Maximising Conservation Impact By Prioritising Islands For .

Island invasives: scaling up to meet the challenge. Ch 3D Strategy: Scaling upis climatically similar (Fig. 1). These islands are mostlytropical and range from small sandy islets of 0.01 ha toislands with mountain ranges and a variety of habitat types 20,000 ha. Only 14 islands are permanently inhabitedby human communities of up to 65,000 people, whilethe remaining islands are either completely uninhabited,function only as tourist resorts or destinations, or arevisited temporarily by fishermen.Selection of potential invasive speciesTo assess biodiversity loss that could result from theinvasion of harmful animal species, we selected the fivemost harmful invasive terrestrial vertebrates (McCreless,et al., 2016) that are widespread in the Caribbean region.Green iguanas (Iguana iguana) are known to hybridiseand compete with native reptiles (Gibbon, et al., 2000;Vuillaume, et al., 2015), small Indian mongoose (Urvaauropunctata) are versatile predators considered one of theworst invasive species (Hays & Conant, 2007; Barun, et al.,2008), brown (Rattus norvegicus) and black rats (R. rattus)and feral cats (Felis catus) are efficient predators that canhave detrimental effects on island biodiversity (Towns,et al., 2006; Jones, et al., 2008; Medina, et al., 2011;Nogales, et al., 2013). These five species are distributedwidely across islands in the Caribbean (Kairo, et al., 2003;Dawson, et al., 2015) and are therefore potential invadersof all islands in the region.Distribution of native and invasive speciesFor each island we previously collated information onthe presence of native and invasive terrestrial vertebratespecies for an eradication prioritisation (Dawson, et al.,2015) and a general inventory of biodiversity (Churchyard,et al., 2016), and updated these previous compilationswith recent information and threat assessments (IUCN,2017). We considered all globally threatened terrestrialvertebrate species (including marine turtles) as listed on theInternational Union for Conservation of Nature Red List ofThreatened Species (IUCN, 2017) and all colonial seabirdspecies and restricted range bird species. We also includedreptiles of conservation concern that are endemic to a singleterritory or inhabit fewer than 15 islands across their range(Dawson, et al., 2015). We updated this information withnew records shared by local partner organizations since2013 (Hedges, 2017). We considered the green iguana thatexists on Montserrat as a genetically distinct conservationmanagement unit, because it is genetically closely relatedto the iguana on Saint Lucia, which is treated as a nativespecies of conservation concern (Powell, 2004; Stephen,et al., 2013; Vuillaume, et al., 2015). Due to the lack ofsufficient distribution data and limited existing knowledgeof interactions, native and invasive plant or invertebratespecies were not considered in this prioritisation.Calculating the conservation threat score of islandsWe followed the approach of Dawson, et al. (2015) tocalculate the conservation threat score (termed ‘conservationvalue’ in Dawson, et al., 2015) of each island based on thesum of each native species’ vulnerability. The vulnerabilitywas calculated as the product of the global threat status,the irreplaceability, which indicates the global significanceof an island’s population, and the severity of impact of themost harmful invasive vertebrate species already presenton an island (i.e. the species with the greatest severity ofimpact score; Dawson, et al., 2015; Stanbury, et al., 2017).We scored threat and impact categories on both a linear andlogarithmic scale to address the arbitrariness of assigningquantitative values to normative categories (Game, et al.,2013; Helmstedt, et al., 2016). The severity of impactwas classified in three categories, depending on whetheran invasive species had no impact on a native species (0),small to moderate impact that would reduce populationsize but allow the native species to persist (1), or a severeimpact that would eventually lead to the local extinction ofthe native species (2). We classified unassessed reptiles as‘At Risk’, which received a numerical value equivalent to‘Vulnerable’ (Dawson, et al., 2015).Simulating the invasion of islands to calculate increasein conservation threat scoreTo