Demographic status of Komodo dragons populations in Komodo National Park
ABSTRACT The Komodo dragon (Varanus komodoensis) is the world's largest lizard and endemic to five islands in Eastern Indonesia. The current management of this species is limited by a paucity of demographic infor-mation needed to determine key threats to population persistence. Here we conducted a large scale trap-ping study to estimate demographic parameters including population growth rates, survival and abundance for four Komodo dragon island populations in Komodo National Park. A combined capture mark recapture framework was used to estimate demographic parameters from 925 marked individuals monitored between 2003 and 2012. Island specific estimates of population growth, survival and abun-dance, were estimated using open population capture–recapture analyses. Large island populations are characterised by near or stable population growth (i.e. k $ 1), whilst one small island population (Gili Motang) appeared to be in decline (k = 0.68 ± 0.09). Population differences were evident in apparent sur-vival, with estimates being higher for populations on the two large islands compared to the two small islands. We extrapolated island specific population abundance estimates (considerate of species habitat use) to produce a total population abundance estimate of 2448 (95% CI: 2067–2922) Komodo dragons in Komodo National Park. Our results suggest that park managers must consider island specific population dynamics for managing and recovering current populations. Moreover understanding what demographic, environmental or genetic processes act independently, or in combination, to cause variation in current population dynamics is the next key step necessary to better conserve this iconic species.
- SourceAvailable from: Tim S. Jessop[Show abstract] [Hide abstract]
ABSTRACT: Finding practical ways to robustly estimate abundance or density trends in threatened species is a key facet for effective conservation management. Further identi-fying less expensive monitoring methods that provide adequate data for robust population density estimates can facilitate increased investment into other conservation initiatives needed for species recovery. Here we evaluated and compared inference-and cost-effec-tiveness criteria for three field monitoring-density estimation protocols to improve con-servation activities for the threatened Komodo dragon (Varanus komodoensis). We undertook line-transect counts, cage trapping and camera monitoring surveys for Komodo dragons at 11 sites within protected areas in Eastern Indonesia to collect data to estimate density using distance sampling methods or the Royle–Nichols abundance induced het-erogeneity model. Distance sampling estimates were considered poor due to large confi-dence intervals, a high coefficient of variation and that false absences were obtained in 45 % of sites where other monitoring methods detected lizards present. The Royle–Nichols model using presence/absence data obtained from cage trapping and camera monitoring produced highly correlated density estimates, obtained similar measures of precision and recorded no false absences in data collation. However because costs associated with Communicated by Indraneil Das.Biodiversity and Conservation 07/2014; 23(10). · 2.07 Impact Factor
Demographic status of Komodo dragons populations
in Komodo National Park
Deni Purwandanaa, Achmad Ariefiandya, M. Jeri Imansyaha, Heru Rudihartob, Aganto Senob,
Claudio Ciofic, Damien A. Fordhamd, Tim S. Jessope,⇑
aKomodo Survival Program, Denpasar, Bali, Indonesia
bKomodo National Park, Labuan Bajo, Flores, Indonesia
cDepartment of Biology, University of Florence, 50019 Sesto Fiorentino, FI, Italy
dEnvironment Institute and School of Earth and Environmental Sciences, University of Adelaide, North Terrace, SA 5005, Australia
eDepartment of Zoology, University of Melbourne, Parkville, Vic 3010, Australia
a r t i c l e i n f o
Received 27 July 2013
Received in revised form 11 January 2014
Accepted 12 January 2014
a b s t r a c t
The Komodo dragon (Varanus komodoensis) is the world’s largest lizard and endemic to five islands in
Eastern Indonesia. The current management of this species is limited by a paucity of demographic infor-
mation needed to determine key threats to population persistence. Here we conducted a large scale trap-
ping study to estimate demographic parameters including population growth rates, survival and
abundance for four Komodo dragon island populations in Komodo National Park. A combined capture
mark recapture framework was used to estimate demographic parameters from 925 marked individuals
monitored between 2003 and 2012. Island specific estimates of population growth, survival and abun-
dance, were estimated using open population capture–recapture analyses. Large island populations are
characterised by near or stable population growth (i.e. k ? 1), whilst one small island population (Gili
Motang) appeared to be in decline (k = 0.68 ± 0.09). Population differences were evident in apparent sur-
vival, with estimates being higher for populations on the two large islands compared to the two small
islands. We extrapolated island specific population abundance estimates (considerate of species habitat
use) to produce a total population abundance estimate of 2448 (95% CI: 2067–2922) Komodo dragons in
Komodo National Park. Our results suggest that park managers must consider island specific population
dynamics for managing and recovering current populations. Moreover understanding what demographic,
environmental or genetic processes act independently, or in combination, to cause variation in current
population dynamics is the next key step necessary to better conserve this iconic species.
? 2014 Elsevier Ltd. All rights reserved.
