This article was originally published in a journal published by Elsevier, and the attached copy is provided by Elsevier for the author’s benefit and for the benefit of the author’s institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues that you know, and providing a copy to your institution’s administrator. All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier’s permissions site at: http://www.elsevier.com/locate/permissionusematerial
B I O L O G I CA L C O N S E RVAT I O N
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biocon
co
py
Island differences in population size structure and catch per unit effort and their conservation implications for Komodo dragons Tim S. Jessopa,b,*, Thomas Madsenc, Claudio Ciofie, M. Jeri Imansyaha, Deni Purwandanaa, Heru Rudihartod, Achmad Arifiandya, John A. Phillipsa a
Conservation and Research of Endangered Species, Zoological Society of San Diego, Escondido, CA 92027, USA Department of Wildlife Conservation and Research, Zoos Victoria, P.O. Box 74, Elliot Avenue, Parkville, Vic. 3052, Australia c Department of Biological Sciences, University of Wollongong, Wollongong, NSW 2522, Australia d Taman National Komodo, Labuan Bajo, Flores, NTT, Indonesia e Department of Animal Biology and Genetics, University of Florence, Via Romana 17, 50125 Florence, Italy
on
al
b
A B S T R A C T
Article history:
Species inhabiting archipelagos are often characterised by high levels of interpopulation
Received 1 June 2006
divergence (e.g. size related traits). This divergence may, in turn, influence their life-history.
Received in revised form
To facilitate better management and conservation of the Komodo dragon (Varanus komodo-
12 October 2006
ensis), an island endemic, we identified demographic differences between two island pop-
Accepted 15 October 2006
ulations in Komodo National Park, Indonesia. Comparison of data collected from dragon
Available online 4 December 2006
populations inhabiting Rinca Island and the much smaller Gili Motang Island indicated that
pe
rs
A R T I C L E I N F O
between 1994 and 2004, the Komodo dragon population on Gili Motang significantly decreased its: (1) mean body mass, (2) body condition and (3) relative abundance. These
Varanus komodoensis
results suggest that the numerically small Gili Motang population was oscillating down-
Komodo dragon
wards; in contrast, the Rinca Island population had been relatively stable. More importantly
Islands
these results emphasize the necessity for managers of this priority conservation species to
Population divergence
r's
Keywords:
understand further the inherent functional differences among dragon populations to
Conservation implications
develop island specific management units. Current management practices (e.g. monitortions and thus run the risk of being unable to detect adverse effects for populations that are potentially most prone to decline.
Introduction
Au
1.
th o
ing) instigated by Komodo National Park management ignore small island dragon popula-
Species inhabiting archipelagos are often distributed on multiple and variable island land masses exhibiting heterogeneity in environment and community structure (MacArthur and Wilson, 1967). This environmental variation, along with the relative degrees of isolation facilitates evolutionary and ecological processes that are responsible for often significant lev-
2006 Elsevier Ltd. All rights reserved.
els of population divergence leading to both increased endemism and biodiversity associated with archipelagos (Grant, 1998). However, insular species are particularly susceptible to threatening processes including habitat loss, harvesting and invasive species because they are isolated and occur on smaller land masses (Reid and Miller, 1989; Burkey, 1995). In addition the demographic and genetic quality of island populations often facilitates their susceptibility to threats (Frank-
* Corresponding author: Address: Department of Wildlife Conservation and Research, Zoos Victoria, P.O. Box 74, Elliot Avenue, Parkville, Vic. 3052, Australia. Tel.: +61 3 92859387; fax: +61 3 92859360. E-mail address:
[email protected] (T.S. Jessop). 0006-3207/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2006.10.025
248
B I O L O G I C A L C O N S E RVAT I O N
on
al
co
py
measures. By adopting a fisheries type approach used in stock assessment, we evaluated changes in morphological and catch per unit effort (from trapping) to make inferences about the temporal dynamics of Komodo dragon populations inhabiting the islands of Gili Motang and Rinca. We compared data collected from this population in 1994 to complementary data collected annually in 2002, 2003 and 2004. Specifically, we assessed whether the Gili Motang and Rinca populations could be described as stable, or oscillating, and whether they exhibited concordance in their dynamics over the time period between 1994 and 2004 by assessing differences in five parameters—(a) size frequency distributions, (b) mean snout–vent length (SVL) and mass, (c) body condition, (d) relative population abundance derived from catch per unit effort data and (e) abundance estimates of the Gili Motang population in 2004. Frequency differences in body size distributions over time could indicate relative changes in survivorship, recruitment and migration. Temporal changes in body condition might indicate the populations health influenced by such factors as food availability, disease or parasitism. Differences in catch per unit effort could provide an indication whether the population numbers are rising, falling or stable. An estimate of population abundance, on a small isolated island, could provide a rough gauge for assessing the theoretical extinction proneness of the Gili Motang dragon population. Collectively, temporal differences in these parameters between islands could provide the first insights into potential differences in the dynamics and thus status of populations of the Komodo dragon populations, a species of unparalelled national status and high conservation value in Indonesia.
