078 1047 1032 1028 1029 SW Rock (Twin Peaks Is.) -33.98 122.90 1446 1447 1420 1400 1258 1253 1248 1264 1237 1223 1234 1225 1177 1141 1118 1138 1097 1066 1052 1048 1047 Taylor Island -33.92 122.87 1450 1451 1423 1404 1262 1257 1251 1268 1240 1226 1238 1228 1180 1144 1121 1141 1100 1068 1054 1051 1049 Kermadec (Wedge) Is. -34.09 122.83 1451 1452 1425 1405 1263 1258 1252 1269 1241 1227 1239 1229 1181 1146 1122 1143 1103 1072 1057 1053 1054 Kimberley Island -33.95 122.47 1486 1487 1460 1440 1298 1293 1288 1304 1277 1263 1274 1265 1217 1181 1158 1178 1137 1106 1091 1088 1087 MacKenzie Island -34.20 122.11 1516 1517 1490 1470 1328 1323 1317 1333 1306 1292 1303 1294 1246 1210 1187 1209 1169 1139 1124 1119 1121 Little Island -34.46 121.99 1525 1526 1498 1478 1336 1331 1325 1341 1314 1300 1311 1302 1254 1218 1195 1218 1181 1152 1135 1130 1134 Rocky Island -34.08 120.87 1632 1633 1605 1585 1443 1438 1433 1449 1422 1408 1419 1410 1362 1326 1302 1324 1284 1254 1239 1235 1235 West Island -34.08 120.49 1667 1668 1640 1620 1478 1473 1468 1484 1457 1443 1454 1445 1397 1361 1337 1359 1320 1289 1274 1270 1270 Red Islet -34.04 119.78 1732 1733 1706 1686 1543 1538 1533 1549 1522 1508 1519 1510 1462 1426 1403 1425 1385 1354 1339 1335 1335 Middle Doubtful Is. -34.37 119.61 1744 1745 1717 1698 1555 1550 1544 1560 1533 1519 1530 1521 1473 1437 1414 1438 1400 1371 1354 1349 1352 Haul Off Rock -34.70 118.66 1828 1829 1801 1781 1638 1633 1627 1643 1616 1602 1613 1604 1556 1520 1497 1522 1487 1459 1441 1435 1441 Buller Island -30.66 115.12 2616 2617 2589 2569 2426 2421 2415 2431 2404 2390 2401 2392 2344 2308 2285 2310 2275 2247 2229 2223 2229 North Fisherman Is. -30.13 114.95 2674 2675 2647 2627 2484 2479 2473 2489 2462 2448 2459 2450 2402 2366 2343 2368 2333 2305 2287 2281 2287 Beagle Island -29.81 114.88 2710 2711 2683 2663 2520 2515 2509 2525 2498 2484 2495 2486 2438 2402 2379 2404 2369 2341 2323 2317 2323 Abrolhos-Southern Group -28.90 113.94 2820 2821 2793 2773 2630 2625 2619 2635 2608 2594 2605 2596 2548 2512 2489 2514 2479 2451 2433 2427 2433 Abrolhos -Easter Group -28.67 113.82 2848 2849 2821 2801 2658 2653 2647 2663 2636 2622 2633 2624 2576 2540 2517 2542 2507 2479 2461 2455 2461
Appendices 72
Appendix 2. Cont.
