ArticlePDF Available

Ecological Role of an Apex Predator Revealed by a Reintroduction Experiment and Bayesian Statistics


Abstract and Figures

Recent studies suggest that apex predators play a pivotal role in maintaining healthy, balanced ecosystems. However, a criticism of studies investigating the ecological role of apex predators is that understanding does not come from manipulative experiments. Here, we use a before-after-control-impact-paired design to test predictions generated from trophic cascade theory (TCT) and mesopredator release hypothesis (MRH) by experimentally introducing dingoes into a 37km 2 paddock and measuring the resultant effects on mammal assemblages. To increase precision of parameter estimates generated by our experiment, we used a Bayesian framework which included prior information recorded from a mensurative study located in a comparable ecosystem that contrasted indices of mammal abundance where dingoes were common and rare. Results of the mensurative study were consistent with TCT and MRH. When using an uninformative prior, results of the experiment showed that dingo addition only had a negative effect on kangaroo activity. Use of an informative prior reduced uncertainty of the posterior mean parameter estimates from the experiment and suggested that red foxes were affected negatively and small mammals and rabbits were affected positively by dingo introduction. However, the prior had a strong influence on the posterior mean effect sizes for small mammals, rabbits and foxes. Opposite polarity of uninformed and prior parameter estimates for rabbits suggests that the prior was incompatible with the uninformed estimates from the manipulative experiment. Our study shows how use of logical informative priors can help to overcome statistical issues associated with low-replication in large-scale experiments, but the strong influence of the prior, means that our findings were driven largely by the mensurative study.
This content is subject to copyright. Terms and conditions apply.
Ecological Role of an Apex Predator
Revealed by a Reintroduction
Experiment and Bayesian Statistics
K. E. Moseby,
* M. S. Crowther,
and M. Letnic
Arid Recovery Ltd., P.O. Box 147, Roxby Downs, South Australia 5725, Australia;
Centre for Ecosystem Science, University of New
South Wales, Sydney, New South Wales 2052, Australia;
School of Life and Environmental Sciences, University of Sydney, Sydney,
New South Wales 2006, Australia
Recent studies suggest that apex predators play a
pivotal role in maintaining healthy, balanced
ecosystems. However, a criticism of studies inves-
tigating the ecological role of apex predators is that
understanding does not come from manipulative
experiments. Here, we use a before-after-control-
impact-paired design to test predictions generated
from trophic cascade theory (TCT) and meso-
predator release hypothesis (MRH) by experimen-
tally introducing dingoes into a 37-km
and measuring the resultant effects on mammal
assemblages. To increase precision of parameter
estimates generated by our experiment, we used a
Bayesian framework which included prior infor-
mation recorded from a mensurative study located
in a comparable ecosystem that contrasted indices
of mammal abundance where dingoes were com-
mon and rare. Results of the mensurative study
were consistent with TCT and MRH. When using
an uninformative prior, results of the experiment
showed that dingo addition only had a negative
effect on kangaroo activity. Use of an informative
prior reduced uncertainty of the posterior mean
parameter estimates from the experiment and
suggested that red foxes were affected negatively
and small mammals and rabbits were affected
positively by dingo introduction. However, the
prior had a strong influence on the posterior mean
effect sizes for small mammals, rabbits and foxes.
Opposite polarity of uninformed and prior param-
eter estimates for rabbits suggests that the prior was
incompatible with the uninformed estimates from
the manipulative experiment. Our study shows
how use of logical informative priors can help to
overcome statistical issues associated with low-
replication in large-scale experiments, but the
strong influence of the prior means that our find-
ings were driven largely by the mensurative study.
Key words: apex predator; trophic cascade; me-
sopredator release; Bayesian prior; BACIP; land-
scape-scale experiment.
Apex predators can play a pivotal role in main-
taining, healthy, balanced ecosystems, and recent
research suggests that the magnitude of their effects
as ecosystem regulators has been enormously
underestimated (Ripple and others 2014). A key
reason for the relatively recent realisation of the
magnitudes of apex predators’ effects on ecosys-
Received 9 December 2017; accepted 28 May 2018
Electronic supplementary material: The online version of this article
( contains supplementary
material, which is available to authorized users.
Authors’ Contribution KEM and ML conceived and designed the
study. KM and ML performed the field research. MSC analysed the data.
All authors wrote the paper.
*Corresponding author; e-mail:
Ó2018 Springer Science+Business Media, LLC, part of Springer Nature
tems is that the temporal and spatial scales required
to conduct controlled experiments on large carni-
vores are logistically prohibitive (Letnic and others
2012). Additionally, in many jurisdictions legal
impediments and ethical concerns prevent the
undertaking of experimental manipulations of apex
predator abundance (Estes and others 2011).
Consequently, our understanding of apex preda-
tors’ effects on ecosystems comes primarily from
the accumulation of evidence stemming from un-
planned natural experiments made possible by the
extirpation of predators from ecosystems or fluc-
tuations in the their abundance associated with
climatic variability and disease epidemics (Ripple
and others 2014). However, the evidence regarding
the ecological role of apex predators is not experi-
mental in many studies, and thus, the reported
effects might be due to other underlying factors
(Mech 2012; Kauffman and others 2013). Planned
manipulative experiments with adequate controls
and before/after experimental designs are required
to provide empirical evidence for apex predator
effects and validate the outcomes derived from
natural experiments.
Current evidence suggests that apex predators
can shape ecological assemblages via a multitude of
interaction pathways (Estes and others 2011; Ordiz
and others 2013). Two key interaction pathways
via which apex predators shape ecosystems are
through their suppressive effects on large and
medium herbivores and smaller predators (Ritchie
and Johnson 2009; Colman and others 2014).
Trophic cascade theory predicts that suppression of
apex predators’ ‘‘top-down’’ effects will result in
the irruption of herbivore populations and subse-
quent depletion of plant biomass (Schmitz 2008).
The mesopredator release hypothesis predicts that
when the abundance of apex predators is reduced,
the abundance or activity of smaller predators in-
creases owing to the relaxation of direct killing and
interference competition by apex predators (Crooks
and Soule
´1999; Moseby and others 2012; Pasanen-
Mortensen and others 2013). In turn, irruptions of
mesopredators are often accompanied by declines
in the abundances of their prey owing to elevated
rates of predation by mesopredators (Ritchie and
Johnson 2009).
One potential way to overcome the typically low
levels of replication and hence statistical power
available when conducting studies of apex preda-
tors’ effects on ecosystems is to employ Bayesian
statistical frameworks which incorporate prior
information (McCarthy and Masters 2005; Morris
and others 2015). Modelling using Bayesian infer-
ence can increase the precision of model parameter
estimates by combining previous knowledge with
newly collected data to produce a ‘‘posterior’’ dis-
tribution (Martin and others 2013). In this study,
we use a Bayesian framework to evaluate the ef-
fects of manipulating the abundance of an apex
predator, the dingo (Canis dingo), on mammal
assemblages in arid Australia, by combining
experimental data on dingoes’ effects with ‘‘prior’’
data from a correlative study investigating differ-
ences in mammal assemblages in areas where
dingoes were common and rare on either side of a
dingo-proof fence.
At our experimental site, situated at Arid
Recovery in South Australia (Figure 1), we intro-
duced dingoes into a 37-km
dingo-proof enclosure
housing natural populations of mesopredators and
mammalian prey species. Because we only had the
resources to construct one enclosure, we used a
Before-After-control-impact-paired (BACIP) design
for our experiment and compared the responses of
mammals in our experimental enclosure with a
nearby control area where dingoes were excluded
by the dingo fence and any incursions subject to
ongoing population control. To increase the preci-
sion of the parameter estimates generated by our
BACIP analysis, we used ‘‘prior’’ information on
the effect of manipulating dingo abundance de-
rived from a ‘‘natural experiment’’ which com-
pared mammal assemblages on either side of
Australia’s dingo-proof fence. The dingo fence sites
were located approximately 400 km to the east of
the experimental site in the Strzelecki Desert and
had similar landforms, flora and fauna to the
experimental site (Figure 1B, C).
Applying the mesopredator release hypothesis,
trophic cascade theory and knowledge gleaned
from earlier correlative studies (Letnic and others
2009; Colman and others 2014), we expected that
the magnitude of the effect of dingoes on other
mammals should scale with their body size.
Specifically, we predicted that: (1) the activity of
herbivorous kangaroos (15–90 kg) should be lower
where dingoes are abundant as they are within the
preferred prey size range for dingoes; (2) the
activity of mesopredators (foxes 5–7 kg and cats 3–
6 kg) should be lower where dingoes are abundant
owing to previous evidence of the dingoes’ direct
suppressive effects on populations of these species
and (3) activity of smaller prey species, namely
small mammals (<0.1 kg) and rabbits (0.7–1.5 kg)
should be greater where dingoes are abundant be-
cause their populations should benefit indirectly
from dingoes’ suppressive effects on mesopreda-
K. E. Moseby and others
Strzelecki Desert Observation Site
We measured the differences between indices of
abundance derived from track surveys on either
side of the dingo fence that straddles the border
between New South Wales and South Australia in
the Strzelecki Desert, Australia (Figure 1A). These
data were our prior information on the effects that
dingoes have on other mammal species for our
Bayesian analyses. The climate, vegetation, fauna
and geomorphology of the Strzelecki Desert are
similar to the Roxby Downs study area situated
approximately 400 km to the south-west which
receives a mean annual rainfall of 200 mm. The
dingo fence was built along the border of South
Australia and New South Wales in the early
twentieth century for excluding dingoes from New
South Wales. The position of the NSW/SA border,
and hence dingo fence, following the meridian
141°E was established by a decree from the King of
England in 1837 prior to the region’s exploration
and thus does not reflect any natural biological or
geographical boundary. Dingoes are present but
rare on the New South Wales side of the dingo
fence where intensive efforts are undertaken to
control their numbers but are relatively common in
South Australia (Caughley and others 1980; Letnic
and others 2009; Letnic and Koch 2010).
