Ian Robertson

Conservation of the platypus (Ornithorhynchus anatinus): Development of a framework to assess the health of wild platypus populations. A wide range of factors, including individual animal health, genetic diversity and demographics, are associated with wildlife population declines and investigation of these factors may be more sensitive in detecting early impacts on wildlife populations, than estimates of population size alone. Defining wildlife population health as the ability of a wildlife population to remain viable in the long term, this project developed and implemented a holistic health assessment framework for platypuses to gather baseline data, to investigate environmental, temporal and individual patterns within this data, and to provide insights into potential threatening processes. Platypus distribution and population density in two river catchments in northwest Tasmania were investigated in a live capture/release field study during which 154 individuals were captured. The effect on capture numbers of broad habitat characteristics was investigated. A survey of public sightings provided additional information on platypus distribution and population density. The novel use of in-stream microchip readers to monitor platypus movements/survivorship was developed. Data was collected on the timing and frequency of platypus movements, as well as continued use of monitoring sites by individuals captured in this study and in a study three to six years earlier. The timing of the breeding season in Tasmania was investigated using hormonal, ultrasonographic and remote monitoring observations. Genetic diversity and geographical distribution of alleles at the Major Histocompatibility Complex Class II DZB locus was also investigated. The reliability of existing and novel body condition indices was studied. The prevalence of exposure to a range of parasitic, fungal and bacterial agents was determined. Haematology and biochemistry reference intervals were produced. Little evidence was found that the two study populations were in poor health. Baseline population health data, that for many species has been absent when population declines have occurred, was collected for platypuses; and the project’s general approach will serve as a template for similar research in other species.

Research into the diseases of free–ranging wildlife requires targeted surveillance for pathogen(s) of interest, however also relies on reference data of general health indicators against which findings can be interpreted. Wildlife reintroductions and translocations in New Zealand introduce specific disease risks related to spread of disease agents from the source site, exposure to novel disease at the destination, or changes in transmission factors for existing diseases at the destination. The discovery of Beak and feather disease virus (BFDV) causing clinical disease in wild Red-crowned Parakeets (RCP, Cyanoramphus novaezelandiae) intended for translocation in 2008 led to an increased interest from conservation managers in the pathogens affecting this species.

This study aimed to investigate health and disease in a free-ranging population of RCP experiencing feather loss, with an epidemiological focus to determine temporal trends in disease prevalence, and infer risk factors for disease expression. We captured 229 individuals over 5 sampling sessions between 2011-2013, including 4 surveys on Tiritiri Matangi Island (n=184), and one survey on Hauturu-o-Toi/Little Barrier Island (LBI) (n=45).

Normal haematological and biochemical reference ranges were described. The relatively high creatine kinase results highlighted a potential susceptibility to capture myopathy in this species, and warrant further investigation. Comparison of DNA-based sexing results to beak measurements demonstrated high concordance between the two sexing methods.

We investigated BFDV in the population by PCR of blood and feather, and found a low total prevalence on Tiritiri Matangi Island of 1.09% (0.1-3.9%) for 2011-12, and on Hauturu-o-Toi/LBI of 4.4% (0.5%-15.1%) in 2013. Screening by PCR of opportunistic samples from the Wellington region of the North Island, and phylogenetic analysis of the full viral genome sequences from all positive samples, revealed ongoing evidence of viral flow between RCP and Eastern Rosellas (Platycercus eximius) in the Hauraki Gulf/Auckland region, with separate but closely related strains from the Wellington region. These findings, combined with the first report of seroprevalence data for a New Zealand parrot using the haemagglutination inhibition test, suggest RCP may be a dead-end or spill over host for BFDV, with implications for a downgrading of the conservation threat this pathogen currently poses.

The collection of skin biopsies in all birds during the 2012 and 2013 (n=135) following an outbreak of feather loss led to the discovery of a knemidokoptinid mite associated with mange in this species. The mite, Procnemidocoptes janssensi has only been previously described once from a lovebird (Agapornis nigrigenis) in Zambia in 1967. We present the first report of the mite in New Zealand, including histopathological stages of infestation, and the first epidemiological study on mange in a wild parrot globally. Relative mite abundance (number of mites per skin biopsy) was found to be associated with likelihood of feather loss and clinical signs. Findings were also consistent with the presence of a carrier or endemic state, with emergence of clinical disease and epizootics triggered by as yet unknown host, environment or mite factors.

