Estuarine crocodiles in a tropical coastal floodplain obtain nutrition from terrestrial prey
Estuarine crocodiles in a tropical coastal floodplain obtain nutrition from terrestrial prey
Maria Fernanda Adame 0 1 2
Timothy D. Jardine 0 2
Brian Fry 0 1 2
Dominic Valdez 0 1 2
Garry Lindner 0 2
Jonathan Nadji 0 2
Stuart E. Bunn 0 1 2
0 Australian Government- National Environmental Research Program , SE Bunn
1 Australian Rivers Institute, Griffith University , Nathan, QLD , Australia , 2 School of Environment and Sustainability, University of Saskatchewan , Saskatoon, SK , Canada , 3 Kakadu National Park, Jabiru, NT , Australia
2 Editor: Hamish Campbell, Charles Darwin University , AUSTRALIA
The estuarine crocodile (Crocodylus porosus) is one of the largest and most widespread crocodilians in the world. Although considered an apex species, the role of the estuarine crocodile in aquatic foodwebs is poorly understood; we know what crocodiles ingest, but not what nourishes them. In this study, we used a combination of stable isotope measurements (δ13C, δ15N, and δ34S) and direct feeding observations to identify the source of nutrition of estuarine crocodiles in Kakadu National Park, Northern Australia. Our results show that most crocodiles sampled (size 850 ± 4200mm, with 76% of them being > 2.5 m) consume a large variety of prey, however a large proportion of their nutrition is derived from terrestrial prey. Introduced species such as water buffaloes (Bubalus bubalis) and pigs (Sus scrofa) could contribute between 53 and 84% to the nutrition of the sampled crocodiles. The isotopic composition of large crocodiles (total length > 3 m) suggested possible increase in marine prey consumption with size (R2 = 0.30; p = 0.005). Additionally, we found crocodiles sampled in the dry season had on average higher terrestrial contributions compared to crocodiles sampled during the wet season (84.1 ± 2.4% versus 55.4 ± 7.0%). Overall, we found that terrestrial prey are important source of nutrition for many crocodiles in this region where introduced herbivorous mammals are abundant.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
The estuarine crocodile is an iconic and widespread predator in the tropics, occurring from
southern India to Northern Australia [
]. The estuarine crocodile is one of the largest living
crocodilians, reaching 7m in length. It has a reputation of being aggressive, feeding on a wide
variety of prey that can include humans [1±3]. The diet of the estuarine crocodile is highly
variable among populations and organisms, and commonly includes ontogenetic shifts [4±6]. In
Australia, juveniles and adult crocodiles feed on low trophic levels, while medium-sized
individuals feed on higher trophic levels [
]. Stomach contents of estuarine crocodiles in Australia
analysis, decision to publish, or preparation of the
and Malaysia show that juveniles feed on crabs, prawns, insects, and small fish, while sub
adults (120±180 cm total length, TL) feed on a mixture of crustaceans, mammals, and birds
[6,8±10]. The diet of adults (TL > 180 cm) has been mainly inferred from direct observations
and includes a wide range of riverine, terrestrial, and marine organisms. The data available
show that estuarine crocodiles ingest a large variety of foods, however, which prey are the most
important resource for their nutrition is still unknown.
Contrary to direct observations of feeding activity or stomach contents, stable isotopes help
discern assimilated prey that contributes to nutrition rather than prey that is simply ingested
]. For example, because plant material has been observed in the stomachs of crocodiles, we
could falsely infer that crocodiles feed on plants, an unlikely source of nutrition for a
carnivorous predator [
]. Analyzing the gut contents of large, aggressive crocodiles is difficult because
animals are hard to capture and subdue. Thus, many studies have concentrated on studying
smaller individuals [
]. It is important to include adult individuals in studies of crocodile
diets, because they comprise the majority of organisms in healthy populations . The
combination of direct dietary observations with ecological tracers can help identify the diet and
nutrition sources of predators that are difficult to sample.
The diet of a top predator has implications at the ecosystem level. For example, adult
Alligator mississippiensis inhabit freshwaters, but consume substantial amounts of marine prey; thus,
they play an important role in the exchange of nutrients and carbon between freshwater and
marine ecosystems [
]. Another species, the Australian freshwater crocodile, C. johnstoni,
has a diet consisting mostly of aquatic prey, such as fish, and crustaceans [
C. johnstoni could be important in regulating the local freshwater food web. The diet of
predators has cascading effects onto ecosystem functions, such as in nutrient cycling and carbon
sequestration [16±18]. Understanding the diet of crocodiles can provide information on their
role in nutrient transport, top-down regulation of prey populations, and terrestrial-marine
The role of crocodiles in aquatic and terrestrial ecosystems can be disrupted (e.g. [
example, in northern Australia, introduced terrestrial species are abundant, with a population
of water buffalo (Bubalus bubalis) currently numbering more than 150,000; and with feral pigs
(Sus scrofa) reaching 10±20 individuals per square kilometer [
]. As a result, prey biomass of
introduced mammals is large and vastly exceeds that of native mammals in the region .
Currently, the population of estuarine crocodiles in Northern Australia is growing [
hunting was banned in the 1970s, crocodile numbers have recovered from an estimated 3,000
in 1984 to 70,000 in 1998 [
]. It is unknown whether these two factors, abundant terrestrial
prey and a growing population, are somehow related.