Arid Recovery Experimental Site
We used a before-after-control-impact paired de-
sign (Stewart-Oaten and others 1986) to evaluate
the effect of dingoes on mammal assemblages in an
Figure 1. AMap showing the location of the Strzelecki Desert study site where the prior information for this study was
obtained and the experimental study site at Roxby Downs. Sites where dingoes are common and rare are denoted by
closed and open symbols, respectively. The dashed line indicates the dingo fence. BOblique aerial photograph of the
landscape at Roxby Downs showing linear sand dunes and inter-dunal swales. COblique aerial photograph of the
Strzelecki Desert landscape showing linear sand dunes and inter-dunal swales.
Ecological Role of an Apex Predator
arid landscape at Arid Recovery in the state of
South Australia. There was a fenced treatment area
into which dingoes were added, and one unfenced
control area where dingoes were present but oc-
curred at very low densities due to culling activities
which were intensified when dingoes were de-
tected. Sampling was conducted simultaneously in
each treatment area on three occasions prior to
dingo addition and eight occasions after dingo
addition. Using this experimental design, an effect
of dingo addition would be indicated by a signifi-
cant interaction between Treatment (dingo, con-
trol) and Period (before, after dingo addition). The
study was conducted over 4 years between
December 2008 and December 2012. The climate at
the study area is arid, with a mean annual rainfall
of only 166 mm. Rainfall is unpredictable and
In December 2008, two dingoes were experi-
mentally added to the 37-km
paddock surrounded
by a dingo-proof fence constructed of wire netting
(1.6 m high). The upper margin of the fence was a
‘‘floppy top’’ curving inwards to keep dingoes, cats
and foxes within the paddock but still allowing cats
and foxes to climb in (Moseby and others 2012).
The fence was only 1.6 m high, and foxes and cats
readily accessed the dingo pen by climbing up the
wire mesh. Remote cameras recorded foxes climb-
ing into the paddock on several occasions (K. Mo-
seby pers obs). The netting fence was built from 50-
mm netting. This allowed small rabbits and small
mammals to pass through the fence. The fence was
impervious to kangaroos. Foxes, cats, red kanga-
roos (Macropus rufus), small mammals (Notomys
alexis and Pseudomys spp.) and European rabbits
(Oryctolagus cuniculus) were all naturally present
within the dingo paddock when it was constructed
and at the time of dingo addition.
Several factors suggest that the experimental
paddock closely resembled a natural environment
and the introduced dingo pair behaved like wild
dingoes. Dingo home ranges in the arid zone can
reach 272 km
(Eldridge and others 2002) and
even 403 km
(Newsome and others 2013), but
home ranges vary considerably across Australia
(10–272 km
, Robley and others 2010) with arid
zone female home ranges reported as 56 km
(Thomson 1992) and 47 km
(Corbett 1995) and
even 10 km
when food resource availability is
high (Newsome and others 2013). The size of the
paddock, the presence of permanent water and the
high density of prey items suggest that our paddock
was large enough to support natural home range
sizes, particularly of female dingoes, as they
approximate the lower arid zone home range of
dingoes. The range in density of the dingo popu-
lation in the experimental paddock over the course
of the study (5–27 dingoes per 100 km
) was sim-
ilar to that recorded in a study of free-ranging
dingoes in the arid zone (1–22 per 100 km
(Thomson 1992). Dingoes were recorded howling
and scent marking, displayed normal pack beha-
viour and were captured from the wild adjacent to
the experimental paddock (Moseby and others
The paddock was large enough to contain several
home ranges of cats and foxes; previous home
range studies in the same region estimated average
home range sizes of 17.4 and 16.6 km
for cats and
foxes, respectively (Moseby and others 2009) and
cat home ranges overlapped. Track counts and
previous research suggest that the 37-km
was large enough to encompass thousands of rab-
bits (home range 2.1–4.2 ha; (Moseby and others
2005) and small mammals and hundreds of kan-
The southern section of the dingo paddock
comprised a clay inter-dunal swale more than 2 km
wide and vegetated with chenopod shrubs, Atriplex
spp. and Maireana astrotricha. Longitudinal sand
dunes supporting woody shrubs, Acacia ligulata and
Dodonaea viscosa were present in the northern sec-
tions, separated by 100–400-m-wide swales. A dam
within the paddock contained water throughout
the study.
The control area was situated 5 km east of the
dingo paddock, a distance considered sufficient to
ensure independence but close enough to contain
similar habitat types and reflect similar climatic
events. Dingoes were excluded from the control
area by a dog proof fence to the north. If incursions
of dingoes into the control area were detected, they
were culled by land-managers via shooting or
opportunistic laying of poison meat baits. Habitats
within the control area were similar to those
available in the dingo paddock with a large clay
swale, an area of closely spaced sand dunes, a dam,
and an area of stony desert. The dam within the
control area contained water throughout the study.
The paddock was conservatively stocked with
domestic cattle (Bos taurus). The control area was
part of a larger stock paddock fenced using a stan-
dard barbed-wire cattle fence which did not pre-
vent the movements of dingoes, foxes, cats,
kangaroos, rabbits or small mammals.
Experimental Addition of Dingoes
In December 2008, male and female dingoes were
captured approximately 50 km north of the dingo
K. E. Moseby and others
paddock and released into the dingo paddock.
These wild dingoes were captured using soft catch
Jake äfoot hold traps set around a cattle carcass.
We lightly anaesthetised the captured adult dingoes
using a mixture of 1 ml of medetomidine
hydrochloride and 0.5 ml of ketamine, adminis-
tered intramuscularly. The anaesthetic was re-
versed using 0.5 ml of atipamezole hydrochloride.
Dingoes were fitted with GPS Argos satellite collars,
transported in an air-conditioned car and released
within the dingo paddock on the same morning as
capture. Dingoes were checked after 2 h and were
then radio-tracked daily for the first week. Radio-
tracking fixes indicated that both dingoes began
moving throughout the paddock within a few
hours of release (Moseby and others 2012).
Activity Indices at the Strzelecki Desert
We used track counts to index the abundance of
dingoes, foxes, cats, kangaroos, rabbit and small
mammals on either side of the dingo fence on 11
occasions between 2007 and 2014. At each site on
each occasion, we established 25 track detection
stations at 1-km intervals along low-use, single-
lane dirt roads (Letnic and others 2009). The track
stations (2 m long) were raked and smoothed to
produce a smooth surface that spanned the width
of the road (ca. 3 m on average). The tracking
stations were monitored and swept daily for three
consecutive nights. Each morning a unique mark
was made in the corner of each tracking station.
This mark was used to determine if wind, rain or
vehicles might affect the observer’s ability to
interpret the plot the following morning. If the
mark could not be detected clearly the next
morning, the previous night’s record for that sta-
tion was considered invalid. For each track detec-
tion station, an index of the activity of each taxon
(dingo, fox, cat, rabbit, kangaroo, small mammal)
was expressed as the number of nights a plot was
visited by each species divided by the number of
nights that the plot was considered valid. An index
of activity for each taxon on each side of the dingo
fence on each occasion was calculated subse-
quently as the mean of all track plots.
Activity Indices at the Experimental Site
The activity of predator and prey species was
sampled in the dingo and control paddock three
times during the 11 months prior to the addition of
dingoes (February 2008, July 2008 and November
2008) and 10 times whilst dingoes were present in
the dingo pen (December 2008–July 2012). Indices
of dingo, fox, cat, kangaroo, small rodent and rabbit
activity were derived from the presence of spoor
along 200-m track transects (Read and Eldridge
2010) established away from roads in both the
control and dingo paddock in the three main
habitat types; sand dune, swale and creek-line. A
total of 39 transects (20 sand dune, 10 creek-lines
and 9 swale) were established in the dingo paddock
and 38 (20 dune, 8 creek-line and 10 swale) in the
control area. The 1-m-wide transects were swept
clean using a metal bar dragged behind an all-ter-
rain vehicle the night before the first of two con-
secutive mornings of track counts. The transects
were accessed using an all-terrain vehicle. For each
track transect, an index of the activity of each
taxon (dingo, fox, cat, rabbit, kangaroo, small
mammal) was expressed as the number of nights a
plot was visited by each species divided by the
number of nights that the plot was considered va-
lid. An index of activity for each taxon on each side
of the dingo fence on each occasion was calculated
subsequently as the mean of all track plots.
Statistical Analyses
The informative prior we used for our analyses was
the difference in the track-based index of abun-
dance of each taxon in the Strzelecki Desert,
treating each sampling occasion as a replicate. For
each variable (dingo, fox, kangaroos, cat, rabbit and
small mammals), we calculated the effect size (r),
precision and 95% Bayesian confidence intervals
(credible intervals) of the informative prior using
the data from the Strzelecki Desert. We used a
generalised linear model with a normal distribu-
tion, with Treatment and time being fixed factors
using the program OpenBUGS (Spiegelhalter and
others 2014). Data and code for the models that we
performed for the informative prior from the Strz-
elecki Desert are provided in the supplementary
We analysed the Arid Recovery experimental
data for each taxon using a Bayesian BACIP anal-
ysis (Crome and others 1996), to evaluate support
for the following two alternative models: (1) that
there is no prior information predicting magnitude
or direction of the experiment (that is, using an
uninformative prior with a mean of 0 and SD of
1000, equivalent to a frequentist analysis (McCar-
thy 2007); (2) that dingo addition had an effect on
the activity of kangaroos, foxes, cats, rabbits and
small mammals that was consistent with the prior
information from the Strzelecki Desert. We in-
cluded the variable ‘‘dingo’’ in our Bayesian anal-
Ecological Role of an Apex Predator
yses to evaluate whether the introduction of din-
goes led to a measurable increase in dingo activity.