Au
th o
r's
pe
rs
ham, 1998). Combined, such processes have led to, or are expected to contribute to, much greater rates of extirpation or extinction in island vertebrates compared to continental forms (Diamond, 1989; Case et al., 1992; Cardillo et al., 2006). Consequently managing island endemics for conservation is a high priority particularly in tropical archipelagos where both diversity and threatening processes are greatest. However, even within a single endemic, which in archipelagos can be represented by disjunct populations distributed across multiple islands, conservation can be complicated by divergence in organismal traits that can lead to different impacts (i.e. from threatening processes) among populations. For example, when exposed to natural perturbations such as El Nino, which can dramatically reduce food availability, birds and lizards occupying different islands in the Galapagos archipelago may vary in their susceptibility to mortality and the degree to which their independent populations decline (Wikelski and Trillmich, 1997; Grant, 1998). In essence, this variation in a population’s decline reflects respective differences in life-history traits underpinning population dynamic processes. Subsequently, the capacity to detect intraspecific differences in population dynamics would seem essential to conserving island endemics and particularly necessary for delivering demographic based wildlife-management strategies to mitigate specific threatening processes (Caswell et al., 1999). Despite the application of population dynamic studies for guiding the conservation and management of species, there are few comparisons between multiple populations of the same species living among islands (Coulson et al., 2005), especially for species of high conservation value. The Komodo dragon (Varanus komodoensis) is a large and robust species of monitor lizard endemic to five islands in south-east Indonesia (Auffenberg, 1981; Ciofi and De Boer, 2004). At present, data pertaining to the demography of this species is largely absent (Auffenberg, 1981; Jessop et al., 2004). Variation in population dynamics, manifested as temporal differences in survival and fecundity among island populations could be promoted in response to differences in island-specific evolutionary and ecological processes. Across its restricted range, this large predator maintains populations on islands ranging in area from approximately 10 km2 up to 6000 km2 (Ciofi and Bruford, 1999). Existing studies indicate that there are significant differences among populations with respect to population genetics and morphology, particularly between large and small island populations, reflecting dispersal capacity and resource dynamics (Ciofi and Bruford, 1999; Jessop et al., 2006). However, despite these examples of interpopulation divergence, there is little functional evidence for how the populations vary demographically over time, a crucial indicator of overall population status. In this preliminary study we investigated the potential for insular differences in the temporal population dynamics of dragons in Komodo National Park. This World Heritage protected area contains four extant populations distributed across two large and two small islands. Using a comparison between one small island (Gili Motang) and one large island (Rinca) we examined demographic differences in these two populations over a 10 year period. If insular differences exist, then management of insular meta-populations of V. komodoensis will need to consider island (i.e. population) specific conservation
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
2.
Methods
This study was undertaken on Komodo dragon populations on the islands of Rinca (278.0 km2) and Gili Motang (10.3 km2) in Komodo National Park in the Lesser Sunda region of Southeastern Indonesia (Fig. 1). Within Rinca, four study sites were selected and subsequently combined to represent a total island sample in which to assess inter-island population variation and included the valleys of Loh Buaya, Loh Baru, Loh Tongker and Loh Dasami (Fig. 1). On the small mountainous island of Gili Motang, field work was confined to the coastal flats and adjacent hills representing approximately 20% of available island habitat. The primary study area on this island consisted of a triangular shaped wedge, 2.1 km2 in area, on the north-west side of the island. Outside this core study area a number of randomly positioned trapping sites were used to verify that our trapping area provided a representative population sample. Field work for this study consisted of four trapping sessions (8–14 days duration per site); first in 1994 as part of a population genetic study, and then again annually from 2002 through to 2004 during routine population monitoring within Komodo National Park. All field work was conducted in August/September during the mid dry season to standardise capture protocols. Trapping sites in the 2002– 2004 period incorporated those sites used in the 1994 sampling period providing an overlapping sampling regimen. Komodo dragons were captured in 300 cm · 50 cm · 50 cm long box traps (N = 4–8 traps) baited with goat meat (0.5 kg)
249
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
on
al
co
py
B I O L O G I C A L C O N S E RVAT I O N
rs
Fig. 1 – Maps of Rinca (a) and Gili Motang (b) in relation to the other major islands in Komodo National Park (Inset). Numbers 1–4 refer to the field sites of Loh Buaya, Loh Baru, Loh Tongkir and Loh Dasami, respectively on Rinca Island and number 5 to the field site on Gili Motang.