Finally, we conducted a nesting study using natural and artificial nests on Tiritiri Matangi Island from 2012-2013, to infer disease risks for nestling and fledgling RCP, and describe baseline health parameters for this group. Overall nesting success was low, and likely to be related to the drought experienced that year. Nest mites were detected in 64% (95%CI: 41-83%) of nests, however their presence did not significantly affect fledging success or numbers. Chicks from nest mite positive nests had higher mean absolute heterophil counts (p=0.04), suggesting an inflammatory response to the presence of these mites, and a potential fitness cost that warrants further study. We did not detect BFDV in any nests studied, and therefore cannot infer the impact of this pathogen on this age class, although the study provided further evidence BFDV is present at a low prevalence on Tiritiri Matangi Island.

The highly pathogenic avian influenza virus H5N1 was the cause of a pandemic of avian influenza in poultry throughout many parts of the world. The role of wild birds in the transmission and cycling of this virus has been uncertain and the current study was designed to collect further data on the role of wild birds in the transmission of H5N1 in Thailand. The study site for the current study was located in Nakorn Pathom province, the central part of Thailand, where both backyard poultry and low biosecurity poultry farms are common and co-exist. The analysis of existing extensive data from the national wild bird surveillance program for HPAI H5N1 virus in Thailand, found that since 2004 the prevalence of infection with H5N1 in wild birds was low (1.0% 95%CI (0.7, 1.2). However, the annual prevalence varied considerably over this period with a peak of 2.7% (95%CI 1.4, 4.1) in 2004, which dropped to 0.5% (95%CI 0.3, 0.8) and 0.6% (95%CI 0.3, 1.0) in 2005 and 2006, respectively, and then rose again to 1.8% (95%CI 1.0, 2.6) in 2007. During this period, sixteen species of wild birds tested positive for H5N1 virus infection. All samples from juvenile birds were negative for H5N1 virus, whereas the virus prevalence in pooled samples from adult birds was 0.6% (95%CI 0.4, 0.9). The positive birds belonged to twelve species which were mainly resident species that are commensal with human activities. Infected wild bird samples were only found in provinces where poultry outbreaks had occurred. A risk factor study conducted in this project using a questionnaire for villagers on farm practices and wild birds observed in the area revealed that factors associated with disease included replacing poultry individually into households/farms, buying native chickens and/or fighting cocks from commercial hatcheries and the presence of lesser whistling ducks (Dendrocygna javanica) on farms. Selecting healthy poultry when purchasing replacement birds was identified as a protective factor in this study.

The longitudinal wild bird surveillance programs conducted in this study revealed that the serological and virological prevalence of H5N1 virus were low in the wild bird population. The seroprevalence as tested by the H5N1 serum neutralization test (NT) was 2.1% (95% CI 0.7, 3.5). Species that tested positive to NT were rock pigeon (Columba livia), Asian pied starling (Gracupica contra), spotted dove (Streptopelia chinensis), oriental magpie robin (Copsychus saularis), blue-tailed bee-eater (Merops philippinus), myna (Acridotheres spp.), and pond heron (Ardeola spp.). The prevalence of H5N1 virus detection was 0.5% (95% CI 0.0, 1.1); the two H5N1 virus -positive samples were from Asian pied starling (Gracupica contra) and white vented myna (Acridotheres grandis). Wild birds that tested positive to H5N1 virus were mostly common terrestrial birds, some of which showed no clinical signs of disease. Molecular epidemiology showed that the viruses isolated from the survey were most closely related to poultry viruses isolated in Thailand (A/chicken/Thailand/PC-168/2006, A/chicken/Phichit/NIAH606988 /2006, and A/quail/Thailand /CU-333/06). There was no evidence to support the presence of unique strains in wild birds in Thailand.

