Preventing Freshwater Turtle Extinctions

Critically Evaluating Best Management Practices for 

Preventing Freshwater Turtle Extinctions 

Spencer R-J 1 *, Van Dyke JU 1 , Michael B. Thompson 2 

1School of Science and Health, Hawkesbury Institute for the Environment, Western Sydney University 

2 School of Life and Environmental Sciences, Heydon-Laurence Building (A08), University of Sydney, NSW 

2006, Australia 

*Corresponding Author: r.spencer@uws.edu.au 

Summary 

Ex situ conservation tools, such as captive breeding for reintroduction, are considered last 

resort to help recover threatened or endangered species. However, they may also provide 

alternative strategies where reducing threats directly is difficult or ineffective. Headstarting, 

or captive rearing of eggs or neonate animals and subsequent release into the wild, has 

been controversial for decades. A major criticism is that headstarting is a symptomatic 

treatment of conservation problems (halfway technology), however, it may provide a 

mechanism to address multiple threats, particularly in close proximity of population centres. 

Here we conduct Population Viability Analyses (PVA) to assess the risk of extinction of 

Australia’s most widespread freshwater turtle, Chelodina longicollis, to increasing adult road 

mortality and reduced recruitment through nest predation from introduced foxes. We also 

model a range of management scenarios to test the effectiveness of headstarting, fox 

management, and measures that reduce adult road mortality. We show that headstarting 

should be a primary tool for managing freshwater turtles under threats that affect multiple 

life history stages. Headstarting from harvest populations were the only scenarios that 

eliminated all risks of extinction, while also maintaining population growth. Small 

increments in adult mortality have greatest effect on population growth and extinction risk, 

however, where threats simultaneously affect other life history stages (e.g.. recruitment), 

eliminating harvest pressures on adult females alone will not eliminate the risk of 

population extinction. In our models, one harvest population could supply enough 

hatchlings to supplement 25 other similar sized populations at an annual rate to maintain 

population growth and eliminate the risk of population extinction. We advocate the creation 

of harvest populations for managing freshwater turtles facing significant threats to multiple 

life history stages. 

1 

Spencer R-J et al. 2017 Critically Evaluating Best Management Practices for Preventing Freshwater Turtle 

Extinctions. Conservation Biology. In Press.

To Headstart or Not to 

Headstart? 

Headstarting, or captive rearing of 

eggs or hatchling turtles for release 

into the wild, is a conservation 

practice that has long been 

controversial (e.g., Huff 1989; Burke 

1991; Frazer 1992; Seigel and Dodd 

Jr 2000; Bell et al. 2005;). 

Here we conduct Population Viability 

Analyses (PVA) to assess the risk of 

extinction of Freshwater Turtles to 

increasing adult mortality, 

particularly focusing on harvesting 

of females, and reduced recruitment 

from invasive nest predators. 

We also model a range of 

management scenarios to test the 

effectiveness of headstarting, 

reduction of nest destruction via fox 

management, and measures that 

reduce adult mortality via roadkill. 

Captive breeding for reintroduction and 

population augmentation is an ex situ 

conservation strategy sometimes used to 

help recover threatened or endangered 

species (Bowkett 2009; Conde et al. 

2011). 

Ex situ strategies provide a 

valuable management tool where 

reducing threats directly is difficult or 

ineffective. In essence, ex situ 

conservation strategies aim to replace 

individuals that are impacted by ongoing 

threats, but they have been 

criticized for low rates of success and 

high costs (Wolf et al. 1996; Fischer and 

Lindenmayer 2000) and are unlikely to 

work unless integrated into a broader 

recovery plan (IUCN 2013). Although the 

benefits of ex situ conservation 

strategies are only now being critically 

evaluated (e.g. Canessa et al. 2015), 

they have long been the basis of 

conservation strategies for many taxa; 

few more high profile than headstarting 

in marine turtles. 

The scientific merit of headstarting 

as a conservation tool to manage turtle 

populations has been questioned 

because: 

2 



1) it is a symptomatic treatment of 

conservation problems (halfway 

technology: Frazer 1992); 2) it may alter 

behavior of head-started individuals 

(Woody, 1991); 3) it may disrupt 

population genetics (Rasberry 2015); and 

4) it may disrupt ecological function 

(Bowen et al. 1994). From a population 

point of view, the egg and juvenile stages 

have been considered life-history stages 

that have minimal impact on population 

growth (Heppell 1998). 

Thus, headstarting has been seen 

primarily as a strategy to promote ecotourism 

and attract funding, especially 

for marine turtle conservation (Tisdell 

and Wilson, 2005). The effectiveness of 

headstarting marine turtles as a 

population management strategy has 

never been truly assessed, largely 

because the time between release and 

maturity (when a released juvenile may 

return to nest) is over 20 years in some 

species (Ernst and Barbour 1989). 

However, recent studies suggest that 

headstarting (in conjunction with other 

management strategies) may have been 

effective, and once declining or 

endangered marine turtle populations are 

now stable or increasing (Crowder and 

Heppell 2011). 

Headstarting is not commonly used as a 

conservation strategy for freshwater turtles. 

