Final Report No. 13/45

Australian Marine JVIammal Centre

Season 2004/15

• Title - Maintaining the monitoring of pup production at key Australian sea lion colonies in South Australia (2014/15). • Chief Investigator - Prof Simon Goldsworthy • Organisation - SARDI Aquatic Sciences

Activity Period - April 2014 to May 2015

Table of contents

1. Activity Summary 2. The Outcomes/Objectives 3. Appropriateness 4. Effectiveness

1. Activity Summary

A clear summary of approximately 500 words outlining the work undertaken and any significant findings (for publication on the Department's web site)

This project undertook the first contemporaneous survey over 1-2 breeding seasons of Australian sea lion (ASL) pup abundance, providing the first base-line estimate ofpup abundance across the species South Australia (SA) range. In all, over 120 individual surveys across 54 known breeding sites were conducted over the 18 month period of the project. A total of 2,417 pups were recorded at 44 sites, including two newly identified breeding sites located at Curta Rocks (7 pups) and Williams Island (5 pups). For five sites, surveys were conducted outside the breeding season, so earlier survey estimates were used. For colonies along the Bunda Cliffs where cliff-top surveys underestimate pup production, adjustments based on verified cliff-top and ground surveys were incorporated. With these adjustments, the current total pup abundance for ASL breeding sites in SA was estimated to be 2,619. Based on available survey data available, primarily between 2004 and 2008 (although data for some colonies was from as early as 1990), Shaughnessy et al. (2011) estimated ASL pup abundance in SA to be 3,119. However, this figure did not include five newly identified breeding colonies and three sites which were poorly surveyed and have since been comprehensively surveyed. This adds a minimum of 175 pups to the assessment published in 2011 and leads to a total of 3,294. The 2014/15 estimate of 2,619 is therefore 20% lower (675 pups) than the 2011 assessment.

For 54 ASL sites where current or previous historic pup production has been reported, a subset of 32 sites provide at least two comparable surveys undertaken at least six years apart (mean interval 14.5 years, range 6-44.5 years), enabling trends in abundance to be estimated. These sites account for 82% (2,157) of the estimated SA pup abundance in 2014/15, and include eight key ASL monitoring sites from four regions: Seal Bay and Seal Slide (Kangaroo Island); Dangerous Reef and English Island (Spencer Gulf); Olive and Jones Islands (Chain of Bays) and Lilliput and Blefuscu Islands (Nuyts Archipelago). Analyses of historic time-series indicates that pup numbers at most breeding sites (78%) and regions have declined between the comparable survey periods. The overall rate of decline is 3.7% per year, or 5.5% per breeding cycle, but the rate of decline varied considerably between sites and regions. Following IUCN Redlist assessment methods, and using a mean generation time of 12.6 years, the change in pup abundance over three generations (37.8 years) was extrapolated based on the observed change in pup abundance with the exponential rate of change assumed to be constant. Following this approach, the total percentage change in estimated pup abundance over three generations was estimated to be -69.7%. This global assessment approach based on 82% of the estimated SA ASL pup production, suggests the species has declined by -70% in 3 generations, meeting the IUCN Endangered category based on criteria A2(b) and (d).

The regional analysis indicates that declines have occurred across the range of the species in SA. The extent of these declines at individual colonies over the last 3 generations is estimated to range between 22% and 85%. Western populations offEyre Peninsula (Nuyts Archipelago and Chain of Bays) and in the GAB (Bunda Cliffs) have shown the greatest rates of decline, including >50% decline in pup abundance over the last ten years for breeding sites in the western Nuyts Archipelago. Other major colonies off the western Eyre Peninsula at Olive, Nicolas Baudin and West Waldegrave Islands have undergone a -37% decline in pup abundance since surveys in the early to late 2000s.

An evaluation of survey methodologies and strategies was undertaken and recommendations on future monitoring strategies have been proposed for the species. More details will be provided in a peer-reviewed SARDI report on the project, which is currently being drafted.

2. The Outcomes/Objectives

List of the Project Objectives

The objectives of the project were to: 1. Continue to provide data on the status and trends in abundance of Australian sea lions (ASL) at key monitoring sites in SA, as well as a number of other locations where information on breeding status and trends in abundance are poor. 2. Archive the first State-wide survey of all ASL breeding sites within an 18m period, providing a contemporary assessment of the species status, and analyse historic time-series data on all sites to provide an assessment of their trends in abundance. 3. Review and evaluate sites, survey methodologies and strategies developed over the last decade. Based on these analyses, provide options and make recommendations on future monitoring strategies for the species. The degree to which the Activity has achieved each of the objectives

Obiective 1: Continue to provide data on the status and trends in abundance ofASL at key monitoring sites in SA, as well as a number of other locations where information on breeding status and trends in abundance are poor.