A wild bird observational study undertaken demonstrated that habitats which contain the potential for a high risk of interspecies transmission of HPAI H5N1 viruses were open system duck farms and household/backyard areas. In these areas wild birds were commonly observed feeding together and in close contact with domestic poultry and pigs. Common terrestrial birds considered as bridge species (e.g. pigeons, sparrows, mynas, starlings, and doves) were likely to be involved in the disease transmission. Moreover, a qualitative risk assessment conducted in this study showed that the risk of wild birds transmitting the disease to poultry was low with an overall risk ranking of ―Medium severity‖. For quantitative risk assessment conducted, the risk of an infected lesser whistling duck defaecating an infectious dose of HPAI H5N1 virus close to a domestic duck in an open system duck farm was 5.8 x 10-6. This risk increased to 2.5 x 10-1 when all ducks visiting an open system duck farm were considered in a year.

In conclusion, wild birds can help maintain the virus in wild and domestic bird populations through spill back and spill over. However, risk of wild birds transmitting HPAI H5N1 virus to poultry in the current study was considered to be low. Monitoring of the disease in wild birds and poultry should be performed in Thailand, and the biosecurity of small and backyard poultry farms should be improved.

The woylie or brush‐tailed bettong (Bettongia penicillata ogilbyi) has recently undergone a dramatic decline (approximately 80% between 2001 and 2006). The Woylie Conservation and Research Project (WCRP) was established to investigate possible causes of this decline. It was hypothesised that predators and/or a disease may be a concomitant cause if not the primary cause(s) of the decline, based on the peculiar temporal and spatial characteristics of the decline and available associative evidence.

This research project is an integrated and collaborative component of the WCRP and its broad aim was to contribute to the knowledge on the general health and ecological attributes of woylie populations that were considered directly relevant for the conservation and recovery of the species.

Initially, the WCRP in collaboration with several researchers supported the investigation of specific pathogens. These projects were ongoing when the research described in this thesis began, however there had been no disease risk assessment prior to these ongoing pathogen studies. Therefore, a formal qualitative assessment of the disease risks potentially relevant to the woylie declines was undertaken in this study to ensure a systematic evaluation and to prioritise allocation of resources. Several pathogens were identified as a high priority for further investigation including, but not limited to, Macropod Herpesvirus (MaHV), Macropod Orbivirus (Wallal and Warrego serogroups), and Encephalomyocarditis virus (EMCV).

A haematological investigation was carried out and reference ranges were established. An overall increase of the leukocytic response in animals trapped in Upper Warren (8%, n=23) compared to woylies in Karakamia (0%) was demonstrated. Gender differences were also recorded, namely males had higher red blood cell, white blood cell and lymphocyte counts than females.

No clear evidence was found that supported an association between changes in the health status of woylies and the decline. Nevertheless, the increased proportion of lymphocytosis (p<0.0005) in Perup, which includes two forest blocks that underwent a decline during sample collection, and the higher prevalence of health problems identified during the physical examinations of animals trapped in Upper Warren (41.3%, n=557) as compared to those from Karakamia (10%, n=80, p<0.0005. Odds Ratio=6.33, 95% CI 2.99‐13.40), justified further disease investigations.

Based on the results of the disease risk assessment and haematological analysis, the serological response to Macropod Herpesvirus (MaHV 1 and 2), Encephalomyocarditis virus (EMCV) and Orbivirus (Wallal and Warrego serogroups) was investigated. There was no serological evidence of any of these viruses affecting woylie populations. Nevertheless, due to sample size limitations, it was not possible to confirm the absence of these diseases with a high level of confidence (i.e. >90%). Additionally, the absence of detection of seropositive individuals does not necessarily imply absence of the pathogen in the population.

Genetic profiles of indigenous (extant wild populations) and translocated woylie populations were examined in order to assess whether woylie populations were suffering from a reduced genetic “health”, as a consequence of the bottleneck that occurred after European settlement. In order to do this a preliminary investigation of the cross‐species performance of 32 primer pairs was carried out to assess their suitability for the aims of this study. Twelve microsatellite primer sets were identified as polymorphic and reliable for genetic analysis in woylie. Additionally, the cross‐species performance of the 32 primer pairs was analysed within the potoroines species to facilitate future ecological and genetic studies in bettongs and potoroos. A 50% reduction in amplification success of polymorphic loci for every 1 million years of evolutionary distance from taxa was found and a “priority‐list” of markers for use in potoroines was identified.

