Fauna Spotters logo

Golden-backed Tree Rat (Mesembriomys macrurus)

The Golden-backed Tree-rat is a large, mainly nocturnal rodent of northern Australia. It now occurs mainly in the north-west Kimberley of Western Australia, including rugged mainland areas and offshore islands, and is listed as Critically Endangered in the Northern Territory. On development sites, the main risks are removal of hollow-bearing and fruiting trees, loss of dense cover, intense fire, introduced herbivores and increased exposure to feral cats.

Conservation status

Northern Territory, Territory Parks and Wildlife Conservation Act 1976: Critically endangered.

Breeding season

The species is active throughout the year. The clearest seasonal signal is the presence of pregnant females and juveniles during the mid to late dry season, although breeding may be opportunistic and can occur year round in captivity.

The monthly pattern is indicative rather than a confirmed population-wide calendar. Breeding timing in free-ranging animals is poorly known, and the species may breed opportunistically.

  • Breeding: Pregnant females have been recorded in August, October and November. These records indicate mid to late dry-season breeding, but do not define a strict breeding season.
  • Young: Juveniles have been recorded in August, October and November. Young are weaned at about 6 to 7 weeks and are fully grown at about 4 months.
  • Nightly foraging and local movement: The species is mainly nocturnal throughout the year and may travel more than 1 km from a refuge to forage. It can be active shortly before dusk and soon after sunrise.
  • Fire and cover sensitivity: Late dry-season fires are the main fire risk because intense fires can remove fruiting and hollow-bearing trees and dense cover. The exact timing varies across the tropical north.
  • Highest clearing risk: Clearing is most likely to harm the species from August to November, when pregnant females, juveniles and occupied dens are documented. Disturbance can also remove food and shelter needed before and after this period.

Identification

This is a large semi-arboreal rodent with a distinctive orange-brown stripe along the middle of the back. It can move quickly on the ground as well as through trees and rocky habitat.

The upper body and sides are grey, the underparts are creamy white and the feet are white. The basal third of the long tail is grey, while the remainder is white and ends in a conspicuous white brush.

The orange-brown mid-dorsal stripe extends from the crown along the back to the base of the tail. This stripe is the most reliable field feature because it distinguishes the species from other Australian tree-rats.

Head and body: 188 to 245 mm

Tail: 291 to 360 mm

Weight: 207 to 330 g

Identification: Distinguishing it from similar rodents

The Golden-backed Tree-rat is midway in size between the smaller Brush-tailed Rabbit-rat, Conilurus penicillatus, and the larger Black-footed Tree-rat, Mesembriomys gouldii.

The combination of a grey body, white feet, a grey-based white tail with a terminal brush, and the continuous orange-brown back stripe should be used when separating it from other large northern rodents. Sexes are not reliably separated by external field characters, although tracked males had larger home ranges than females.

Sex-related field information: No reliable external sex difference is provided. In one study, males had larger average home ranges than females.

Identification: Signs and detection

Direct sightings are most likely at night, around dusk or shortly after sunrise. Animals may be seen moving rapidly through dense grass, along rock faces, in trees or across open ground, often with the tail held high.

Useful shelter signs include tree hollows, loose woven nests under pandanus crowns, occupied logs, scree, rock castles and cliff crevices. The species has been recorded denning in cliffs and live tree hollows, so these features should be inspected during daytime habitat searches.

Scats, tracks, calls and eyeshine are not sufficiently described in the supplied sources for reliable species-level identification. Hair sampling can be used as a supporting method by experienced practitioners, but the back stripe and a confirmed photograph or other diagnostic evidence are preferable.

  • Search hollow-bearing trees, pandanus crowns, cliffs, scree and rock crevices during the day.
  • Use spotlighting, cage trapping and camera trapping in suitable habitat, with attention to dense vegetation, rainforest edges and rocky areas.
  • Treat an unconfirmed small mammal sighting as provisional unless the orange-brown dorsal stripe, white feet and distinctive tail pattern are recorded.

Distribution: States and regions

The species formerly occupied a much larger area of northern Australia. Its confirmed current range is concentrated in the north-west Kimberley, while its status in the Northern Territory remains uncertain and the species is treated as Critically Endangered there.

