Context. Amphibians are particularly susceptible to the effects of habitat loss and fragmentation. The construction and use of roads is among the most common sources of habitat fragmentation and can lead to serious population declines. Unused resource access roads, such as those formerly used for logging, can still negatively impact salamanders and reduce habitat quality through edge effects. Unfortunately, habitat rehabilitation and enhancement is rarely attempted on unused forest roads.

Aims . Our aim was to elaborate on a previous study that tested several types of woody debris to mitigate the negative impacts of forest roads by creating a novel habitat on an unused forest road in Algonquin Provincial Park. Here we focus solely on the use of large, squared timbers and their use by salamanders.

Methods . We tested the application of coarse woody debris (CWD) to the surface of an unused forest road. CWD were sampled for salamanders seven times during the 2011 field season. Local climatic variables were tested against salamander captures, and CWD size preferences and patterns of salamander aggregation under CWD were assessed.

Key results . We observed five salamander species and 415 individuals under timbers in the 2011 field season. Larger timbers (>1 m³) were preferred by all species observed and a significant proportion of animals were found in groups of two or more under larger timbers. High ambient temperature and low relative humidity negatively affected the number and species composition observed under timbers, suggesting that the efficiency of CWD as a survey method and enhanced habitat is season dependent.

Implications . Large timbers placed on unused forest roads may provide suitable refuges for migrating or dispersing forest salamanders while they attempt to cross the road. The tendency of salamanders to aggregate under CWD allows individual red efts to reduce water loss; however, red-backed salamanders are territorial and may drive off conspecifics. The use of large CWD may be an effective and low-cost method to rehabilitate unused forest roads and can be used to promote habitat connectivity for salamanders in targeted habitats, such as near wetlands, or for other species of concern.

Aims . Sea-level rise is one of the most certain consequences of global warming and is predicted to exert significant adverse effects on wildlife in coastal habitats worldwide. Terrestrial fauna inhabiting low-lying islands are likely to suffer the greatest loss to habitat from sea-level rise and other oceanographic impacts stemming from anthropogenic climate change. Bramble Cay (Maizab Kaur), an ∼4 ha, low elevation sand cay located in Torres Strait, Australia, supports the only known population of the endangered Bramble Cay melomys Melomys rubicola Thomas, 1924. As a result of a decline in this population noted during previous monitoring to 2004, habitat loss due to erosion of the cay and direct mortality from storm surges were implicated as major threats to this species. This study aimed to confirm the current conservation status of the species, to seek information about the key factor or factors responsible for the population decline and to recover any remaining individuals for a captive insurance population.

Methods . During three survey periods (December 2011, March 2014 and August–September 2014), a total of 1170 small mammal trap-nights, 60 camera trap-nights, 5 h of nocturnal searches and 5 h of diurnal searches were undertaken on Bramble Cay.

Key results . All three survey periods failed to detect any Bramble Cay melomys. The island had experienced a recent, severe reduction in vegetation, which is the primary food resource for the Bramble Cay melomys. Herbaceous cover on the cay decreased from 2.16 ha in 2004 to 0.065 ha in March 2014 before recovering somewhat to 0.19 ha in August–September 2014.

Conclusions . These results demonstrate that this rodent species has now been extirpated on Bramble Cay. The vegetation decline was probably due to ocean inundation resulting from an increased frequency and intensity of weather events producing extreme high water levels and storm surges, in turn caused by anthropogenic climate change.

Implications . The loss of the Bramble Cay melomys from Bramble Cay probably represents the first documented mammalian extinction due to human-induced climate change. This event highlights the immediate need to mitigate predicted impacts of sea-level rise and ocean inundation on other vulnerable species occurring on low lying islands and in susceptible coastal zones through captive breeding and reintroduction or other targeted measures.

Context . Boom and bust population cycles are characteristic of many arid-zone rodents, but it is unknown to what extent these dynamics might be influenced by the presence of invasive rodents, such as the house mouse (Mus musculus) in Australia.

