The Recovery of Rat Island Following the Eradication of Introduced Predators

THE RECOVERY OF RAT ISLAND FOLLOWING THE ERADICATION OF INTRODUCED PREDATORS

Rat Island Recovery Project

The Rat Island Recovery Project (RIRP) began with a feasibility study published in January 2004 and provided to the Department of Fisheries. Since 2008 it has been part of the Conservation Council (WA) Citizen Science for Ecological Monitoring Program. The project seeks to monitor and document the recovery of the Rat Island seabird colonies and the terrestrial ecosystem following the successful eradication of Black Rats and feral cats in the 1990s and to facilitate restoration projects that may enhance the recovery process. The current team consists of Nic Dunlop, Elizabeth & John Rippey, Laura Bradshaw, Josie Walker, Alaya Spencer-Cotton, Jenita Enevoldsen and Andrew Burbidge. Access to the Saville-Kent Research Centre was provided by the Department of Fisheries (WA). The production of this contribution to the management of Rat Island was supported by a NACC Coastal Grant with the GIS work done by Fisheries WA. The work is dedicated to the memory of our colleague Colin Chalmers.

CONTENTS

1. RAT ISLAND 1

2. THE ENVIRONMENTAL HISTORY OF RAT ISLAND 2 2.1 Original Terrestrial Ecosystem on Rat Island 3 2.3 Impact of Guano Mining 3 2.3 Impact of Introduced Predators 4 2.4 Impact of Fishing Communities 5

3 RAT ISLAND RECOVERY 7 3.1 Eradication of Introduced Predators 7

3.2 The Terrestrial Ecosystem Following Predator Eradication 8 3.2.1 Rat Island Recovery Project (RIRP) - Survey & Assessment Methods 8 3.2.2 Vegetation 10 3.2.3 Invertebrate Fauna 11 3.2.4 Vertebrate Fauna 12 3.2.5 Stable Isotope Analysis of the Terrestrial Food Web 13 3.3 Seabird Re-colonization 15 3.3.1 Seabird Survey Methods 15 3.3.2 Seabird Re-colonization Trends 15

4 RAT ISLAND MANAGEMENT ʹ RECOVERY OR RESTORATION? 18

4.1 Managing Recovery 18 4.1.1 Emerging Wildlife and Ecosystem Management Issues 18 4.1.2 Management of Ecological Weeds 21 4.2 Restoration Projects 23 4.2.1 Eradication of the House Mouse 23 4.2.2 Rehabilitation of habitat for burrow-nesting species 24 4.2.3 Re-introduction of extirpated reptile species REFERENCES 26

Figure 1 - Map showing Abrolhos Islands, Easter Group and Rat Island 1 Figure 2 ʹ Isoscape of Rat Island 14

GIS Map 1- Heritage features of Rat Island 29 GIS Map 2- Rat Island Recovery study sites 30 GIS Map 3- Vegetation of Rat Island 31 GIS Map 4- Seabird colonies up to 2011/12 32 GIS Map 5- Seabird colonies in 2012/13 33 GIS Map 6- Management actions on Rat Island 34 GIS Map 7- Distribution of Bryophyllum infestation on Rat Island 35

1. RAT ISLAND

Rat Island is an elevated (3-4 metres above MSL) island in the Easter Group of the Houtman Abrolhos archipelago 78 km west of Geraldton, Western Australia (Figure 1). The island is comprised of coralline limestone (Wallabi Limestone) formed during the high stand of the Eemian stage 125 000 years BP. The area would have been a low coastal ridge during the latter part of the Pleistocene and then isolated from the mainland by the most recent marine transgression about 7000 years ago. Colonization by seabirds and the accumulation of its historical mantle of guano would have occurred during the latter Holocene period.

Rat Island is broadly rectangular in shape with its long axis oriented north-south. It has a supra-tidal area of 61Ha and is sparsely vegetated (Harvey et. al 2001).

The Abrolhos archipelago is perched on the edge of the continental shelf and adjacent to the southward flowing Leeuwin Current and consequently provides important breeding sites for seabirds, particularly tropical species (Storr et al. 1986, Gaughan et al. 2002).

2. THE ENVIRONMENTAL HISTORY OF RAT ISLAND

Rat Island was surveyed and named by Lieutenant John Lort Stokes on the HMS Beagle in April 1840. Evidently the island was already populated with Black Rats to the extent that they gave it name.

͞dŚĞĐĞŶƚƌĞŝƐůĂŶĚǁĞŶĂŵĞĚZĂƚ/ƐůĂŶĚ͕ĨƌŽŵƚŚĞquantity of that ǀĞƌŵŝŶǁŝƚŚǁŚŝĐŚŝƚǁĂƐŝŶĨĞƐƚĞĚ͘͟ John Lort Stokes,1846

The presence of Black Rats prior to any European inhabitation of the Abrolhos is somewhat puzzling but was presumably the result of some undocumented visit by a sailing ship or of a shipwreck. Stokes visited in April. This was outside the spring-summer breeding period of the large colonies of tropical terns on Rat Island. These were documented later in the colonial period.

Rat Island was the scene for what has become arguably the best-documented ecological calamity in the history of European settlement in Western Australia. In 1889, A.J. Campbell estimated that the mixed colony of Common Noddies (Anous stolidus) and Sooty Terns (Sterna fuscata) held 1 452 000 birds (Serventy et al.1971). These minions were completely extirpated by the late 1930s through the combined effects of guano mining, the introduction of cats and egg ʹcollecting by fishermen (Burbidge et al. 1996).

During the colonial period the Rat Island tern colony was at least 3 times the size of the spectacular breeding aggregations that still occur on the southern end of Pelseart Island (Burbidge et al.1996). Gibson (1908) visited both islands and noted that the Sooty Tern colony on Rat Island was much larger than the one on Pelseart. His observations effectively confirm that the colonies we have on Pelseart Island today are not the result of displacement from the Rat Island population.

As well as the spectacular tern colonies, Rat Island had thousands of Wedge-tailed Shearwaters (Ardennia pacificus) burrows until at least 1913 in its guano mantle and low sand-dunes (Stokes 1840, Stanbury 1993 and Alexander 1922). Other Abrolhos seabird species may also have nested on Rat Island at the time, particularly those (of larger size) that were least vulnerable to rat predation.

By the time the rock-lobster fishery became established at the Abrolhos in the 1950s Rat Island was a worked-out exhausted environment, now silent with the loss of its great tern colonies. Some terrestrial plants and animals hung on, whilst others like the Spiny-tailed Skink, had been extirpated. The camp settlement on Rat Island developed during a period when the Rat Island ecosystem lacked many obvious natural values and with little regard for environmental impact.

2.1 Original Terrestrial Ecosystem on Rat Island

There are no records of the flora and vegetation of Rat Island prior to guano mining and the introduction of cats. The first systematic surveys of Abrolhos vegetation were commenced in the 1990s (Harvey et al. 2001). The undisturbed platform islands of most comparable size are located in the Pelsaert Group. These islands lack a guano mantle and have no history of high seabird concentrations so they are not particularly useful as ecological reference sites.

It is likely that much of Rat Island was covered with Nitre Bush Nitraria dominated heathland at varying densities, with patches of Samphire (Tecticornia succulent heath) in the more saline, low lying areas. These vegetation types occupy much of remnant unmined surface today. Atriplex cinerea low heath would have occupied any low accumulations of dune sand (as it still does along the western margin of the island). Numerous exotic, annual herbs and grasses have been introduced to the island along with a range of exotic succulent perennials, ĐƵůƚŝǀĂƚĞĚĂƐ͚ŐĂƌĚĞŶ͛ƉůĂŶƚƐĂƌŽƵŶĚƚŚĞƌŽĐŬ-lobster fishing camps.

The terrestrial invertebrate fauna of the Abrolhos Islands has not been systematically investigated. However observations from the Rat Island Recovery Project provide some information on the probable indigenous verses introduced components in the terrestrial invertebrate community (see Section 3.2.3).