In Western Australia, the current range is in the north-west Kimberley, from near Kalumburu south to the Yampi Peninsula. Records include the Mitchell Plateau, Prince Regent National Park, the Artesian Range, Yampi Peninsula and suitable gorge country in the King Leopold Ranges.

The species has also been recorded on Kimberley islands in the Buccaneer Archipelago and nearby coastal waters, including Augustus, Carlia, Chambers, Conilurus, Hidden, Jar, Lachlan, an unnamed island in Scott Strait, Uwins, Wollaston and Buckle Head.

Historical records extend into the Pilbara and across the Top End of the Northern Territory. In the Northern Territory, records include the upper McArthur area, upper Mary River, South Alligator River and Deaf Adder Gorge. Repeated recent surveys have not confirmed a continuing NT population.

Current Western Australian stronghold: North-west Kimberley, including Mitchell Plateau, Prince Regent, Artesian Range, Yampi Peninsula and King Leopold Ranges

Northern Territory status: Critically Endangered under NT legislation; presence is uncertain and may be restricted to historical records

Historical range outside the Kimberley: Top End of the Northern Territory, northern Pilbara and drier parts of the Kimberley

Distribution: Strongholds and subpopulations

Important mainland areas include the Mitchell Plateau, Prince Regent National Park, the Artesian Range and Yampi Sound Training Area. The species was locally common at some Yampi Sound camera sites, with detections at 34 of 47 sites surveyed over two wet seasons.

On the mainland, it was recorded at 30 of 91 north Kimberley survey sites between 2011 and 2017, and at 15 of 119 sites surveyed between 1994 and 2011. The increase in records at Mitchell Plateau coincided with improved early dry-season fire management and cattle control.

Island populations are distributed across more than 10 Kimberley islands. Rocky habitat appears particularly important on islands, and many islands may provide partial protection from cats and introduced herbivores.

Distribution: Population size and trend

There is no reliable range-wide population estimate. The Australian Government assessment found no evidence of a continuing decline over the previous 10 years, but recognised substantial uncertainty because long-term monitoring has been limited.

Local abundance can be high in suitable Kimberley habitat. A density estimate averaging 6.09 animals per square kilometre, applied to an assumed area of occupancy of at least 2,000 square kilometres, produced an indicative estimate of about 12,000 mature individuals. This is an assumption-based estimate, not a direct population count.

Extent of occurrence in the 2019 assessment: 78,086 km²

Estimated area of occupancy in the 2019 assessment: 244 km² using a 2 by 2 km grid method

Indicative mature population estimate: About 12,000, based on density and area assumptions

Generation length: About 3 years

Habitat: Habitat types

The Golden-backed Tree-rat uses a broad range of tropical habitats, but current records are concentrated in areas with high rainfall, rocky terrain, dense vegetation and reliable shelter and food resources.

Recorded habitats include rainforest and riparian areas on volcanic, lateritic, sandstone and floodplain surfaces; eucalypt-dominated woodland and savanna; Livistona palm woodland; pandanus-associated lowland plains; rugged sandstone plateaux and screes; granite and quartzite hills; and plains beside hills and escarpments.

The species commonly uses habitat edges, including the edge of rainforest patches adjoining eucalypt woodland. In the Artesian Range it selected rainforest over recently burnt savanna, while also using savanna, river edges, cliffs, scree and open woodland.

Habitat: Vegetation and structure

Important structural features include dense understorey, fruiting trees, flowering plants, hollow-bearing trees, pandanus, perennial grasses, rocky cover and fire-sensitive rainforest patches.

In the Kimberley, records include King Leopold and Warton sandstone with eucalypt woodland over hummock grassland, vine thicket, Livistona palms and rainforest or riparian vegetation. The species can use open savanna where food, shelter and cover remain available.

  • Retain connected rainforest patches and their edges where they adjoin woodland or savanna.
  • Treat hollow-bearing trees, fruiting trees, pandanus and dense understorey as potential habitat resources.
  • Inspect rocky plateaux, scree, cliffs, creek lines and gorge country even where vegetation is sparse.

Habitat: Climate and terrain

Most records are from areas with mean annual rainfall above 600 mm and the current range is mainly in higher-rainfall parts of the north-west Kimberley. The type specimen came from the wetter part of the otherwise relatively dry Pilbara coast, where annual rainfall was about 320 mm.