High financial costs, as well as landscape level 

disconnectivity among populations, have 

probably restricted its use, and past 

population modelling suggests that 

conservation efforts are more effective when 

focused on reducing adult mortality (e.g., 

Heppell et al. 1996; Heppell 1998). Elasticity 

values of adult survival are orders of 

magnitude greater than for any other stage 

(Heppell 1998), meaning that the survival of 

one adult has far greater impact on 

population stability than the survival of 

individuals from any other stage. Elasticity 

quantifies how many individuals of each stage 

would be required to maintain population 

stability or growth. Although increasing 

population growth by 1% requires far fewer 

adults surviving than eggs (e.g., Heppell et al. 

1996; Heppell 1998)., increasing adult 

survival, from a management perspective, 

may be far more costly and/or complicated 

than harvesting eggs from nests or gravid 

females, or protecting nests. Thus the 

benefits of protecting many eggs over fewer 

adults may prove more cost-effective and 

logistically feasible. Headstarting has become 

an acceptable method for turtle management 

and conservation despite model projections. 

The Turtle Survival Alliance, is currently 

involved in headstarting for at least 11 

species, including freshwater turtles and land 

tortoises (Burke 2015).. However, there is 

little evidence demonstrating the 

effectiveness of headstarting for restoring 

freshwater turtle populations (Burke 2015). 

Freshwater turtles face threats that permeate 

all life history stages, from egg to adult. Turtle 

life histories are characterised by high, 

fluctuating rates of egg and juvenile 

mortality. That mortality is balanced by 

extreme iteroparity (i.e. long-lived, highly 

fecund), where threats to adult survival are 

generally low. 

3 



In Australia, humans have impacted this 

regime because mortality of eggs and young 

has increased, primarily because of predation 

by invasive foxes, Vulpes vulpes (Thompson 

1983), and increased adult mortality through 

motor vehicle mortality on roads and direct 

predation by foxes on nesting females 

(Spencer 2002). In the Murray River in 

Australia, mortality rates of eggs by foxes 

have increased to over 93% (Thompson 

1983; Spencer 2002a). As a result, turtles in 

the River Murray are in serious decline, with 

abundances 69-91% lower than 40 years ago 

(Chessman, 2011). Apart from road deaths 

and predation, turtles are also struck by 

boats, drowned in fishing nets or in irrigation 

pumps, killed by anglers and human 

activities that affect water quality are 

increasing the prevalence of wildlife diseases 

(Kennett et al. 2009). 

With multiple threats impacting multiple life 

history stages of freshwater turtles, the 

dilemma for conservation is the capacity to 

implement diverse, broad-scale management 

strategies to address each threat. Thus, it is 

time to evaluate whether headstarting may 

allow for simultaneously managing a range of 

threats affecting freshwater turtles. Our aim 

is to evaluate a range of management 

strategies using Australia’s most common 

and widespread freshwater turtle, the 

Eastern Long-necked Turtle (Chelodina 

longicollis), which is highly mobile and 

inhabits wetlands throughout the south-east 

of Australia, including cities and urban areas 

(Cann 1998). Nest predation rates are high 

because invasive foxes occur throughout 

their range, but they are also at particular 

risk of road mortality and habitat 

fragmentation arising from urban 

infrastructure (Hamer et al. 2016) because 

they frequently move between wetlands 

(Spencer and Thompson 2005). They occupy 

a wide range of wetland habitats and are a 

late maturing species, with males maturing 

at 7-8 years of age and females at 10-12 

years of age (Kennett et al. 2009). Mortality 

of adult turtles can drive species to 

extinction (Heppell 1998), and combined 

with reduced recruitment levels because of 

invasive predators, a species like Chelodina 

longicollis may be at particular risk of 

extinction. Given their proximity to population 

centres, management of their multiple threats 

is complex and difficult. 

Here we conduct Population Viability 

Analyses (PVA) to assess the risk of 

extinction of C. longicollis to 

increasing adult mortality, 

particularly focusing on harvesting of 

females, and reduced recruitment 

through fox predation. We also model 

a range of management scenarios to 

test the effectiveness of 

headstarting, reduction of nest 

destruction via fox management, and 

measures that reduce adult mortality 

via roadkill. 

Chelodina longicollis is the most 

widespread freshwater turtle in Australia, 

with an extensive range across eastern 

Australia (Cann 1998). Their range 

broadly overlaps human populated areas 

and includes the capital cities of Brisbane, 

Melbourne, Canberra, Adelaide and 

Sydney (Cann 1998; Kennett et al. 2009). 

Chelodina longicollis occupies a wide 

range of habitats, such as shallow 

ephemeral swamps, farm dams 

(Chessman 1988; Wong and Burgin 

1997), and flowing rivers (Chessman 

1988). 

4 



Their estimated population densities vary 

between 26 and 400 turtles ha-1 

(Parmenter 1976; Kennett et al. 2009, 

Ferronato 2015). Their lifespan is not 

known, but they are slow growing and 

may live more than 100 years (Parmenter 

1976; Thompson 1993). Females emerge 

from the water to oviposit one clutch 

annually of 10-20 eggs (Parmenter 1976; 

Thompson 1993; Vestjens 1969) and both 

sexes frequently migrate overland among 

water bodies to exploit ephemeral 

swamps (Kennett and Georges 1990; 

Stott 1987). 

They are carnivorous, consuming 

primarily aquatic insects and carrion 

(Chessman 1988). Estimates of 

population densities range from 26 to 

400 turtles ha-1 (Parmenter 1976; 

Chessman 1978; Georges 1982a). These 

estimates are based on trappable 

populations (ie. not including early 

juvenile or embryonic stages). Based on 

our model parameters, an initial 

population size 1000 produces a stable 

stage trappable population of 250 turtles. 

