Andrew M. Gormley

Using presence-only and presence-absence data to estimate the current and potential distributions of established invasive species.

1.Predicting the current and potential distributions of established invasive species is critical for evaluating management options, but methods for differentiating these distributions have received little attention. In particular, there is uncertainty among invasive species managers about the value of information from incidental sightings compared to data from designed field surveys. This study compares the two approaches, and develops a unifying framework, using the case of invasive sambar deer Cervus unicolor in Victoria, Australia. 2.We first used 391 incidental sightings of sambar deer and 12 biophysical variables to construct a presence-only habitat suitability model using Maxent. We then used that model to stratify field sampling, with proportionately greater sampling of cells with high predicted habitat suitability. Field sampling, consisting of faecal pellet surveys, sign surveys and camera trapping, was conducted in 80 4-km2 grid cells. A Bayesian state-space occupancy model was used to predict probability of suitable habitat from the field data. 3.The Maxent and occupancy models predicted similar spatial distributions of habitat suitability for sambar deer in Victoria and there was a strong positive correlation between the rankings of cells by the two approaches. The congruence of the two models suggests that any spatial and detection biases in the presence-only data were relatively unimportant in our study. 4.We predicted the extent of suitable habitat from the occupancy model using a threshold that gave a false negative error rate of 0·05. The current distribution was the suitable habitat within a kernel that had a 99·5% chance of including the presence locations pooled from incidental sightings and field surveys: the potential distribution was suitable habitat outside that kernel. Several discrete areas of potential distribution were identified as priorities for surveillance monitoring with the aim of detecting and managing incursions of sambar deer. 5.Synthesis and applications.Our framework enables managers to robustly estimate the current and potential distributions of established invasive species using either presence-only and/or presence–absence data. Managers can then focus control and/or containment actions within the current distribution and establish surveillance monitoring to detect incursions within the potential distribution.

A novel approach to assess the probability of disease eradication from a wild-animal reservoir host.

Surveying and declaring disease freedom in wildlife is difficult because information on population size and spatial distribution is often inadequate. We describe and demonstrate a novel spatial model of wildlife disease-surveillance data for predicting the probability of freedom of bovine tuberculosis (caused by Mycobacterium bovis) in New Zealand, in which the introduced brushtail possum (Trichosurus vulpecula) is the primary wildlife reservoir. Using parameters governing home-range size, probability of capture, probability of infection and spatial relative risks of infection we employed survey data on reservoir hosts and spillover sentinels to make inference on the probability of eradication. Our analysis revealed high sensitivity of model predictions to parameter values, which demonstrated important differences in the information contained in survey data of host-reservoir and spillover-sentinel species. The modelling can increase cost efficiency by reducing the likelihood of prematurely declaring success due to insufficient control, and avoiding unnecessary costs due to excessive control and monitoring.

Movements and habitat selection by wild dogs in eastern Victoria

To investigate movements and habitat selection by wild dogs we attached satellite-linked global positioning system (GPS) units to nine wild dogs (Canis lupus dingo and Canis lupus familiaris) captured in eastern Victoria in summer 2007. Units estimated locations at 30-min intervals for the first six months and then at 480-min intervals for six more months. DNA testing revealed all these wild dogs to be related. Home ranges of males were almost three times larger than those of females (males: 124.3 km2 ± 56.3, n = 4; females: 45.2 km2 ± 17.3, n = 5) and both sexes preferred subalpine grassland, shrub or woodland at the landscape and home-range scales. Wild dogs were recorded more often than expected within 25 m of roads and less often than expected within 25 m of watercourses. Wild dogs displayed higher-velocity movements with shallow turning angles (generally forwards) that connected spatial and temporal clusters comprising slower-velocity, shorter, and sharper turning movements. One wild dog travelled 230 km in 9 days before returning to its home range and another travelled 105 km in 87 days. The home-range sizes reported in this study are much larger than previously reported in south-eastern Australia. This finding, together with previous studies, suggests that the spatial scale at which wild dog management occurs needs to be reconsidered.

Century-scale effects of invasive deer and rodents on the dynamics of forests growing on soils of contrasting fertility

Understanding the long-term impacts of invasive mammalian browsers and granivores in mixed forests is difficult due to the many processes potentially affecting the demography of long-lived trees. We constructed individual-based spatially explicit simulation models of two mixed conifer–angiosperm forests, growing on soils of contrasting phosphorus (P) availability, to investigate how browsing by invasive red deer (Cervus elaphus scoticus) and granivory by invasive rodents (primarily house mouse Mus musculus) might alter forest dynamics. Models were parameterized with field data. Seedling growth and survival rates were estimated inside and outside deer exclosures. Seed predation rates were estimated at high and low rodent densities in mast and non-mast seeding years. For the alluvial terrace forest, which grew on P-rich soil, our model contained 15 tree species dominated by angiosperms; our model of the P-poor marine terrace forest contained seven species dominated by conifers. The two forest models were us...

