Nest site selection by flatback sea turtles: Characterization of nesting beach topography with airborne LiDAR

Rushan bin Abdul Rahman This thesis is presented as part of the requirement for the degree of Bachelor of

Science (Honours) in Environmental Science, School of Veterinary and Life Sciences, of Murdoch University

2018

i

Declaration

I declare this thesis is my own account of my research and contains as its main content work which

has not previously been submitted for a degree at any tertiary education institute.

Rushan bin Abdul Rahman (28 May, 2018)

Thesis Word Count: 9,371

Literature Review Word Count: 6,247

Total Word Count: 15,618

ii

Acknowledgements

I would like to first thank my supervisors Dr. Margaret E. Andrew and Prof. Lynnath E.

Beckley. I asked for a project involving sea turtles, geographic information systems, and the use of the

R statistical software, and I got exactly what I wanted. Thank you for the opportunity to work on and

learn from an incredibly interesting and complicated project, as well as your unending patience,

guidance, advice, and feedback over the past two years. Both of you have pushed and challenged me

to produce work I never thought I could produce, and I am a better scientist now because of it.

I want to thank Dr. Tony Tucker and Dr. Scott Whiting from the Department of Biodiversity,

Conservation and Attractions (DBCA) for always being available and swiftly responding to my

enquiries. I would also like to thank Dr. Scott Whiting and Prof. Lynnath E. Beckley for providing the

airborne LiDAR data flown by Airborne Research Australia as well as Pendoley Environmental Pty.

Ltd. for the airborne sea turtle survey of Eighty Mile Beach provided by Dr. Tony Tucker and Dr.

Scott Whiting. I would like to thank Murdoch University for partial funding support.

My deepest gratitude to my family, whose support allowed me to be in Australia. Without you,

this entire university experience would not have happened. Thank you for entertaining all my

nonsense, tears, highs, and lows. Most of all, thank you for always believing in me, even during the

times when I did not.

To my partner Zubaidah for having the patience to be away from me for so long, and for the

strength to stay with me for as long as you have. You are the rock that keeps me grounded while my

head is in the clouds. Thank you for reminding me to keep my eyes on the horizon and being there

when I needed it, even when there were times I could not do the same for you.

To my friends in Australia, you are my family away from home. Thank you for reminding me

that the only way to experience the world is not through maps and models, but to actually go out and

enjoy it. There are not enough words to express my gratitude to each and every one of you. There are

too many people to name, but you know who you are.

I cannot forget the academic and professional staff and students in Murdoch University who

have provided guidance, an outlet, and a shoulder to cry on throughout my time here. Many, if not all,

have helped me in ways you could not begin to imagine.

Last, but not least, to the healthcare staff in Murdoch and in Singapore, I thank you from the

bottom of my heart. These years were the most difficult years I ever had to face for reasons outside of

academia. Thank you for helping me face these years with courage during the times I frantically

searched for an exit. You know who you are, and I could not have done this without you.

iii

Abstract

Protecting sea turtle nesting beaches is a commonly used management strategy for sea turtle

conservation as gravid sea turtles are more vulnerable to predation and disturbance during oviposition.

Therefore, understanding where sea turtles nest and the characteristics of their nesting sites is

fundamental for their conservation. However, the process of nest site selection is not well understood.

Beach topography is one factor that is believed to drive nest site selection, but considering that many

nesting beaches, particularly in Australia, are large and remote, effectively acquiring beach

topographic information is challenging. Airborne light detection and ranging (LiDAR) data can obtain

accurate information on beach topography and can be used to understand how topography influences

nest site selection. The aims of this study were to (i) characterize beach topography using LiDAR data

of Eighty Mile Beach, a remote 220 km long beach in north-west Australia, and (ii) determine which

features influenced nest site selection of the flatback sea turtle (Natator depressus) population nesting

there. Metrics of beach features that were believed to be relevant for nest site selection were

calculated from piecewise linear regression models fitted to beach profiles extracted from transects

generated along Eighty Mile Beach. Aerial photographic survey data were used to quantify flatback

sea turtle nesting and were converted to presence-absence and nesting density classes per photograph.

Classification tree models were used to model the influence of beach topography on the density and

presence-absence of nests. Nests were predicted to be present if the highest elevation in the profile

was 9 to 12 m (overall accuracy = 60%, detection rate = 80%), though density could not be modelled

successfully (overall accuracy = 18%, detection rate = 80%). These results agree with the hypothesis

that the silhouette in front of a sea turtle influences nest placement, though moderate model

performance suggests that nest site selection by flabtack sea turtles at Eighty Mile Beach are also

influenced by other factors besides beach topography. The methods used in this study successfully

characterized the beach topography of Eighty Mile Beach using LiDAR data and have the potential to

be used on remote and large beaches in other parts of the world.

iv

Table of Contents

1 INTRODUCTION ..................................................................................................................... 1

1.1 SEA TURTLE NEST SITE SELECTION ...................................................................... 2

1.2 FLATBACK SEA TURTLES ......................................................................................... 3

1.3 APPLICATIONS OF ESTIMATING TOPOGRAPHY WITH LIGHT DETECTION

AND RANGING (LiDAR) .............................................................................................. 4

2 METHODS ................................................................................................................................ 5

2.1 STUDY SITE ................................................................................................................... 5

2.2 QUANTIFYING FLATBACK TURTLE NESTING ..................................................... 8

2.3 CHARACTERIZING BEACH TOPOGRAPHY .......................................................... 10

2.3.1 Airborne LiDAR Survey ....................................................................................... 10

2.3.2 Extraction of Beach Profiles ................................................................................. 11

2.3.3 Characterization of Beach Topography ................................................................ 11

2.4 MODELLING NEST SITE SELECTION .................................................................... 14

3 RESULTS ................................................................................................................................ 18

3.1 FLATBACK SEA TURTLE NESTING ALONG EIGHTY MILE BEACH ............... 18

3.2 EIGHTY MILE BEACH TOPOGRAPHIC CHARACTERISTICS ............................. 20

3.3 MODELLING NEST SITE SELECTION BY FLATBACK SEA TURTLES ............. 25

3.3.1 Presence and Absence of Flatback Sea Turtle Nesting Pits .................................. 25

3.3.2 Density of Flatback Sea Turtle Nesting Pits ......................................................... 29

4 DISCUSSION .......................................................................................................................... 33

4.1 CHARACTERIZATION OF BEACH TOPOGRAPHY .............................................. 33

4.2 FLATBACK SEA TURTLE NEST SITE SELECTION .............................................. 34

v

4.3 MODEL PERFORMANCE ........................................................................................... 37

5 CONCLUSION ........................................................................................................................ 39

6 REFERENCES ........................................................................................................................ 40

7 APPENDIX I: LITERATURE REVIEW ................................................................................ 49

vi

List of Tables

Table 1. Topographic metrics used to characterize beach topography with, descriptions and justification of

the metrics and references for their importance in nest site selection .......................................................... 14

Table 2 Assigned category of each photograph relative to the number of pits present, their frequency

within the dataset, the proportion of the dataset each category constitutes, and the weighting applied. ...... 17

Table 3. Range and mean (± standard deviation) of the beach metrics along Eighty Mile Beach derived

from LiDAR data. ......................................................................................................................................... 24

Table 4. Confusion matrix for the predicted and observed presence/absence of pits in a given photograph

when evaluated on the testing data. Values indicate number of photographs. ............................................. 25

Table 5. Confusion matrix for the predicted and observed density of pits in a given photograph when

evaluated on the testing data. Values indicate number of photographs. ....................................................... 29

vii

List of Figures

Figure 1. Boundary of Eighty Mile Beach Marine Park and its major management zones. PSZ = Pardoo

Sanctuary Zone, CKSZ = Cape Keraudren Sanctuary Zone, WRZ = Wallal Recreational Zone, WSPZ =

Waru Special Purpose Zone, PSPZ = Pilyarlkarra Special Purpose Zone, ASPZ = Anna Plains Sanctuary

Zone, JSPZ = Jangyjartiny Special Purpose Zone. ......................................................................................... 7

Figure 2. Regions of Eighty Mile Beach sampled during the aerial survey for flatback sea turtle pits by

Pendoley Pty. Ltd. on the 7th of January, 2014. .............................................................................................. 9

Figure 3. Example photos illustrating how overlaps between aerial survey photographs were managed.

Yellow circle highlights a pit, and yellow box highlights sea turtle tracks. The size of the sampling unit

equals the size of the photo minus the region of overlap on one side. The pit depicted in the photo is

assigned to the count for photo 2. ................................................................................................................. 10

Figure 4. Diagram to explain metrics used in analysis of beach topography. (a) Areas of the seaward side

of a profile that would be characterized in a piecewise linear regression (grey line). (b) Piecewise linear

regression segments fitted on the entire beach profile (black line), resulting in the seaward side of the

primary dune being poorly fitted. (c) Identification of the peak of the primary dune (yellow circle) with the

removal of the landward side of the primary dune and retention of the seaward side of the primary dune.

(d) Beach metrics derived from the piecewise linear regression. The only beach metric retained from the

landward side of the primary dune was the highest elevation in the beach profile (blue circle). ................. 13

Figure 5. The number of photographs relative to the number of pits in each photograph taken in the aerial

survey along Eighty Mile Beach. .................................................................................................................. 15

Figure 6. Number of pits in each photograph along Eighty Mile Beach, relative to distance from Cape

Keraudren. Each point in the graph represents the number of pits in each photograph, with darker shades of

grey indicating greater density of pits. The pink shaded area is the Wallal Recreation Zone, while the blue

shaded region is the Anna Plains Sanctuary Zone. ....................................................................................... 18

Figure 7. Spatial running mean of pits present and the observed presence-absence of nests in two sites at

Eighty Mile Beach. Photographs were assigned a value between 0 and 1 based on the proportion of

photographs that had a pit present within a 2.5 km radius of a given photograph (a). Insets map the

presence/absence of pits within photographs for the Wallal Recreational Zone (b) and the south-eastern

region of Anna Plains Sanctuary Zone (c). ................................................................................................... 19

Figure 8. Examples of beach profiles (grey lines) along Eighty Mile Beach with fitted piecewise linear

regression models (dashed red lines) segmented at breakpoints (yellow circles). RMSE for the fitted

viii

piecewise linear regression model is also provided. a = 0 km from Cape Keraudren, b = 75 km, c = 121

km, d = 157 km. ............................................................................................................................................ 20

Figure 9. Beach metric measures relative to their distance from Cape Keraudren. Greater opacity indicates

greater density of points Pink shaded region is the Wallal Recreation Zone, while the blue shaded area is

the Anna Plains Sanctuary Zone. (a) Angle of beach face before primary dune. (b) Angle of primary dune.

(c) Length of primary dune. (d) Elevation of primary dune. (e) Highest elevation in profile. (f) Number of

breaks in beach profile before primary dune. ............................................................................................... 22

Figure 10. Variable importance for jack-knifed models with photographs predicted to have pits. Overall

error (left), commission error (middle), and omission error (right) are provided for models that only use

one variable (top row) and models that remove one variable (bottom row). Each variable (blue bars) is

displayed along the x-axis and compared to the full models (orange bar). .................................................. 26

Figure 11. Classification tree modelling the presence or absence of a pit within a given photograph.

Terminal nodes identify the predicted response and the number of photographs classified into each node.

For these photographs, pie charts show the proportion with pits observed to be present (black) and absent

(white). .......................................................................................................................................................... 27

Figure 12. Mapped predictions for classification tree model. Photographs are assigned a value between 0

and 1 based on the proportion of photographs that are predicted to have a pit present within a 2.5 km radius

of a given photograph (a). Insets show photographs with the predicted presence/absence of pits at Wallal

Recreational Zone (b) and the south-eastern region of Anna Plains Sanctuary Zone (c). ............................ 28

Figure 13. Variable importance for jack-knifed models with photographs predicted to high pit density.

Overall error (left), commission error (middle), and omission error (right) are provided for models that

only use one variable (top row) and models that remove one variable (bottom row). Each variable (blue

bars) is listed along the x-axis and compared to the full (orange bar). ......................................................... 30

Figure 14. Classification tree modelling of the pit density categories within a given photograph. Terminal

nodes contain the predicted response and the number of observations classified into each node. For these

photographs, pie charts show the proportion with pits observed to be present (black) and absent (white). 31

Figure 15. Mapped predictions for classification tree model of different densities. Photographs are assigned

a value between 0 and 1 based on the proportion of photographs that are predicted to have a high pit

densities within a 2.5 km radius of a given photograph (top panel). Insets show areas where photographs

would be predicted to have high pit density at Wallal Recreational Zone (a) and the south-eastern region of

Anna Plains Sanctuary Zone (b). .................................................................................................................. 32

ix

1

1 INTRODUCTION

Sea turtles are a group of threatened charismatic megafauna with global distributions.

With the exception of the Polar Regions, their foraging, nesting, and migratory range span all

oceans of the world (Wallace et al. 2010). Seven species of sea turtle have been documented, all

of which are listed on the International Union for the Conservation of Nature Red List of

Threatened Species (Marine Turtle Specialist Group 1996; Red List Standards & Petitions

Subcommittee 1996; Abreu-Grobois and Plotkin 2008; Mortimer and Donnelly 2008; Seminoff

2004; Wallace et al. 2013; Casale and Tucker 2015). Though there are areas of the globe where

sea turtle populations are secure (e.g. green sea turtles in Hawaii, Balazs and Chaloupka 2004;

and the Great Barrier Reef, Hof et al. 2017; or Kemp’s ridley sea turtles in Florida, Shaver et

al. 2016), sea turtles in other regions face significant threats to their persistence. These include

loss of foraging and nesting habitats (e.g. Polidoro et al. 2017), poaching (Koch et al. 2006;

Mancini and Koch 2009; Guebert et al. 2013; Nuno et al. 2018), unregulated exploitation

(Tapilatu et al. 2017) and fisheries bycatch (Peckham et al. 2007; Riskas et al. 2018). These

threats can expose sea turtles to greater cumulative impacts throughout their life as they are

long-lived, wide-ranging animals with different life history stages occupying a variety of

habitats (Bolten 2003; Bjorndal et al. 2005).

Breeding and reproduction can be a particularly vulnerable life history stage (e.g.

Blamires et al. 2003; Sarahaizad et al. 2012) as females must emerge onto the beach to nest

while the laid clutch does not receive parental care during the entire duration of incubation.

Thus, successful nesting sites must maintain the safety of the nesting sea turtle during the

processes of digging the body pit, nest chamber construction, egg laying, and camouflaging

(Sarahaizad et al. 2012), as well as ensure the nest has the greatest chance of surviving heat

stress, predation, and tidal inundation during the incubation period (Leighton et al. 2011). This

means that the cues for nest placement are a conservation priority (Hamann et al. 2010) in order

to protect current nesting beaches (Wallace et al. 2011) as well as predict where nesting may

2

occur as coastlines are altered by sea level rise or more violent and frequent storms

(Poloczanska et al. 2010; Fuentes et al. 2011; Pike 2013). Protection of breeding grounds is a

common conservation strategy for the population recovery of wide ranging, migratory animals

(e.g. migratory birds, Wade and Hickey 2008; and nesting sea turtles, Nel et al. 2013). Indeed,

management plans around the world ensure that the protection of nesting beaches are a priority

for protected areas (e.g. flatback sea turtles, Department of Biodiversity, Conservation and

Attractions 2017; or other marine turtle species, Asian Development Bank 2011).

1.1 SEA TURTLE NEST SITE SELECTION

At the local scale, vegetation, sand characteristics, and beach topography have been

associated with nest site selection. Sea turtles appear to nest within or near vegetation (Kamel

and Mrosovsky 2005; Leighton et al. 2011; da Silva et al. 2016), while the amount of moisture

retained in the sand may aid in nest chamber construction (Bustard and Greenham 1968;

Garmestani et al. 2000; Chen et al. 2007). However, sand that is too moist can impede gas

exchange (Ackerman 1977) and could be an indicator the sea turtle is below the high tide line.

Beach topography might be a strong driver for nest site selection, as cues indicating the nesting

sea turtle is above the high tide line means the nest clutch would not be inundated by the tide

during incubation (Whitmore and Dutton 1985; Miller et al. 2003). Beach elevation may be an

important factor for this reason; studies at resolutions of the nesting site (Kikukawa et al. 1999;

Katselidis et al. 2013) and at the scale of beaches (Yamamoto et al. 2012; Dunkin et al. 2016)

have found beach elevation to be an important predictor of nest site selection. Another study

postulates that landward silhouettes in front of the sea turtle are an important cue for nest

placement (Salmon et al. 1995; Witherington et al. 2011a, 2011b), whether it is a dune

(Witherington et al. 2011b), a cliff-face (Mazor et al. 2013), clumps of vegetation (Hays and

Speakman 1993), an unused building (Salmon et al. 1995), or a temporarily erected barrier

(Witherington et al. 2011a). Topographic heterogeneity was also found to be an important driver

for nest site selection. For example, leatherback sea turtles in Florida, USA, nested in areas of

the beach that were higher than the surrounding area (Yamamoto et al. 2012), while loggerhead

3

sea turtles on the Sunshine Coast, Australia nested on areas of the dune where there was greater

concavity at the dune toe and were not highly rugged (Kelly et al. 2017).

Although considerable research has investigated sea turtle nest site selection,

uncertainties still remain. There are also gaps in the distribution of research effort by species

and by nesting region. Of the 34 studies reviewed from the literature (see Appendix I: Literature

Review), 19 of these studies were conducted on loggerhead sea turtles, of which 13 came from

the United States. In contrast, flatback sea turtles, a species of sea turtle that can only be found

in Australia, was investigated by only 4 studies (Limpus 1971; Blamires et al. 2003; Koch and

Guinea 2006; Bannister et al. 2016).

1.2 FLATBACK SEA TURTLES

Flatback sea turtles (Natator depressus) are the only species of sea turtle without an

oceanic post-hatchling or juvenile phase (Walker and Parmenter 1990), resulting in their

endemism to the continental shelf waters of Australia (Limpus 2007). The IUCN lists them as

data deficient (Red List Standards & Petitions Subcommittee 1996) but they are listed as

vulnerable throughout Australia (Department of Environment and Energy 2018). There have

been several recent studies on the flatback sea turtle populations in Western Australia (see

Appendix I: Literature Review), but few studies investigated nest site selection for this species

throughout Australia. These have generally supported the role of beach topography: initial

observations by Limpus (1971) found that flatback sea turtles in southern Queensland selected

nesting sites on the top of dunes. Another study found the steepness of the dunes in Fog Bay,

Northern Territory, confined the flatback sea turtles to nest at the dune base (Blamires et al.

2003). Two studies at Bare Island, Northern Territory, found flatback sea turtles nested on

beaches with gently sloping dunes (Bannister et al. 2016; Koch and Guinea 2006).

These studies were carried out on beaches that were relatively short (between 1 and 30

km), but other flatback sea turtle nesting beaches can be expansive and remote, such as Cape

Domett (Whiting et al. 2008), Barrow Island (Pendoley et al. 2014), or the 220 km long Eighty

Mile Beach (Department of Parks and Wildlife 2014) in Western Australia. Long and remote

4

beaches used for other sea turtle species can also be found around Australia. For example, the

closest settlement to loggerhead, green, and possibly hawksbill sea turtle nesting sites at

Gnaraloo Bay and Cape Farquhar (Riskas 2014; Thomson et al. 2016) is Carnarvon at 120 km.

Obtaining data relevant for nest site selection from such remote and long beaches can be

logistically challenging.

1.3 APPLICATIONS OF ESTIMATING TOPOGRAPHY WITH LIGHT DETECTION AND RANGING (LiDAR)

Advances in technology that remotely obtain information on beach topography can be

useful for studies investigating nest site selection by sea turtles. Traditional methods for

measuring beach topography are generally restricted in the number of observations that can be

made and are generally limited to the nest site itself (Wood and Bjorndal 2000; Mazaris et al.

2006; Zavaleta-Lizarraga and Morales-Mavil 2013), and are less able to comprehensively

characterize the entire beach. Such techniques may be more difficult to conduct at remote and

expansive beaches, such as many nesting beaches in Australia. Existing global elevation

datasets have resolutions that are too coarse to capture the subtle nuances of the beach profile

that may be relevant for a sea turtle in its selection of a nesting site. For example, the Shuttle

Radar Topography Mission as a resolution of 30 m and does not have data for intertidal habitats

(NASA Jet Propulsion Laboratory (JPL) 2013).

Light detection and ranging (LiDAR), which captures detailed 3D measurements of the

Earth’s surface may be ideal for these purposes. LiDAR technology is an active sensor that fires

pulses of light to measure the distance of objects. Distance to objects are determined by the time

it takes for the pulse of light to return to the sensor. When attached to an aircraft, elevation can

be determined with geo-referenced points where the lasers contact the ground. These can be

used to obtain data that captures fine-scale topographic variability. This detailed information on

landscape topography has been used to model suitable habitats. This has been particularly useful

when the distribution of an organism is reliant on the topography of the landscape, such as

predicting the distribution of plants (Sellars and Jolls 2007; Andrew and Ustin 2009; Questad et

5

al. 2014) or mapping suitable nesting sites for sea turtles (Yamamoto et al. 2012; Dunkin et al.