Au
th o
r's
pe
according to methods outlined in Jessop et al. (2006). Following capture, dragons were restrained with rope and their mouths taped shut. Several morphological characters, including head length, and snout to vent length (SVL) were measured using calipers and a fiberglass tape. Body mass was obtained using digital scales. After 1994, dragons were permanently marked using passive integrated transponders (Trovan ID100). To prevent violation of statistical independence, only the first capture is used of each dragon captured repetitively among the 2002, 2003 and 2004. Variations in frequency distribution of SVL and mass between sampling periods are presented as measures of skewness (i.e. g1; refer to formula 6.5 in Zar, 1999). We used Kolmogorov–Smirnov tests to compare for significant differences between frequencies in population size structure at sampling periods. Temporal differences in the mean SVL and body mass for both populations were analyzed by parametric t-test (data ln transformed to eliminate violations of normality). A comparative measure of body condition was calculated using the annual regression equation of mass (ln transformed) plotted against SVL (ln transformed). Analysis of covariance (ANCOVA) analysis was used to explore variation in body condition between 1994 and 2002–2004. Given that we only obtained a small population sample from Gili Motang in 1994 (N = 12) there is a risk that our results could be confounded, as by chance, this small sample size could be unrepresentative of the real population. To increase the validity of our results we used bootstrapping techniques (1000 resampling iterations with replacement; Efron and Tibshirani, 1993) to produce a pseudo-population sample for SVL and mass from the 1994 Gili Motang data. We com-
pared these data sets with their real population samples to determine if the respective populations means were statistically similar using t-tests. Another issue with the small sample collected in 1994 was reduced statistical power to infer significant temporal differences in frequency distributions in population size structure, which in the context of this study, is a more valuable indicator of population change than estimates based on sample means. Again, we used bootstrapping techniques to produce pseudo-population data sets to provide a basis for assessing temporal differences (i.e. 1994 versus 2002–2004) in the population size structure of Gili Motang dragons. To assess temporal variation in population numbers, an index of population size based on catch per unit effort was calculated by dividing the daily catch of new dragons (including recaptures from previous years) by the number of open traps. Daily catch per unit effort was summed and divided by the number of field days for each field trip to provide a mean annual estimate of catch per unit effort for the years—1994, 2002, 2003 and 2004. One of the well documented limitations in using catch per unit effort as a proxy of true abundance is its tendency to result in population estimates that remain high (i.e. hyperstability) while actual abundance is declining (Harley et al., 2001). Thus this method may overestimate true abundance and reduce the capacity to detect both presence and severity of a population decline. Nevertheless, this type of data is still able to detect correlated changes in abundance and could provide a preliminary basis for detecting relative trends in abundance between islands. The Jolly-Seber method was used to determine a basic estimate of population size (N) (using the software Ecological
250
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
B I O L O G I C A L C O N S E RVAT I O N
Gili Motang: one-way ANOVA; F2,36 = 1.24; P = 0.30; Rinca: F2,126 = 1.78; P = 0.40; mass: Gili Motang: F2,36 = 1.35; P = 0.27; Rinca: F2,126 = 1.66 P = 0.37). The low population number on Gili Motang, resulted in small annual sample sizes. Data was therefore pooled across years in order to increase the statistical power of the subsequent analyses. In 1994 the Gili Motang population’s SVL distribution was negatively skewed (g1 = 0.14 ± 0.64; Fig. 2a) whereas in 2002–2004 it was strongly positively skewed (g1 = 0.66 ± 0.38). A similar trend of increasing positive skew was recorded for mass (Fig. 2b); with the 2002–2004 data (g1 = 1.29 ± 0.38) being more positively skewed than the mass data distribution observed in 1994 (g1 = 0.75 ± 0.64). The increase in positive skew indicates that in 2002–2004 the Gili Motang population had undergone a frequency shift towards proportionally smaller and lighter animals. However, these temporal frequency dif-
Results
3.1.
Change in size related frequency distributions
co
3.
py
methodology version 2) within the Gili Motang study area in 2004. To estimate the effective size of the area trapped and compensate for an ‘‘edge effect’’ (movement of individuals outside of the core area) a boundary strip was included around the trapping area (Krebs, 1999). Once the boundary area was included, the effective trapping area represented 3.5 km2 (34% of the island area). Total island population size was estimated, by multiplying the numbers captured within the effecting trapping area with the area of the entire island.
We did not detect any significant among-year variation (2002, 2003 and 2004) in SVL and mass on Gili Motang or Rinca (SVL:
25
20
Frequency %
12 10
15
on
8 6
10
5
4
rs
2
0
0 50 55 60 65 70 75 80 85 90 95 100 105110115 120
50 45 40 35 30 25
r's
20 15 10 5 0
60 70 80 90 100 110 120 130 140 150
SVL cm
35 30 25 20 15 10 5 0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
th o
30 40 50
40
Frequency %
pe
SVL cm
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
60
40
100 20
SVL mass 15
*
20
0
1994
2002-2004
Time Period
SVL mass
80
10
20
15 SVL cm
SVL cm
80
Mass kg
Mass kg
100
Au
Mass kg
60
10
Mass kg
Frequency %
14
al
20
16
Frequency %
1994 (n = 32) 2002-2004 (n = 128)
1994 (n = 12) 2002-2004 (n =39)
18
40
5
20
0
0
5
0 1994
2002-2004
Time Period
Fig. 2 – A comparison of frequency distributions in snout–vent length (SVL- top row), body mass (middle row) and the mean population SVL and body mass (bottom row) of the Gili Motang (a–c) and Rinca (d–f) Komodo dragon population sampled in 1994 and 2002–2004. Error bars represent the standard error of the mean.