Rugged terrain can break up fire, restrict introduced herbivores and provide cover from predators. The species is not known to be restricted to a particular elevation, and the supplied sources do not provide a reliable elevation range.

Rainfall associated with most records: More than 600 mm mean annual rainfall

Recorded low-rainfall exception: Type specimen locality near Roebourne, about 320 mm mean annual rainfall

Elevation limit: No reliable elevation limit is provided

Foraging habitat: Food

The species feeds both in trees and on the ground. It follows patchy food resources across a relatively large home range and can travel well beyond its refuge at night.

Recorded foods include flowers, fruits and termites, with smaller contributions from grass, leaves, ants and beetles. Other summaries also report seeds and a broad range of invertebrates.

In the Mitchell Plateau, recorded foods included Persoonia falcata flowers, Hypoestes floribunda flowers, Canarium australianum fruit, Eucalyptus tetrodonta leaves and termites.

Main recorded dry-season foods: Flowers, fruits and termites

Additional foods: Grass, leaves, seeds, ants and beetles

Named food plants: Persoonia falcata, Hypoestes floribunda, Canarium australianum and Eucalyptus tetrodonta

Foraging habitat: Foraging behaviour

The species forages mainly at night, in trees, on the ground, along river edges, on cliff faces and among boulder scree. In one study, 66 percent of foraging observations were in trees, 13.9 percent were on the ground, 11 percent were on cliff faces and 8.3 percent were in boulder scree along the river edge.

It uses rainforest and adjacent woodland, and individuals can travel more than 1 km from a refuge into open savanna to forage. It may move rapidly through dense grass and can forage shortly before dusk or soon after sunrise.

Recorded distance from refuge to foraging area: More than 1 km

Average home range: 6.3 ha, with recorded values from 0.24 to 30.44 ha

  • Search fruiting and flowering trees, rainforest edges, creek lines, dense grass, cliffs and scree.
  • Include both trees and ground-level resources in habitat assessments because the species switches between arboreal and terrestrial foraging.
  • Consider food trees and refuges beyond the construction footprint because individuals may travel more than 1 km during nightly foraging.

Breeding habitat: Breeding biology

Breeding information is limited. Available records indicate breeding or the presence of dependent animals during the dry season, with shelter in tree hollows, rock features and occasional pandanus nests.

Pregnant females and juveniles have been recorded from August to November, suggesting mid to late dry-season breeding. Captive animals can breed throughout the year, but the timing of breeding in free-ranging populations is not fully known.

Females have four teats. Litters usually contain two young, with records of one and three young. Young are reported to be weaned at six to seven weeks, fully grown at about four months and sexually mature at around 12 months.

Breeding period indicated by field records: Mid to late dry season, with pregnant females and juveniles recorded in August, October and November

Litter size: Usually 2, with records of 1 and 3

Gestation: Not stated in the supplied documents

Weaning: About 6 to 7 weeks

Fully grown: About 4 months

Age at sexual maturity: About 12 months

Longevity: 7 years in captivity

Breeding habitat: Breeding habitat

Breeding habitat needs secure den sites close to food, including hollow-bearing live trees, dead trees, cliffs, scree, logs, rock castles and dense pandanus foliage. Females and young have been recorded in woven nests in pandanus.

The availability of hollows and fruiting trees is likely to affect breeding and juvenile survival. Large, intense fires can remove both resources, while rocky country and rainforest patches can provide shelter after fire.

  • Identify and retain hollow-bearing trees, pandanus crowns, cliffs, scree and logs within and beside the work area.
  • Avoid disturbing occupied dens during the period when pregnant females and juveniles are most likely to be present, particularly August to November.
  • Maintain nearby food trees and dense vegetation so that breeding animals do not lose both shelter and foraging habitat at the same time.

Sheltering habitat: Known shelter types

The Golden-backed Tree-rat shelters above ground and in rocky features. It does not depend solely on tree hollows, so habitat searches must include cliffs, scree, logs and pandanus as well as trees.

Recorded day shelters include live tree hollows, dead-standing trees, logs, cliffs, scree, rock castles and crevices among rocks. Loose woven nests under the spiky crown of pandanus are used occasionally.