How et al. (2004) recorded eight reptile species from Rat Island, based on specimens in the Western Australian Museum collection and reliable reports of the former population of Spiny-tailed Skinks Egernia stokesii. Recent observations suggest that some of these species may have been lost and additional species recently introduced (Section 3.2.4).

2.2 Impact of Guano Mining

Mining began in earnest on Rat Island in 1885 and wound-up in 1915 (Stanbury 1993, Burbidge et al.1996). The predominately Chinese mine-workers (Coolies) dug and swept-up the guano enriched soil, levering-out the surface limestone in order to get access to the material sequestered in fissures and cracks. Low embankments were constructed for trolley lines that carried the excavated material back to the shipping stockpiles at the north- eastern end of the Island. The product was then loaded onto vessels via trolley lines that ran out onto a stone jetty.

The habitat available to nesting seabirds was drastically altered over time. The soil was effectively removed from about 81 йŽĨƚŚĞŝƐůĂŶĚ͛ƐƐƵƌĨĂĐĞ, leaving limestone pavement, sink holes, rock-piles of coralline slabs, piles of smaller diameter screened rock material, a system off anastomosing low embankments and little perennial woody vegetation.

There appear to be no records of the sequencing of mining fields, however there is variation in the degree of grey discolouration on the excavated slabs of limestone (probably resulting from weathering and colonization by biofilm and cryptogams) that may indicate different time periods. One area, ƌĞĨĞƌƌĞĚƚŽĂƐƚŚĞ͚ĐĞŶƚƌĂůǁĞƐƚĞƌŶĨŝĞůĚ͛ŝŶƚŚĞƌĞĐŽƌĚƐ;^ƚĂŶďƵƌǇ 1993, GIS Map 1 ), was mined after 1897 and the workings indicate this involved the removal of larger limestone slabs and more screening of smaller rocks from the raised material compared to earlier quarries. This is consistent with the increased effort and cost required to recover the product towards the end of the mine life. It may have been the area re-worked during World War 2 with more mechanized mining machinery. In any event this area has a much whiter background colour, deeper relief, higher rock-piles and less vegetation cover than other mined areas on the island.

At the northern end of the island a significant proportion of the un-mined area (eg. football field area) was probably used to graze sheep for food (Stanbury 1993) or possibly horses used for pulling the trolleys (GIS Map 1). Later this area was used for the first airstrip. The succulent-shrub (Nitre Bush) vegetation in this area was mostly replaced by grasses and annuals (particularly invasive iceplant) and shows little sign of natural recovery.

Many surface nesting seabirds such as terns may have been able to utilize the mined-out landscape for breeding in the absence of the introduced predators. However the thorough removal of the guano mantle, and overlying low sand dunes, effectively eliminated the nesting habitat for burrowing species (shearwaters and storm-petrels) over much of the surface.

2.3 Impact of Introduced Predators

Despite the presence of Black Rats, the constant disturbance of the guano diggers and dramatic changes in nesting habitat the tern colonies on Rat Island were still reported to be ͚prodigious͛ in November 1913 (Alexander 1922). Some egging by fishermen is thought to have occurred both during and after the guano mining years but this was unlikely to have been beyond sustainable levels. However cats, introduced by the guano miners sometime between 1889 and 1913 (Alexander 1922) to control the rats became established and clearly predated nesting terns during the breeding season. Alexander (1922) predicted the fate of the colony in 1913 as he wrote:

͞Many of them (Common Noddies) undoubtedly fall victims to the cats, and were it not that these are kept down by the difficulty in finding food at other seasons of the year, when they appear to feed chiefly on crabs, it would doubtless not be long before the Noddies were exterminĂƚĞĚ͘͟

A small and dwindling number of Common Noddies were still present on Rat Island in 1936 and Sooty Terns were still present in 1938 but both species were probably extirpated around that time (Burbidge et al.1996). The Spiny-tailed Skink Egernia stokesii (a form endemic to the Abrolhos Islands) occurred on Rat Island. It also appears to have been extirpated from Rat Island by the cats. dŚĞƉƌŽĐĞƐƐƚŚĂƚŽĐĐƵƌƌĞĚŽŶZĂƚ/ƐůĂŶĚĂƉƉĞĂƌƐƚŽďĞĂŶĞǆĂŵƉůĞŽĨ͚ŚǇƉĞƌƉƌĞĚĂƚŝŽŶ͛ (Russell & Le Corre 2009, Russell 2011) caused by the interaction of two introduced predators, in this case the Black or Ship Rat Rattus rattus and the Cat Felis catus.

Hyperpredation (diagram reproduced from Russell & Le Corre (2009).

High concentrations of seabirds were persisting in the presence of the Black Rats until the introduction of cats by the guano miners. These seabirds had a limited, highly synchronized breeding season swamping the rats with enormous numbers over a short period of time. As indicated by Alexander (1922) the prey availability for cats outside this season was limited. Low food abundance on Rat Island during the autumn and winter may have reduced rat numbers to relatively low levels at the start of spring and the onset of seabird breeding. However, the cats could switch from seabird to rat (and rabbit) consumption in autumn and winter and consequently maintain their numbers between seabird breeding seasons. The depredations from the relatively high cat populations during the breeding seasons were then sufficient to drive rapid decline in the Common Noddy and Sooty Tern colonies on Rat Island. As the seabirds declined cat predation on the rats increased and both populations were probably reduced to lower levels.

2.4 Impact of Fishing Communities

The eastern edge of Rat Island provides sheltered access to deep water and has provided suitable sites for fishing camps since at least the 1940s. Since World War II these camps have been occupied by Rock Lobster fishermen. There have been up to 59 Rock Lobster camps on Rat Island (currently over 70 shacks) and one camp associated with the operations of a cultured pearl lease. The foundations of a number of abandoned camps occur to the

5 south of the current settled area. The settlement footprint of Rat Island, including the airstrip and access pathways, is approximately 14 Ha or 22.9% of the total area. However much of this area was formerly mined for guano and would have little additional impact on ƚŚĞ/ƐůĂŶĚ͛ƐďŝŽĚŝǀĞƌƐŝƚǇ.

Both the guano miners and fishers living on Rat Island in the early 1900s probably harvested seabird eggs. Sooty Tern eggs have been a subsistence food and commercially harvested in the Western Indian Ocean for at least a century, a practice that continues today (Ridley & Percy 1958, Feare & Gill 1997). It is unlikely that this was a significant cause of colony decline.

Gardens would be virtually impossible to maintain in the exposed, worked out and saline surface of Rat Island. However, apparently to increase the local amenity the lobster fishers have translocated a number of exotic succulent plants and cultivated them around the camps. Some of these are becoming naturalised and invading areas outside the settlement, mainly from abandoned camps. Three such ƐƉĞĐŝĞƐ͚DŽƚŚĞƌŽĨDŝůůŝŽŶƐ͛Bryophyllum delagoensi , Prickley Pear Opuntia stricta and Cotyledon obiculare have the potential to become serous ecological weeds on Rat Island and elsewhere in the Abrolhos. Many fishing camps will require decommissioning in the next few years due to a contraction in the number of operating Zone A licences in the fishery. This process will need to be carefully managed to prevent further impacts on the occupied islands.

3. R A T ISL A ND R E C O V E R Y

The demise of Rat Island from arguably the most important seabird colony in the eastern Indian Ocean to a barren and relatively lifeless guano quarry has been quite well documented. Key processes that created and maintained WKH,VODQG¶Vterrestrial ecosystem were suspended by the loss of its marine fauna (seabirds) following the introduction and interaction of exotic predators. Thousands of years of soil development, driven by the transfer of marine nutrients from the ocean to the land by the seabirds, was reversed with the removal of guano by the 6WDWH¶VILUVWH[SRUWPLQLQJLQGXVWU\

In 1991 a group of inspired government officers carried out a successful program to eradicate the Black Rats, and ultimately the feral cats from Rat Island. The objective of this intervention was probably more in line with removing Rat Island as a foothold for these predators in the Abrolhos, to prevent them island-hopping elsewhere in the archipelago. The outcome however has probably exceeded those expectations as their intervention probably represents the first critical step in the ecological restoration of Rat Island itself. This section in the management plan documents the rat-cat eradication program and the current (2011- 2013) condition of the Rat Island ecosystem as determined by the Rat Island Recovery Project (RIRP).