In this study, we use a combination of stable isotope values (δ13C, δ15N, and δ34S) of
crocodiles and their potential prey to identify the diet and source of nutrition of a large and growing
population of estuarine crocodiles in Northern Australia. We included a large number of
individual crocodiles (n = 45) ranging from juveniles of 85 cm TL to adults of 4.2 m TL. We
sampled potential prey from the rivers (barramundi, Lates calcarifer, and mullet, Liza ordensis),
land (wallaby, Macropus agilis; buffalo, B. bubalis; and feral pigs, S. scrofa), and ocean (giant
sea catfish, Netuma thalassina). We also included direct field observations of feeding behavior
and gut contents. Our study sites are within three rivers in Northern Australia (Fig 1) which
maintain intact hydrology and have large and highly productive seasonal floodplains [
The main goals of the study were: first, to identify the sources of nutrition of the estuarine
crocodile population in Kakadu National Park, Northern Australia, and second, to investigate
how introduced and abundant terrestrial prey may have affected the ecological role of this
iconic predator in the region.
2 / 16
Fig 1. Sampling sites within rivers Kakadu National Park, Northern Territory, Australia. Crocodiles were sampled
in three rivers and their respective floodplain areas (grey): Wildman, South and East Alligator Rivers.
Materials and methods
All handling of animals was conducted under Griffith University animal's ethics protocol
approved by the Animal Ethics Committee (ENV/08/11/AEC) and Kakadu National Park
permit guidelines (RK 786). Sampling was conducted in Kakadu National Park in northern
Australia. Samples were collected from the Wildman River (WR), South Alligator River (SAR) and
East Alligator River (EAR) all of which have undisturbed flooding regimes (Fig 1). The rivers
are bounded by the Van Diemen Gulf to the north and the Kakadu escarpment to the south.
The region is a tropical savannah characterized by a monsoonal climate with two distinct
seasons, wet and dry. The wet season usually starts between October and December, typically
lasting three to four months. During this period, the region receives 1,300±1,500 mm of rain [
which results in the inundation of the floodplains (Fig 1)
Sampling was conducted from 2012 to 2014. Crocodiles were opportunistically sampled in the
SAR, EAR, and the WR. Most samples were obtained from the crocodile management tagging
program that surveys the crocodile population and distribution throughout Kakadu National
Park. These crocodiles were trapped, measured and released for monitoring purposes. Most of
the individuals sampled through this program (44%) are around 2.5 m, which is the most
common crocodile size in the park (44%, G. Lindner pers. comm). Other samples were obtained
from crocodiles that were relocated due to their overt interest towards people or boats, or were
identified as nuisance animals that needed to be removed from the local community or tourist
locations. A few samples were obtained from traps in the upper reaches of rivers, from dead
crocodiles that were killed by other aggressive crocodiles, or by vehicles on the road.
Each sample consisted of a piece of scute taken from the mid-section of the tail (S1 Fig).
Scutes are made of a combination of keratin and collagen [
]. Our samples consisted
3 / 16
predominantly of keratin. Similar to other studies [
], we found no significant difference
between the isotope value of keratin and collagen (D. Valdez, unpublished data). When
possible, the total length (TL) of the crocodile was measured (n = 25) and sex was determined by
probing the cloaca with a finger. We sampled 45 crocodiles that ranged from 85 cm to 4.2 m
TL, from which 14 male and 5 female individuals were identified (S1 Table). Crocodiles were
sampled during the dry (n = 20) and the wet season (n = 25). We sampled 31 crocodiles from
the EA, 9 from the SA, and 2 from the WR. We also analyzed samples of 3 individuals from
unknown locations within the National Park (S1 Table).
We classified potential prey as terrestrial, riverine, riverine-marine, or marine. Terrestrial
prey included feral pigs (S. scrofa, n = 58), wallaby (Macropus agilis, n = 35), and buffalo (B.
bubalis, n = 8); riverine prey were represented by mullet (L. ordensis, n = 258); riverine-marine
by barramundi (Lates calcarifer, n = 385); and marine prey by the giant catfish (Netuma
thalassina, n = 5). Muscle tissue from terrestrial animals was opportunistically collected from road
kill and from individuals killed as part of the management program of feral species of Kakadu
National Park. Buffalo tissue samples were also collected from a nearby farm. Riverine and
marine fish were collected using multiple methods including hand line, rod and reel, cast net,
10 and 16-cm gill nets that were 20m in length, and backpack and boat electro-fishers
(SmithRoot, Inc. Vancouver, WA, and U.S.A.). For large fish, caudal fin tissue was sampled
nonlethally for stable isotope analyses because fin tissue is a reliable surrogate for muscle tissue
]. Other potential prey such as birds, reptiles, and sharks were opportunistically collected
Tissue samples were rinsed and transported frozen. Samples were oven-dried at 60ÊC for 24
h before being ground to a fine powder and homogenised with a ball and mill grinder.
Analyses for δ13C, δ15N, and δ34S were conducted with an elemental-analyser isotope ratio mass
spectrometer system (EA-IRMS, Sercon System, Griffith University). Analytical errors, based
on standard deviations of in-house standards, were < 0.1½ for δ13C, < 0.2 ½ for δ15N,
and < 0.5½ for δ34S.