The Indonesian archipelago possesses extraordinary species
diversity and endemism (McKinnon, 1996; Mittermeier et al.,
1998; Whitten, 2000). However, rapid and wide spread habitat
loss, variable capacity in natural reserve management, alongside
the looming challenges of climate change pose considerable risk
to the nation’s biodiversity (Sodhi et al., 2004; Fordham and Brook,
2010). Establishing the threat human impacts pose to biodiversity
in Indonesia is made difficult by the absence of important demo-
graphic data needed to assess conservation status (Harris et al.,
2011). For example, even for the iconic Komodo dragon (Varanus
komodoensis), the world’s largest lizard lacks estimates of rates of
population growth or abundance (Jessop et al., 2007). The Komodo
dragon is an apex predator and has an isolated island distribution
making it particularly sensitive to global change (Cardillo et al.,
Komodo dragons inhabit five islands in eastern Indonesia, with
four island populations located within Komodo National Park
(KNP) and several fragmented populations persisting on the larger
island of Flores (Ciofi and De Boer, 2004; Jessop et al., 2004, 2007).
Komodo dragon range area is thus small, with isolated populations.
Its distribution is suspected to have decreased substantially in re-
cent decades (Ciofi, 2002) as a consequence of prey removal by hu-
mans (e.g. Timor deer) and habitat loss (forest conversion to
agriculture) (Ciofi, 2002; Jessop et al., 2007), especially outside pro-
tected areas. The Komodo dragon is classified by the International
Union for the Conservation of Nature (IUCN) as ‘vulnerable’ due to
demographic decline and limited distribution (IUCN, 2012).
We undertook a large-scale field study to estimate island spe-
cific demographic parameters in Komodo National Park in Eastern
Indonesia. Komodo National Parks remains the largest and best
0006-3207/$ - see front matter ? 2014 Elsevier Ltd. All rights reserved.
⇑Corresponding author. Address: Department of Zoology, University of Mel-
bourne, Parkville, Vic 3052, Australia. Tel.: +61 83440206.
E-mail address: email@example.com (T.S. Jessop).
Biological Conservation 171 (2014) 29–35
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/biocon
resourced of the protected areas existing within the distribution of
Komodo dragons. Intuitively, given the large difference in island
area and proximity among the park’s extant Komodo dragon pop-
ulations, we hypothesised that demographic parameters could
vary among populations due to different evolutionary histories
and contemporary variation in habitat quality or demographic
and genetic processes (Frankham, 1998; Whittaker and Fernandez-
Palacios, 2007; Laver et al., 2012). Our aims were three fold. The
first was to assess the status of the four island populations by eval-
uating averaged population growth rates to summarise population
trajectories as exhibiting positive, stable or negative population
growth. Our second aim was to estimate mean survival for each
Komodo dragon populations to ascertain further potential island
differences in population demography. Our final aim estimated
population abundance at ten sites on four islands and extrapolated
these island estimates to produce a total Komodo dragon popula-
tion abundance for Komodo National Park. These demographic
estimates represent a long-standing goal of our field monitoring
activities and a key request of the Indonesian National Park
Authority. Ultimately having access to this important demographic
information enabled us to discuss the ensuing implications for cur-
rent management and prioritization of conservation activities for
Komodo dragons within Komodo National Park.
2. Materials and methods
2.1. Study area
Fieldwork for this study was conducted from April 2003–
November 2006 and again from April 2009–April 2012 in Komodo
National Park, Eastern Indonesia (8?3504000S, 119?2505100E). The to-
tal area of Komodo National Park is 1817 km2of which 603 km2is
land and 1214 km2is sea (Fig. 1). Komodo National Park consists of
the two large islands of Komodo and Rinca (311.5 km2and
204.8 km2, respectively)and
14.1 km2; Gili Motang, 9.5 km2; and Nusa Kode, 7.8 km2). There
are four main vegetation communities in Komodo National Park.
Tropical monsoon forest dominates areas above 500–700 m. At
lower elevations deciduous monsoon forest (primarily Tamarindus
indica) occurs in valley floors holding permanent and ephemeral
water courses. Savannah woodland and savannah grassland occu-
py drier areas distal from water courses. The climate is highly sea-
sonal and dominated by a long dry season (March–November) and
a short summer wet season (December–February). The annual
rainfall is less than 1000 mm. The mean monthly maximum and
minimum temperature ranges from 31–38 ?C and 19.5–25.6 ?C,
2.2. Trapping protocols
We captured Komodo dragons at 273 fixed trapping locations
established at ten study sites on the four islands (Komodo, Rinca,
Nusa Kode and Gili Motang) that have extant lizard populations
(Figs. 1 and 2a). The Komodo Island sites: Loh Liang (K1), Loh Lawi
(K2), Loh Sebita (K3) and Loh Wau (K4) consisted of 32, 32, 21 and
9 trapping locations respectively. The Rinca Island sites: Loh Buaya
(R1), Loh Baru (R2), Loh Tongker (R3) and Loh Dasami (R4) con-
sisted of 22, 22, 13 and 24 trapping locations respectively. Gili Mo-
tang (GM) and Nusa Kode (NK) sites consisted of 16 and 12
trapping locations. The variation in sampling effort reflected differ-
ences in the area of the study sites (that often consisted of discrete
valleys). Trapping locations within each study site were located in
landscapes dominated by deciduous monsoon forest that is consid-
ered high quality habitat for Komodo dragons (Fig. 2b).