B I O L O G I C A L C O N S E RVAT I O N
ferences in the population were not significantly different (although close) for either SVL (Kolmogorov–Smirnov test: D = 0.351, P = 0.12) or mass (Kolmogorov–Smirnov test: D = 0.375, P = 0.094). The same analyses conducted on the Rinca population revealed that the 1994 population’s SVL distribution data was strongly positively skewed (g1 = 1.88 ± 0.41; Fig. 2d) whereas 2002–2004 was less so (g1 = 0.76 ± 0.41). A similar trend of decreasing positive skew was recorded for mass (Fig. 2e); with the 2002–2004 data (g1 = 0.313 ± 0.21) being less positively skewed than the mass data frequency distribution observed in 1994 (g1 = 1.155 ± 0.21). The decrease in skew suggests there was an increased frequency of larger and heavier animals in the Rinca population between the two sampling periods. These temporal differences in the Rinca population size structure were not significantly different for either SVL (Kolmogorov–Smirnov test: D = 0.172, P = 0.401) or mass (Kolmogorov–Smirnov test: D = 0.133, P = 0.726).
population has undergone changes in its population size structure between 1994 and 2002–2004.
3.4.
co
py
We examined if the 2002–2004 Gili Motang and Rinca populations exhibited any difference in body condition compared to 1994. To permit data pooling among years, we first checked that data did not violate the homogeneity of regression slopes (Gili Motang: F2,37 = 1.375, P = 0.267; Rinca: F2,126 = 0.999, P = 0.42) and that there was no significant annual difference among slopes (Gili Motang: ANCOVA, F2,37 = 0.155, P = 0.86; Rinca: F2,126 = 0.799, P = 0.75). Subsequent comparison of body condition between the 1994 and 2002–2004 on Gili Motang revealed a highly significant difference between the intercepts of the two samples (ANCOVA F1,49 = 12.64, P < 0.001; test for homogeneity of slopes F1,49 = 1.375, P = 0.267; Fig. 3) strongly suggesting that dragons, independent of their size, were significantly thinner in 2000–2004 compared to 1994. In contrast there was no difference in body condition between sampling periods in the Rinca population (ANCOVA F1,156 = 0.728,
al
Mean differences in SVL and body mass
4
3.3.
Inferences from bootstrapped data
ln mass kg
pe
rs
The mean SVL of the Gili Motang population decreased from 85.52 ± 5.70 cm in 1994, to 75.68 ± 2.46 in 2002–2004 (Fig. 2c) (a reduction in SVL by 12%; t-test = t1,49 = 1.81; P = 0.077). The mean body mass also decreased from 15.17 ± 3.13 kg in 1994, to 8.14 ± 0.86 kg in 2002–2004 (a reduction by 46%; t-test = t1,49 = 2.34; P = 0.021). In contrast the mean SVL of the Rinca did not change between the 1994 and 2002–2004, 82.00 ± 4.85 cm in 1994, and 85.68 ± 2.79 in 2002–2004 (t-test = t1,158 = 0.59; P = 0.556; Fig. 2f). Similarly, we did not detect any change in mean body mass between the two sampling periods; 16.24 ± 3.14 kg in 1994, and18.88 ± 1.70 kg in 2002–2004 (t-test = t1,158 = 0.71; P = 0.438).
Differences in body condition
on
3.2.
251
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
3
2
1 Gili Motang 1994 Gili Motang 2002-2004
0 3.8
4.0
4.2
4.4
4.6
4.8
5.0
ln SVL cm 5 4 3
l n m a ss k g
Au
th o
r's
To determine the robustness of our results, given that 1994 Gili Motang data was taken from a small sample (N = 12), we bootstrapped the data (using 1000 resampling iterations with replacement) and determine if the means were significantly similar between and real- and pseudo-population data sets. For the sample taken in 1994 there was no significant differences between the means estimated from real or bootstrapped data for SVL (real mean = 85.52 ± 5.70 cm versus bootstrapped mean = 85.82 ± 0.17 cm; t-test; t = 0.19, P = 0.85) or mass (real mean = 15.17 ± 3.13 versus bootstrapped mean 15.17 ± 0.09 kg; t-test; t = 0.01, P = 0.98). This congruence between real and bootstrapped means suggested that the small sample collected on Gili Motang in 1994 was likely to be representative of a non-biased population sample. We again used bootstrapped data produced from the 1994 and 2002–2004 Gili Motang samples to compare temporal frequency differences in the in SVL and mass. Using this data it was evident that there were highly significant differences in the frequency distributions for SVL (Kolmogorov–Smirnov test: D = 0.847, P < 0.001) and mass (Kolmogorov–Smirnov test: D = 0.8955, P < 0.001) between the 1994 and 2002–2004 sampling periods. These tests on bootstrapped data confirmed the general trends from the real data that the Gili Motang
2 1 0 Rinca 1994 Rinca 2002-2004
-1 -2 3
4
5
6
ln SVL cm Fig. 3 – A comparison of the mean population body condition represented by relationships between mass (ln transformed) versus SVL (ln transformed) of Gili Motang (a) and the Rinca populations (b) sampled in 1994 and 2002– 2004.
