The world's islands are of singular importance from a conservation perspective because they hold a disproportionate share of global biodiversity. However, island ecosystems are fragile, and the introduction of invasive species poses an especially grave threat to their species. My research took place on the island of Guam, which has been particularly hard hit by invasive species in recent history. The introduction of the invasive Brown Treesnake (BTS; Boiga irregularis) to Guam in the mid-20th century caused the extinction of the vast majority of its avifauna. Of the remaining bird species, the native Såli (Aplonis opaca) is a forest starling now limited to breeding and roosting in certain urbanized areas on the island. Previous research has shown that although Såli adults appear to have high survival, their fledglings are highly vulnerable to predators. My research focused on refining our understanding of how BTS are impacting Såli ecology and survival, and how BTS management can benefit Såli conservation. In my first chapter, I conducted the first field study to investigate the influence of both prey and predator sizes on the frequency of ingestion success by snakes. I found that nearly half of ingestion attempts by BTS on fledgling Såli were unsuccessful, and that BTS routinely pushed the upper limits of their capacity to swallow prey. These results suggest that relatively small BTS pose a greater predation risk to birds than previously recognized, and broaden the size range of BTS that should be targeted in management efforts for avian conservation. In my second chapter, I investigated a potential life-history trade-off for Såli between the proximity of nest sites to foraging areas and the risk of fledgling predation by BTS. I found that Såli nesting closer to the forest make more yearly nesting attempts, but that fledgling survival was higher the farther their nest site was from the forest. Såli therefore balance potential fecundity with predation risk to fledglings in their selection of nest sites. In my third chapter, I examined the survival response of fledgling Såli to the suppression of BTS. I found that snake control did not lead to a change in fledgling survival from the pre-treatment baseline or in comparison to areas where snakes were not controlled, and that there was no shift from snake predation to other causes of mortality. Changes to the strategies or tools employed will therefore be required in order for future BTS control efforts within similar habitats to effectively improve avian survival. This body of work demonstrates that BTS are affecting Såli in unexpected ways, and offers new guidance on how to manage damaging invasive species for native species conservation. / Doctor of Philosophy / Islands are of high value important for conservation because they hold a more than their share of the world's animal and plant species. However, island ecosystems are fragile, and the introduction of damaging exotic species is especially threatening to species that live there. My research took place on the island of Guam, which has been particularly hard hit by damaging non-native species in recent history. The introduction of the Brown Treesnake (BTS; Boiga irregularis) to Guam in the mid-20th century caused the extinction of the vast majority of its bird species. Of the remaining bird species, the native Såli (Aplonis opaca) is now limited to breeding and roosting in certain developed areas on the island. Previous research has shown that although Såli adults appear to have high survival, their fledglings (young birds that recently left the nest) are highly vulnerable to predators. My research focused on refining our understanding of how BTS are impacting Såli life cycles, and how BTS management can benefit Såli conservation. In my first chapter, I conducted the first field study to investigate the influence of both prey and predator sizes on the capacity for snakes to swallow their prey. I found that nearly half of attempts by BTS to swallow fledgling Såli were unsuccessful, and that BTS routinely pushed the upper limits of their capacity to swallow prey. These results suggest that relatively small BTS pose a greater predation risk to birds than previously recognized, and broaden the size range of BTS that should be targeted in management efforts for avian conservation. In my second chapter, I investigated a potential life-history trade-off for Såli between how close their nest sites are to foraging areas and the risk of fledglings being killed by BTS. I found that Såli nesting closer to the forest make more nesting attempts per year, but that fledgling survival was higher the farther their nest site was from the forest. Såli therefore balance the potential number of young they can raise with the predation risk to those same young in their choice of nest sites. In my third chapter, I examined whether the suppression of BTS led to a change in the survival of fledgling Såli. I found that snake control did not lead to a change in fledgling survival from the period before the treatment or in comparison to areas where snakes were not controlled, and that there was no shift from snake predation to other causes of death. Changes to the strategies or tools employed will therefore be required in order for future BTS control efforts within similar habitats to effectively improve avian survival. This body of work demonstrates that BTS are affecting Såli in unexpected ways, and offers new guidance on how to manage damaging invasive species for native species conservation.