Within each study site, we used purpose built traps set at differ-
ent trapping locations to capture Komodo dragons (Fig. 2c). Traps
Fig. 1. Map of Komodo National Park and the location of the 10 trapping sites used in this study. Four sites were each located on the large islands of Komodo (sites denoted as
K1, K2, K3, K4); and Rinca (R1, R2, R3, R4). A single trapping site was located each on the small islands of Gili Motang and Nusa Kode. The shaded area constitutes the effective
trapping area within each site.
D. Purwandana et al./Biological Conservation 171 (2014) 29–35
comprised aluminium boxes (300 cm L ? 50 cm H ? 50 cm W) fit-
ted with a wire activated front door (Fig. 2d). The distance between
trap locations was set at approximately 500 m in order to maintain
independence among traps (based on telemetry data; e.g.
Imansyah et al., 2008). Traps were positioned in shaded areas to
avoid the potential overheating of trapped individuals. Fresh goat
meat (? 0.5 kg) was used as bait to lure lizards into traps.
Field work durations varied between 8 and 14 days per site (in
each year) to enable coverage of all trapping locations within each
site. Field work was conducted during the dry season with each
study site monitored sequentially at the same time each year.
Within sites, at each trapping location, trapping occurred for three
consecutive days, with each trap checked twice daily (8:00–11:00
and 14:00–17:00). The time interval between the morning and
afternoon daily check for each trap was ?6 h. We also attempted
to directly capture, via noose pole, any Komodo dragon that was lo-
cated within a 50 m radius of each trap to increase sample size at
each trapping location. Both traps and noose pole permitted cap-
ture of all Komodo dragon size classes, with the exception of hatch-
ling and small juvenile lizards (<1.5 kg) that exhibit an arboreal life
phase before exploiting terrestrial habitat use (Imansyah et al.,
2.3. Lizard handling and restraint protocols
Following capture, dragons were restrained with rope and their
mouths taped closed. Snout to vent length (SVL) was measured
using a flexible plastic tape between the tip of the snout (i.e. junc-
ture between upper and lower jaw) and the cloaca. The SVL re-
corded for each individual represented the average of two
measurements that were within 0.5 cm of each other. Body mass
was obtained using digital scales. Each dragon was permanently
identified using a passive integrated transponder (Trovan ID100a,
Micorchips Australia Pty Ltd., Australia) inserted between the der-
mis and muscle of the upper right hind leg. In addition we painted
each lizard’s back with a unique paint code (using non-toxic fabric
marker) to increase our ability to recognise individual lizards and
make general observations of their behaviour whilst conducting
our study. Processing time was usually less than 20 min before liz-
ards were released at their point of capture.
2.4. Demographic analyses
We collated captures of individually marked lizards to develop
capture histories. All analyses were conducted in Program Mark
(White and Burnham, 1999). Analyses were performed on a single
combined data set in which each island population was coded as
an individual group.
For estimates of population growth, survival and abundance, we
developed a set of candidate models for analysis, evaluated good-
ness-of-fit, and estimated an overdispersion parameter ð^ cÞ for the
data set. We used an information theoretic approach to select the
most parsimonious model, based on the AICc model selection crite-
rion (lower AICc values represented better fitting models) (Burnham
and Anderson, 2002). Models were then ranked using the quasi-
likelihood AICcvalue (QAICc) to account for the any overdispersion
and their respective model weights (w) estimated to evaluate their
strength of model support (Anderson et al., 1994). We used the
model-averaging approach (conducted in program Mark) that
incorporated model selection uncertainty for all models with sub-
stantial model support (DAIC < 2 from the top ranked model).
We calculated rates of finite population growth (k) time-aver-
aged across the study for each Komodo dragon population. Popula-
tion growth rate (k = Nt1/Nt) describes the per capita rate of growth
of population, either as the factor by which population size in-
creases (k > 1.0), decreases (k < 1.0), or is stable (k = 1.0) per unit
time (Sibly and Hone, 2002). Using the combined mark recapture
data with each island identified as a group we applied the Pradel
Survival and Lambda analysis (Pradel, 1996) to estimate time aver-
aged population specific growth rates and survival estimates.
Fig. 2. Depicts (a) Komodo dragon, the target species of this study; (b) typical site
specific habitat used to conduct trapping in Komodo National Park; (c) an example
of a trapping grid design used to conduct mark recapture analyses; (d) and a
purpose built trap (3 m length) used to capture lizards in this study.