2016).

LiDAR was also the preferred method of obtaining topographic information of coastlines

because of its high vertical accuracy and ease at which large areas can be mapped quickly

(Mason et al. 2000). For these dynamic landscapes, LiDAR scans taken over time (Woolard and

Colby 2002; Shrestha et al. 2005; Kerfoot et al. 2012) or before and after storms (Zhang et al.

2005; Stockdon et al. 2009) can accurately quantify changes in the beach topography. These

studies were able to rapidly obtain data on topography across large areas of beach, which

highlights the potential for the use of airborne laser scans of expansive and remote beaches in

Australia. Further, given that many of these beaches are nesting sites for endemic flatback sea

turtles, there is also an opportunity to determine the influences of topography on their nest site

selection.

The overall aim of the current study was to determine how beach topography affects

flatback sea turtle nest site selection. Specifically the study objectives were to (i) characterize

the topography of the extensive and remote beach with airborne LiDAR data; and (ii) to

determine the topographic feature(s) that influence nest site selection by flatback sea turtles.

2 METHODS

2.1 STUDY SITE

This study investigated sea turtle nest site selection at Eighty Mile Beach, a 220 km long

beach situated in northwest Western Australia within the Eighty Mile Beach Marine Park

(Figure 1). It is considered to be a non-continuous, tide-dominated, ultra-dissipative beach

broken by bluffs and tidal and river inlets (Short 2006a) with high wave energy in the southwest

(Cape Keraudren) but lower wave energy in the north (Cape Missiessy) (Piersma et al. 1999).

The tidal range along Eighty Mile Beach varies from 4 to 10 m, resulting in wide intertidal areas

6

that can reach 4 km (Short 2006a, 2006b), such as those found at the Anna Plains Sanctuary

Zone in the north. These shallow and wide intertidal zones are thought to attenuate wave energy

(Piersma et al. 1999), allowing for the deposition of marine-derived carbonate mud in the

middle and upper intertidal environments (Semeniuk 2008). On the other hand, the high-energy

south-western region of Eighty Mile Beach has much of its sand derived from Mesozoic-tertiary

bedrock or reworked quaternary desert dunes (Semeniuk 2008). A majority of the Eighty Mile

Beach coastal areas are backed by an expansive fore dune system (Short 2006b) formed from

sedimentation deposited into the Canning Basin (Department of Parks and Wildlife 2014;

Semeniuk 2008). The exception to this is an area south of the Anna Plains Sanctuary Zone,

where a mangrove meets the ocean and has no dune development (Short, 2006b).

Climate in north-west Australia is characterized by hot and wet summers and dry winters

(Bureau of Meteorology 2015; O’Donnell et al. 2015). Precipitation in the region has been

increasing in the last two centuries (Smith 2004; Feng et al. 2013; O’Donnell et al. 2015; Ren

and Leslie 2015) and mainly occurs during the austral summer-autumn period (December –

March) with average annual rainfall exceeding 310 mm (O’Donnell et al. 2015; Bureau of

Meterology 2018). Tropical cyclones are common during this period and are the greatest

contributor to rainfall in north-western Australia (Ng et al. 2015). This period also experiences

the highest average monthly maximum air temperatures, reaching up to 40⁰ Celsius (O’Donnell

et al. 2015). The dry season (April – October) has less rainfall and cooler temperatures, with

average monthly maximum air temperatures reaching a minimum of 25⁰ and an average

monthly rainfall not exceeding 10 mm (Bureau of Meteorology 2015; O’Donnell et al. 2015).

Flatback sea turtles nest across Eighty Mile Beach (Pendoley Environmental 2017) during

the austral summer (Department of Parks and Wildlife 2014). The highest density of nests

occurs within the Wallal Recreation Zone adjacent to Eighty Mile Beach Caravan Park and

within the Anna Plains Sanctuary Zone (Figure 1; Department of Parks and Wildlife 2014).

7

Figure 1. Boundary of Eighty Mile Beach Marine Park and its major management zones. PSZ = Pardoo Sanctuary Zone, CKSZ = Cape Keraudren Sanctuary Zone, PwSPZ

= Paruwuturr Special Purpose Zone, WRZ = Wallal Recreational Zone, WSPZ = Waru Special Purpose Zone, PSPZ = Pilyarlkarra Special Purpose Zone, ASPZ = Anna

Plains Sanctuary Zone, JSPZ = Jangyjartiny Special Purpose Zone.

Cape Missiessy

Cape Keraudren

PwSPZ

8

2.2 QUANTIFYING FLATBACK TURTLE NESTING

An aerial survey conducted by Pendoley Pty. Ltd. was flown along the shoreline from

Broome to Port Hedland at low altitude (200 m) during the flatback sea turtle nesting season on

7th January 2014 to document nesting activity on the beach. Three areas with a cumulative total

distance of about 25 km (or 11%) of Eighty Mile Beach were not surveyed during the flight

mission as these were characterized as tidal or river inlets, which are not normal nesting areas

for sea turtles (Figure 2; Weishampel et al. 2003; Yalçın-Özdilek et al. 2007) and to conserve

battery power. Geotagged photographs (sampling units) were taken with a Sony SLT-A99V

digital SLR camera at a frequency of 40 photos/minute with ~60 m spacing between photograph

centroids and approximately 10 – 20% overlap between photographs. Only nesting pits (herein

known as “pits”) within each photograph were manually counted to determine the distribution

of pits across Eighty Mile Beach. A pit was counted when a stark depression in the sand was

seen within the photograph and were considered proxies for ideal nesting habitat (as pits could

be abandoned nesting attempts). The pits within the overlapped regions between two

photographs were managed by assigning the pit to the count of the north-eastern photograph in

order to maintain consistency of sampling units (approximately 60 m) and avoid double

counting (Figure 3).

9

Figure 2. Regions of Eighty Mile Beach sampled during the aerial survey for flatback sea turtle pits by Pendoley Pty. Ltd. on the 7th of January, 2014.

Wetland area

10

Figure 3. Example photos illustrating how overlaps between aerial survey photographs were

managed. Yellow circle highlights a pit, and yellow box highlights sea turtle tracks. The size of the

sampling unit equals the size of the photo minus the region of overlap on the north-eastern side. The

pit depicted in the photo is assigned to the count for photo 2.

2.3 CHARACTERIZING BEACH TOPOGRAPHY

2.3.1 Airborne LiDAR Survey

The airborne LiDAR flight mission was flown between Port Hedland and Broome by

Airborne Research Australia (www.airborneresearch.org.au) on 24th and 25th September, 2015

using a light aircraft equipped with the LiDAR RIEGL Q680i-S scanner. The flight path was

centred above the dune crest to ensure the swath (approximately 450 m) included the upper

intertidal zone, beach, primary dune, and some landward dunes. Point clouds were returned at

approximately 10 points/m2 with a nominal vertical precision of 0.1 m (2 standard deviations)

and vertical accuracy of 0.25 m (1 standard deviation). These point clouds were then used to

generate digital elevation models (0.5 m horizontal resolution), which were provided by the

contractor. Though the time between the aerial photographic survey of turtle nesting and the

11

LiDAR flight mission was 20 months, there was no cyclone activity at Eighty Mile Beach in

between these times (Bureau of Meterology 2018), so the coastline is expected to have

undergone little change

2.3.2 Extraction of Beach Profiles

The digital elevation models were used to obtain beach profiles along the entire length of

Eighty Mile Beach. A contour line was first generated from the digital elevation model to

delineate the general direction of the beach. Contour line elevations were specific to 10 km

stretches of beach as continuous contour lines of on 10-km section were different from the

continuous contour line of another section. Transects were then established perpendicular to the

contour line at a 10-m spacing, a distance argued to be wide enough before variability between

transects would increase. The elevation and coordinates of the digital elevation model were

extracted from each pixel along each transect to represent vertical profiles of the beach. Contour

lines and transects were made in ArcMap 10.3.1 (ESRI, Redlands CA) while the extraction of

beach profiles was carried out using the Region of Interest tool within ENVI 5.3.1 (Exelis

Visual Informaiton Solutions, Venice). As Eighty Mile Beach Marine Park is within two UTM

zones, digital elevation models were split and processed into their respective coordinate

systems: WGS 84 UTM50S and UTM51S. A UTM projected coordinate system was used over

a geographic coordinate system as UTM minimized visual distortion within a localized area.

2.3.3 Characterization of Beach Topography

Beach profiles can have complex shapes that may provide relevant cues to nesting sea

turtles. For example, the beach profile from sea to land will change in angle and elevation

between the beach face, the primary dune toe, and the peak of the primary dune. Piecewise

linear regression models were used to characterize the complexity of the beach profile as this

method is able to distinguish several regression models within a dataset without being prompted

where specific breaks between regression models may lie (Muggeo, 2003). Using a beach

profile as an example (Figure 4a), a linear piece would be generated along the beach face. As

the beach profile changes from the beach face to the primary dune face, the model can no longer

12

maintain the best linear fit, and a breakpoint is generated at the dune toe. This is followed by the

generation of another linear segment to be fitted on the primary dune face. This process

continues piecewise along the entire beach profile, resulting in several linear segments

connected at breakpoints. The segmented package (Muggeo, 2003, 2008) in R version 3.4.3 (R

Core Team 2017) was used to carry out the piecewise linear regression modelling.

For each profile, a series of models were generated with sequentially increasing numbers

of breakpoints up to a maximum of 15 breakpoints; among these, the model providing the best

fit of the beach profile was selected using the Akaike Information Criterion. Preliminary results

revealed that, for many profiles, the majority of the allowed breakpoints were used to

characterize the complex terrain on the landward side of the primary dune, resulting in poor

model fit on the seaward side (Figure 4b). As a result, two rounds of piecewise linear

regression models were conducted. The first round identified the peak of the primary dune,

which allowed for the landward side of the profile to be removed. The second round of

piecewise linear regression modelling was then carried out on the seaward side of the primary

dune, allowing all breakpoints to be used to effectively characterize subtle changes in the beach

profile (Figure 4c). The breakpoints and slopes of different parts of the beach profile derived

from the piecewise linear regression models were used to quantify beach topographic

characteristics expected to be relevant for sea turtle nest site selection, such as the angle of the

primary dune and the elevation of the primary dune peak (Figure 4d; Table 1). The only

information retained from the landward side of the primary dune was the highest peak in the

profile.

13

Figure 4. Diagram to explain metrics used in the analysis of beach topography. (a) Areas of the

seaward side of a profile that would be characterized in a piecewise linear regression (grey line). (b)

Piecewise linear regression segments fitted on the entire beach profile (black line), resulting in the

seaward side of the primary dune being poorly fitted. (c) Identification of the peak of the primary

dune (yellow circle) with the removal of the landward side of the primary dune and retention of the

seaward side of the primary dune. (d) Beach metrics derived from the piecewise linear regression.

14

Table 1. Topographic metrics used to characterize beach topography with descriptions and

justification of the metrics and references for their importance in nest site selection.

Metric

Justification

Source Angle of beach face

before primary dune

A smaller angle at the beach face followed by a steeper angle at the primary dune would be an indicator to a sea turtle that it has reached an area above the high tide line

Wood and Bjorndal 2000; Katselidis et al. 2013; Roe et al. 2013

Angle of primary

dune

A steeper angle may be an indicator the

sea turtle transitioned to an area of beach above the high tide line

Horrocks and Scott 1991; Wood

and Bjorndal 2000; Cuevas et al. 2010; Spainer 2010; Zare et al. 2012; Katselidis et al. 2014

Length of primary

dune face The space on the primary dune on

which the sea turtle would be able to nest in

Limpus 1971; Salmon and Witherington 1995; Blamires et al. 2003

Elevation of

primary dune peak

Tall landward objects in front of a sea turtle have been found to be an important factor in nest site selection in other studies

Kikukawa et al. 1999; Yamamoto et al. 2012; Dunkin et al. 2016

Highest elevation in profile

Silhouettes in front of the sea turtle are hypothesized to be an environmental cue for nest placement

Salmon et al. 1995; Witherington et al. 2011b, 2011a; Mazor et al. 2013

Number of breaks before primary dune

An index for the complexity that indicates beach morphology

Whitmore and Dutton 1985; Hays et al. 1995; Rizkalla and Savage 2011; Yamamoto et al. 2012; Kelly et al. 2017

2.4 MODELLING NEST SITE SELECTION

As 25 km of Eighty Mile Beach was not surveyed during the 2014 aerial survey, 195 km

(88%) of Eighty Mile Beach was used in the analysis. Pits found within each photograph were

visually interpreted as nesting pits of flatback sea turtles. Counts of pits within each photograph

ranged from 0 to 8 (Figure 5) with the majority showing no pits. These counts were converted to

the presence or absence of a pit or the density class of pits in a photograph, using the Jenks

natural breaks optimization in ArcMap 10.3.1 to divide nesting density into low (1 pit) or high

(2 – 8 pits) density classes (Jenks 1967). Nest presence/absence and density classes were

modelled against characteristics of beach topography using classification tree analyses.

15

Figure 5. The number of photographs relative to the number of pits in each photograph taken in the

aerial survey along Eighty Mile Beach.

Classification tree modelling is a non-parametric machine learning technique that can be

used in predictive modelling (Breiman et al. 1983). This modelling technique has been found to

perform well in species distribution models (Seguardo and Araujo 2004; Elith and Graham

2009; Elith et al. 2006; Andrew and Ustin 2009) and has shown good predictive power in

determining sea turtle nest site selection under current climate conditions (Mazaris et al. 2006),

after anthropogenic impacts (Rizkalla and Savage 2011), and in future climate-change

conditions (Mazaris et al. 2015). Classification trees recursively split a dataset at nodes along

the explanatory variable that groups the data into as homogenous groups as possible. The

process continues iteratively until the data can no longer be partitioned into homogenous

groups, ending at a terminal node with a predicted value (Franklin, 2009; Loh, 2011; Strobl et

al., 2009; Therneau and Atkinson, 2018a). In this study, classification trees were used to

determine how beach topography influences nest site selection by flatback sea turtles using the

rpart package (Therneau and Atkinson 2018b) in R version 3.4.3 (R Core Team 2017). The

16

analysis units for these models were the beach lengths sampled by individual photographs in the

flatback sea turtle aerial nesting surveys; henceforth, analysis units will be referred to as

“photographs” for brevity.

Classification trees are sensitive to imbalances within the dataset (Chawla 2003; Menardi

and Torelli 2014), which were prevalent in this study as photographs with no pits were

substantially more frequent than photographs with pits (Figure 5). Imbalance can be corrected

by weighting rare classes to give them equal influence to the model as common classes.

Weights used in presence/absence and density class models were calculated from the frequency

of each class in the photographs and are provided in Table 2.

The data was geographically sub-sampled into training and testing sets. Spatial

autocorrelation between photograph centroids was accounted for by grouping photographs into

10 km segments, followed by subsampling every fourth segment as testing data. This method

has shown to reduce spatial autocorrelation between testing and training data (Boria et al. 2014;

Radosavljevic and Anderson 2014; Jiménez-Alfaro et al. 2018). The same training and test

subsets were used when modelling for both the presence or absence of pits and the pit densities.

Model performance was evaluated by calculating the overall error, omission error, and

commission error. For the presence/absence model in the current study, omission error is the

proportion of photographs with pits observed but were predicted to be absent, while the

commission error is the proportion of photographs predicted to have pits but were not observed

to have pits. For the density model, omission error is the proportion of photographs with high

pit density that were predicted to be other classes, while the commission error is the proportion

of photographs that were predicted to have high density but were observed to be other classes.

Variables were jack-knifed to determine their importance in the classification tree

models. This technique creates alternative models with subsets of the variables, and assesses

changes in model performance when leaving out a focal variable, and when creating univariate

models with only the focal variable. Variable importance was evaluated with the overall error,

commission error, and omission error of jack-knifed models applied to testing data relative to

17

the full model, though focus was given to omission error, which was judged to be the most

important type of prediction error in this context. For the density class models, focus was given

to the high-density class as it was of interest to predict where areas of high nesting density was

occurring across Eighty Mile Beach.

Table 2 Assigned category of each photograph relative to the number of pits present, their frequency

within the dataset, the proportion of the dataset each category constitutes, and the weighting applied.

Model Number of pits Category Frequency Weight

Presence/absence 0 Absent 1890 1.00

1 or more Present 586 3.23

Nest density/photograph

0 None 1890 1.00

1 Low 343 5.51

2 - 8 High 243 7.78

18

3 RESULTS

3.1 FLATBACK SEA TURTLE NESTING ALONG EIGHTY MILE BEACH

The photographic aerial survey yielded 2,476 photographs of Eighty Mile Beach. Of

these photographs, 1,890 had no evidence of any pits and 586 with visible pits. Within each

photograph that had pits visible, the number of pits ranged from 1 (n = 343) to 8 (n = 1), with a

total of 972 pits recorded. Figure 6 shows the number of pits within each photograph along

Eighty Mile Beach relative to their distance from Cape Keraudren, the southern-most landmark

in the study extent. Several of the photographs with high pit densities were found within the

Wallal Recreation Zone and Anna Plains Sanctuary Zone. Figure 7 shows the spatial running

mean of photographs with pits across Eighty Mile Beach.

Figure 6. Number of pits in each photograph along Eighty Mile Beach, relative to distance from

Cape Keraudren in the southwest. Each point in the graph represents the number of pits in each

photograph, with darker shades of grey indicating greater density of pits. The pink shaded area is the

Wallal Recreation Zone, while the blue shaded region is the Anna Plains Sanctuary Zone.

19

Figure 7. Spatial running mean of pits present and the observed presence-absence of nests in two sites at Eighty Mile Beach. Photographs were assigned a value between 0

and 1 based on the proportion of photographs that had a pit present within a 2.5 km radius of a given photograph (a). Insets map the presence/absence of pits within

photographs for the Wallal Recreational Zone (b) and the south-eastern region of Anna Plains Sanctuary Zone (c).

a

b

c

b c

PwSPZ

20

3.2 EIGHTY MILE BEACH TOPOGRAPHIC CHARACTERISTICS

Using the LiDAR data, six beach metrics were calculated for 20,105 cross-shore transects

along Eighty Mile Beach. Examples of beach profiles and the piecewise linear regression

models fitted to them are given in Figure 8, which show how well the model fits to the beach

profile. The root mean square error (RMSE) of the piecewise linear regression models along

transects ranged from 0.00 to 2.27 m (mean ± s.d. = 0.30 ± 0.24 m), indicating they successfully

characterized the shape of the profiles.

Figure 8. Examples of beach profiles (grey lines) along Eighty Mile Beach with fitted piecewise

linear regression models (dashed red lines) segmented at breakpoints (yellow circles). RMSE for the

a RMSE = 0.15 m Length of primary dune face = 34 m Angle of primary dune = 12⁰

b RMSE = 0.29 m Length of primary dune face = 128 m Angle of primary dune = 5⁰

c RMSE = 0.13 m Length of primary dune face = 34 m Angle of primary dune = 11⁰

d RMSE = 0.07 m Length of primary dune face = 136 m Angle of primary dune = 1⁰

Distance Along Profile (m) Distance Along Profile (m)

Distance Along Profile (m) Distance Along Profile (m)

21

fitted piecewise linear regression model is also provided. a = 0 km from Cape Keraudren, b = 75 km,

c = 121 km, d = 157 km.

Figure 9 shows how the measures of the beach metrics vary with position along the

beach, measured as the distance from the start of the beach at Cape Keraudren. Overall, there is

large variability in the measurements of the beach metrics across Eighty Mile Beach except for

the regions in Anna Plains Sanctuary Zone and about 130 to 150 km from Cape Keraudren. The

area 130 to 150 km from Cape Keraudren is characterized by a long section of marsh and saltl

flats that extend directly onto the intertidal zone, and hence has very little dune development.

The angle of the beach face was steepest between 60 and 110 km from Cape Keraudren, with a

decline in the beach angle between 120 km and Anna Plain Sanctuary Zone (Figure 9a). This

was followed by an increase within the sanctuary zone until Cape Missiessy in the north. Some

beach faces had negative angles, or angles of depression, which generally occurred in areas of

river and tidal inlets near the Waru Special Purpose Zone (Figure 9a). The steepest primary

dunes were found between 60 and 100 km from Cape Keraudren, near Pardoo Station, while

primary dunes with the smallest angles were found between 130 and 160 km (Figure 9b).

Primary dune faces were longer the farther they were from Cape Keraudren, with the longest

primary dune face (270 m) occurring in Anna Plains Sanctuary Zone 191 km from Cape

Keraudren (Figure 9c). The tallest primary dunes were between 60 and 100 km from Cape

Keraudren in the western region of the beach, followed by the primary dunes in Anna Plains

Sanctuary Zone in the north. The shortest primary dunes were found between 130 and 150 km

from Cape Keraudren (Figure 9d). Profiles had higher peaks between 60 and 80 km from Cape

Keraudren, followed by a decrease between 130 and 150 km (Figure 9e). Beach profiles

appeared to be more complex, with a greater number of breakpoints used in the piecewise

regression models, between 90 and 130 km from Cape Keraudren, followed by lower

complexity at 180 km and a slight increase number of breaks from 200 km to Cape Missiessy

(Figure 9f).