Thus we set an upper carrying capacity of 

2000 individuals to reflect the upper 

limits of the trappable population in the 

wild. 

We constructed matrix population projection 

models (PopTools; Hood 2010) and used 

Vortex 10.0 (Lacy and Pollack 2015) to assess 

population viability analysis of current 

Chelodina longicollis populations (Table 1). 

We ran the baseline scenario with parameters 

from Table 1. This represents the ‘Pre- 

European Settlement’ scenario where nest 

predation levels were high but highly variable, 

and minimal adult mortality occurred on land. 

Secondly, we ran a scenario in the ‘Post- 

European Settlement’ environment where 

nest predation rates are very high and less 

variable, with 0%, 1% and 2% of the adult 

female starting population being 

harvested/depredated each year (or at 3 year 

intervals). We then modelled a range of 

management scenarios where (1) hatchling 

turtles were supplemented from a separate 

harvest population at annual (and at 5-year 

intervals) rates of 15%, 30%, 45%, 60% and 

75% of the initial adult population size; and 

(2) fox management tools were used to 

reduce nest predation rates by increments of 

5% up to a 70% reduction from current rates 

(e.g., Spencer et al. 2016). 

We investigated the relationship between fox 

activity and nest predation rate throughout 

much of the range of C. longicollis in south 

eastern Australia. From 2014-2016, we 

created nest sites around wetlands along the 

Murray River (35.9561° S, 144.3693° E), 

Winton Wetlands (36.5496° S, 145.9834° E) 

and Hawkesbury (33.5965° S, 150.7505° E) 

regions of south-eastern Australia, using 

commercially purchased unfertilised chicken 

and quail eggs. Two to five eggs were placed 

in an artificial nest and 5 - 30 nests were 

created in 200 m2 areas. Camera traps were 

placed in each area and a fox activity index 

was created by calculating the number of 

days (24 h period) and dividing by the 

number of days that the trial was conducted. 

Trials were conducted for 4 - 9 weeks and 

nest predation rates were assessed at the end 

of each trial. Logistic regression analyses 

were conducted to examine the relationship 

between fox activity and nest predation rate. 

5 



1 

Table 1. Base parameters used for modelling population growth and risk of extinction of C. longicollis 

Vortex Parameter Wild population estimate Reference 

# of populations 1 - 

Inbreeding depression No 

included? 

Concordance of 

No 

environmental variation (EV) 

and reproduction 

Age of first reproduction 6/10 Kennett et al. 2009 

(♂ / ♀) 

Maximum age of 

70 

reproduction 

Annual % adult females 100 

breeding 

Kennett et al. 2009 


Density dependent 

None 

reproduction? 

% males in breeding pool 100% 

Maximum number of broods 1 Cann 1998 

per year 

Clutch size Cann 1998 

Offspring sex ratio 50/50 

Adult mortality 5% (±5%) both Shine and Iverson 1995; 

Spencer 2002; Spencer 

and Thompson 2005 

0-1 years (pre-/post- 

European Settlement) 

75%(±25%)/95%(±10%) Spencer 2002; Spencer 


1-2 years 50%(±10%) Spencer 2002; Spencer 










6-7 years (♀) 10%(±10%) Spencer 2002; Spencer 






Initial population size 

1000 (Stable age distributed) 


Carrying capacity (K) 2000(±500) 


2 

Table 2. Population growth and risk of extinction. Pre-European conditions includes variable nest predation rates (75% ± 25%), whereas the 

post-European environment includes high nest predation rates (95% ± 10%). Headstarting, fox management, and adult harvesting scenarios 

are compared. Rates of hatchling supplementation are based on initial adult population size. 

Stochastic Population Growth Rate (R) Risk of Extinction (0-1) 

Reducing Adult Mortality to 0% 

Pre-European Settlement 0.0485 0.02 

Post-European Settlement -0.02 0.32 

Headstarting- Post-European Settlement (7.5% supplementation per year) 0.0032 0 

Headstarting- Post-European Settlement (15% supplementation per five years) -0.0027 0 

Headstarting- Post-European Settlement (30% supplmentation per five years) 0.0013 0 

Fox Management- Post European Settlement (10% reduction in Nest Predation) 0.0309 0.01 


Adult Mortality at 1% of the Female Population 


Post-European Settlement -0.0442 1 

Headstarting- Post-European Settlement (15%supplemation per year) 0.006 0 



Adult Mortality at 2% of the Female Population 


Post-European Settlement (High Fox Predation) -0.0603 1 

Headstarting- Post-European Settlement (15% supplementation per year) -0.0012 0 

Headstarting- Post-European Settlement (30% supplementation per year) 0.0155 0 

Headstarting- Post-European Settlement (45% supplmentation per 5 years) -0.006 0 

Headstarting- Post-European Settlement (60% supplmentation per 5 years) -0.003 0 

Headstarting- Post-European Settlement (75% supplmentation per 5 years) 0.0017 0 



Spencer Fox Management- R-J et al. Post 2017 European Critically Settlement (70% Evaluating reduction in Best Nest Predation) Management Practices for Preventing 0.0793 Freshwater Turtle 0.32 

Extinctions. Pre-European Conservation Settlement (harvesting Biology. at three year In intervals) Press. 