Oliv
e Is
land
Eas
t Fra
nklin
Ree
f
Wes
t Fra
nklin
Ree
f
Bre
akw
ater
Isla
nd
Glid
don
Ree
f
Loun
ds Is
land
Fene
lon
Isla
nd
Mas
illon
Isla
nd
Wes
t Isl
and
Pur
die
Isla
nd
Wes
tern
Nuy
ts R
eef
Bun
da C
liffs
B1
Bun
da C
liffs
B2
Bun
da C
liffs
B3
Bun
da C
liffs
B5
Bun
da C
liffs
B6
Bun
da C
liffs
B8
Bun
da C
liffs
B9
North Page -35.76 138.30 556 596 596 613 613 627 603 605 611 635 713 833 860 893 900 921 949 955 South Page -35.78 138.29 557 597 597 614 614 628 604 606 612 636 714 834 861 894 901 922 950 956 Seal Slide -36.03 137.54 530 570 570 587 587 601 577 579 585 609 687 807 834 867 874 895 923 929 Seal Bay (Kangaroo Is) -36.00 137.33 511 550 550 567 567 581 557 559 565 589 667 788 814 847 854 875 903 909 Peaked Rocks -35.19 136.49 388 427 427 444 444 458 433 435 441 464 540 659 685 717 724 744 771 777 North Island -35.13 136.45 383 422 422 439 439 453 427 430 436 459 535 654 680 712 719 739 766 772 Dangerous Reef -34.82 136.22 378 417 417 434 434 447 422 424 430 454 529 649 674 706 713 733 760 767 English Island -34.64 136.20 394 433 433 450 450 463 438 440 446 470 545 665 690 723 729 749 777 783 Albatross Island -35.07 136.18 367 406 406 423 423 436 411 413 419 443 518 638 663 696 702 722 749 756 Lewis Island -34.98 136.03 353 392 392 409 409 422 397 399 405 429 504 624 649 682 688 708 735 742 South Neptune (Main) Is. -35.33 136.11 364 403 403 420 420 433 408 410 416 440 515 635 660 693 699 719 747 753 North Neptune (East) Is. -35.23 136.07 355 394 394 411 411 424 399 401 407 431 506 626 651 684 690 710 737 744 Liguanea Island -35.00 135.62 307 346 346 363 363 376 351 353 359 383 458 578 603 635 642 662 689 696 Price Island -34.71 135.29 271 310 310 327 327 340 315 317 323 347 422 542 567 600 606 626 654 660 Four Hummocks (North) Is. -34.76 135.04 247 287 286 304 303 317 292 294 300 324 399 519 544 576 583 603 630 636 Rocky Island (North) -34.26 135.26 209 250 250 267 267 282 262 264 269 291 376 498 528 562 570 591 621 627 West Waldegrave Island -33.60 134.76 122 163 164 181 180 197 178 180 185 205 296 418 450 487 495 517 549 555 Jones Island -33.19 134.37 64 105 105 122 122 138 122 123 128 147 241 363 397 434 443 466 498 505 Ward Island -33.74 134.29 117 155 154 172 171 184 159 161 167 191 270 392 421 456 463 485 515 521 Pearson Island -33.95 134.26 139 176 175 192 192 204 177 179 185 210 284 405 432 466 473 494 523 530 Nicolas Baudin Island -33.02 134.13 36 77 77 94 94 109 93 94 100 119 212 334 369 407 415 439 472 479 Olive Island -32.72 133.97 0 41 42 58 58 75 66 67 71 86 185 306 342 381 390 414 448 455 East Franklin Reef -32.45 133.67 41 0 3 17 17 34 39 38 40 46 149 267 306 346 354 379 414 421 West Franklin Reef -32.46 133.64 42 3 0 17 17 33 36 35 37 44 147 266 304 343 352 377 411 419 Breakwater Island -32.32 133.56 58 17 17 0 0 19 39 37 36 32 136 253 292 332 341 367 401 409 Gliddon Reef -32.32 133.56 58 17 17 0 0 19 39 37 36 32 137 253 293 333 341 367 402 409 Lounds Island -32.27 133.37 75 34 33 19 19 0 35 33 29 13 117 234 273 313 322 348 382 390 Fenelon Island -32.58 133.28 66 39 36 39 39 35 0 2 8 35 120 241 277 315 324 348 382 389 Masillon Island -32.56 133.28 67 38 35 37 37 33 2 0 6 32 119 240 276 314 323 348 381 388 West Island -32.51 133.25 71 40 37 36 36 29 8 6 0 27 114 235 271 310 319 343 377 384 Purdie Island -32.27 133.23 86 46 44 32 32 13 35 32 27 0 105 222 261 300 309 335 369 377 Western Nuyts Reef -32.12 132.13 185 149 147 136 137 117 120 119 114 105 0 122 157 197 205 230 265 272
Appendices 73
Appendix 2. Cont.