To estimate the contribution of prey to the crocodile's diet, we corrected the isotope values
of the crocodile for isotopic discrimination that occurs during prey consumption. We
subtracted 1.4 ½ from δ13C and 3.0 ½ from δ15N for each individual based on isotopic
discrimination (Δ) of keratin in captive estuarine crocodiles in Northern Australia [
]. Because the
trophic correction for wild crocodiles is uncertain, we assessed the effect of using a range of
isotopic discrimination factors, including those estimated for other species of crocodilians
(American alligator, Alligator mississippiensis and broad-snouted caiman, Caiman latirostris),
which have been reported to be lower (Δ = 0.6 and 0.9½ for δ13C, and 1.2 and 0.8½ for δ15N,
]). We did not apply a correction to δ34S data because trophic isotopic
discrimination of δ34S is minimal (Δ = 0.5½, ).
Isotope turnover in crocodilians is likely to be slow, for example, in captive American
alligators (Alligator mississippiensis), δ13C has a half-live of 142 days and δ15N of 277 days [
Thus, we expected difficulties in assessing changes in diet at seasonal time scales. However,
wild estuarine crocodiles have food conversion rates twice as high as captive ones [6,29],
which could translate into faster isotopic turnover rates. To test whether we could identify
seasonal differences in the diet, we compared isotope values of crocodiles sampled during the wet
versus crocodiles sampled during the dry season.
We compiled observations of feeding activities from records held by National Park rangers of
Kakadu National Park. Direct feeding of crocodiles is likely to be biased towards crocodiles
4 / 16
hunting terrestrial prey in the water-edge. Gut contents were also recorded from individual
crocodiles that were killed for management purposes.
We tested for differences among isotope values with Analyses of Variance (ANOVA), where
isotope value (δ13C, δ15N, and δ34S) was the dependent variable, and sex, catchment, and
season were the independent and/or random factors of the model. We also tested for difference
among isotopic values of prey (terrestrial, riverine, riverine-marine, and marine); when
differences were significant, Bonferroni post-hoc tests were conducted. Linear regressions were
conducted to assess the relation between isotope value and crocodile size, and between food
source contributions and crocodile size. Normality was assessed with probability plots and
Shapiro-Wilk tests. When the variable was not normally distributed (e.g. % food contribution),
it was transformed (log10). The statistical tests were performed with SPSS Statistics (v21, IBM,
New York, USA). Values reported are means and standard errors, unless specified.
The relative contribution of each prey group to the nutrition of the crocodiles was first
assessed by plotting the mean values of crocodiles versus possible food sources. Isotope mixing
models for interpreting feeding relations in animals are generally based on the assumption
that important food sources are included and well-characterized, and that animals have isotope
values that are intermediate between source values of included prey [
]. However, some
individual crocodiles were not clearly intermediate between potential prey sources, so that some
sources may have been missed in spite of multi-year sampling. Also source mixing implies that
animals have some access to all sources, but for crocodiles that can vary widely in their
residency and mobility patterns, access to all sources did not seem assured [
]. The possibility of
missing sources and lack of access to other sources led us to use two different but related
isotope modelling approaches for interpreting crocodile diets.
The first approach used mean isotope values and proximity of crocodile and source values
to assess similarity or affinity. This ªproximityº technique was developed in the 1990s
(Reviewed in [
]) and gives results generally similar to mixing models. The technique differs
from mixing models in that it does not require that animals are intermediate between source
values, and gives a somewhat more qualitative estimate of similarity or affinity between
samples and potential sources, somewhat like cluster analysis results. This looser similarity
assessment is more consistent with our sampling strategy and limited knowledge of crocodile
behaviour. This approach can give biased results if a nearby source is split into several similar
groups, with a result that the new cluster of sources acquires an inflated importance.
Recognizing this problem, we were careful to not split sources into too many categories in our
proximity analyses. In this proximity approach, distance from the crocodile mean to each source was
calculated in 3D isotope space using Pythagorean distances:
d34Scroc d34Ssourcei Eq 1
where di is the Euclidean distance from the crocodile value to the value of the ith source. The
distances to all sources were inverted and summed, then the inverted distances divided by this
sum and multiplied by 100 to give percent estimates of each source value. Results from this
inverse distance or proximity approach were used to understand population-level mean trends
in crocodile feeding. We performed two separate analyses, one for the EAR and one for the
SAR. Due to low number of samples for the WR (n = 2), the crocodiles from this river were
not included in this analysis.
The second approach was to focus on individuals where sources were aggregated into four
categories, with a standardized Bayesian mixing model (SIAR Solo) applied to each animal
5 / 16
]. We did this because crocodiles have a wide variety of diets and tend to be individualistic
in their diet choices [
]. Unknown and missed sources are inherently aggregated into four
groups: terrestrial, riverine, riverine-marine, and marine. The application of SIAR Solo gave
standardized profiles of resource use that often differed among individual crocodiles. SIAR
and other Bayesian mixing models commonly have a bias towards ªall sources are equalº or
generalist solutions [34±36], so we used the SIAR Solo approach as a conservative way to find
instances where individual differences were still apparent for crocodiles. To assess the effect of
seasonality, we separated crocodiles sampled during the dry season from those sampled during
Overall, the first affinity approach was applied to means for a generalized overview of
average crocodile feeding at the whole landscape level. The SIAR Solo mixing model approach
probed for variations in resource use by individuals, with this Bayesian approach expected to
identify the more robust feeding differences among individuals. We also applied the first
proximity approach to individuals, finding that overall it gave similar indications of differences
among individuals evident in the SIAR Solo approach. This congruence of results from the
proximity and Bayesian approaches has also been previously observed [
Direct observations of crocodile feeding activities confirmed that crocodiles consumed a wide
variety of organisms (Table 1). Crocodiles in the study area were observed to feed on cattle,
pigs, water buffalo, as well as flying fox (Pteropus sp) and snakes (brown snake, Pseudonaja
textilis, and pythons). They are also known to feed on dogs and humans, with four fatal attacks on
humans in the region since 1978. Crocodiles fed on birds including magpie goose (Anseranas
semipalmata), heron (Ardea sp), cormorants (Phalacrocorax sp), egrets (Egretta sp), ibis
(Threskiornis moluccus), and spoonbills (Platalea sp). Riverine prey included mullet and barramundi.
