This is the first study to investigate and provide definitive evidence that the plant pathogen Phytophthora cinnamomi is a significant threat to the mammal fauna of Western Australia. This study investigated the impact of P. cinnamomi-induced habitat disturbance and degradation on Antechinus flavipes leucogaster (yellow-footed antechinus) or mardo. Phytophthora cinnamomi is an introduced and invasive soil-borne plant pathogen that kills many common and structurally important plant species, which results in significant changes to the structural characteristics of affected areas. An evaluation of P. cinnamomi affected and unaffected areas of the northern jarrah (Eucalyptus marginata) forest revealed significant declines in the structure, composition and complexity of all areas affected by P. cinnamomi. Dieback Expression Score values ranged from a mean value of 1.88 ± 1.01 to 3.8 ± 0.41 at the P. cinnamomi affected sites, indicating a high degree of disturbance. A non-metric multidimensional scaling (MDS) analysis using 16 habitat variables identified significant (ANISOM: R=0.343, P<0.003) separation among affected and unaffected sites. A SIMPER analysis revealed that ground and shrub cover vegetation, small and total log densities, percentage leaf litter cover, and the densities of small, medium, tall single crowned and total Xanthorrhoea preissii were the greatest contributors separating affected and unaffected areas. Presently, our understanding of how P. cinnamomi affects the fauna of Western Australia is limited. This providing a unique opportunity to examine how P. cinnamomi-induced disturbance impacts upon the mardo. The mardo is a small insectivorous marsupial that is regarded as being common and a habitat generalist that occupies a broad range of forest and woodland habitats throughout the south-west of Western Australia. Until the present study, the specific habitat requirements, and therefore the factors limiting the present distribution of the mardo have received little attention. Therefore, in addition to being the first study to evaluate the impact of P. cinnamomi on Western Australian fauna, this study also provides important information about the present distribution of the mardo. Detection-nondetection mark-release surveys conducted in P. cinnamomi affected and unaffected regions of the northern jarrah forest, revealed that although, mardos were recorded at most sites, the number of mardo individuals, captures and detections were considerably lower at P. cinnamomi affected areas. Patch Occupancy analysis, using an information theoretic approach, revealed that the probability of a mardo occupying a region of the northern jarrah forest affected by P. cinnamomi ranged from a likelihood of 0.0 to 25.0%, while in contrast there was a 41.0 to 51.0% likelihood of a mardo occurring among unaffected regions. This discovery supports the hypothesis that P. cinnamomi-induced habitat disturbance impacts upon the distribution of the mardo. An evaluation of the micro-habitat features important to the mardo using Patch Occupancy modelling using an information theoretic approach identified large logs and X. preissii densities as positive contributors to the present distribution of the mardo in the northern jarrah forest. Indeed, the likelihood of a mardo occupying an area with large logs and dense patches of X. preissii ranged from 62.2% to 85.0%. In contrast, in the P. cinnamomi affected sites with lower X. preissii densities the patch occupancy probabilities ranged from 0.0% to 45.7%. Logs and X. preissii strongly contribute to the understorey and may increase nest locations and cover while offering protection from predators. Mardos may avoid P. cinnamomi affected areas because of lower X. preissii densities, which may result in fewer nest locations, reduced cover and an increased likelihood of predation. However, the results of the study must be treated as preliminary findings, therefore there may be additional environmental related or unrelated to P. cinnamomi factors that may also contribute to the occupancy rates of the mardo. Therefore, further studies and research on the ecology and biology of the mardo is strongly encouraged. Until this research is conducted, P. cinnamomi most be considered as significant threat to the conservation of the mardo. Therefore, the conservation of the mardo in the northern jarrah forest depends on limiting the spread and impact of P. cinnamomi, as well as the retention of large logs and tall X. preissii. Given that large logs and tall X. preissii contribute to the distribution of the mardo, strong consideration must be given to using these natural elements to rehabilitate the most severely disturbed areas of the northern jarrah forest. Consideration must be given to the conservation of other small and threatened mammal species that inhabit susceptible plant communities in the south-west of Western Australia. An understanding of how P. cinnamomi impacts on the mardo and other native mammals will contribute to our ability to control, protect and manage vulnerable communities and ecosystems in Western Australia. If the spread and impact of this pathogen is left unchecked, the ultimate consequence to the conservation of many small to medium native mammals that are dependant on structurally complex habitat may be devastating.

<p>In this work suitable habitats for hazel grouse (<i>Bonasa bonasia</i>) were identified using ecological niche factor analysis (ENFA). The results from ENFA reveal that hazel grouse utilize a different and more restricted niche than what is generally available in the study area. When a landscape is fragmented the amount of edge increases, which is negative for many species and thus will affect the amount of available area. The perimeter-area ratio was used to analyze the relative importance of geometric shape. In order to estimate a correlation between incidence of hazel grouse and landscape features census data and land cover maps were analyzed with logistic regression models. It is concluded that hazel grouse is tied to coniferous forest and avoid open areas. However, the result indicates that there is a scale effect that should be considered. The amount of edge in a landscape seems to be important and shape of patches could be a better measure in metapopulation dynamics. In this study the Incidence function model was used to estimate occupancy levels and capacity of a landscape to sustain a metapopulation according to four different area measurement scenarios. Results from the simulations indicate that perimeter-area related measures of patch size combined with capacity could be a more important measure for estimation of population dynamics compared to a basic area measurement. Using a landscape genetic approach, hazel grouse genetic structure, neighbourhood size and dispersal distance were estimated. Genetic estimates of dispersal were in concordance with previous ecological estimates. The results indicate evidence of a population structure reminiscent of what has been found in many other Scandinavian animals with a basic north-south divide. No evidence was found that geographic and environmental structures affected gene flow and dispersal patterns for the hazel grouse.</p>

In this work suitable habitats for hazel grouse (Bonasa bonasia) were identified using ecological niche factor analysis (ENFA). The results from ENFA reveal that hazel grouse utilize a different and more restricted niche than what is generally available in the study area. When a landscape is fragmented the amount of edge increases, which is negative for many species and thus will affect the amount of available area. The perimeter-area ratio was used to analyze the relative importance of geometric shape. In order to estimate a correlation between incidence of hazel grouse and landscape features census data and land cover maps were analyzed with logistic regression models. It is concluded that hazel grouse is tied to coniferous forest and avoid open areas. However, the result indicates that there is a scale effect that should be considered. The amount of edge in a landscape seems to be important and shape of patches could be a better measure in metapopulation dynamics. In this study the Incidence function model was used to estimate occupancy levels and capacity of a landscape to sustain a metapopulation according to four different area measurement scenarios. Results from the simulations indicate that perimeter-area related measures of patch size combined with capacity could be a more important measure for estimation of population dynamics compared to a basic area measurement. Using a landscape genetic approach, hazel grouse genetic structure, neighbourhood size and dispersal distance were estimated. Genetic estimates of dispersal were in concordance with previous ecological estimates. The results indicate evidence of a population structure reminiscent of what has been found in many other Scandinavian animals with a basic north-south divide. No evidence was found that geographic and environmental structures affected gene flow and dispersal patterns for the hazel grouse.