Oliv
e Is
land
Eas
t Fra
nklin
Ree
f
Wes
t Fra
nklin
Ree
f
Bre
akw
ater
Isla
nd
Glid
don
Ree
f
Loun
ds Is
land
Fene
lon
Isla
nd
Mas
illon
Isla
nd
Wes
t Isl
and
Pur
die
Isla
nd
Wes
tern
Nuy
ts R
eef
Bun
da C
liffs
B1
Bun
da C
liffs
B2
Bun
da C
liffs
B3
Bun
da C
liffs
B5
Bun
da C
liffs
B6
Bun
da C
liffs
B8
Bun
da C
liffs
B9
Bunda Cliffs B1 -31.49 131.07 306 267 266 253 253 234 241 240 235 222 122 0 47 87 97 124 160 168 Bunda Cliffs B2 -31.59 130.58 342 306 304 292 293 273 277 276 271 261 157 47 0 41 50 77 114 122 Bunda Cliffs B3 -31.58 130.15 381 346 343 332 333 313 315 314 310 300 197 87 41 0 10 36 73 81 Bunda Cliffs B5 -31.59 130.05 390 354 352 341 341 322 324 323 319 309 205 97 50 10 0 27 63 71 Bunda Cliffs B6 -31.61 129.77 414 379 377 367 367 348 348 348 343 335 230 124 77 36 27 0 37 44 Bunda Cliffs B8 -31.64 129.38 448 414 411 401 402 382 382 381 377 369 265 160 114 73 63 37 0 8 Bunda Cliffs B9 -31.65 129.30 455 421 419 409 409 390 389 388 384 377 272 168 122 81 71 44 8 0 Bunda Cliffs B10 -32.28 126.01 748 719 717 710 710 691 683 683 680 678 576 485 438 398 389 362 326 318 Spindle Island -33.76 124.16 920 898 895 891 891 874 859 859 858 862 766 694 648 611 602 577 543 535 Ford (Halfway) Island -33.77 124.04 931 909 906 902 902 885 870 871 869 873 777 705 658 621 612 587 553 546 Six Mile Island -33.64 123.97 936 914 911 907 907 890 875 876 874 877 781 707 660 623 614 588 554 546 Round Island -34.11 123.89 948 928 925 922 922 905 889 889 888 893 799 731 685 649 640 615 582 575 Cooper Island -34.23 123.61 976 955 952 949 950 933 917 917 916 921 827 760 714 678 669 645 611 604 Salisbury Island -34.36 123.55 983 963 960 957 957 941 924 925 923 929 836 771 724 689 680 655 622 615 Poison Creek Island -33.92 123.33 998 976 973 969 969 953 937 938 936 940 845 773 726 689 680 654 620 612 Stanley (Wickham) Is. -34.02 123.29 1002 981 978 974 975 958 942 943 941 945 850 780 733 696 687 662 628 620 Glennie Island -34.10 123.11 1020 999 996 993 993 976 960 961 959 964 869 799 752 715 706 681 647 639 SW Rock (Twin Peaks Is.) -33.98 122.90 1038 1016 1013 1010 1010 993 978 978 976 981 885 813 766 729 719 694 659 652 Taylor Island -33.92 122.87 1040 1018 1015 1011 1011 994 979 980 978 982 886 813 766 729 719 694 659 652 Kermadec (Wedge) Is. -34.09 122.83 1045 1024 1021 1017 1017 1001 985 985 984 988 893 822 775 738 729 704 669 662 Kimberley Island -33.95 122.47 1077 1055 1052 1049 1049 1032 1017 1017 1016 1019 923 849 803 765 756 730 695 688 MacKenzie Island -34.20 122.11 1112 1091 1088 1085 1085 1068 1053 1053 1052 1056 961 889 842 805 796 771 736 728 Little Island -34.46 121.99 1126 1106 1103 1100 1100 1084 1067 1068 1067 1072 978 909 862 825 816 791 757 749 Rocky Island -34.08 120.87 1226 1204 1201 1197 1197 1181 1165 1166 1164 1168 1072 996 949 911 902 876 840 833 West Island -34.08 120.49 1261 1239 1236 1232 1233 1216 1201 1201 1199 1203 1107 1030 983 945 936 910 874 867 Red Islet -34.04 119.78 1326 1304 1301 1297 1297 1280 1265 1266 1264 1267 1170 1093 1046 1007 998 972 936 928 Middle Doubtful Is. -34.37 119.61 1344 1323 1320 1316 1316 1300 1284 1284 1283 1287 1191 1117 1069 1031 1022 996 961 953 Haul Off Rock -34.70 118.66 1433 1413 1410 1407 1407 1390 1374 1374 1373 1378 1283 1210 1163 1125 1115 1090 1054 1047 Buller Island -30.66 115.12 2221 2201 2198 2195 2195 2178 2162 2162 2161 2166 2071 1998 1951 1913 1903 1878 1842 1835 North Fisherman Is. -30.13 114.95 2279 2259 2256 2253 2253 2236 2220 2220 2219 2224 2129 2056 2009 1971 1961 1936 1900 1893 Beagle Island -29.81 114.88 2315 2295 2292 2289 2289 2272 2256 2256 2255 2260 2165 2092 2045 2007 1997 1972 1936 1929 Abrolhos-Southern Group -28.90 113.94 2425 2405 2402 2399 2399 2382 2366 2366 2365 2370 2275 2202 2155 2117 2107 2082 2046 2039 Abrolhos -Easter Group -28.67 113.82 2453 2433 2430 2427 2427 2410 2394 2394 2393 2398 2303 2230 2183 2145 2135 2110 2074 2067
Appendices 74
Appendix 2. Cont.