Livestock guardian dogs – or “LGDs” – are commonly used by domestic sheep ranchers and reduce the need for killing wild carnivores to protect livestock. LGDs are mostly used in the United States to reduce the number of livestock killed by coyotes, but whether they can prevent killing by larger carnivores like wolves and grizzly bears is unclear. It is important to identify which behavioral traits and LGD breeds work best for guarding livestock so that ranchers can protect their stock and environmentalists can enjoy a greater number of wild animals on the landscape. This study investigated the effectiveness of different LGD breeds in the Western U.S. to help determine how best to use LGDs. I investigated (1) which LGD breed works best for each predator, (2) if LGD breeds behave differently, (3) how carnivores respond when LGDs and sheep move through their home ranges, and (4) whether having good LGDs makes ranchers more accepting of predators. I compared common U.S. breeds of LGD with three exotic breeds used primarily in other countries with wolves and grizzly bears. From 2013 – 2016 data was collected on sheep that were killed and what killed them, how different LGD breeds behaved, what carnivore species were present near sheep grazing with LGDs, and ranchers’ attitudes towards LGDs and large carnivores throughout Idaho, Montana, Oregon, Washington, and Wyoming. Results of the study show that all three of the exotic breeds of LGD are better at protecting sheep from certain predators than LGD breeds commonly used in the U.S. There are also some breed differences in LGD behavior that may help ranchers make better decisions about which LGD breed is best for them. Sheep grazing with LGDs seemed to drive-off wolves, but they also attracted smaller carnivores. Also, ranchers’ attitudes about LGDs are generally very positive, but they don’t affect attitudes about wolves and grizzly bears. Below, I discuss these and other findings in terms of both ecology and wildlife management.

Predators play a critical role in regulating the structure and function of ecosystems by exhibiting top-down forces on lower trophic levels. Despite their important contributions in maintaining ecosystem health, lethal predator control remains a global wildlife management strategy to reduce predation on livestock, culturally and/or economically important species, and species at risk, as well as to reduce conflict with humans. Predator control has received criticism due in part to a paucity of rigorous research on the community-level impacts of this practice, beyond the target prey species. Specifically, there is a lack of understanding of the behavioural consequences of predator control on the wider ecological community. In this thesis, I used a multi-year camera trap dataset to evaluate how government-mandated grey wolf (Canis lupus) population reduction to conserve boreal woodland caribou (Rangifer tarandus caribou) could impact the resource selection of the carnivore community in northeast Alberta, Canada. In my second chapter, I investigated whether perceived persecution risk due to predator control may alter wolf habitat selection. I found that wolves switched from positively associating with roads before predator control to avoiding anthropogenic linear features and selecting for block features after predator control. These results suggest that lethal control may prompt wolves to prioritize local prey acquisition near block features over movement on linear features. In my third chapter, I examined whether coyote, lynx, and black bear exhibited shifts in co-occurrence with habitat features, competitors, and prey consistent with a release from top-down suppression in response to predator control. I found that predator control triggered unexpected behavioural changes among coyote and lynx consistent with a release from top-down suppression, but not among black bears. Non-apex predator response to predator control may depend on the strength of competition between the apex and non-apex predator, emphasizing the need to consider bottom-up processes when trying to understand the indirect effects of predator control. This research demonstrates that predator control can have trickle-down effects within the larger ecological community, specifically affecting how species utilize resources. As predator control continues to be a recommended wildlife management strategy, it is imperative to continue investigating its unintended consequences throughout the ecological community. / Graduate

