Skip to main content
Dryad logo

Carcasses attract invasive species and increase artificial nest predation in a desert environment


Newsome, Thomas; Spencer, Emma (2021), Carcasses attract invasive species and increase artificial nest predation in a desert environment, Dryad, Dataset,


In addition to feeding on animal remains, many scavengers also function as predators. Carcasses may therefore affect local animal communities by attracting facultative scavengers and increasing predation risk for other species in the vicinity of the carcasses. This risk may be elevated in low productivity environments, especially where humans increase carcass production and where facultative scavengers include invasive species. In June and October 2018, we monitored experimentally placed red kangaroo (Osphranter rufus) carcasses and artificial bird nests in two different habitats in the Simpson Desert, Australia, to identify the nest predators attracted to the carcasses, and to determine how carcasses affect overall and predator-specific nest predation. We modelled our nests to approximate those of the ground nesting little buttonquail (Turnix velox) and the endangered night parrot (Pezoporus occidentalis). Native Corvus spp. and then invasive red foxes (Vulpes vulpes) were the top carcass visitors and nest egg predators. Carcass presence and open habitat increased overall nest predation and fewer artificial parrot nest eggs were depredated compared to those of quail. Open habitat and carcass presence only increased predator-specific nest predation by foxes, but corvid nest predation was highest in June 2018, and for the artificial quail nest types. Foxes were the main predator of eggs from night parrot nests. Our study shows that carcass provisioning by humans may have indirect, deleterious effects on ground nesting birds, and indicates that foxes might pose a greater threat to night parrot populations than previously recognised.


Site set-up and carcass provisioning

We conducted our experiments in June and October 2018 (Australia winter and spring, respectively), to account for seasonal differences. Eighty sites were established, split evenly across the two study periods and generally alternating between carcass ‘present’ (n = 40) and carcass ‘absent’ (n = 40) sites (Fig. 1b). Within each study period, sites were spaced ~0.50 km (range: 0.49–1.68 km) apart, with distances of about 1 km (range: 0.91–2.59 km) between each carcass ‘present’ site. These distances were chosen to minimise carcass scent travel between sites, such that scavengers would have to actively forage and seek out carcass sites rather than being able to move quickly from one site to another. Carcass-present and absent sites were also split evenly between dune crest and valley habitats. We classed dune crests as “open” and the inter-dune valleys as “closed” habitats. Open habitats were devoid of tree cover, while closed habitats were sparsely populated by gidgee trees.

The centre point of both carcass-present and absent sites was marked with a single black 1 m high stake (Fig. 1c). At carcass-present sites, we placed a single red kangaroo (Osphranter rufus) carcass (mean 25 kg +/- 0.5 se) sourced from pre-planned local culls, and as such were not killed for the purpose of this study. Any carcass displaying evidence of disease (e.g. heavy parasite loads) was not used. Following collection, carcasses were placed in the field without freezing within 24 h (June) or 36 h (October) of collection. Scientific licences were obtained to relocate and monitor the carcasses (SL WA0006737), and all research was approved by the University of Sydney Animal Ethics Committee (Project number 2017/1173).

Artificial nests

Artificial nests were installed on NW–SE transects in both carcass-present and absent sites, following the direction of the sand dunes. Transects were designed to intersect either the central carcass subsidy, or the stake used to mark each carcass-absent site. In June 2018, 6 quail nests (see design below) were positioned at 10 m, 30 m and 50 m along the transect such that each distance was sampled in both the NW and SE direction (total of 240 quail nests in June 2018). In October 2018, this method was replicated, with the addition of two parrot nests (see design below) set up 10 m on either side of the centre points in the carcass-present and absent sites. Parrot nests were set up only in dune crest sites, as these nests were constructed within spinifex hummocks only found consistently in the dune crest habitat (total of 240 quail and 40 parrot nests in October 2018).

We modelled artificial nests to replicate those of two species of ground nesting birds in the region. The first nest model mimicked the nest of the little buttonquail (hereafter ‘quail’ nests), while the second mimicked a night parrot nest (‘parrot’ nests). Little buttonquails are commonly sighted in our study area and build their nests on the ground in grassland, usually at the edge of small shrubs and overhanging grasses. In contrast, night parrots build their nests in the centre of dense shrubs and, in particular, large partially dead spinifex hummocks (Murphy et al., 2017). While night parrots have not been sighted in our study area, they have been recorded in nearby areas with comparable habitats (Murphy et al., 2017); as an endangered species, they provide a relevant conservation model to work with (Murphy et al., 2018).