In a study of 363 den sites, 53 percent were in cliffs, 33 percent were in live tree hollows, 5 percent were in logs, 4 percent were in scree, 4 percent were in rock castles and 0.3 percent were in dead-standing trees.

Cliff dens: 53 percent of recorded dens in one study

Live tree hollow dens: 33 percent

Log dens: 5 percent

Scree dens: 4 percent

Rock castle dens: 4 percent

Dead-standing tree dens: 0.3 percent

  • Search cliffs, rock castles, scree slopes, logs, live and dead trees, and pandanus crowns during daytime inspections.
  • Inspect hollow-bearing trees before clearing, including trees that appear partly dead or have hollows that are not visible from ground level.
  • Treat connected rock features and nearby dense vegetation as a shelter system rather than assessing individual rocks or trees in isolation.

Sheltering habitat: Shelter quality

Den trees were not selected simply by tree availability. The study found that tree species and the number of hollows were important, with 64 percent of used trees having more than three hollows. Several preferred den trees were fruiting or riparian species.

Rocky habitat can provide shelter from fire, feral cats and introduced herbivores. Rainforest patches may act as refuges when surrounding savanna has recently burnt.

Used den trees with more than three hollows: 64 percent in the study sample

Known nest material: Pandanus leaves

  • Map hollow-bearing trees and rock shelters before vegetation removal or earthworks.
  • Keep fire, machinery and people away from identified dens where practicable, especially during the likely breeding period.
  • Prevent slash, spoil, fencing or stored materials from blocking rock crevices, scree or movement routes between shelter and feeding areas.

Threats: Fire and vegetation loss

The species has undergone a major historical range contraction. Current threats act together, especially where fire or grazing removes cover and makes animals more exposed to feral cats.

Frequent, extensive and intense late dry-season fires reduce fruiting trees, hollow-bearing trees, dense understorey and fire-sensitive rainforest. The species selects long-unburnt rainforest over recently burnt savanna, although it can use recently burnt savanna when other resources remain.

Fire management is linked to local recovery. At Mitchell Plateau, the species reappeared at survey sites after increased early dry-season burning and cattle control. A return to extensive late-season fires could cause local declines or loss from affected sites.

  • Avoid clearing or burning rainforest patches, fruiting trees, hollow-bearing trees and dense refuges.
  • Use the local fire management plan to reduce the extent and intensity of late dry-season fires near occupied or potential habitat.
  • Keep firebreaks, access tracks and construction areas from creating continuous ignition pathways through rocky woodland, rainforest edges and riparian vegetation.

Threats: Clearing, development and fragmentation

Clearing can remove den trees, fruiting plants, dense cover and connected routes between refuges and feeding areas. Bauxite mining on the Bougainville Peninsula was identified as a potential threat that could severely affect some subpopulations if it proceeds.

The species is not considered severely fragmented across its current range, but project-scale removal of rocky habitat, rainforest patches, creek lines or hollow-bearing trees can reduce the resources available to a local group.

  • Survey rocky hills, gorges, rainforest edges, creek lines and hollow-bearing trees before finalising the clearing footprint.
  • Retain habitat links between denning features and nearby fruiting or flowering trees.
  • Treat offshore islands and isolated mainland refuges as high-consequence areas because replacement habitat and recolonisation may be limited.

Threats: Predators and introduced herbivores

Predation by feral cats is considered highly plausible on the mainland, although direct predation on this species has not been demonstrated. Rocky terrain and partly arboreal behaviour provide some protection, but reduced vegetation cover can increase exposure.

Cattle, pigs and donkeys can reduce shrub cover and alter the availability of tall fruit-bearing understorey. Grazing and fire can act together with cat predation.

  • Limit feral cat access and avoid actions that remove rocky cover or dense vegetation around known habitat.
  • Control cattle and other introduced herbivores where they are damaging shrub cover, riparian vegetation or pandanus.
  • Do not leave food waste, carcasses or other attractants that could increase predator activity around work areas.

Threats: Roads, hydrology and disease

The supplied documents do not identify roads, vehicle strike, disease or hydrology change as established species-specific threats. Even so, new roads and access tracks can remove habitat, increase fire access and increase human activity on islands and in remote refuges.

Creek lines and riparian vegetation are used for shelter and foraging, so drainage changes, sedimentation or loss of river-edge vegetation should be treated as potential local impacts where these habitats occur.

