Modeling the distribution of a wide-ranging invasive species using the sampling efforts of expert and citizen scientists

In its invasive range in Australia, the European rabbit threatens the persistence of native flora and fauna and damages agricultural production. Understanding its distribution and ecological niche is critical for developing management plans to reduce populations and avoid further biodiversity and economic losses.We developed an ensemble of species distribution models (SDMs) to determine the geographic range limits and habitat suitability of the rabbit in Australia. We examined the advantage of incorporating data collected by citizens (separately and jointly with expert data) and explored issues of spatial biases in occurrence data by implementing different approaches to generate pseudo-absences. We evaluated the skill of our model using three approaches: cross-validation, out-of-region validation, and evaluation of the covariate response curves according to expert knowledge of rabbit ecology.Combining citizen and expert occurrence data improved model skill based on cross-validation, spatially reproduced important aspects of rabbit ecology, and reduced the need to extrapolate results beyond the studied areas.Our ensemble model projects that rabbits are distributed across approximately two thirds of Australia. Annual maximum temperatures >25°C and annual minimum temperatures >10°C define, respectively, the southern and northern most range limits of its distribution. In the arid and central regions, close access to permanent water (≤~ 0.4 km) and reduced clay soil composition (~20%-50%) were the major factors influencing the probability of occurrence of rabbits. Synthesis and applications. Our results show that citizen science data can play an important role in managing invasive species by providing missing information on occurrences in regions not surveyed by experts because of logistics or financial constraints. The additional sampling effort provided by citizens can improve the capacity of SDMs to capture important elements of a species ecological niche, improving the capacity of statistical models to accurately predict the geographic range of invasive species.

Changes in small native animal populations following control of European rabbits (Oryctolagus cuniculus) by warren ripping in the Australian arid zone

Context European rabbits have a great impact on native vegetation and small vertebrates in Australia. Rabbits consume vegetation and promote invasive plants and invasive predators, and compete directly and indirectly with native animals suppressing those populations. Aims We explored the changes in small native vertebrates and invertebrates following the removal of rabbits. Methods Warren ripping was undertaken on a property in south-western Queensland at four sites and the results of pitfall trapping were compared with four nearby paired control sites. Invertebrates and small mammals were counted in pitfall traps, and bird surveys were conducted in all treatment and control sites. Key results Following a rabbit-control program, we observed a four-fold increase in the number of dunnarts trapped in treatment plots, whereas no change was observed in control plots. The spring following the rabbit-control program also saw an increase in some lizards in treatment plots. Conclusions The presence of rabbits in arid-zone Australia can suppress native animal populations. Implications Many species of small native mammals and lizards rely on food sources that fluctuate greatly with environmental conditions. The presence of rabbits altering the landscape, supporting introduced predators, reducing vegetation and, therefore, insects, adds increased pressure for insectivorous species. Rabbit control through warren ripping in arid-zone Australia is an effective method to reduce rabbit numbers, and allowed for an increase in small vertebrates in treated areas.

A Landscape Approach to Invasive Species Management

Biological invasions are not only a major threat to biodiversity, they also have major impacts on local economies and agricultural production systems. Once established, the connection of local populations into metapopulation networks facilitates dispersal at landscape scales, generating spatial dynamics that can impact the outcome of pest-management actions. Much planning goes into landscape-scale invasive species management. However, effective management requires knowledge on the interplay between metapopulation network topology and management actions. We address this knowledge gap using simulation models to explore the effectiveness of two common management strategies, applied across different extents and according to different rules for selecting target localities in metapopulations with different network topologies. These management actions are: (i) general population reduction, and (ii) reduction of an obligate resource. The reduction of an obligate resource was generally more efficient than population reduction for depleting populations at landscape scales. However, the way in which local populations are selected for management is important when the topology of the metapopulation is heterogeneous in terms of the distribution of connections among local populations. We tested these broad findings using real-world scenarios of European rabbits (Oryctolagus cuniculus) infesting agricultural landscapes in Western Australia. Although management strategies targeting central populations were more effective in simulated heterogeneous metapopulation structures, no difference was observed in real-world metapopulation structures that are highly homogeneous. In large metapopulations with high proximity and connectivity of neighbouring populations, different spatial management strategies yield similar outcomes. Directly considering spatial attributes in pest-management actions will be most important for metapopulation networks with heterogeneously distributed links. Our modelling framework provides a simple approach for identifying the best possible management strategy for invasive species based on metapopulation structure and control capacity. This information can be used by managers trying to devise efficient landscape-oriented management strategies for invasive species and can also generate insights for conservation purposes.