3.1 E radication of Introduced Predators

In 1991 Andrew Burbidge and Phil Fuller of the Department of Conservation & Land Management, Randal Owens from the Department of Fisheries and Ken Johnson of the CSIRO undertook a program to eradicate Black Rats & cats from Rat Island. The objective of this intervention was primarily to prevent Black Rats from using the Rat Island group from becoming a platform for the invasion of other islands in the Easter Group and ultimately the rest of the Abrolhos, i.e. it was conceived as a bio-security rather than a recovery measure.

Rat Island and its near neighbours (Bushby, Little Rat, Roma, Little Roma and Dry) were baited with oats vacuum-impregnated with the anticoagulant rodenticide Pindone. White Bank, just north of Rat Island, was inspected for rat sign, but no sign was found and it was not baited.

Baiting commenced on 30 November and was completed on 7 December 1991. About one cupful of bait was placed into thin plastic bags, which were placed on the ground, in a 50 m grid on the larger islands, and more densely spaced on smaller islands. Baits were inspected every three to four days and replaced if partially or completely consumed. All partly- consumed bait bags were replaced with full bags at the end of the project and an additional bag of bait was placed in the centre of each 50 m by 50 m grid at project completion.

Cat control was conducted for several months after the rat baiting using traps. No rats were observed after the baiting program was completed but one or two cats persisted for some years. The last surviving cat died in around 2000 (Russell Dyson pers.com).

A study of New Zealand seabird islands where rats have been eradicated used the stable isotope methodology to demonstrate that the recovery of impaired terrestrial ecological processes occurred with 12-22 years (Holly 2010). Rats were eradicated from Rat Island 22 years ago and cats ultimately about 12 years ago. Seabird return was first observed in 2003 and has accelerated significantly in the last two years. The RIRP has been benchmarking the Rat Island ecosystem at the point where seabird re-colonization may become a significant driver of changes in the recovering terrestrial ecosystem.

3.2 The Terrestrial Ecosystem Following Predator Eradication

3.2.1 Rat Island Recovery Project (RIRP) - Survey & Assessment Methods

The Rat Island Recovery Project has not at this stage attempted to inventory the entire biodiversity of Rat Island post the eradication of introduced predators. Sampling has been primarily directed at documenting the process of seabird re-colonisation, re-surveying the flora and vegetation and collecting samples of plant and animal tissues for carbon & nitrogen stable-isotope analysis.

Vegetation Sampling & Mapping

RIRP Flora Survey

The August 2012 field visit was used to update the survey previously conducted by RIRP in 2003 (Dunlop & Rippey 2004) and previous published flora surveys (particularly Harvey et al. 2001). RIRP has lodged sixty-four (64) species with WA Herbarium, 29 not vouchered before and 15 not recorded previously for Rat Island. The majority of the 83 species recorded for the Island have now been vouchered with the WA Herbarium.

Vegetation Mapping

Harvey et al. (2001) produced a vegetation map of Rat Island which remains a useful representation today. Most of the boundaries on this map remain clearly visible on the most recent geo-referenced aerial photography ( Coastline Aerial Imagery ʹ Landgate WA), 14 years after their survey in 1999. Harvey et al. (2001) used the life-form density classes of Beard (1981) to document the floristics and structure of each vegetation stratum. The RIRP has opted to simplify this classification by describing vegetation types using only the dominant stratum and removing the need to use complex codes.

In February 2013 the RIRP nested fifteen 20m line intercept transects within the vegetation types delineated by Harvey et al. (2001) (GIS Map 2). The percentage cover intercepted by

8 each perennial species was recorded within each consecutive 1m section of each transect. A mean percentage cover was calculated for each perennial species on each transect. Also included in the vegetation survey were 20 m line-intercept transects placed diagonally across the three detailed ecological study sites RB, RNB and MNB (GIS Map 2)

Stable Isotope Sampling

Carbon and nitrogen stable isotope analysis was used to provide a baseline food web structure for the Rat Island ecosystem for future monitoring of the anticipated changes in nutrient sources, energy flow and productivity associated with the resumption of the marine subsidy from seabird re-colonization (Mulder et al. 2011). For details of the application of stable isotope methods to terrestrial food webs see for example Dunlop & Bullen 2011.

In February 2011 RIRP established three 400m2 detailed study sites in which future stable- isotope sampling was to be concentrated (GIS Map 2). Two sites were located on in remnant unmined Nitraria billarieri heath (RB, RNB) and the third in a mined quarry area sparsely vegetated with Atriplex bunburyana (MNB). At that time the RNB site had no history of nesting seabirds and the RB sites was occupied by Sooty Terns for the first time in that season. Site MNB was outside the area occupied by Sooty or Bridled Terns in 2011. During December 2012 RB and MNB were occupied by nesting Sooty Terns and a few nests encroached at site RNB late in the season (February 2013).

To capture invertebrates and reptiles RIRP established traplines at RB at RNB each consisting of four (160mm diameter, 145mm diameter) dry pitfall traps along an 8m length of drift fence (gutter-guard). Two funnel traps were also placed on each drift-fence. This form of trapping was not possible at MNB due to a lack of friable substrate. Day and night searches were also conducted for invertebrates and reptiles on the study sites and elsewhere.

House Mice were sampled in all three detailed ecological study sites using plastic break-back traps set only at night to avoid non-target vertebrates.

Breast feathers were collected from samples of Bridled and Sooty Tern chicks in February 2012, along with Flying Fish regurgitated by the Sooty Terns. These tissues were used to provide the stable isotope signatures of animal tissues transferred from the surrounding oceanic ecosystem.

The following tissues were prepared for stable isotope analysis by drying and then grinding to a powder.

Nitre Bush (Nitraria billardieri )- leaves from branch tips

Saltbush (Atriplex bunburyana) - leaves from branch tips

Golden-rumped Ant (Polyrachis ammonoeides ʹchitin exoskeleton

Trap-jaw Ants (Odontomachus ruficep ) ʹ chitin exoskeleton

Skink (Ctenotus fallens) ʹ amputated tail tips

House House (Mus domesticus) ʹ dorsal fur

Sooty Tern (Onychoprion fuscata) ʹ fledgling breast feathers

Bridled Tern (Onychoprion anaethetus) - fledgling breast feathers

Regurgitated Flying Fish (Cypselurus sp) ʹ pectoral fins

The same terrestrial plant and animal tissues were collected from the detailed study sites in April 2013, one growing season after the occupation of site RB with Sooty Terns. This sampling was used to increase RIRPs existing sample sizes and to test for short-term changes in nutrient availability and trophic relationships.

The carbon and nitrogen analysis stable isotope analysis was conducted by the specialist laboratory at the Faculty of Science, Monash University.

3.2.2 Vegetation

Harvey et al (2001) reported 71 species from Rat Island, 37 native and 34 (48%) introduced. The RIRP has increased this flora to 96 recorded species of which 45 (47%) are introduced (Appendix 1). This ratio of native to introduced species is fairly typical of continental seabird islands of south-western Australia (Rippey et al 2002).

The results from the line intercept transects are summarized in Appendix 2, with each transect allocated to its vegetation type.

Twelve vegetation types have been utilised in the revised vegetation map (GIS Map 3). Based on the line intercept sampling some of the areas drawn by Harvey et al. 2001 have been merged with adjoining units. One new vegetation type has been recognized (6 - Atriplex bunburyana sparse heath). This unit broadly corresponds with the centre of the Western Central Guano Field (GIS Map 3).

Six dominant species were identified within twelve vegetation units. These are identified by their vegetation type numbers on GIS Map 3.