  • Keep new tracks and laydown areas out of rainforest edges, creek lines, riparian vegetation and rocky shelter complexes where practicable.
  • Maintain drainage patterns and avoid concentrating runoff onto scree, creek banks or denning features.
  • Record unusual mortality or signs of illness and seek specialist advice, because disease information for this species is limited.

Survey methodology: 1. Habitat assessment and daytime searches

Survey this species where works may affect north-west Kimberley habitat or any plausible Northern Territory record. It is largely nocturnal, difficult to detect, and uses both trees and rocky features, so surveys should combine targeted habitat searches with more than one detection method.

Map rainforest and riparian edges, eucalypt woodland and savanna with fruiting trees, Livistona or pandanus woodland, rugged sandstone plateaux, scree, cliffs, rock piles and creek lines before selecting survey sites.

Search for tree hollows, hollow logs, pandanus nests, cliffs, scree and rock shelters during the day, then target these features during night surveys.

Treat all occupied sites and nearby suitable habitat as important until survey results demonstrate otherwise, particularly on islands and in rugged country.

Typical body mass: 207 to 330 g

Head and body length: 188 to 245 mm

Tail length: 291 to 360 mm

Generation length: Approximately 3 years

  • Use vegetation maps, aerial imagery, topographic maps and field inspection to stratify survey effort across habitat types and likely refuge features.
  • Record fire history, cattle or other herbivore impacts, feral cat activity, fruiting trees and hollow-bearing trees at each survey site.
  • Do not rely on habitat inspection alone to establish presence or absence.

Survey methodology: 2. Night spotlighting and direct observation

Conduct night searches because the species is mainly nocturnal, while also checking around dusk and soon after sunrise when individuals may forage.

Search slowly along tracks, creek lines, rainforest edges, rocky slopes and beneath likely den trees, checking trees, cliffs, scree and the ground.

Look for the distinctive orange-brown stripe along the back, white feet, grey basal tail and white terminal tail brush.

Activity period: Mainly nocturnal, with some activity shortly before dusk and soon after sunrise

Useful search settings: Rainforest edges, riparian areas, fruiting trees, hollow-bearing trees, cliffs, scree and rugged sandstone

  • Record the date, start and finish time, observers, route, weather, habitat, fire history and search distance or area for every search.
  • Use a quiet approach and avoid repeatedly spotlighting or chasing an animal around a den or refuge.
  • Treat an uncertain sighting as unconfirmed until supported by clear photographs, a capture, diagnostic hair or another reliable record.

Survey methodology: 3. Cage trapping and release

Use cage trapping only under an approved survey design and by experienced personnel, because the species is semi-arboreal, mobile and associated with complex rocky habitat.

The published Kimberley study used 30 cage traps at each site, spaced 30 m apart, baited at dusk and checked before dawn; it accumulated 4,470 trap nights across the study.

Published trap layout: 30 cage traps per site, 30 m apart

Published trap checking period: Set at dusk and checked the following morning before dawn

Published study effort: 4,470 trap nights

  • Place traps near likely travel routes, rainforest edges, fruiting trees, creek lines, rock shelters and den features rather than in open, recently burnt ground alone.
  • Use the study bait mixture of peanut butter, oats, honey and apple only where it is permitted by the project approval and fauna ethics requirements.
  • Check traps before dawn and remove traps promptly when heat, rain, flooding, fire or predator activity creates a welfare risk.
  • Record trap locations, trap nights, weather, bait, captures, non-target animals, body measurements, sex, age class, microchip or other identification and release location.
  • Do not infer absence from a short trapping session, low trap success or a failure to capture animals in dense or inaccessible habitat.

Survey methodology: 4. Camera traps and supporting detection methods

Use baited or unbaited camera traps at rocky hills, springs, rainforest edges, creek lines, likely den features and fruiting trees where cage trapping is unsuitable.

Camera trapping detected the species at 34 of 47 Yampi Sound sites over two wet seasons in 2016 to 2017, and at 10 of 22 Mitchell Plateau sites in 2016, showing the value of repeated remote detection.