252
B I O L O G I C A L C O N S E RVAT I O N
0.8
0.6
0.2
py
0.4 Gili Motang Rinca
0.0 1994
2002
2003
2004
Year
rs
Relative population abundance
on
P = 0.395; test for homogeneity of slopes F1,156 = 0.563, P = 0.454; Fig. 3).
al
Fig. 4 – A comparison between the catch per unit effort for Komodo dragons trapped on Gili Motang and Rinca at four sampling periods over a 10 year period between 1994 and 2004.
3.5.
(Fig. 2a and b). Despite the small sample size collected in 1994, these changes approached significance (i.e. P 0.1), and when bootstrapped these frequency differences in morphological data were significantly different. Furthermore, between 1994 and 2002–2004, the mean body mass of the dragons has declined by 46.3%, and body condition has also decreased between sampling periods (Figs. 2c and 3). Catch per unit effort estimates of abundance declined by 68.33% between 1994 and 2002–2004 (Fig. 4). Together these results suggest that the Gili Motang population has been demographically unstable over this 10-year period, and is declining. In contrast, the Rinca population has been stable over this period with only minor changes in population parameters. The differences in population variation between these two islands could potentially arise due to various island specific density dependent and/or environmental processes that regulate the dynamics of populations (Coulson et al., 2000). Given the downward trend in catch per unit effort, decreased body size and an overall reduction in body condition, food availability, disease or inbreeding depression could be highly ranked among possible factors underpinning these changes to the Gili Motang population. In particular, temporal differences in food availability, more often than disease, are known to influence population characteristics and ultimately abundance in reptiles (Laurie and Brown, 1990; Andrews, 1991; Wikelski and Trillmich, 1997; Madsen and Shine, 2000; Madsen et al., 2006). Gili Motang island exhibits the lowest density of large prey among the four populations within Komodo National Park (Jessop et al., 2006). At present we have little recent evidence to suggest that anthropic threats such as deer hunting are responsible for the low ungulate density on this island. Energetic constraints on a population can lead to increased mortality, reduced recruitment, emigration and ultimately decreased abundance (Laurie and Brown, 1990; Preen and Marsh, 1995; Rhind and Bradley, 2002; Amar et al., 2003; Madsen et al., 2006). If greater temporal variation in food availability (i.e. prey), compared to Rinca, is the process responsible for the changes in the Gili Motang Komodo dragon population it could explain several specific features of our results. For example, a decrease in large prey could explain the decrease in the proportion of large Komodo dragons in the population. If large prey, like deer, are at low density, large dragons could be challenged with insufficient prey to meet energetic requirements and consequently mortality rates may be increased through starvation. In addition, because absolute energetic requirements scale with body size, heavier dragons (up to 35.0 kg in 1994 compared to 24.0 kg in 2004) would require greater absolute energetic requirements to survive on this small isolated island relative to smaller individuals. In other vertebrates, food limitations have been implicated in increased male mortality, because males often have a larger body sizes, lower fat reserves and greater nutritional and energetic requirements (Laurie and Brown, 1990). For example, in marine iguana (Amblyrhynchus cristatus) populations subjected to food limitations, induced by severe El Nin˜o southern oscillation events, large males were subject to the biggest loss in body condition and the highest mortality relative to other size classes (Wikelski and Trillmich, 1997). However, to verify our prediction that
co
Mean annual CPUE/day
1.0
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
3.6.
th o
r's
pe
For the Gili Motang population catch per unit effort in 1994, 2002, 2003 and 2004, was 0.60 ± 0.06, 0.29 ± 0.07, 0.23 ± 0.05 and 0.19 ± 0.04 dragons/trap-day, respectively. This represents a 68.3%, decrease in catch per unit effort between 1994 and 2004, and a significant decline over time (ANOVA, F3,34 = 7.47; P = 0.001; Fig. 4). Post-hoc analysis (Tukey’s test) indicated that there was a significant difference (P < 0.05) between the catch per unit effort measured in 1994 with that measured in 2002, 2003 and 2004 that comprised a homogenous sub-group. In contrast we did not observe any difference in catch per unit effort for Rinca population in 1994, 2002, 2003, and 2004, which was estimated at 0.81 ± 0.13, 0.67 ± 0.19, 0.72 ± 0.09 and 0.75 ± 0.06 dragons/trap-day, respectively (ANOVA, F3,93 = 0.16; P = 0.913).