0.0271 0.39 

Post-European Settlement (harvesting three year intervals) -0.044 1 

6

TurtleSAT is a citizen science tool where 

incidental sightings of turtles (or their 

activity) can be recorded via website or 

iOS or Android App 

(http://TurtleSAT.org.au). Users are 

taken through a series of questions via a 

user-friendly interface and a photo can be 

attached. The whole form is geo-located 

automatically using the device’s in-built 

GPS. TurtleSAT has been active since May 

2014 and has more than 4700 recordings 

(as of 14/10/2016). TurtleSAT is 

promoted by an active social media 

campaign (Facebook and Twitter) and 

numerous community groups and 

government agencies in each state 

actively use it to record incidental 

sightings of turtles. We used TurtleSAT 

data to map terrestrial turtle deaths 

(primarily road mortalities) throughout 

eastern Australia. The exercise was not to 

quantify the rates of mortality per 

population, but rather to demonstrate the 

spatial scale over which that mortality is 

occurring. 

RESULTS 

With no adult mortality in a pre-European 

environment, population growth is 

positive and there is a 2% risk of 

extinction (Table 2). As adult mortality 

increases, population growth rates decline 

(Fig. 1) and the risk of extinction increases 

by ~40% for each 1% increase in adult 

harvesting per year (Table 1). In the post- 

European environment, fox predation on 

nests increases significantly and even if 

there is no harvesting on adult females, 

populations decline (Fig. 2) and the risk of 

extinction increases by 30% (Table 1). As 

expected, the combination of both sources of 

mortality leads to extinction within 100 years 

(Table 2). 

Fig. 1. Impact of mortality to adults in a pre-European 

environment where nest predation rates are variable (75% ± 

25%) and adult females are harvested at the rates of 0% (dark 

grey), 1% (grey) and 2% (light grey) of the starting adult 

female population. 

Population Size (N) 

1400 

1200 

1000 

800 

600 

400 

200 

0 

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 201 

Fig. 2. Impact of nest predation on population growth. In a pre- 

European environment where nest predation rates are variable 

(75% ± 25%) (dark grey) and post-European environment 

(light grey) where nest predation rates are high (95 ± 10%). 

No adults are harvested. 

Headstarting 

Supplementing hatchlings at a rate of 30% of 

the initial population size can negate the 

impact of adult mortality and high nest 

predation rates from foxes (Fig. 3). Based on 

an average clutch size of 15 eggs, 

supplementation of eggs from only 4% of the 

adult female population would replicate pre- 

European population growth (Fig. 3). Lower 

levels of hatchling supplementation can even 

buffer the combined impacts of higher adult 

mortality and high nest predation rates (Fig. 

3). Supplementing hatchlings into a 

population can stabilise populations even if 

the rate of supplementation is not annual. 

Supplementing populations at the rate of 

45%, 60% and 75% of the starting adult 

population size every 5 years eliminates the 

risk of extinction (Table 2), although this 

pulse recruitment will not replicate pre- 

European conditions (Fig. 4). 

Year 

7 



Reducing Nest Predation 

The effect of reducing nest predation rates 

is highly dependent on the degree of adult 

mortality per year. In areas where adult 

mortality rates are high, nest predation 

rates must be reduced by 70% to 

effectively replicate pre-European 

population growth (Fig. 5), but the risk of 

extinction still remains at 32% (Table 2). In 

populations with lower adult mortality 

rates, a 15% reduction in nest predation 

rates can replicate pre-European 

population conditions; but extinction 

probabilities still remain at 20% (Table 2). 

Fig. 4. Headstarting as a management option in an 

environment where adult mortality is high (2% annual 

mortality) and nest predation is high because of foxes. 

Population numbers based on pre-European conditions are 

shown in black; 5 year pulse headstarting with hatchlings 

representing 75% of the initial adult population size in light 

grey; 5 year pulse headstarting with hatchlings representing 

60% of the initial adult population size in dark grey; 5 year 

pulse headstarting with hatchlings representing 45% of the 

initial adult population size in grey. 

(a) 

1 

Fig. 3. Headstarting as a management option in an environment 

where adult mortality is high (2% annual harvesting) and nest 

predation is high because of foxes. Population numbers based on 

pre-European conditions are shown in dark grey; Annual 

headstarting with hatchlings representing 30% of the initial adult 

population size in light grey; Annual headstarting with hatchlings 

representing 15% of the initial adult population size in dark grey 

(dashed); Annual headstarting with hatchlings representing 

7.5% of the initial adult population size in light grey (dashed). 

Preventing Adult Mortality 

Eliminating adult mortality only from a 2 

population does not reduce the risk of 

extinction (Table 2). Populations still decline 

with high nest predation rates even if old age 

is the only source of female mortality (Fig. 6). 

Fig. 5. Population numbers based on pre-European conditions 

are shown in dark grey. (a) Adult mortality rates at 2% per 

year. Nest predation rates are reduced to 25% (dotted line); 

50% (light grey) and 65% (grey) (b) Adult harvest rates at 1% 

per year. Nest predation rates are reduced to 80% (dark grey); 

85% (light grey). 

Table 3. Review of available literature assessing efficacy of standard fox control methods in Australia. 