22

Figure 9. Beach metric measures relative to their distance from Cape Keraudren. Greater opacity

indicates greater density of points Pink shaded region is the Wallal Recreation Zone, while the blue

shaded area is the Anna Plains Sanctuary Zone. (a) Angle of beach face before primary dune. (b)

Angle of primary dune. (c) Length of primary dune.

a

b

c

23

Figure 9 (cont.). Beach metric measures relative to their distance from Cape Keraudren. Greater

opacity indicates greater density of points Pink shaded region is the Wallal Recreation Zone, while

the blue shaded area is the Anna Plains Sanctuary Zone. (d) Elevation of primary dune. (e) Highest

elevation in profile. (f) Number of breaks in beach profile before primary dune.

d

e

f

24

Table 3 shows the summary statistics for each beach metric. These metrics were not

highly correlated with each other; the minimum correlation was –0.55 (primary dune length and

breaks in beach profile) and the maximum correlation was 0.41 (elevation of primary dune peak

and primary dune length).

Table 3. Range and mean (± standard deviation) of the beach metrics along Eighty Mile Beach

derived from LiDAR data.

Range Mean ± standard deviation

Angle of beach face before primary dune

-19⁰ – 28⁰ 2⁰ ± 2⁰

Angle of primary dune

0⁰ – 74⁰ 10⁰ ± 9⁰

Length of primary dune

1 – 270 m 51 ± 47 m

Elevation of primary dune peak

0 – 25 m 8 ± 3 m

Highest elevation in profile

2 – 26 m 11 ± 4 m

No. breaks before primary dune

1 – 15 5 ± 3

25

3.3 MODELLING NEST SITE SELECTION BY FLATBACK SEA TURTLES

3.3.1 Presence and Absence of Flatback Sea Turtle Nesting Pits

The distribution model performed moderately with an overall error (when applied to the

test dataset) of about 40%, while the omission error (20%) was substantially lower than its

commission error (66%, Table 4). When applying the jack-knifed one-variable models to the

testing data, the angle of the primary dune, the highest elevation in the profile, and the number

of breaks before the primary dune were the most important variables in predicting the presence

of pits, as their use resulted in a low omission error comparable to that of the full model (Figure

10). Removed-variable models (when evaluated against the testing data) indicated the most

important variable was the highest elevation in the profile as its removal resulted in the greatest

gain in omission error relative to the full model.

Table 4. Confusion matrix for the predicted and observed presence/absence of pits in a given

photograph when evaluated on the testing data. Values indicate number of photographs.

Observed

Presence Absence Row total Commission Error

Pred

icte

d

Presence 118 230 348 66.1%

Absence 29 258 287 10.1%

Column Total 147 488 635

Omission Error 19.7% 47.1%

Accuracy 59.2%

26

Figure 10. Variable importance for jack-knifed models with photographs predicted to have pits.

Overall error (left), commission error (middle), and omission error (right) are provided for models

that only use one variable (top row) and models that remove one variable (bottom row). Each

variable (blue bars) is displayed along the x-axis and compared to the full models (orange bar).

27

Consistent with the variable importance results, the full model only included the highest

elevation in the beach profile term, despite being supplied with all six variables to use in

constructing the model. The model predicted that profiles with intermediate heights (9 – 12 m)

would have pits present; beach profiles less than or greater than this were predicted to not have

any pits (Figure 11). When mapping the model results across Eighty Mile Beach, profiles were

predicted to have suitable nest sites around the Wallal Recreation Zone, the Waru Special

Purpose Zone, and the Anna Plains Sanctuary Zone. A small region with several photographs

predicted to have pits was found south-west of the Paruwuturr Special Purpose Zone in the

south-eastern region of Eighty Mile Beach (Figure 12).

Figure 11. Classification tree modelling the presence or absence of a pit within a given photograph.

Terminal nodes identify the predicted response and the number of photographs classified into each

node. For these photographs, pie charts show the proportion with pits observed to be present (black)

and absent (white).

< 9 m ≥ 9 m

< 12 m ≥ 12 m

28

Figure 12. Mapped predictions for classification tree model. Photographs are assigned a value between 0 and 1 based on the proportion of photographs that are predicted to

have a pit present within a 2.5 km radius of a given photograph (a). Insets show photographs with the predicted presence/absence of pits at Wallal Recreational Zone (b) and

the south-eastern region of Anna Plains Sanctuary Zone (c).

a

b

c

b c

PwSPZ

29

3.3.2 Density of Flatback Sea Turtle Nesting Pits

The density model for flatback sea turtle nesting performed poorly, with an overall error

of 82% when applied to the testing data (Table 5). The omission error for high pit density (20%)

was substantially lower than the commission error (85%), while both error metrics were

considerably worse when modelling low pit density (omission error = 69%, commission error =

90%). The high omission error associated to modelling low densities is due to the model not

being able to distinguish between areas of low pit densities with areas of high pit densities,

while the high commission errors are driven by the model predicting there are areas of high and

low pit density where none were observed. Jack-knifed one-variable models show that the most

important variable was the number of breaks in the profile before the primary dune as it resulted

in the lowest omission error for the high-density class when it was the only variable used in the

model (Figure 13). The most important variables in jack-knifed removed-variable tests for high

pit densities were the length of the primary dune face, the elevation of the primary dune peak,

and the number of breaks in the profile before the primary dune. The removal of these variables

resulted in greater omission errors when compared to the full model.

Table 5. Confusion matrix for the predicted and observed density of pits in a given photograph when

evaluated on the testing data. Values indicate number of photographs.

Observed

High Low None Row total Commission Error

Pred

icte

d

High 49 54 231 334 85.33

Low 10 27 218 255 89.4%

None 2 5 39 46 15.2%

Column Total 61 86 488 635

Omission Error 19.7% 68.6% 92.0%

Accuracy 18.1%

30

Figure 13. Variable importance for jack-knifed models with photographs predicted to have high pit

density. Overall error (left), commission error (middle), and omission error (right) are provided for

models that only use one variable (top row) and models that remove one variable (bottom row). Each

variable (blue bars) is listed along the x-axis and compared to the full model (orange bar).

31

The full model of nest density classes made use of two variables: highest elevation in the

beach profile and the elevation of the primary dune. This discrepancy from the variable

importance results likely arose because the assessment of variable importance emphasized

omission error rates of the high density class, while model construction attempted to minimize

confusion between all density classes. Photographs were predicted to have lower nesting

densities when profiles had higher peak elevations (greater than 9 m) and lower primary dune

peak elevations (< 1 m). In cases where the primary dunes were higher (≥ 1 m), photographs

were predicted to have high nesting densities when the highest elevation in the profile was less

than 12 m. Profiles with a peak elevation lower than 9 m or greater than 12 m did not have any

nests predicted in their photographs(Figure 14), similar to the distribution model (Figure 11).

Mapping the model predictions across Eighty Mile Beach reveals that areas predicted to have

high pit densities were within the Anna Plains Sanctuary Zone, the north-eastern part of the

Wallal Recreation Zone, and south-east of the Paruwuturr Special Purpose Zone (Figure 15).

Figure 14. Classification tree modelling of the pit density categories within a given photograph.

Terminal nodes contain the predicted response and the number of observations classified into each

node. For these photographs, pie charts show the proportion with pits observed to be in high density

(black), low density (grey), and where none were observed (white).

< 9 m ≥ 9 m

< 1 m ≥ 1 m

< 12 m ≥ 12 m

32

Figure 15. Mapped predictions for classification tree model of different densities. Photographs are assigned a value between 0 and 1 based on the proportion of photographs

that are predicted to have a high density of pits within a 2.5 km radius of a given photograph (top panel). Insets show areas where photographs would be predicted to have

high pit density at Wallal Recreational Zone (a) and the south-eastern region of Anna Plains Sanctuary Zone (b).

a

b

c

b c

PwSPZ

33

4 DISCUSSION

The aims of this study were to develop a method to characterize the topography of a

remote and expansive beach and determine which features of its topography influenced sea

turtle nest site selection. The use of airborne LiDAR addressed the challenge of efficiently

obtaining beach topography data from the 220-km long coastline of Eighty Mile Beach at

resolutions high enough to capture fine-scale variations. Piecewise linear regression models

were then used to quickly calculate ecologically relevant metrics that were tailored to the needs

of modelling flatback sea turtle nest site selection. This study highlights the ease at which

beaches of this type can be swiftly characterized.

4.1 CHARACTERIZATION OF BEACH TOPOGRAPHY

This study demonstrates that the use of airborne LiDAR can be useful in obtaining

information that is relevant for ecological research, especially for studies of this type. Such

LiDAR data sources are becoming readily available; for example, more than 700 km of the

south-west coast of Australia has been mapped with airborne LiDAR (Department of Planning,

Lands and Heritage 2017), while United States coastlines are periodically surveyed by LiDAR,

and the data made publicly available (NOAA 2018). However, to effectively extract information

from LiDAR-derived digital elevation models in a manner that characterizes different features

of the beach can be challenging. Moving window analyses are commonly performed across

rasterized digital elevation models and can estimate topographic features such as slope or

relative elevation (e.g. Sellars and Jolls 2007). However, beaches have distinct, intuitively

recognizable features, such as beach slopes and dunes, which are organized in a way that is

consistent with other beach features in the profile. These would not easily be captured by a

moving window approach because it ignores the essentially linear nature of beach systems. The

extraction of profiles along the beach and fitting piecewise linear segments to these profiles

allows these distinct beach features to be identified and measured. Quantitative information on

34

each feature can then be quickly calculated from the fitted linear segments, which can be

customized to the requirements of coastal or conservation managers (Mason et al. 2000).

Here, six beach metrics were calculated from fitted piecewise linear regression models of

20,105 coastal profiles at a spacing of 10 m across the 220-km coastline of Eighty Mile Beach.

These metrics were tailored to capture features hypothesised to be relevant for nest site

selection: the angle of the beach face before the primary dune, the angle of the primary dune,

the elevation of the primary dune peak, length of the primary dune, the highest elevation in the

profile, and the number of breaks in the profile before the primary dune. The good fit of the

piecewise linear regression models (mean RMSE ± s.d. = 0.30 ± 0.24 m) suggests that the

measurements calculated from the fitted linear segments are a reasonable reflection of the

LiDAR scans of Eighty Mile Beach topography. These can then be corroborated by comparison

of these results to previous site visits and descriptions of Eighty Mile Beach (Short 2006a,

2006b). For example, in the current study, the angle of the beach face before the primary dune

was consistently low (mean ± s.d. = 2⁰ ± 2⁰), which agrees with previous work describing

Eighty Mile Beach as an ultra-dissipative beach (Short 2006a, 2006b). Another detail captured

by the LiDAR data was an area lacking dunes approximately 130 to 150 km from Cape

Keraudren, which was described previously as a 16 km stretch of coastline characterized by

mangroves and several small drainage creeks where dunes are absent (Short 2006b). This

method thus shows the use of piecewise linear regressions to extract several metrics has the

potential to be useful in several areas of ecological research, such as sea turtle habitat use.

4.2 FLATBACK SEA TURTLE NEST SITE SELECTION

The current study found the flatback sea turtle population at Eighty Mile Beach nested in

areas where the highest elevation in the profile was between 9 and 12 m. Other studies indicate

flatback sea turtles preferentially nested on the dune face above the high tide line, such as those

in Mon Repos, Queensland (Limpus 1971), Crab Island, Queensland (Limpus et al. 1983),

Mundabullangana, Western Australia (Blamires et al. 2003), and Bare Sand Island, Northern

Territory (Bannister et al. 2016). The selection of a particular site on the dune may be dependent

35

on the ease at which the flatback sea turtle is able to scale the dunes (Koch and Guinea 2006;

Bannister et al. 2016). Flatback sea turtles in Fog Bay, Northern Territory, were confined to the

dune base as they could not scale the steep gradient of the dune (Blamires et al. 2003), while

those at Bare Island, Northern Territory, nested on the top of the dune because of the low dune

gradient (Koch and Guinea 2006; Bannister et al. 2016). Though the flatback sea turtles at

Curtis Island, south-east Queensland were found to scale the steep dunes to nest at the dune

peak, this is not a usual occurrence (Limpus 1971).

However, the results of these other studies are not directly comparable with the current

study as they described the exact nesting site relative to the cross-section of the beach. In

contrast, the current study considered nest site selection in the perpendicular direction,

determining how variation between beach profiles influences nest placement of flatback sea

turtles in Eighty Mile Beach. By considering the metrics of the entire beach profile, the current

study obtained results that are consistent with the conclusions of other studies that the silhouette

in front of the sea turtle is an important environmental cue for nest site selection (Mazaris et al.

2006; Witherington et al. 2011a, 2011b). Witherington et al. (2011a) erected barriers in front of

loggerhead sea turtles as they were emerging onto the beach found that the sea turtles began nest

construction about 3 m away from the barrier, making no attempt to move around or climb over

the barrier (e.g. as leatherbacks did with an escarpment, Roe et al. 2013). They concluded that

the silhouette of a landward feature in front of the sea turtle was the cue for nesting

(Witherington et al. 2011a). These results may explain why the highest elevation in the profile

may be the most important factor in the presence or high density of flatback sea turtle nests at

Eighty Mile Beach. Differences between the present findings and previous investigations of

flatback turtle nesting behaviour may also be due to population-specific characteristics. The

flatback sea turtle population at Eighty Mile Beach is its own genetic stock (Tucker et al. 2015),

and thus the influences that drive nest site selection by the flatback sea turtles at Eighty Mile

Beach may be unique to this population. Population-specific nesting behaviour has also been

seen in hawksbill sea turtles, for example, where those in French West Indies preferred nesting

36

in low-lying vegetation or on sand adjacent to the forest edge (Kamel and Delcroix 2009), while

those in El Salvador preferred to nest amongst the dense woody vegetation (Liles et al. 2015).

Two other studies have utilized LiDAR-derived beach measurements to determine the

influence of beach topography on nest site selection on other sea turtle species. Both these

studies were conducted on the east coast of Florida (Yamamoto et al. 2012: approximately 85

km; Dunkin et al. 2016: 200 km), but obtained different results to the present study. Yamamoto

et al. (2012) found that higher beach elevations were the strongest predictor for loggerhead sea

turtle nesting, that green sea turtle nesting was correlated with deeper offshore approaches, and

that leatherback sea turtles nested on areas of beach with a higher topographic positional index.

The second study (Dunkin et al. 2016) found that a beach elevation of 5 m was the optimal

elevation for nesting by loggerhead sea turtles, which also yielded the lowest omission error

(11%). Both these studies appear to mirror other findings that found beach elevation was

important for nest placement by sea turtles, as this may be a cue the sea turtle is above the high

tide line (Horrocks and Scott 1991; Kikukawa et al. 1999; Katselidis et al. 2013).

However, direct comparisons between the two Florida studies and the current study are

difficult. It should be noted that the current study did not use beach elevation as an explanatory

variable as it was collinear with other beach metrics of the model. In addition to using different

suites of explanatory variables than the current study, the previous studies also used different

scales of analyses and modelled different response variables. Both studies in Florida predicted

density of nests in 20 km segments of beach, while explanatory variables were summarized

statistics at the same sampling unit. This was due to the fact that information on sea turtle

nesting was only available at beach level. In contrast, sea turtle nesting data used in the current

study was available at a resolution of 60 m. These differences in resolution can yield different

results. Models developed from coarser resolutions can perform poorly because the factors that

influence species distributions at larger scales may not be the same processes at work with

smaller resolutions (Nyström Sandman et al. 2013). In other words, the predicted distribution of

sea turtle nesting pits modelled from summarized elevations across 20 km might not be

37

ecologically relevant for a sea turtle that is selecting a suitable nesting site at a scale of 5 m (but

see below for discussion on modelling distribution vs. density). Modelling nest site selection at

larger resolutions may be appropriate for coarse-scale biogeographical studies (see Austin and

Van Niel 2011), but not for the purpose of predicting suitable habitat (e.g. Nezer et al. 2017).

Species distribution models that are developed on data available at finer resolutions can reveal

the processes that drive nest site selection at these smaller scales.

4.3 MODEL PERFORMANCE

The presence/absence model for flatback sea turtle nest site selection performed

moderately (overall error = 40%), but the density model performed poorly (overall error =

80%). The difference in performance between the two models may be based on the resolution of

the response variable; distribution models have a tendency to perform better than models that

attempt to predict abundance (Pearce and Ferrier 2001). Such examples can be found when

predicting the abundance and occurrence of red kites across the Iberian Peninsula (Seoane et al.

2003), breeding aggregations of raptors (Estrada and Arroyo 2012), seabirds in marine protected

areas (Oppel et al. 2012), or moose in Canada (Michaud et al. 2014). The better performance of

distribution models may be attributed to the fact that the presence, rather than abundance, of an

organism can be driven by less complex multivariate relationships between environmental

variables at a given place or time (Pearce and Ferrier 2001). On the other hand, attempting to

predict abundance can prove difficult as varying abundances may be driven by complex

combinations between environmental parameters (Michaud et al. 2014) or because confounding

variables not accounted for can result in models that are oversimplified (Merow et al. 2014). In

such cases, predicting the occurrence of an organism would allow for lower overall error, as in

the current study.

There are several types of errors in classification analyses, which have different

implications to the useability of the results. In the current study, omission error was given

greater emphasis in evaluating model performance as it was determined to be more important to

ensure that areas where nests were observed were correctly predicted by the model (Franklin

38

2009). If the purpose was to ensure that the majority of nesting areas were correctly predicted

by the model, both the distribution and density model can be useful in predicting the presence or

the high density of pits (omission error = 20%). Though commission error was not given as

much importance as omission error, the high commission errors in both the distribution and

density models in predicting presence or high density of pits indicate the models over-estimated

where pits may occur across Eighty Mile Beach. This could be due to areas with nesting pits

were similar to other possibly suitable beaches that did not have any pits observed, and hence

predicting those areas to have nests when none were observed. However, the nest survey data

used for the current study was a single day within an entire nesting season. Thus, recorded

absences in this dataset may not reflect a true absence of pits, and hence a truly unsuitable site

for nesting (MacKenzie et al. 2002). The addition of nesting data evenly distributed in time

across the season could improve the performance of the model.

The moderate performance of the presence/absence model in this study, however, may be

an indicator that other variables besides beach topography may influence nest site selection. At

coarser scales, it has been hypothesized that offshore bathymetry was an important driver for

nesting beach selection (Mortimer 1982; Yamamoto et al. 2012). At the scale of the nesting site

itself, vegetation appears to play an important role in nest site selection, as sea turtles may nest

within or near vegetation (Kamel and Mrosovsky 2005; Leighton et al. 2011; da Silva et al.

2016) as this may be an indicator to the sea turtle that is it above the high tide line. Sea turtles

may also nest within the vicinity of vegetation due to the sand being more stabilized by

vegetation roots (Bustard and Greenham 1968; Chen et al. 2007), making construction of the

nest chamber easier as it would not continuously collapse during construction. In this regard,

sand moisture might also play an important role in easing nest construction, as drier sand tends

to be looser (Garmestani et al. 2000; Wood and Bjorndal 2000; Chen et al. 2007). Sand moisture

might be helpful to a certain threshold (Chen et al. 2007); sand that is too moist can impede gas

exchange (Ackerman 1977) or be an indicator to the sea turtle that it is below the high tide line.

Nesting density also appeared to decrease in closer proximity to river inlets (Weishampel et al.

2003; Yalçın-Özdilek et al. 2007). Vegetation and river inlets can be detected by LiDAR, so

39

even if the current project only focused on the influence of topography on nest site selection,

including vegetation and river inlets in the presence/absence and density models might improve

their performance.

5 CONCLUSION

The approach developed in this study allowed the topography of a remote and long beach

to be characterized and models to be developed that predicted the presence/absence and density

of flatback turtle nests along Eighty Mile Beach. The presence/absence of nests was found to be

associated with the highest elevation in a profile between 9 and 12 m. These findings agree with

the conclusion of other studies that the silhouette of a landward feature in front of a sea turtle is

an important environmental cue for nest placement. The presence/absence model performed

only moderately, while the density of nests could not be successfully modelled. This restricts

the usefulness of the density model to managers as identifying areas of high density may be of

greater conservation interest. The poor and moderate performance of the density and

presence/absence models may be attributed to the study using sea turtle nesting data that was

from a single day in an entire nesting season or that nest site selection is driven by

environmental cues other than topography. The inclusion of nesting data from more days in the

nesting season as well as other variables that can be estimated from the LiDAR data – such as

the presence of vegetation or river inlets – may improve the performance of both models. The

current study demonstrates that characterizing long and remote beaches is possible, and that the

methods used here can be applied to similar beaches in other parts of the world. Given the

importance of coastal environments, such as habitat for the threatened charismatic megafauna

like sea turtles, other studies such as these can be invaluable to informing management for their

conservation.

40

6 REFERENCES

Abreu-Grobois, A., Plotkin, P. (2008) Lepidochelys olivacea. IUCN Red List of Threatened Species, e.T11534A3292503.