Finally, crocodiles were observed to consume marine prey, including prawns, octopus, flatback
turtle (Natator depressus) and turtle hatchlings. Stomach contents of crocodiles and specific
sightings from this study are shown in Table 1.
Crocodiles had a wide range of isotopic values (Table 2). The δ13C and δ15N values of
crocodile were similar between catchments, season and sex of the crocodiles (catchment season sex
F1, 31 = 0.93, p = 0.34; F1, 31 = 1.50, p = 0.23; and F1, 31 = 0.74, p = 0.79, respectively). However,
there were significant differences in δ34S values between sexes in different seasons (sex season
F2, 2 = 69.2, p = 0.028). Finally, δ13C and δ15N, but not δ34S, significantly increased with body
size (R2 = 0.22, p = 0.01; R2 = 0.35, p < 0.01; R2 = 0.05 p = 0.08, respectively; Fig 2), similar to
data of C. porosus in other regions of Australia [
] (Fig 2)
The isotopic composition of the prey was significantly different for terrestrial, riverine,
riverine-marine and marine animals for δ13C (F198, 3 = 55.77, p <0.001, riverine and river-marine
different from marine and terrestrial), δ15N (F198, 3 = 70.85, p <0.001, all different), and δ34S
(marine different from the rest, F198, 3 = 4.59, p = 0.004; S3 Table.
Crocodile isotopic composition was closest to that of terrestrial animals in dual isotope
space (δ13C vs δ15N, Fig 3A and 3B), especially pigs and buffaloes. When plotting δ15N vs δ34S,
crocodiles in both EAR and SAR were closest to pigs and mullet (Fig 3C and 3D). In 3D
isotope plots (Fig 4), crocodiles most closely plot around terrestrial and riverine sources. Using
the proximity-based method for evaluating diet relationships in 3D isotope space (δ13C, δ15N,
and δ34S), terrestrial animals had the highest contribution to the diet of crocodiles in both
EAR and SAR with contributions of 58.9 ± 2.0 and 69.0 ± 0.6%, respectively (Fig 4).
Contributions for pigs and buffalo were particularly high with 23.0 ± 1.2 and 22.2 ± 1.7% for the EAR,
and 27.3 ± 2.9 and 24.0 ± 2.2% for the SAR. Contributions of other prey were similar between
6 / 16
the EAR and SAR with 18.0 ± 1.8 and 9.2 ± 0.3% for riverine, 15.1 ± 0.5 and 12.0 ± 0.5% for
riverine-marine, and 7.9 ± 0.3 and 9.7 ± 0.7% for marine prey.
When assessing the difference between different isotopic discrimination factors for δ15N,
we found that if we used the minimum discrimination factor (Δ = 0.8 ½) the contribution of
river-marine prey increased by 28% (20.6 ± 0.9 and 16.9 ± 1.5% for EAR and SAR, respectively;
Fig 3B). However, with either discrimination factor (minimum of 0.8 and maximum of 3.0½),
terrestrial prey was the dominant contributor to the crocodile's diet (53.0 ± 1.7, and 58.6 ±
2.6% for the EAR and SAR, respectively), even when considering other possible prey such as
birds, reptiles, crabs, shrimp, snails and sharks (S1 Table).
7 / 16
The SIAR solo Bayesian model applied to individuals showed similar results, with terrestrial
prey having the highest contribution, although values were higher with a mean of 72.9 ± 4.2%.
Riverine prey had the second largest contribution with 14.4 ± 4.1%, followed by marine with
9.4 ± 4.3% and river-marine with 3.4 ± 0.7%. The stomach contents of one crocodile (number
23 in the wet season in Fig 5) were matched to its diet. This crocodile had wallaby, turtle and
vegetation material in its stomach and isotope values that suggested a predominately terrestrial
We found a difference in diet with crocodile size, with larger crocodiles having a higher
contribution of marine prey to their diet compared to smaller ones (R2 = 0.30; p = 0.005; Fig
6D). Additionally, the source of nutrition of crocodiles was variable between crocodiles
sampled in different seasons (Fig 5). The crocodiles caught during the dry season had isotopic
values that suggested a heavy reliance on terrestrial prey with a mean contribution of 84.1 ± 2.4%.
The crocodiles caught during the wet season had a mean contribution of 55.4 ± 7.0% of
terrestrial prey, 25.1 ± 8.1% of riverine prey, and 12.6 ± 3.9% of marine prey. However, this
difference in wet versus dry season was mostly driven by six crocodiles, five individuals caught in
the Magela floodplain in the EAR with predominately a riverine diet and one individual caught
at Cahill's crossing (EAR) which had a predominately a marine diet (Fig 5).
The proximity analyses for seasonal comparisons showed similar results than the SIAR Solo
model, but with slightly different contributions. Terrestrial prey contributed 52% in the dry
season versus 36% in the wet season; river prey contributed 19% in the dry season and 28% in
the wet season; river-marine prey contributed 23% in the dry and 30% in the wet season; and
finally, marine prey contributed 7% in the dry season and 6% in the wet season.