(b) 

8 



1400 

1200 

1000 

Population Size (N) 

800 

600 

400 

200 

0 

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 201 

Year 

Fig. 6. Population numbers based on pre-European conditions are shown in dark grey. (a) Adult mortality 

rates at 2% per year and nest predation rates are high (light grey- dashed); b) Adult mortality rates at 1% per 

year and nest predation rates are high (dark grey- dashed); and c) Adult mortality rates at 0% per year and 

nest predation rates are high (dotted). 

Nest Predation Rate 

1 

0.9 

0.8 

0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 

0 

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 

Fox Activity (mean presence of foxes per day) 

Fig. 7. Relationship between fox activity and egg predation throughout eastern Australia. 

9 



DISCUSSION 

Here we show that headstarting could be an 

effective tool for managing freshwater turtles 

against multiple threats that affect multiple 

life history stages. In our models, 

headstarting from harvest populations is the 

only tool that eliminates all risks of 

extinction, while maintaining small 

increments in population growth. Similar to 

previous studies (e.g., Heppell et al. 1996; 

Heppell 1998), small increments in adult 

mortality have greatest effect on population 

growth and extinction risk, but if other 

threats simultaneously affect other life 

history stages (e.g., recruitment), 

eliminating harvest pressures on adult 

females alone will not eliminate the risk of 

population extinction. 

Our study clearly shows that headstarting can 

effectively halt or reverse declines of C. 

longicollis populations in the face of multiple 

threats, especially in cases where external 

threats affect multiple life history stages of 

freshwater turtles. In our case study, factors 

that have impacted C. longicollis will never 

dissipate until populations are extinct, or 

technology to reduce threats from invasive 

predators becomes more effective. 

Headstarting should be the primary 

conservation tool for managing C. longicollis 

in decline and similar analyses should be 

conducted to evaluate its value for each 

species requiring active management. 

Management of a population or a species 

under threat often focuses directly on 

reducing impacts on the life history stage(s) 

affected. In doing so, focus inevitably is 

directed to the threat, rather than on the 

impacts on the affected population. Plant 

biologists and conservationists have long 

criticised classical biocontrol for lacking 

quantitative assessments of effectiveness, 

especially post-release (McEvoy & Coombs 

1999), yet invasive vertebrate pest 

management primarily focuses on reducing 

densities of invasive predators or herbivores. 

The core components of conservation policy 

to manage their impacts is to reduce predator 

numbers in an area using lethal methods, 

such as poison baiting, shooting, and 

trapping. The actual efficacy (e.g., reduced 

impact on target species or increases in 

biodiversity) of such programs are rarely 

assessed (Walsh et al. 2012) and success is 

determined by the number of carcasses, 

reduced activity of the target species or the 

number of baits taken. Efficacy of these 

programs is vital given the limited resources 

available for most conservation programs and 

the high costs associated with lethal control. 

AU$21.3m was spent on labor costs alone for 

red fox control in Australia in 1998–2003, but 

the benefits to native prey are largely not 

known (Reddiex et al. 2006). Management 

often assumes that there is a linear 

relationship between impact and invasive 

species densities; hence successes in 

reducing invasive species are also lauded as 

conservation outcomes. Yet, rarely are the 

relationships between densities and impact 

linear, and in some rare instances, reductions 

in invasive animals may result in increased 

impacts on native species (Spencer et al. 

2016). Clearly, impacts on turtle nests are not 

a direct function of increased activity or 

densities of foxes in south-eastern Australia 

(Fig. 7), and management focusing on 

reducing foxes must achieve nothing short of 

complete eradication from an area to truly 

reduce nest predation rates (Fig. 7); an 

achievement that appears impossible for any 

extended period of time (Table 3). 

Conservation outcomes and management 

decisions should be based on the relationship 

between the level of threat and impact on the 

affected species. The impacts of introduced 

species are a product of their density and 

functional response to prey densities. 

Functional responses are the changes in the 

rate of prey consumption by individual 

predators, relative to prey abundance 

(Holling, 1959). Numerical responses are the 

aggregative rates of prey consumption by all 

predators relative to prey density, which 

10 



Fig. 8. Locations of turtle mortality sightings in south-eastern Australia from September 2014-September 2016 in TurtleSAT 

(http://TurtleSAT.org.au). Photo of a dead C. longicollis from Victoria uploaded to TurtleSAT. 

change with predator density via reproduction or migration, in response to changes in prey 

density (Holling, 1959). Traditional management of introduced predators focuses on reducing 

predator populations, and thus focuses on numerical responses, which fails to manage for 

functional responses, and may not eliminate impacts of highly-efficient individual predators 

(Spencer et al. 2016). For conservation, functional responses become critical when 

introduced species densities are declining, as functional responses essentially define the 

levels (numerical reductions) that management must achieve if resources are directed into 

removing introduced species. For example, if one predator in a system can achieve 

extremely high levels of predation, only management options that completely eliminate that 

individual from the system will be effective. In open systems at a landscape level, this level 

of management is impossible to achieve, or not cost effective (Doherty and Ritchie 2016). In 

a review of the literature to determine efficacy of standard lethal control methods in 

Australia, we found that few studies were able to reduce fox numbers for extended periods, 

and none were able to eradicate them (Table 3). Thus, freshwater turtle populations are 

unlikely to get any relief from standard fox control programs (assuming functional responses 

remain constant after a reduction of foxes) and alternative management techniques should 

be considered to manage turtle populations (or other species) under threat by predation 

from invasive predators like foxes. 