Ackerman, R. A. (1977) The respiratory gas exchange of sea turtle nests. Respiration Physiology. 31, 19-38.

Andrew, M. E., Ustin, S. L. (2009) Habitat suitability modelling of an invasive plant with advanced remote sensing data. Diversity and Distributions. 15, 627–640.

Asian Development Bank (2011) Comprehensive action plans of the Sulu–Sulawesi Marine Ecoregion: A priority seascape of the Coral Triangle Initiative. Asian Development Bank, Mandaluyong City, Philippines.

Austin, M. P., Van Niel, K. P. (2011) Improving species distribution models for climate change studies: Variable selection and scale. Journal of Biogeography. 38, 1–8.

Balazs, G. H., Chaloupka, M. (2004) Thirty-year recovery trend in the once depleted Hawaiian green sea turtle stock. Biological Conservation. 117, 491–498.

Bannister, N., Holland, J., Farrelly, T. (2016) Nest site fidelity of flatback turtles (Natator depressus) on Bare Sand Island, Northern Territory, Australia. Northern Territory Naturalist. 27, 47–53.

Pendoley Environmental. (2017) Eighty Mile Beach Marine Park marine turtle aerial surveys 2012, 2014, & 2016. Pendoley Environmental Pty. Ltd., Perth, Australia.

Bjorndal, K., Bolten, A., Chaloupka, M. (2005) Evaluating trends in abundance of immature green turtles, Chelonia mydas, in the Greater Caribbean. Ecological Applications. 15, 304–314.

Blamires, S. J., Guinea, M. L., Prince, R. I. T. (2003) Influence of nest site selection on predation of flatback sea turtles (Natator depressus) eggs by varanid lizards in Northern Australia. Chelonian Conservation and Biology. 4, 1–7.

Bolten, A. B. (2003) Variation in sea turtle life history patterns: Neritic vs. oceanic developmental stages. In: The Biology of Sea Turtles: Volume II (eds P. L. Lutz, J. A. Musick, & J. Wyneken) pp. 243–257, CRC Press, Boca Raton, FL.

Boria, R. A., Olson, L. E., Goodman, S. M., Anderson, R. P. (2014) Spatial filtering to reduce sampling bias can improve the performance of ecological niche models. Ecological Modelling. 275, 73–77.

Breiman, L., Friedman, J. H., Olshen, R. A., Stone, C. J. (1983) Classification and Regression Trees. Chapman & Hall, Boca Raton, Florida.

Bureau of Meteorology (2015) Australian Climate Influences. [online]. Available from: http://www.bom.gov.au/watl/about-weather-and-climate/australian-climate-influences.shtml [Accessed May 10, 2018].

Bureau of Meterology (2018) Southern Hemisphere Tropical Cyclone Data Portal. [online]. Available from: http://www.bom.gov.au/cyclone/history/tracks/index.shtml.

Bustard, H. R., Greenham, P. (1968) Physical and chemical factors affecting hatching in the green sea turtle, Chelonia mydas (L .). Ecology. 49, 269–276.

41

Casale, P., Tucker, A. D. (2015) Caretta caretta. IUCN Red List of Threatened Species, e.T3897A83157651.

Chawla, N. V. (2003) C4.5 and imbalanced data sets: Investigating the effect of sampling method, probabilistic estimate, and decision tree structure. In: Workshop on Learning from Imbalanced Datasets II pp. 1–8Washington, DC.

Chen, H.-C., Cheng, I.-J., Hong, E. (2007) The influence of the beach environment on the digging success and nest site distribution of the green turtle, Chelonia mydas, on Wan-an Island, Penghu Archipelago, Taiwan. Journal of Coastal Research. 5, 1277–1286.

Cuevas, E., de los Ángeles Liceaga-Correa, M., Mariño-Tapia, I. (2010) Influence of beach slope and width on hawksbill (Eretmochelys imbricata) and green turtle (Chelonia mydas) nesting activity in El Cuyo, Yucatán, Mexico. Chelonian Conservation and Biology. 9, 262–267.

Department of Biodiversity, Conservation and Attractions (2017) North West Shelf Flatback Turtle Conservation Program Strategic Conservation Plan 2014–21. Department of Biodiversity, Conservation and Attractions, Perth, Australia

Department of Environment and Energy (2018) Natator depressus – Flatback Turtle. Species Profile Threat. Database. [online]. Available from: http://www.environment.gov.au/cgi-bin/sprat/public/publicspecies.pl?taxon_id=59257 [Accessed May 10, 2018].

Department of Parks and Wildlife (2014) Eighty Mile Beach Marine Park Management Plan 80 2014–2024. Department of Parks and Wildlife, Perth, Australia.

Department of Planning Lands and Heritage (2017) LiDAR Survey and Bathymetric Mapping. Coastal vulnerability - LiDAR Survey and Bathymetric Mapping [online]. Available from: https://www.planning.wa.gov.au/675.aspx [Accessed May 19, 2018].

Dunkin, L., Reif, M., Altman, S., Swannack, T. (2016) A spatially explicit, multi-criteria decision support model for loggerhead sea turtle nesting habitat suitability: A remote sensing-based approach. Remote Sensing. 8, 1–22.

Elith, J., Graham, C. H. (2009) Do they? How do they? WHY do they differ? On finding reasons for differing performances of species distribution models. Ecography. 32, 66–77.

Elith, J., Graham, C. H., Anderson, R. P. R. P., Dudik, M., Ferrier, S., Guisan, A., Hijmans, R. J., Huettmann, F., Leathwick, J. R., Lehmann, A., Li, J., Lohmann, L. G., Loiselle, B. A., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J. M. C., Peterson, A. T., Phillips, S. J., et al. (2006) Novel methods improve prediction of species’ distributions from occurrence data. Ecography. 29, 129–151.

ESRI (2011) ArcGIS v 10.3.1. Environmental Systems Research Institute, Redlands, CA.

Estrada, A., Arroyo, B. (2012) Occurrence vs abundance models: Differences between species with varying aggregation patterns. Biological Conservation. 152, 37–45.

Exelis Visual Informaiton Solutions (2010) ENVI 5.3.1. Exelis Visual Information Solutions, Boulder, CO.

Feng, J., Li, J., Xu, H. (2013) Increased summer rainfall in northwest Australia linked to southern indian ocean climate variability. Journal of Geophysical Research: Atmospheres. 118, 467–480.

Franklin, J. (2009) Mapping species distributions: Spatial inference and prediction. Cambridge University Press, New York.

42

Fuentes, M. M. P. B., Limpus, C. J., Hamann, M. (2011) Vulnerability of sea turtle nesting grounds to climate change. Global Change Biology. 17, 140–153.

Garmestani, A. S., Percival, H. F., Portier, K. M., Rice, K. G. (2000) Nest-site selection by the loggerhead sea turtle in Florida’s Ten Thousand Islands. Journal of Herpetology. 34, 504–510.

Guebert, F. M., Barletta, M., da Costa, M. F. (2013) Threats to sea turtle populations in the Western Atlantic: Poaching and mortality in small-scale fishery gears. Journal of Coastal Research. 65, 42–47.

Hamann, M., Godfrey, M. H., Seminoff, J. A., Arthur, K., Barata, P. C. R., Bjorndal, K. A., Bolten, A. B., Broderick, A. C., Campbell, L. M., Carreras, C., Casale, P., Chaloupka, M., Chan, S. K. F., Coyne, M. S., Crowder, L. B., Diez, C. E., Dutton, P. H., Epperly, S. P., Fitz Simmons, N. N., Formia, A., et al. (2010) Global research priorities for sea turtles: Informing management and conservation in the 21st century. Endangered Species Research. 11, 245–269.

Hays, G. C., Mackay, A., Adams, C. R., Mortimer, J. A., Speakman, J. R., Boerema, M. (1995) Nest site selection by sea turtles. Journal of the Marine Biological Association of the United Kingdom. 75, 667–674.

Hays, G. C., Speakman, J. R. (1993) Nest placement by loggerhead turtles, Caretta caretta. Animal Behaviour. 45, 47–53.

Hof, C., Smallwood, E., Meager, J., Bell, I. (2017) First citizen-science population abundance and growth rate estimates for green sea turtles Chelonia mydas foraging in the northern Great Barrier Reef, Australia. Marine Ecology Progress Series. 574, 181–191.

Horrocks, J. A., Scott, N. M. (1991) Nest site location and nest success in the hawksbill turtle Eretmochelys imbricata in Barbados West-Indies. Marine Ecology Progress Series. 69, 1–8.

Jenks, G. F. (1967) The data model concept in statistical mapping. International Yearbook of Cartography. 7, 186-190.

Jiménez-Alfaro, B., Suárez-Seoane, S., Chytrý, M., Hennekens, S. M., Willner, W., Hájek, M., Agrillo, E., Álvarez-Martínez, J. M., Bergamini, A., Brisse, H., Brunet, J., Casella, L., Dítě, D., Font, X., Gillet, F., Hájková, P., Jansen, F., Jandt, U., Kącki, Z., Lenoir, J., et al. (2018) Modelling the distribution and compositional variation of plant communities at the continental scale. Diversity and Distributions. 1–13.

Kamel, S. J., Delcroix, E. (2009) Nesting ecology of the hawksbill turtle, Eretmochelys imbricata, in Guadeloupe, French West Indies from 2000-07. Journal of Herpetology. 43, 367–376.

Kamel, S. J., Mrosovsky, N. (2005) Repeatability of nesting preferences in the hawksbill sea turtle, Eretmochelys imbricata, and their fitness consequences. Animal Behaviour. 70, 819–828.

Katselidis, K. A., Schofield, G., Stamou, G., Dimopoulos, P., Pantis, J. D. (2013) Evidence-based management to regulate the impact of tourism at a key marine turtle rookery on Zakynthos Island, Greece. Oryx. 47, 1–11.

Katselidis, K. A., Schofield, G., Stamou, G., Dimopoulos, P., Pantis, J. D. (2014) Employing sea-level rise scenarios to strategically select sea turtle nesting habitat important for long-term management at a temperate breeding area. Journal of Experimental Marine Biology and Ecology. 450, 47–54.

43

Kelly, I., Leon, J. X., Gilby, B. L., Olds, A. D., Schlacher, T. A. (2017) Marine turtles are not fussy nesters: A novel test of small-scale nest site selection using structure from motion beach terrain information. PeerJ. 5, e2770.

Kerfoot, W. C., Yousef, F., Green, S. A., Regis, R., Shuchman, R., Brooks, C. N., Sayers, M., Sabol, B., Graves, M. (2012) Light detection and ranging (LiDAR) and multispectral studies of disturbed Lake Superior coastal environments. Limnology and Oceanography. 57, 749–771.

Kikukawa, A., Kamezaki, N., Ota, H. (1999) Factors affecting nesting beach selection by loggerhead turtles (Caretta caretta): A multiple regression approach. Journal of Zoology. 249, 447–454.

Koch, A. U., Guinea, M. L. (2006) Lower nesting success of flatback turtles caused by disorientation. Marine Turtle Newsletter, 114, 16.

Koch, V., Nichols, W. J., Peckham, H., De La Toba, V. (2006) Estimates of sea turtle mortality from poaching and bycatch in Bahia Magdalena, Baja California Sur, Mexico. Biological Conservation. 128, 327–334.

Leighton, P. A. Ã., Horrocks, J. A., Kramer, D. L. (2011) Predicting nest survival in sea turtles: When and where are eggs most vulnerable to predation? Animal Conservation. 14, 186–195.

Liles, M. J., Peterson, M. J., Seminoff, J. A., Altamirano, E., Henríquez, A. V., Gaos, A. R., Gadea, V., Urteaga, J., Torres, P., Wallace, B. P., Peterson, T. R. (2015) One size does not fit all: Importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the eastern Pacific Ocean. Biological Conservation. 184, 405–413.

Limpus, C. J. (1971) The flatback turtle, Chelonia depressa Garman in southeast Queensland, Australia. Herpetologica. 27, 431–446.

Limpus, C. J. (2007) A biological review of Australian marine turtles. 5. Flatback turtle Natator depressus (Garman). (ed L. Fien). Queensland Environmental Protection Agency, Brisbane.

Limpus, C. J., Parmenter, C. J., Baker, V., Fleay, A. (1983) The Crab Island sea turtle rookery in the north-eastern Gulf of Carpentaria. Australian Wildlife Research. 10, 173–184.

Loh, W.-Y. (2011) Classification and regression trees. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery. 1, 14–23.

MacKenzie, D. I., Nichols, J. D., Lachman, G. B., Droege, S., Royle, A. A., Langtimm, C. A. (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology. 83, 2248–2255.

Mancini, A., Koch, V. (2009) Sea turtle consumption and black market trade in Baja California Sur, Mexico. Endangered Species Research. 7, 1–10.

Marine Turtle Specialist Group (1996) Lepidochelys kempii. IUCN Red List of Threatened Species, e.T11533A3292342.

Mason, D. C., Gurney, C., Kennett, M. (2000) Beach topography mapping: A comparison of techniques. Journal of Coastal Conservation. 6, 113–124.

Mazaris, A. D., Matsinos, Y. G., Margaritoulis, D. (2006) Nest site selection of loggerhead sea turtles: The case of the island of Zakynthos, W Greece. Journal of Experimental Marine Biology and Ecology. 336, 157–162.

44

Mazaris, A. D., Vokou, D., Almpanidou, V., Türkozan, O., Sgardelis, S. (2015) Low conservatism of the climatic niche of sea turtles and implications for predicting future distributions. Ecosphere. 6, 1–12.

Mazor, T., Levin, N., Possingham, H. P., Levy, Y., Rocchini, D., Richardson, A. J., Kark, S. (2013) Can satellite-based night lights be used for conservation? The case of nesting sea turtles in the Mediterranean. Biological Conservation. 159, 63–72.

Menardi, G., Torelli, N. (2014) Training and assessing classification rules with imbalanced data. Data Mining and Knowledge Discovery. 28, 92–122.

Merow, C., Smith, M. J., Edwards, T. C., Guisan, A., Mcmahon, S. M., Normand, S., Thuiller, W., Wüest, R. O., Zimmermann, N. E., Elith, J. (2014) What do we gain from simplicity versus complexity in species distribution models? Ecography. 37, 1267–1281.

Michaud, J. S., Coops, N. C., Andrew, M. E., Wulder, M. A., Brown, G. S., Rickbeil, G. J. M. (2014) Estimating moose (Alces alces) occurrence and abundance from remotely derived environmental indicators. Remote Sensing of Environment. 152, 190–201.

Miller, J. D., Limpus, C. J., Godfrey, M. H. (2003) Nest site selection, oviposition, eggs, development, hatching and emergence of loggerhead turtles. In: The Biology of Loggerhead Sea Turtles (eds A. Bolten & B. Witherington) pp. 125–143, Smithsonian Institution Press, Washington, DC.

Mortimer, J. A. (1982) Factors influencing beach selection nesting turtles. In: Biology and Conservation of Sea Turtles (ed K. A. Bjorndal) pp. 45–51Smithsonian Institution Press Pages, Washington, DC.

Mortimer, J. A., Donnelly, M. (2008) Eretmochelys imbricata. IUCN Red List of Threatened Species, e.T8005A12881238.

Muggeo, V. M. R. (2003) Estimating regression models with unknown break-points. Statistics in Medicine. 22, 3055–3071.

Muggeo, V. M. R. (2008) segmented: An R package to fit regression models with broken-line relationships. R News. 8, 20–25.

NASA Jet Propulsion Laboratory (JPL) (2013) NASA Shuttle Radar Topography Mission United States 1 arc second. NASA EOSDIS L. Process. DAAC. [online]. Available from: http://dx.doi.org/10.5067/ASTER/AST_L1T.003 [Accessed May 20, 2018].

Nel, R., Punt, A. E., Hughes, G. R. (2013) Are coastal protected areas always effective in achieving population recovery for nesting sea turtles? PLoS One. 8, e63525.

Nezer, O., Bar-David, S., Gueta, T., Carmel, Y. (2017) High-resolution species-distribution model based on systematic sampling and indirect observations. Biodiversity Conservation. 26, 421–437.

Ng, B., Walsh, K., Lavender, S. (2015) The contribution of tropical cyclones to rainfall in northwest Australia. International Journal of Climatology. 35, 2689–2697.

NOAA (2018) United States Interagency Elevation Inventory. DIGITALCOAST. [online]. Available from: https://coast.noaa.gov/inventory/ [Accessed April 17, 2018]

Nuno, A., Blumenthal, J. M., Austin, T. J., Bothwell, J., Ebanks-Petrie, G., Godley, B. J., Broderick, A. C. (2018) Understanding implications of consumer behavior for wildlife farming and sustainable wildlife trade. Conservation Biology. 32, 390–400.

Nyström Sandman, A., Wikström, S. A., Blomqvist, M., Kautsky, H., Isaeus, M. (2013) Scale-

45

dependent influence of environmental variables on species distribution: A case study on five coastal benthic species in the Baltic Sea. Ecography. 36, 354–363.

O’Donnell, A. J., Cook, E. R., Palmer, J. G., Turney, C. S. M., Page, G. F. M., Grierson, P. F. (2015) Tree rings show recent high summer-autumn precipitation in northwest Australia is unprecedented within the last two centuries. PLoS One. 10, 1–19.

Oppel, S., Meirinho, A., Ramírez, I., Gardner, B., O’Connell, A. F., Miller, P. I., Louzao, M. (2012) Comparison of five modelling techniques to predict the spatial distribution and abundance of seabirds. Biological Conservation. 156, 94–104.

Pearce, J., Ferrier, S. (2001) The practical value of modelling relative abundance of species for regional conservation planning: A case study. Biological Conservation. 98, 33–43.

Peckham, S. H., Maldonado Diaz, D., Walli, A., Ruiz, G., Crowder, L. B., Nichols, W. J. (2007) Small-scale fisheries bycatch jeopardizes endangered Pacific loggerhead turtles. PLoS One. 2, 1–6.

Pendoley, K. L., Bell, C. D., McCracken, R., Ball, K. R., Sherborne, J., Oates, J. E., Becker, P., Vitenbergs, A., Whittock, P. A. (2014) Reproductive biology of the flatback turtle Natator depressus in Western Australia. Endangered Species Research. 23, 115–123.

Piersma, T., Pearson, G. B., Hickey, R., Lavaleye, M. (1999) The long mud: Benthos and shorebirds of the foreshore of Eighty-mile Beach, Western Australia. Royal Netherlands Institute for Sea Research, Texel, Netherlands.

Pike, D. A. (2013) Climate influences the global distribution of sea turtle nesting. Global Ecology and Biogeography. 22, 555–566.

Polidoro, B. A., Ralph, G. M., Strongin, K., Harvey, M., Carpenter, K. E., Arnold, R., Buchanan, J. R., Camara, K. M. A., Collette, B. B., Comeros-Raynal, M. T., De Bruyne, G., Gon, O., Harold, A. S., Harwell, H., Hulley, P. A., Iwamoto, T., Knudsen, S. W., Lewembe, J. de D., Linardich, C., Lindeman, K. C., et al. (2017) The status of marine biodiversity in the Eastern Central Atlantic (West and Central Africa). Aquatic Conservation: Marine and Freshwater Ecosystems. 27, 1021–1034.

Poloczanska, E. S., Limpus, C. J., Hays, G. C. (2010) Vulnerability of marine turtles to climate change. In: Advances in Marine Biology (ed D. W. Sims) pp. 151–211. Elsevier Ltd, London, UK.

Questad, E. J., Kellner, J. R., Kinney, K., Cordell, S., Asner, G. P., Thaxton, J., Diep, J., Uowolo, A., Brooks, S., Inman-Narahari, N., Evans, S. A., Tucker, B. (2014) Mapping habitat suitability for at-risk plant species and its implications for restoration and reintroduction. Ecological Applications. 24, 385–395.

R Core Team (2017) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Australia. https://www.R-project.org/.

Radosavljevic, A., Anderson, R. P. (2014) Making better MAXENT models of species distributions: Complexity, overfitting and evaluation. Journal of Biogeography. 41, 629–643.

Red List Standards & Petitions Subcommittee (1996) Natator depressus. IUCN Red List of Threatened Species, e.T14363A4435952.

Ren, D., Leslie, L. M. (2015) Changes in tropical cyclone activity over Northwest Western Australia in the past 50 years and a view of the future 50 years. Earth Interactions. 19, 1–22.

46

Riskas, K. (2014) Discovery of new nesting beaches for loggerhead turtles (Caretta caretta) at Gnaraloo on the Ningaloo Coast, Western Australia. Marine Turtle Newsletter. 140, 4–6.

Riskas, K. A., Tobin, R. C., Fuentes, M. M. P. B., Hamann, M. (2018) Evaluating the threat of IUU fishing to sea turtles in the Indian Ocean and Southeast Asia using expert elicitation. Biological Conservation. 217, 232–239.

Rizkalla, C. E., Savage, A. (2011) Impacts of seawalls on loggerhead sea turtle (Caretta caretta) nesting and hatching success. Journal of Coastal Research. 27, 166–173.