Published camera result: 34 of 47 Yampi Sound sites over two wet seasons

Published Mitchell Plateau result: 10 of 22 camera sites in 2016

  • Keep cameras operating through both wet and dry conditions where practicable, and record camera model, placement, height, lure or bait, operating dates and effort in camera nights.
  • Use diagnostic photographs to confirm the orange-brown dorsal stripe, tail pattern and white feet, and retain original image files with the record.
  • Hair collection can support mammal surveys generally, but the survey guidelines do not identify hair sampling as a confirmed identification method for this species, so do not use it as the sole basis for a record.
  • Combine camera results with spotlighting, habitat searches or cage trapping when a decision about presence or absence is required.

Survey methodology: Minimum effort and reporting

  • Prepare a species-specific survey design that covers all potentially suitable habitat and explains why the selected method, timing and effort can detect a largely nocturnal, low-detection tree-rat.
  • Use more than one method where habitat is dense, rocky, remote or difficult to access, and repeat surveys across suitable weather and seasonal conditions rather than treating one negative survey as proof of absence.
  • Report survey effort as person hours, kilometres or area searched, trap nights, camera nights, number of cameras and nights deployed, and the number and location of habitat features checked.
  • Submit confirmed records, photographs and negative survey data to the relevant state or territory wildlife database, including the Northern Territory database where work occurs in the NT, and retain the project spatial data.
  • Report any capture, injury, mortality, translocation or suspected NT occurrence to the relevant wildlife authority under the project approval and permit conditions.

Before habitat disturbance: Avoid and minimise habitat loss

The species is not listed under the EPBC Act on the supplied current sources, but it is listed as Critically Endangered in the Northern Territory. Its remaining Australian range is concentrated in the north-west Kimberley, with historical and uncertain records in the NT, so project decisions must be based on current habitat and state or territory requirements.

Recorded home range: 0.24 to 30.44 ha by minimum convex polygon, with a mean of 6.3 ha in the Artesian Range

Recorded foraging movement: Individuals may travel more than 1 km from a refuge into open savanna at night

  • Avoid clearing rainforest patches, rainforest edges, riparian vegetation, fruiting and hollow-bearing trees, pandanus stands, rugged sandstone, cliffs, scree and rock shelters wherever practicable.
  • Retain connected vegetation between rainforest, riparian areas, woodland, rocky refuge areas and foraging habitat, rather than isolating individual trees or small habitat patches.
  • Keep intact den features and nearby foraging habitat together, because individuals may travel more than 1 km from a refuge into open savanna to forage at night.
  • Avoid works in recently occupied habitat where fire, cattle, feral cats or other disturbance has already reduced cover and food resources.

Before habitat disturbance: Approvals and project planning

  • Check the EPBC Act referral requirements for the whole project and other protected matters; the Golden-backed Tree-rat itself is not currently listed under the EPBC Act in the supplied sources.
  • In the Northern Territory, obtain the required approvals and fauna permits before disturbing potential habitat, because the species is listed as Critically Endangered under the Territory Parks and Wildlife Conservation Act 1976.
  • Confirm current Western Australian and Northern Territory species status directly with the relevant wildlife authority before finalising the impact assessment, because older documents record different jurisdictional statuses.
  • Prepare a fauna management plan covering survey results, retained habitat, pre-clearance checks, spotter catcher roles, stop-work triggers, handling, release, injury response, records and reporting.
  • Stage clearing so that retained habitat is available before adjoining habitat is disturbed, and schedule night-sensitive works to avoid active periods where the risk assessment identifies a likely encounter.

Before habitat disturbance: Prepare and mark the site

Den tree finding: The species used 11 of 78 available tree species in the published study, and trees with more than three hollows were more likely to be used

Recorded den sites: 53 percent cliffs, 33 percent live tree hollows, 5 percent logs, 4 percent scree, 4 percent rock castles and 0.3 percent dead-standing trees

  • Complete a pre-clearance survey in all potential habitat, including trees, cliffs, scree, logs, rock piles, pandanus crowns, creek lines and rainforest edges.
  • Mark no-go areas, retained trees, retained rocky refuges, access routes and fauna release areas before machinery enters the site.
  • Identify hollow-bearing trees and trees with multiple hollows, as the study found that 64 percent of used den trees had more than three hollows.
  • Install exclusion fencing where it can prevent entry into retained habitat, but do not fence animals into a work area or block access to water, food or known refuge features.
  • Brief machine operators and supervisors on the species' appearance, nocturnal activity, likely den features, stop-work triggers and the spotter catcher's authority.