Data and code for the models that we performed for
the experimental site at Roxby Downs are provided
in the supplementary information.
We calculated the informative priors from the
Strzelecki Desert by calculating the effect size (r)
which is the effect of being in the presence of
abundant dingoes (DingoFence
) on the indices of
dingo, kangaroo, rabbit, fox, cat and small mam-
mal, accounting for fixed differences among sam-
pling times (TimeEffect), using means of 0 and
precision (s) of 0.001. The model for the calculation
the expected differences in indices is represented
mi¼aþrDingoFenceiþTimeEffect ti
where a(the intercept) Normal(0,0.001), r (the
effect size) Normal(0,0.001), t
represents sam-
pling time, TimeEffect (the differences in indices
among times) Normal(0,0.001).
For the manipulative experimental site, treat-
ment (dingo addition, control) was considered a
repeated measure, with Before-After (before dingo,
after dingo BA
) and Time (sampling occasion, t
fixed factors (Stewart-Oaten and others 1986). We
used OpenBUGS to generate 100,000 samples from
the posterior distributions for each analysis, after
discarding the initial 10,000 samples as a ‘‘burn in’’
(following McCarthy and Parris 2004; McCarthy
and Masters 2005).
The effect size (r) of the experiment was calcu-
lated as follows:
1. the mean indices in each time prior to the
where y= the global reference capture
rate Normal(0,0.001), r0 = mean difference be-
tween treatment and control plots prior to the
manipulation Normal(0,0.001), t
= sampling
time Normal(0,0.001)
2. mean indices in each time in the first period
after the treatment
where rNormal(l
) (for the informative prior
models) or rNormal (0,0.001) (for uninforma-
tive models), Treatment Normal(0,0.001) where
represents the species in question (that is, dingo,
cat, fox, rabbit, small mammal or kangaroo)
The mean and 95% credible intervals were then
calculated for each effect size (r), which is the effect
size of the Before-After factor, in relation to
Treatment. The experiment was considered to have
a positive or negative effect on the effect size of the
response variables if the credible intervals did not
cross zero. We assessed convergence of the models
using by visual assessment of trace plots of each
estimated parameter (Hobbs and Hooten 2015).
We evaluated the models with and without
informative prior information by using the preci-
sion of the effect size (r). A reduction in the range
of the 95% credible interval of r indicates a higher
precision of r.
Strzelecki Desert Observation Site Time
Series Plots
In the Strzelecki Desert, dingo activity was consis-
tently higher ‘‘outside’’ the dingo fence in South
Australia than ‘‘inside’’ the dingo fence in New
South Wales (Figure 2A). The activity of kangaroos
and red foxes was consistently higher in New South
Wales where dingoes were rare (Figure 3A, C). Cat
activity was low at the beginning of the study and
did not differ markedly on each side of the dingo
fence (Figure 3E).
Rabbit activity was higher at sites in South
Australia where dingo activity was high on every
sampling occasion except April 2013 when activity
levels were similar on each side of the dingo fence
(Figure 3G). Similarly, small mammal activity was
higher at sites outside the dingo fence in South
Australia where dingoes were common on all
sampling occasions except June 2011 when activity
levels were similar on both sides of the dingo fence
following high rainfalls received in 2010 (Fig-
ure 3I).
Arid Recovery Experimental Site Time
Series Plots
Dingo spoor indicated that both the control and
dingo paddock contained similar levels of dingo
activity in the Period between the start of the study
and the introduction of dingoes to the experimen-
tal paddock in December 2008 (Figure 2B). Fol-
lowing the introduction of dingoes in December
2008, dingo activity was approximately 4 times
higher in the dingo paddock than the control
paddock until the end of the study in July 2012
(Figure 2B). The pair of dingoes introduced to the
experimental paddock bred during the study Peri-
od. The dingo population was estimated to be 10
individuals at the end of the experiment in July
K. E. Moseby and others
2012. Spoor counts suggested that dingoes existed
at low numbers in the control paddock throughout
the study.
The red kangaroo, Macropus rufus, was the only
species of kangaroo detected in the study areas via
remote game cameras or anecdotal observations. A
similar percentage of transects in the dingo and
control paddocks contained kangaroo spoor in the
three sampling sessions prior to the addition of
dingoes (Figure 3B). After the addition of dingoes,
kangaroo activity declined within the dingo pad-
dock but their activity increased over time in the
control paddock (Figure 3B). Towards the end of
the study in September and December 2012, red
kangaroo activity in both the experimental and
observational studies was more than 5 times higher
at sites where dingoes were controlled.
Fox activity within the dingo and control pad-
docks was similar at the start of the study (Fig-
ure 3B). Following the introduction of dingoes in
December 2008 and a significant rainfall event in
April 2010, fox activity increased markedly at the
control site but not the experimental site (Fig-
ure 3D). From July 2010 onwards, fox activity in
the dingo paddock was approximately half that of
the control paddock (Figure 3D). Similar trends in
cat activity were recorded in the dingo and control
paddock both prior to and after the addition of
dingoes to the dingo paddock (Figure 3F).
Rabbit activity was generally higher in the dingo
paddock than the control site both before and after
the introduction of dingoes (Figure 3H). Activity of
small mammals did not differ markedly between
the dingo paddock and control site before or after
the addition of dingoes, respectively (Figure 3J).
Comparison of Effect Sizes with Informed
and Uninformed Priors
The experimental data using both informed and
uninformed prior distributions showed an effect of
the dingo manipulation on dingo activity in com-
parison with the control site where dingoes were
rare (Figure 4). These results were consistent with
the observational prior information from the Strz-
elecki Desert (Figure 4). Also in line with our a
priori prediction, dingo addition had a strong neg-
ative effect on the uninformed effect size for kan-
garoo activity (Figure 4). Counter to our
predictions, the effect of dingo introduction on the
uninformed effect sizes for other variables was
ambiguous (Figure 4).
The uncertainty in the estimates of effect size in
the experimental data was considerably reduced by
using the data from the observational study from
the Strzelecki Desert as a prior. This reduction in
uncertainty is evidenced by the narrowing of the
95% credible intervals around the effect size for
each of the variables (Figure 4). Use of the in-
formed prior had a negligible effect on the posterior
distribution of the effect size for cats (Figure 4), and
there was no difference in cat activity in the dingo
experiment, whether or not an informative prior
was used (Figure 4).
The effect of dingoes did not have a clear effect
on abundance indices of rabbits, small mammals
and foxes when we used an uninformative prior
(Figure 4). When prior information was used,
dingo introduction as predicted had a positive effect
on small mammals and rabbits and a negative effect
on foxes (Figure 4). However, the prior informa-
Figure 2. Indices of dingo activity derived from track counts at Athe Strzelecki Desert observational site and Bthe
experimental site where dingoes were introduced. Closed symbols/unbroken lines denote sites where dingoes were
common/introduced. Open symbols/dashed lines denote sites where dingoes were rare. The arrow in Bindicates when
dingoes were introduced into the fenced paddock at the experimental site.
Ecological Role of an Apex Predator
Figure 3. Indices of activity derived from track counts at the Strzelecki Desert observational site and the experimental site
where dingoes were introduced for kangaroos, foxes, cats, rabbits and small mammals. Closed symbols/unbroken lines
denote sites where dingoes were common/introduced. Open symbols/dashed lines denote sites where dingoes were rare.
The arrows depicted on graphs for the experimental site indicate when dingoes were introduced into the fenced paddock.
K. E. Moseby and others
tion on the differences in fox, small mammal and
rabbit abundance indices on either side of the dingo
fence had a much larger effect on the posterior
distribution than the manipulative experiment.
This was because the experiment had a relatively
weak effect on these variables and the low level of
replication coupled with considerable between
times variation in population indices, particularly
during the before dingo addition Period, resulted in
the precision of the parameter estimates being low.
In the case of rabbits, the polarity of the unin-
formed effect size was inconsistent with posterior
distribution. This suggests that the prior informa-
tion for rabbits was incompatible with the unin-
formed estimates from the manipulative
experiment and had an overwhelming effect on the
posterior. In the case of small mammals, the
polarity of the posterior effect size was in the same
direction as the uninformed effect size; however,
the discrepancy between the effect sizes was
marked, indicating the prior information had more
of an effect on the posterior distribution than the
manipulative experiment. In the case of foxes, the
polarity of the uninformed and posterior estimates
was in the same direction, but the marked differ-
ence in the effect sizes suggests again that the prior
had a much stronger effect on the posterior distri-
bution than the manipulative experiment.
A constraint facing many large-scale manipulative
experiments, particularly those on the effects of
apex predators, is obtaining enough replication at
appropriate temporal and spatial scales to meet the
requirements of frequentist statistical analyses
(Oksanen 2001). Our study demonstrates that the
use of logical informative priors in a Bayesian
framework can help to overcome the issues of low
power in a large-scale unreplicated manipulative
experiment (McCarthy and Masters 2005). The
benefits of including logical priors to analyse our
dataset were evidenced by reduced uncertainty of
the posterior mean estimates of the effects of
experimentally introducing dingoes. The unin-
formed estimate of the effect of dingo introduction
revealed a negative effect on kangaroo activity, but
for the other variables the uninformed effect sizes
were ambiguous. Using data from the cross-dingo-
fence comparison as informative priors reduced
uncertainty in the posterior estimates to such an
extent, that the effect of dingo introduction became
credible for small mammals, rabbits and red foxes.