This project continued monitoring pup production and trends in abundance at eight key ASL monitoring sites from four regions across SA. These include: Seal Bay and Seal Slide (Kangaroo Island); Dangerous Reef and English Island (Spencer Gulf); Olive and Jones Island (Chain of Bays) and Lilliput and Blefuscu Islands (Nuyts Archipelago). Monitoring of pup production for consecutive (or near consecutive) breeding seasons has been maintained at these sites since at least 2007. For this project, monitoring ofpup production at Seal Bay started in December 2014 when the breeding season commenced and was still underway at the time of writing this report (May 2015). As such, survey results for the previous (2013/14) breeding have been used here, but will be updated when the 2015 breeding season is completed. Surveys at the Seal Slide were undertaken between December 2014 and May 2015. At Dangerous Reef and English Island surveys were undertaken between March and October 2015; at Olive and Jones Islands between September and December 2014 and at Lilliput and Blefuscu Islands between December 2014 and April 2015.

Ground and aerial surveys were undertaken ofpup numbers at another 46 ASL sites (and numerous non-breeding sites) in SA, many of which had not been surveyed comprehensively for many years (including two previously unknown breeding sites). Surveys also provide a comprehensive assessment of the timing ofASL breeding across all SA breeding sites and its asynchrony.

Objective 2: Archive the first State-wide survey^fall ASL breeding sites within an 18m period, providing a contemporary assessment of the species status, and analyse historic time- series data on all sites to provide an assessment of their trends in abundance.

Surveys were conducted at 54 ASL sites across SA, including a number ofnon-breeding sites. Pups were recorded at 44 sites. This is the first time that contemporaneous surveys have been undertaken across the species SA range over 2-3 breeding seasons, providing the first contemporaneous base-line in pup abundance for the species in SA. Time series data were available for 32 sites, with the maximum interval between 'good' surveys ranging between 6 and 44.5 years. The total quantum and rate of change in estimated pup abundance over time was calculated for these sites.

Objective 3: Review and evaluate sites, survey methodologies and strategies developed over the last decade. Based on these analyses, provide options and make recommendations on future monitoring strategies for the species.

An evaluation of survey methodology has been undertaken. Options for future monitoring strategies are discussed below.

More detail will also be provided in a peer-reviewed SARDI report on the project, which is currently being drafted.

3. Appropriateness

The appropriateness of the approaches used in the development and implementation of the Activity

Survey methodology for ASL has progressively been developed and improved over the last 20 years. Given the asynchronous breeding pattern for the species, a critical aspect of the success of any ASL survey, irrespective of the methodology being used, is getting the timing of the survey right with respect to the breeding season. Fortunately with ongoing monitoring by SARDI scientists over the last decade there have been many opportunities to assess breeding schedules for many sites, which has increased the likelihood of survey success. However, for some sites that had not been surveyed for some time, there were issues with survey timing. Another critical approach has been the almost universal use of helicopters to access sites. With the exception of a few colonies with vehicle access or where boat access can be achieved reliably, all other surveys utilised helicopters for access. This meant that surveys could be planned and conducted at specific times that were not contingent on weather. The use of helicopters also increased efficiency within field trips, as multiple islands could be quickly assessed over a short time, providing essential information on the timing and progression of the breeding season at different sites.

4. Effectiveness

The degree to which the Activity has effectively met its stated objectives Obiective 1: Continue.tQjirovide data on the status and trends in abundance ofASL at key monitoring sites in SA, as well as a number other locations where information on breeding status and trends in abundance are poor.

This project surveyed all previously known ASL breeding sites as well as many other possible or potential breeding sites across SA. Most sites were visited at least twice over the period of the project and a range of survey methods was used, from aerial to ground surveys (pup counts, mark-recapture procedures, and cumulative pup production methods) (see Table 1). This included continued monitoring of pup production and trends in abundance at eight key ASL monitoring sites from four regions across SA: Seal Bay and Seal Slide (Kangaroo Island); Dangerous Reef and English Island (Spencer Gulf); Olive and Jones Islands (Chain of Bays) and Lilliput and Blefuscu Islands (Nuyts Archipelago) (Figure 1).

Monitoring ofpup production for consecutive (or near consecutive) breeding seasons has been maintained at the key monitoring sites since at least 2007. For this project, monitoring of pup production at Seal Bay started in December 2014 when the breeding season commenced and was still underway at the time of writing this report (May 2015). As such, survey results for the previous (2013/14) breeding season have been used here, but will be updated when the 2015 breeding season is completed. Surveys at the Seal Slide were undertaken between December 2014 and May 2015. At Dangerous Reef and English Island surveys were undertaken between March and October 2015; at Olive and Jones Islands between September and December 2014 and at Lilliput and Blefuscu Islands between December 2014 and April 2015. Updates on pup numbers for the 2014/15 breedmg seasons for these key sites are presented in Figure 1, specific information on their status and trends in abundance is addressed in the following section, and more detail will be provided in the final SARDI report for the project. The project activities has fully met the stated aims of Objective 1.