Why 0.02%? A review of the basis for current practice in aerial 1080 baiting for rabbits in New Zealand

Abstract. Sodium fluoroacetate (1080) has been used as an aerially distributed toxin against mammalian pests in New Zealand since the 1950s. Although its use for rabbit (Oryctolagus cuniculus) control ceased temporarily after the illegal release of rabbit haemorrhagic disease virus (RHDV) in 1997, there has been a recent resurgence in the use of aerial baiting with 1080 to control rabbits as the efficacy of RHDV has fallen. Current practices for rabbit control using 1080 have changed little since the 1980s, with high sowing rates and low toxin loadings commonplace. The lack of ongoing development in baiting practices for rabbit control contrasts sharply with continued improvements in the aerial 1080 baiting practices for brushtail possums (Trichosurus vulpecula) in New Zealand, such as a shift to a comparatively high 1080 loading and using much lower application rates of prefeed and toxic bait. These modifications have resulted in an overall reduction in the amount of toxin used for possum control. The disparity in these two approaches prompted a formal review of the rationale on which the current 1080 baiting practices for rabbits are based. Two issues emerged strongly. First, the current low toxin loading used (0.02–0.04% 1080 in bait) is not based on experimental optimisation in New Zealand but, rather, on research conducted several decades earlier in Australia. Second, despite long-standing concerns about the quality of carrot bait used in New Zealand, current bait manufacturing and distribution practices still produce large numbers of small sublethal fragments. Thus, the current New Zealand practice of multiple prefeeds and very high sowing rates of bait with a low 1080 loading used against rabbits seems to have resulted from the need to compensate for the low toxic loading and poor quality control of the bait (carrots). We, therefore, suggest that there is considerable potential to improve current aerial 1080 baiting practices for controlling rabbits in New Zealand. More generally, these findings also help illustrate that ‘best’ pest-management practice may sometimes be based on pragmatic solutions aimed at overcoming unrecognised internal constraints that are in fact avoidable. Refining and modernising vertebrate pest-control programs, so that they better meet efficacy requirements and contemporary public expectations, therefore requires understanding not just that a solution works, but also how it works.

Effects of large-scale high-severity fire on occupancy and abundances of an invasive large mammal in south-eastern Australia

Abstract Context. Despite large mammals being an important component of many ecosystems, there is little information on the impacts of fire on large mammal populations. Aims. We evaluated the effects of the large-scale high-severity ‘Black Saturday’ fires of 7 February 2009 on occupancy and abundances of an invasive large mammal, the sambar deer (Cervus unicolor), in south-eastern Australia. Methods. The effects of the Black Saturday fires on the abundance of sambar deer were assessed using repeated annual counts of faecal pellets during 2007–11 in Kinglake National Park, which was burnt, and in Mount Buffalo National Park, which was not burnt. Pre-fire occupancy was modelled from data collected at 80 4-km2 cells using three survey methods. The same survey methods were used at 15 burnt (n = 9 sampled pre-fire) and 15 unburnt (n = 5 sampled pre-fire) cells 16–24 months after Black Saturday. Because multiple surveys were performed in each cell, we used a Bayesian state–space site-occupancy model to partition changes in the probability of occupancy from changes in the probability of detection. Key results. Counts of sambar deer pellets increased linearly during 2007–11 in the unburnt Mount Buffalo National Park. Pellet counts also increased linearly in Kinglake National Park from 2007 to 2008, and then decreased (to zero) following Black Saturday; pellet counts increased again in 2010 and 2011. Sambar deer occupancy was weakly reduced (from 0.99 to 0.88) in burnt cells 16–24 months after Black Saturday, but was little changed in unburnt cells (from 0.99 to 0.98). Conclusions. We conclude that the abundance of sambar deer was substantially reduced by the large-scale high-severity Black Saturday fires, but that most burnt habitat was reoccupied 16–24 months later. Implications. There is concern about the negative impacts of invasive sambar deer on native biodiversity, particularly immediately post-fire. Our study suggests that it takes at least 8 months before sambar deer recolonise areas burnt by a large-scale high-severity fire; however, a risk-averse approach would be to act (e.g. by erecting fences or culling) sooner than that.

A modelling framework for predicting the optimal balance between control and surveillance effort in the local eradication of tuberculosis in New Zealand wildlife.