Likewise, reducing adult road mortality for turtle species that are highly mobile across 

terrestrial habitats, like C. longicollis, has been historically difficult. Fences are required to 

funnel turtles to suitable crossing locations (ecopassages) and prevent them from crossing 

the road, but may themselves cause mortality (Ferronato 2014). In addition, some turtles 

can climb fences, which limits their effectiveness (Baxter-Gilbert et al. 2015). If the fence is 

effective, turtles are then funnelled to cross underneath the road via an ecopassage, but 

there is conflicting evidence whether they are always willing to do so (Woltz et al. 2008; 

Baxter-Gilbert et al. 2015), and some evidence that the ecopassage itself may act as a prey 

trap (Little et al. 2002). Furthermore, constructing suitable road crossings for wildlife at 

suitable densities on a roadway is expensive (Mata et al. 2008; Baxter-Gilbert et al. 2015). 

Thus, there are substantial feasibility challenges in reducing or preventing adult turtle 

11 


mortality on roadways. 


Our models demonstrate that headstarting would be an effective and cost-efficient primary 

management tool in a broad-scale integrated conservation plan for C. longicollis. There is no 

well recognized and accepted definition of headstarting. The IUCN-SSC (2013) states that 

head-starting reptiles "avoids the heavy mortality of young age classes in the wild; wild 

hatchlings are reared in protective enclosures before release at less susceptible size/age". 

Burke (2015) redefined headstarting as the practice of protecting especially vulnerable life 

stages of a species to increase the likelihood of survivorship for conservation purposes. 

Success of population manipulations such as reintroductions and head-starting requires 

appropriate planning to minimize negative genetic and disease consequences, as well as 

impacts to other native species. Captive populations are subject to problems such as 

inbreeding depression, loss of genetic diversity, and adaptation to captivity. Thus, it is 

important to manage captive populations in a way so that introduced individuals resemble 

the original founders as closely as possible, increasing the probability of success (Kleiman et 

al. 2010). With critically endangered species, this is extremely difficult given the low 

numbers, and likely genetic diversity, of breeder animals. Still, headstarting programs with 

small captive populations of Galapagos tortoises have proved successful at restoring 

population numbers in some situations (Jensen et al 2015). Importantly, headstarting as a 

preventative tool must also be tuned to maintain genetic diversity across a wide area 

(Jensen et al 2015). In cases where turtles are declining due to nest predation by invasive 

predators, like C. longicollis, developing suitable harvest populations in situ is the key. Many 

common species of turtle occur in integrated wetlands and water treatment plants (eg., 

constructed wetlands) throughout their range, and these facilities may provide a tool for low 

cost headstarting programs for widespread but declining populations. 

Constructed wetlands are small water bodies that have enormous biodiversity potential. The 

Integrated Constructed Wetland (ICW) concept is a refinement of storm and wastewater 

treatment and effluent reuse facilities, where water and effluent are treated through a series 

of ponds. Final tertiary treatment ponds can contain some of the highest aquatic biodiversity 

in a region, and well-designed wastewater treatment and effluent reuse plants are potentially 

important sanctuaries for local biodiversity, including turtles. They often also have existing 

infrastructure including fences, roads, etc. that facilitate protection from invasive species like 

foxes. The reproductive potential of turtle populations in constructed wetlands therefore 

represents a potential pre-existing resource for developing localized headstarting programs in 

situ. 

Fig. 9. Graphic demonstration of how one harvest population can supply a large region of turtles each year for headstarting. Each population 


represents a water body on the landscape, the size of which is random. The dark grey population represents a potential harvest population, 

where Extinctions. hatchlings can Conservation be translocated Biology. to surrounding In Press. populations (light grey) to eliminate extinction risk and maintain population growth at 

pre-European settlement levels. One hectare of turtles can supply enough hatchlings to release into 25ha elsewhere. 

12

Constructed wetlands could be used as breeder facilities for headstarting turtles throughout a 

local region, where the genetic diversity of turtles within the constructed wetland suitably 

matched that of natural wild populations. Our models show that a single population of females 

could supply enough hatchlings to supplement 25 other similar sized populations at an annual 

rate to maintain population growth at pre-European levels and completely eliminate the risk 

of population extinction. A simplistic model where relative turtle densities are based on surface 

area of water demonstrates that all eggs/hatchlings collected from 1 ha of water can provide 

sufficient numbers of turtles to headstart ~25 ha of water in a region (Table 4; Fig. 9). Our 

models also demonstrate that periodic increases in recruitment can sustain populations, 

potentially allowing populations in a region to be managed in a mosaic fashion. In other words, 

not all populations need to be actively managed each year. 

In conclusion, we demonstrate a theoretical basis for why headstarting programs are a useful 

conservation strategy for stopping the declines of freshwater turtles, especially those primarily 

threatened by both adult mortality (due to roads, predation, or harvest) and invasive predators 

affecting multiple life history stages. Chelodina longicollis is Australia’s most common species 

of freshwater turtle but it is clearly at risk because of adult female mortality rates and reduced 

juvenile recruitment. Its longevity has hidden the impact of these threats, but Australia is now 

at the stage since post-European settlement where the effects of foxes and urban population 

sprawl (and associated infrastructure) are being observed through large declines of the 

population of adult turtles. While ours are predictive models, they reflect what is occurring in 

some populations, where there have been declines of up to 91% (Chessman 2011). Perhaps 

it is time to introduce headstarting as a primary management tool for actively managing 

declining turtle populations, rather than as a tool that is only used once a species becomes 

critically endangered. 