Recolonisation of rabbit warrens following coordinated ripping programs in Victoria, south-eastern Australia

Context Warren ripping has been demonstrated to be an effective tool for controlling rabbit populations. However, few studies have examined factors influencing the rate at which ripped warrens are likely to be recolonised (i.e. be re-opened). Aims To examine factors influencing the recolonisation of ripped warrens by rabbits by using data collected on 555 warrens for up to 15 years following coordinated ripping programs at 12 sites in Victoria, south-eastern Australia. Methods Warren-monitoring data (number of active and inactive warren entrances) were analysed using discrete-time survival analysis to determine the effects of warren-level and site-level covariates on the recolonisation of ripped warrens. Key results Warren recolonisation was related to the distance between the ripped warren and the nearest active warren, the number of active entrances in the nearest warren, the initial number of active entrances in the ripped warren and the rabbit spotlight abundance index at the site. The probability of warren recolonisation was highest for ripped warrens within 1 km of an active warren and negligible beyond 3 km. The probability of warren recolonisation also increased by 22% for every increase in the rabbit spotlight count at the site by 10 rabbits km–1. Conclusions The recolonisation of ripped warrens was highly influenced by both the distance to, and size of, neighbouring active warrens. Larger warrens also appear to be preferentially recolonised compared with smaller warrens, suggesting that recolonisation of ripped areas may be related to habitat quality. The present results are consistent with ideas from classical metapopulation theory predicting that the rates of colonisation of vacant patches are dependent on both the proximity and size of the source population as well as the quality of habitat patches. Implications Although coordinated warren ripping programs are effective at achieving long-term control of rabbits, their efficiency at maintaining low rabbit populations can be increased by adopting an adaptive monitoring program that incorporates warren size and the spatial relationships among warrens, and using this information to better target maintenance-control activities.

Modelling landscape-level numerical responses of predators to prey: the case of cats and rabbits

Predator-prey systems can extend over large geographical areas but empirical modelling of predator-prey dynamics has been largely limited to localised scales. This is due partly to difficulties in estimating predator and prey abundances over large areas. Collection of data at suitably large scales has been a major problem in previous studies of European rabbits (Oryctolagus cuniculus) and their predators. This applies in Western Europe, where conserving rabbits and predators such as Iberian lynx (Lynx pardinus) is important, and in other parts of the world where rabbits are an invasive species supporting populations of introduced, and sometimes native, predators. In pastoral regions of New Zealand, rabbits are the primary prey of feral cats (Felis catus) that threaten native fauna. We estimate the seasonal numerical response of cats to fluctuations in rabbit numbers in grassland–shrubland habitat across the Otago and Mackenzie regions of the South Island of New Zealand. We use spotlight counts over 1645 km of transects to estimate rabbit and cat abundances with a novel modelling approach that accounts simultaneously for environmental stochasticity, density dependence and varying detection probability. Our model suggests that cat abundance is related consistently to rabbit abundance in spring and summer, possibly through increased rabbit numbers improving the fecundity and juvenile survival of cats. Maintaining rabbits at low abundance should therefore suppress cat numbers, relieving predation pressure on native prey. Our approach provided estimates of the abundance of cats and rabbits over a large geographical area. This was made possible by repeated sampling within each season, which allows estimation of detection probabilities. A similar approach could be applied to predator-prey systems elsewhere, and could be adapted to any method of direct observation in which there is no double-counting of individuals. Reliable estimates of numerical responses are essential for managing both invasive and threatened predators and prey.