D. Purwandana et al./Biological Conservation 171 (2014) 29–35
A candidate set of 18 models were assessed to evaluate average
population specific growth rates and survival estimates. Here mod-
els varied in parameter combinations with survival (U) and recap-
ture (p), parameters considering all combinations of time constant
(.) or variable (t) time and group (i.e. island) effects. The population
growth (k) parameter was modelled as function of group (i.e. kg).
Additionally null and global models were estimated in model com-
parisons. We assessed goodness-of-fit of the global model and esti-
mated overdispersion (^ c =v2/df using combined v2values and
degrees of freedom from tests 2 and 3) in Program RELEASE (Burnham
et al., 1987).
Population abundance estimates were derived using the POPAN
formulation of the Jolly–Seber (JS) method (Arnason and Schwarz,
1995; Schwarz and Arnason, 1996). The following parameters
could be estimated from POPAN JS models: U (apparent survival),
p (recapture probability), PENT (probability of entry into the pop-
ulation at each occasion) and N (size of super-population, i.e. the
total number of individuals present within the population during
the entire study period). This JS model is assumed to be ‘open’,
and allows additions (births and immigration) and losses (deaths
and emigration) between successive sampling occasions (Schwarz
and Arnason, 1996).
A set of candidate 32 Jolly–Seber models were tested on the to-
tal data set (with each island coded as a individual group) of all
captured lizards where survival (U), recapture (p), parameters
were estimated by considering all combinations of time constant
(.) or variable (t) time and group (i.e. island) effects (models). The
PENT parameter was modelled as time variable or group by time
combination. The N parameter was modelled as function of group
(i.e. Ng). Popan produces two different estimates of population
abundance the super population estimate (N) and the derived an-
nual abundance parameter (Ng,t) for each island (g) at each annual
sampling period (t). In this study we refer to the latter as we only
wanted to use the final annual sampling estimate of abundance to
calculate total abundance for the four islands. We assessed good-
ness-of-fit of the global model and estimated overdispersion using
the median ð^ c ¼ 3:4Þ implemented in program Mark.
Total island and Komodo National Park Komodo dragon popula-
tion abundances were then estimated for each island using only
the 2011/2012 (i.e. final sampling period) derived estimate of an-
nual population abundance (plus SEM; upper and lower 95% CI).
To extrapolate to island-wide abundances we divided the 2011/
2012 estimates by the proportion of habitat sampled (i.e. the ratio
of the effective trapping area to the total dragon specific habitat on
each island). To estimate the effective trapping area for each island
we summed each site’s trapping grid area and inflated this with a
boundary layer to consider capture of individuals moving into the
trapping grid area from external habitat. The area of the boundary
layer was estimated as half the radius of the mean linear distance
between all individual recapture locations within a site’s trapping
grid (Krebs, 1999).
Next we estimated the total dragon specific habitat on each is-
land. We felt it prudent to produce an estimate that considered
variation in lizard habitat use. This was deemed necessary as our
published telemetry studies (e.g. Imansyah et al., 2008; Harlow
et al., 2010a,b) have clearly indicated that Komodo dragons do
not occupy all island habitats equally (i.e. density is not uniform
across habitats). To estimate the dragon specific habitat for each is-
land, we first calculated the area of each of the 5 major vegetation
communities using satellite imagery. We then adjusted (a penalty
between 0 [absence of dragons in habitat] and 1 [maximum pre-
ferred habitat] for each vegetation community area in accordance
to their respective lizard occupancy and use. The degree of habitat
penalisation was based on percentage frequency of respective hab-
itat obtained by radio telemetry of lizards (?2300 observations
from 53 individuals) and visual recordings of lizard-habitat use
associations collected during routine visual surveys conducted be-
tween 2002–2013 (e.g. Imansyah et al., 2008). Poorly utilised hab-
itat types included savannah grassland/savannah woodland,
mangrove forests, and quasi cloud forest. These habitats permit
some lizards to forage or transition through them, but are other-
wise not considered good quality resident habitat compared to
open deciduous forest and closed low altitude dense canopy forest
(<300 m elevation). As a consequence, by penalising unfavourable
habitats there was a considerable reduction in the total island area
(uncorrected = 533.6 km2versus corrected = 300.3 km2).
Finally we divided the abundance estimates by the proportion
of habitat sampled on each island and calculated an island-wide
Komodo dragon population abundance estimate; and by summing
four islands together we produced a total estimate for Komodo
dragons inhabiting Komodo National Park in 2011/2012.
3.1. Island specific population growth rate and survival estimates
From 2003–2012, we encountered 925 individual lizards during
2002 capture events. This data was analysed using a candidate set
of 16 Pradel lambda and survival models that were ranked by mod-
el fit criteria to estimate average population and survival for each
island across all sampling periods (Appendix A). Several models re-
ceived high model support and hence the derived estimates were
calculated using weighted model averaging techniques. We ad-
justed our population growth and survival parameter estimates
to accommodate overdispersion in the data set ð^ c ¼ 3:4Þ.