Bun
da C
liffs
B10
Spi
ndle
Isla
nd
Ford
(Hal
fway
) Isl
and
Six
Mile
Isla
nd
Rou
nd Is
land
Coo
per I
slan
d
Sal
isbu
ry Is
land
Poi
son
Cre
ek Is
land
Sta
nley
(Wic
kham
) Is.
Gle
nnie
Isla
nd
SW
Roc
k (T
win
Pea
ks Is
.)
Tayl
or Is
land
Ker
mad
ec (W
edge
) Is.
Kim
berle
y Is
land
Mac
Ken
zie
Isla
nd
Littl
e Is
land
Roc
ky Is
land
Wes
t Isl
and
Red
Isle
t
Mid
dle
Dou
btfu
l Is.
Hau
l Off
Roc
k
Bul
ler I
slan
d
Nor
th F
ishe
rman
Is.
Bea
gle
Isla
nd
Abr
olho
s-So
uthe
rn G
roup
Abr
olho
s-Ea
ster
Gro
up
North Page -35.76 138.30 1205 1334 1345 1354 1354 1379 1382 1408 1410 1426 1446 1450 1451 1486 1516 1525 1632 1667 1732 1744 1828 2616 2674 2710 2820 2848 South Page -35.78 138.29 1206 1335 1346 1355 1355 1380 1383 1409 1411 1427 1447 1451 1452 1487 1517 1526 1633 1668 1733 1745 1829 2617 2675 2711 2821 2849 Seal Slide -36.03 137.54 1179 1308 1319 1327 1328 1352 1356 1382 1384 1400 1420 1423 1425 1460 1490 1498 1605 1640 1706 1717 1801 2589 2647 2683 2793 2821 Seal Bay (Kangaroo Is) -36.00 137.33 1159 1288 1299 1308 1308 1332 1336 1362 1364 1380 1400 1404 1405 1440 1470 1478 1585 1620 1686 1698 1781 2569 2627 2663 2773 2801 Peaked Rocks -35.19 136.49 1022 1147 1158 1167 1166 1190 1194 1220 1222 1238 1258 1262 1263 1298 1328 1336 1443 1478 1543 1555 1638 2426 2484 2520 2630 2658 North Island -35.13 136.45 1017 1142 1153 1162 1161 1185 1189 1215 1217 1233 1253 1257 1258 1293 1323 1331 1438 1473 1538 1550 1633 2421 2479 2515 2625 2653 Dangerous Reef -34.82 136.22 1011 1136 1147 1156 1156 1180 1183 1210 1212 1227 1248 1251 1252 1288 1317 1325 1433 1468 1533 1544 1627 2415 2473 2509 2619 2647 English Island -34.64 136.20 1028 1152 1163 1172 1172 1196 1199 1226 1228 1244 1264 1268 1269 1304 1333 1341 1449 1484 1549 1560 1643 2431 2489 2525 2635 2663 Albatross Island -35.07 136.18 1001 1125 1136 1145 1145 1169 1172 1199 1201 1217 1237 1240 1241 1277 1306 1314 1422 1457 1522 1533 1616 2404 2462 2498 2608 2636 Lewis Island -34.98 136.03 987 1111 1122 1131 1131 1155 1158 1185 1187 1203 1223 1226 1227 1263 1292 1300 1408 1443 1508 1519 1602 2390 2448 2484 2594 2622 South Neptune (Main) Is. -35.33 136.11 998 1122 1133 1142 1142 1166 1169 1196 1198 1214 1234 1238 1239 1274 1303 1311 1419 1454 1519 1530 1613 2401 2459 2495 2605 2633 North Neptune (East) Is. -35.23 136.07 989 1113 1124 1133 1133 1157 1160 1187 1189 1205 1225 1228 1229 1265 1294 1302 1410 1445 1510 1521 1604 2392 2450 2486 2596 2624 Liguanea Island -35.00 135.62 940 1065 1076 1085 1085 1109 1112 1139 1141 1156 1177 1180 1181 1217 1246 1254 1362 1397 1462 1473 1556 2344 2402 2438 2548 2576 Price Island -34.71 135.29 905 1029 1040 1049 1049 1073 1076 1103 1105 1121 1141 1144 1146 1181 1210 1218 1326 1361 1426 1437 1520 2308 2366 2402 2512 2540 Four Hummocks (North) Is. -34.76 135.04 881 1006 1017 1026 1026 1050 1053 1079 1081 1097 1118 1121 1122 1158 1187 1195 1302 1337 1403 1414 1497 2285 2343 2379 2489 2517 Rocky Island (North) -34.26 135.26 888 1025 1036 1044 1046 1071 1075 1099 1102 1118 1138 1141 1143 1178 1209 1218 1324 1359 1425 1438 1522 2310 2368 2404 2514 2542 West Waldegrave Island -33.60 134.76 830 981 992 999 1006 1032 1037 1057 1061 1078 1097 1100 1103 1137 1169 1181 1284 1320 1385 1400 1487 2275 2333 2369 2479 2507 Jones Island -33.19 134.37 788 949 960 967 975 1002 1008 1026 1030 1047 1066 1068 1072 1106 1139 1152 1254 1289 1354 1371 1459 2247 2305 2341 2451 2479 Ward Island -33.74 134.29 788 936 947 955 960 986 991 1012 1015 1032 1052 1054 1057 1091 1124 1135 1239 1274 1339 1354 1441 2229 2287 2323 2433 2461 Pearson Island -33.95 134.26 790 933 944 952 956 982 987 1009 1012 1028 1048 1051 1053 1088 1119 1130 1235 1270 1335 1349 1435 2223 2281 2317 2427 2455 Nicolas Baudin Island -33.02 134.13 765 930 941 947 957 984 990 1007 1011 1029 1047 1049 1054 1087 1121 1134 1235 1270 1335 1352 1441 2229 2287 2323 2433 2461 Olive Island -32.72 133.97 748 920 931 936 948 976 983 998 1002 1020 1038 1040 1045 1077 1112 1126 1226 1261 1326 1344 1433 2221 2279 2315 2425 2453 East Franklin Reef -32.45 133.67 719 898 909 914 928 955 963 976 981 999 1016 1018 1024 1055 1091 1106 1204 1239 1304 1323 1413 2201 2259 2295 2405 2433 West Franklin Reef -32.46 133.64 717 895 906 911 925 952 960 973 978 996 1013 1015 1021 1052 1088 1103 1201 1236 1301 1320 1410 2198 2256 2292 2402 2430 Breakwater Island -32.32 133.56 710 891 902 907 922 949 957 969 974 993 1010 1011 1017 1049 1085 1100 1197 1232 1297 1316 1407 2195 2253 2289 2399 2427 Gliddon Reef -32.32 133.56 710 891 902 907 922 950 957 969 975 993 1010 1011 1017 1049 1085 1100 1197 1233 1297 1316 1407 2195 2253 2289 2399 2427 Lounds Island -32.27 133.37 691 874 885 890 905 933 941 953 958 976 993 994 1001 1032 1068 1084 1181 1216 1280 1300 1390 2178 2236 2272 2382 2410 Fenelon Island -32.58 133.28 683 859 870 875 889 917 924 937 942 960 978 979 985 1017 1053 1067 1165 1201 1265 1284 1374 2162 2220 2256 2366 2394 Masillon Island -32.56 133.28 683 859 871 876 889 917 925 938 943 961 978 980 985 1017 1053 1068 1166 1201 1266 1284 1374 2162 2220 2256 2366 2394 West Island -32.51 133.25 680 858 869 874 888 916 923 936 941 959 976 978 984 1016 1052 1067 1164 1199 1264 1283 1373 2161 2219 2255 2365 2393 Purdie Island -32.27 133.23 678 862 873 877 893 921 929 940 945 964 981 982 988 1019 1056 1072 1168 1203 1267 1287 1378 2166 2224 2260 2370 2398 Western Nuyts Reef -32.12 132.13 576 766 777 781 799 827 836 845 850 869 885 886 893 923 961 978 1072 1107 1170 1191 1283 2071 2129 2165 2275 2303
Appendices 75
Appendix 2. Cont.