Acknowledgements 

We thank Heather Cameron and the Murray River Turtle Team for their support, as well as, 

many other colleagues for constructive discussion on this subject. Financial support was 

provided by the Australian Research Council Linkage Grant Program (LP150100007), North- 

Central Catchment Management Authority, Yorta Yorta Aboriginal Corporation, Foundation for 

National Parks and Wildlife, Victorian Department of Land, Environment, Water and Planning, 

Winton Wetlands, Turtles Australia, Inc. and Save Lake Bonney Group Inc. TurtleSAT is 

supported by the Invasive Animals CRC and NSW Department of Primary Industries. We 

especially thank Peter West and Michael Newton (Newton Green Technologies) for on-going 

support. We also thank Turtles Australia for their ongoing support of TurtleSAT. 

13 


Extinctions. Conservation Biology. In Press. 

https://www.youtube.com/watch?v=WNdGs2H2P8g

References 

Baxter-Gilbert JH, Riley JL, Lesbarrères D, 

Litzgus JD. 2015. Mitigating Reptile Road 

Mortality: Fence Failures Compromise 

Ecopassage Effectiveness. PLoS ONE 10(3): 

e0120537. doi:10.1371/journal.pone.0120537 

Bell, CDL, Parsons J, Austin TJ, Broderick AC, 

Ebanks-Petrie G, and Godley BJ. 2005. 

Some of them came home: the Cayman Turtle 

Farm headstarting project for the green turtle 

Chelonia mydas. Oryx 39:137-148. 

Bowen BW, Conant TA, and Hopkins Murphy 

SR. 1994. Where Are They Now? The Kemp’s 

Ridley Headstart Project. Conservation 

Biology 8:853-856. 

Bowkett AE. 2009. Recent captive-breeding 

proposals and the return of the Ark concept to 

global species conservation. Conservation 

Biology 23:773–776. 

Burke, RL. 1991. Relocations, repatriations and 

translocations of amphibians and reptiles: 

taking a broader view. Herpetologica 47:350- 

357. 

Burke, RL. 2015. Head-starting turtles: learning 

from experience. Herpetological Conservation 

and Biology 10:299-308. 

Cann, J. 1998. Australian freshwater turtles. 

Beaumont Publishing, Singapore. 

Canessa S, Converse SJ, West M, Clemann N, 

Gillespie G, McFadden M, Silla AJ, Parris 

KM. and McCarthy MA. 2016. Planning for 

ex-situ conservation in the face of uncertainty. 

Conservation Biology 30(3): 599-609. 

Chessman, BC. 1988. 'Habitat preferences of 

freshwater turtles in the Murray Valley, 

Victorian and New South Wales'. Australian 

Wildlife Research 15: 485-491. 

Chessman, BC. 2011. Declines of freshwater 

turtles associated with climatic drying in 

Australia’s Murray-Darling Basin. Wildlife 

Research 38:664–671. 

Conde DA, Flesness N, Colchero F, Jones O, 

Scheuerlein A. 2011. An emerging role of 

zoos to conserve biodiversity. Science 

331:1390–1391. 

Crowder LB and Heppell SS. 2011. The decline 

and rise of a sea turtle: how Kemp’s ridleys are 

recovering in the Gulf of Mexico. Solutions 

2:67-73 

Doherty T S, and Ritchie EG. 2016. Stop 

Jumping the Gun: A Call for Evidence-Based 

Invasive Predator Management. Conservation 

Letters. doi:10.1111/conl.12251 

Ferronato BO, Roe JH, and Georges A. 2014. 

Reptile bycatch in a pest-exclusion fence 

established for wildlife reintroductions. J. Nat. 

Conserv. 22: 577-85. 

Frazer NB. 1992. Sea Turtle Conservation and 

Halfway Technology. Conservation Biology 

6:179-184. 

Fischer J, and Lindenmayer DB. 2000. An 

assessment of the published results of animal 

relocations. Biological Conservation 96:1–11. 

Hamer JA, Harrison LJ, and Stokeld D. 2016. 

Road density and wetland context alter 

population structure of a freshwater turtle. 

Austral Ecology. 41: 53-64. 

Heppell SS. 1998. An application of life history 

theory and population model analysis to turtle 

conservation. Copeia 1998(2):367-375. 

Heppell SS, and Crowder LB. 1998. Prognostic 

evaluation of enhancement programs using 

population models and life history analysis. 

Bulletin of Marine Science 62:495-507. 

Heppell SS, Limpus CJ, Crouse DT, Frazer NB, 

and Crowder LB. 1996. Population model 

analysis for the loggerhead sea turtle, Caretta 

caretta, in Queensland. Wildlife Research. 23: 

143-159. 

Hood GM, Barry SC, and Martin PAJ. 2009. 

Alternative Methods for Computing the 

Sensitivity of Complex Surveillance Systems. 

Risk Analysis 29(12): 1686-1698. 

Holling CS. 1959. The components of predation as 

revealed by a study of small-mammal 

predation of the European pine sawfly. 

Canadian Entomologist 91:293–320. 

Huff JA. 1989. Florida (USA) terminates 

“headstart” program. Marine Turtle Newsletter 

46:1-2. 