Aim . To determine whether the presence of M. musculus can have negative consequences on the population abundance and reproduction of two old Australian endemic rodents (the spinifex hopping mouse, Notomys alexis, and sandy inland mouse, Pseudomys hermannsburgensis).

Methods . The study took place on the sand dunes of a cattle station in central Australia. Population abundance was estimated as the number of individuals caught in small mammal traps, and female reproductive condition by external examination and, in a few cases, euthanasia and inspection of the reproductive tract.

Key results . Two synchronous periods of high abundance of N. alexis and M. musculus occurred several months after significant rainfall events, whereas the abundance of P. hermannsburgensis was consistently low. No reproduction took place in N. alexis or M. musculus when populations had reached high abundance. During low-rainfall periods, M. musculus was not detected on the sand dunes, and the two endemic species were sparsely distributed, with reproduction occasionally being evident.

Conclusions . During dry periods, M. musculus contracted back to refuges around the homestead and, after significant rainfall, it expanded onto the sand dunes and became abundant at the same time as did N. alexis. In contrast, and unlike in areas where M. musculus was generally rare, P. hermannsburgensis always remained at a low abundance. These patterns suggest that in areas of the natural environment close to human-modified sites, populations of at least one species of an old endemic rodent are supressed by the presence of M. musculus. Reproduction did not occur in the old endemics at times of high M. musculus abundance, but did take place in spring/early summer, even in some dry years.

Implications . The spread of M. musculus into the Australian arid zone may have had negative impacts on the population dynamics of P. hermannsburgensis. These findings suggest that the presence of human settlements has resulted in refuges for house mice, which periodically spread out into the natural environment during ‘boom’ times and adversely affect the natural population cycle of ecologically similar species such as P. hermannsburgensis.

Aims . Indicators of pending state-shifts carry value for policy makers. Predator–prey relations reflect key ecological processes that shape ecosystems. Variance in predator–prey relations may serve as a key indicator of future state-shifts.

Methods . Lion (Panthera leo) diet in the Kruger National Park was evaluated as such an indicator. Over the three-decade time span reviewed, variance in diet in relation to rainfall, prey abundance, management strategies and disease emergence were reviewed.

Key results . Rainfall patterns, both seasonal and cyclical, were identified as key drivers of predator–prey selection. However, the intensity of management in the form of artificial waterpoints overrode and confounded natural process. The results suggest that savanna systems are stable and punctuated by climatic events in the form of extreme above-average rainfall that temporarily destabilises the system. However, droughts are a cyclical part of the savanna system.

Conclusion . Lion prey selection did fluctuate with changing environmental conditions. Abrupt state shifts did occur; however, the ecosystem returned to a stable state.

Implications . State shifts in ecosystems pose key challenges to conservation managers. State shifts appear to be primarily associated with management interventions and environmental factors.

Context . Long-term viability of wildlife populations may be influenced by the adaptive or maladaptive nature of behavioural shifts. Yet, in the short term, implications of novel behaviour are often uncertain, as they were for a newly formed urban nesting colony of pied imperial-pigeons (PIPs) on the mainland coast of north-eastern Australia. It represented unprecedented behaviour, as most of PIPs, also known as Torresian imperial-pigeons, Ducula bicolor/spilorrhoa, breed colonially on remote small islands.

Aims . The present study would (1) determine whether aggregated mainland nesting continued, (2) evaluate reproductive success, (3) evaluate spatial distribution of nests and (4) explore possible association of reproductive success with predator presence and broad indicators of food availability.

Methods . With assistance from volunteers, we found mainland PIP nests and revisited them intermittently to monitor progress. We calculated quantitative estimates of nest survival by the Mayfield method and evaluated reproductive output in relation to environmental data from independent sources.

Key results . During the 2012, 2013 and 2014 breeding seasons, we recorded 436, 387 and 417 PIP nest events at the new mainland colony. Daily nest-survival rates declined progressively and estimated fledgling output decreased over successive seasons from 0.66 to 0.44 per nest event. The highest mainland output was below that estimated for an island PIP colony (0.78) on the basis of sparse prior data. Across potential foraging grounds, there was no negative change in land use and no widespread adverse weather to account for diminishing success. We identified important causes of nest failure among mainland nesting PIPs to be predation, predominantly by birds, and anthropogenic hazards, including tree pruning and collisions with vehicles and windows of buildings.