New Zealand provides an excellent example of the effect of exotic predators on native reptile populations. Prior to human arrival, reptiles evolved in the absence of mammalian predators but are now sympatric with 11 introduced mammalian predators. New Zealand’s reptile populations have declined over the past millennium because they have few defensive behaviours against this new predator guild. However, relatively few studies have investigated the effects of introduced mammalian predators on skinks. In this thesis, I studied the life history of several skink species and the long-term demographic changes in these species to evaluate population trends. I evaluated the effectiveness of mammalian predator control in the Rotoiti Nature Recovery Project (RNRP) for restoring skink populations and I investigated the potential sub-lethal effects of mammalian predators on skinks. Finally, I tested whether two skink species had developed behaviours to avoid the scent of introduced rats or hedgehogs. I estimated size at sexual maturity, birthing season and habitat preferences for speckled skinks (Oligosoma infrapunctatum) in the Nelson Lakes area. My research shows that skink populations are declining both inside and outside of the RNRP. The largest declines are seen in the rarer species and even within the predator-controlled area of the RNRP the speckled skink is nearing extirpation. In addition, the proportion of female northern grass skinks (O. polychroma) and larger individuals of both sexes has decreased since 1970; suggesting females and larger individuals are more vulnerable to predation. An investigation of three fitness surrogates (body condition, parasite load and prevalence of caudal autotomy) showed that for both northern grass and glossy brown skinks (O. zelandicum), body condition was significantly lower in populations with mammalian predators than without. This has serious conservation implications because it shows that lizard populations may not only be in decline from direct predation, but also additional stresses associated with predation that may lead to reduced reproductive output. Neither parasite load, nor the prevalence of caudal autotomy, appears to be good indicators of fitness for northern grass or glossy brown skinks. I found no evidence of substantial avoidance behaviours in glossy brown and northern grass skinks to either rat or hedgehog odour. Two hypotheses are suggested to explain this. Firstly, there may not have been enough time for these species to evolve avoidance behaviours, and secondly, there may be insufficient selection pressure due to the high efficiency of the alien predators, or because rats and hedgehogs are active foragers and thus scent gives the prey limited information on immediate predation pressure. This lack of evidence for the evolution of antipredator behaviour, in addition to reduced body condition and population decline in areas with mammalian predators present, highlights the importance of intensive mammalian predator control for the continued survival of skink species on mainland New Zealand.

Mule deer (Odocoileus hemionus) are an iconic species of wildlife, and populations of mule deer across much of the western U.S. have experienced recent fluctuations in size. Factors that affect the survival and subsequent recruitment of juveniles may be the preeminent cause of population fluctuations for mule deer in many areas. Many factors, including habitat loss, extreme weather, intense predation, timing and synchrony of parturition, and competition with other species may be influencing these changes. We studied two potential factors that can influence the survival of neonate mule deer in southern Utah. To better understand how predation affects mule deer, we first implemented a study of the response of mule deer to removal of coyotes in southern Utah. We monitored survival and cause-specific mortality of neonate mule deer in areas where coyotes were removed and where they were not removed. We used multi-model inference within Program MARK and a known-fate model to estimate survival of neonate mule deer in both treatments (removal and non-removal), and to investigate factors potentially influencing survival. Our results indicated that coyote control can decrease mortality and increase survival of neonate mule deer in some situations. Removal of coyotes was most effective when removal efforts occurred for multiple consecutive years, and when control efforts occurred in or near fawning habitat. Second, we examined how synchrony of parturition affects the survival and cause-specific mortality of neonate mule deer. Reproductive synchrony is a strategy that influences the survival of juveniles and the growth of populations. Our objective was to test three possible explanations for the synchrony of parturition in mule deer; 1) pressure of predation on newborns, 2) a hybrid of predation and environmental effects, and 3) weather and food availability. To determine the effects of the timing of parturition on the survival and predator-related mortality of neonate mule deer, we used multi-model inference within Program MARK and a known-fate model. Our results indicated that the timing of parturition influenced survival and predator-related mortality of neonate mule deer. There was a lag between the onset of parturition of mule deer and predation of mule deer by fawns; individuals born close to the onset of parturition had higher survival and lower predator-related mortality than those whose births were delayed relative to the onset of parturition. Since predators selected for neonate mule deer that were born late, predator learning may partially explain reproductive synchrony in mule deer. Environmental factors may have a greater effect than predation on the survival of early-born individuals.