Lllllput Island Blefuscu Island

_- - • '•""\

I 60 c a. 3 0. 40

20 ~~ ~ y = 0.003 " e^O.OOSx) r'= 0.012 - -y=g.963e-11 •eil(0.014x) ^ 0,192

2008 2010 2012 2014 2016 2008 2010 2012 2014 2016

Year a. Year b.

Olive Island Jones island

.~"t~-< • • __------•--

S 10

- - y = 8.296«»W • e"(.O.M46l) I'- 0.457 - -y«1.8a3e-l6-e«(O.OI9») r'=0.210

2006 2008 2010 2012 2014 2016 2(XM 2006 2008 2010 2012 20M 2016

Year Year c. d. Dangerous Reef ;-v English Island / t 70

_.*-•' g 50

a: 30

10 - y » 4.763e-f7 • eA(0.055x) r- 0.618

0 1990 1995 2000 2005 2010 2015 2004 2006 2008 20)0 2012 2014 2016

Year Year e. f.

350 Seal Bay • Pup production estimate •• Maximum Ifre pup count 300 • ---.- 250 • • 200

• • 150 » • - '-- •• • • ' • • • • <00 • - - y . 0.014 • e"(0.003«) r- 0.055 y a 1.6345*13 •e"(.

1990 1995 2000 2005 2010 2015 2000 2002 2004 2006 2008 2010 2012 2014

Year h. Year

Figure 1. Trends in pup abundance at eight key ASL monitoring sites from four regions across SA: Lilliput and Blefuscu Islands (a, b - Nuyts Archipelago); Olive and Jones Islands (c, d. Chain of Bays); Dangerous Reef and English Island (e, f- Spencer Gulf); and Seal Bay and Seal Slide (g, h - Kangaroo Island). Red points indicate unreliable data due to a likely influx ofpups from neighbouring colonies.

Obiective 2: Archive the first State-wide sui-vey of all ASL breeding sites within an 18m period, providing a contemporary assessment of the species status, and analyse historic time- series data on all sites to provide an assessment of their trends in abundance,

Over 120 individual surveys across 54 known breeding sites were conducted over the 18 month period of the project (Table 1). This does not include the weekly surveys conducted at Seal Bay on Kangaroo Island during the core part of its breeding season, or surveys conducted at haul-out sites. This is the first time that contemporaneous surveys have been undertaken across the species SA range over 1-2 breeding seasons, providing the first contemporaneous base-line estimate ofpup abundance for SA, including the best estimate of the timing of breeding and the degree of its asynchrony across SA (Table 1). Previous assessments have relied on limited available survey data for many sites, spanning up to 18 years, to provide a best estimate ofpup and population abundance (Gales et al. 1994, Goldsworthy et al. 2009, Shaughnessy et al. 2005, Shaughnessy et al.2011).

A total of 2,417 pups were recorded at 44 sites (Table 1). Two sites, Curta Rocks (7 pups) and Williams Island (5 pups) were identified as new breeding sites (Table 1). To provide the best estimate ofASL pup abundance, adjustments have been made for some sites. For the Bunda Cliffs breeding sites (GAB B1-B9, H12), cliff-top surveys underestimate pup production so estimates have been adjusted based on verified cliff-top and ground surveys by Dennis & Shaughnessy (1996) for the 1995 breeding season (see Goldsworthy et al. 2003), adjusted based on the 39% decline reported for all sites between 1993/94 and 2012/13 by Mackay et al. (2013). For five other sites (Albatross and English Islands, Peaked Rocks and North Islet, and The Pages), pup production surveys were conducted well after the breeding season had ended based on the number ofmoulted pups sighted, or were not able to be verified by ground counts. The previous surveys were undertaken at Albatross Island, Peaked Rocks and North Islet in 2011, and at The Pages in 2010. As such, the timing of breeding had to be extrapolated forward without the benefit of more recent information. In addition, dispersal of pups from Dangerous Reef potentially inflated counts at Albatross and English Island (one tagged pup from Dangerous Reef was sighted at English Island during the survey). For these sites, estimates from previous ground surveys were used (see Goldsworthy et al. 2012, Shaughnessy et al. 2013). Black Point (on Kangaroo Island) is an isolated promontory between Seal Bay and Cape Gantheaueme, and has previously been recognised as a haul-out site with occasional pupping (Shaughnessy et al. 2009) and was not surveyed. A previous estimate of one pup based on a survey in 2002 was used for this site (Shaughnessy et al. 2009).