Time averaged population growth rates (k) derived from model
averaged estimates indicated clear differences among the four is-
land populations (Table 1). The two large populations on Komodo
(k = 0.97 ± 0.02; 95%CI = 0.84–1.00)
(k = 1.00 ± 0.03; 95% CI = 0.93–1.06) exhibited near-stable or stable
time averaged growth rates. The estimates of population growth
on the small island of Motang (k = 0.68 ± 0.09; 95% CI = 0.47–
0.83) indicated that this population was in decline. The estimate
for Kode (k = 0.97 ± 0.15; 95% CI = 0.16–1.00) was associated with
a large margin of error making it difficult to infer the status of pop-
ulation growth for this island population.
Differences in average apparent survival (U) derived from mod-
el averaged Pradel models were also evident among the four island
populations. Survival estimates were much higher for the Komodo
and Rinca Komodo dragon populations compared to the small is-
land Kode and Motang populations (Table 1).
3.2. Estimates of Komodo dragon population abundance
Abundance estimates were obtained from a candidate set of 32
open population Jolly–Seber models (POPAN) that were ranked by
model fit criteria to estimate a derived population abundance for
each island at each annual sampling period (Appendix B). Several
models received high model support and hence the derived
2011/2012 abundance estimates were calculated using weighted
Time averaged estimates of population growth rate (k) and survival (U) for the four
extant Komodo dragon populations inhabiting Komodo National Park. Standard error
of the mean (SEM) and the lower and upper 95% confidence intervals (CI) are
presented as measures of error.
? xk ? SEM
1.00 ± 0.03
0.96 ± 0.03
0.68 ± 0.1
0.97 ± 0.16
? xU ? SEM
0.72 ± 0.05
0.75 ± 0.05
0.54 ± 0.18
0.52 ± 0.23
D. Purwandana et al./Biological Conservation 171 (2014) 29–35
model averaging techniques. Derived annual abundance estimates
indicated clear differences among islands over the duration of the
study (Appendix C). The estimates of median ð^ c ¼ 3:4Þ was
1.39 ± 0.04 suggesting limited overdispersion, but nevertheless,
model output was adjusted to accommodate this value.
Site specific population abundances were then used to produce
extrapolated estimates of total island and National Park population
abundance. These estimates were considerate of habitat suitability
and hence only considered vegetation communities known to har-
bour resident Komodo dragons. These estimates indicated that the
total island population abundances were much greater on the large
islands compared to small islands due to higher densities as well as
their considerably larger habitat areas (Table 2). Summing the four
island population abundances produced a total population esti-
mate for Komodo National Park of 2448 ± 229 (95% CI: 2067–
2922) Komodo dragons (Table 2).
Using a large scale mark recapture study; we provide the first
spatiotemporal estimates of demographic rates for the Komodo
dragon sampled over 10 years. Our results indicated obvious
demographic differences between lizard populations on the two
large islands (i.e. Komodo and Rinca) and those on the two small
islands (i.e. Kode and Motang). Key differences included higher
and more stable population growth for the large island popula-
tions. These differences culminated in large islands having stable
population growth and higher survival. In contrast, the population
on the small Motang Island appeared to be experiencing a substan-
tial decline and lower survival relative to the large island popula-
tions. The other small island population inhabiting Kode Island
was also associated with broad confidence intervals around the
population growth and survival estimates making it less easy to
determine the status of this population. Such findings provide
important insights into the population status, prioritization and
ultimately how managers could better conserve Komodo dragons
in Komodo National Park.
A major challenge to management of often small island popula-
tions, especially those in decline, is to understand what processes
influence demography (Caughley, 1994). For example, the apparent
decline in the Motang population could be conceivably triggered
by one or more processes that arise due to its geographic isolation,
small island area and presumably historically small population size
(Jessop et al., 2007). For example, spatial or temporal variation in
food availability can induce dramatic population fluctuations on is-
lands as a result of altered survival, fecundity and recruitment
(Laurie and Brown, 1990; Grant et al., 2000). Demographic re-
sponses to food availability are often more pronounced in island
systems given relatively depauperate food webs limiting forage
choice. Similarly it is conceivable that inbreeding depression could
negatively influence survival of Komodo dragons particularly on
small isolated island populations (Frankham, 1998, 2005; Eldridge
et al., 1999). For example, inbreeding could decrease population
survival via higher rates of abnormal offspring or reduced clutch
sizes as observed in other small isolated populations (including
reptiles) (Madsen et al., 1996; Saccheri et al., 1998; Keller and
Weller, 2002). Inbreeding could also covary with other ecological
pressures to exacerbate mortality (Lande, 1988; O’grady et al.,
2004, 2006). Additionally an Allee effect may arise and contribute
to demographic stochasticity and inbreeding again increasing
extinction risk in small and especially declining island populations
(Courchamp et al., 2008).