The difference in diet between the dry and wet season was also noted in the δ15N values,
which were on average 0.6½ higher in the wet compared to the dry season, a difference that
was slight, but significant (F1, 43 = 4.28, p = 0.04). When considering sex and catchment in the
model, the difference between seasons was not significant (season sex catchment F1, 31 = 1.50,
8 / 16
Fig 2. Correlation between isotopic value of crocodile scute (δ13C and δ15N; ½) and body size (mm). Red circles
are values from this study, blue circles are values from C. porosus in Cape York, Australia, from Hanson et al. 2015 [
Both datasets follow a linear trend of higher isotope values with increasing total body length (mm).
p = 0.23). Nevertheless, low δ15N values seemed the best overall indication of dietary
dependence on terrestrial prey in this study. Low δ15N values were found only in terrestrial
herbivorous animals (pigs, buffalo and wallabies), but not in other vertebrates and invertebrates
collected (birds, reptiles, shrimp, crabs, snails, shark, S1 Table).
Most estuarine crocodiles sampled in this study derived a large proportion of their nutrition
from terrestrial prey. This result is surprising, because in general, crocodilians are considered
apex aquatic predators (e.g. [
]). However, our results suggest that this population is having
less influence in the aquatic food web than previously thought. Although estuarine crocodiles
in this region consume a wide range of prey, it appears that for a large number of individuals,
terrestrial animals are an important source of nutrition.
9 / 16
Fig 3. Isotopic composition (δ34S, δ13C and δ15N) of crocodiles (black diamonds) and potential prey. Samples were
obtained fromthe East Alligator River (A,C, E) and South Alligator River (B,D,F), terrestrial prey includes water
buffalo (orange circle), pigs (red circle) and wallabies (yellow circle); riverine is represented by mullet (green circle);
riverine-marine prey is represented by barramundi (purple circle) and marine prey is represented by giant sea catfish
(blue circle). Crocodile data were corrected to the level of prey by subtracting 1.4 ½ from δ13C values [
]. The effect on
the crocodile value of different fractionation factors is shown as a box of possible values around the crocodile mean in
Previous studies have shown that terrestrial animals could provide an important source of
nutrition for the estuarine crocodile in the region [
]. For example, native rats (Rattus
colletti and Xeromys myoides) have been found in the stomachs of juvenile crocodiles (0.9±1.3 m
]). Larger crocodiles (> 3m), are also known to hunt and kill large pigs and cattle at
the water edge (G. Lindner pers. comm), and smaller individuals have been observed feeding
on pig and cattle carrion left by larger dominant males [
] (G. Lindner, pers comm.). An
estuarine crocodile requires approximately 4% of their body weight every week to maintain its
body mass . As such, a large crocodile of 1,000 kg would need 40 kg of food per week. Feral
pigs weigh between 50±100 kg; thus, many crocodiles could take advantage of this prey that is
large, abundant, and relatively easy to catch to satisfy their dietary requirements.
Despite the reliance on terrestrial prey for most of the sampled crocodiles, we found a
difference in diet associated with body size. Previous results have suggested ontogenetic changes
in the diet of estuarine crocodiles in Australia [
]. Our results support this finding; we found a
significant increase in δ13C and δ15N values with body size. The difference in isotopic
composition appears to be associated with higher marine contribution in the large crocodiles (3 to
4.5m TL). Similar ontogenetic shifts in diets have been found in other crocodilians such as
Crocodylus niloticus in Botswana [
], Alligator mississippiensis in the US, and Crocodylus acutus
10 / 16
Fig 4. 3D plot of the isotopic composition (δ13C, δ15N and δ34S) of crocodiles (diamonds) and potential prey. Samples are
from the East and South Alligator Rivers; terrestrial prey includes water buffalo, pigs and wallabies; riverine prey is represented
by mullet; riverine-marine prey is represented by barramundi and marine prey is represented by giant sea catfish.
Fig 5. Proportion of contribution of prey to the nutrition ofestuarine crocodiles in Kakadu National Park.
Crocodiles were caught during the dry and wet season and isotope values were analysed with SIAR solo. Terrestrial
prey includes water buffalo, pigs and wallabies; riverine prey is represented by mullet; riverine-marine prey is
represented by barramundi and marine prey is represented by giant sea catfish. Box plots represent the 5, 25, 75 and
95% credibility intervals.
PLOS ONE | https://doi.org/10.1371/journal.pone.0197159
11 / 16
Fig 6. Contribution of (A) terrestrial, (B) riverine, (C) riverine-marine and (D) marine prey to the diet of
estuarine crocodiles. The contribution was assessed with SIAR solo using data from crocodiles ranging from 0.85 to
4.2 m long (TL).
in Belize; similar to our results, adults of the latter species consume more marine prey than
]. However, the difference in δ13C and δ15N values could also mean different
metabolic processes associated with body size .