1. Atriplex cinerea low heath 2. Tecticornia indica succulent heath 3. Frankenia pauciflora, Tecticornia indica low open heath 4. Atriplex bunburyana low open heath 5. Atriplex bunburyana heath

6. Atriplex bunburyana sparse heath 7. Threlkeldia diffusa, Atriplex bunburyana low open heath 8. Nitraria billardierei tall heath 9. Nitraria billardierei tall closed heath. 10. Frankenia pauciflora low open heath. 11. Annual herbfield 12. Settled area

3.2.3 Invertebrate Fauna

During the course of the Rat Island Recovery Project two relatively abundant ant species collected for use in stable isotope analysis and other invertebrates were recorded during the sampling process.

A few indigenous species of ants dominated the epigaeic invertebrate fauna, the most abundant and widespread being the Golden-rumped Ant Polyrachis ammonoeides and Trap- jaw Ant Odontomachus ruficep (Heterick 2009). P. ammonoeides has also been collected from Morley Island (B. Heterick pers.com), observed on Wooded Island and is common on coastal environments between Dongara and Shark Bay. The Iridomyrmex spp. that dominate mainland ant communities are scarce on Rat Island although one species in the I.rufoniger group was eventually collected. The other ants recorded during the current survey were species of Solenopsis, Camponotus and Rhytidoponera (Heterick 2009).

Polyrachis ammoneoides a coastal ant species from the Mid-west Region.

Low abundances of beetles, grasshoppers, earwigs and spiders have also been observed. The invertebrate top-predator recorded on Rat Island was an indigenous Long-legged Centipede (Scutigeridae) in the genus Thereuopoda.

Some of the invertebrates recorded were introduced species including the land snail Theba pisana that may have a significant impact on the islands vegetation and the domestic Cockroach Biatta orientalis.

3.2.4 Vertebrate Fauna

Reptiles There are historical records (WA Museum collection) of 7 reptile species from Rat Island How et al. 2004). Of these only four were observed during field surveys in 2012 and 2013 which included 96 trap days with pitfall & drift traplines and 62 trap/days with funnel traps. These species were Ctenotus fallens, Cryptoblepharus carnabyi, Delma australis and Christinus marmoratus. Other species such as Crenodactylus ocellatus, Lerista distinguenda and Menetia greyii must either, be present at extremely low densities making detection difficult, or have recently been extirpated from the Island. The explosion in the abundance of House Mice Mus domesticus following the eradication of the rats in 1991 may be a contributing factor.

Maryan et al. (2005) recently added two gecko species to the Rat Island reptile fauna, Gehyra variegata and Heteronotia binoei. We observed both of these indigenous geckoes to be associated with human habitation on the Island and these may have been accidently introduced by the operators of the Rock Lobster fishing camps.

Land Birds The land-bird fauna of Rat Island would always have been limited to a few species, including the Western Silvereyes, Welcome Swallows, Brush Bronzewings and Spotless Crakes that have been observed during this survey. Habitat changes (clearing and human habitation) may have allowed the Australian Kestrels to become resident. A vagrant Shining Bronze Cuckoo was present on the Island in August 2012.

House Mice House Mice Mus domesticus were not documented as being present on Rat Island until 1989 although they are likely to have been present for much longer as Black Rats are significant predators of mice and suppress their numbers (Caut et al.2007). Mice are currently abundant on Rat Island with high densities of tracks and burrows in areas with remaining sandy substrate. All break-back traps set overnight either trapped or were sprung by mice.

3.2.5 Stable Isotope Analysis of the Terrestrial Food Web

The vegetation-transect data from within the mapped vegetation types (GIS Maps 1 and 2 ) was used to estimate the total perennial vegetation cover. About 35 % of the ground- surface of Rat Island is vegetated with perennial plants.

The same data set was used to estimate the proportion of the perennial vegetation cover made up of C3 or C4 plant species. Plants with the C4 photosynthetic pathway (Atriplex bunburyana, A. cinerea and Tecticornia indica) make up 67.8 % of perennial vegetation cover. The dominance of C4 plants appears to be an unusual characteristic at this latitude and may be a combined response to climate and the extensive mined-out area in the landscape. Nitre Bush Nitraria billardieii is the dominant shrub on most of the other smaller ͚ƐĞĂďŝƌĚŝƐůĂŶĚƐ͛͘dŚŝs C3 plant is currently restricted to 3.12 Ha of unmined surface on Rat Island.

The bio-chemical photosynthetic pathways used by C3 and C4 plants produce contrasting fractionations of heavy 13C verses normal 12C carbon atoms. This results in distinctive carbon stable isotope ratios (Delta Carbon, ɷ13C) in C3 and C4 plants. Figure 2 shows the carbon signatures from Nitraria billardieri samples from the detailed study plots RNB and RB and Ariplex bunburyana samples from MNB. The Nitraria signature lies within the typical ɷ13C range for C3 plants and Atriplex for C4. The ɷ13C values for the Nitraria samples from RNB and RB were not significantly different.

The ɷ13C values in the tissues of producers (plants) are transmitted more or less unaltered to consumers in the food chain (e.g. to grazers, omnivores, predators, parasites and top- predators). The suite of predators sampled within the detailed study sites on Rat Island all have ɷ13C values intermediate between Atriplex (C4) and Nitraria (C3) indicating dependence on both sets of producers for energy but perhaps a disproportionate reliance on the C3 food chain in the remnant unmined vegetation as this perennial vegetation provided only about 32% of the available plant biomass.

Consumer organisms differentially excrete the normal 14N atoms in their diet and concentrate the heavier 15N stable isotope in their proteins. Consequently the stable isotope ratio for nitrogĞŶ;ĞůƚĂEŝƚƌŽŐĞŶ͕ɷ15N) undergoes at stepwise increase (enrichment) of o about 2.5 /00, with each trophic level. This pattern in fractionation can be used to infer ƚƌŽƉŚŝĐůĞǀĞůŝĨƚŚĞďĂƐĞůŝŶĞɷ15N in the producers is known (Kolb 2011).

13 15 o Figure 2: The ɷ ĂŶĚɷ N values ( /oo) of plant and animal tissues collected from the detailed ecological study sites on Rat Island in February 2012 (pre seabirds) and April 2013 (post seabirds). Values are means and error bars are standard deviations.

Nitraria pre seabirds 25 Nitraria post seabirds Atriplex pre seabirds

Odontomachus pre 20 seabirds Polyrhachis pre seabirds Nitrogen

Polyrachis post

15 seabirds Ctenotus pre 15 seabirds Ctenotus post

Delta seabirds Mus pre seabirds 10 Mus post seabirds

5 -26 -21 -16 -11 -6 -1 Delta 13 Carbon

Consumer organisms differentially excrete the normal 14N atoms in their diet and concentrate the heavier 15N stable isotope in their proteins. Consequently the stable isotope ratio for nitrogen (Delta Nitrogen, ɷ15N) undergoes at stepwise increase (enrichment) of o about 2.5 /00, with each trophic level. This pattern in fractionation can be used to infer trophic level if the baseline ɷ15N in the producers is known (Kolb 2011).

Amongst the consumer organisms sampled on Rat Island the omnivorous Golden-rumped Ants (Polyrachis) occupied the lowest mean trophic level whilst the highest order consumers were the predatory Trap-Jaw Ants (Odontomachus). The insectivorous skinks (Ctenotus fallens) occupied a significantly lower mean trophic level than the Trap-jaw ants indicating that these ants probably take more predacious invertebrates such as spiders. Like most ground spiders the Trap-Jaw Ants were nocturnal.

House Mice had broad habitat preferences in terms of individual variation in energy sources (ɷ13C values) and the results indicate that they were largely insectivorous (Figure 2) on Rat Island.

The results from the re-sampling in the detailed study sites in April 2013 suggested 15N enriched nitrogen contributed by the colonizing seabirds might already be present in the tissues of the Nitre Bush, Polyrachis ants, Coast Skinks (Ctenotus) and House Mice. However the differences in ɷ15N between pre and post seabird study sites are not as yet statistically significant (Figure 2). The results do not indicate much direct utilization of seabird tissues or prey remains, even in the House Mice at this stage.