quantify the magnitude of biodiversity loss thatcould result from invasion, we first assessed which of thefive selected invasive species were already present on anisland in 2016, and then simulated the arrival and invasionof those species that were not yet present in 2016. We thenre-calculated the conservation threat score of each islandas described above, where the vulnerability of each nativespecies was adjusted to reflect the most harmful invasivespecies on the island, which may be one of the simulatedinvaders. We assumed that all invasive species not yetpresent on an island would invade, because biosecuritymeasurements should, in our opinion, not be tailored for asingle species but guard against the arrival of a broad suiteof species. However, we emphasise that our prioritisationcould also be performed for single species invasions, butassessing the merits of guarding against one or anotherinvasive species would require information about therelative invasion risk of various species.The calculation of the conservation threat scoredepends on a classification of the threat posed by eachinvasive species to each native species, but these threatscan be hypothetical for interactions between certain islandendemic species and invasive species that have so far notinvaded the respective island. Consequently, we drew ontaxonomically related or otherwise very similar speciesto specify the potential threat that would result frominvasion. For example, if black rats adversely affect a smallSphaerodactylus gecko on one island, we assumed that asimilarly sized Sphaerodactylus species that is endemic toan island without any rats would suffer similar effects if theisland were invaded by rats (Case & Bolger, 1991).Prioritising islands for biosecurityFig. 1 Location of 318 islands (black dots) in six UnitedKingdom Overseas Territories where the priority forbiosecurity was assessed. Circles around islandsindicate the location of the highest priority islands listedin this paper.664Islands that should receive the most immediateinvestment into biosecurity are those where the nativefauna would face the greatest increase in conservationthreat score if the five selected vertebrate species invaded.We therefore calculated the difference in conservation

Oppel, et al.: Prioritising islands for biosecurityvalue at present and after the simulated invasion of thefive vertebrate species, and ranked islands based on themagnitude of this difference. We present the results as aranking table and include information on island size andhuman population size for each island. These aspects willaffect the complexity and cost of biosecurity measures,as well as the probability of invasive species arrival andestablishment, but they did not factor into our prioritisationof islands for biosecurity, which was entirely based on thepotential threat to native biodiversity. All calculations wereperformed in R 3.2.5 (R Development Core Team 2015)based on the code provided by Dawson, et al. (2015).RESULTSOf the 318 islands in our assessment, 125 did not haveany invasive species on them, and 150 (47%) did not haveany of the five focal invasive species. Of the islands withany of the five focal invasive species, 31 (10%) had oneinvasive, 117 (37%) had two, 12 (4%) had three, 6 (2%)had four, and only two islands (Tortola and Virgin Gorda,British Virgin Islands) had all five of the focal invasivespecies. On 183 islands (57.5%) the invasion of any of thefive focal invasive species would not lead to an increase inthe conservation threat score, because the native vertebrateson these islands were not at greater risk of predation fromthose invasive species that have not yet invaded. Thus,biosecurity measures to prevent the invasion of at least oneof the five focal species would be useful on 133 islands inour assessment.We identified several important islands across theCaribbean UKOTs and Bermuda where biosecurity couldhelp prevent the loss of globally important biodiversity(Table 1). Two islands emerged where an invasion ofnon-native vertebrates could lead to an increase in theconservation threat score more than five times greaterthan on any other island included in our study, mostly dueto the potential loss of Critically Endangered endemicreptiles (Table 1): Sombrero (Anguilla), and Cayman Brac(Cayman Islands).Among the most important