Crocodiles are highly individualistic in their diets [
], which is reflected in the wide
range of isotopic values within their tissues. Large variability in isotope values from crocodiles
also suggests a wide diversity of habitats, diversity of prey available, and different home range
]. For instance, some crocodiles are highly mobile (e.g. travelling 1000 km in 6
months), while others are fairly resident within their home ranges . It has been
hypothesized that differences in homing strategies result in differences in diet [
]. In our dataset, five
crocodiles were caught in the inundated floodplains during the wet season . These
crocodiles could be nomadic individuals traveling onto the floodplain and taking advantage of its
high productivity [
]. These individuals had primarily a riverine diet. Contrary, all the
individuals sampled in the dry season had primarily a terrestrial diet. During the dry season,
crocodiles are concentrated in waterholes along with terrestrial animals, including introduced pigs
and buffaloes, which depend on these waterholes for drinking water [
]. During this period,
terrestrial animals seem to provide abundant and relatively easy prey for these crocodiles that
are hunting at the land-water interface [
Crocodile populations have declined in many areas of the world. In South Africa, a decline
in population numbers of C. niloticus has been associated with water pollution [
northern Australia, the decline of the freshwater crocodile, C. johnstoni, was related to the
introduction of the poisonous cane toad (Bufo marinus. [
]).The estuarine crocodile population in
Kakadu National Park appears to be stable [
], with individuals having a wide trophic
niche and large geographical distribution. It seems plausible that the capacity of the crocodiles
12 / 16
to feed on introduced terrestrial prey has played a role in the recovery of their population
], probably because the carrying capacity of their habitat has increased due to the
increase in prey, particularly pigs. Crocodiles appear to be exerting top-down pressure on pigs
and buffaloes, which are considered serious pests in the area. Our results provide an empirical
example of a native predator adapting to an invasive prey and capitalizing on it [
]. A similar
phenomenon might be occurring in other areas where introduced terrestrial prey has become
abundant. For example, estuarine crocodiles in the Gulf of Carpentaria (northeast Australia)
also have relatively low δ15N values (2 to 5½) and high δ13C values (-26 to -18½, Fig 2 [
These values are similar to those in the current study, and are consistent with the isotopic
composition of terrestrial herbivores.
Based on our isotopic analyses and direct observations, we found that estuarine crocodiles in
Kakadu National Park, Northern Australia consume a large variety of prey. However, most of
the crocodiles sampled in this study derive a large proportion of their nutrition from terrestrial
S1 Fig. Crocodile scute sampled for isotope analyses of a Crocodylus porosus individual.
S1 Table. Characteristics of Crocodylus porosus individuals from in Kakadu National Park,
Australia. EAR = East Alligator River; SAR = South Alligator River, WR = Wildman River;
M = Male; F = Female.
S2 Table. Isotope values of potential prey for estuarine crocodiles (Crocodylus porosus) in
Kakadu National Park, Australia. Values are mean ± standard error.
S3 Table. Isotope values of potential prey for estuarine crocodiles (Crocodylus porosus) in
Kakadu National Park, Australia: Marine, marine-riverine, riverine and terrestrial.
We acknowledge the Traditional Owners, the Bininj/Mungguy people, of the land in which
this research was done. We thank Kakadu National Park staff for logistic support in the field.
We thank ERRISS for providing accommodation in the Park. This research was conducted
with the support of funding from the Queensland Government through the Advance
Queensland Fellowship granted to MFA.
Conceptualization: Maria Fernanda Adame, Timothy D. Jardine, Brian Fry, Stuart E. Bunn.
Data curation: Maria Fernanda Adame, Timothy D. Jardine, Dominic Valdez.
Formal analysis: Maria Fernanda Adame, Timothy D. Jardine, Brian Fry.
Funding acquisition: Stuart E. Bunn.
13 / 16
Investigation: Maria Fernanda Adame, Timothy D. Jardine, Dominic Valdez, Garry Lindner,
Jonathan Nadji, Stuart E. Bunn.
Methodology: Maria Fernanda Adame, Timothy D. Jardine, Brian Fry, Garry Lindner,
Jonathan Nadji, Stuart E. Bunn.
Project administration: Stuart E. Bunn.
Resources: Dominic Valdez.
Supervision: Maria Fernanda Adame, Stuart E. Bunn.
Validation: Dominic Valdez, Garry Lindner, Jonathan Nadji.
Visualization: Garry Lindner, Jonathan Nadji.
Writing ± original draft: Maria Fernanda Adame.
Writing ± review & editing: Maria Fernanda Adame, Timothy D. Jardine, Brian Fry, Dominic
Valdez, Stuart E. Bunn.
14 / 16
15 / 16
1. Ross J. Crocodiles. Status Sruvey and Conservation Plan . 2nd ed. Specialist IC, editor. Gland, Switzerland and Cambridge, UK; 1998 .
2. Kar SK , Bustard HR . Saltwater crocodile attacks on man . Biol Conserv . 1983 ; 25 : 377 ± 382 . https://doi. org/10.1016/ 0006 - 3207 ( 83 ) 90071 -X
Webb GJW , Yerbury M , Onions V. A record of a Crocodylus porosus (Reptilia, Crocodylidae) attack . J Herpetol . 1978 ; 12 : 267 ± 268 .
4. Magnusson W , Viera da Silva E , Lima A . Diets of Amazonian crocodilians . J Herpetol . 1987 ; 21 : 85 ± 95 .
5. Radloff FGT , Hobson KA , Leslie AJ . Characterising ontogenetic niche shifts in Nile crocodile using stable isotope (δ13C, δ15N) analyses of scute keratin . Isotopes Environ Health Stud . 2012 ; 48 : 439 ± 56 . https://doi.org/10.1080/10256016. 2012 .667808 PMID: 22462522
Webb GJW , Hollis GJ , Manolis C . Feeding, growth and food conversion rates of wild juvenile saltwater crocodiles (Crocoylus porosus) . J Herpetol . 1991 ; 462 ± 473 .