Roe, J. H., Clune, P. R., Paladino, F. V (2013) Characteristics of a leatherback nesting beach and implications for coastal development. Chelonian Conservation and Biology. 12, 34–43.

Salmon, M., Reiners, R., Lavin, C., Wyneken, J. (1995) Behavior of loggerhead sea turtles on an urban beach. I. Correlates of nest placement. Journal of Herpetology. 29, 560–567.

Salmon, M., Witherington, B. E. (1995) Artificial lighting and seafinding by loggerhead hatchlings: Evidence for lunar modulation. Copeia. 4, 931–938.

Sarahaizad, M. S., Shahrul Anuar, M. S., Mansor, Y. (2012) Nest site selection and digging attempts of green turtles (Chelonia mydas, Fam. Cheloniidae) at Pantai Kerachut and Telok Kampi, Penang Island, Peninsular Malaysia. Malaysian Applied Biology. 41, 31–39.

Seguardo, P., Araujo, M. B. (2004) An evaluation of methods for modelling species distributions. Journal of Biogegoraphy. 31, 1555–1568.

Sellars, J. D., Jolls, C. L. (2007) Habitat modeling for Amaranthus pumilus: An application of light detection and ranging (LIDAR) data. Journal of Coastal Research. 23, 1193–1202.

Semeniuk, V. (2008) Holocene sedimentation, stratigraphy, biostratigraphy, and history of the Canning Coast, north-western Australia. Journal of the Royal Society of Western Australia. 91, 53–148.

Seminoff, J. A. (2004) Chelonia mydas. IUCN Red List of Threatened Species, e.T4615A11037468.

Seoane, J., Viñuela, J., Díaz-Delgado, R., Bustamante, J. (2003) The effects of land use and climate on red kite distribution in the Iberian peninsula. Biological Conservation. 111, 401–414.

Shaver, D. J., Lamont, M. M., Maxwell, S., Walker, J. S., Dellingham, T. (2016) Head-started Kemp’s ridley turtle (Lepidochelys kempii) nest recorded in Florida: Possible implications. Chelonian Conservation and Biology. 15, 138–143.

Short, A. D. (2006a) Australian beach systems—nature and distribution. Journal of Coastal Research. 221, 11–27.

Short, A. D. (2006b) Beaches of the Western Australian Coast: Eucla to Roebuck Bay. Sydney University Press, Sydney, Australia.

Shrestha, R. L., Carter, W. E., Sartori, M., Luzum, B. J., Slatton, K. C. (2005) Airborne laser swath mapping: Quantifying changes in sandy beaches over time scales of weeks to years. ISPRS journal of Photogrammetry & Remote Sensing. 59, 222–232.

da Silva, V. R. F., Mitraud, S. F., Ferraz, M. L. C. P., Lima, E. H. S. M., Melo, M. T. D., Santos, A. J. B., da Silva, A. C. C. D., de Castilhos, J. C., Batista, J. A. F., Lopez, G. G., Tognin, F., Thomé, J. C., Baptistotte, C., da Silva, B. M. G., Becker, J. H., Wanderline,

47

J., de Vasconcellos Pegas, F., Róstan, G., dei Marcovaldi, G. G., dei Marcovaldi, M. Â. G. (2016) Adaptive threat management framework: integrating people and turtles. Environment, Development, and Sustainability. 18, 1541–1558.

Smith, I. (2004) An assessment of recent trends in Australian rainfall. Australian Meteorological Magazine. 53, 163–173.

Spainer, M. J. (2010) Beach erosion and nest site selection by leatherback sea turtles (Dermochelys coriacea) and implications for management practices At Playa Gandoca, Costa Rica. Revista de Biología Tropical. 58, 1237–1246.

Stockdon, H. F., Doran, K. S., Sallenger Jr., A. H. (2009) Extraction of lidar-based dune-crest elevations for use in examinin the vulnerability of beaches to inundation during hurricanes. Journal of Coastal Research. 53, 59–65.

Strobl, C., Malley, J., Tutz, G. (2009) An introduction to recursive partitioning: Rationale, application, and characteristics of classification and regression trees, bagging, and random forests. Psychological Methods. 14, 323–348.

Tapilatu, R. F., Wona, H., Batubara, P. P. (2017) Status of sea turtle populations and its conservation at Bird’s Head Seascape, Western Papua Indonesia. Biodiversitas, J. Biol. Divers.. 18, 129–136.

Therneau, T. M., Atkinson, E. J. (2018a) An introduction to recursive partitioning using the rpart routines. Compr. R Arch. Netw., 1–62.

Therneau, T., Atkinson, B. (2018b) rpart: Recursive partitioning and regression trees. R package version 4.1-12. https://CRAN.R-project.org/package=rpart.

Thomson, J. A., Hajnoczky, N., Hattingh, K. (2016) The sea turtle rookery at Gnaraloo Bay, Western Australia: Using nocturnal observations to validate diurnal track interpretations. Chelonian Conservation and Biology. 15, 187–196.

Tucker, T., Whiting, S., Mitchell, N., Berry, O., FitzSimmons, N., Pendoley, K. (2015) The Kimberley marine turtle project under the Western Australian Marine Sciences Institution (WAMSI). In: Proceedings of the Second Australian and Second Western Australian Marine Turtle Symposia, Perth 25-27 August 2014 (eds S. D. Whiting & A. Tucker) pp. 79–81 Science Division, Department of Parks and Wildlife, Perth.

Wade, S., Hickey, R. (2008) Mapping migratory wading bird feeding habitats using satellite imagery and field data, Eighty-Mile Beach, Western Australia. Journal of Coastal Research. 243, 759–770.

Walker, T. A., Parmenter, C. J. (1990) Absence of a pleagic phase in the life cycle of the flatback turtle, Natator depressa (Garman). Journal of Biogeography. 17, 275–278.

Wallace, B. P., DiMatteo, A. D., Bolten, A. B., Chaloupka, M. Y., Hutchinson, B. J., Abreu-Grobois, F. A., Mortimer, J. A., Seminoff, J. A., Amorocho, D., Bjorndal, K. A., Bourjea, J., Bowen, B. W., Dueñas, R., Casale, P., Choudhury, B. C., Costa, A., Dutton, P. H., Fallabrino, A., Finkbeiner, E. M., Girard, A., Girondot, M., Hamann, M., Hurley, B.J., López-Mendilaharsu, M., Marcovaldi, M.A., Musick, J.A., Nel, R. Pilcher, N.J., Troëng, S., Witherington, B., Mast, R.B. (2011) Global conservation priorities for marine turtles. PLoS One. 6, e24510.

Wallace, B. P., DiMatteo, A. D., Hurley, B. J., Finkbeiner, E. M., Bolten, A. B., Chaloupka, M. Y., Hutchinson, B. J., Alberto Abreu-Grobois, F., Amorocho, D., Bjorndal, K. A., Bourjea, J., Bowen, B. W., Dueñas, R. B., Casale, P., Choudhury, B. C., Costa, A., Dutton, P. H., Fallabrino, A., Girard, A., Girondot, M., Godfrey, M.H., Hamann, M., López-Mendilaharsu, M., Marcovaldi, M.A., Mortimer, J.A. Musick, J.A., Nel, R.,

48

Pilcher, N.J., Seminoff, J.A., Troëng, S., Witherington, B., Mast, R.B. (2010) Regional management units for marine turtles: A novel framework for prioritizing conservation and research across multiple scales. PLoS One. 5, 1–11.

Wallace, B. P., Tiwari, M., Girondot, M. (2013) Dermochelys coriacea. IUCN Red List of Threatened Species.

Weishampel, J. F., Bagley, D. a., Ehrhart, L. M., Rodenbeck, B. L. (2003) Spatiotemporal patterns of annual sea turtle nesting behaviors along an East Central Florida beach. Biological Conservation. 110, 295–303.

Whiting, A. U., Thomson, A., Chaloupka, M., Limpus, C. J. (2008) Seasonality, abundance and breeding biology of one of the largest populations of nesting flatback turtles, Natator depressus: Cape Domett, Western Australia. Australian Journal of Zoology. 56, 297–303.

Whitmore, C. P., Dutton, P. H. (1985) Infertility, embryonic mortality and nest-site selection in leatherback and green sea turtles in Suriname. Biological Conservation. 34, 251–272.

Witherington, B., Hirama, S., Mosier, A. (2011a) Sea turtle responses to barriers on their nesting beach. Journal of Experimental Marine Biology and Ecology. 401, 1–6.

Witherington, B., Hirama, S., Mosier, A. (2011b) Barriers to sea turtle nesting on Florida (United States) beaches: Linear extent and changes following storms. Journal of Coastal Research. 27, 450–458.

Wood, D. W., Bjorndal, K. A. (2000) Relation of temperature, moisture, salinity, and slope to nest site selection in loggerhead sea turtles. Copeia. 1, 119–119.

Woolard, J. W., Colby, J. D. (2002) Spatial characterization, resolution, and volumetric change of coastal dunes using airborne LIDAR: Cape Hatteras, North Carolina. Geomorphology. 48, 269–287.

Yalçın-Özdilek, Ş., Özdilek, H. G., Ozaner, F. S. (2007) Possible influence of beach sand characteristics on green turtle nesting activity on Samandağ Beach, Turkey. Journal of Coastal Research. 236, 1379–1390.

Yamamoto, K. H., Powell, R. L., Anderson, S., Sutton, P. C. (2012) Remote sensing of environment using LiDAR to quantify topographic and bathymetric details for sea turtle nesting beaches in Florida. Remote Sensing of Environment. 125, 125–133.

Zare, R., Vaghefi, M., Kamel, S. (2012) Nest location and clutch success of the hawksbill sea turtle (Eretmochelys imbricata) at Shidvar Island, Iran. Chelonian Conservation and Biology. 11, 229–234.

Zavaleta-Lizarraga, L., Morales-Mavil, J. E. (2013) Nest site selection by the green turtle (Chelonia mydas) in a beach of the north of Veracruz, Mexico. Revista Mexicana De Biodiversidad. 84, 927–937.

Zhang, K., Whitman, D., Leatherman, S., Robertson, W. (2005) Quantification of beach changes caused by Hurricane Floyd along Florida’s Atlantic coast using airborne laser surveys. Journal of Coastal Research. 211, 123–134.

APPENDIX I: LITERATURE REVIEW

49

7 APPENDIX I: LITERATURE REVIEW

INTRODUCTION

The conservation of sea turtles is an objective shared amongst scientists, indigenous

peoples, and the general public alike (Campbell and Smith 2006; Smith-Marder 2006; Morgan

2007; Wallace et al. 2011). Conservation programs around the world are underway to ensure the

continued persistence of sea turtles, though becoming increasingly difficult with limited

monetary resources. Obtaining information to make informed decisions is critical to ensure that

conservation outcomes are maximized (Wallace et al. 2011). One such conservation strategy is

the protection of coastal areas as several sea turtle life-history processes, such as breeding and

nesting, can be protected within a single area (Wallace et al. 2011; Nel et al. 2013). New

technologies such as airborne laser scanning are tools used in ecological and coastal research to

characterize forest structure and beach topography (Mason et al. 2000; Lefsky et al. 2002;

Zhang et al. 2005; Simonson et al. 2014), and has already been extended to understand the

importance of beach topography for the nest site selection by sea turtles (Long et al. 2011;

Yamamoto et al. 2012, 2015).

The use of airborne laser scanning has the greatest benefits in expansive landscapes, such

as Australian beaches, some of which are also nesting beaches to endemic flatback sea turtles.

This species of sea turtle nest in several important rookeries on the continent, one of which is

Eighty Mile Beach (Limpus 2007; Department of Parks and Wildlife 2014). This literature

review focuses on nest site selection by flatback sea turtles and the use of airborne laser

scanning in characterizing topography of nesting beaches in Eighty Mile Beach Marine Park.

This review will cover the anthropogenic threats facing the seven species of sea turtles first,

followed by a focus on flatback sea turtles, including their biology, distribution, and the specific

threats they face in Western Australia. A review of the literature covering what is known about

sea turtle nest site selection will also be discussed, with a particular focus on topography and a

brief discussion on other factors of nest site selection. Airborne laser scanning will then be

APPENDIX I: LITERATURE REVIEW

50

introduced, with a brief explanation on how the technology works, followed by the applications

of airborne laser scanning in ecology. The review then transitions into the brief evolution of

Eighty Mile Beach, followed by the ecological, cultural, and recreational values of the marine

park itself. The review is concluded with the aims of the current project, which are focused on

determining how beach topography is used by flatback sea turtles as a factor for nest site

selection.

SEA TURTLES AND ANTHROPOGENIC THREATS

The seven extant species of sea turtles are faced with the threat of extinction. Each of

these species of sea turtles has undergone rapid and large population declines (Table 1),

resulting in being listed as either vulnerable, endangered, critically endangered, or data deficient

under the International Union for the Conservation of Nature (IUCN) Red List of Threatened

Species (Marine Turtle Specialist Group 1996; Red List Standards & Petitions Subcommittee

1996; Seminoff 2004; Abreu-Grobois & Plotkin 2008; Mortimer & Donnelly 2008; Wallace et

al. 2013; Casale & Tucker 2015). Green sea turtles, for example, have seen 50% declines of

nesting populations in three generations (Seminoff 2004), while it is estimated that the Kemp’s

ridley sea turtles have lost 80% of their nesting population within three generations (Marine

Turtle Specialist Group 1996). Future population losses are also predicted to be substantial, with

hawksbill and green sea turtles predicted to lose an additional 80% of their nesting populations

within three generations (Meylan and Donnelly 1999; Seminoff 2004; Mortimer and Donnelly

2008).

This is not true for other sea turtle populations around the world. For example, the green

sea turtle nester abundance in Hawaii has increased from less than 100 individuals to more than

500 over thirty years, faster than what was predicted since their harvesting was made illegal in

the 1970s (Balazs and Chaloupka 2004). Intensive protection of nesting beaches in the

Caribbean has seen an increase in leatherback sea turtle nester abundance over 20 years, with a

predicted annual increase in population abundance of 13% (Dutton et al. 2005). Hawksbill sea

turtles, which are listed as Critically Endangered by the IUCN, have seen an increase in their

APPENDIX I: LITERATURE REVIEW

51

nesting abundance over 15 years in Barbados, West Indies, which was a testament to the local

and regional protection of this species (Beggs et al. 2007). Intensive captive-rear and release

programs conducted on the Kemp’s ridley sea turtle in the United States has even seen the

return of a reproductive female to nest (Shaver et al. 2016).

Australian sea turtle populations are protected under federal and several state laws, which

have allowed their populations to persist. Green sea turtles in Edgecumbe Bay, Queenland, have

seen an annual increase of 8.3%/year, which was testament to the involvement of the indigenous

community in sea turtle conservation (Hof et al. 2017). An earlier study in the southern Great

Barrier Reefs have seen green sea turtle population abundances increase by 11%/year

(Chaloupka and Limpus 2001), but was not necessarily true for the loggerhead sea turtles,

which saw population declines of possibly 3% (Chaloupka and Limpus 2001). The authors

attributed these declines to recruitment failure due to fox predation of nests and the mortality of

juveniles during their pelagic phase to incidental long line fishing (Chaloupka and Limpus

2001). This study highlights that though there are increases in population abundance of some

sea turtle populations, but this increase is dependent on the successful execution of conservation

programs and protection of foraging and breeding sites (Wallace et al. 2011; Wallace et al.

2013; Nel et al. 2013;). Without the successful execution of such conservation programs, sea

turtle populations may continue to decline.

The conservation of sea turtles is complicated by the geographically expansive nature of

their life cycle. For example, sea turtle nesting beaches can occur several thousands of

kilometres from their foraging grounds, and sea turtles must undertake long-distance migrations

between these sites approximately every two years (Limpus et al. 1992; Bowen et al. 1995;

Hughes et al. 1998; Miller et al. 1998; Luschi et al. 2003). This means that the protection of

both foraging grounds, nesting beaches, and migratory corridors used by sea turtles are

necessary to ensure the continued viability of the species (Wallace et al. 2011; Nel et al. 2013;

Pendoley et al. 2014b). Further, very little is known about where and how hatchling sea turtles

disperse once they enter offshore oceanic currents, or how long they remain within them (Carr

APPENDIX I: LITERATURE REVIEW

52

1954; Bolten 2003; McClellan et al. 2010; Mansfield et al. 2014; Scott et al. 2014). The

inability to identify important oceanic habitats for these pelagic “lost years” and potentially

providing insufficient protection for these habitats can jeopardize the viability of sea turtle

species (Hamann et al. 2010; Wallace et al. 2011).

Table 1 The seven species of sea turtles (taxonomic and common names), occurrence and nesting regions, IUCN status, and estimated percent-population decline over three generations.

Species Common Name

Nesting Regions IUCN Status

Estimated percent-

population decline

Source

Dermochelys coriacea

Leatherback turtle

Tropical Vulnerable 40.1% Wallace et al. (2013)

Chelonia mydas

Green turtle Tropical Endangered 48% to 67%

Seminoff (2004)

Natator depressus

Flatback turtle

Australian Tropical

Data deficient (requires updating)

Needs updating

Red List Standards & Petitions Subcommittee (1996)

Eretmochelys imbricata

Hawksbill turtle

Tropical Critically endangered

84% to 87%

Mortimer & Donnelly (2008)

Caretta caretta

Loggerhead turtle

Sub-tropical and Temperate

Vulnerable 47%

Casale & Tucker (2015)

Lepidochelys olivacea

Olive ridley turtle

India and west coast of Mexico

Vulnerable 63% to 83% Abreu-Grobois & Plotkin (2008)

Lepidochelys kempii

Kemp’s ridley turtle

Gulf of Mexico, United States

Critically Endangered

80% Marine Turtle Specialist Group (1996)

The seven species of sea turtles share very similar life histories and many of the

anthropogenic threats are associated to the different life-history stages (Fig 1). Anthropogenic

lighting from onshore coastal estates can deter gravid sea turtles from nesting on a beach

(Salmon, Reiners, et al. 1995; Mazor et al. 2013). Egg clutches laid on beaches are vulnerable to

being illegally or legally harvested by people (Marcovaldi and Chaloupka 2007; Valverde et al.

2012; Stringell et al. 2013; Humber et al. 2014), damage by tourists as well as beach users

(Katselidis et al. 2013), or feral animal predation (Brown and MacDonald 1995; Blamires et al.

APPENDIX I: LITERATURE REVIEW

53

2003; Mendonça et al. 2010). Tyre ruts and pedestrian tracks left behind by off-road vehicles

and beach visitors on beaches can prolong the time taken by newly-emerged hatchlings to reach

the water (Hosier et al. 1981; van de Merwe et al. 2012), increasing the time they are vulnerable

to predation. Anthropogenic lights on the landward side of the beach can cause sea hatchlings to

move towards the light inland rather than towards the ocean (Salmon and Witherington 1995;

Salmon, Tolbert, et al. 1995; Tuxbury and Salmon 2005; Berry et al. 2013), while offshore

lights (such as those on oilrigs) can cause sea turtles to swim towards them instead of swimming

out to open ocean (Thums et al. 2016). Hatchlings, neonates, juveniles, and sub-adults in their

pelagic phase are vulnerable to becoming victims of by-catch (Polovina et al. 2003; Wallace et

al. 2008; Levy et al. 2015) and may also ingest marine debris (Bjorndal et al. 1994; Bugoni et

al. 2001; Lazar and Gracan 2011; Schuyler et al. 2012). Sub-adults and adults that recruit into

neritic habitats are also at risk of poaching (Garcia-Martinez and Nicols 2000; Mancini and

Koch 2009), legal harvesting (Humber et al. 2014), or being injured and killed by boat strikes

(Davenport and Davenport 2006). Coastal development can also impact sea turtles, especially

during breeding. The building of seawalls can cause beaches to passively erode, reducing

nesting success while increasing the likelihood that clutches are washed out when these beaches

completely erode (Rizkalla and Savage 2011). Continued development of coastlines can reduce

the length of available beaches to nest on, which can result in sea turtles “squeezing” nests onto

available beaches (Mazaris et al. 2009) or remove these beaches entirely (Lai et al. 2015). These

impacts are only the direct impacts that humans have on sea turtles, as climate change poses a

different suite of impacts that also threaten sea turtles.

APPENDIX I: LITERATURE REVIEW

54

Fig 1 The life cycle of a sea turtle with various threats faced in different life history stages. Their life cycle begins with a clutch of eggs in the sand incubated between 45 – 65 days (a). Emergence of hatchlings (b) is followed by a dash into the surf and a swimming frenzy (c). They live their pelagic stage (d) until they are large enough to recruit into neritic environments (e). After developing sufficient energy stores, they migrate back to their natal site to nest (f).