Bun
da C
liffs
B10
Spi
ndle
Isla
nd
Ford
(Hal
fway
) Isl
and
Six
Mile
Isla
nd
Rou
nd Is
land
Coo
per I
slan
d
Sal
isbu
ry Is
land
Poi
son
Cre
ek Is
land
Sta
nley
(Wic
kham
) Is.
Gle
nnie
Isla
nd
SW
Roc
k (T
win
Pea
ks Is
.)
Tayl
or Is
land
Ker
mad
ec (W
edge
) Is.
Kim
berle
y Is
land
Mac
Ken
zie
Isla
nd
Littl
e Is
land
Roc
ky Is
land
Wes
t Isl
and
Red
Isle
t
Mid
dle
Dou
btfu
l Is.
Hau
l Off
Roc
k
Bul
ler I
slan
d
Nor
th F
ishe
rman
Is.
Bea
gle
Isla
nd
Abr
olho
s-So
uthe
rn G
roup
Abr
olho
s-Ea
ster
Gro
up
Bunda Cliffs B1 -31.49 131.07 485 694 705 707 731 760 771 773 780 799 813 813 822 849 889 909 996 1030 1093 1117 1210 1998 2056 2092 2202 2230 Bunda Cliffs B2 -31.59 130.58 438 648 658 660 685 714 724 726 733 752 766 766 775 803 842 862 949 983 1046 1069 1163 1951 2009 2045 2155 2183 Bunda Cliffs B3 -31.58 130.15 398 611 621 623 649 678 689 689 696 715 729 729 738 765 805 825 911 945 1007 1031 1125 1913 1971 2007 2117 2145 Bunda Cliffs B5 -31.59 130.05 389 602 612 614 640 669 680 680 687 706 719 719 729 756 796 816 902 936 998 1022 1115 1903 1961 1997 2107 2135 Bunda Cliffs B6 -31.61 129.77 362 577 587 588 615 645 655 654 662 681 694 694 704 730 771 791 876 910 972 996 1090 1878 1936 1972 2082 2110 Bunda Cliffs B8 -31.64 129.38 326 543 553 554 582 611 622 620 628 647 659 659 669 695 736 757 840 874 936 961 1054 1842 1900 1936 2046 2074 Bunda Cliffs B9 -31.65 129.30 318 535 546 546 575 604 615 612 620 639 652 652 662 688 728 749 833 867 928 953 1047 1835 1893 1929 2039 2067 Bunda Cliffs B10 -32.28 126.01 0 239 247 243 283 312 325 309 319 338 346 345 358 379 421 445 519 552 612 639 733 1521 1579 1615 1725 1753 Spindle Island -33.76 124.16 239 0 11 22 46 73 87 79 85 104 119 120 128 158 195 214 306 341 406 425 516 1304 1362 1398 1508 1536 Ford (Halfway) Island -33.77 124.04 247 11 0 16 40 65 80 68 75 94 108 109 117 147 184 204 295 330 394 414 505 1293 1351 1387 1497 1525 Six Mile Island -33.64 123.97 243 22 16 0 52 74 89 67 75 94 106 106 116 143 182 204 291 325 389 410 502 1290 1348 1384 1494 1522 Round Island -34.11 123.89 283 46 40 52 0 29 42 56 56 72 92 96 97 132 164 179 278 313 378 395 484 1272 1330 1366 1476 1504 Cooper Island -34.23 123.61 312 73 65 74 29 0 15 43 37 49 71 76 73 109 138 151 253 288 353 368 456 1244 1302 1338 1448 1476 Salisbury Island -34.36 123.55 325 87 80 89 42 15 0 53 45 51 73 79 73 110 134 144 249 284 349 362 450 1238 1296 1332 1442 1470 Poison Creek Island -33.92 123.33 309 79 68 67 56 43 53 0 12 29 40 42 50 80 117 137 228 