Hence limited dispersal, simple trophic dynamics, inbreeding
and Allee effects could independently or additively decrease sur-
vival in small island Komodo dragon populations. Thus managers
may well need to identify and then rank these different processes
to best determine their relative threat and hence choose one or
more courses of action to best promote population recovery. This
could be done using a population viability analysis framework
(e.g. Beissinger, 2002). For example, if there was clear evidence
that low prey availability was responsible for poor survival then
prey supplementation would seem warranted. However if evi-
dence for inbreeding depression was found it would require trans-
location of individuals from genetically appropriate populations to
increase genetic diversity and effective population size.
Importantly, given the apparent strength of decline in the Mo-
tang population, we encourage conservation managers to rapidly
consider options for recovery of this population if this in fact
deemed necessary (e.g. under legislation). However, we also
acknowledge that consideration of recovery options for this popu-
lation might also be influenced by the relative impact of this pop-
ulation’s extirpation on the Park’s total Komodo dragon population
size which appears minimal (i.e. extirpation of the Motang popula-
tion represents ?2% of the park’s total population size); and that
costs of management actions to conserve this small population
could inadvertently compete for resources needed for other impor-
tant conservation activities including ranger patrolling that en-
sures high quality habitat is maintained across the park.
However, even if this were true, a ‘‘do nothing’’ rational would
be inappropriate if there was clear evidence for functional evolu-
tionary novelty in this population for which we currently do not
know. Examples of functional evolutionary novelty could include
identification of novel genes (e.g. conferring immune tolerance);
or evolution of different life-history attributes that could underpin
large body size variation observed among different island popula-
tions (Jessop et al., 2006). If present such attributes would be worth
conserving as they could maintain functional diversity important to
larger populations that will need to contend with dramatic climate
change consequences over the next century (Moritz, 2002).
We extrapolated site specific abundance to produce a total pop-
ulation abundance estimate for each island and by summing these
for Komodo National Park. We estimated that 2448 ± 229 (95% CI:
2067–2922) Komodo dragons were present in Komodo National
Park. This estimate is ecologically plausible, as it considers that
all islands possess varying areas of poorly utilised habitat due to
absence of prey (e.g. coastal mangrove forest or elevated quasi-
cloud forest) or adequate shade (e.g. savannah grassland) to permit
efficient thermoregulation (Harlow et al., 2010a,b).
Total island population abundance estimates for Komodo dragons in Komodo National Park in 2011–2012. Table reports data necessary to estimate total population abundance
drawn from island specific population abundances within island trapping areas then corrected to total island area after considering the availability of suitable lizard habitat.
Island (area km2)
?XN ? SEM Island
Corrected island area
for lizard habitat (km2)
% trapped area/
Extrapolated island (corrected)
N ± SEM (95% CI)
166.8 ± 13.5 (132–179)
168.2 ± 13.5 (159–212)
18.7 ± 6.0 (5–34)
12.1 ± 5.8 (3–22)
1166 ± 94.4 (923–1252)
1185 ± 95.1 (1119–1493)
44 ± 14.1 (12–80)
53 ± 25.6 (13–97)
Total National Park Abundance2448 ± 229 (2067–2922)
D. Purwandana et al./Biological Conservation 171 (2014) 29–35
This extrapolated abundance estimate is encouraging from a
population viability perspective because species population sizes
that exceed several thousand individuals are thought to confer
substantially greater evolutionary resilience to extinction (Reed,
2005; Traill et al., 2010). However, it is important to remember
for Komodo dragons, and indeed for many non-volant terrestrial is-
land species, that there is a strong likelihood that different island
populations do not function as metapopulations due to infrequent
dispersal and gene flow (Whittaker and Fernandez-Palacios, 2007;
Ciofi, 2002; Jessop et al., 2008). This means that in the advent that
all island are closed to dispersal and gene flow, mangers must con-
sider each island as independent demographic management units.
Equally important is that there now consideration for how to pri-
oritize conservation investment into these different island popula-
tions (Wilson et al., 2006). In the absence that either small island
population lacks unique functional attributes that could confer
broader evolutionary potential, then it would seem straightfor-
ward to prioritize the two largest islands that possess substantially
larger populations as a demographic based prioritization strategy
to ensure evolutionary resilience (Traill et al., 2010).
How do our results interact with current park management? At
present the park management focuses its conservation activities
via regular ranger patrols conducted primarily across the two large
islands of Komodo and Rinca. These patrols provide habitat secu-
rity and should be maintained indefinitely to help promote stable
lizard populations on large islands. However, as discussed above,
it will now be necessary for park managers to consider additional
strategies for how to manage small island populations. A central
tenant of future park managers will thus be contingent on deter-
mining the demographic and evolutionary value of small island
populations to the population viability of Komodo National Park
populations as a whole. This is far from a trivial task and will re-
quire the park to enhance technical capacity and collaborate with
other national and international research organisations to help ac-
quire the evidence necessary to make a well informed decision on
the management value of small island populations.