The ɷ15N values in the Rat Island terrestrial ecosystem were extraordinarily high. Total N and ɷ15N normally increase in environments receiving a nitrogen subsidy from seabird colonies (Kolb 2011). The high ɷ15N enrichment values reflect the accumulation of 15 N in the marine food chain and the rapid loss of 14N through volatizing ammonia, leaving inorganic nitrogen rich in 15N (Kolb 2011). However the very high ɷ15N on Rat Island cannot reflect a recent (last century) heavy source of nitrogen deposition. Nitrogen is almost certainly extremely limiting on Rat Island at present but what remains must be entirely enriched inorganic nitrogen from the original guano deposits. Nitrogen mineralization perhaps associated with phosphate minerals that formed at the base of the original guano mantle may have prevented the loss of nitrate to leaching.

3.3 Seabird Re-colonization

3.3.1 Seabird Survey Methods

The RIRP made systematic day and night searches for breeding seabirds on each of its field surveys conducted in December 2003, December 2008, February 2012, August 2012, December 2012 and February 2013. The number of breeding pairs was censused by locating and counting nest sites, aerial counts of birds flying over colonies or by mapping colony boundaries and transect-based density sampling. The size of the large Sooty Tern colony in December 2012 was estimated by GPS mapping the colony boundaries and then estimating nest densities with seven randomly located 50 by 2m (100m2) belt transects. These transects were done with head-torches at night when birds tended to remain incubating on the nest in the presence of the observers.

Adult and fledgling Sooty and Bridled Terns were individually marked with numbered alloy bands (supplied by the Australian Bird & Bat Banding Scheme) to enable future estimates of site and area fidelity in the adults and colony philopatry in the progeny reared on Rat Island.

3.3.2 Seabird Re-colonization Trends

A survey in December 2003 recorded the first tentative indications of seabird re- colonization and this was attributed to the successful eradication of the Black Rat in 1991 by the then Department of Conservation and Land Management and the demise of the last feral cat around 2000 (Dunlop & Rippey 2004). At that stage there were 6 breeding pairs of Bridled Terns nesting near the southern end of the Island. By 2008 there were between 50 and 100 pairs breeding at the southern end of the Island most with nests under rock-piles left by the guano miners.

Fairy Terns naturally nest in the open on beaches, coral rubble or other reflective substrates with a preference for sites that provide nearby cover for chicks. The denuded, mined out limestone surfaces (with nearby rock-piles) on Rat Island would appear to be signature habitat for this species as between 500 and 1000 pairs were nesting there in December 2008 (Figure 3).

Fairy Tern nest in mined out habitat.

The numbers of seabirds recorded breeding on Rat Island since 2003 are presented in Table 1. The spatial areas of the colonies prior to 2012 are presented on GIS Map 4. The colonies documented over the 2012/13 season are outlined on GIS Map 5. By 2012 (12 years after the demise of the last feral cat) eight seabird species had returned to breed on Rat Island. In 2012/13 this involved an estimated 72,923 breeding pairs with the vast majority being Sooty Terns.

During December 2012 RIRP checked most of the other Easter Group islands known to have housed Sooty Tern colonies (Burbidge et al.1996, C.A, Surman pers.comm and pers.obs.) including Wooded, Morley, Leo, Campbell, Suomi, Keru, Serventy, Alexander and Gilbert Islands. All these islands were vacant suggesting that the entire Easter Group Sooty Tern population was on Rat Island in that year.

Table 1: RIRP records of breeding seabird numbers, colony areas and estimated colony densities on Rat Island from 2003-2013.

SEABIRD SPECIES DATE AREA OCCUPIED ESTIMATED COLONY SIZE ESTIMATED (hectares) DENSITY (pairs / (Pairs on Rat hectare Island)

Bridled Tern 2003 Dec-2003 - - 6

Bridled Tern 2008 Dec-2008 - - 50-100

Bridled Tern 2011 Feb - 2012 6.9 25 174

Bridled Tern 2012 Dec - 2012 8.4 25 210

Fairy Tern 2008 Dec - 2008 3.0 - 750

Sooty Tern 2011 Feb-2012 2.0 - 5000

Sooty Tern 2012 Dec-2012 29.52 2400 72324

Roseate Tern Dec-2012 0.09 - 300

White-faced Storm- Aug-2012 0.14 51 27 petrel 2012

Caspian Tern 2012 Aug & Dec- - - 1 2012

Pacific Gull 2012 Dec-2012 - - 1

Silver Gull April 2013 1.35 44 60

Bridled Tern populations breeding off south-western Australia have expanding due to trophic shifts associated with regional changes in ocean climate (Dunlop & Surman 2012). It is therefore not surprising that these were the first seabirds to recolonise and their presence ŝŶƚŚĞŶĞƐƚŝŶŐŚĂďŝƚĂƚŵĂǇŚĂǀĞĨĂĐŝůŝƚĂƚĞĚƚŚĞƐĞƚƚůŝŶŐŽĨŽƚŚĞƌƐƉĞĐŝĞƐ;ƐĞĞ͛ŝŶĨŽƌŵĂƚŝŽŶ ďĂƌƌŝĞƌ͛ĞĨĨĞĐƚŝŶƵŶůŽƉϮϬϬϵͿ͘

The establishment of a colony of White-faced Storm-petrels is particularly remarkable as these small, burrow nesting seabirds are particularly vulnerable to Black Rat and even House Mouse predation and would not have been able to persist on Rat Island even prior to the introduction of the cats (Towns et al. 2011).

4. RAT ISLAND MANAGEMENT ʹ RECOVERY OR RESTORATION?

The eradication of Black Rats and feral cats from Rat Island in the 1990s was the initial critical intervention in the recovery of the terrestrial ecosystem by providing for the return of breeding seabirds. Ironically the relatively rapid return of some tropical species such as the Sooty Tern may have resulted from colony instability due to recurrent breeding failure on the other islands in the Easter Group caused by a probable decline in marine productivity in the oceanic waters off mid ʹ Western Australia (Surman et al. 2012, Dunlop 2011).

The marine nutrient and energy resources, transferred by these birds from sea to land, are anticipated to drive the partial recovery of the terrestrial ecosystem by rebuilding the soil organic matter and increasing primary productivity. Changes in the flora might also be predicted with the increased nutrient availability favouring faster-growing species including regional ornithocoprophilic plants. Invertebrate and reptile populations are likely to increase in abundance in response to the increased primary productivity and more species, from the regional pool of volant species, may be able to colonise and increase the islands biodiversity.

Ecological recovery could however be limited or potentially derailed by some un- remediated human induced changes to Rat Island including the removal of much of the soil by guano mining, introduced plants and animals, habitat degradation, human disturbance and the extent of the settlement footprint.

Future objectives for the management of Rat Island could settle for accommodating the changes now being driven by the resumption of natural processes (recovery) or they could specify further interventions towards ecological restoration (Clewell & Aronson 2013). That is, taking actions intended to return the biodiversity of Rat Island closer to its original undisturbed condition. These interventions might include the eradication of some potential ecological weeds, the eradication of the House Mouse population, the restoration of some habitat areas with remnant soil for burrow-nesting seabird species and the re-introduction of the extirpated reptiles.

4.1 Managing Recovery

4.1.1 Emerging Wildlife and Ecosystem Management Issues

Communications Tower

The wire stays supporting the FESA (Fire & Emergency Services) communications tower (GIS Map 6) on Rat Island have been a minor nocturnal collision hazard for seabirds since at least December 2003, when a number of Shearwater carcasses were observed (J.N. Dunlop pers.obs). However with the rapid increase in the number of breeding seabirds over the last two seasons the problem of protected species mortality from aerial collisions with these stays has become significant.

This season (2012/13) the RIRP counted the fresh kills and maimed birds in the area beneath and immediately surrounding the array of stays supporting the tower. By 9 December there had been 49 Sooty Tern casualties with a further 59 between that date and 1 February. Sooty Terns are particularly vulnerable because they are particularly active overhead at night. RIRP also recorded collision - related mortalities in 2 Wedge-tailed Shearwaters, a juvenile White-breasted Sea-eagle and a Pacific Gull around the communications tower.