7. Hanson JO , Salisbury SW , Campbell HA , Dwyer RG , Jardine TD , Franklin CE . Feeding across the food web: The interaction between diet, movement and body size in estuarine crocodiles (Crocodylus porosus) . Austral Ecol . 2015 ; 40 : 275 ± 286 . https://doi.org/10.1111/aec.12212
8. Taylor JA . The food and feeding habits of subadult Crocodylus porosus in Northern Australia . Aust Wildl Res . 1979 ; 6 : 347 ± 360 .
9. Sah SAM , Stuebing RB . Diet, growth and movements of juvenile crocodiles Crocodylus porosus Schneider in the Klias River , Sabah, Malaysia. J Trop Ecol . 1996 ; 12 : 651 ± 662 . https://doi.org/10.1017/ S0266467400009858
Webb GJW , Manolis SC , Buckworth R . Crocodylus johnstoni in the McKinlay River Area , N.T.I. Variation in the diet, and a new method of assessing the relative importance of prey . Aust J Zool . 1982 ; 30 : 877 ± 899 .
11. Fry B . Stable isotope ecology . New York, U.S.A .: Springer; 2006 . 295pp.
12. Fukuda Y , Webb G , Manolis C , Delaney R , Letnic M , Lindner G , et al. Recovery of saltwater crocodiles following unregulated hunting in tidal rivers of the Northern Territory, Australia . J Wildl Manage . 2011 ; 75 : 1253 ± 1266 . https://doi.org/10.1002/jwmg.191
13. Nifong JC , Layman CA , Silliman BR . Size, sex and individual-level behaviour drive intrapopulation variation in cross-ecosystem foraging of a top-predator . J Anim Ecol . 2015 ; 84 : 35 ± 48 . https://doi.org/10. 1111/ 1365 - 2656 .12306 PMID: 25327480
14. Tucker AD , Limpus CJ , Mccallum HI , Mcdonald KR , Tucker AD , Limpus CJ , et al. Ontogenetic dietary partitioning by Crocodylus johnstoni during the dry season . Copeia . 1996 ; 1996 : 978 ± 988 .
Webb GJW , Messel H. Movement and dispersal patterns of Crocodylus porosus in some rivers of Arnhem Land, Northern Australia . Wildl Res . 1978 ; 5 : 263 ± 283 . https://doi.org/10.1071/WR9780263
16. Schmitz O , Suttle K. Effects of top predator species on direct and indirect interactions in a food web . Ecology . 2001 ; 82 : 2072 ± 2081 .
17. McCann KS , Rasmussen JB , Umbanhowar J. The dynamics of spatially coupled food webs . Ecol Lett . 2005 ; 8 : 513 ±523 https://doi.org/10.1111/j.1461- 0248 . 2005 . 00742 . x PMID : 21352455
18. Atwood TB , Hammill E , Greig HS , Kratina P , Shurin JB , Srivastava DS , et al. Predator-induced reduction of freshwater carbon dioxide emissions . Nat Geosci . 2013 ; 6 : 191 ± 194 . https://doi.org/10.1038/ ngeo1734
19. Letnic M , Webb JK , Shine R . Invasive cane toads (Bufo marinus) cause mass mortality of freshwater crocodiles (Crocodylus johnstoni) in tropical Australia . Biol Conserv . 2008 ; 141 : 1773 ± 1782 . https://doi. org/10.1016/j.biocon. 2008 . 04 .031
20. Australian Government . Department of Sustainability, Environment, Water, Population and the Communities. The feral pig (Sus scrofa) . Canberra, Australia; 2011 .3 pp.
Woinarski JCZ , Legge S , Fitzsimons JA , Traill BJ , Burbidge AA , Fisher A , et al. The disappearing mammal fauna of northern Australia: Context, cause, and response . Conserv Lett . 2011 ; 4 : 192 ± 201 . https:// doi.org/10.1111/j. 1755 - 263X . 2011 . 00164 .x
Ward DP , Pettit NE , Adame M , Douglas MM , Setterfield SA , Bunn SE . Seasonal spatial dynamics of floodplain macrophyte and periphyton abundance in the Alligator Rivers region (Kakadu) of northern Australia . Ecohydrology . 2016 ; 9 : 1675 ± 1686 . https://doi.org/10.1002/eco.1757
23. Adame M , Pettit N , Valdez D , Ward D , Burford M , Bunn S. The contribution of epiphyton to the primary production of tropical floodplain wetlands . Biotropica . 2017 ; 49 : 461 ± 471 .
24. Australian Government . Australian Bureau of Meteorology. Available: http://www.bom.gov.au/. Viewed January 2017 .
25. Rosenblatt AE , Heithaus MR . Slow isotope turnover rates and low discrimination values in the American Alligator: Implications for interpretation of ectotherm stable isotope data . Physiol Biochem Zool . 2013 ; 86 : 137 ± 148 . https://doi.org/10.1086/668295 PMID: 23303328
26. Jardine TD , Hunt RJ , Pusey BJ , Bunn SE . A non-lethal sampling method for stable isotope studies of tropical fishes . Mar Freshw Res . 2011 ; 62 : 83 ± 90 .
27. Marques T , Bassetti L , Lara N , ArauÂjo M , Piña C , Camargo P , et al. Isotopic Discrimination Factors (Δ13C and Δ15N) between tissues and diet of the broad-snouted Caiman (Caiman latirostris) . J Herpetol . 2014 ; 48 : 332 ± 337 .
28. Nehlich O. The application of sulphur isotope analyses in archaeological research: A review . Earth Sci Rev . 2015 ; 142 : 1± 17 .
Webb GJW , Reynolds S , Brien M , Manolis C , Brien J , Christian K. Improving Australia's crocodile industry productivity . Canberra, Australia; 2013 .