Artificial quail nests were positioned no more than 5 cm from the base of spinifex hummocks, small shrubs, low lying gidgee trees or logs, by creating a small indent (~10 cm diameter) in the ground with the palm of a hand (Fig. 1d). Artificial parrot nests were placed in spinifex hummocks (1–4 m diameter) using a broom handle to force a 15 cm wide hole in the hummock. A broom pole was then used to create a small chamber (< 30 cm deep and wide) at the base of the plant where the roots enter the sand (Fig. 1e). Each nest contained 1 artificial egg, made from plasticine modelling clay and coated with Plasti-dip™, and 1 commercial quail egg, collected fresh and then refrigerated until use. Nests were also scented with ~5 g of feathers and droppings collected from domestically reared quails (Game Farm Pty Ltd., Galston NSW). The quail egg, droppings and feathers were used as attractants. Quail droppings were replenished at any sites where eggs remained 6 days after placement. We used latex gloves smeared in quail droppings to reduce human scent on nests and, when possible, human footprints were smoothed using a rake to minimise any obvious pathways in the sand that might lead predators to the nests. A GPS unit (Garmin GPSMAP 64sc) was used to mark each nest site.

All nests were set out between ~11 am and 8 pm, 1 day after carcass placement to ensure that scents associated with transporting the carcasses were kept separate from the nest sites. When setting the nests, we ensured that any nearby scavengers and predators, especially corvids, were flushed before starting, to ensure that they would not be able to watch nest placement. As nest placement took ~15 min per site, it was not expected that this had any significant impact on scavengers using the carcass sites.

Carcass and nest site monitoring

To allow ongoing monitoring and detection of scavengers visiting the carcasses, a remote camera (Reconyx Hyperfire PC800) was positioned on the stake marking the centre of each site, 3–4 m from the kangaroo carcass (Fig. 1c). The camera was programmed to take continuous photographs when triggered by thermal movement around the carcass (rapidfire, no wait period). To prevent removal of carcasses from the camera monitoring frame, carcasses were secured to the ground by wire attaching the neck and Achilles tendon of the animal to two metal stakes spaced ~0.6 m apart. We monitored carcasses over the same period we monitored nests (14 days). We examined all photographs collected by the cameras and tagged them by the animal species present in the frame and whether it was engaged in scavenging or not. We then examined the tagged images and compiled a list of all vertebrate species observed feeding on at least one carcass. Species that we suspected to be feeding on insects on the carcass, but not the carcass itself, were included on this list.

Nests were monitored in the landscape for 14 days, as little buttonquails and other ground nesting birds in the region generally hatch in an equivalent time (Higgins et al., 1990). Every 2 days, we visited each nest and recorded evidence of predation. Sites where either clay or quail eggs had been bitten into or removed were considered depredated. The plastic-coated clay enabled identification of species attempting to depredate the eggs (i.e. from the tooth or beak marks in the clay) and helped to minimise the scent of the clay, which has been linked to higher rates of predation (Purger et al., 2012). In addition to the plastic-coated clay eggs, identification of nest predators was aided by smoothing sand around the nest site and through the use of remote cameras (Reconyx Hyperfire PC800), which were placed 3 m away from select nests at carcass and control sites and programmed to take pictures continuously when any animal motion was detected (79 nests were monitored by cameras). As camera placement may influence nest predation rates, cameras were hidden in adjacent vegetation and set on nests at random sites (Richardson et al., 2009). No cameras were placed on artificial parrot nests.

When reviewing the carcass and nest site camera images, differentiating corvid species was challenging. In most cases, we could identify Australian ravens by their larger size and stature, smaller group numbers and long throat hackles that form a layered beard. We identified crows by their thinner appearance, smaller throat hackles, white feather bases and generally larger group numbers. While both little crows and Torresian crows occur in our study area and are difficult to differentiate without hearing their calls, we are fairly confident that all crows appearing in our camera images were little crows. This assumption is based on the fact that no Torresian crows were detected in the field during the two 2-week periods of fieldwork carried out as part of this study (i.e. based on calls), or by concurrent bird surveys carried out in the study area and surrounding region (pers. comm. Ayesha Tulloch). Further, continuous bird surveys in the region typically only place Torresian crows in the area following high rainfall events and we conducted our study during a dry period. At nest sites, corvid predation events could only be specified to species when using camera images (and not by prints in the sand or beak marks in the clay eggs).