It is suggested that white or fluorescent rubber hosing could be threaded onto the wire stays. This would increase the diameter of the lines making them more visible and also soften the impact of any collisions that did occur.

Airstrip Operations

RIRP observations this season indicate that there was little interaction between the large numbers of nesting Sooty Terns and aircraft during landing or take-off. Geraldton Air Charter report only occasional collisions with seabirds. The cleared buffer zones around the strip appear to be adequate to maintain separation between birds and aircraft at the current seabird densities.

However whilst flat open areas prevent occupation by most breeding seabirds they may be attractive to others, particularly if these areas have a pale, reflective substrate colour resembling a beach or shoreline. Fairy Terns Sternula nereis are sometimes attracted to nest in such areas and have recently nested in flat, open areas in the mined out landscape on Rat Island (GIS Map 4). In February 2013 a post- breeding flock of 200-300 individuals was roosting at night on the recently constructed taxi apron at the northern end of the airstrip. This signals that this area may be attractive as a breeding site in future seasons. It is recommended that the crushed limestone surface be spray-painted with black to produce a checkerboard of zebra pattern to reduce its attractiveness.

Another approach might be to prepare an alternative area elsewhere on the island away from the airstrip and its approaches. One recently quarried site of white worked limestone in the western central mining field might be further enhanced by spreading shell material (sourced from a Pacific Gull anvil site) to provide an attractive surface for breeding Fairy Terns (GIS Map 6).

It is recommended that:

1. Air charter operators maintain a record of any bird-strike incidents at Rat Island.

2. The existing buffer zone / apron be maintained free of cover including rocks, junk and woody shrubs that may spread the nesting habitat. 3. The airstrip area is searched for bird casualties on a monthly basis between October and March. 4. If monitoring shows a significant increase in bird-strike incidents the situation should be investigated and a management strategy devised. 5. That the Fairy Tern management measures described above be implemented.

Human Disturbance in Seabird Colonies

Rat Island is a multiple use island with a settlement of WRL camps, an aquaculture camp, the Saville-Kent Research Facility and an airstrip servicing the Easter Group. The return of the seabirds to Rat Island will increase the frequency of interactions between breeding birds and human activities. The agencies responsible for management now must consider how to ensure compliance with the Wildlife Conservation Act 1950 by protecting breeding seabirds from excessive disturbance by people.

Recent changes in the management of the WRL Fishery have significantly reduced the number of fishers and this will ultimately reduce the number of camps on islands. These changes have also lead to the removal of the closed season meaning the fishers could now occupy their camps at any time during the year. A pattern of fishing and therefore occupancy will eventually emerge but at present this is difficult to predict. Previously Abrolhos (A Zone) fishers operated from March to June, a period outside the breeding period for all the seabirds currently using Rat Island. As this is no longer the case some consideration needs to be given to managing human activities in and around seabird breeding colonies.

At present the Abrolhos Islands have no regulations that provide for zoning or access controls on the islands of the Ministerial Reserve (as is the case for DEC managed Nature Reserves) other than the prohibition of camping. This limits the current capacity to regulate public behaviour and protect critical wildlife habitats or aggregations. The introduction of appropriate wildlife protection regulations would appear an essential precursor to the promoted increase in tourism activities in the Archipelago.

In the case of multiple use islands where human activities must be accommodated the best strategy is to manage human movement patterns to promote ͚habituation͛. This is a dampening of flight or defensive behaviour in response to familiar, benign stimuli. The result manifests in increased site tenacity during periods of human intrusion and a reduction in egg and chick losses caused by extreme environmental conditions, territorial disputes and predation.

To promote habituation human incursions into or adjacent to breeding colonies need to be regular ʹ occur at sufficient frequency to become familiar,

20 consistent ʹ using an established route or pattern of movement, moderated ʹ activity should be quiet, passive, involve small groups and be sensitive to seabird flight responses and predator behaviour, and time limited ʹ duration of human presence should be short.

The following measures are recommended for Rat Island to minimise the disturbance to seabird colonies and promote habituation.

1. The current perimeter access track (GPS Map 6) should be formalised with explanatory and interpretive signage, guide posts or other markers, and hardening (e.g. boardwalks) in some locations. 2. The southern access track to the airstrip should be closed between October and March except for emergency vehicles or for research and monitoring activities. 3. Consideration should be given to opening and managing the main trolley-line embankment (GIS Maps 1 & 6) as a central access way and as a part of a heritage trail explaining the islands history colonial as a guano mine. 4. Movements off the proposed access track system should be prohibited between October and March. 5. The boundaries to the White-faced Storm-petrel colony should be marked to protect burrows from trampling.

4.1.2 Management of Ecological Weeds

A number of species of exotic succulent (rockery) plants have been cultivated around the rock-lobster camps and commo use buildings over time. These include Houseleek Tree Aeonium arboretum, Agave Agave americana, Prickly Pear Opuntia stricta, Money Tree Portulacaria afra, Pigs Ear Cotyledon orbiculata and Mother of Millions Brypohyllum delagoense. Opuntia and Cotyledon are known invasive plants in a variety of contexts including in island ecosystems. Of particular concern at present is Bryophyllum as it is clearly spreading outside its original area of cultivation on the Island. The current extent of the outbreak is shown on GIS Map 7. Bryophyllum is a declared agricultural weed in Queensland partly because it is poisonous and a risk to stock (Queensland Department of Employment, Economic Development & Innovation 2011).

During 2012 the RIRP established 3 (100m2) trial treatment plots to investigate possible methods for Bryophyllum eradication on Rat Island (GIS Map 2). The treatments were as follows:

Hand weeded: In February 2012 all the Bryophyllum plants (including fragments) were hand ǁĞĞĚĞĚĨŝůůŝŶŐϭϰůĂƌŐĞďůĂĐŬƉůĂƐƚŝĐŐĂƌďĂŐĞďŝŶůŝŶĞƌďĂŐƐ͘dŚĞƉůĂŶƚƐǁĞƌĞƐŽůĂƌ͚ĐŽŽŬĞĚ͛ŝŶ the bags for a couple pf months and then incinerated.

Herbicide: The recommended herbicide for Bryophyllum control (Grazon) was not used on Rat Island due to its high solubility and eco-toxicity in aquatic environments. Instead RIRP

21 sprayed the Bryophyllum foliage with 1.5 litres of 7.2 g/L glyphosate (isopropylamine salt) during a period of dry and still weather on 18 August 2012.

Black Plastic: On 10 December 2012 the third treatment plot was covered with heavy duty black landscaping plastic after pruning the woody vegetation to ground level. The plastic was removed 3 February 2012 after nearly 2 months of solar heat treatment.

The treatment plots were located in an area of dense Bryophyllum infestation between the Fishing and Aquaculture Body Corporate leases (GIS Map 2).

Hand-weeding Bryophyllum from a treatment plot

Table 2; The results of the trial Bryophyllum treatments are presented in Table 2. The cover of Bryophyllum at each stage is the mean percentage cover per metre along a 15m diagonal line-intercept transect.

Plot / Treatment Date Established Date Monitored % Cover Bryophyllum (percentage reduction) Plot 1/hand weeded 12 Feb 2012 13.9 18 Aug 2012 1.9 (86.4) 3 Feb 2013 4.3 (69.1) Plot 2/ Glyphosate 18 Aug 2012 7.93 3 Feb 2013 0.93(88.3) Plot 3 / Black Plastic 10 Dec 2012 28.40 18 Apr 2013 9.8 (65.5)

The results indicate that hand-weeding left enough fragmented material or seed to support regeneration over time and was of course very labour intensive. The herbicide treatment

22 was effective but would require retreatment persistent re-treatment to effect eradication over time. The Bryophyllum plants responded to the heating and shading from the black plastic with an increased growth rate through stem elongation. . However significant mortality occurred after the black plastic was removed. This method could be useful around decommissioned lobster camps.