Les rats introduits (Rattus spp.) sont des espèces invasives majeures menaçant la biodiversité sur la plupart des îles de laPlanète. Deux espèces, le rat noir (R. rattus) et le rat du Pacifique (R. exulans) vivent en sympatrie au sein des forêts deNouvelle-Calédonie, où la question de la faisabilité et de l' intérêt de leur contrôle (i.e. limitation locale de leur 'abondance)pour la conservation de la biodiversité native est posée. En raison d'un manque de cadre conceptuel des projets de contrôle,nous avons d'abord réalisé une synthèse et une analyse des opérations de contrôle de rats invasifs dans les milieux naturelsdes îles du monde. Puis, nous avons cherché à caractériser et à comprendre la dynamique des populations de ces deux espècesde rats sympatriques ainsi que leurs interactions avec la biodiversité native en forêt dense humide du massif du Mont Panié.Des opérations de piégeage létal et de capture-marquage-recapture ont montré que les rats noirs étaient plus abondants que lesrats du Pacifique. Les analyses de leur régime alimentaire ont révélé que les deux espèces ont à la fois des proies communeset des proies qui leur sont propres impliquant un renforcement ainsi qu'un élargissement de leurs impacts sur la biodiversiténative. Les rats consomment une grande majorité de fruits et de graines, d'invertébrés et de Squamates mais les oiseaux, quijustifient souvent la mise en place de projets de gestion de rats, ne semblent pas ici être une de leur proie préférentielle. Depotentiels effets positifs des rats sur la dispersion des graines ont également été mis en évidence au travers d'une comparaisondu potentiel germinatif de graines après passage par leur tractus digestif et celui de frugivores natifs. Enfin, nous avons puproposer des stratégies de piégeage létal afin de contrôler efficacement les populations de rats invasifs. Une meilleurecompréhension des impacts des rats en situation de sympatrie ainsi qu' une meilleure connaissance du lien entre densité de ratset intensité des effets sur la biodiversité permettraient d'optimiser les stratégies de contrôle de rats invasifs lorsquel'éradication n'est pas envisageable. / Introduced rats (Rattus spp.) are one of the major invasive species threatening native biodiversity on islands worldwide. Twospecies, the black rat (R. rattus) and the Pacifie rat (R. exulans) are sympatric in New Caledonian rainforests, where questionsasto the feasibility and the utility of their control (i.e. local limitation of their abundance) for the conservation of nativebiodiversity remain unanswered. ln response to the lack of a conceptual framework for control projects, we fi rst conducted areview of invasive rat control operations in island natural areas worldwide. Then we sought to characterize and understand thepopulation dynamics of these two sympa tric rat species and the ir interactions with native biodiversity in the rainforest of MontPanié mountain. Lethaltrapping operations and capture-mark-recapture showed that black rats were more abundant thanPacifie rats. Diet analysis revealed that the two species consume both shared and unshared prey likely resulting in astrengthening and a broadening oftheir impacts on native biodiversity, relative to the impact that each species would haveal one. Rats consume a large quantity of fruits and seeds, invertebrates, and Squamates. However, birds, which often justify theimplementation of rat management projects, do not appear here to be one of the ir preferred prey, either as adults or throughnest predation. A potential positive rat impact on seed dispersal was a Iso highlighted through a comparison of seedgermination after seeds had passed through rat versus native frugivore digestive tracts. Finally, we propose lethal trappingstrategies to efficiently control invasive rat populations. A better understanding ofboth rat impacts in sympatric situations,and the link between rat density and the intensity of their effects on biodiversity would allow optimizing rat control strategieswhen eradication is not feas ible.

Le chat haret est l’un des prédateurs invasifs les plus dommageables pour la biodiversité insulaire. Sa présence est associée à une perte de biodiversité sur l’ensemble des îles sur lesquelles il est établi, et où il constitue une menace pour de nombreuses espèces de vertébrés souvent endémiques et menacés. En Nouvelle-Calédonie des populations de chats harets sont présentes dans tous les milieux et habitats et l’étude de son écologie et de ses impacts sur la faune ont fait l’objet de ce travail de thèse. L’analyse du régime alimentaire sur 14 sites d’études représentatifs des 4 habitats majeurs a révélé un régime très diversifié et une forte prédation sur les vertébrés natifs et notamment sur le groupe des scinques, des roussettes et des pétrels. Parmi les 44 espèces de vertébrés retrouvées dans le régime alimentaire de ce prédateur invasif, la plupart sont endémiques et 20 listées comme menacées sur la liste rouge mondiale de l’UICN. Le suivi des déplacements d’individus équipés de colliers GPS au niveau d’une presqu’île abritant une importante colonie d’oiseaux marins, a permis de mettre en évidence de grands domaines vitaux pour les mâles, des domaines vitaux petits pour les femelles et des patrons de déplacements liés aux différentes étapes du cycle reproducteur des oiseaux marins. Ces éléments, couplés aux analyses de régime alimentaire suggèrent une prédation à une large échelle géographique, à la fois sur les adultes reproducteurs mais également sur les jeunes oiseaux