A population estimate for Gili Motang
4.
Au
Mark recapture data resulted in a population estimate of 53.2 ± 23.7 within the Gili Motang effective trapping site of 3.5 km2. When extrapolated to the entire island the number of Komodo dragons in 2004 was estimated to 156.6 ± 67.7.
Discussion
Compared to the large island of Rinca, the Komodo dragon population on the small island of Gili Motang exhibited significant difference in both individual and population level parameters between 1994 and 2002–2004. Temporal decreases in the mean mass and body size (i.e. SVL) suggested that this population had altered its population size structure. In particular frequency differences over time suggested that the Gili Motang Population had lost its largest and heaviest individuals resulting in differences in population size structure
B I O L O G I C A L C O N S E RVAT I O N
on
al
co
py
(Purvis et al., 2000). In particular their large body size, consistent with a slow life-history, presumably means their capacity to rebound from a severe populations decline will require considerable time. Do managers of KNP currently need to intervene and manipulate conditions to enhance the Gili Motang Komodo dragon population? One difficulty in answering this question is that without prior knowledge of the population dynamics on this island it is impossible to gauge whether our current results (i.e. decreased catch per unit effort and a single estimate of population abundance) are indicative of a temporary downward oscillation or a progressive decline in population abundance. Also given its small area and approximately 10,000 years of isolation from other insular populations (Chappel and Shackleton, 1986; McCulloch et al., 1999), densities on Gili Motang are likely to have always been less (at least based on biogeographic theory) than the larger islands where prey resources are significantly greater. Thus a lack of long term data and the expected potential for a very different prey resource dynamic on this small island due to biogeographical differences makes it difficult to calibrate the current risk of extirpation for this population based on the decline inferred from the catch per unit effort data. Nevertheless, the implementation of long term population monitoring would ensure that managers have robust data to address population trends. If the downward trend continues and the population estimate fell to about 100 individuals, we encourage park management to instigate manipulative conservation strategies to begin recovery actions on Gili Motang. In the first instance, supplemental feeding by releasing of large natural prey (i.e. Timor deer) could be initial options to facilitate population recovery. If such measures failed to increase the dragon population, and evidence of inbreeding depression is recorded, in which case translocation of dragons from other island populations that are the most genetically and ecologically analogous to Gili Motang may be another management option to be considered as part of the recovery process (Crandall et al., 2000; Madsen et al., 1999; Wolf et al., 1996; Nelson et al., 2002). Our preliminary results suggest insular variation in the population dynamics of Komodo dragons within Komodo National Park. Previous research has indicated genetic and morphological differences among populations (Ciofi and Bruford, 1999; Jessop et al., 2006). Together these results are all factors that could represent a basis for designating different management units as they provide data to suggest evolutionary and/ or ecological differences among populations (Crandall et al., 2000). At present, Komodo National Park does not differentiate among island populations from a management perspective, but in doing so runs the risk of failing to recognise inherent differences necessary to optimise conservation and management of Komodo dragons. With respect to monitoring the status of extant populations, the Park currently focuses effort on the two large islands of Komodo and Rinca, but not the two small islands (Fig. 1). Needless to say, given the increased potential for genetic and demographic processes underpinning small population extinctions on islands, managers of Komodo National Park should invest in sustained annual monitoring of their small island populations (Lande, 1988; Frankham, 1998; Eldridge et al., 1999). This will enable
Au
th o
r's
pe
rs
the temporal variation in prey abundance is a potential mechanism responsible for the observed changes in the Gili Motang dragon population considerable long term monitoring of both dragons and their prey must be conducted on this island. Gili Motang is a small and isolated island and its population of Komodo dragons exhibits the lowest genetic diversity among all extant populations as a result of limited insular gene flow and genetic drift (Ciofi and Bruford, 1999). Genetic isolation and drift increase the probability of expression of delaterious recessive allelles and may therefore lead to inbreeding depression (Frankham et al., 2002; O’Grady et al., 2006). It is conceivable that inbreeding depression could also compound any decline (be it food related or otherwise) with higher rates of abnormal offspring or reduced clutch sizes as observed in other small isolated reptile populations (Madsen et al., 1999). Although there is no evidence of reduced fitness of Komodo dragons on Gili Motang this current population may exhibit negligible through to