30. Smith JA , Mazumder D , Suthers I , Taylor M. To fit or not to fit: evaluating stable isotope mixing models using simulated mixing polygons . Methods Ecol Evol . 2013 ; 4 : 612 ± 618 .
31. Campbell HA , Dwyer RG , Irwin TR , Franklin CE . Home range utilisation and long-range movement of estuarine crocodiles during the breeding and nesting season . PLoS One . 2013 ; 8 . https://doi.org/10. 1371/journal.pone. 0062127 PMID: 23650510
32. Phillips D , Gregg J. Uncertainty in source partitioning using stable isotopes . Oecologia . 2011 ; 127 : 171 ± 179 .
33. Parnell AC , Inger R , Bearhop S , Jackson AL . Source partitioning using stable isotopes: Coping with too much variation . 2010 ; 5: e9672 . https://doi.org/10.1371/journal.pone. 0009672 PMID: 20300637
34. Fry B . Alternative approaches for solving underdetermined isotope mixing problems . Mar Ecol Prog Ser . 2013 ; 472 : 1± 13 . https://doi.org/10.3354/meps10168
35. Fry B . Minmax solutions for underdetermined isotope mixing problems: Reply to Semmens et al . Mar Ecol Prog Ser . 2013 ; 490 : 291 ± 294 .
36. Brett MT . Resource polygon geometry predicts Bayesian stable isotope mixing model bias . Mar Ecol Prog Ser . 2014 ; 514 : 1± 12 .
37. Taylor J . A rapid approach to evaluate putative nursery sites for penaeid prawns . J Sea Res . 2016 ; 114 : 26 ± 31 .
38. Magnusson WE , Webb GJW , Taylor JA . Two new locality records, a new habitat and a nest description for Xeromys myoides thomas (rodentia: Muridae) . Wildl Res . 1976 ; 3 : 153 ± 157 . https://doi.org/10.1071/ WR9760153
39. Manolis SC , Webb GJW , Britton AR . Crocodiles and other reptiles: bioindicators of pollution . In: The Finnis River: A natural laboratory of mining impacts . Sydney, Australia; 2002 . pp. 65 ± 69 .
40. Platt SG , Rainwater TR , Finger AG , Thorbjarnarson JB , Anderson TA , McMurry ST . Food habits, ontogenetic dietary partitioning and observations of foraging behaviour of Morelet's crocodile (Crocodylus moreletii) in northern Belize . Herpetol J . 2006 ; 16 : 281 ± 290 .
41. Villamarin F , Jardine TD , Bunn SE , Marioni B , Magnusson WE . Body size is more important than diet in determining stable-isotope estimates of trophic position in crocodilians . Sci. Reports 2018 ; 8: 2020
42. Delany M , Linda S , Moore C . Diet and condition of American alligators in 4 Florida Lakes . Proc Annu Southeast Assoc Fish Wild Agencies . 1999 ; 53 : 375 ± 389 .
43. Layman CA , Quattrochi JP , Peyer CM , Allgeier JE . Niche width collapse in a resilient top predator following ecosystem fragmentation . Ecol Lett . 2007 ; 10 : 937 ± 944 . https://doi.org/10.1111/j.1461- 0248 . 2007 . 01087 . x PMID : 17845294
Ward DP , Petty A , Setterfield S. , Douglas MM , Ferdinands K , Hamilton SK , et al. Floodplain inundation and vegetation dynamics in the Alligator Rivers region (Kakadu) of northern Australia assessed using optical and radar remote sensing . Remote Sens Environ . 2014 ; 147 : 43 ± 55 . https://doi.org/10.1016/j.
rse. 2014 . 02 .009
45. Jardine TD , Rayner TS , Pettit NE , Valdez D , Ward DP , Lindner G , et al. Body size drives allochthony in food webs of tropical rivers . Oecologia . Springer Berlin Heidelberg; 2017 ; 505. https://doi.org/10.1007/ s00442-016-3786-z PMID: 27896479
46. Davenport J , Grove D , Cannon J , Ellis TR , Stables S. Food capture, appetite, digestion rate and efficiency in hatchling and juvenile Crocodylus porosus . J Zoo Lond . 1990 ; 220 : 569 ± 592 .
47. Botha H , van Hoven W , Guillette LJ . The decline of the Nile crocodile population in Loskop Dam , Olifants River, South Africa. Water SA . 2011 ; 37 : 103 ± 108 . https://doi.org/10.4314/wsa.v37i1. 64109
48. Nichols T , Letnic M. Problem crocodiles: reducing the risk of attacks by Crocodylus porosus in Darwin Harbour, Northern Territory, Australia . In: Mitchell J, Brown R , Bartholomew B , editors. Herpetological Conservation . 2008 . pp. 503 ± 512 .
49. Purvis A , Gittleman JL , Cowlishaw G , Mace GM . Predicting extinction risk in declining species . Proc R Soc London B . 2000 ; 267 : 1947 ± 1952 . https://doi.org/10.1098/rspb. 2000 .1234 PMID: 11075706
50. Barber N , Marquis R , Tori W. Invasive prey impacts the abundance and distribution of native predators . Ecology . 2008 ; 89 : 2678 ± 2683 . PMID: 18959305
51. Carlsson NOL , Sarnelle O , Strayer DL . Native predators and exotic preyÐAn acquired taste? Front Ecol Environ . 2009 ; 7 : 525 ± 532 . https://doi.org/10. 1890 /080093