Genetics does not appear to be a contributing factor to the present woylie decline. Expected heterozygosity (HE) was around 80%, ranging from 42.3% to 83.6% and the allelic richness (NAR) was around 6, ranging from 2.67 to 9.72. Nevertheless, among the indigenous populations particular concern was raised for woylies at Tutanning Nature Reserve, and for the translocated populations on the South Australian islands. These populations have a substantially reduced genetic diversity (Tutanning: HE = 0.64; St Peter island: HE = 0.631; Wedge island: HE = 0.602; Venus Bay island “A”: HE = 0.423).

Important insights were gained into woylie population structure and dynamics through the analysis of molecular data. Four genetically distinct indigenous populations were identified (i.e. Dryandra woodland and Tutanning Nature Reserve in the wheatbelt region and two discrete populations in the Upper Warren in the south‐west forests of Western Australia). The mtDNA analysis showed historical connections between populations in Dryandra and the Upper Warren region (Kingston and Perup). These connections no longer exist as a result of habitat fragmentation caused by agriculture and farming land use. Additionally, substantial gene flow was identified between Kingston and Perup and was supported and quantified by microsatellite analyses in the order of 2‐3% migration rate. The evidence of current gene flow within and between populations (i.e. up to 60 km) signifies that direct transmission of an aetiological agent would be possible throughout the whole Upper Warren region within the time frame experienced in the decline.

Analysis of genetic data indicated also that the woylie population in Kingston had already undergone a decline. As a consequence of this change in population abundance, the spatial genetic structure of this population changed, generating a significant correlogram up to 6 km. In other words, in this population, two woylies trapped within a radius of 6 km are likely to be related as opposed to other populations where the genetic signal drops between 1 and 3 km. Additionally, and consistent with previous ecological studies, female philopatry was confirmed and genetic consequences of this behaviour were identified. Population viability analysis (PVA) demonstrated that the main threatening process for woylie populations is the result of the interaction of various variables (in particular predation and inbreeding) that acquire a considerable strength together, whilst not being greatly significant by themselves. It also quantified the minimum mortality rates necessary for the decline to occur (an average juvenile and subadult mortality rate of 28% and 22% for adults per 91 day time period). The minimum viable population size (MVP) estimated through PVA was consistent with the empirical evaluation based on molecular data (i.e. 1,000‐2,000 individuals). As a consequence of the inherent inability of satisfactorily predicting stochastic events and incomplete knowledge on important factors that may affect population size a conservative approach should be adopted. On this basis, a population size of more than 8,000 individuals should be targeted to maximise the likelihood of positive conservation outcomes.

In light of the results of this research project, disease can not be completely dismissed as a possible cause of decline, in particular in association with predation. Haematological, serological and genetic information generated by this study greatly improved the available knowledge on the health and viability of woylie populations and represent baseline data that will enable monitoring and detection of changes in the health status in these populations, as well as contribute to the refinement of the disease risk assessment and quarantine protocols.

The haematological data will also facilitate and improve the interpretation of disease investigations carried out by the additional collaborative components of the WCRP. Moreover, information obtained on woylie ecology through the analysis of genetic molecular data will assist such interpretations, for example by conveying indications on the frequency and extent of animal movements.

This research also provided suggestions for critical management decisions. For example, the identification of woylie populations at risk of substantial loss of genetic diversity and possibly inbreeding depression calls for appropriate management actions. Where there is no indication of any factor limiting the demographic growth of the populations (i.e. populations on South Australian islands) supplementation was identified as the most suitable management option. Based on the detailed knowledge obtained on the spatial organization of woylie populations, it is now possible to adequately source animals from indigenous populations to augment genetic diversity. Animals should be trapped at a distance of at least 1‐3 km in order to maximise the probability that individuals are unrelated. On the other hand, it is critical to identify the causes currently limiting population growth prior to the implementation of these management actions, especially where limited population size (in respect to the carrying capacity) and consequent genetic drift is the main reason for the poor genetic profile of the population rather than isolation.

The PVA also helped to determine critical requirements for the establishment of new, and maintenance of, populations; more specifically that sites should be able to support a minimum population size of 8,000 individuals and that the average mortality rate should be maintained below 22% for juveniles and 28% for adults per 91 day period.