The human-induced climate change of recent times also has far reaching effects on sea

turtles in all stages of their life cycle. For example, sea turtle egg clutches can only develop

within a range of temperatures, with warmer temperatures generating more females and cooler

temperatures creating more males (Mrosovsky and Provancha 1992; Mrosovsky et al. 1992;

Godfrey et al. 1999; Broderick et al. 2000; Tomillo et al. 2014). Rising temperatures as a result

of climate change can result in 100% mortality within egg clutches (Segura and Cajade 2010;

Howard et al. 2014), entirely female-biased clutches (Broderick et al. 2000), or hatchlings that

do not perform adequately while swimming to offshore currents (swimming frenzy) (Burgess et

al. 2006; Sim et al. 2014). Rising temperatures from global warming can also affect offshore

• Tyre rut and pedestrian tracks prolonging time on beach

• Predation by feral animals • Onshore light causing

misorientation and disorientation

• Loss of nesting beaches to coastal development or increased storms

• Increase temperatures • Egg poaching

• Ingestion of marine debris • By-catch • Poaching

APPENDIX I: LITERATURE REVIEW

55

carbonate producers and reef-forming communities, causing changes in sediment accretion of

mainland and island beaches used by nesting sea turtles (Fuentes et al. 2010). Stronger storms

and rising sea levels can also wash out and drown entire clutches of eggs (Milton et al. 1994;

Fuentes et al. 2011; Dewald and Pike 2014; Patino-Martinez et al. 2014) or erode current and

potential nesting beaches (Mazaris et al. 2009; Rizkalla and Savage 2011). Foraging habitats,

such as seagrass communities, can also be degraded or lost (Short and Neckles 1999; Koch et al.

2013). This can lead to an abundance of sea turtles with poor body condition and increased

mortality of reproductive adults (Poloczanska et al. 2010), which can result in reduced breeding

and lower reproductive output. These threats are not unique to any one species of sea turtle, and

even sea turtles that do no display an oceanic phase, such as the flatback sea turtle of Australia,

are vulnerable.

FLATBACK SEA TURTLES

Flatback sea turtles (Natator depressus) are endemic to tropical Australia (Fig 2) and are

markedly different from other species of sea turtles. They are flatter than other sea turtle species

(Bustard and Limpus 1969; Limpus 2007) and lay clutches with fewer eggs (Bustard and

Limpus 1969; Pendoley et al. 2014a), while the eggs and hatchlings themselves are considerably

larger (Bustard and Limpus 1969; Limpus et al. 1984; Hewavisenthi and Parmenter 2002;

Limpus 2007; Pendoley et al. 2014a). Flatback sea turtles also have no oceanic post-hatchling or

juvenile phase (Limpus et al. 1983; Walker and Parmenter 1990), a trait that is shared in the

other species of sea turtles. Indeed, the colouration and size of flatback sea turtle hatchlings and

neonates would allow them to quickly evade predators while living in their turbid neritic habitat

of Australian continental waters (Salmon et al. 2009, 2010).

APPENDIX I: LITERATURE REVIEW

56

Fig 2 Distribution of nesting flatback sea turtle populations throughout tropical Australia. Red circles indicate nesting population, with size of each circle representing the size of the nesting population (Limpus 2007, p.11).

Western Australia is home to Southeast Indian Ocean Management Unit of flatback sea

turtles, one of the four management units within Australia (Table 2) (Dutton et al. 2002;

Limpus 2007; Whittock et al. 2016). Initially, very little was known about the flatback sea

turtles of this management unit (Limpus 2007), but many publications in recent years have

addressed several of these knowledge gaps (Sperling et al. 2010; Pendoley et al. 2014a;

Pendoley et al. 2014b; Whittock et al. 2014; Pendoley and Kamrowski 2015; Whittock et al.

2016; Pendoley et al. 2016; Pendoley and Kamrowski 2016). For example, flatback sea turtle

nesting has been found to be evenly distributed throughout different subregions of the Pilbara

(Fig 3) (Pendoley et al. 2016), while the ranges of female flatback sea turtles between nesting

events (inter-nesting period) has been shown to be fairly consistent (Sperling et al. 2010;

Whittock et al. 2014). Migration corridors between foraging and feeding grounds were

identified along the northwest shelf for individuals from the Pilbara (Pendoley et al. 2014b),

with their foraging grounds found near Eighty Mile Beach, offshore from Quondong Point, and

in the Lynher and Holothuria Banks (Whittock et al. 2016). Addressing these knowledge gaps

A biological review of Australian marine turtle species. 5. Flatback turtle, Natator depressus (Garman)

11

Figure 3a. Distribution of Natator depressus nesting beaches. The data are incomplete for the western part of Arnhem Land and Western Australia.

Figure 3b. Post-nesting dispersal of Natator depressus from Queensland rookeries to their respective foraging areas. Symbols denote rookery of origin for the females: circles - eastern Australian rookeries of Mon Repos, Curtis Island, Peak Island and Wild Duck Island; squares – Crab Island in NE Gulf of Carpentaria; triangle – Bountiful Island, SE Gulf of Carpentaria. Figure 3. Distribution of nesting beaches and post nesting dispersal for Natator depressus.

APPENDIX I: LITERATURE REVIEW

57

was important, particularly for the flatback sea turtle population in the Pilbara, as this

population often interacts with the oil and gas industry of the region.

Table 2 Management units of flatback sea turtles throughout Australia. Management Unit Major rookeries Breeding

migration Peak

nesting Post-hatchling

migration Location of

foraging adults

East Australian Management Unit

Peak Island, Wild Duck Island

Up to 1,300 km

Late Nov to

Early Dec

Neritic waters within Great Barrier Reef lagoon

Captured in trawl fisheries over soft-bottom waters

Gulf of Carpentaria and Torres Strait Management Unit

Crab Island, Deliverance Island, Kerr Island

Out of Gulf to Arafura Sea and northern Great Barrier Reef

July

Gulf of Carpentaria & Southern Arafura Sea

Open water soft-bottom habitats of the Arafura Sea, Gulf of Papua, and Great Barrier Reef

Western Northern Territory Breeding Unit

Coburg Peninsula /adjacent islands, Cape Domett

Unknown July Recorded in coastal waters

Occur at low densities within coastal waters such as Darwin Harbour

Southeast Indian Ocean Management Unit

Barrow Island, Thevenard Island, Dampier Archipelago, Mundabullangana

Utilize a 1,200 km corridor in NW Australia

Jan Unknown

Eighty Mile Beach, Lacepede Islands, Lynher Banks

Data obtained from Pendoley et al. (2014a, 2014b, 2016) and Limpus (2007)

APPENDIX I: LITERATURE REVIEW

58

Fig 3 Distribution of nesting flatback sea turtle populations throughout the Pilbara region as surveyed by Pendoley et al. (2016). Black circles indicate presence of nesting activity, while crosses indicate absence of turtle nesting.

Many of the threats faced by the flatback sea turtles of the Pilbara come from their

interactions with the urban and oil and gas activities of the region. Hatchlings emerging from

their nests have shown to be misoriented in the presence of offshore lights (Thums et al. 2016).

The inter-nesting ranges of flatback turtles within the Pilbara were found to be in close

proximity of several major resource developments and within offshore lease title areas,

suggesting that flatback sea turtles were potentially exposed to multiple risks during their inter-

nesting period (Whittock et al. 2014). This is of concern as the inter-nesting period is used for

rest while female flatback sea turtles undergo ovulation, which may be sensitive to disturbance

(Sperling et al. 2010). Almost half (48%) of the migratory corridor used by flatback sea turtles

between their foraging and nesting grounds are in unprotected waters, potentially leaving them

abundance (tracks night–1) was assigned to one of five broadcategories: 0 (none), 1–10 (low), 11–100 (medium), 100–500(high) and >500 (very high) tracks night–1 for each species, as perLimpus (2009).

Survey location, technique, duration (n) and species-specificpresence/absence at all locations are given in the SupplementaryMaterial.

ResultsFlatback turtles

Regional flatback nesting activity was recorded at 58% of allsurveyed locations (n = 89) (Table 2). Most activity (76%) wasisland-based, distributed across both islands situated closer toshore (<25 km) within the Dampier, Onslow and Port Hedlandsubregion andon the outer islands of theBarrowGroup subregion>50 km from the mainland coastline (Figs 3–6). The Onslow andDampier subregions had the equal highest regional proportionof flatback nesting among surveyed locations (Table 2). Regionalflatback turtle nesting at locations where abundance was

quantified was 877.4! 29.5 tracks night–1 (range = 0.2–221.0,n = 77).

Green turtles

Regional green turtle nesting activity was recorded at 36% of allsurveyed locations (n= 55) (Table 2). Most locations (93%) wereislandswith activity primarily located on outer islands away fromthe mainland coastline (Figs 3–6). Activity at surveyed mainlandlocations was low (7%). Subregional nesting activity was mostwidespreadwithin the BarrowGroup subregion, recorded at 71%of all surveyed locations (Table 3). Regional overnight nesting atlocations where abundance was quantified was 1200.5! 62.0tracks night–1 (range = 0.1–313.0, n = 47).

Hawksbill turtlesRegionally, hawksbill turtle nesting was recorded at 29% of allsurvey locations (n= 45), concentrated within the Onslowsubregion (42% of all regional nesting) (Tables 2 and 3). Therewas no hawksbill nesting in the Port Hedland subregion or at any

0 2010 30 40 50 km

0 10 20 km

0 2010 30 40 50 km

0 2010 30 40 50 km

Legend:Evidence of nestingactivity (any species) :

Present

Absent

BarrowGroup

Onslow

Dampier Port Hedland

Fig. 2. Survey locations in each subregion, showing presence and absence of marine turtle nesting activity. Black dot, present;grey cross, absent.

220 Australian Journal of Zoology K. L. Pendoley et al.

APPENDIX I: LITERATURE REVIEW

59

vulnerable to the impacts by industrial activities (Pendoley et al. 2014b). Though adult flatback

sea turtles displayed foraging behaviour that make them robust against offshore developments,

their very nature of being endemic can reduce their ability to adapt should developments expand

(Bonn et al. 2002; Whittock et al. 2016).

Managing these threats must be prioritized to ensure that conservation outcomes are

maximized with limited financial resources. As mentioned before, protecting a species with a

wide geographic distribution can be difficult (Wallace et al. 2011). As suggested by Pendoley et

al. (2014b), protecting migratory corridors can ensure the safe migration of flatback sea turtles

between their foraging grounds and breeding sites (Pendoley et al. 2014b). The development of

additional protected coastal areas provide habitat that will assist in their protection (though other

factors can influence population decline), especially if nesting beaches are protected (Nel et al.

2013).

NEST SITE SELECTION

The protection of high-density nesting beaches can increase the reproductive output of

sea turtles by ensuring they have a safe habitat for nesting, clutch incubation, hatching, and

hatchling emergence (Mortimer et al. 2011; Nel et al. 2013). Therefore, to protect or mitigate

human impacts upon sea turtles, it is necessary to understand which beaches are used by sea

turtles as nesting habitats (Hamann et al. 2010; Wallace et al. 2011). This can be difficult as

how a sea turtle selects one nest site over another is not well understood, and has been identified

as a priority for research to determine which beaches require protection (Hamann et al. 2010).

Nest site selection between species and between different populations of the same species

appear to have some similarities, particularly with respect to beach topography (Mortimer 1982;

Wood and Bjorndal 2000; Katselidis et al. 2013; Roe et al. 2013; Dunkin et al. 2016). Elevation

of the nesting site above sea level appeared to be important, but the profile of the beach may

also be a determining factor, while interactions between topographic variables may also exist

(Horrocks and Scott 1991; Roe et al. 2013; Katselidis et al. 2014). For example, loggerhead sea

APPENDIX I: LITERATURE REVIEW

60

turtles in Sanibel and Captiva in Florida, United States, always above the most recent high tide

line (Hays et al. 1995). An experimental study in the Archie Carr National Wildlife Refuge in

Florida extended this finding and found the loggerhead sea turtles began nest construction when

a temporary barrier was erected in front of a loggerhead sea turtle (Witherington et al. 2011).

Loggerhead sea turtle populations on the Sunshine Coast, Queensland, however, were found to

preferentially nest on areas of beach where there was greater concavity at the dune toe (Kelly et

al. 2017). In comparison, hawksbill sea turtles in Barbados appeared to prefer sites 1 m above

the high tide line (Horrocks and Scott 1991). The distance travelled inland by these hawksbill

sea turtles was dependent on the slope of the beach; steeper slopes meant shorter distances

travelled across the beach, while beaches with lower slopes meant the sea turtle needed to travel

greater distances across the beach in order to reach an elevation of 1 m (Horrocks and Scott

1991). In contrast, hawksbill sea turtles in El Salvador preferred to nest within dense vegetation

(Liles et al. 2015). A separate study in Greece modelled the importance of beach orientation,

width, length, elevation, and slope for where loggerhead sea turtles would place their nests

(Katselidis et al. 2013). Though an elevation of 1 m above the mean sea level was the most

reliable indicator for where a sea turtle would nest, the authors also found that the loggerhead

sea turtles had a preference for steeper slopes (Katselidis et al. 2013).

Studies that modelled nest site selection over larger areas detected single variables such

as elevation as important factors for nest site selection (Yamamoto et al. 2012; Dunkin et al.

2016). A study in Florida using airborne laser scanning on 87 km of beach showed that

minimum elevation on the beach dune was the most important factor in nest site selection

(Yamamoto et al. 2012). A separate study conducted in Florida across 200 km of coast using a

multi-criteria decision support model showed that the model was sensitive to the removal of

elevation, but not to other morphological characteristics, indicating that elevation was an

important factor for nest site selection by sea turtles (Dunkin et al. 2016).

Topography may be an important factor for where a sea turtle may decide to nest on a

beach, but other factors appear to be important to varying spatial extents. The region which sea

APPENDIX I: LITERATURE REVIEW

61

turtles nest in is driven by natal homing, where males and females use geomagnetic cues to

return to regions they hatched from (Allard et al. 1994; Lohmann et al. 2008; Lohmann et al.

2013; Brothers and Lohmann 2015). Beach selection appears to be driven by where the paths of

storms and cyclones cross the coast. Beach selection has been suggested to be driven by the

location of storms and tropical cyclone pathways crossing the coast; important nesting sites

between Brisbane and Prosepine in Queensland appeared to be just south of areas that had

higher storm frequencies (Fuentes et al. 2011). This may be attributed to the selection of

beaches that were not undergoing erosion (Koch et al. 2007) or sudden changes in beach volume

(Long et al. 2011; Yamamoto et al. 2015). Offshore approaches that are open and have fewer

rocks are also preferred as sea turtles may sustain fewer injuries when approaching the beach to

nest and returning to the ocean after depositing its eggs (Mortimer 1982). This preference is also

apparent with flatback sea turtles in eastern Australia, where many rookeries are fronted by

mudflats rather than coral reef flats (Limpus 2007). Green turtles on Ascension Island

(Mortimer 1982) and leatherback turtles nesting in the Pacific coast of Costa Rica (Roe et al.

2013) appear to prefer deeper offshore approaches, which may be due to deeper offshore

approaches correlating with steeper beach slopes (Roe et al. 2013). Within the nesting site itself,

the moisture of the sand is important during the construction the egg chamber as nests in drier

sand collapse more often than moister sand (Mortimer 1990; Chen et al. 2007). Plant roots also

stabilize sand, which may explain why green sea turtles in Taiwan (Chen et al. 2007) or

hawksbill sea turtles in Nicaragua (Liles et al. 2015) show a preference for nesting at the edge of

vegetation. However, other authors noted that highly dense vegetation will have higher root

density and impede nest digging (Hays et al. 1995).

How a sea turtle may decide to nest in one site on the beach over another is further

complicated by the dynamic nature of beach systems. Beaches can vary between dissipative,

reflective, or various intermediate beach types (Woodroffe 2002). Dissipative beaches

experience high wave energy that is initially broken by an offshore bar and is continually lost

along a wide surf zone (Bird 2008). Because wave energy is dissipated across the wide surf

zone, these beaches are usually flat with a wide intertidal zone (Woodroffe 2002). Reflective

APPENDIX I: LITERATURE REVIEW

62

beaches are on the opposite end of the spectrum, where waves surge onto the beach and wave

energy is reflected off the beach face itself (Bird 2008). Such beaches are steep and occur in

protected regions (e.g. embayments) and often have a deeper offshore bathymetry than

dissipative beaches. Wave periods also tend to be longer on reflective with lower wave energy

(Woodroffe 2002). Intermediate beach types occur in between these two extremes, with an

unstable beach shifting through intermediate types and acquiring dissipative and reflective

beach characteristics at any single time. For example, an intermediate beach can have an upper

portion that is reflective while the lower portion is dissipative (Woodroffe 2002). Such beaches

are also characterised by offshore bars (Woodroffe 2002), which can build or erode over time,

shifting to become more dissipative or reflective. The response of the beach, and the beach front

behind the beach, would be to erode, accrete, or stabilize (Woodroffe 2002).

The response of gravid sea turtles to a highly dynamic environment would be to move

where erosion occurs the least over generational time (Spanier 2010; Fuentes et al. 2011).

Green, flatback, hawksbill, and loggerhead sea turtles from eastern Queensland nest in areas

with low occurrence of tropical cyclones to nest on beaches that have a lower potential to erode

(Fuentes et al. 2011). Indeed, it has been demonstrated that sea turtles nesting several times

during a season in different locations is a legacy of natural selection as a response to beaches

that are more dynamic than others (Hart et al. 2013). If one area of beach is washed out and

entire clutches are lost, a single sea turtle would have egg clutches in other areas of beach that

would not have eroded (Hart et al. 2013). Loggerhead sea turtles in the Gulf of Mexico, for

example, have been shown to travel great distances during their inter-nesting periods to nest in

sites separated by hundreds of kilometres (Hart et al. 2013). This has implications for

conservation, as the availability of beaches does not necessarily indicate the availability of

preferred beaches to nest on (Fuentes et al. 2011).

Of the various conditions for nest site selection, beach topography appears to be

important in determining where a sea turtle will initiate digging. It is then important to predict

where a sea turtle may initiate pit construction in order to protect and mitigate impacts from

APPENDIX I: LITERATURE REVIEW

63

human development on current and potential nesting beaches (Salmon, Reiners, et al. 1995;

Mazaris et al. 2009; Rizkalla and Savage 2011; Mazor et al. 2013;). However, sea turtle nesting

can occur over large ranges, whether a single sea turtle travels expansive distances to nest

several times in a single season (Hart et al. 2013) or if a breeding population nests along a vast

stretch of beach (Dunkin et al. 2016). Attempting to collect detailed information on beach

topography over large scales can be nearly impossible with traditional, ground-based methods.

Airborne laser scanning can resolve this, as described below.

LIGHT DETECTION AND RANGING (LiDAR)

Airborne laser scanning is an excellent source of consistent, detailed topographic data

over large extents. Light detection and ranging (LiDAR) technology uses light pulses to

measure distance (Fig 4). When attached to a plane with a global positioning system (GPS) and

side-to-side scanner, it is able to obtain elevation data of the ground as well as the heights of

objects on the surface of the earth (Lefsky et al. 2002; Woolard and Colby 2002; Beraldin et al.

2010). The combined result of geo-referenced laser points over an expanse of land generates a

point-cloud, where each individual point on the ground represents where a laser has struck the

ground (Fig 5). Because the accuracy of point-clouds can be within centimeters (Beraldin et al.

2010), they can be converted into detailed digital surface models (DSMs) or digital elevation

models (DEMs) (Xiaoye 2008; Lindenbergh and Pietrzyk 2015) that capture fine-scale

topographic variation (Andrew and Ustin 2009; Long et al. 2011; Montaghi et al. 2013), an

attribute that has great appeal to coastal managers and researchers (Mason et al. 2000).

APPENDIX I: LITERATURE REVIEW

64

Fig 4 Schematic of airborne laser scanning over a landscape. Landscape (right) being scanned by a LiDAR scanner attached to the plane, with the point-cloud data forming behind the plane (left) (Rogalski and Chrzanowski 2014).

Fig 5 An example of a point-cloud obtained from the flight mission during this study. Each point in the image represents where the laser struck the Earth’s surface. Warmer colours are higher elevations (red is highest), while cooler colours are lower elevations (green in this image is the lowest).

APPENDIX I: LITERATURE REVIEW

65

Airborne laser scanning has been useful in several applications to coastal management

and research. Out of various techniques to quantify beach topography, airborne LiDAR met the

needs of the various groups of people (e.g. coastal defence, economic exploitation, etc.) based

on cost efficiency and vertical accuracy (Mason et al. 2000). This made LiDAR appealing to

coastal managers and researchers who wished to determine the rate or extent of coastal erosion

over time or after cyclonic events. For example, airborne LiDAR can detect small-scale changes

to beach volume after hurricane events (Zhang et al. 2005) or changes to beach volume over

various time scales (e.g. weeks or years) (Shrestha et al. 2005). Airborne LiDAR-derived DEMs

can also be coupled with multispectral imaging and aerial photography to successfully detect

erosion on coasts with greater accuracy, accurately classify substrate types (Kerfoot et al. 2012),

and correctly characterize scours along the beach landscape (Sherman et al. 2013).

The ability of LiDAR to capture fine-scale variation in the landscape has seen its

applications extended to investigating habitat suitability and species diversity. In terrestrial

habitats, forest structure and its relation to species assemblages has often been investigated as

LiDAR is able to detect attributes of vertical forest structure (Vierling et al. 2008; Bergen et al.