However, the prior had a strong influence on the
posterior mean effect sizes for small mammals,
rabbits and foxes. Below we discuss our results in
the context of previous literature on the ecological
effects of dingoes and the constraints on our study,
and in particular, the strong influence that the
prior information had on our results.
A caveat of our study is that we used track counts
to index mammal activity as opposed to population
estimates or indices derived from direct observation
or mark-recapture methods. A shortcoming of
using track-based indices of mammal activity as a
proxy of abundance is that they can be inflated by
making multiple observations of tracks left by the
same individual and are also influenced by the
characteristics of the substrate at each site and the
weather at time of sampling (Funston and others
2010; Stephens and others 2015). In addition, it
remains conceivable that differences in track in-
dices between areas with and without dingoes were
due to differences in habitat use of mammals rather
Figure 4. The mean effect of manipulating dingo abundance (the bars are 95% credible intervals) for each response
variable. The grey dots represent the prior information used to inform the experimental data. Prior information was
obtained by contrasting activity indices of the response variables across the dingo fence in the Strzelecki Desert. Black dots
and white dots represent the uninformed and informed estimates from the Roxby Downs sites comparing the effects of
experimentally manipulating dingo abundance at Roxby Downs on activity estimates of dingoes, small mammals, rabbits,
cats, foxes and kangaroos.
Ecological Role of an Apex Predator
than shifts in their abundance. Indeed, numerous
studies have shown that fear of predators can drive
shifts in the behaviour and habitat use of prey and
competitors (Ford and others 2014; Painter and
others 2015; Ordiz and others 2017). Thus, using
track-based indices to evaluate predators’ effects on
prey and competitors could be confounded by shifts
in habitat use. However, in our experimental study
area, telemetry of dingoes, foxes and cats following
the experimental reintroduction of dingoes pro-
vided no support for the hypothesis that foxes and
cats avoided dingoes in time or space (Schroeder
and others 2015). Furthermore, the idea that the
discrepancy in track counts of mammals on either
side of the dingo fence in our observational study
was due to behavioural shifts rather than differ-
ences in their abundance is not supported by sim-
ilarly large differences in abundance indices
derived from spotlight surveys (Letnic and Koch
2010) and in the case of small mammals, live
trapping (Letnic and others 2009). Nonetheless, we
recommend that where possible future studies
examining predators’ effects on ecosystems should
use methods that enable robust population esti-
mates of prey and competitors. If abundance in-
dices are used to evaluate predators’ effects, we
recommend that measures are taken to also deter-
mine the effects that predators have on habitat use
by prey and competitors.
A further caveat of our study is that the fenced
experimental exclosure may have influenced the
results by being too small to accommodate dingoes.
However, at 37 km
the exclosure was within
previous estimates, albeit at the lower end, of home
range size for desert dwelling dingoes (10–403 km
(Thomson 1992; Corbett 1995; Newsome and oth-
ers 2013), and thus provided sufficient space for a
dingo pack to exist as evidenced by the increase in
their numbers during the study. We contend also
that despite its caveats, the findings of our experi-
ment are significant because to date no experi-
mental reintroduction of dingoes has been
undertaken to investigate their ecological role de-
spite many calls to do so (Dickman and others
2009; Newsome and others 2015). No such exper-
iment has been undertaken because it is against
government policy to conduct a reintroduction of
dingoes outside of fenced exclosures in all Aus-
tralian jurisdictions.
A concern regarding the use of prior information
in Bayesian models is that priors can overwhelm
posterior estimates (Martin and others 2013). In
our study, the prior information on the differences
in fox, small mammal and rabbit abundance indices
on either side of the dingo fence had a much larger
effect on the posterior distribution than the
manipulative experiment. However, when consid-
ering the influence of the prior, it is important to
note that differences in the activities of kangaroos,
foxes, rabbits and small mammals across the dingo
fence at the Strzelecki Desert site were the result of
continuous culling of dingoes on the eastern side of
the dingo fence for over 100 years (Letnic and
others 2012). In contrast, the manipulation of
dingo abundance at the experimental site reported
in this study lasted only 3 years and hence would
be expected to have a smaller effect on mammal
assemblages than the long-term exclusion of din-
goes. Moreover, a short coming of our experi-
mental manipulation was that we only had three
sampling periods in the period prior to dingoes
being introduced. Thus, the precision for our
parameter estimates of the uninformed effect sizes
was poor because in our BACIP design, sampling
periods in the before and after dingo introduction
periods were the replicates. Nonetheless, we con-
tend that prior information on the responses of
mammals to manipulation of dingo abundance in
the Strzelecki Desert site was an appropriate choice
for the experimental site because of the similarities
in climate and ecological processes between the
two areas.
The informed posterior estimates of the effect of
dingo introduction were generally consistent with
our a priori predictions generated from trophic
cascade theory, the mesopredator release hypoth-
esis and previous studies (Wallach and others 2010;
Letnic and others 2012; Moseby and others 2012).
Dingoes’ effects on abundance indices of sympatric
mammals scaled with body size. Larger mammals,
kangaroos and red foxes were negatively affected
by dingoes due presumably to direct lethal and
non-lethal interactions with dingoes. Consistent
with the prediction that smaller mammals benefit
from indirect interactions with dingoes, rabbits and
small mammals responded positively to the intro-
duction of dingoes. Counter to our predictions,
feral cat activity was unaffected by dingoes. How-
ever, as discussed above and below, an alternative
interpretation of this trend in body size is that the
prior had an overwhelming influence on the pos-
terior estimates.
Our prediction that dingoes would have a sup-
pressive effect on kangaroo abundance was borne
out in both the experimental and observational
studies and was the strongest effect detected.
Kangaroos are frequently preyed upon by dingoes
(Morris and Letnic 2017), and previous correlative
studies have reported dingoes to have strong sup-
pressive effects on kangaroo abundance (Caughley
K. E. Moseby and others
and others 1980; Pople and others 2000). Taken
together, the results of this study and previous
studies indicate that suppressing dingo populations
will lead to a substantial increase in kangaroo
abundance. In turn, increased grazing pressure
associated with the irruptions of herbivores can
have cascading effects on other organisms (Ripple
and others 2014). Earlier studies have linked
irrupting kangaroo populations to a reduction in
the biomass of palatable plants (Letnic and others
2009) and negative effects on populations of
smaller animals due to competition for food re-
sources and or loss of shelter habitat (Howland and
others 2016). Increased abundances of kangaroos
where dingo populations have been suppressed
have implications for livestock production, because
livestock producers who control dingo populations
are likely to exacerbate competition for forage be-
tween kangaroos and livestock (Prowse and others
2015). Indeed, dingoes can improve pasture pro-
duction by regulating kangaroo populations (Mor-
ris and Letnic 2017) and thus potentially provide a
net gain for cattle producers (Prowse and others
2015) whose livestock are at relatively low risk of
being attacked by dingoes (Campbell and others
The experimental data suggested that dingoes
had an ambiguous effect on fox activity, although
the time series plots show that fox activity was
similar at the control and experimental sites until
approximately 1.5 years following the introduction
of dingoes. After this time, fox numbers increased
at the control site but showed a muted increase at
the experimental site where dingoes had been
introduced (Figure 3D). Incorporation of data from
the cross-dingo-fence comparison as an informa-
tive prior reduced uncertainty in the posterior
estimates for fox activity and suggested that foxes
responded negatively to the introduction of din-
goes. We assume that the increase in fox numbers
at the control site was due to increased breeding
and immigration in response to increases in the
abundances of rodents following heavy rainfalls in
April 2010 and that top-down regulation by din-
goes prevented a similar increase in fox activity
within the experimental site. The hypothesis that
dingoes suppressed fox activity within the experi-
mental exclosure is strengthened by the fact that
the fence surrounding the experimental site was
not a barrier that prevented foxes from moving into
the paddock. Indeed, foxes were able to gain access
to the paddock by climbing the fence and were
recorded to do so by infra-red game cameras (KM
personal observations). The marked difference in
fox activity between the experimental and control
sites following July 2010 is also consistent with the
results from previous studies examining numerical
relationships between indices of dingo and fox
abundance (Newsome and others 2001; Letnic and
others 2011; Colman and others 2014) and
telemetry observations conducted in our dingo
introduction area which showed that dingoes killed
foxes (Moseby and others 2012).
Counter to our predictions, cat activity indices
did not differ between areas with and without
dingoes (Figure 4). This contradicts concurrent
observations that dingoes killed 50% of teleme-
tered cats within the experimental exclosure (Mo-
seby and others 2012). Previous studies have
shown indices of cat abundance to display mixed
responses to the manipulations of dingo abundance
(Kennedy and others 2012; Letnic and others 2012;
Colman and others 2014). The wide ranging and
cryptic behaviour of cats and the likely ability of
surviving cats to quickly increase their home range
size or reinvade after the death of sympatric con-
specifics (Moseby and Hill 2011) suggest that spoor
counts may not fully reflect the impacts that din-
goes have on cats. Further studies are recom-
mended to better understand the interactions
between dingoes and cats and to investigate the
relationship between cat density and track-based
abundance indices.