This is the first comprehensive and rigorous attempt to estimate
the demographic status of extant Komodo dragon populations
across the species’ distribution. These demographic estimates pro-
vided for Komodo National Parks are clearly essential for prioritis-
ing conservation actions within the largest protected area afforded
to extant lizard populations. Our results indicate clear differences
in the population demography of Komodo dragons within Komodo
National Park. Currently, Komodo National Park does not differen-
tiate among island populations from a management perspective,
but in doing so increases the risk of failing to recognise inherent is-
land specific differences necessary to optimise conservation and
management of Komodo dragons. Needless to say, given the in-
creased potential for genetic and demographic processes underpin-
ning small population extinctions on islands, managers of Komodo
National Park should consider ongoing annual demographic and
genetic monitoring of all island populations (Lande, 1988; Frankham,
1998; Eldridge et al., 1999). Such information is essential to guide
what strategies may be best suited to attempt to reduce population
declines across the range of this species.
We thank Komodo National Park staff and volunteers who as-
sisted us in the fieldwork.
Major funding for this study (2002–2006) was awarded to T.S.J.
via a Conservation Research Postdoctoral Fellowship from the Zoo-
logical Society of San Diego. Later funding support (2007 onwards)
was provided by the Komodo Species Survival Plan of the American
Zoo and Aquarium Association and is gratefully acknowledged.
This research was conducted via Memorandum of Understanding
(MOU) between the Zoological Society of San Diego, and the Indo-
nesian Department of Forestry and Conservation (PHKA) or via an
MOU between Komodo Survival Program and Komodo National
Park. We thank the two anonymous reviewers for their construc-
tive and insightful feedback that helped improve the quality of
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biocon.2014.
Anderson, D.R., Burnham, K.P., White, G.C., 1994. AIC model selection in
overdispersed capture–recapture data. Ecology 75, 1780–1793.
Arnason, A.N., Schwarz, C.J., 1995. POPAN-4: enhancements to a system for the
analysis of mark–recapture data from open populations. J. Appl. Stat. (22), 785–
Beissinger, S.R., 2002. Population Viability Analysis. University of Chicago Press,
Burnham, K.P., Anderson, D.R., White, G.C., Brownie, C., Pollock, P.H., 1987. Design
and Analysis Methods for Fish Survival Experiments Based on Release–
Recapture. American Fisheries Society Monograph No. 5, Bethesda, 737 p.
Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: A
Practical Information-Theoretic Approach,
Cardillo, M., Purvis, A., Sechrest, W., Gittleman, J.L., Bielby, J., Mace, G.M., 2004.
Human population density and extinction risk in the world’s carnivores. PLoS
Biol. 2 (7), e197.
Caughley, G., 1994. Directions in conservation biology. J. Anim. Ecol. 63, 215–244.
Ciofi, C., 2002. Conservation genetics. In: Murphy, J.B., Ciofi, C., de la Panouse, C.,
Walsh, T. (Eds.), Komodo Dragon. Smithsonian Institution Press, Washington,
Ciofi, C., de Boer, M.E., 2004. Distribution and conservation of the Komodo Monitor
(Varanus komodoensis). Herpetol. J. 14, 99–107.
Courchamp, F., Berec, L., Gascoigne, J., 2008. Allee effects in ecology and
conservation. Environ. Conserv. 36, 80–85.
Eldridge, M.D.B., King, J.M., Loupis, A.K., 1999. Unprecedented low levels of genetic
variation and inbreeding depression in an island population of the black-footed
rock-wallaby. Conserv. Biol. 13, 531–541.
Frankham, R., 1998. Inbreeding and extinction: island populations. Conserv. Biol.
Frankham, R., 2005. Genetics and extinction. Biol. Conserv. 126, 131–140.
Fordham, D.A., Brook, B.W., 2010. Why tropical island endemics are acutely
susceptible to global change. Biodivers. Conserv. 19, 329–342.
Grant, P.R., Grant, B.R., Keller, L.F., Petren, K., 2000. Effects of El Ni~ no events on
Darwin’s finch productivity. Ecology 81, 2442–2457.
Harlow, H.J., Purwandana, D., Jessop, T.S., Phillips, J.A., 2010a. Size-related
differences in the thermoregulatory habits of free-ranging Komodo dragons.
Int. J. Zool. 2010, 1–9.
Harlow, H.J., Purwandana, D., Jessop, T.S., Phillips, J.A., 2010b. Body temperature and
thermoregulation of Komodo dragons in the field. J. Therm. Biol. 35 (7), 338–
Harris, J.B.C., Sekercioglu, C.H., Sodhi, N.S., Fordham, D.A., Paton, D.C., Brook, B.W.,
2011. The tropical frontier in avian climate impact research. Ibis 153, 877–882.
Imansyah, M.J., Jessop, T.S., Ciofi, C., Akbar, Z., 2008. Ontogenetic differences in the
spatial ecology of immature Komodo dragons. J. Zool. 274 (2), 107–115.