islands we identified forbiosecurity, three were inhabited by 1000 people andhave existing populations of rats, feral cats, and greeniguanas (Cayman Brac, Grand Cayman, and Montserrat,Table 1). However, the small Indian mongoose is so farabsent from those islands and reducing the risk of invasionof this efficient predator on islands that already haveother harmful invasive species could help secure globallyimportant biodiversity. Together with Montserrat, Anegadain the British Virgin Islands was among the top prioritiesfor biosecurity to reduce the risk of invasion of black ratsand small Indian mongoose, despite both islands alsobeing a high priority for the eradication of already existinginvasive species (Dawson, et al., 2015).DISCUSSIONWe show that effective biosecurity on islands in theCaribbean UK Overseas Territories could reduce the riskof further spread of harmful invasive vertebrates to islandswhere globally threatened reptiles and birds would be atrisk. Investing in effective biosecurity procedures andeducating the public and policy makers about the risksto their national heritage when no biosecurity is in placeshould be the immediate next steps of UK and localgovernments, private island owners, and internationalfunding bodies. Our approach offers the guidance to focuson a limited number of vulnerable islands, as more thanhalf of the islands we evaluated are not at immediate riskof further biodiversity loss from the invasion of the fiveinvasive vertebrate species that we selected.Similar to other prioritisations identifying islands foreradication of invasive species (e.g., Harris, et al., 2012;Dawson, et al., 2015; Stanbury, et al., 2017), our list issubject to incomplete information about the distribution ofboth native and invasive species. The distribution of severalreptile species is poorly documented across many islandsof the Caribbean, and their threat status is also poorlyassessed on the IUCN Red List, both of which may affectour assessment of their local importance and thereforeintroduce bias to our projections of loss in conservationvalue (Russell, et al., 2017). Further surveys to increasethe knowledge of native and invasive species on islandswould be beneficial but should not be used as an argumentto delay the immediate adoption of effective biosecurityprotocols to safeguard the most important islands that weidentified.Besides thorough knowledge about the native andinvasive species occurring on an island, our approach alsorequires a classification of the interactions between nativeand invasive species. Because these interactions can behypothetical for single-island endemic native species thathave not been exposed to invasive species, due caution isnecessary when interpreting the output of our prioritisation.We used the response of taxonomically similar species tothe same invasive species to predict biologically plausibleconsequences of an invasion, but interactions betweennative and invasive species are often complex andunpredictable (Simberloff & Von Holle, 1999; Simberloff,2006). We encourage researchers to provide robust andreliable predictions about the potential consequences ofinvasions to assist with strategic investment decisions forreducing the risk of invasive species becoming establishedon islands harbouring globally important biodiversity(Moore, et al., 2010).In summary, we demonstrated that biosecurity is notonly important on small uninhabited islands or privatelyowned tourist resorts where natural habitats remain andendemic and globally threatened species persist. Evenon large and populated islands such as Grand Cayman,Cayman Brac, and Montserrat, the invasion of smallIndian mongoose could result in a significant deteriorationof the conservation status of several globally threatenedvertebrates (Hays & Conant, 2007). We therefore urgelocal governments, private island owners (e.g. MosquitoIsland) and communities to carefully inspect all incomingcargo and people and establish ongoing measures to detectand remove any new invasive species. Training of borderofficials and conservation staff, public education andawareness campaigns targeting the accidental introductionof invasive species onto uninhabited islands by visitingpeople (e.g. fishermen, tourists) should also be implemented,because international and domestic biosecurity measuresare currently