Rabbits and small mammals displayed ambigu-
ous responses to the experimental introduction of
dingoes, but when prior information was consid-
ered, indices of small mammal and rabbit abun-
dances responded positively to the introduction of
dingoes. These findings are consistent with previ-
ous studies that have found small prey species to
exhibit higher population growth, high abundance
indices and be less wary in ecosystems where apex
predators are abundant (Palomares and others
1995; Crooks and Soule
´1999; Gordon and others
2015), but are likely to be due to the strong effect
that dingo exclusion had on small mammals and
rabbits in the mensurative study. However, in the
case of rabbits the opposite polarity of the unin-
formed and prior estimates of effect size suggests
that the prior information from the mensurative
study was incompatible with the manipulative
experiment. Given the ambiguous response of
rabbits and small mammals in our experimental
system and the strong influence of the prior, we
recommend that further studies are conducted to
evaluate apex predators’ direct and indirect effects
on these taxa with increased replication over large-
time scales.
Our study highlights the value of incorporating
prior information into large-scale studies, where
Ecological Role of an Apex Predator
logistical constraints mean that it is often impossi-
ble to design and conduct studies that employ tra-
ditional experimental designs. Although we do not
argue that prior information is a substitute for well-
replicated experimental studies, it is important to
acknowledge that many valid hypotheses which
require manipulation of factors such as pollutants,
fire, land-use and the abundance of large predators
often cannot be tested experimentally at multiple
locations because replication is impossible due to
geographical differences, expense or social and le-
gal considerations (Oksanen 2001). However, de-
spite our aim to experimentally evaluate the effects
of manipulating dingo abundance on other mam-
mals, the strong influence of the prior means that
our findings were driven largely by the mensura-
tive study. Thus, our study highlights an inferential
problem that can arise when applying prior infor-
mation derived from correlative studies to large-
scale unreplicated field experiments. Nonetheless,
we contend that making use of carefully selected
prior information within Bayesian frameworks,
such as we have done in this study, can enable
large-scale manipulations to be conducted, inter-
preted and subsequently provide managers with
the information they require to implement evi-
dence-based management strategies, with the ca-
veat that the influence of the prior is disclosed
when interpreting the results.
Funding was provided by the South Australian Arid
Lands NRM Board, BHP Billiton and Arid Recovery.
ML was supported by the Australian Research
Council. Part of this study was conducted at Arid
Recovery, a conservation partnership between BHP
Billiton, The South Australian Department for
Environment, University of Adelaide and local
community. We thank the landowners who pro-
vided access to the study area and many staff and
volunteers who assisted with fencing and data
collection. We thank two anonymous reviewers for
their constructive comments on the manuscript.
This study was conducted under ethics approval
(permit no. 6/2007-M3) from the South Australian
Wildlife Ethics Committee.
Campbell G, Coffey A, Miller H, Read JL, Brook A, Fleming PJS,
Bird P, Eldridge S, Allen BL. 2018. Dingo baiting did not re-
duce fetal/calf loss in beef cattle in northern South Australia.
Animal Prod Sci.
Caughley G, Grigg GC, Caughley J, Hill GJE. 1980. Does dingo
predation control the densities of kangaroos and emus? Aust
Wildlife Research 7:1–12.
Colman NJ, Gordon CE, Crowther MS, Letnic M. 2014. Lethal
control of an apex predator has unintended cascading effects
on forest mammal assemblages. Proc R Soc B Biol Sci
Corbett LK. 1995. Dingo in Australia and Asia. Sydney: UNSW
Crome FHJ, Thomas MR, Moore LA. 1996. A novel Bayesian
approach to assessing impacts of rain forest logging. Ecol Appl
Crooks KR, Soule
´ME. 1999. Mesopredator release and avifaunal
extinctions in a fragmented system. Nature 400:563–6.
Dickman CR, Glen AS, Letnic M. 2009. Reintroducing the dingo:
can Australia’s conservation wastelands be restored. Reintrod
Top-order Predat 7:238.
Eldridge SR, Shakeshaft BJ, Nano TJ. 2002. The impact of wild
dog control on cattle, native and introduced herbivores and
introduced predators in central Australia. Final report to the
Bureau of Rural Sciences. Alice Springs: Parks and Wildlife
Commission of the Northern Territory.
Estes JA, Terborgh J, Brashares JS, Power ME, Berger J, Bond
WJ, Carpenter SR, Essington TE, Holt RD, Jackson JBC,
Marquis RJ, Oksanen L, Oksanen T, Paine RT, Pikitch EK,
Ripple WJ, Sandin SA, Scheffer M, Schoener TW, Shurin JB,
Sinclair ARE, Soule
´ME, Virtanen R, Wardle DA. 2011.
Trophic downgrading of planet earth. Science 333:301–6.
Ford AT, Goheen JR, Otieno TO, Bidner L, Isbell LA, Palmer TM,
Ward D, Woodroffe R, Pringle RM. 2014. Large carnivores
make savanna tree communities less thorny. Science
Funston P, Frank L, Stephens T, Davidson Z, Loveridge A,
Macdonald D, Durant S, Packer C, Mosser A, Ferreira S. 2010.
Substrate and species constraints on the use of track inci-
dences to estimate African large carnivore abundance. J Zool
Gordon CE, Feit A, Gru
¨ber J, Letnic M. 2015. Mesopredator
suppression by an apex predator alleviates the risk of preda-
tion perceived by small prey. Proc R Soc Lond B Biol Sci.
Hobbs NT, Hooten MB. 2015. Bayesian models: a statistical pri-
mer for ecologists. Princeton (NJ): Princeton University Press.
Howland BW, Stojanovic D, Gordon IJ, Radford J, Manning AD,
Lindenmayer DB. 2016. Birds of a feather flock together:
Using trait-groups to understand the effect of macropod
grazing on birds in grassy habitats. Biol Cons 194:89–99.
Kauffman MJ, Brodie JF, Jules ES. 2013. Are wolves saving
Yellowstone’s aspen? A landscape-level test of a behaviorally
mediated trophic cascade: reply. Ecology 94:1425–31.
Kennedy M, Phillips BL, Legge S, Murphy SA, Faulkner RA.
2012. Do dingoes suppress the activity of feral cats in northern
Australia? Austral Ecol 37:134–9.
Letnic M, Greenville A, Denny E, Dickman CR, Tischler M,
Gordon C, Koch F. 2011. Does a top predator suppress the
abundance of an invasive mesopredator at a continental scale?
Glob Ecol Biogeogr 20:343–53.
Letnic M, Koch F. 2010. Are dingoes a trophic regulator in arid
Australia? A comparison of mammal communities on either
side of the dingo fence. Austral Ecol 35:167–75.
K. E. Moseby and others
Letnic M, Koch F, Gordon C, Crowther MS, Dickman CR. 2009.
Keystone effects of an alien top-predator stem extinctions of
native mammals. Proc R Soc B Biol Sci 276:3249–56.
Letnic M, Ritchie EG, Dickman CR. 2012. Top predators as
biodiversity regulators: the dingo Canis lupus dingo as a case
study. Biol Rev 87:390–413.
Martin TG, Arcese P, Kuhnert PM, Gaston AJ, Martin J-L. 2013.
Prior information reduces uncertainty about the conse-
quences of deer overabundance on forest birds. Biol Cons
McCarthy MA. 2007. Bayesian methods for ecology. Cambridge:
Cambridge University Press.
McCarthy MA, Masters P. 2005. Profiting from prior information
in Bayesian analyses of ecological data. J Appl Ecol 42:1012–
McCarthy MA, Parris KM. 2004. Clarifying the effect of toe clip-
ping on frogs with Bayesian statistics. J Appl Ecol 41:780–6.
Mech LD. 2012. Is science in danger of sanctifying the wolf? Biol
Cons 150:143–9.
Morris T, Letnic M. 2017. Removal of an apex predator initiates a
trophic cascade that extends from herbivores to vegetation
and the soil nutrient pool. Proc R Soc B 1854:20170111.
Morris WK, Vesk PA, McCarthy MA, Bunyavejchewin S, Baker
PJ. 2015. The neglected tool in the Bayesian ecologist’s shed: a
case study testing informative priors’ effect on model accu-
racy. Ecol Evol 5:102–8.
Moseby K, De Jong S, Munro N, Pieck A. 2005. Home range,
activity and habitat use of European rabbits (Oryctolagus
cuniculus) in arid Australia: implications for control. Wildl Res
Moseby K, Hill B. 2011. The use of poison baits to control feral
cats and red foxes in arid South Australia I. Aerial baiting
trials. Wildl Res 38:338–49.
Moseby KE, Neilly H, Read JL, Crisp HA. 2012. Interactions
between a top order predator and exotic mesopredators in the
Australian Rangelands. Int J Ecol 2012:250352. https://doi.
Moseby K, Stott J, Crisp H. 2009. Movement patterns of feral
predators in an arid environment–implications for control
through poison baiting. Wildl Res 36:422–35.
Newsome AE, Catling PC, Cooke BD, Smyth R. 2001. Two
ecological universes separated by the Dingo Barrier Fence in
semi-arid Australia: interactions between landscapes, her-
bivory and carnivory, with and without dingoes. Rangel J
Newsome TM, Ballard GA, Crowther MS, Dellinger JA, Fleming
PJ, Glen AS, Greenville AC, Johnson CN, Letnic M, Moseby
KE. 2015. Resolving the value of the dingo in ecological
restoration. Restor Ecol 23:201–8.
Newsome TM, Ballard GA, Dickman CR, Fleming PJ, van de Ven
R. 2013. Home range, activity and sociality of a top predator,
the dingo: a test of the Resource Dispersion Hypothesis.
Ecography 36:914–25.
Oksanen L. 2001. Logic of experiments in ecology: Is pseu-
doreplication a pseudoissue? Oikos 94:27–38.
Ordiz A, Bischof R, Swenson JE. 2013. Saving large carnivores,
but losing the apex predator? Biol Cons 168:128–33.
Ordiz A, Sæbø S, Kindberg J, Swenson J, Støen OG. 2017.