With the above adjustments, the current total pup abundance for ASL breeding sites in SA is estimated to be 2,619 (Table 1). This is 500 pups (16%) less than the 3,119 estimated by Shaughnessy et al. (2011), based on a range of SA surveys conducted primarily between 2004 and 2008, although data for some colonies was from as early as 1990. However, this is likely to have underestimated pup abundance over that period because since then, five new breeding colonies have been identified and surveyed (Cap Island, Rocky Island (South), Little Hammock Island, Curta Rocks and Williams Island accounting for 58 pups); and three sites inadequately surveyed in the past have had comprehensive surveys undertaken since (most as part of the current study). Previous estimates ofNuyts Reef (15 pups). Rocky Island (North) (16 pups) and North Casuarina Island (3 pups) were considerably less than reported in this study (105, 35 and 11 pups, respectively; Table 1). Accounting for these adjustments, a further 175 pups could be added to the estimate of Shaughnessy et al. (2011) increasing it to 3,294. This figure is likely to provide the best comparative figure to the current estimate, increasing the difference to 675 pups (20%).

The key finding of the State-wide surveys is that the current size of the ASL population in SA is much smaller than previous estimates. Although some of this may be explained by previous estimates taking the best (highest pup numbers) surveys available over two decades, most of the reduced estimate is attributed to a decline in population size (discussed in more detail below).

For 49 ASL sites where current or previous pup production has been reported, a subset of 32 sites provided at least two comparable surveys at least a six years apart (mean interval 14.5 years, range 6 - 44.5 years). These sites account for 82% (2,157) of the estimated pup abundance in SA, and also provide indicators of trends in abundance across six of the seven meta-populations (regions) across SA (Table 2). The key findings of these analyses are that pup numbers at most breeding sites (78%) and regions have declined between the survey periods. The overall rate of decline is -3.7% per year, or -5.5% per breeding cycle, but the rate of decline varies considerably between sites and regions (Table 2). Following FLJCN Redlist assessment methods, and using a mean generation time of 12.6 years, extrapolations were made to the change in pup abundance over three generations (37.8 years) based on the observed change in pup abundance and assuming a constant exponential rate of change (Table 2). Following this approach, the total percentage change in estimated pup abundance over three generations was estimated to be -69.7%.

This global assessment approach based on 82% of the estimated SA ASL pup production, suggests the species has declined by -70% in 3 generations, meeting the IUCN Endangered category based on criteria A2(b) and (d). Table 1. Survey effort across 54 Australian sea lion sites in South Australia between November 2012 and May 2015. Blue shading indicates either the observed or estimated timing of breeding among sites. Survey type is indicated: ground count (G); mark-recapture (M); aerial survey by helicopter (A) and cliff-top survey (C). The eight key ASL monitoring sites are shaded grey. The number ofpups recorded in surveys at each site is indicated, and the overall estimate ofpup abundance is provided. Where this differs from the number surveyed, the year and source of the estimate is indicated.

12 13 14 15 Pups Pups Site [reedlngsite N D F M A M J J A S OIIE FMAMJ J AS OHD F f.' 1 ;AB B9 C : " c c • 10 2014 Adjusted - Matkay et al. 2013 2 •ABHI2 0 2014 Ttii 3 iABBS c ^ c ' 22 2014 ited-Mackayetal.2013 4 iABB7.1 0 2014 This survey 5 iABB7 c 2 2014 Adjusted - Mackayet al.2013 6 iABB5 c 7 2014 Adjusted - Macloy et aL 2013 7 .ABB5 c 25 20U Adjusted - Mackayet al.2013 8 >ABE4 1 2014 Adjusted - Matkay et al. 2013 9 iAB63 c IS 3014 Adjusted - Msckayet al.2013 10 ;ABB2 c 3 2014 Adjusted - Machay et al. 2013 11 iABBLl c 3 2014 Adjusted - Mackay et al. 2013 12 ;ABB1 0 2014 TNs sun/ey 13 '/estemNuytsReef GA" G G G A G 105 2015 m 14 t Fowler (Camel-Foot Bay) A 0 2015 Ttii 15 urdiE G" 67 2015 TTiis survey 16 /estls. G G "G 20 20 2015 nils survey 17 enelon Is. G " G 'G IS 19 2015 Till IS )unds Is. G 20 20 2015 Thi 6 '' 19 realwaterls. A 27 27 2015 TT1; 20 liddon Reef A G 0 0 2015 TM 21 lefmm (3 ' GM GM A A GM GM 1D5 105 2015 TIii 22 llllput A " G G^ 6M A <5M GM 74 74 2015 TM 23 live \s. A '' G^ GM G S ' G^ G; 142 142 2014 Ttit 24 icolas Baudn Is. A G^ G A - G 63 63 2014 TN urvey G'1 25 /ard is. G G 44 44 2015 Thi urvey 26 G^ G G G G ' 30 30 2015 TW 27 orothee Is G 0 0 2013 Gol worthy et at. 2014 28 olnt Labatt A A'1 0 2 2013 m 29 >nei Is. 6 19 19 2014 7US G ' 30 festV/aldegravels. G G A 89 89 2015 W 31 -ip Island G 31 31 2014 TTii 32 scky (North) Is. G'1 G G '' 35 35 2014 W G" G '' 33 scky Island (South) II 20147 urvey 34 lurHummocks Island G' GA' G 6 2014 T jr^ey 35 ttle Hummock Island G '' 4 2014 T 36 •ice Is. G 31 2014 T un/ey 37 G G 25 2015 T uney 38 Jrta Rocks A A 7 2014 T urvey 39 fiHiams Island AG G s 2014 T 40 iwls Island G 'G' Q ~ G 92 3314 T 41 orth Meptune, East G " G 9 2014 T 42 1 uth Neptune. Mafn 2 2014 T >y 43 batross Is. G ' G 35 2011 G worthy et al. 2012 44 iglishh. A G 64 2011 G worthy etsl. 2012 45 aneerous Rtef A" G'G^G.^ G^ 2014 T 4G orthls. G '' 6A 32 2011 G worthy etal.2012 47 ?aked Rods A" A ' 17 2011 G worthy etal. 2012 48 Drth Casuarinals.fKangaroo Is.) 11 2014 T1 49 ipe Bouguer tKangaroo Is.) G '' 9 2014 T 50 ive Point (Kangaroo Is.) 0 2014 Tl jney 51 >alBay{Kanearoolt.) 6G3G6GGGGG G G GO G G G G 268 2014 T1 jrvey 52 ackPoint(Kangarools.) 2002 S! finessyetal.2009 53 »1 Slide (Kangarooll.) G ' G '' G ' G G G 10 2015 T1 54 lePaies G 309 478 2010 Shaughnessyetal.2013 Total 2,417 Z.G19