weak across all Caribbean UK OverseasTerritories (Key, 2017; RSPB, 2017). Laws governingbiosecurity measures in the Caribbean UK OverseasTerritories and Bermuda are disjointed, not comprehensiveand scattered through various environmental, agriculturaland customs regulations. Collaboration under existingnational legislative mechanisms may improve the situationquickly prior to enacting any new legislation (RSPB,2017). We would also encourage regional collaboration indeveloping biosecurity measures, information sharing andlearning from any existing biosecurity initiatives.ACKNOWLEDGEMENTSWe greatly appreciate the support of all the UK OverseasTerritory governments and BirdLife International partnerorganisations for providing data on the distribution ofnative and invasive species. We are grateful to MontserratNational Trust, Anguilla National Trust, National ParksTrust of the British Virgin Islands, Turks and Caicos665

66600200CaymanIslandsBritish VirginIslandsBritish VirginIslandsBritish VirginIslandsBermudaGrandCaymanSalt IslandCarval RockAnegadaNonsuchIsland053,1600AnguillaLittle ScrubIsland0Cayman Brac CaymanIslandsBritish 149.23,889.429.2Islandarea 44.9737.0737.6927.55,514.55,729.2Potentialincrease inconservationthreat scoreChelonia mydas, Eretmochelys imbricata,Dermochelys coriacea, Cyclura pinguis,Spondylurus anegadaePlestiodon longirostris, Pterodroma cahowBrown rat, black rat, feralcat, green iguana, smallIndian mongooseAmphisbaena fenestrata, Sphaerodactylus sp.Eretmochelys imbricata, Chelonia mydas,Caretta caretta, Dendrocygna arborea,Crocodylus acutus, Cyclura lewisi, Anolisconspersus, Crocodylus rhombifer, Typhlopscaymanensis, Tropidophis caymanensisEretmochelys imbricata, Spondylurussemitaeniatus, Spondylurus sloanii,Amphisbaena fenestrataAmeiva coraxBlack rat, small IndianmongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseSmall Indian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseSmall Indian mongooseSphaerodactylus sp., Ameiva corvinaBrown rat, black rat, feralcat, green iguana, smallIndian mongooseDendrocygna arborea, Crocodylus acutus,Anolis luteosignifer, Typhlops epactius,Tropidophis schwartzi, Anolis maynardii,Celestus maculatus, Cyclura nubilacaymanensisSpondylurus semitaeniatus, Sphaerodactylusparthenopion, Cyclura pinguis, AmphisbaenafenestrataGlobally threatened species at risk frominvasionPotential invadersTable 1 The top 25 islands in Caribbean UK Overseas Territories and Bermuda where the invasion of five common vertebrate species could potentially cause the greatest increasein threats to native biodiversity. Note that some invasive species are already present on some islands (i.e. those with current threat score 0), and only the potential new invadersare listed; islands with a current threat score 0 would also benefit from the removal of already existing invasive species. Human population size and island area are provided forinformation, as they will affect invasion risk and effort required for biosecurity. The current and post-invasion threat scores are calculated as the sum of all impact scores of invasivespecies on all threatened native species present on an island before and after potential invasion.Island invasives: scaling up to meet the challenge. Ch 3D Strategy: Scaling up

6.4British Virgin 0IslandsMontserratRound RockMontserratandBig Sand Cay TurksCaicos7.80Turks andCaicosBush Cay054.510157.48.50Turks andCaicosFish Cay492219.50.9British Virgin 0Islands00.3FallenJerusalemBermudaInner PearRock0102.2BermudaHorn RockIslandarea (ha)Virgin 0Ginger Island alincrease inconservationthreat scoreChelonia mydas, Cyclura carinata,Leiocephalus psammodromusGreen iguana, small IndianmongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseSpondylurus semitaeniatus, AmphisbaenafenestrataEretmochelys imbricata, Cyclura carinataEretmochelys imbricata, Cyclura carinataEretmochelys imbricata, Spondylurussemitaeniatus, Amphisbaena fenestrataEretmochelys imbricata, Spondylurussemitaeniatus, Amphisbaena fenestrataPlestiodon longirostris, Pterodroma cahowPlestiodon longirostris, Pterodroma cahowGlobally threatened species at risk frominvasionDermochelys coriacea, Eretmochelysimbricata, Chelonia mydas, Caretta caretta,Turdus lherminieri, Icterus oberi, Diploglossusmontisserrati, Leptodactylus fallax, Anolislividus, Mabuya montserratae, Iguana