IUCN (International Union for Conservation of 

Nature). 2013. Guidelines for reintroductions 

and other conservation translocations. Species 

Survival Commission, IUCN, Gland, 

Switzerland. 

Jensen EL, Tapia W, Caccone A and Russello 

MA. 2015. Genetics of a head-start program 

to guide conservation of an endangered 

Galápagos tortoise (Chelonoidis ephippium). 

Conserv Genet 16: 823. 

Kennett R and Georges A. 1990. 'Habitat 

utilisation and its relationship to growth and 

reproduction of the eastern long-necked turtle 

Chelodina longicollis (Testudinata: Chelidae), 

from Australia'. Herpetologica. 46: 22-33. 

Kennett R, Roe J, Hodges K and Georges A. 

2009. Chelodina longicollis (Shaw 1784)- 

eastern long-necked turtle, common longnecked 

turtle, common snake-necked turtle. 

Chelonian Research Monographs 5: 031.1- 

031.8. 

Kleiman DG, Beck BB, Baker AJ, et al. 1990. 

The conservation program for the golden lion 

tamarin, Leontopithecus rosalia. Endang Sp 

Update 8:82–85. 

Lacy RC, and Pollak JP. 2015. Vortex: A 

Stochastic Simulation of the Extinction 

Process. Version 10.1. Chicago Zoological 

Society, Brookfield, Illinois, USA. 

Little SJ, Harcourt RG, Clevenger AP. 2002. 

Do wildlife passages act as prey-traps? Biol 

Conserv. 107:135–145. 

Mata C, Hervás I, Herranz J, Suáraz F, and 

Malo J. 2008. Are motorway wildlife passages 

worth building? Vertebrate use of roadcrossing 

structures on a Spanish motorway. J. 

Environmental Management 88: 407-415. 

Parmenter CJ. 1976. The natural history of the 

Australian freshwater turtle, Chelodina 

longicollis (Shaw) (Testudina, Chelidae). 

Doctor of Philosophy, The University of New 

England, Armidale, NSW. 

Rasberry A. 2015. Can headstart programs truly 

be effective? Genetic assessment of Cyclura 

collei, the headstarted Jamaican iguana. Doctor 

of Philosophy, Mississippi State University, 

Starkville, MS. 

Reddiex B, Forsyth DM, McDonald-Madden 

E. et al. 2006. Control of pest mammals 

for biodiversity protection in Australia. I. 

Patterns of control and monitoring. 

Wildlife Res. 33: 691-709. 

Seigel RA, and Dodd Jr CK. 2000. Manipulation 

of turtle populations for conservation: halfway 

technologies or viable options? Pp. 218-238 In 

Turtle Conservation. Klemens, M.W. (Ed.). 

Smithsonian Institution Press, Washington. 

Shine R and Iverson JB. 1995. Patterns of 

survival, growth and maturation in turtles. 

Oikos 72(3):343-348. 

Spencer R-J. 2002. Experimentally testing nest 

site selection: fitness trade-offs and predation 

risk in turtles. Ecology. 83: 2136-2144. 

Spencer R-J, and Thompson MB. 2005. 

Experimental analysis of the impact of foxes 

on freshwater turtle populations. Conservation 

Biology 19:845–854. 

Spencer R-J, Van Dyke JU. and Thompson MB. 

2016. The ethological trap: functional and 

numerical responses of highly efficient 

invasive predators driving prey extinctions. 

Ecol Appl. 26: 1969–1983. 

Stott P. 1987. 'Terrestrial movements of the 

freshwater tortoise Chelodina longicollis 

(Shaw) as monitored with a spool tracking 

device'. Australian Wildlife Research. 14: 559- 

567. 

Thompson MB. 1983. Populations of Murray river 

tortoise, Emydura (Chelodina): the effect of 

egg predation by the red fox, Vulpes vulpes. 

Wildlife Research 10: 363–371. 

Tisdell C. and Wilson C. 2005. Do opencycle 

hatcheries relying on tourism 

conserve sea turtles? Sri Lankan 

developments and economic–ecological 

considerations. Environmental 

Management. 35: 441-452. 

Vestjens W. 1969. 'Nesting, egg-laying and 

hatching of the snake-necked tortoise at 

Canberra, A.C.T'. Australian Zoologist. 15: 

141-149. 

Walsh JC, Wilson KA, Benshemesh J, and 

Possingham HP. 2012. Unexpected outcomes 

of invasive predator control: the importance of 

evaluating conservation management actions. 

Animal Conservation 15: 319-328. 

Wolf CM, Griffith B, Reed C, Temple SA. 1996. 

Avian and mammalian translocations: update 

and reanalysis of 1987 survey data. 

Conservation Biolog. 10:1142–1154. 

Woltz HW, Gibbs JP, Ducey PK. 2008. Road 

crossing structures for amphibians and reptiles: 

Informing design through behavioral analysis. 

Biological Conservation 141: 2745–275. 

Woody JB. 1991. Guest editorial: it’s time to stop 

headstarting Kemp’s ridley. Marine Turtle 

Newsletter. 55: 7-8. 

14 

Spencer R-J et al. 2017 Critically Evaluating Best Management Practices for Preventing Freshwater 

Turtle Extinctions. Conservation Biology. In Press.

Preventing Freshwater Turtle Extinctions

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?