Seasonality and human disturbance alter brown bear activity
patterns: implications for circumpolar carnivore conservation?
Anim Conserv 20:51–60.
Painter LE, Beschta RL, Larsen EJ, Ripple WJ. 2015. Recovering
aspen follow changing elk dynamics in Yellowstone: Evidence
of a trophic cascade? Ecology 96:252–63.
Palomares F, Gaona P, Ferreras P, Delibes M. 1995. Positive ef-
fects on game species of top predators by controlling smaller
predator populations: an example with lynx, mongooses, and
rabbits. Conserv Biol 9:295–305.
Pasanen-Mortensen M, Pyyko
¨nen M, Elmhagen B. 2013. Where
lynx prevail, foxes will fail-limitation of a mesopredator in
Eurasia. Glob Ecol Biogeogr 22:868–77.
Pople AR, Grigg GC, Cairns SC, Beard LA, Alexander P. 2000.
Trends in the numbers of red kangaroos and emus on either
side of the South Australian dingo fence: Evidence for
predator regulation? Wildl Res 27:269–76.
Prowse TA, Johnson CN, Cassey P, Bradshaw CJ, Brook BW.
2015. Ecological and economic benefits to cattle rangelands of
restoring an apex predator. J Appl Ecol 52:455–66.
Read J, Eldridge S. 2010. An optimised rapid detection technique
for simultaneously monitoring activity of rabbits, cats, foxes
and dingoes in the rangelands. Rangel J 32:389–94.
Ripple WJ, Estes JA, Beschta RL, Wilmers CC, Ritchie EG,
Hebblewhite M, Berger J, Elmhagen B, Letnic M, Nelson MP,
Schmitz OJ, Smith DW, Wallach AD, Wirsing AJ. 2014. Status
and ecological effects of the World’s largest carnivores. Sci-
ence 343:1241484.
Ritchie EG, Johnson CN. 2009. Predator interactions, meso-
predator release and biodiversity conservation. Ecol Lett
Robley A, Gormley A, Forsyth DM, Wilton AN, Stephens D.
2010. Movements and habitat selection by wild dogs in east-
ern Victoria. Aust Mammal 32:23–32.
Schmitz OJ. 2008. Herbivory from individuals to ecosystems.
Annu Rev Ecol Evol Syst 39:133–52.
Schroeder T, Lewis M, Kilpatrick A, Moseby K. 2015. Dingo
interactions with exotic mesopredators: spatiotemporal
dynamics in an Australian arid-zone study. Wildl Res 42:529–
Spiegelhalter D, Thomas A, Best N, Lunn D. 2014. OpenBUGS
User Manual. Version 3.2.3.
Stephens PA, Pettorelli N, Barlow J, Whittingham MJ, Cadotte
MW. 2015. Management by proxy? The use of indices in
applied ecology. J Appl Ecol 52:1–6.
Stewart-Oaten A, Murdoch WW, Parker KR. 1986. Environ-
mental impact assessment: ‘‘Pseudoreplication’’ in time?
Ecology 67:929–40.
Thomson P. 1992. The behavioural ecology of dingoes in north-
western Australia. IV. Social and spatial organistaion, and
movements. Wildl Res 19:543–63.
Wallach AD, Johnson CN, Ritchie EG, O’Neill AJ. 2010. Predator
control promotes invasive dominated ecological states. Ecol
Lett 13:1008–18.
Ecological Role of an Apex Predator
... We collected barn owl pellets across a broader area than that which we surveyed predator and small mammal abundance encompassing two dune field areas that are divided by the dingo barrier fence, the Strzelecki Desert and the Moondiepitchnie Dunefield. We did this because of the difficulty in locating barn owl roosts and because the two areas have similar landforms and ecological communities (Moseby et al. 2018;Rees et al. 2019). Furthermore, previous studies show that the dingo barrier fence has a similar effect on the 1 3 abundances of dingoes, red foxes, feral cats and small mammals in both areas with dingoes and small mammals being less abundant and red foxes and feral cats more abundant where dingoes are functionally extinct (Moseby et al. 2018;Rees et al. 2019). ...
... We did this because of the difficulty in locating barn owl roosts and because the two areas have similar landforms and ecological communities (Moseby et al. 2018;Rees et al. 2019). Furthermore, previous studies show that the dingo barrier fence has a similar effect on the 1 3 abundances of dingoes, red foxes, feral cats and small mammals in both areas with dingoes and small mammals being less abundant and red foxes and feral cats more abundant where dingoes are functionally extinct (Moseby et al. 2018;Rees et al. 2019). We collected pellets (regurgitated pellets of fur, feathers and bone) from beneath barn owl roosts in abandoned man-made structures that were constructed or vacated post-construction of the DBF. ...
Full-text available
In ecosystems, some organisms facilitate others indirectly, by interacting with one or more common mediator organisms. Thus, the indirect effects of introducing or removing species can be resonant, sometimes leading to successional extinctions. The dingo (Canis dingo) is the apex predator in Australian deserts and was introduced to the continent between 3000 and 5000 years ago. Dingoes suppress the abundances of introduced mesopredators, the red fox (Vulpes vulpes) and feral cat (Felis catus) and in so doing mitigate small mammal declines wreaked by these mesopredators. Given the positive association between the abundances of dingoes and small mammals, we predicted that dingoes indirectly facilitate a specialised native predator of small mammals, the Barn owl, Tyto alba. We tested our prediction by monitoring the abundances of dingoes, foxes, cats, small mammals and barn owls and investigating barn owl diets in areas where dingoes were common versus areas where dingoes were functionally extinct on either side of the dingo barrier fence (DBF) in the Strzelecki Desert. Foxes and cats were less abundant in areas where dingoes were common. Conversely, small mammals and barn owls were more abundant where dingoes were common. Owls in areas where dingoes were common fed almost exclusively on small mammals, but owls in areas where dingoes were functionally extinct fed on greater proportion of birds and invertebrates. The findings of our study provide evidence that an introduced apex predator may indirectly facilitate a native predator and illustrates the myriad of far-reaching indirect effects that can result from apex predator suppression.
... ; Ripple and Beschta 2012;Moseby et al. 2018). Therefore, taking advantage of (a) Scatterplot of our field data showing the relationship between quantile regression lines fitted to the 0.95 (P < 0.001) and 0.9 (P = 0.043) quantiles. ...
Full-text available
The mesopredator release hypothesis predicts that abundance of smaller predators should increase in the absence of larger predators due to release from direct killing and competition. However, top predators’ effects on mesopredators are unlikely to operate in isolation but interact with other factors such as primary productivity of the landscape and human activities. We investigate factors influencing activity indices of a top predator (dingo) and an introduced mesopredator (red fox) in forests of south-eastern Australia. We used generalised linear models to investigate the effects that net primary productivity, proximity to freehold land and poison baiting campaigns directed at dingoes had on fox and dingo activity. Baiting was the best predictor of activity for both dingoes and foxes. Dingo activity was variable but typically lower at baited sites. Fox activity varied within a lower range at a majority of sites compared to the dingo but was typically higher at the baited sites. Positive responses of foxes to dingo control are consistent with the mesopredator release hypothesis and suggest in this region dingoes may have greater suppressive effect on fox populations than poisoning campaigns directed towards dingoes. Our results suggest that removal of dingoes may be counter-productive for biodiversity conservation because it may lead to higher activity of foxes.
... On each survey occasion (n = 33), we measured tracks of mammalian predators to estimate activity on either side of the dingo fence. At each of the four sites, 25 track plots in sand dunes (3 m × 1 m) were cleared daily and checked the following morning for tracks (Moseby et al. 2018;Rees et al. 2019a). We derived a passive activity index from the proportion of track plots that each species was detected. ...
Full-text available
Nomadism is an advantageous life history strategy for specialised predators because it enables the predator to respond rapidly to changes in prey populations. The letter-winged kite (Elanus scriptus) is a nomadic nocturnal bird of prey endemic to arid and semi-arid zones of Australia. Letter-winged kites prey almost exclusively on nocturnal rodents and are often associated with rodent irruptions, but little is known about the ecology of letter-winged kites inside their core range. The Strzelecki Desert contains a known dingo-mediated predation refuge for native rodents. In this manuscript, we compare kite sightings, predator activity, and small mammal populations across survey sites in the Strzelecki Desert where dingoes were common and where dingoes were rare and use publicly available data from the Atlas of Living Australia (ALA) to assess trends in the occurrence of kites in the region. Ninety-five percent of ALA observations occurred in areas where dingoes were common. Similarly, all our observations of kites occurred where dingoes were common and during an extended population irruption of Notomys fuscus. Notomys fuscus was the most frequent item in the letter-winged kite diet at our study sites. We suggest that there is significant evidence that these sites in the Strzelecki Desert form part of the core range for the letter-winged kite whose use of this area is facilitated by a predation refuge for rodents mediated by the dingo. We conclude that predation refuges mediated by dingoes could be a factor driving the distributions of letter-winged kites and other predators of rodents, particularly nomadic predators.
... Estimates of the impact of dingo predation support this idea and strongly suggest dingoes now function as a native predator in Australian ecosystems rather than as an alien predator. There have been no replicated experimental manipulations of dingo abundance to quantify their impacts on native prey (but see Moseby et al. 2019), and although some predator removal experiments using 1080 to control foxes Vulpes vulpes have also killed dingoes, these don't pinpoint dingo impacts. There have also been many mensurative studies used to estimate dingo impact, for example comparing prey abundance on either side of the dingo fence or in response to baiting regimes allowing estimates of effect size (e.g. ...