2009; Simonson et al. 2014). For example, different forest attributes such as the presence of

dead trees and openness of canopies can be detected by LiDAR (Martinuzzi et al. 2009). This

data can be used to generate or improve habitat suitability models for birds (Martinuzzi et al.

2009; Tattoni et al. 2012). Airborne laser scanning-derived topographic data of coastal habitats

can also be used to model habitat suitability, such as the case for a species of flowering plant on

the Atlantic barrier islands (Sellars and Jolls 2007) or to determine important measures of beach

topography used by sea turtles to select nesting sites (Yamamoto et al. 2012; Dunkin et al.

2016). The ability to capture fine-scale changes in coastal habitats has also been used to

determine the response of sea turtles to changes in beach volume over time (Yamamoto et al.

2015) or after storm events (Long et al. 2011). Airborne laser scanning-derived benthic rugosity

can also be used to assess fish species assemblages in coral reefs (Wedding et al. 2008) as coral

reef rugosity is positively correlated to fish species assemblages (McCormick 1994; Knudby

and LeDrew 2007).

APPENDIX I: LITERATURE REVIEW

66

It is worth mentioning that although airborne LiDAR may be a panacea in determining

the importance of different topographic variables for nest site selection, the scale at which the

data is analysed needs to be considered (Woolard and Colby 2002). For example, a study in

Florida investigated how nest site selection of loggerhead sea turtles changed before and after a

hurricane event, but averaged the slopes of 25 m sections of beach (Long et al. 2011). Another

study further south of the previous study used 5 m resolution to investigate nest site selection of

loggerhead sea turtles (Yamamoto et al. 2012). Though the studies investigated different

questions and were in different locations, the information obtained based on the scales used

would be significantly different, even if they were investigating the same topic (Woolard and

Colby 2002). Focused on nest site selection, it may be interesting to determine if important

variables or interactions between variables for nest site selection change as the scale used during

analysis changed as well.

Nevertheless, the benefits of using LiDAR in ecosystem studies are two-fold. First, data

can be obtained over large extents, such as 200 km of coastline in the study of nest site selection

of loggerhead sea turtles in Florida (Dunkin et al. 2016). Second, LiDAR can detect fine-scale

variations within the landscape, which can be used to create DEMs that reflect the real-life

topography of the study site (Andrew and Ustin 2009). Such benefits have been highlighted

when used to understand how topography is important for sea turtle nest site selection (Long et

al. 2011; Yamamoto et al. 2015, 2012; Dunkin et al. 2016). The application of LiDAR to

Australian landscapes is ideal, where many long beaches on this continent are remote as well as

inaccessible and cannot be characterized effectively with ground-based methods.

EIGHTY MILE BEACH MARINE PARK

Eighty Mile Beach is a 220-km continuous beach located in the northwest region of

Western Australia at the Canning Basin. The Canning Basin is a sedimentary basin of marine

sediment that began forming 500 million years before present (BP) (Short 2006a). Eighty Mile

APPENDIX I: LITERATURE REVIEW

67

Beach only developed when the glaciers began to form at 7,500 years BP (Semeniuk 2008).

Offshore sand barriers developed 7,500 – 5,000 years BP along areas of coast that were open to

high energy waves (Semeniuk 2008). The coast continued to expand seawards as the glaciers

formed and sea levels dropped (Semeniuk 2008). Being open to high energy waves, much of the

sedimentation of Eighty Mile Beach today is primarily tidally-reworked carbonate and

Mesozoic-Tertiary bedrock (Short 2006b; Semeniuk 2008). The continued formation of offshore

sandbars has made Eighty Mile Beach an ultra-dissipative beach with a low gradient and a wide

intertidal zone (Short 2006a). This has provided a suite of habitats for several animal species

(Wilson 2013).

The intertidal and high tide beach areas of Eighty Mile Beach are recognized as important

habitats for benthic invertebrates and vertebrate animals alike (Short 2006a; Wilson 2013). The

benthic invertebrate species richness found within the Eighty Mile Beach intertidal zones have a

much higher density per square metre than intertidal zones found in other parts of the globe

(Honkoop et al. 2006; Compton et al. 2008). These high-diversity benthic invertebrate

communities sustain a population of 3.5 million migratory shorebirds that travel along the East-

Asian Australasian Flyway (Rogers et al. 2011; Minton et al. 2013). Flatback sea turtle nesting

sites have also been identified along the coast of Eighty Mile Beach (Department of Parks and

Wildlife 2014). This incredible array of biodiversity, amongst other key values, was significant

enough to warrant Eighty Mile Beach to be gazetted as a marine park (Department of Parks and

Wildlife 2014).

Eighty Mile Beach Marine Park was gazetted as the 13th marine park in Western Australia

in 2013. The key ecological values mentioned above can be found in several areas along the

beach, such as the three sanctuary zones: Anna Plains, Kurtamparanya, and the Pananykarra

Sanctuary Zones (Fig 6). Other zones are of cultural significance, such as the four special

purpose zones designated for their importance to the people of the region. The Paruwuturr,

Waru, and Pilyarlkarra Special Purpose Zones are found in Nyangumarta country, while

Jangyjartiny is jointly held by traditional owners of the Nyangumarta and Karajarri people

APPENDIX I: LITERATURE REVIEW

68

(Department of Parks and Wildlife 2014). The Wallal Recreation Zone has been designated for

recreational purposes and is adjacent to the Eighty Mile Beach Caravan Park. This site is a

popular tourist area that experiences high visitation during the dry winter months and provides

easy access onto the beach for vehicles and pedestrians (Short 2006a; Department of Parks and

Wildlife 2014; Beckley et al. 2015). The Wallal Recreation Zone is also an important nesting

site for endemic flatback sea turtles (Department of Parks and Wildlife 2014). Conflict between

flatback sea turtle nesting and recreation is alleviated by the nature of how the area is used;

flatback sea turtles nest from November to December during the hot summer months, while

visitors tend to visit the beach during the cooler winter months (Department of Parks and

Wildlife 2014; Beckley et al. 2015).

Eighty Mile Beach Marine Park is an important nesting site for flatback sea turtles, but

very little information about the nesting population in this area is available. Initial satellite

tagging of sea turtles has been conducted by the Department of Parks and Wildlife (Tucker et al.

2015), but no studies on nest site selection by the flatback sea turtles has been undertaken for

this population so far. Because of the extensive expanse of beach (220 km), it provides the ideal

case study to obtain topographic data over a broad extent with airborne LiDAR scanning, which

can be processed into DEMs. These DEMs can be analysed to determine the importance of

beach topography for flatback sea turtles in selecting nesting sites, which was obtained from

aerial flight surveys by Pendoley Environmental Pty. Ltd.

AIM AND RESEARCH QUESTIONS

The overall aim of this project is to determine the importance of topographical variables

for nest site selection by flatback sea turtles at Eighty Mile Beach Marine Park. The specific

research questions derived from this aim are:

1. To characterize the topography of a remote and expansive beach

2. To determine the influences of different features of beach topography on nest site

selection.

APPENDIX I: LITERATURE REVIEW

69 APPENDIX I: LITERATURE REVIEW

Fig 6 Location of Eighty-Mile Beach Marine Park (blue) in the Kimberley Region of north-west Australia indicating the various sanctuary and special purpose zones. Figure adapted from Eighty Mile Beach Marine Park Management Plan (Department of Parks and Wildlife 2014, p.18). PSZ = Pardoo Sanctuary Zone, CKSZ = Cape Keraudren Sanctuary Zone, PwSPZ = Paruwuturr Special Purpose Zone, WRZ = Wallal Recreational Zone, WSPZ = Waru Special Purpose Zone, PSPZ = Pilyarlkarra Special Purpose Zone, ASPZ = Anna Plains Sanctuary Zone, JSPZ = Jangyjartiny Special Purpose Zone.

Cape Missiessy

Cape Keraudren

PwSPZ

APPENDIX I: LITERATURE REVIEW

70 APPENDIX I: LITERATURE REVIEW

REFERENCES Abreu-Grobois, A., Plotkin, P. (2008) Lepidochelys olivacea. IUCN Red List of Threatened

Species, e.T11534A3292503.

Allard, M. W., Miyamoto, M. M., Bjorndal, K. A., Bolten, A. B., Bowen, B. W. (1994) Support for natal homing in green turtles from mitochondrial DNA sequences. Copeia. 1, 34–41.

Andrew, M. E., Ustin, S. L. (2009) Habitat suitability modelling of an invasive plant with advanced remote sensing data. Diversity and Distributions. 15, 627–640.

alazs, G. H., Chaloupka, M. (2004) Thirty-year recovery trend in the once depleted Hawaiian green sea turtle stock. Biological Conservation. 117, 491–498.

Beckley, L. E., Smallwood, C. B., Fisher, E. A. (2015) Chapter I: A spatially explicit benchmark of human use along Eighty Mile Beach prior to implementation of the Marine Park. In: Final Report of Project 2.1.1 of the Kimberley Marine Research Program Node of the Western Australian Marine Science Institution (ed L. E. Beckley) WAMSI, Perth, Western Australia.

Beggs, J., Horrocks, J., Krueger, B. (2007) Increase in hawksbill sea turtle Eretmochelys imbricata nesting in Barbados, West Indies. Endangered Species Research. 3, 159–168. [

Beraldin, J.-A., Blais, F., Lohr, U. (2010) Laser scanning technology. In: Airborne and Terrestrial Laser Scanning (eds G. Vosselman & H.-G. Maas) pp. 1–42Whittles Publishing, Dunbeath, Scotland.

Bergen, K. M., Goetz, S. J., Dubayah, R. O., Henebry, G. M., Hunsaker, C. T., Imhoff, M. L., Nelson, R. F., Parker, G. G., Radeloff, V. C. (2009) Remote sensing of vegetation 3-D structure for biodiversity and habitat: Review and implications for lidar and radar spaceborne missions. Journal of Geophysical Research: Biogeosciences. 114, 1–13.

Berry, M., Booth, D. T., Limpus, C. J. (2013) Artificial lighting and disrupted sea-finding behaviour in hatchling loggerhead turtles (Caretta caretta) on the Woongarra coast, south-east Queensland, Australia. Australian Journal of Zoology. 61, 137–145.

Bird, E. (2008) Beaches. In: Coastal Geomorphology: An Introduction pp. 133–200. John Wiley & Sons, Ltd., West Sussex, England.

Bjorndal, K. A., Bolten, A. B., Lagueux, C. J. (1994) Ingestion of marine debris by juvenile sea turtles in coastal Florida habitats. Marine Pollution Bulletin. 28, 154–158.

Blamires, S. J., Guinea, M. L., Prince, R. I. T. (2003) Influence of nest site selection on predation of flatback sea turtles (Natator depressus) eggs by varanid lizards in Northern Australia. Chelonian Conservation and Biology. 4, 1–7.

Bolten, A. B. (2003) Active swimmers - passive drifters: The oceanic juvenile stage of loggerheads in the Atlantic System. In: Loggerhead Sea Turtles (ed B. E. Witherington) pp. 63–78Smithsonian Institution Press, Washington, DC.

Bonn, A., Rodrigues, A. S. L., Gaston, K. J. (2002) Threatened and endemic species: Are they good indicators of patterns of biodiversity on a national scale? Ecological Letters. 5, 733–741.

Bowen, B. W., Abreu-Grobois, F. A., Balazs, G. H., Kamezaki, N., Limpus, C. J., Ferl, R. J. (1995) Trans-Pacific migrations of the loggerhead turtle (Caretta caretta) demonstrated with mitochondrial DNA markers. Proceedings of the National Academy of Sciences of the United States of America. 92, 3731–3734.

APPENDIX I: LITERATURE REVIEW

71 APPENDIX I: LITERATURE REVIEW

Broderick, A. C., Godley, B. J., Reece, S., Downie, J. R. (2000) Incubation periods and sex ratios of green turtles: Highly female biased hatchling production in the eastern Mediterranean. Marine Ecology Progress Series. 202, 273–281.

Brothers, J. R., Lohmann, K. J. (2015) Evidence for geomagnetic imprinting and magnetic navigation in the natal homing of sea turtles. Current Biology. 25, 392–396.

Brown, L., MacDonald, D. W. (1995) Predation on green turtle Chelonia mydas nests by wild canids at Akyatan Beach, Turkey. Biological Conservation. 71, 55–60.

Bugoni, L., Krause, L., Petry, M. V (2001) Marine debris and human impacts on sea turtles in southern Brazil. Marine Pollution Bulletin. 42, 1330–1334.

Burgess, E. A., Booth, D. T., Lanyon, J. M. (2006) Swimming performance of hatchling green turtles is affected by incubation temperature. Coral Reefs. 25, 341–349.

Bustard, R., Limpus, C. (1969) Observations of the flatback turtle Chelonia depressa Garman. Herpetologica. 25, 29–34.

Campbell, L. M., Smith, C. (2006) What makes them pay? Values of volunteer tourists working for sea turtle conservation. Environmental Management. 38, 84–98.

Carr, A. (1954) Rips, FADS, and little loggerheads. Bioscience. 36, 91–100.

Casale, P., Tucker, A. D. (2015) Caretta caretta. IUCN Red List of Threatened Species, e.T3897A83157651.

Chaloupka, M., Limpus, C. (2001) Trends in the abundance of sea turtles resident in southern Great Barrier Reef waters. Biological Conservation. 102, 235–249.

Chen, H.-C., Cheng, I.-J., Hong, E. (2007) The influence of the beach environment on the digging success and nest site distribution of the green turtle, Chelonia mydas, on Wan-an Island, Penghu Archipelago, Taiwan. Journal of Coastal Research. 5, 1277–1286.

Compton, T. J., Troost, T. A., Van Der Meer, J., Kraan, C., Honkoop, P. J. C., Rogers, D. I., Pearson, G. B., De Goeij, P., Bocher, P., Lavaleye, M. S. S., Leyrer, J., Yates, M. G., Dekinga, A., Piersma, T. (2008) Distributional overlap rather than habitat differentiation characterizes co-occurrence of bivalves in intertidal soft sediment systems. Marine Ecology Progress Series. doi: 10.3354/meps07732.

Davenport, J., Davenport, J. L. (2006) The impact of tourism and personal leisure transport on coastal environments: A review. Estuarine, Coastal and Shelf Science. 67, 280–292.

Department of Parks and Wildlife (2014) Eighty Mile Beach Marine Park Management Plan 80 2014–2024. Department of Parks and Wildlife, Perth, Australia.

Dewald, J. R., Pike, D. A. (2014) Geographical variation in hurricane impacts among sea turtle populations. Journal of Biogeography. 41, 307–316.

Dunkin, L., Reif, M., Altman, S., Swannack, T. (2016) A spatially explicit, multi-criteria decision support model for loggerhead sea turtle nesting habitat suitability: A remote sensing-based approach. Remote Sensing. 8, 1–22.

Dutton, D. L., Dutton, P. H., Chaloupka, M., Boulon, R. H. (2005) Increase of a Caribbean leatherback turtle Dermochelys coriacea nesting population linked to long-term nest protection. Biological Conservation. 126, 186–194.

Dutton, P., Broderick, A. C., FitzSimmons, N. N. (2002) Defining management units: Molecular genetics. In: Proceedings of the Western Pacific Sea Turtle Cooperative

APPENDIX I: LITERATURE REVIEW

72 APPENDIX I: LITERATURE REVIEW

Research & Management Workshop (ed I. Kinan) pp. 93–101Western Pacific Regional Fishery Management Council, Honolulu, Hawaii, USA.

Fuentes, M. M. P. B., Bateman, B. L., Hamann, M. (2011) Relationship between tropical cyclones and the distribution of sea turtle nesting grounds. Journal of Biogeography. 38, 1886–1896.

Fuentes, M. M. P. B., Dawson, J. L., Smithers, S. G., Hamann, M., Limpus, C. J. (2010) Sedimentological characteristics of key sea turtle rookeries: Potential implications under projected climate change. Marine and Freshwater Research. 61, 464–473.

Garcia-Martinez, S., Nicols, W. J. (2000) Sea turtles of Bahia Magdalena, Baja California Sur, Mexico: Demand and supply of an endnagered species. In: Proceedings of the Tenth Biennial Conference of the International Institute of Fisheries Economics and Trade (ed R. S. Johnston) pp. 1–4. International Institute of Fisheries Economics and Trade (IIFET), Corvallis, Oregon, USA.

Godfrey, M. H., D’Amato, A. F., Marcovaldi, M. Â., Mrosovsky, N. (1999) Pivotal temperature and predicted sex ratios for hatchling hawksbill turtles from Brazil. Canadian Journal of Zoology. 77, 1465–1473.

Hamann, M., Godfrey, M. H., Seminoff, J. A., Arthur, K., Barata, P. C. R., Bjorndal, K. A., Bolten, A. B., Broderick, A. C., Campbell, L. M., Carreras, C., Casale, P., Chaloupka, M., Chan, S. K. F., Coyne, M. S., Crowder, L. B., Diez, C. E., Dutton, P. H., Epperly, S. P., Fitz Simmons, N. N., Formia, A., et al. (2010) Global research priorities for sea turtles: Informing management and conservation in the 21st century. Endangered Species Research. 11, 245–269.

Hart, K. M., Lamont, M. M., Sartain, A. R., Fujisaki, I., Stephens, B. S. (2013) Movements and habitat-use of loggerhead sea turtles in the northern Gulf of Mexico during the reproductive period. PLoS One. 8, e66921.

Hays, G. C., Mackay, A., Adams, C. R., Mortimer, J. A., Speakman, J. R., Boerema, M. (1995) Nest site selection by sea turtles. Journal of the Marine Biological Association of the United Kingdom. 75, 667–674.

Hewavisenthi, S., Parmenter, J. C. (2002) Egg components and utilization of yolk lipids during development of the flatback turtle Natator depressus. Journal of Herpetology. 36, 43–50.

Hof, C., Smallwood, E., Meager, J., Bell, I. (2017) First citizen-science population abundance and growth rate estimates for green sea turtles Chelonia mydas foraging in the northern Great Barrier Reef, Australia. Marine Ecology Progress Series. 574, 181–191.

Honkoop, P. J. C., Pearson, G. B., Lavaleye, M. S. S., Piersma, T. (2006) Spatial variation of the intertidal sediments and macrozoo-benthic assemblages along Eighty-mile Beach, North-western Australia. J. Sea Res.. 55, 278–291.

Horrocks, J. A., Scott, N. M. (1991) Nest site location and nest success in the hawksbill turtle Eretmochelys imbricata in Barbados West-Indies. Marine Ecology Progress Series. 69, 1–8.

Hosier, P. E., Kochhar, M., Thayer, V. (1981) Off-road vehicle and pedestrian track effects on the sea-approach of hatchling loggerhead turtles. Environmental Conservation. 8, 158–161.

Howard, R., Bell, I., Pike, D. A. (2014) Thermal tolerances of sea turtle embryos: Current understanding and future directions. Endangered Species Research. 26, 75–86.

APPENDIX I: LITERATURE REVIEW

73 APPENDIX I: LITERATURE REVIEW

Hughes, G. R., Luschi, P., Mencacci, R., Papi, F. (1998) The 7000-km oceanic journey of a leatherback turtle tracked by satellite. Journal of Experimental Marine Biology and Ecology. 229, 209–217.

Humber, F., Godley, B. J., Broderick, A. C. (2014) So excellent a fishe: A global overview of legal marine turtle fisheries. Diversity and Distributions. 20, 579–590.

Katselidis, K. A., Schofield, G., Stamou, G., Dimopoulos, P., Pantis, J. D. (2013) Evidence-based management to regulate the impact of tourism at a key marine turtle rookery on Zakynthos Island, Greece. Oryx. 47, 1–11.

Katselidis, K. A., Schofield, G., Stamou, G., Dimopoulos, P., Pantis, J. D. (2014) Employing sea-level rise scenarios to strategically select sea turtle nesting habitat important for long-term management at a temperate breeding area. Journal of Experimental Marine Biology and Ecology. 450, 47–54.

Kelly, I., Leon, J. X., Gilby, B. L., Olds, A. D., Schlacher, T. A. (2017) Marine turtles are not fussy nesters: A novel test of small-scale nest site selection using structure from motion beach terrain information. PeerJ. 5, e2770.

Kerfoot, W. C., Yousef, F., Green, S. A., Regis, R., Shuchman, R., Brooks, C. N., Sayers, M., Sabol, B., Graves, M. (2012) Light detection and ranging (LiDAR) and multispectral studies of disturbed Lake Superior coastal environments. Limnology and Oceanography. 57, 749–771.

Knudby, A., LeDrew, E. (2007) Measuring structural complexity on coral reefs. In: Proceedings of the American Acadamy of Underwater Sciences 26th Symposium (eds N. W. Pollock & J. M. Godfrey) pp. 181–188. AAUS, Dauphin Island, AL.

Koch, A. U., Guinea, M. L., Whiting, S. D. (2007) Effects of sand erosion and current harvest practices on incubation of the flatback sea turtle (Natator depressus). Australian Journal of Zoology. 55, 97–105.

Koch, M., Bowes, G., Ross, C., Zhang, X. H. (2013) Climate change and ocean acidification effects on seagrasses and marine macroalgae. Global Change Biology. 19, 103–132.