Conclusions . With ongoing exposure to these hazards, mainland nesting PIPs cannot be expected to increase productivity; hence, the new colony may be a short-term phenomenon. We infer that the historic success of PIP populations in Queensland stems from their selection of breeding sites on remote islands that are largely free of relevant predators and anthropogenic activity.

Implications . Future conservation of extant PIP abundance will depend crucially on protection of island breeding sites, because multiple hazards of mainland nesting make it an unfavourable alternative strategy.

Context . Reducing road mortality is essential to reptile conservation in regions with dense road networks. The Georgian Bay, Ontario population of the eastern massasauga rattlesnake (Sistrurus catenatus) is designated as Threatened, in part because of high road mortality. In Killbear Provincial Park, four ecopassages and barrier fencing were constructed along three busy park roads to reduce road mortality of massasaugas.

Aim . Although mitigation of road mortality has been widely recommended and in some instances implemented for reptiles, effectiveness of mitigation efforts is often inadequately evaluated. The goals of our study were to use long-term data to quantify the effectiveness of ecopassages and barrier fencing in reducing massasauga fatalities on roads, and to evaluate the potential of these structures to serve as movement corridors for individual snakes.

Methods . We used five approaches to assess the overall efficacy of mitigation efforts: (1) comparison of pre- and post-mitigation road mortality; (2) camera traps in ecopassages to document massasauga and predator presence; (3) automated tag readers in ecopassage entrances to detect PIT-tagged individuals; (4) an experiment to assess massasauga willingness to enter and travel through ecopassages; and (5) measurement of temperature fluctuations in ecopassages to assess thermal suitability for massasaugas.

Key results . We found a significant decrease in road mortality of massasaugas on stretches of park roads associated with ecopassages and barrier fencing post construction. Automated tag readers and cameras detected the presence of massasaugas and other animals within the ecopassages, and experimental data showed that massasaugas willingly entered, and in some cases crossed through, ecopassages.

Conclusion . Our evaluation of mitigation structures determined that they successfully reduce road mortality and provide potential movement corridors between bisected habitats, provided that intense maintenance of the fencing is conducted yearly. We also demonstrated the need to utilise a combination of multiple post-monitoring methods to effectively evaluate mitigation structures.

Implications . This study provides a template for construction of similar mitigation in other key locations where reptile road mortality occurs.

Context . Establishing appropriate faunal baselines is critical for understanding and abating biodiversity declines. However, baselines can be highly reliant on historical records that come from already disturbed ecosystems. This is exemplified in the Murray–Darling Depression bioregion of Australia, where European settlement (and accompanying marked land-management changes and the introduction of many species) triggered rapid declines and losses of native species, often before their documentation.

Aims . We aim to establish the mammal fauna present when Europeans settled the Murray Mallee and Murray–Darling Depression bioregion and determine the extent of mammal loss since European settlement.

Methods We describe a dated vertebrate assemblage from Light’s Roost in the lower Murray Mallee region of South Australia. We compare our data with those of modern fauna surveys and historical records to document the extent of change in the mammal fauna since European settlement.

Key results . Radiocarbon ages showed that the assemblage was accumulating, at a minimum, within an interval from 1900 to 1300 years ago. Since this time, the Murray–Darling Depression has lost half of its flightless terrestrial mammals. Species lost include the mulgara (Dasycercus blythi/cristicauda), which places this taxon within only 40 km of Lake Alexandrina, the hitherto-disputed type locality for D. cristicauda. Fossils provided the principal evidence for nearly half of the Murray Mallee fauna and over three-quarters of the fauna are represented in the fossil record.

Conclusions . Late Holocene assemblages provide important archives of species biogeography and diversity. Our revised faunal baseline indicated that the pre-European fauna of the Murray–Darling Depression was more diverse than hitherto understood and its reduction appears largely caused by the impacts of European settlement.