263 328 346 438 1226 1284 1320 1430 1458 Stanley (Wickham) Is. -34.02 123.29 319 85 75 75 56 37 45 12 0 19 36 40 43 76 110 129 223 259 324 341 432 1220 1278 1314 1424 1452 Glennie Island -34.10 123.11 338 104 94 94 72 49 51 29 19 0 23 29 25 61 92 110 206 241 306 323 413 1201 1259 1295 1405 1433 SW Rock (Twin Peaks Is.) -33.98 122.90 346 119 108 106 92 71 73 40 36 23 0 7 13 40 77 99 188 223 288 306 397 1185 1243 1279 1389 1417 Taylor Island -33.92 122.87 345 120 109 106 96 76 79 42 40 29 7 0 19 37 77 101 186 221 286 305 397 1185 1243 1279 1389 1417 Kermadec (Wedge) Is. -34.09 122.83 358 128 117 116 97 73 73 50 43 25 13 19 0 37 68 88 181 216 281 298 389 1177 1235 1271 1381 1409 Kimberley Island -33.95 122.47 379 158 147 143 132 109 110 80 76 61 40 37 37 0 43 72 148 183 248 268 360 1148 1206 1242 1352 1380 MacKenzie Island -34.20 122.11 421 195 184 182 164 138 134 117 110 92 77 77 68 43 0 31 115 150 215 231 321 1109 1167 1203 1313 1341 Little Island -34.46 121.99 445 214 204 204 179 151 144 137 129 110 99 101 88 72 31 0 111 144 208 219 306 1094 1152 1188 1298 1326 Rocky Island -34.08 120.87 519 306 295 291 278 253 249 228 223 206 188 186 181 148 115 111 0 35 100 120 214 1002 1060 1096 1206 1234 West Island -34.08 120.49 552 341 330 325 313 288 284 263 259 241 223 221 216 183 150 144 35 0 65 87 181 969 1027 1063 1173 1201 Red Islet -34.04 119.78 612 406 394 389 378 353 349 328 324 306 288 286 281 248 215 208 100 65 0 40 126 914 972 1008 1118 1146 Middle Doubtful Is. -34.37 119.61 639 425 414 410 395 368 362 346 341 323 306 305 298 268 231 219 120 87 40 0 94 882 940 976 1086 1114 Haul Off Rock -34.70 118.66 733 516 505 502 484 456 450 438 432 413 397 397 389 360 321 306 214 181 126 94 0 788 846 882 992 1020 Buller Island -30.66 115.12 1521 1304 1293 1290 1272 1244 1238 1226 1220 1201 1185 1185 1177 1148 1109 1094 1002 969 914 882 788 0 61 97 225 254 North Fisherman Is. -30.13 114.95 1579 1362 1351 1348 1330 1302 1296 1284 1278 1259 1243 1243 1235 1206 1167 1152 1060 1027 972 940 846 61 0 36 167 196 Beagle Island -29.81 114.88 1615 1398 1387 1384 1366 1338 1332 1320 1314 1295 1279 1279 1271 1242 1203 1188 1096 1063 1008 976 882 97 36 0 135 163 Abrolhos-Southern Group -28.90 113.94 1725 1508 1497 1494 1476 1448 1442 1430 1424 1405 1389 1389 1381 1352 1313 1298 1206 1173 1118 1086 992 225 167 135 0 29 Abrolhos -Easter Group -28.67 113.82 1753 1536 1525 1522 1504 1476 1470 1458 1452 1433 1417 1417 1409 1380 1341 1326 1234 1201 1146 1114 1020 254 196 163 29 0