Bovine tuberculosis (TB) impacts livestock farming in New Zealand, where the introduced marsupial brushtail possum (Trichosurus vulpecula) is the wildlife maintenance host for Mycobacterium bovis. New Zealand has implemented a campaign to control TB using a co-ordinated programme of livestock diagnostic testing and large-scale culling of possums, with the long-term aim of TB eradication. For management of the disease in wildlife, methods that can optimise the balance between control and surveillance effort will facilitate the objective of eradication on a fixed or limited budget. We modelled and compared management options to optimise the balance between the two activities necessary to achieve and verify eradication of TB from New Zealand wildlife: the number of lethal population control operations required to halt the M. bovis infection cycle in possums, and the subsequent surveillance effort needed to confidently declare TB freedom post-control. The approach considered the costs of control and surveillance, as well as the potential costs of re-control resulting from false declaration of TB freedom. The required years of surveillance decreased with increasing numbers of possum lethal control operations but the overall time to declare TB freedom depended on additional factors, such as the probability of freedom from disease after control and the probability of success of mop-up control, i.e. retroactive culling following detection of persistent disease in the residual possum population. The total expected cost was also dependent on a number of factors, many of which had wide cost ranges, suggesting that an optimal strategy is unlikely to be singular and fixed, but will likely vary for each different area being considered. Our approach provides a simple framework that considers the known and potential costs of possum control and TB surveillance, enabling managers to optimise the balance between these two activities to achieve and prove eradication of a wildlife disease, or the pest species that transmits it, in the most expedient and economic way. This cost- and risk-evaluation approach may be applicable to other wildlife disease problems where limited management funds exist.

Optimising the Application of Multiple- Capture Traps for Invasive Species Management Using Spatial Simulation

Internationally, invasive vertebrate species pose a significant threat to biodiversity, agricultural production and human health. To manage these species a wide range of tools, including traps, are used. In New Zealand, brushtail possums (Trichosurus vulpecula), stoats (Mustela ermine), and ship rats (Rattus rattus) are invasive and there is an ongoing demand for cost-effective non-toxic methods for controlling these pests. Recently, traps with multiple-capture capability have been developed which, because they do not require regular operator-checking, are purported to be more cost-effective than traditional single-capture traps. However, when pest populations are being maintained at low densities (as is typical of orchestrated pest management programmes) it remains uncertain if it is more cost-effective to use fewer multiple-capture traps or more single-capture traps. To address this uncertainty, we used an individual-based spatially explicit modelling approach to determine the likely maximum animal-captures per trap, given stated pest densities and defined times traps are left between checks. In the simulation, single- or multiple-capture traps were spaced according to best practice pest-control guidelines. For possums with maintenance densities set at the lowest level (i.e. 0.5/ha), 98% of all simulated possums were captured with only a single capacity trap set at each site. When possum density was increased to moderate levels of 3/ha, having a capacity of three captures per trap caught 97% of all simulated possums. Results were similar for stoats, although only two potential captures per site were sufficient to capture 99% of simulated stoats. For rats, which were simulated at their typically higher densities, even a six-capture capacity per trap site only resulted in 80% kill. Depending on target species, prevailing density and extent of immigration, the most cost-effective strategy for pest control in New Zealand might be to deploy several single-capture traps rather than investing in fewer, but more expense, multiple-capture traps.

Cost-Effective Large-Scale Occupancy–Abundance Monitoring of Invasive Brushtail Possums (Trichosurus Vulpecula) on New Zealand’s Public Conservation Land

There is interest in large-scale and unbiased monitoring of biodiversity status and trend, but there are few published examples of such monitoring being implemented. The New Zealand Department of Conservation is implementing a monitoring program that involves sampling selected biota at the vertices of an 8-km grid superimposed over the 8.6 million hectares of public conservation land that it manages. The introduced brushtail possum (Trichosurus Vulpecula) is a major threat to some biota and is one taxon that they wish to monitor and report on. A pilot study revealed that the traditional method of monitoring possums using leg-hold traps set for two nights, termed the Trap Catch Index, was a constraint on the cost and logistical feasibility of the monitoring program. A phased implementation of the monitoring program was therefore conducted to collect data for evaluating the trade-off between possum occupancy–abundance estimates and the costs of sampling for one night rather than two nights. Reducing trapping effort from two nights to one night along four trap-lines reduced the estimated costs of monitoring by 5.8% due to savings in labour, food and allowances; it had a negligible effect on estimated national possum occupancy but resulted in slightly higher and less precise estimates of relative possum abundance. Monitoring possums for one night rather than two nights would provide an annual saving of NZ

Complementarity of indigenous and western scientific approaches for monitoring forest state

72,400, with 271 fewer field days required for sampling. Possums occupied 60% (95% credible interval; 53–68) of sampling locations on New Zealand’s public conservation land, with a mean relative abundance (Trap Catch Index) of 2.7% (2.0–3.5). Possum occupancy and abundance were higher in forest than in non-forest habitats. Our case study illustrates the need to evaluate relationships between sampling design, cost, and occupancy–abundance estimates when designing and implementing large-scale occupancy–abundance monitoring programs.