The regional analysis indicates that declines have occurred across the range of the species in SA. The extent of these declines over the last 3 generations is estimated to range between 22% and 85% (Table 2). Western populations off the Eyre Peninsula (Nuyts Archipelago and Chain of Bays) and in the GAB (Bunda Cliffs) have shown the greatest rates of decline. A key example of a marked regional decline was noted for breeding sites in the western Nuyts Archipelago, where surveys undertaken in 2015 at Purdie, West, Fenelon and Lounds Islands counted 126 pups in total, which was less than half the number seen in surveys at these sites conducted in 2005 and 2008 (262 pups, Shaughnessy et al. 2011). Similarly, the other major colonies off the western Eyre Peninsula at Olive Island, Nicolas Baudin and West Waldegrave have undergone an overall decline of-37% in pup abundance since surveys in the early to late 2000s.

Further details of site specific changes and trends in pup abundance will be provided in a peer-reviewed SARDI report on the project, which is currently being drafted.

The project activities have fully met the stated aims of Objective 2. Table 2. Regional summary of the number breeding sites with time series data enable estimates of rates of change in pup abundance over time. The number of sites decreasing and increasing, and the mean annual and per breeding season rate of change (r) is also presented. The estimate three generation change in pup abundance is also provided following IUCN assessment methods.

Region No. No. with No. No. r(%)per r(%)per 3-gen breeding sites time series decreasing Increasing year breeding season change (%) Bunda Cliffs 9 9 9 0 -3.7% -5.5% -65.2% NuytsReef 1 0 NuytsArchipelago 8 6 3 3 -5.2% -7.8% -85.1% Chain of Bays 8 6 5 1 -3.0% -4.5% -78.8% SW-Eyre 6 2 1 1 -1.7% -2.6% -8.4% Spencer Gulf 11 6 5 1 -5.0% -7.4% -58.9% Kangaroo Island 6 3 2 1 -0.5% -0.8% -21.8% Sum/mean 49 32 25 (78%) 7 (22%) -3.7% -5.5% -69.7%

Objective 3: Review ancLeyalyate sites, survey methodoloeies and strategies developed over the last decade. Based on these analyses, provide options and make recommendations on future monitoring strategies for the species.

Survey methodology for ASL has progressively been developed and improved over the last 20 years. A discussion of key issues around survey methodology is provided below and options for future monitoring strategies are discussed.

Further detail will be provided in a peer-reviewed SARDI report on the project, which is currently being drafted.

Timing of surveys:

Given the asynchronous breeding pattern for the species, a critical aspect of the success of any ASL survey, irrespective of the methodology being used, is getting the timing of survey right with respect to the breeding season. For large colonies, we typically aim to commence surveys in the 3rd month of breeding and to conduct an additional survey at the end of the breeding season. With small colonies it is possible to undertake suitable ground counts at the end of the breeding season, as the breeding season may only last 3-4 months (see below).

Knowledge of the timing of breeding is the single most important factor in ensuring survey success. Investing in and maintaining this knowledge can save considerable costs in ensuring that the timing of future surveys is appropriate.