Full-text available
Distinguishing between whether a species is alien or native can be problematic, especially for introduced species that are long-established in new areas outside of their natural range. Transport by humans is the criterion for alien status used by many definitions, whereas arbitrary time since arrival to a location is often used to define native status. Here I propose an eco-evolutionary approach to distinguish between alien and native status and use this to resolve uncertainty in the status of the dingo in Australia. Dingoes were transported to mainland Australia by humans, but more than 4000 years ago, and dingoes now interbreed with feral domestic dogs. Legally, this mix of events has the dingo classified as native in some jurisdictions and alien in others. I suggest that native status for introduced species should be based on (1) whether the species has evolved in their new environment; (2) whether local species recognise and respond to them as they do towards deep endemic native species, and; (3) whether their impacts benchmark against those of a native species or are exaggerated like those of other alien species. Dingoes are behaviourally, reproductively and morphologically different to close ancestors from south-east Asia, and this difference has a genetic basis indicative of evolution in Australia. There is abundant evidence that native prey species on mainland Australia recognise and respond to them as a dangerous predator, which they are. But there is strong evidence that dingo impacts on prey are not exaggerated, with effect sizes from mensurative experiments similar to those of experiments on native predators rather than alien predators. These three lines of evidence suggest dingoes should be considered native to mainland Australia. I suggest this eco-evolutionary approach to defining native status can be helpful in resolving the often-heated debates about when an alien species becomes native.
... Although emus are birds, their body-size falls within the same range as many mammalian herbivores whose populations are regulated by dingo predation (Hunter et al., 2018;Letnic and Crowther, 2013;Letnic et al., 2009;Moseby et al., 2019). Thus, it is not altogether surprising that predation by dingoes may suppress emu abundance. ...
In desert ecosystems, some argue that primary productivity controls vertebrate populations while others contend that predators' top-down effects are under-appreciated. While seldom used to explain population dynamics in desert environments, the exploitation ecosystems hypothesis (EEH) unites bottom-up and top-down processes by conceptualizing how the biomass of plants and consumers vary along spatial productivity gradients in the presence and absence of predators. Here, we test predictions from the EEH along temporal primary productivity gradients, by comparing abundances of an apex predator (dingo) and a large omnivorous flightless bird (emu) on either side of Australia's dingo-proof fence over a nine year period. Where dingo populations were not suppressed by humans, their abundance increased with rainfall in the previous 12 months but did not respond to rainfall where their populations were suppressed. Conversely, emu abundance increased in response to antecedent rainfall where dingoes were rare but showed a negligible response to rainfall where dingoes were common. Our results accord with the EEH by suggesting that dingoes' top-down effects can decouple the linkage between emu populations and primary productivity. More generally, our findings show that top-down effects can have primacy over bottom-up effects for both an apex predator and an omnivorous prey species.
... reduce the confidence-interval width) associated with subsequent estimates. Prior information has been particularly effective in increasing precision in examining the responses of small marsupials to disturbance (Masters et al. 2003;McCarthy and Masters 2005), and mammal responses to dingo presence (Moseby et al. 2019). The inclusion of means and variations of koala numbers provided within the present study, including those from the smaller populations that were only distance sampled, can be treated as priors for future Bayesian estimates of koala numbers. ...
Full-text available
Context. Precise and accurate estimates of animal numbers are often essential for population and epidemiological models, as well as for guidance for population management and conservation. This is particularly true for threatened species in landscapes facing multiple threats. Estimates can be derived by different methods, but the question remains as to whether these estimates are comparable. Aims. We compared three methods to estimate population numbers, namely, distance sampling, mark–recapture analysis, and home-range overlap analysis, for a population of the iconic threatened species, the koala (Phascolarctos cinereus). This population occupies a heavily fragmented forest and woodland habitat on the Liverpool Plains, north-western New South Wales, Australia, on a mosaic of agricultural and mining lands. Key results. All three methods produced similar estimates, with overlapping confidence intervals. Distance sampling required less expertise and time and had less impact on animals, but also had less precision; however, future estimates using the method could be improved by increasing both the number and expertise of the observers. Conclusions. When less intrusive methods are preferred, or fewer specialised practitioners are available, we recommend distance sampling to obtain reliable estimates of koala numbers. Although its precision is lower with a low number of sightings, it does produce estimates of numbers similar to those from the other methods. However, combining multiple methods can be useful when other material (genetic, health and demographic) is also needed, or when decisions based on estimates are for high-profile threatened species requiring greater confidence. We recommend that all estimates of population numbers, and their precision or variation, be recorded and reported so that future studies can use them as prior information, increasing the precision of future surveys through Bayesian analyses.
... Dingoes suppress large herbivores (macropodids) and mesopredators (foxes and cats), and indirectly support growth of vegetation and small mammal numbers and diversity in arid (Brook et al. 2012, Gordon et al. 2017, Letnic and Crowther 2013, Letnic et al. 2009, Letnic and Koch 2010, Letnic et al. 2009), temperate mesic (Colman et al. 2015, Colman et al. 2014, Johnson and VanDerWal 2009) and tropical (Leo et al. 2019) environments (summarised in Figure 2). Although most of these studies are based on "mensurative" experiments in response to the dingo fence or baiting regimes, there is also recent manipulative experimental evidence of dingoes suppressing kangaroo and fox numbers while facilitating small mammals (Moseby et al. 2019). ...
Full-text available
The impact of hybridisation between dingoes and domestic dogs, and the subsequent introgression of domestic dog genes into dingo populations, remains a topic of significant impact. It has been claimed, but with little evidence or logical argumentation, that dingoes with significant dog introgression have different effects on agriculture and ecosystems than dingoes with no dog introgression. Introgression is a natural process in evolution, occurring in many species, although this is sometimes human assisted. Canid species in particular show high levels of introgression, due to their genetic and phylogenetic similarities, and human persecution creates scenarios encouraging hybridisation. Dingoes are no exception and demonstrate high levels of introgression of domestic dog genes, particularly in the temperate areas of south-eastern Australia. The available evidence shows that this introgression has minimal effects on the functional morphology of the dingo skull. There is also some preliminary evidence that introgression has not had a major impact on dingo reproductive biology. Studies on the impacts of dingoes on arid, tropical and temperate ecosystems, where levels of introgression vary greatly, all show consistent positive impacts of dingoes, regardless of the amount of domestic dog genes within the dingo population, on these ecosystems. Hence, hybridisation and resultant introgression from domestic dog genes appear to have little effect on aspects of the functional morphology or ecological role of the dingo. Accordingly, introgression does not diminish the conservation status of the dingo.
... In this regard, fenced exclosures achieve rewilding objectives at least for some functions driven by smaller species. Only the largest exclosures can achieve rewilding of devils and dingoes (Moseby et al. 2018), and these would need to be even larger to host self-sustaining populations or multiple groups of devils and dingos. ...
Full-text available
Rewilding is increasingly recognized as a conservation tool but is often context specific, which inhibits broad application. Rewilding in Australia seeks to enhance ecosystem function and promote self-sustaining ecosystems. An absence of large-bodied native herbivores means trophic rewilding in mainland Australia has focused on the restoration of functions provided by apex predators and small mammals. Because of the pervasive influence of introduced mesopredators, predator-proof fences, and establishment of populations on predator-free islands are common rewilding approaches. This sets Australian rewilding apart from most jurisdictions and provides globally relevant insights but presents challenges to restoring function to broader landscapes. Passive rewilding is of limited utility in arid zones. Although increasing habitat extent and quality in mesic coastal areas may work, it will likely be necessary to undertake active management. Because much of Australia's population is in urban areas, rewilding efforts must include urban areas to maximize effectiveness. Thus rewilding is not synonymous with wilderness and can occur over multiple scales. Rewilding efforts must recognize human effects on other species and benefit both nature and humans. Rewilding in Australia requires development of a shared vision and strategy and proof-of-concept projects to demonstrate the benefits. The repackaging of existing conservation activities as rewilding may confuse and undermine the success of rewilding programs and should be avoided. As elsewhere, rewilding in Australia should be viewed as an important conservation tool. © 2019 Society for Conservation Biology.
Full-text available
The impacts of invasive predators can be amplified by high densities of invasive prey species. In Australia, hyper abundant rabbit populations lead to high densities of feral cats and correspondingly high impact of cats on native species, especially small mammals. Therefore, it would be expected reducing rabbits could also reduce abundance of cats, and thereby alleviate predation on native small mammals. However, cats might respond to the loss of rabbits by prey-switching to native small mammals, resulting in increased predation on those species. Our aim was to understand the short-term effects of an experimental reduction of rabbit abundance on feral cats and their small-mammal prey in arid South Australia. We reduced the rabbit population in a 37 km² experimental enclosure by ~ 80%, while monitoring an adjacent unmanipulated area as a control. Cat activity and survival of VHF-collared cats in the enclosure decreased by 40% following the rabbit reduction. Surviving cats increased their consumption of reptiles, birds and invertebrates, but they nevertheless evinced hunger by increased intake of experimentally-supplied sausages. There was no change in either the proportion of cat scats that contained remains of small mammals, or the rate at which video-collared cats were recorded killing small mammals, even though the activity of small mammals declined. Our results demonstrate that individual feral cats prey-switch in response to removal of their primary prey. However, we also show that survival and overall activity of cats decreased, which could result in net, long-term benefits for native prey threatened by cats. Management of feral cats using food lures or baits would also be more effective when introduced prey are scarce, as cats are more likely to eat novel food.