IUCN, 2012. IUCN Red List of Threatened Species, Version 2012.1.
Jessop, T.S., Madsen, T., Ciofi, C., Imansyah, J., Purwandana, D., Rudiharto, H.,
Ariefiandy, A., Phillips, J.A., 2007. Differences in population size structure and
body condition: conservation implications for Komodo Dragons. Biol. Conserv.
Jessop, T.S., Madsen, T., Sumner, J., Rudiharto, H., Phillips, J.A., Ciofi, C., 2006.
Maximum body size among insular Komodo dragon populations covaries with
large prey density. Oikos 112, 422–429.
Jessop, T., Sahu, A., Opat, D., 2008. Varanus komodoensis (Komodo Dragon)
occurrence. Herpetol. Rev. 39 (1), 95–96.
Jessop, T.S., Sumner, J., Rudiharto, H., Purwandana, D., Imansyah, M.J., Phillips, J.A.,
2004. Distribution, use and selection of nest type by Komodo Dragons. Biol.
Conserv. 117 (5), 463–470.
Keller, L.F., Weller, D.M., 2002. Inbreeding effects in wild populations. Trends Ecol.
Evol. 17, 230–241.
Krebs, C.J., 1999. Ecological Methodology. Harper & Row, Publishers, New York, USA.
second ed. Springer-Verlag, New
D. Purwandana et al./Biological Conservation 171 (2014) 29–35
Lande, R., 1988. Genetics and demography in biological conservation. Science 241,
Laurie, W.A., Brown, D., 1990. Population biology of marine iguanas (Amblyrhynchus
cristatus): changes in the annual survival rates and the effects of size, sex, age
and fecundity in a population crash. J. Anim. Ecol. 59, 529–544.
Laver, R.J., Purwandana, D., Ariefiandy, A., Imansyah, J., Forsyth, D., Ciofi, C., Jessop,
T.S., 2012. Life-history and spatial determinants of somatic growth dynamics in
Komodo dragon populations. PLoS ONE 7 (9), e45398.
Madsen, T., Stille, B., Shine, R., 1996. Inbreeding depression in an isolated population
of adders Vipera berus. Biol. Conserv. 75, 113–118.
McKinnon, K., 1996. The Ecology of Kalimantan, vol. 3. Periplus Editions, Hong Kong.
Mittermeier, R.A., Myers, N., Thomsen, J.B., Da Fonseca, G.A.B., Olivieri, S., 1998.
Biodiversity hotspots and major tropical wilderness areas: approaches to
setting conservation priorities. Conserv. Biol. 12, 516–520.
Moritz, C., 2002. Strategies to protect biological diversity and the evolutionary
processes that sustain it. Syst. Biol. 51, 238–254.
O’Grady, J.J., Brook, B.W., Reed, D.H., Ballou, D.J., Tonkyn, D.W., Frankham, R., 2006.
Realistic levels of inbreeding depression strongly affect extinction risk in wild
populations. Biol. Conserv. 133, 42–51.
O’Grady, J.J., Reed, D.H., Brook, B.W., Frankham, R., 2004. What are the best
correlates of extinction risk. Biol. Conserv. 118, 513–520.
Reed, D.H., 2005. Relationship between population size and fitness. Conserv. Biol.
Pradel, R., 1996. Utilization of capture–mark–recapture for the study of recruitment
and population growth rate. Biometrics 52, 703–709.
Saccheri, I., Kuussaari, M., Kankare, M., 1998. Inbreeding and extinction in a
butterfly metapopulation. Nature 392, 491–494.
Schwarz, C.J., Arnason, A.N., 1996. A general methodology for the analysis of open-
model capture recapture experiments. Biometrics 52, 860–873.
Sibly, R.M., Hone, J., 2002. Population growth rate and its determinants: an
overview. Philos. Trans. Roy. Soc. Lond. B Biol. Sci. 357 (1425), 1153–1170.
Sodhi, N.S., Koh, L.P., Brook, B.W., Ng, P.K.L., 2004. Southeast Asian biodiversity: the
impending disaster. Trends Ecol. Evol. 19, 654–660.
Traill, L.W., Brook, B.W., Frankham, R.R., Bradshaw, C.J., 2010. Pragmatic population
viability targets in a rapidly changing world. Biol. Conserv. 143 (1), 28–34.
White, G.C., Burnham, K.P., 1999. Program MARK: survival estimation from
populations of marked animals. Bird Study 46 (S1), S120–S139.
Whittaker, R.J., Fernandez-Palacios, J.M., 2007. Island Biogeography: Ecology,
Evolution and Conservation. Oxford University Press, Oxford, UK.
Whitten, T., 2000. The Ecology of Sumatra, vol. 1. Periplus Edition, Jakarta,
Wilson, K.A., Mcbride, M.F., Bode, M., Possingham, H.P., 2006. Prioritizing global
conservation efforts. Nature 440, 337–340.
D. Purwandana et al./Biological Conservation 171 (2014) 29–35