Finally, this research helped to identify important future areas of investigation. These include longitudinal studies of the health status of individual woylies; epidemiological analysis of the data generated by this study integrated with those generated by other WCRP researchers and a quantification of the influence of ecological factors, such as rainfall and diet, on general health parameters. Additionally, regular genetic monitoring is recommended because the baseline data produced in this study and the associated ecological and demographic data available would provide an optimal opportunity to improve our understanding of genetic consequences of rapid population declines. This monitoring may help to quantify the genetic loss associated with the decline and evaluate the accuracy of PVA predictions. It might be possible to assess the success of management actions (e.g. supplementations) and detect if and when inbreeding depression becomes manifest in populations at lower genetic diversity (e.g. Tutanning).

In addition, the molecular genetic data represents the background work needed to establish the interplay between individual (host) genetic profile and disease susceptibility or fitness (including fecundity and survival). The fact that cross‐species primers were used would make the utility of the knowledge acquired easily and directly applicable to other species of the superfamily Macropodoidae.

The western ringtail possum, Pseudocheirus occidentalis, is classified as threatened, both nationally and internationally. Land clearing for building development threatens the last major coastal population stronghold in and around the town of Busselton in the south-west of Western Australia (WA). Translocation of displaced P. occidentalis from this locality into nearby conservation estates commenced in 1991, in the presence of fox control, with the aim of re-establishing populations of the species within suitable habitat outside its current range. Initial successes (1991-1998) were followed by a major population decline at one site for unclear reasons. The aim of this project was to determine which factors presently limit translocation success for P. occidentalis and thereby provide direction for future management of the species.

Displaced and rehabilitated P. occidentalis were translocated into three sites, two of which were baited for fox control. Survival was monitored weekly, causes of mortality were ascertained and attributes of habitat use were mapped and analysed. Each individual P. occidentalis underwent comprehensive health and disease screening under isoflurane anaesthesia prior to translocation and whenever recaptured for re-collaring. Health, survivorship and habitat use of resident common brushtail possums, Trichosurus vulpecula, were similarly studied at each site. Pilot spotlight surveys using line transect methods were performed at the end of the study to provide provisional data on population densities.

Health screening revealed no evidence that infectious disease currently limits translocation success for P. occidentalis. Possums of both species were negative for toxoplasmosis, leptospirosis, salmonellosis and chlamydiosis. Cryptococcal antigen was detected in one individual T. vulpecula but was not of pathological significance. Endoparasite levels were negatively correlated with body condition. Differences between pre- and post-translocation haematological values were found, suggesting that habitat quality or nutrient intake were lower at the translocation sites than at the sites of origin.

Mortality rates of translocated P. occidentalis were high. The majority of P. occidentalis deaths were attributed to predation, with foxes, cats, pythons and raptors all implicated. Some P. occidentalis died in poor body condition from apparent hypothermia/hypoglycaemia, with moderate to heavy parasite burdens present at necropsy. Most T. vulpecula mortality was attributable to fox predation. Survivorship analyses were carried out using information-theoretic techniques to investigate which, if any, of a suite of hypothesised factors most influenced post-translocation survival of P. occidentalis. The most highly ranked models were those that included pre-release white blood cell counts and/or numbers of T. vulpecula at the release site. Survivorship of P. occidentalis was negatively correlated with each of these factors, and the two together acted in a synergistic fashion. Effects of fox control on P. occidentalis survivorship were equivocal. The average annual survival rate of established P. occidentalis was less than half that of resident T. vulpecula.

Post-translocation dispersal distances varied among individual P. occidentalis. Mean home range sizes of translocated P. occidentalis were larger than those reported for other coastal populations. Individual home ranges overlapped one another, both within and between possum species. Vegetation dominated by peppermint (Agonis flexuosa) was utilised by translocated P. occidentalis where available, and habitat partitioning between the two possum species was observed in some areas. A greater range of diurnal rest site types were utilised by P. occidentalis than T. vulpecula. Spotlight surveys revealed presence of low density P. occidentalis populations, including juveniles, at two sites but numbers remained negligible in the site at which the post-1998 decline had occurred.

Complex interactions involving health, predation, habitat quality and inter-specific competition influence the success or otherwise of wildlife translocation programs. The results of this project suggest that all these factors, particularly predation, affected translocation outcomes for P. occidentalis during the period of study. Complete exclusion of exotic predators (foxes and cats) from the translocation sites may be necessary in future, especially given the numbers of native predators (pythons and raptors) present. In addition to heavy predation pressure, the small size and apparently low carrying capacity of the translocation sites for P. occidentalis, along with high numbers of resident T. vulpecula, currently appears to limit P. occidentalis survival and population growth.