During habitat disturbance: Daily controls and spotter catcher procedure

Disturbance controls should prevent animals being trapped, exposed to machinery or deprived of nearby refuge. A fauna spotter catcher must be able to stop work and must coordinate closely with the operator before each habitat feature is disturbed.

  • Inspect the work area each day before clearing, including retained trees, hollows, logs, rock piles, cliffs, scree, pandanus foliage, creek lines and dense vegetation.
  • Use a qualified and authorised fauna spotter catcher in accordance with the approval and state or territory requirements, with clear authority to stop machinery.
  • Clear progressively towards retained habitat and away from escape routes where the site layout allows, while avoiding pushing animals into active machinery, roads or areas occupied by cats or other predators.
  • Keep machinery stopped while the spotter catcher checks a habitat feature, confirms the escape path and gives the operator permission to proceed.
  • Do not undertake night works, hot works or clearing during severe weather where these conditions increase the likelihood of harming a nocturnal animal or make detection unreliable.

During habitat disturbance: Habitat feature handling

Shelter sites recorded: Tree hollows, cliffs, logs, scree, rock castles, dead-standing trees and loose woven nests beneath pandanus crowns

  • Retain hollow-bearing trees wherever practicable and do not fell them until they have been inspected for animals by the spotter catcher.
  • If a hollow-bearing tree must be removed, use a staged, controlled procedure approved in the fauna management plan, with the tree directed towards retained habitat and checked after it is lowered.
  • Inspect logs, rock piles, scree, cliffs and pandanus crowns before moving them, because the species uses all of these features for shelter or denning.
  • Do not roll or crush rocks, logs or hollow sections with machinery before the spotter catcher has completed the inspection.
  • Protect creek lines, rainforest edges, fruiting trees and dense refuge vegetation from sediment, dust, fire, storage areas and unnecessary vehicle access.

During habitat disturbance: Animal finds, stop-work and records

  • Stop work immediately if a Golden-backed Tree-rat is seen, captured, injured, trapped in a hollow or refuge, or suspected to be in a habitat feature being disturbed.
  • Keep people, machinery, dogs and vehicles away from the animal and allow it to move into nearby retained habitat where it can do so safely.
  • Do not handle the animal unless the person is authorised, trained and acting under the relevant permit, approval and fauna management plan.
  • If handling is required, use a secure, ventilated container, minimise holding time, keep the animal cool and quiet, and release it into suitable nearby retained habitat with shelter and connectivity.
  • Do not relocate the animal long distances or across unsuitable habitat without specific approval and wildlife advice.
  • Send injured animals to an appropriately authorised veterinarian or wildlife carer as soon as practicable, and record the handover.
  • Treat injury, mortality, repeated disturbance of a den, loss of a hollow-bearing tree outside the approved area, or a tree-rat found in the active clearing area as a stop-work and incident-reporting trigger.
  • Record the date, time, coordinates, habitat feature, photographs, animal condition, identification method, handling, release or treatment location, personnel, machinery involved and corrective action.

After habitat disturbance: Post-clearance checks

Post-clearance work must confirm that no animal remains in the disturbed area and that retained habitat controls are still working. Monitoring should be scaled to the habitat affected, the level of uncertainty and any permit or approval conditions.

  • Inspect the cleared footprint, retained trees, logs, rock features, pandanus crowns, drainage lines and exclusion fencing after each clearing stage and again before the next stage begins.
  • Check for live animals, carcasses, trapped fauna, damaged hollows, displaced rocks, damaged logs, fire impacts and new access by cats, cattle or other herbivores.
  • If a den feature is damaged or an animal is detected after clearing, stop the relevant work and follow the fauna management plan and incident response procedure.
  • Remove temporary fencing and equipment only after the spotter catcher or ecologist confirms that animals cannot be trapped by the controls.

After habitat disturbance: Monitoring and reporting

  • Repeat camera trapping, spotlighting or targeted cage trapping in retained habitat where clearing removed known or suspected habitat, where an animal was found, or where approval conditions require monitoring.
  • Use fixed camera locations, consistent deployment periods and the same effort measures so that changes in detection can be compared between monitoring events.
  • Monitor retained rainforest, riparian vegetation, fruiting and hollow-bearing trees, rocky refuges and nearby savanna rather than only the cleared footprint.
  • Report confirmed records, injuries, deaths, relocations, breaches, monitoring results and corrective actions to the relevant authority and wildlife database.
  • Maintain fire, predator, herbivore, weed and access controls required by the project approval, and review them if monitoring shows reduced habitat use.