proches de l’envol. Une opération expérimentale de contrôle d’une population de chats harets a été conduite sur cette presqu’île et a montré une faible durabilité des effets de la suppression des individus sur les densités observées et une rapide recolonisation du site. Les résultats de ce travail plaident pour la mise en place de mesures de limitation des impacts occasionnés et permettent en particulier de cibler les habitats de maquis et forêt humide comme prioritaires en matière de limitation des abondances. / Feral cat (Felis catus) is one of the most successful and harmful invasive predator species for island biodiversity. The presence of this alien predator species generally lead to dramatic loss of native island biodiversity and represents a serious threat for numerous endemic and threatened species. Feral cats have invaded the whole New-Caledonian archipelago and all its habitats. In this study, we focused on the ecology and impacts of this invasive predator on the outstanding endemic fauna found in the different habitats of the exceptional New-Caledonia biodiversity hotspot. Feral cat diet analyses on 14 selected sites representing the 4 main natural habitats revealed a high diversified diet and high predation rates on native species particularly on squamates, flying foxes and petrels. Among the 44 vertebrates species found into the feral cat diet, 20 are IUCN red-listed threatened species. Cat movements of eleven feral cats fitted with GPS collars have been studied in a western coast Peninsula hosting an important seabird colony. Male cats showed large home ranges while female showed small home ranges. Feral cats exhibited important movements within the studied peninsula linked with the breeding cycles of seabirds. GPS data coupled with dietary informations suggested a predation that concerned both breeding adults and fledgings bird, and at a large geographic scale as some feral cats have their core home range distant to the colony (>3km). We evaluated the effects of a high level but intense cat control on this site that showed a low sustainability of feral cat culling and a rapid recolonization process. Our results are pleading for the future limitation of feral cat impacts and call to focus first abundance limitation measures on maquis mosaic and humid forest habitats This study also provided information on the spatial extent and intensity of future control measures in the special context of a large and highly invaded island.

The Blue Penguin (Eudyptula minor) is assumed to be declining over much of its range, largely due to introduced predators. Anecdotal evidence suggests that one of the areas of declining population is the West Coast of the South Island. The purpose of this study was to determine the reasons for the assumed decline of blue penguins in South Westland. This was done by studying breeding ecology at several blue penguin colonies to assess the importance of breeding success and adult mortality on the penguin population. Three blue penguin colonies in South Westland, at Five Mile and Three Mile beaches south of Okarito, and at the Wanganui River mouth near Harihari, were monitored throughout the 2008/09 breeding season. During each burrow visit the number of eggs and chicks were recorded as well as the date of laying, hatching or fledging. Five colonies of blue penguin were also monitored in Buller over the same breeding season in a study conducted by the West Coast Blue Penguin Trust, a community trust based on the West Coast. The results of both studies were compared to determine the effect of predator control on breeding parameters, such as breeding success. Of 137 eggs laid in South Westland, 108 chicks survived until fledging, giving an overall breeding success of 78.8%. In Buller, 64 chicks survived to fledging from 101 eggs laid, resulting in an overall breeding success of 63.4%. Breeding success was significantly higher at penguin colonies in South Westland, compared to the Buller colonies. There was no evidence that predator control had an effect on breeding success in South Westland or Buller. The mean number of chicks fledged per pair that produced eggs was 1.55 in South Westland and 1.16 in Buller. The overall proportion of occupied breeding burrows compared to the total number of suitable burrows at the South Westland sites was 73.8% (n = 103). At the Buller sites, only 60.3 % (n = 151) of the total number of burrows was occupied. Road kills are a major threat to blue penguins in Buller due to the proximity of colonies to the state highway. Fortunately, incidences of road death in South Westland are rare and due to the distance from roads, do not pose a significant threat to South Westland blue penguins. Further study of blue penguin colonies in South Westland is needed to learn more about annual variation in breeding productivity and to determine if breeding success is consistently high over an extended time period. If this is the case, then the cause of blue penguin decline on the West Coast is unlikely to be due to problems with breeding as the breeding success during this study is one of the highest recorded for blue penguins. Although there was no apparent effect of predator control on breeding productivity during this study there is evidence from other locations that predators, in particular stoats, have contributed to the decline of blue penguin populations. More research into the impact of predators on penguins over a longer period of time is needed on the West Coast before a change is made to how predators are managed.