It is recommended that:

1. A program to eradicate Bryophyllum and other introduced succulents outside the Rat Island Body Corporate Area be commenced as soon as resources become available. 2. The introduced succulents within the Body Corporate Area be removed progressively beginning with those around WRL fishing camps that are due for decommissioning following the retirement of fishing entitlements.

4.2 Restoration Projects

4.2.1 Eradication of the House Mouse

House Mice are rarely as significant as predators of seabirds as rats although exceptions have been reported mainly from sub-Antarctic Islands with extreme environmental conditions (Angel et al. 2009, Towns et al.2011). However in the absence of indigenous mammals or introduced rats they can become extremely abundant on islands. At high population densities mice have negative effects on a wide range of ecosystem components as they may function as grazers, seed harvesters, insectivores, predators on the soil invertebrates that drive nutrient cycling or as predators of small vertebrates such as geckoes and skinks (Towns et al. 2011, Mackay et al. 2011, Cory et al. 2011). dŚĞ,ŽƵƐĞDŽƵƐĞƉŽƉƵůĂƚŝŽŶŽŶZĂƚ/ƐůĂŶĚǁĂƐƉƌŽďĂďůǇ͚ƌĞůĞĂƐĞĚ͛ďǇƚŚĞĞƌĂĚŝĐĂƚŝŽŶŽĨƚŚĞ Black Rats and currently dominates the terrestrial ecosystem on Rat Island. It is therefore likely that the House Mice will have a significant influence on the trajectory of ecological recovery initiated by the return of the breeding seabirds. At present the House Mice on Rat Island are mainly insectivorous and are probably competing directly with the remaining small reptiles for limited resources. The apparent disappearance of a number of small lizard species including Crenodactylus ocellatus, Lerista distinguenda and Menetia greyii may be the result of direct competition from, or predation by, House Mice. /Ĩ͚ƌĞƐƚŽƌĂƚŝŽŶ͛ŝƐĂŶ objective then House Mouse eradication would appear to be an important first step.

The success rate for House Mouse eradication on islands has been lower than that for the three invasive rat species (Mackay et al. 2011) partly because mice have smaller home ranges. However eradication on Rat Island using the new generation anticoagulant ƌŽĚĞŶƚŝĐŝĚĞƐǁŽƵůĚĂƉƉĞĂƌƚŽŚĂǀĞĂŚŝŐŚƉƌŽďĂďŝůŝƚǇŽĨƐƵĐĐĞƐƐďĞĐĂƵƐĞŽĨƚŚĞŝƐůĂŶĚ͛Ɛ relatively small area, flat uniform terrain, lack of any significant risk to non-target species (thus avoiding the need for complex bait-stations) and controlled public access. The

23 program would need to incorporate the other nearby smaller islands on the Rat Island platform (Little Rat, Roma and Bushby) to avoid reintroduction.

The methods recently used to successfully eradicate House Mice from Montague Island in New South Wales would seem to be appropriate for Rat Island (Cory et al. 2011). This program used commercially available extruded cereal baits containing 20 ppm or 0.02g/Kg of brodifacoum. Bait was distributed by a spreader bucket on a GPS guided helicopter with a maximum spacing on 10m squares. Around buildings the bait was dispersed by hand. Two winter baiting missions were conducted a fortnight apart. No live mice were detected after the first bait application (Cory et al 2011).

Rat Island is not as topographically complex as Montague Island and it should be possible to conduct all the baiting by hand. This may be a lower cost option if supported by volunteers.

4.2.2 Rehabilitation of habitat for burrow-nesting species

Rat Island once supported large numbers of burrow-nesting shearwaters, probably both Wedge-tailed Shearwaters Ardennia pacificus and Little Shearwaters Puffinus assimilis. The complete restoration of these colonies is not feasible due to the soil being removed from ŵŽƌĞƚŚĂŶϴϬйŽĨƚŚĞŝƐůĂŶĚ͛ƐƐƵƌĨĂĐĞ͘,ŽǁĞǀĞƌĂƌĞĂƐŽĨƐĂŶĚǇƐƵďƐƚƌĂƚĞŽĨĨƐƵĨĨŝĐŝĞŶƚ depth remain along the western seaward edge of the island and at the northern end in an open area now vegetated with only exotic annual herbs and grasses (GPS Map 3).

Any action involving the modification of the land surface of Rat Island needs to consider the location of heritage features. No sites on Rat Island currently have current legal heritage protection but some features may do so in the future (GIS Map 1). A flat sandy area of about 1 Ha near the north-eastern corner of the island with little remaining perennial vegetation appears to be free of heritage constraints (GIS Map 6). This area would be suitable for burrow-nesting seabirds if rehabilitated.

The suggested treatment of this area involves several steps.

1. Early spring spraying of the annual cover of iceplant and exotic grasses with Round-up. 2. The physical removal of the dead (salt-laden) vegetation with a bobcat or rake. 3. The mounding up of the flat sandy area into a series of low berms and swales. 4. Planting and seeding with Atriplex cinerea, Nitraria billardieri and Carpobrotus virescens (Rat Island provenance material).

A small borrow-pit area on the edge of the island west of the airstrip (GIS Map 6) could be similarly rehabilitated to provide extra habitat for colonizing White-faced Storm-petrels.

4.2.3 Re-introduction of extirpated reptile species.

At least one reptile species, the Spiny-tailed Skink Egernia stokesii, was extirpated from Rat number of other reptiles is currently uncertain but these may have been casualties of the House Mouse population explosion that followed the eradication of Black Rats.

The Spiny-tailed Skink (including its sub-species) is a threatened species under both Commonwealth and State legislation and the subject of a recovery plan (Pearson 2012). Populations persist on a number of other Abrolhos Islands particularly in the Wallabi Group (How et al. 2004). This provides an opportunity to re-establish local E. stokesii on Rat Island at the appropriate stage in the restoration process once the ecological conditions are considered suitable. House Mouse eradication would be an important precursor to reptile re-introductions including that of the Spiny-tailed Skink.

4.3 Research Opportunities

The Saville - Kent Field Station (Abrolhos Islands Research institute) was established on Rat Island in 2003 and since then has hosted a variety of mainly marine research projects in the Houtman Abrolhos. However until recently there have been few attractions for the research community on Rat Island itself. The recent events, triggered by rat and cat eradication two decades ago, have changed that situation significantly.

Very few studies have documented the long-term ecological outcomes of successful introduced predator eradication programs on islands (Jones 2010). The Rat Island Recovery Project (RIRP) may well be the first to attempt the long-term monitoring of ecological ĐŚĂŶŐĞŽŶĂ͚ƌĞƐĐƵĞĚ͛ŝƐůĂŶĚĨƌŽŵƚŚĞƉŽŝŶƚǁŚĞŶƐĞĂďŝƌĚƌĞ-colonization commenced. Continuing government and community support for predator eradication projects may depend on the outcomes of successful programs being scientifically documented and communicated in the public arena. The Saville-Kent Field Station could become an important centre for the research into the terrestrial ecology of islands and the processes of recovery and restoration.

Seabird research on the Abrolhos has always been a logistical challenge due to the lack of operational bases and vessels. The re-colonization of Rat Island by 8 seabird species over recent years provides the opportunity to undertake much more time-consuming and detailed investigations than was previously possible, including the use modern GPS tracking technologies. Seabirds are well-tested indicators of changes in the marine environment. The colonies on Rat Island are strategically positioned to monitor changes in ocean climate, fisheries, mercury, hydrocarbons and plastics pollution and the effectiveness of Commonwealth Marine Reserves.

REFERENCES

Angel, A., Wanless, R.M. & J. Cooper (2009). Review of the impact of the introduced house mouse on islands in the Southern Ocean: are mice equivalent to rats. Biological Invasions 11, 1743-1754.

ůĞǆĂŶĚĞƌ͕t͘͘;ϭϵϮϮͿ͘dŚĞǀĞƌƚĞďƌĂƚĞĨĂƵŶĂŽĨ,ŽƵƚŵĂŶ͛ƐďƌŽůŚŽƐ;ďƌŽůŚŽƐ/ƐůĂŶĚƐͿ͕ Western Australia. Journal of the Linnean Society (Zoology) 34, 457-486.