After habitat disturbance: Offsets and residual impacts

  • Apply the avoid, minimise, rehabilitate and offset hierarchy required by the approval pathway before considering an offset.
  • Where residual impacts remain, document the area and condition of habitat lost, the affected refuge features, the number of animals detected, the duration of disturbance and the management actions used to address the impact.
  • Do not treat an offset as a substitute for retaining occupied habitat, hollow-bearing trees, rocky refuges, rainforest patches or connected foraging areas.

Rehabilitation actions: Restore habitat structure and food resources

Rehabilitation should restore the shelter, food and fire protection used by the Golden-backed Tree-rat. It should connect retained refuges with productive vegetation and be monitored for actual use, not just plant establishment.

Recorded food: Flowers, fruits and termites are main dry-season foods, with grasses, ants and beetles also recorded

Published foraging observations: 66 percent in trees, 13.9 percent on the ground, 11 percent on cliff faces and 8.3 percent in boulder scree along a river edge

  • Replant or encourage native rainforest edge, riparian, eucalypt woodland, savanna, Livistona and pandanus vegetation appropriate to the local site and reference community.
  • Prioritise native fruiting and flowering trees and shrubs that provide food, including locally appropriate species documented as foraging resources such as Terminalia hadleyana, Planchonia rupestris, Celtis philippensis and Owenia vernicosa.
  • Protect planted and regenerating vegetation from cattle, donkeys, pigs and other introduced herbivores until it is established and can provide cover and food.
  • Reconnect retained rainforest, riparian areas, rocky refuges and savanna so animals can move between denning and foraging areas.

Rehabilitation actions: Reinstate shelter and refuge features

Published den use: Most recorded dens were cliffs and live tree hollows, with additional use of logs, scree, rock castles and dead-standing trees

  • Retain natural hollow-bearing trees and allow replacement trees to mature; do not remove standing dead trees where they provide safe shelter unless they pose a demonstrated safety risk.
  • Reinstate locally sourced logs, stable rocks and scree in configurations that provide shelter and do not create hazards or block drainage.
  • Use nest boxes only where an ecologist and the relevant authority consider them suitable, because the species' documented den use includes hollows, cliffs, logs, scree and pandanus nests rather than a demonstrated nest-box standard.
  • Protect restored refuge features from machinery, repeated fire, weed invasion, cattle and feral cats.

Rehabilitation actions: Fire, predator and weed management

  • Use local Indigenous and ecological fire knowledge to reduce frequent, extensive and intense late dry-season fires around retained rainforest, fruiting trees, hollow-bearing trees and rocky refuge areas.
  • Plan a patchy, low-intensity burning regime where fire is required, with unburnt refuge areas and protection for fire-sensitive vegetation.
  • Control feral cats in accordance with approved site programs, particularly around recently burnt areas, islands and retained rocky habitat.
  • Control weeds and maintain native understorey cover, because reduced vegetation cover can increase predation risk and reduce food resources.
  • Prevent new cattle, pig, donkey and horse access to rehabilitation areas and repair exclusion fencing after storms or fire.

Rehabilitation actions: Rehabilitation monitoring and success measures

Generation length: Approximately 3 years

Published habitat-use study: 18 individuals were tracked over seven tracking periods between June 2012 and January 2014, producing 945 location points

  • Inspect rehabilitation after each wet season and after fire, record survival and recruitment of planted trees and shrubs, and repair erosion, fencing and failed plantings.
  • Use cameras and targeted night searches to test whether the species returns to restored habitat, with effort reported as camera nights, person hours and search distance.
  • Measure retention and condition of hollow-bearing trees, logs, rocks, pandanus, rainforest patches, fruiting plants, understorey cover and connectivity between refuge areas.
  • Set success measures around habitat condition and confirmed fauna use, not only the number of plants installed.
  • Continue monitoring long enough to detect establishment of replacement hollows and stable use of restored habitat, recognising that generation time is approximately 3 years.

Sources