high levels of inbreeding. In the first instance, detecting evidence of inbreeding depression by way of offspring survival could be a means to resolve the contribution of inbreeding depression to exacerbating population declines on this island. Gili Motang is a small and relatively isolated island and as such climatic induced variation in habitat quality could be relatively dynamic, relative to large islands, and ultimately drive processes (e.g. habitat quality) influencing prey, and inturn, the dragon populations in both evolutionary (see Jessop et al., 2006) and ecological time frames. Annual rainfall patterns are extremely important in regulating the abundance of terrestrial prey populations (Coulson et al., 2000; Georgiadis et al., 2003; Ogutu and Owen-Smith, 2003; Madsen et al., 2006). Rainfall variation across the wet/dry tropics of Northern Australia (a biogeographical zone similar to Komodo national park) (Monk et al., 1997) greatly influences prey density and in turn the population dynamics of their reptile predators (Madsen and Shine, 2000; Madsen et al., 2006). Understanding the potential for similar patterns of annual rainfall in this region of Indonesia to influence the population dynamics of prey and in turn dragons is another key research priority for aiding the conservation and management within Komodo national park. The number of dragons on Gili Motang in 2004 was estimated at 157 ± 68 (i.e. 88.9–224.4). This estimate approximates, or falls below (using the lower standard error), several theoretical thresholds used to flag extinction proneness (O’Grady et al., 2004). In particular demographic and environmental stochasticity and the loss of genetic variability are key factors underpinning extinction in small populations (Frankham, 1998). Demographic stochasticity is usually the major component threatening population viability when the population size is in the order of 100 individuals or smaller (Lande, 1988; Reed, 2005). The relative isolation of this island alongside relatively few commercially valuable terrestrial resources (e.g. timber, commercially valuable species) may provide some respite from anthropic disturbances (e.g. hunting or fire, etc.) for which there is only limited evidence. However, Komodo dragons despite being ectothermic display many intrinsic biological traits that are recognised as factors for increasing extinction proneness in predatory vertebrates
253
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
254
B I O L O G I C A L C O N S E RVAT I O N
them to measure demographic trends, and as inferred by our results that the Gili Motang Komodo dragon population is in decline, provide them with a basis for gauging appropriate management responses, should it be necessary, in the face of a continued decline.
Au
th o
r's
al
on
pe
rs
Amar, A., Redpatha, S., Thirgood, S., 2003. Evidence for food limitation in the declining hen harrier population on the Orkney Islands, Scotland. Biol. Conserv. 111, 377–384. Andrews, R.M., 1991. Population stability of a tropical lizard. Ecology 72, 1204–1217. Auffenberg, W., 1981. The Behavioral Ecology of the Komodo Monitor. University Presses of Florida, Gainesville, Florida, USA. Burkey, T.V., 1995. Extinction rates in archipelagos: implications for populations in fragmented habitats. Conservation Biology 9, 527–541. Cardillo, M., Mace, G.M., Gittleman, J.L., Purvis, A., 2006. Latent extinction risk and the future battlegrounds of mammal conservation. Proceedings of the National Academy of Sciences, USA 103, 4157–4161. Case, T.J., Bolger, D.L., Richman, A.D., 1992. Reptilian extinctions: the last ten thousand years. In: Fielder, P.L., Jain, S.K. (Eds.), Conservation Biology: The Theory and Practice of Nature Conservation, Preservation, and Management. Chapman and Hall, New York, pp. 91–125. Caswell, H., Fujiwara, M., Brault, S., 1999. Declining survival probability threatens the North Atlantic right whale. Proceedings of the National Academy of Sciences, USA 96, 3308–3313. Chappel, J., Shackleton, N.J., 1986. Oxygen isotopes and sea level. Nature 324, 137–140. Ciofi, C., Bruford, M.W., 1999. Genetic structure and gene flow among Komodo dragon populations inferred by microsatellite loci analysis. Molecular Ecology 8, S17–S30. Ciofi, C., De Boer, M.E., 2004. Distribution and Conservation of the Komodo monitor Varanus komodoensis. Journal of Herpetology 14, 99–107. Coulson, T., Milner-Gulland, E., Clutton-Brock, T., 2000. The relative roles of density and climatic variation on population dynamics and fecundity rates in three contrasting ungulate species. Proceeding of the Royal Society London B 267, 1771– 1779. Coulson, T., Gaillard, J., Festa-Bianchet, M., 2005. Decomposing the variation in population growth into contributions from
co
We thank Komodo National Park rangers, technical staff and volunteers, for field assistance. Graeme Gillespie and Joanna Sumner made useful corrections on the manuscripts. Approval for research was conducted under a collaborative program between the Zoological Society of San Diego, The Nature Conservancy (Indonesia program) and the Indonesian Department of Forest Protection and Nature Conservation (PHKA). Financial support for this research was provided to T.S.J. by the Zoological Society of San Diego, the Offield Family Fund, and to CC by Chester Zoo the Zoological Society of London, and Komodo Species Survival Plan of the American Zoo and Aquarium Association (AZA).