Some key points on survey timing include:

Interbreedmg interval is not consistent within and between colonies - although available data suggests the breeding interval is generally between 17 and 18 months, it can vaiy between breeding seasons and appears to vary among colonies. The extent to which the variation in interbreeding interval relates to the duration of a breeding season is unclear.

Breeding season duration varies relative to the size of the colony - breeding seasons are longer for large colonies (6-9 months) but may be as short as 3-4 months in small colonies. In small colonies, surveys conducted at the end of the breeding season often do not record any moulted pups, indicating that breeding is over before the oldest pups have moulted, which we think occurs around 5 months of age (Shaughnessy et al. 2005). At larger colonies, moulted pups are present while some pups are still being born. The accepted wisdom of a 5 month breeding season for the species is probably only applicable to larger colonies. As a consequence there has been a tendency to underestimate the progress of breeding at small breeding sites.

Breeding season duration can vary among seasons - survey data from Seal Bay and Dangerous Reef indicates marked variability in the duration of successive breeding seasons within sites (which can range between 6-12 months - see Goldsworthy et al. 2014). This can impact on the timing and the number and spread of surveys required to adequately survey some sites - especially larger colonies.

Colony access Even if the knowledge of the timing of breeding at sites is excellent, this counts for little if access to islands is reliant on boat access, and therefore governed by sea conditions. In SA, most ASL colonies (with a few exceptions) cannot be reliably or safely accessed by sea. Helicopters, where they can be used, bring great efficiency and capacity to survey multiple sites quickly and efficiently, and enable regional monitoring of abundance and breeding schedules without necessarily having land.

Proximity to other colonies Where neighbouring colonies have earlier or similar breeding seasons, estimates ofpup abundance can be confounded if there is the potential for movement ofpups between sites. This is especially so where small colonies neighbour large colonies. For example tagged pups from Dangerous Reef have been sighted while undertaking surveys at English and Lewis Island. Although the number of tagged individuals that move between colonies can be readily determined, the number ofuntagged individuals that may be present is often unknown, although efforts to estimate the proportion of these based on the number of tagged individuals have been made (Goldsworthy et al. 2008). Tagged pups from Nicolas Baudin Island have been sighted regularly at Jones Island (Goldsworthy et al. 2014), and inter-colony movement of pups is suspected to have lead to higher pup counts for late season surveys at English and Albab-oss Island during recent surveys. These colonies are only ~2 Okm from Dangerous Reef.

'Alternate states' in breeding At a number colonies where consecutive surveys have been undertaken, there can be marked inter-seasonal variability in pup counts. At some sites, the variability in the counts appears to alternate in relation to the alternate breeding seasons (e.g. summer followed by winter; spring followed by autumn). For example, at Seal Bay, counts of live pups are typically greater, and pup mortality rates lower in winter breeding seasons compared to summer breeding seasons (Goldsworthy et al. 2014). In contrast, live pup counts tend to be lower, and pup mortality rates higher in winter breeding seasons at Dangerous Reef compared to summer breeding seasons (Goldsworthy et al. 2014).

At Dangerous Reef, the impact of alternate states on survey methodology and pup abundance estimates has been examined. Goldsworthy et al. (2011) compared three key pup production metrics assessed over four consecutive breeding seasons at Dangerous Reef (2006/07, 2008, 2009/10 and 2011). These were: 1) minimum live and dead pups (cumulative marked [tagged] pups plus cumulative dead [unmarked] pups plus maximum unmarked pups counted); 2) maximum Petersen estimates and 3) cumulative pup production. Comparison of these metrics indicated an apparent altemation between two states in these metrics. In winter breeding seasons, the pup estimates based on all three metrics are very similar to each other (2008, 2011), but in summer breeding seasons, estimates based on all three metrics are very different (2006/07, 2009/10), with cumulative pup production estimates being greater than the Petersen estimates, which are greater than the estimates based on the minimum estimate of live and dead pups. The latter (summer) conforming pattern makes intuitive sense, where Petersen estimates provide a greater estimate based marked and counted pups, and where cumulative pup production estimates (which estimate net-pup production between successive Petersen estimates) provide a greater estimate compared to an isolated Petersen estimate. Goldsworthy et al. (2011) questioned why in the summer (non-conforming) breeding seasons these latter two metrics provided no better estimates than the minunum count of live and dead pups. They suggested the difference may relate to changes in pup behaviour and breeding chronology between summer and winter breeding seasons that affect the sightability of pups and impact on the assumptions of the Petersen estimate (including temporary migration, survival and sightability).