Full-text available
It is widely assumed that organisms at low trophic levels, particularly microbes and plants, are essential to basic services in ecosystems, such as nutrient cycling. In theory, apex predators’ effects on ecosystems could extend to nutrient cycling and the soil nutrient pool by influencing the intensity and spatial organization of herbivory. Here, we take advantage of a long-term manipulation of dingo abundance across Australia’s dingo-proof fence in the Strzelecki Desert to investigate the effects that removal of an apex predator has on herbivore abundance, vegetation and the soil nutrient pool. Results showed that kangaroos were more abundant where dingoes were rare, and effects of kangaroo exclusion on vegetation, and total carbon, total nitrogen and available phosphorus in the soil were marked where dingoes were rare, but negligible where dingoes were common. By showing that a trophic cascade resulting from an apex predator’s lethal effects on herbivores extends to the soil nutrient pool, we demonstrate a hitherto unappreciated pathway via which predators can influence nutrient dynamics. A key implication of our study is the vast spatial scale across which apex predators’ effects on herbivore populations operate and, in turn, effects on the soil nutrient pool and ecosystem productivity could become manifest. © 2017 The Author(s) Published by the Royal Society. All rights reserved.
Full-text available
There is global interest in restoring populations of apex predators, both to conserve them and to harness their ecological services. In Australia, reintroduction of dingoes (Canis dingo) has been proposed to help restore degraded rangelands. This proposal is based on theories and the results of studies suggesting that dingoes can suppress populations of prey (especially medium- and large-sized herbivores) and invasive predators such as red foxes (Vulpes vulpes) and feral cats (Felis catus) that prey on threatened native species. However, the idea of dingo reintroduction has met opposition, especially from scientists who query the dingo’s positive effects for some species or in some environments. Here, we ask ‘what is a feasible experimental design for assessing the role of dingoes in ecological restoration?’We outline and propose a dingo reintroduction experiment—one that draws upon the existing dingo-proof fence—and identify an area suitable for this (Sturt National Park, western New South Wales). Although challenging, this initiative would test whether dingoes can help restore Australia’s rangeland biodiversity, and potentially provide proof-of-concept for apex predator reintroductions globally.
Full-text available
Predators can impact their prey via consumptive effects that occur through direct killing, and via non-consumptive effects that arise when the behaviour and phenotypes of prey shift in response to the risk of predation. Although predators' consumptive effects can have cascading population-level effects on species at lower trophic levels there is less evidence that predators' non-consumptive effects propagate through ecosystems. Here we provide evidence that suppression of abundance and activity of a mesopredator (the feral cat) by an apex predator (the dingo) has positive effects on both abundance and foraging efficiency of a desert rodent. Then by manipulating predators' access to food patches we further the idea that apex predators provide small prey with refuge from predation by showing that rodents increased their habitat breadth and use of 'risky' food patches where an apex predator was common but mesopredators rare. Our study suggests that apex predators' suppressive effects on mesopredators extend to alleviate both mesopredators' consumptive and non-consumptive effects on prey. © 2015 The Author(s) Published by the Royal Society. All rights reserved.
Full-text available
Despite benefits for precision, ecologists rarely use informative priors. One reason that ecologists may prefer vague priors is the perception that informative priors reduce accuracy. To date, no ecological study has empirically evaluated data-derived informative priors' effects on precision and accuracy. To determine the impacts of priors, we evaluated mortality models for tree species using data from a forest dynamics plot in Thailand. Half the models used vague priors, and the remaining half had informative priors. We found precision was greater when using informative priors, but effects on accuracy were more variable. In some cases, prior information improved accuracy, while in others, it was reduced. On average, models with informative priors were no more or less accurate than models without. Our analyses provide a detailed case study on the simultaneous effect of prior information on precision and accuracy and demonstrate that when priors are specified appropriately, they lead to greater precision without systematically reducing model accuracy.
Beef cattle production is the major agricultural pursuit in the arid rangelands of Australia. Dingo predation is often considered a significant threat to production in rangeland beef herds, but there is a need for improved understanding of the effects of dingo baiting on reproductive wastage. We experimentally compared fetal/calf loss on baited and non-baited treatment areas within three northern South Australian beef herds over a 2-4-year period. At re-musters, lactation was used to determine the outcomes of known pregnancies. Potential explanatory factors for fetal/calf loss (dingo baiting, dingo activity, summer heat, cow age, seasonal conditions, activity of dingo prey and selected livestock diseases) were investigated. From 3145 tracked pregnancies, fetal/calf loss averaged 18.6%, with no overall significant effect of baiting. Fetal/calf loss averaged 27.3% for primiparous (first pregnancy) heifers and 16.8% for multiparous (2nd or later calf) cows. On average, dingo-activity indices were 59.3% lower in baited treatments than in controls, although background site differences in habitat, weather and previous dingo control could have contributed to these lower indices. The overall scale and timing of fetal/calf loss was not correlated with dingo activity, time of year, a satellite-derived measure of landscape greenness (normalised difference vegetation index), or activity of alternative dingo prey. Limited blood testing suggested that successful pregnancy outcomes, especially in primiparous heifers, may have been reduced by the livestock diseases pestivirus and leptospirosis. The percentage occurrence of cattle hair in dingo scats was higher when seasonal conditions were poorer and alternative prey less common, but lack of association between fetal/calf loss and normalised difference vegetation index suggests that carrion feeding, rather than calf predation, was the more likely cause. Nevertheless, during the fair to excellent prevailing seasons, there were direct observations of calf predation. It is likely that ground baiting, as applied, was ineffective in protecting calves, or that site effects, variable cow age and disease confounded our results.
Wildlife may adapt activity patterns to daily and seasonal variations in environmental factors and human activity. At the daily scale, diurnal or nocturnal activity can be a response to variations in food availability and/or human avoidance. At the seasonal scale, variation in prey vulnerability underlies the influence of predators on prey population dynamics, which is of management concern when predation affects domestic species. We analyzed the movement patterns of 133 GPS-collared brown bears in three study areas in Sweden in spring, when bears prey on the calves of domestic reindeer and moose, and in summer-early fall, when bears rely mostly on berries, in three areas with a gradient of human disturbance. In spring, the bears' daily movement patterns and time of predation on ungulates overlapped. In summer-early fall, when bears are hyperphagic to store fat for hibernation and reproduction, variation in the degree of nocturnal behavior among study areas likely reflected behavioral adjustments to reduce the risk of encountering people. Flexibility in daily movement patterns by large carnivores may help them survive in human-dominated landscapes, but behavioral changes may also reflect environmental degradation, for example human disturbance influencing foraging opportunities. Diurnal human activity disturbs the carnivores, but that does not hinder depredation on reindeer, because it occurs mostly at night. Thus, ideally carnivores and reindeer should be separated spatially to reduce depredations. A zoning system prioritizing carnivore conservation and reindeer herding in different areas might help reduce a long-lasting conflict.
Bayesian modeling has become an indispensable tool for ecological research because it is uniquely suited to deal with complexity in a statistically coherent way. This textbook provides a comprehensive and accessible introduction to the latest Bayesian methods-in language ecologists can understand. Unlike other books on the subject, this one emphasizes the principles behind the computations, giving ecologists a big-picture understanding of how to implement this powerful statistical approach. Bayesian Models is an essential primer for non-statisticians. It begins with a definition of probability and develops a step-by-step sequence of connected ideas, including basic distribution theory, network diagrams, hierarchical models, Markov chain Monte Carlo, and inference from single and multiple models. This unique book places less emphasis on computer coding, favoring instead a concise presentation of the mathematical statistics needed to understand how and why Bayesian analysis works. It also explains how to write out properly formulated hierarchical Bayesian models and use them in computing, research papers, and proposals. This primer enables ecologists to understand the statistical principles behind Bayesian modeling and apply them to research, teaching, policy, and management. Presents the mathematical and statistical foundations of Bayesian modeling in language accessible to non-statisticians. Covers basic distribution theory, network diagrams, hierarchical models, Markov chain Monte Carlo, and more. Deemphasizes computer coding in favor of basic principles. Explains how to write out properly factored statistical expressions representing Bayesian models.
Context Apex predators occupy the top level of the trophic cascade and often perform regulatory functions in many ecosystems. Their removal has been shown to increase herbivore and mesopredator populations, and ultimately reduce species diversity. In Australia, it has been proposed that the apex predator, the dingo (Canis dingo), has the potential to act as a biological control agent for two introduced mesopredators, the red fox (Vulpes vulpes) and the feral cat (Felis catus). Understanding the mechanisms of interaction among the three species may assist in determining the effectiveness of the dingo as a control agent and the potential benefits to lower-order species. Aims To test the hypotheses that feral cats and foxes attempt to both temporally avoid dingoes and spatially avoid areas of high dingo use. Methods Static and dynamic interaction methodologies based on global positioning system (GPS) telemetry data were applied to test temporal and spatial interactions between the two mesopredators (n≤15) and a dingo pair (n≤2). The experimental behavioural study was conducted in a 37-km2 fenced enclosure located in arid South Australia. Key results The dynamic interaction analysis detected neither attraction nor avoidance between dingoes and cats or foxes at short temporal scales. There was no suggestion of delayed interactions, indicating that dingoes were not actively hunting mesopredators on the basis of olfactory signalling. However, static interaction analysis suggested that, although broad home ranges of cats and foxes overlapped with dingoes, core home ranges were mutually exclusive. This was despite similar habitat preferences among species. Conclusions We found that avoidance patterns were not apparent when testing interactions at short temporal intervals, but were manifested at larger spatial scales. Results support previous work that suggested that dingoes kill mesopredators opportunistically rather than through active hunting. Implications Core home ranges of dingoes may provide refuge areas for small mammals and reptiles, and ultimately benefit threatened prey species by creating mesopredator-free space. However, the potential high temporal variation in core home-range positioning and small size of mutually exclusive areas suggested that further work is required to determine whether these areas provide meaningful sanctuaries for threatened prey. Journal compilation