R E F E R E N C E S
multiple demographic rates. Journal of Animal Ecology 74, 789–801. Crandall, K.A., Beninda-Emonds, O.R.P., Mace, G.M., Wayne, R.K., 2000. Considering evolutionary processes in conservation biology. Trends in Ecology and Evolution 15, 290– 295. Diamond, J.M., 1989. The present, past and future of human caused extinctions. Philosophical Transactions of the Royal Society of London B 325, 469–477. Efron, B., Tibshirani, R.J., 1993. An Introduction to the Bootstrap. Chapman and Hall, USA. 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. Conservation Biology 13, 531–541. Frankham, R., 1998. Inbreeding and extinction: island populations. Conservation Biology 12, 665–675. Frankham, R., Ballou, J.D., Briscoe, D.A., 2002. Introduction to Conservation Genetics. Cambridge University Press, UK. Georgiadis, N., Hack, M., Turpin, K., 2003. The influence of rainfall on zebra population dynamics: implications for management. Journal of Applied Ecology 40, 125–136. Grant, P.R., 1998. Evolution on Islands. Oxford University Press, UK. Harley, S.J., Myers, R.A., Dunn, A., 2001. Is catch-per-unit-effort proportional to abundance. Canadian Journal of Fisheries Aquatic Sciences 58, 1760–1772. Jessop, T.S., Sumner, J., Imansyah, M.J., Puwandana, D., Rudiharto, H., Phillips, J.A., 2004. Distribution, use and selection of nest type by Komodo dragons. Biological Conservation 117, 463– 470. 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. Krebs, C.J., 1999. Ecological Methodology. Addison-Wesley educational Publishers Inc. Lande, R., 1988. Genetics and demography in biological conservation. Science 241, 1455–1460. Laurie, W.A., Brown, D., 1990. Population Biology of Marine Iguanas (Amblyrhynchus cristatus). 11. Changes in the annual survival rates and the effects of size, sex, age and fecundity in a population crash. Journal of Animal Ecology 59, 529–544. MacArthur, R.H., Wilson, E.O., 1967. The Theory of Island Biogeography. Princeton Univ. Press. Madsen, T., Shine, R., 2000. Rain, fish and snakes: climatically driven population dynamics of Arafura filesnakes in tropical Australia. Oecologia 124, 208–215. Madsen, T., Shine, R., Olsson, M., Wittzell, H., 1999. Restoration of an inbred adder population. Nature 402, 34–35. Madsen, T., Ujvari, B., Shine, R., Olsson, M., 2006. Rain, rats and pythons: climate driven population dynamics of predators and prey in tropical Australia. Austral Ecology 31, 30– 37. McCulloch, M.T., Tudhope, A.W., Esat, T.M., Mortimer, G.E., Chappel, J., Pillans, B., Chivas, A.R., Omura, A., 1999. Coral record of equatorial sea-surface temperatures during the penultimate deglaciation at Huon peninsula. Science 283, 202– 204. Monk, K.A., de Fretes, Y., Reksodiharjo-Lilley, G., 1997. The Ecology of Nusa Tenggara and Maluku. Oxford University Press, Oxford, UK. Nelson, N.J., Keall, S.N., Brown, D., Daugherty, C.H., 2002. Establishing a wild population of Tuatara. Conservation Biology 16, 887–894. O’Grady, J.J., Reed, D.H., Brook, B.W., Frankham, R., 2004. What are the best correlates of extinction risk. Biological Conservation 118, 513–520.
py
Acknowledgements
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5
B I O L O G I C A L C O N S E RVAT I O N
Reid, W.V., Miller, K.R., 1989. Keeping Options Alive: The Scientific Basis for Conserving Biodiversity. World Resources Institute, Washington, DC. Rhind, S.G., Bradley, J.S., 2002. The effect of drought on body size, growth and abundance of wild brush-tailed phascogales (Phascogale tapoatafa) in south-western Australia. Wildlife Research 29, 235–245. Wikelski, M., Trillmich, F., 1997. Body size and sexual size dimorphism in marine iguanas fluctuate as a result of opposing natural and sexual selection, an island comparison. Evolution 51, 922–936. Wolf, C.M., Griffith, B., Reed, C., Temple, S.A., 1996. Avian and mammalian translocations: update and reanalysis of 1987 survey data. Conservation Biology 10, 1142–1154. Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice Hall Inc., New Jersey, USA.
th o
r's
pe
rs
on
al
co
py
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. Biological Conservation 133, 42–51. Ogutu, J.O., Owen-Smith, N., 2003. ENSO, rainfall and temperature influences on extreme population declines among African savannah ungulates. Ecology Letters 6, 412–419. Preen, A., Marsh, H., 1995. Responses of Dugongs to large scale loss of sea grass from Hervey Bay, Queensland. Australian Wildlife Research 22, 507–519. Purvis, A., Gittleman, J.L., Cowlishaw, G., Mace, G.M., 2000. Predicting extinction risk in declining species. Proceedings of the Royal Society of London B 267, 1947–1952. Reed, D.H., 2005. Relationship between population size and fitness. Conservation Biology 19, 563–568.
Au
255
1 3 5 ( 2 0 0 7 ) 2 4 7 –2 5 5