To assess this, Goldsworthy et al. (2014) undertook analyses to determine if inter-breeding season differences in individual detection heterogeneity (IDH) may contribute to biasing estimates of pup production and trend analyses at Dangerous Reef. Data from the four seasons were analysed using the captire-mark-recapture (CMR) program MARK to estimate survival and capture probabilities using the Cormack-Jolly-Seber (CJS) model. No significant differences were detected in re-sight probability between breeding seasons (essentially equal to 1 in all breeding seasons). However, survival (across the period of surveys within breeding seasons) varied markedly between breeding seasons, with no apparent relationship between survival in the conforming and non-conforming seasons.

Use of colony appropriate survey methods Survey choice is contingent on many factors; most important are funding and time available to commit to single or multiple surveys. Generally, all sites are amenable to direct counting methods, although the suitability and reliability of direct counts will depend on the site terrain and sightability ofpups, which are affected by vegetation and the amount of other cover present. Reliable and consistent counts can best be derived from small, isolated and exposed sites where all pups can be readily found and there is limited opportunity for pups from neighbouring colonies to be present. Smaller colonies have shorter breeding seasons where pups will still be in their brown lanuago by the end of the breeding season, and therefore show less propensity to disperse.

As colony size increases, so does the length of the breeding season, and the likelihood that pups born early in the breeding season will have moulted before the breeding season ends, be more aquatic and be absent from the colony on foraging trips or disperse to neighbouring sites. These factors all contribute to the availability biases in surveys. For such sites, multiple surveys will provide better estimates ofpup abundance, and can include capture mark and resights (CMR), mark-recapture methods following the Petersen estimate, and the cumulative pup production (CPP) methods which not only estimates the number ofpup present at each survey period, but also the net pup production that has occurred between surveys. These methods have been extensively detailed in previous SARDI reports (e.g. Goldsworthy et al. 2014).

For the larger key monitoring sites, we typically utilise three survey methods; ground counts, mark-recapture and cumulative pup production. These are undertaken over a minimum of two, and up to five trips, depending on the size of the colony and length of the breeding season. There is clearly some trade-off between survey effort and survey precision/accuracy. At this stage we have not been able to readily assess the happy medium between the two. However, given the length ofpupping seasons, we feel that 2-3 surveys are the minimum if the full extent of breeding is to be monitored. We have focused efforts on optimising survey conditions and reducing sources of variability, with an overall aim to reduce survey effort without compromising accuracy.

Recommendations on future monitoring strategies for the species This study identified that ASL populations in SA are much smaller than previously estimated, and are in decline. A future monitoring strategy is essential to meet the Australian sea lion Recoveiy Plan objectives (DSEWPaC 2013) and to ensure that management measures to mitigate threatening processes, particularly bycatch mortality in demersal gillnet fisheries, are adequate to ensure recoveiy of the species.

SARDI researchers have contributed extensively to the Department of the Environment's development of an Australian sea lion Monitoring Framework (including a workshop held in Adelaide April 1-3 2014, and a draft report). Much of our knowledge and experience from survey development and design has been incorporated into that process.

Our general recommendations for future monitoring ofpup abundance for the SA populations include the following. Continue ongoing monitoring of key regionally representative sites across the species range to provide ongoing time-series of trends in abundance. The peer-reviewed SARDI report on the current project will provide a full review, assessment and evaluation of data from the eight current key monitoring sites. It will also assess which sites should be retained and/or replaced, and recommend replacement sites. For regions where there is currently no ongoing monitoring, it will identify potential new sites as regionally representative colonies, including potential synergies with marine park monitoring. Maintaining fine scale population monitoring at Seal Bay, including pup production and population vital rates (survival, fecundity etc). Shifting away from past survey design that targeted only a few selected colonies with high survey effort, to a more mixed survey design. With the shift from boats to helicopters to access colonies, there is greater potential to include low level survey effort of sites near key monitoring sites. For other breeding sites, surveys should be repeated eveiy 3-6 years. Continue trials of remote cameras to monitor the timing of breeding seasons in order to improve survey efficiency.

The project activities have fully met the stated aims of Objective 3.

References: 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.

DSEWPaC. (2013). Recovery Plan for the Australian Sea Lion (Neophoca cinered).

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 ofZoology, London, 234: 353-370.

Goldsworthy S.D., 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. (cds). Marine mammals and humans: fisheries, tourism and management issues. CSIRO Publishing, Melbourne, pp 62-99.

Goldsworthy SD, Shaughnessy PD, Page B, Lowther A, Bradshaw CJA (2008) Developing population monitoring protocols for Australian sea lions: enhancing large and small colony survey methodology. Final Report to the Australian Centre for Applied Marine Mammal Science (ACAMMS), Department of the Environment, Water, Heritage and Arts. SARDI Aquatic Sciences Publication Number F2008/000633-1, SARDI Research Report Series No. 297, 44 pp Goldsworthy, S. D., McKenzie, J., Shaughnessy, P. D., Mclntosh, R. R., Page, B. and Campbell, R. (2009a). An Update of the Report: Understanding the Impediments to the Growth of Australian Sea Lion Populations. Report to the Department of the Environment, Water, Heritage and the Arts. SARDI Publication Number F2008/00847-1. SARDI Research Report series No. 356. 175pp.