While, in the short term, the most efficient use of funds and the best option for the species in its current coastal strongholds might be to put greater effort into conserving P. occidentalis in its natural environment, there could also be value in carrying out further experiments to determine whether or not translocation success can be improved through use of particular management actions. The principles of adaptive management apply both to management of P. occidentalis in its natural environment and to conduction of translocation programs. Possible experimental approaches are outlined and recommendations for further research proposed.

Most of the existing sea turtle populations worldwide are in decline, and loggerhead turtles (Caretta caretta), in particular, are listed as Endangered. The loggerhead nesting population in Western Australia is the largest nesting population in Australia and one of the largest in the Indian Ocean and the world.

This research project investigated hatching success and health, two critical aspects for loggerhead turtle conservation and management, on two important nesting sites in Western Australia: Turtle Bay on Dirk Hartog Island and Bungelup Beach in Cape Range National Park. This project undertook an inter-disciplinary approach encompassing the disciplines of conservation medicine, ecology and epidemiology to investigate questions about sea turtle conservation that could not be addressed by any of these disciplines alone.

Morphological and reproductive measurements were collected during two nesting seasons, between 2006 and 2008, obtaining important baseline reproductive data about the Western Australian population. It was found that the presence of deformed and yolkless eggs was associated with smaller clutch size. At the same time several biotic and abiotic factors were assessed in relation to embryonic and hatchling mortality. Hatching success was significantly reduced by high temperatures during the pre-emergent period, the presence of roots in the nest and nest location along the beach. Results suggest that bacterial contamination of sand on the high nesting density beach sectors may, in part, be responsible for the differences in hatching success along the beach. High nest temperature during the pre-emergent period also significantly reduced emergence success and influenced emergence patterns and duration. Prolonged emergence duration, associated with increased nest temperatures, may further reduce hatchling survival due to diminished energy reserves and increased risk of predation.

On the mainland nesting site, Bungelup Beach, predation of eggs and hatchlings severely limited reproductive success with over 80% of the monitored nests showing signs of partial or complete predation. In contrast with that reported in the literature, ghost crabs (Ocypode spp) were the main predator at this site and the first among the identified predators to dig into nests. Perentie (Varanus giganteus) and introduced foxes (Vulpes vulpes) also predated on eggs and hatchlings, making the level of predation recorded unlikely to be sustainable in the long term.

Health monitoring of the nesting populations enabled the determination of baseline blood health parameters and toxin levels in blood. Two cases of fibropapillomatosis were confirmed for the first time in Western Australian loggerhead turtles. Changes in the leukogram and some biochemical parameters were detected in association with the presence of barnacles, in particular burrowing barnacles. In order to establish the connection between the nesting turtles’ health and reproductive success, maternal health indices were compared to hatching success and reproductive output. Several blood health parameters, including alpha and gamma proteins, iron, zinc and vitamin E levels, were correlated with hatching success or clutch size, suggesting that these parameters influence reproduction in loggerhead turtles. Additionally, sea turtles with reproductive abnormalities, such as soft-shelled or deformed eggs, had higher blood mercury levels than turtles without any egg or clutch abnormalities. This finding raises important questions about the toxic effect of mercury, at low blood concentrations, on sea turtle reproduction.

During the health assessment, a novel intraerythrocytic protozoal parasite species, similar to a malaria parasite (Haemoproteus and Plasmodium spp), was identified for the first time in sea turtles and was described through the use of light microscopy and diagnostic molecular techniques. The phylogenetic analysis indicated that this new parasite is closely related to other haemosporidia isolated from chelonians, but is well separated from malaria parasites isolated from other hosts (e.g. mammals, birds and other reptiles). This parasite appears to be largely benign. Although parasitaemia was low in all infected individuals, further studies are required to assess the potential impact of this haemoparasite on sea turtle fitness.

In conclusion, this study provided further understanding of factors affecting reproductive success, identified threats to the Western Australian nesting population whilst at the same time enabling assessment of the general health of nesting loggerhead turtles in Western Australia.