Beard J.S. (1990). Plant Life of Western Australia. Kangaroo Press NSW.

Burbidge,A.A., Johnstone,R.E. & P.J. Fuller (1996). The status of seabirds in WesternAustralia. Pp 57-71 in Ross,G.J.B., Weaver,K. and J.C. Greig (eds) The Status ŽĨƵƐƚƌĂůŝĂ͛Ɛ^ĞĂďŝƌĚƐ͗WƌŽĐĞĞĚŝŶŐƐŽĨƚŚĞEĂƚŝŽŶĂů^ĞĂďŝƌĚtorkshop, Canberra, 1-2(2003). November 1993. Biodiversity Group, Environment Australia: Canberra.

Caut.,S., J. Casanova, E. Virgos, J. Lozano, G.Witmer G.W. & F. Courchamp (2007). Rats dying formice: modeling the competitor release effect. Austral Ecology 32, 858-868.

Cory, F., Wilson, A., Priddel, D., Carlisle, N. & N. Klomp (2011). Eradication of the House Mouse (Mus musculus) from Montague Island, New South Wales, Australia. Ecological Management & Restoration 12, 102-109.

Clewell, A.F. & J. Aronson (2013). Ecological Restoration. Principles, Values, and Structure of an Emerging Profession. Second Edition. Island Press: Washington.

Dunlop, J.N. (1996). Habituation to human disturbance by breeding Bridled Terns Sterna anaethetus. Corella 20, 13-16.

Dunlop, J.N. (2009). The population dynamics of tropical seabirds establishing frontier colonies on islands of south-western Australia. Marine Ornithology 37, 99-106.

Dunlop, J.N. (2011). Comparative foraging ecology in the dark tern guild breeding off south- western Australiaʹ Insights from stable isotope analysis. Marine Ornithology 39, 201-206.

Dunlop, J.N. & C.A. Surman (2012). The role of foraging ecology in the contrasting responses two dark terns to a changing ocean climate. Marine Ornithology 40, 105-110.

Dunlop, J.N. & R.D. Bullen (2011). Habitat use and trophic structure in a microbat assemblage on the edge of the southern rangelands, Western Australia: insights from stable isotope analysis. The Rangeland Journal 33, 1-7.

Dunlop, J.N. & E. Rippey (2004). Rat Island Recovery Project. A Feasibility Study. Conservation Council (WA) for the Department of Fisheries, Western Australia.

Feare, C.J. (1999). The sustainable exploitation of Sooty Tern eggs in the Seychelles. Seventh Annual Report to the Seychelles Ministry of Environment. Wildwings Bird Management, Haslemere, Surrey, UK.

Gaughan, D.,Surman, C., Moran, M., Burbidge, A. & R.D. Wooller (2002). Feeding ecology of seabirds nesting at the Abrolhos Islands, Western Australia. FRDC Report 1998/203.

Gibson, C.G. (1908). Notes on some birds of the Abrolhos Islands (WA). Emu 8, 64-66.

Harvey, J.M., Alford, J.J., Longman, V.M. & G.Keighery (2001). A flora and vegetation survey of the Houtman Abrolhos, Western Australia. CALMScience 3, 521-623.

How, D. J. Pearson, A. Desmond and B. Maryan (2004). Reappraisal of the reptiles on the islands of the Houtman Abrolhos, Western Australia. Western Australian Naturalist 24, 172- 178.

Heterick, B.E. (2009). A Guide to the Ants of South-western Australia. Records of the Western Australian Museum Supplement 76, 205pp.

Jones, H.P. (2010). Seabird islands take mere decades to recover following rat eradication. Ecological Applications 20, 2075-2080.

Kolb, G.S., Young,H.S. & W.B. Anderson (2011). Effects of Seabirds on Island Consumers. ŚĂƉƚĞƌϳ͘/ŶDƵůĚĞƌ͕W͘,͘ŶĚĞƌƐŽŶ͕t͕͘͘dŽǁŶƐ͕͘Z͘ΘW͘:͘ĞůůŝŶŐŚĂŵĞĚƐ͚͘^ĞĂďŝƌĚ /ƐůĂŶĚƐ͗ĐŽůŽŐǇ͕/ŶǀĂƐŝŽŶĂŶĚZĞƐƚŽƌĂƚŝŽŶ͛͘KǆĨŽƌĚhŶŝǀĞƌƐŝƚǇWƌĞƐƐ͗KǆĨŽƌĚ͘

Mackay, J.W.B., Russell, J.C. & E.C. Murphy (2007). Eradicating House Mice from islands: successes, failures and the way forward. Managing Vertebrate Invasive Species. Paper 27, 294-304.

Maryan, B, Stevenson, C. Pearson, D.J. How R.A. & L.H. Schmitt (2009). Additions to the Reptiles known from islands in the Houtman Abrolhos. Western Australian Naturalist 26

Mulder, P.H. Anderson, W.B., Towns, D.R. & P.J. Bellingham eds. (2011). Seabird Islands: Ecology, Invasion and Restoration. Oxford University Press: Oxford.

Nisbet,I.C.T. (2000). Disturbance, habituation, and management of waterbird colonies. Colonial Waterbirds 23, 312-332.

Pearson, D. (2012). Western Spiny-tailed Skink (Egernia stokesii) Recovery Plan. Wildlife Management Program No.53. Department of Environment & Conservation, Western Australia.

Queensland Department of Employment, Economic Development & Innovation (2011). Fact Sheet. Mother of Millions. Declared Class 2 Pest Plant. Government of Queensland.

Ridley, M.W. & R. Percy (1958). Exploitation of seabirds in the Seychelles. Colonial Research Studies 25, 1-78.

Rippey,E.,Rippey,J.J.,Green,B. & J.N.Dunlop (2002). Comparison of the vegetation of the islands in Shoalwater Bay (Rockingham, Western Australia) with that of the coastal bushland. Journal of the Royal Society of Western Australia 85, 169-179.

Russell, J.C. (2011). Indirect Effects of Introduced Predators on Seabird Islands. Chapter 9. In Mulder, P.H. Anderson, W.B., Towns, D.R. & P.J. Bellingham eds. ͚Seabird Islands: Ecology, /ŶǀĂƐŝŽŶĂŶĚZĞƐƚŽƌĂƚŝŽŶ͛͘KǆĨŽƌĚhŶŝǀĞƌƐŝƚǇPress: Oxford.

Russell, J.C. & Le Corre, M. (2009). Introduced mammal impacts on seabirds in the Iles Eparses, Western Indian Ocean. Marine Ornithology 37, 121-129.

Stanbury, M. (1993). Historic sites of the Easter Group, Houtman Abrolhos,WA. Department of Maritime Archaeology, Western Australian Maritime Museum, Report No.66, Perth.

Surman, C.A., Nicholson, L.W. & J.A. Santora (2012). Effects of climate variability on breeding phenology and performance of tropical seabirds in the eastern Indian Ocean. Marine Ecology Progress Series 454, 147-157.

Storr,G.M., Hanlon, T.M.S & J.N. Dunlop (1983). Herpetofauna of the Geraldton Region, Western Australia. Records of the Western Australian Museum 10, 215-234.

Storr, G.M., Johnstone, R.E. & P.Griffen (1986). Birds of the Houtman Abrolhos, Western Australia. Records of the West Australian Museum Supplement 24, 42pp.

Towns, D.R. Byrd, G.V., Jones, H.R., Rauzon, M.J. Russell, J.C. & C. Wilcox (2011). Impacts of Introduced Predators on Seabirds. Chaper 3. In Mulder, P.H. Anderson, W.B., Towns, D.R. & P.J. Bellingham eds. ͚^ĞĂďŝƌĚ/ƐůĂŶĚƐ͗ĐŽůŽŐǇ͕/ŶǀĂƐŝŽŶĂŶĚZĞƐƚŽƌĂƚŝŽŶ͛͘KǆĨŽƌĚhŶŝǀĞƌƐŝƚǇ Press: Oxford.