Goldsworthy, S. D., Page, B., Lowther, A. D. and Shaughnessy, P. D. (2012). Maintaining the monitoring ofpup production at key Australian sea lion colonies in South Australia (2010/11). South Australian Research and Development Institute (Aquatic Sciences), Adelaide. SARDI Publication No. F2010/000665-2. SARDI Research Report Series No. 601. 64pp.

Goldsworthy, S. D., Kennedy, C, Shaughnessy, P. D., and Mackay, A. I. (2014). Monitoring of Seal Bay and other pinniped populations on Kangaroo Island: 2012-2015. Report to the Department of Environment, Water and Natural Resources. South Australian Research and Development Institute (Aquatic Sciences), Adelaide. SARDI Publication No. F2014/000332- 1. SARDI Research Report Series No.782.39pp

Mackay, A.L, Shaughnessy, P.D. Goldsworthy, S.D. (2013). Australian sea lion abundance in the Bunda Cliffs region, GAB Marine Park. Final Report to the Department of Environment, Water, and Natural Resources. South Australian Research and Development Institute (Aquatic Sciences), Adelaide. SARDI Publication No. F2013/000642-1. SARDI Research Report Series No. 746. 25pp

Shaughnessy, P. D., Dennis, T. E. and Seager, P. G. (2005). Status of Australian sea lions, Neophoca cinerea, and New Zealand fur seals, Arctocephahisforsteri, 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., Mdntosh, R.R. (2009) Status of small colonies of the Australian sea lion Neophoca cinerea on Kangaroo Island, south Australia. Australian Zoologist. 35: 82-89.

Shaughnessy, P. D., Goldsworthy, S. D., Hamer, D. J, Page, B. and Mcliitosh, R. R. (2011). Australian sea lions Neophoca cinerea at colonies in South Australia: distribution and abundance, 2004 to 2008. Endangered Species Research, 13: 87-98.

Shaughnessy PD, Goldsworthy SD, Burch P, and Demiis TE (2013) Pup numbers of the Australian sea lion (Neophoca cinerea) at The Pages Islands, South Australia, over two decades. Australian Journal ofZoology. 61: 112-18.

5. Financial Account of the Activity An income and expenditure statement for the period April 2014 to May 2015 is attached. Note, remaining unspent money will be fully expended by the time the peer reviewed SARDI report is submitted to the AMMC. An report from an approved Auditor as specified in subclause 9.2(f) will also be provided at that time. AUSTRALIAN MARINE MAMMAL Centre

STATEMENT OF INCOME & EXPENDITURE

Name of Grantee: A/Prof Simon Goldsworthy Organisation: Project No.: Project Title: Amount of Grant Exclusive GST Total including ofGST GST Commonwealth Government $116,231.00 $11,623.10 $127,854.10 Contributions Organisation Contributions $46,773.00 $4,677.30 $51,450.30 Other Contributions $94,374.00 $9,437.40 $103,811.40

A. Item B. Expenditure C. Committed ($ GST excl) ($ GST excl) Labour Costs $47,472.97

Operating Costs $50,846.48

Total expenditure $98,319.45

Legal Commitments

A statement ofliow much the Organisation needs to meet current liabilities under legal commitments entered into for the purposes of the Activity and pursuant to the Funding Agreement.

I certify that: (a) all Funding received was spent for the puipose of the Activity and in accordance with the Funding Agreement and that the Organisation has complied with the Funding Agreement; (b) salaries and allowances paid to persons involved in the Activity are in accordance with any applicable award or agreement in force under any relevant law on industrial or workplace relations; (c) unless the Activity Period has expired or the Agreement has been terminated, the unspent portion of the Funds (if any) is available for use within the next Reporting period; (d) the financial information is presented in accordance with any other financial Reporting requirements the Department may notify to the Organisation; (e) where an Asset has been acquired with the Funds, paragraphs 7.5 (d) and (g) (where applicable) have been complied with in respect to the Asset. Signature of Responsible Person (as defined by sub clause 9.6 of the Funding Agreement) Name Prof GavinB egg Research Chief, SARDI Aquatic Sciences

Date 22 May 2015

Please forward this financial acquittal with the Final Report and a copy of a letter, or a report, from the Organisation's Approved Auditor as specified in subclause 9.2(f) of the Funding Agreement to:

Please forward 4 hard copies, and one electronic Word document of this report to:

Robyn Goyen Coordinator Australian Marine IVIammal Centre Australian Antarctic Division 203 Channel Highway KINGSTON TAS 7050 ^inmccoor([email protected]\au { Field Code