Implications . Baselines for species distributions derived from historical records and modern faunal surveys are likely to be incomplete and warrant revision, particularly for smaller and more cryptic species. Deficiencies in regional records mask the extent of mammal declines caused by European colonisation and associated agricultural practices, and thus vulnerability to anthropogenic disturbance.

Context . Live traps are regularly used in field and enclosure studies with mammals. In some scenarios, such as, for example, when the focus is on temporal patterns or to minimise the time animals are contained inside the trap for animal-ethics reasons, it can be highly useful to be alerted immediately when an individual is trapped.

Aims . In the present study, an automated system was trialed that is designed to automatically send a signal to a receiving device (pager, computer, mobile phone) when the body heat or movement of a trapped small mammal is registered by an infrared sensor (ERMINEA permanent monitoring system for rodent detection).

Methods . Sensors were attached to Ugglan multiple-capture traps and used in laboratory conditions and in semi-natural outdoor enclosures with common voles (Microtus arvalis) and bank voles (Myodes glareolus), as well as in the field with bank voles, Apodemus species and common voles. Sensor readings were compared to visual observation and trapping results.

Key results . In enclosure and field conditions, 100% and 98.7% of traps recorded captured animals correctly. There were no sensor signals when rodents moved along the outside or in the entrance compartment of the traps. Rodents sitting on the trap door triggered the sensor in 50% of cases when there was no bedding in the trap; however, there were no sensor signals if bedding was present. In laboratory trials, 20–70% of traps were falsely triggered by large insects (crickets), depending on ambient temperature and whether bedding was in the trap.

Conclusions . Generally, the system was a reliable, flexible and easy-to-handle tool to monitor live captures. To minimise false negatives (animals trapped without signal), testing sensor function in the pre-baiting phase and software adjustments are recommended.

Implications. The sensors are compatible with various trapping and other monitoring devices, providing the potential to be used in a wide range of applications. Their use is likely to optimise study designs, especially when temporal patterns are recorded or animals or samples need to be obtained soon after capture, and to minimise stress of trapped animals because they can be removed shortly after capture.

Context . Species richness is a relevant diversity component of community ecology and many standardised techniques are available for data estimations. However, each technique is appropriate to a few environment types and has its own sampling biases. Thus, it is necessary to test the effectiveness of traditional and heterodox sampling techniques in different habitat types, especially for highly diverse taxonomic groups, such as anurans.

Aims . We present a comparison based on species richness and detection between the following techniques: acoustic survey with visual encounter of adults (ASVE), automated digital recorders (ADR) and dip net survey of larvae (DSL). We sought to determine: (1) the most efficient sampling technique to survey species richness in ponds of grasslands habitats; and (2) whether efficiency is related to the particular life history traits of species.

Methods . During 2014 and 2015, we sampled 47 ponds distributed in vulnerable Brazilian grassland areas using ASVE, ADR and DSL. Anuran species were surveyed across two seasons that coincide with the peak of anuran breeding activity in the region.

Key results . Species richness recorded by ADR and ASVE was higher when compared with DSL. In terms of combined utilisation, ADR with DSL was as efficient as using all techniques together. However, species detection differed among sampling techniques. ADR had the higher percentage of species that were sampled exclusively in at least one breeding pond and DSL detected two species that were not detected by the other techniques.

Conclusions . Our findings suggest that ADR includes most species whose males call for only a few hours during the night or day, and DSL exclusively detects ‘explosive breeders’, incorporating a wide range of life history traits. ASVE becomes unnecessary since it does not include specific variations of species’ calling behaviour. It is susceptible to a discrepancy of survey data among observers and potentially causes a human disturbance effect in the estimated data.

Implications . We strongly recommended the use of ADR rather than ASVE, and we recommended DSL as a supplementary technique for population monitoring and surveys in grassland areas. We encourage researchers to include ADR and DSL in future works to survey biologic data outside of the short sampling event for which ASVE is commonly utilised, thereby improving the interpretation of biological patterns.