iguana(Montserrat)Brown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongoosePotential invadersTable 1 (continued) The top 25 islands in Caribbean UK Overseas Territories and Bermuda where the invasion of five common vertebrate species could potentially cause the greatestincrease in threats to native biodiversity. Note that some invasive species are already present on some islands (i.e. those with current threat score 0), and only the potential newinvaders are listed; islands with a current threat score 0 would also benefit from the removal of already existing invasive species. Human population size and island area are providedfor information, as they will affect invasion risk and effort required for biosecurity. The current and post-invasion threat scores are calculated as the sum of all impact scores of invasivespecies on all threatened native species present on an island before and after potential invasion.Oppel, et al.: Prioritising islands for biosecurity667

668000000Turks andCaicosTurks andCaicosTurks andCaicosTurks andCaicosTurks andCaicosTurks andCaicosSix HillsWest CayMiddletonCayLong Cay(Turks)White CayIndian CayPlandon CayMiddle Creek Turks andCayCaicos00UKOTSix Hills East Turks 6.23.1Islandarea crease inconservationthreat scoreBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongooseBrown rat, black rat, feralcat, green iguana, smallIndian mongoosePotential invadersCyclura carinataCyclura carinataCyclura carinataCyclura carinataCyclura carinata, LeiocephaluspsammodromusCyclura carinata, Tropidophis greenwayiCyclura carinata, Aristelliger hechtiCyclura carinata, Aristelliger hechtiGlobally threatened species at risk frominvasionTable 1 (continued) The top 25 islands in Caribbean UK Overseas Territories and Bermuda where the invasion of five common vertebrate species could potentially cause the greatestincrease in threats to native biodiversity. Note that some invasive species are already present on some islands (i.e. those with current threat score 0), and only the potential newinvaders are listed; islands with a current threat score 0 would also benefit from the removal of already existing invasive species. Human population size and island area are providedfor information, as they will affect invasion risk and effort required for biosecurity. The current and post-invasion threat scores are calculated as the sum of all impact scores of invasivespecies on all threatened native species present on an island before and after potential invasion.Island invasives: scaling up to meet the challenge. Ch 3D Strategy: Scaling up

Oppel, et al.: Prioritising islands for biosecurityNational Trust, Bermuda Audubon Society, UKOTAmembers, C. Stringer, J. Hall, J. Millett, G. Gerber, and K.Newton for providing valuable assistance. Two anonymousreviewers and Dick Veitch provided useful comments toimprove this contribution.Keitt, B., Holmes, N., Hagen, E., Howald, G. and Poiani, K. (2019).‘Going to scale: Reviewing where we’ve been and where we need to goin invasive vertebrate eradications’. In: C.R. Veitch, M.N. Clout, A.R.Martin, J.C. Russell and C.J. West (eds.) Island invasives: scaling up tomeet the challenge, pp. 633–636. Occasional Paper SSC no. 62. Gland,Switzerland: IUCN.REFERENCESKey, J. (2017). Tackling Invasive Non-native Species in the UK OverseasTerritories – Gap Analysis of Biosecurity Capacity. Sand Hutton, UK:Animal and Plant Health Agency.Atkinson, I.A.E. (1985). ‘The spread of commensal species of Rattus tooceanic islands and their effects on island avifaunas’. In: P.J. Moors(ed.) Conservation of Island Birds. pp. 35–81. Cambridge, UnitedKingdom: International Council for Bird Preservation.Barun, A., Budinski, I. and Simberloff, D. (2008). ‘A ticking time-bomb:The small Indian mongoose in Europe’. Aliens 26: 14–16.Brooke, M.d.L., Hilton, G.M. and Martins, T.L.F. (2007). ‘Prioritizingthe world’s islands for vertebrate-eradication programmes’. AnimalConservation 10: 380–390.Broome, K. (2007). ‘Island biosecurity as a pest management tactic inNew Zealand’. In: G. Wi

We used all 318 islands in the fi ve Caribbean UKOTs (Anguilla, British Virgin Islands, Cayman Islands, Montserrat, and Turks and Caicos Islands) and in Bermuda, which is situated 1,500 km north of the Caribbean but Maximising conservation impact by prioritising islands for biosecurity