Lai, S., Loke, L. H. L., Hilton, M. J., Bouma, T. J., Todd, P. A. (2015) The effects of urbanisation on coastal habitats and the potential for ecological engineering: A Singapore case study. Ocean and Coastal Management. 103, 78–85.

Lazar, B., Gracan, R. (2011) Ingestion of marine debris by loggerhead sea turtles, Caretta caretta, in the Adriatic Sea. Marine Pollution Bulletin. 62, 43–47.

Lefsky, M. A., Cohen, W. B., Parker, G. G., Harding, D. J. (2002) Lidar remote sensing for ecosystem studies. Bioscience. 52, 19–30.

Levy, Y., Frid, O., Weinberger, A., Sade, R., Adam, Y., Kandanyan, U., Berkun, V., Perry, N., Edelist, D., Goren, M., Bat-Sheva Rothman, S., Stern, N., Tchernov, D., Gil, R. (2015) A small fishery with a high impact on sea turtle populations in the eastern Mediterranean. Zoology in the Middle East. 61, 300–317.

Liles, M. J., Peterson, M. J., Seminoff, J. A., Altamirano, E., Henríquez, A. V., Gaos, A. R., Gadea, V., Urteaga, J., Torres, P., Wallace, B. P., Peterson, T. R. (2015) One size does not fit all: Importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the eastern Pacific Ocean. Biological Conservation. 184, 405–413.

Limpus, C. J. (2007) A biological review of Australian marine turtles. 5. Flatback turtle Natator

APPENDIX I: LITERATURE REVIEW

74 APPENDIX I: LITERATURE REVIEW

depressus (Garman). (ed L. Fien). Queensland Environmental Protection Agency, Brisbane.

Limpus, C. J., Fleay, A., Baker, V. (1984) The flatback turtle, Chelonia depressa, in Queensland: Reproductive periodicity, philopatry and recruitment. Australian Wildlife Research. 11, 579–587.

Limpus, C. J., Miller, J. D., Parmenter, C. J., Reimer, D., McLachlan, N., Webb, R. (1992) Migration of green (Chelonia mydas) and loggerhead (Caretta caretta) turtles to and from Eastern Australian Rookeries. Wildl. Res.. 19, 347–358.

Limpus, C. J., Parmenter, C. J., Baker, V., Fleay, A. (1983) The flatback turtle, Chelonia depressa, in Queensland: Post-nesting migration and feeding ground distribution. Australian Wildlife Research. 10, 557–561.

Lindenbergh, R., Pietrzyk, P. (2015) Change detection and deformation analysis using static and mobile laser scanning. Applied Geomatics. 7, 65–74.

Lohmann, K. J., Lohmann, C. M. F., Brothers, J. R., Putman, N. F. (2013) Natal homing and imprinting in sea turtles. In: Biology of Sea Turtles: Volume III (eds J. Wyneken, K. J. Lohmann, & J. A. Musick) pp. 59–78. CRC Press Taylor & Francis Group, Boca Raton, FL.

Lohmann, K. J., Putman, N. F., Lohmann, C. M. F. (2008) Geomagnetic imprinting: A unifying hypothesis of long-distance natal homing in salmon and sea turtles. Proceedings of the National Academy of Sciences of the United States of America. 105, 19096–19101.

Long, T. M., Angelo, J., Weishampel, J. F. (2011) LiDAR-derived measures of hurricane- and restoration-generated beach morphodynamics in relation to sea turtle nesting behaviour. International Journal of Remote Sensing. 32, 231–241.

Luschi, P., Hays, G. C., Papi, F. (2003) A review of long-distance movements by marine turtles, and the possible role of ocean currents. Oikos. 103, 293–302.

Mancini, A., Koch, V. (2009) Sea turtle consumption and black market trade in Baja California Sur, Mexico. Endangered Species Research. 7, 1–10.

Mansfield, K. L., Wyneken, J., Porter, W. P., Luo, J. (2014) First satellite tracks of neonate sea turtles redefine the ‘lost years’ oceanic niche. Proceedings of the Royal Society: Biology. 281, 20133039.

Marcovaldi, M., Chaloupka, M. (2007) Conservation status of the loggerhead sea turtle in Brazil: An encouraging outlook. Endangered Species Research. 3, 133–143.

Marine Turtle Specialist Group (1996) Lepidochelys kempii. IUCN Red List of Threatened Species, e.T11533A3292342.

Martinuzzi, S., Vierling, L. A., Gould, W. A., Falkowski, M. J., Evans, J. S., Hudak, A. T., Vierling, K. T. (2009) Mapping snags and understory shrubs for a LiDAR-based assessment of wildlife habitat suitability. Remote Sensing of Environment. 113, 2533–2546.

Mason, D. C., Gurney, C., Kennett, M. (2000) Beach topography mapping: A comparison of techniques. Journal of Coastal Conservation. 6, 113–124.

Mazaris, A. D., Matsinos, G., Pantis, J. D. (2009) Evaluating the impacts of coastal squeeze on sea turtle nesting. Ocean and Coastal Management. 52, 139–145.

APPENDIX I: LITERATURE REVIEW

75 APPENDIX I: LITERATURE REVIEW

Mazor, T., Levin, N., Possingham, H. P., Levy, Y., Rocchini, D., Richardson, A. J., Kark, S. (2013) Can satellite-based night lights be used for conservation? The case of nesting sea turtles in the Mediterranean. Biological Conservation. 159, 63–72.

McClellan, C. M., Braun-McNeill, J., Avens, L., Wallace, B. P., Read, A. J. (2010) Stable isotopes confirm a foraging dichotomy in juvenile loggerhead sea turtles. Journal of Experimental Marine Biology and Ecology. 387, 44–51.

McCormick, M. I. (1994) Comparison of field methods for measuring surface topography and their associations with a tropical reef fish assemblage. Marine Ecological Press Series. 112, 87–96.

Mendonça, V. M., Al Saady, S., Al Kyumi, A., Erzini, K. (2010) Interactions between green turtle (Chelonia mydas) and foxes (Vulpes vulpes arabica, V. rueppellii sabaea, and V. cana) on turtle nesting grounds in the northwestern Indian Ocean: Impacts of the fox community on the behavior of nesting sea turtles at the Ras Al Hadd Turtle Reserver, Oman. Zoological Studies. 49, 437–452.

van de Merwe, J. P., West, E. J., Ibrahim, K. (2012) Effects of off-road vehicle tyre ruts on the beach dispersal of green sea turtle Chelonia mydas hatchlings. Endangered Species Research. 18, 27–34.

Meylan, A. B., Donnelly, M. (1999) Status justification for listing the hawksbill turtle (Eretmochelys imbricata) as Critically Endangered on the 1996 IUCN Red List of Threatened Animals. Chelonian Conservation and Biology. 3, 200–224.

Miller, J. D., Dobbs, K. A., Limpus, C. J., Mattocks, N., Jr, A. M. L. (1998) Long-distance migrations by the hawksbill turtle, Eretmochelys imbricata, from north-eastern Australia. Wildlife Research. 25, 89.

Milton, S. L., Leone-Kabler, S., Schulman, A., Lutz, P. L. (1994) Effects of Hurricane Andrew on the sea turtle nesting beaches of South Florida. Bulletin of Marine Science. 54, 974–981.

Minton, C., Connor, M., Price, D., Jessop, R., Collins, P., Sitters, H., Hassell, C., Pearson, G., Rogers, D. (2013) Wader numbers and distribution on Eighty Mile Beach, north-west Australia: Baseline counts for the period 1981-2003. Conservation Science Western Australia. 8, 345–366.

Montaghi, A., Corona, P., Dalponte, M., Gianelle, D., Chirici, G., Olsson, H. (2013) Airborne laser scanning of forest resources: An overview of research in Italy as a commentary case study. International Journal of Applied Earth Observation and Geoinformation. 23, 288–300.

Morgan, R. C. (2007) Property of spirits: Hereditary and global value of sea turtles in Fiji. Human Organization. 66, 60–68.

Mortimer, J. A. (1982) Factors influencing beach selection nesting turtles. In: Biology and conservation of sea turtles (ed K. A. Bjorndal) pp. 45–51. Smithsonian Institution Press Pages, Washington, DC.

Mortimer, J. A. (1990) The influence of beach sand characteristics on the nesting behavior and clutch surivival of green turtles (Chelonia mydas). Copeia. 1990, 802–817.

Mortimer, J. A., Camille, J.-C., Boniface, N. (2011) Seasonality and status of nesting hawksbill (Eretmochelys imbricata) and green turtles (Chelonia mydas) at D’Arros Island, Amirantes Group, Seychelles. Chelonian Conservation and Biology. 10, 26–33.

APPENDIX I: LITERATURE REVIEW

76 APPENDIX I: LITERATURE REVIEW

Mortimer, J. A., Donnelly, M. (2008) Eretmochelys imbricata. IUCN Red List of Threatened Species, e.T8005A12881238.

Mrosovsky, N., Bass, A., Corliss, L. A., Richardson, J. I., Richardson, T. H. (1992) Pivotal and beach temperatures for hawksbill turtles nesting in Antigua. Canadian Journal of Zoology. 70, 1920–1925.

Mrosovsky, N., Provancha, J. (1992) Sex ratio of hatchling loggerhead sea turtles: Data and estimates from a 5-year study. Canadian Journal of Zoology. 70, 530–538.

Nel, R., Punt, A. E., Hughes, G. R. (2013) Are coastal protected areas always effective in achieving population recovery for nesting sea turtles? PLoS One. 8, e63525.

Patino-Martinez, J., Marco, A., Quiñones, L., Hawkes, L. A. (2014) The potential future influence of sea level rise on leatherback turtle nests. Journal of Experimental Marine Biology and Ecology. 461, 116–123.

Pendoley, K., Kamrowski, R. L. (2015) Influence of horizon elevation on the sea-finding behaviour of hatchling flatback turtles exposed to artificial light glow. Marine Ecology Progress Series. 529, 279–288.

Pendoley, K., Kamrowski, R. L. (2016) Sea-finding in marine turtle hatchlings: What is an appropriate exclusion zone to limit disruptive impacts of industrial light at night? Journal for Nature Conservation. 30, 1–11.

Pendoley, K. L., Bell, C. D., McCracken, R., Ball, K. R., Sherborne, J., Oates, J. E., Becker, P., Vitenbergs, A., Whittock, P. A. (2014a) Reproductive biology of the flatback turtle Natator depressus in Western Australia. Endangered Species Research. 23, 115–123.

Pendoley, K. L., Schofield, G., Whittock, P. A., Ierodiaconou, D., Hays, G. C. (2014b) Protected species use of a coastal marine migratory corridor connecting marine protected areas. Mar. Biology. 161, 1455–1466.

Pendoley, K. L., Whittock, P. A., Vitenbergs, A., Bell, C. (2016) Twenty years of turtle tracks: Marine turtle nesting activity at remote locations in the Pilbara, Western Australia. Australian Journal of Zoology. 64, 217–226.

Poloczanska, E. S., Limpus, C. J., Hays, G. C. (2010) Vulnerability of marine turtles to climate change. In: Advances in Marine Biology (ed D. W. Sims) pp. 151–211. Elsevier Ltd, London, UK.

Polovina, J., Howell, E., Parker, D., Balazs, G. (2003) Dive-depth distribution of loggerhead (Carretta carretta) and olive ridley (Lepidochelys olivacea) sea turtles in the Central North Pacific: Might deep longline sets catch fewer turtles? Fishery Bulletin. 101, 189–193.

Red List Standards & Petitions Subcommittee (1996) Natator depressus. IUCN Red List of Threatened Species, e.T14363A4435952.

Rizkalla, C. E., Savage, A. (2011) Impacts of seawalls on loggerhead sea turtle (Caretta caretta) nesting and hatching success. Journal of Coastal Research. 27, 166–173.

Roe, J. H., Clune, P. R., Paladino, F. V (2013) Characteristics of a leatherback nesting beach and implications for coastal development. Chelonian Conservation and Biology. 12, 34–43.

Rogalski, A., Chrzanowski, K. (2014) Infrared devices and techniques (revision). Metrology and Measurement Systems. 21, 565–618.

APPENDIX I: LITERATURE REVIEW

77 APPENDIX I: LITERATURE REVIEW

Rogers, D. I., Hassell, C. J., Boyle, A., Gosbell, K., Minton, C., Rogers, K. G., Clarke, R. H. (2011) Shorebirds of the Kimberley coast - populations, key sites, trends and threats. Journal of the Royal Society of Western Australia. 94, 377–391.

Salmon, M., Hamann, M., Wyneken, J. (2010) The development of early diving behavior by juvenile flatback sea turtles (Natator depressus). Chelonian Conservation and Biology. 9, 8–17.

Salmon, M., Hamann, M., Wyneken, J., Schauble, C. (2009) Early swimming activity of hatchling flatback sea turtles Natator depressus: A test of the ‘predation risk’ hypothesis. Endangered Species Research. 9, 41–47.

Salmon, M., Reiners, R., Lavin, C., Wyneken, J. (1995) Behavior of loggerhead sea turtles on an urban beach. I. Correlates of nest placement. Journal of Herpetology. 29, 560–567.

Salmon, M., Tolbert, M. G., Painter, D. E., Goff, M., Reiners, R. (1995) Behaviour of loggerhead sea turtles on an urban beach. II. Hatchling orientation. Journal of Herpetology. 29, 568–576.

Salmon, M., Witherington, B. E. (1995) Artificial lighting and seafinding by loggerhead hatchlings: Evidence for lunar modulation. Copeia. 4, 931–938.

Schuyler, Q., Hardesty, B. D., Wilcox, C., Townsend, K. (2012) To eat or not to eat? Debris selectivity by marine turtles. PLoS One. 7, e40884.

Scott, R., Biastoch, A., Roder, C., Stiebens, V. A., Eizaguirre, C. (2014) Nano-tags for neonates and ocean-mediated swimming behaviours linked to rapid dispersal of hatchling sea turtles. Proceedings of the Royal Society: Biology, 1–9.

Segura, L. N., Cajade, R. (2010) The effects of sand temperature on pre-emergent green sea turtle hatchlings. Herpetological Conservation and Biology. 5, 196–206.

Sellars, J. D., Jolls, C. L. (2007) Habitat modeling for Amaranthus pumilus: An application of light detection and ranging (LIDAR) data. Journal of Coastal Research. 23, 1193–1202.

Semeniuk, V. (2008) Holocene sedimentation, stratigraphy, biostratigraphy, and history of the Canning Coast, north-western Australia. Journal of the Royal Society of Western Australia. 91, 53–148.

Seminoff, J. A. (2004) Chelonia mydas. IUCN Red List of Threatened Species, e.T4615A11037468.

Shaver, D. J., Lamont, M. M., Maxwell, S., Walker, J. S., Dellingham, T. (2016) Head-started Kemp’s ridley turtle (Lepidochelys kempii) nest recorded in Florida: Possible implications. Chelonian Conservation and Biology. 15, 138–143.

Sherman, D. J., Hales, B. U., Potts, M. K., Ellis, J. T., Liu, H., Houser, C. (2013) Impacts of Hurricane Ike on the beaches of the Bolivar Peninsula, TX, USA. Geomorphology. 199, 62–81.

Short, A. D. (2006a) Australian beach systems—nature and distribution. Journal of Coastal Research. 221, 11–27.

Short, A. D. (2006b) Beaches of the Western Australian Coast: Eucla to Roebuck Bay. Sydney University Press, Sydney, Australia.

Short, F. T., Neckles, H. A. (1999) The effects of global climate change on seagrasses. Aquatic Botany. 63, 169–196.

APPENDIX I: LITERATURE REVIEW

78 APPENDIX I: LITERATURE REVIEW

Shrestha, R. L., Carter, W. E., Sartori, M., Luzum, B. J., Slatton, K. C. (2005) Airborne laser swath mapping: Quantifying changes in sandy beaches over time scales of weeks to years. ISPRS journal of Photogrammetry & Remote Sensing. 59, 222–232.

Sim, E. L., Booth, D. T., Limpus, C. J., Guinea, M. L. (2014) A comparison of hatchling locomotor performance and scute pattern variation between two rookeries of the flatback turtle (Natator depressus). Copeia. 2, 339–344.

Simonson, W. D., Allen, H. D., Coomes, D. A. (2014) Applications of airborne lidar for the assessment of animal species diversity. Methods in Ecology and Evolution. 5, 719–729.

Smith-Marder, P. (2006) The turtle and the psyche. Psychological Perspectives. 49, 228–248.

Spanier, M. J. (2010) Beach erosion and nest site selection by the leatherback sea turtle Dermochelys coriacea (Testudines: Dermochelyidae) and implications for management practices at Playa Gandoca, Costa Rica. Revista de Biología Tropical. 58, 1237–46.

Sperling, J. B., Grigg, G. C., Limpus, C. J. (2010) Diving behaviour in two distinct populations of gravid Flatback turtles Natator depressus. Australian Zoologist. 35, 291–306.

Stringell, T. B., Calosso, M. C., Claydon, J. A. B., Clerveaux, W., Godley, B. J., Lockhart, K. J., Phillips, Q., Ranger, S., Richardson, P. B., Sanghera, A., Broderick, A. C. (2013) Marine turtle harvest in a mixed small-scale fishery: Evidence for revised management measures. Ocean and Coastal Management. 82, 34–42.

Tattoni, C., Rizzolli, F., Pedrini, P. (2012) Can LiDAR data improve bird habitat suitability models? Ecological Modelling. 245, 103–110.

Thums, M., Whiting, S. D., Reisser, J., Pendoley, K. L., Pattiaratchi, C. B., Proietti, M., Hetzel, Y., Fisher, R., Meekan, M. G. (2016) Artificial light on water attracts turtle hatchlings during their near shore transit. Royal Society Open Science. 3, 160142.

Tomillo, P. S., Oro, D., Paladino, F. V., Piedra, R., Sieg, A. E., Spotila, J. R. (2014) High beach temperatures increased female-biased primary sex ratios but reduced output of female hatchlings in the leatherback turtle. Biological Conservation. 176, 71–79.

Tucker, T., Whiting, S., Mitchell, N., Berry, O., FitzSimmons, N., Pendoley, K. (2015) The Kimberley marine turtle project under the Western Australian Marine Sciences Institution (WAMSI). In: Proceedings of the Second Australian and Second Western Australian Marine Turtle Symposia, Perth 25-27 August 2014 (eds S. D. Whiting & A. Tucker) pp. 79–81. Science Division, Department of Parks and Wildlife, Perth.

Tuxbury, S. M., Salmon, M. (2005) Competitive interactions between artificial lighting and natural cues during seafinding by hatchling marine turtles. Biological Conservation. 121, 311–316.

Valverde, R. A., Orrego, C. M., Tordoir, M. T., Gómez, F. M., Solís, D. S., Hernández, R. A., Gómez, G. B., Brenes, L. S., Baltodano, J. P., Fonseca, L. G., Spotila, J. R. (2012) Olive ridley mass nesting ecology and egg harvest at Ostional Beach, Costa Rica. Chelonian Conservation and Biology. 11, 1–11.

Vierling, K. T., Vierling, L. A., Gould, W. A., Martinuzzi, S., Clawges, R. M. (2008) Lidar: Shedding new light on habitat characterization and modeling. Frontiers in Ecology and the Environment. 6, 90–98.

Walker, T. A., Parmenter, C. J. (1990) Absence of a pleagic phase in the life cycle of the flatback turtle, Natator depressa (Garman). Journal of Biogeography. 17, 275–278.

APPENDIX I: LITERATURE REVIEW

79 APPENDIX I: LITERATURE REVIEW

Wallace, B. P., DiMatteo, A. D., Bolten, A. B., Chaloupka, M. Y., Hutchinson, B. J., Abreu-Grobois, F. A., Mortimer, J. A., Seminoff, J. A., Amorocho, D., Bjorndal, K. A., Bourjea, J., Bowen, B. W., Dueñas, R., Casale, P., Choudhury, B. C., Costa, A., Dutton, P. H., Fallabrino, A., Finkbeiner, E. M., Girard, A., et al. (2011) Global conservation priorities for marine turtles. PLoS One. 6, e24510.

Wallace, B. P., Heppell, S. S., Lewison, R. L., Kelez, S., Crowder, L. B. (2008) Impacts of fisheries bycatch on loggerhead turtles worldwide inferred from reproductive value analyses. Journal of Applied Ecology. 45, 1076–1085.

Wallace, B. P., Tiwari, M., Girondot, M. (2013) Dermochelys coriacea. IUCN Red List of Threatened Species.

Wedding, L. M., Friedlander, A. M., McGranaghan, M., Yost, R. S., Monaco, M. E. (2008) Using bathymetric lidar to define nearshore benthic habitat complexity: Implications for management of reef fish assemblages in Hawaii. Remote Sensing of Environment. 112, 4159–4165.

Whittock, P. A., Pendoley, K. L., Hamann, M. (2014) Inter-nesting distribution of flatback turtles Natator depressus and industrial development in Western Australia. Endangered Species Research. 26, 25–38.

Whittock, P. A., Pendoley, K. L., Hamann, M. (2016) Flexible foraging: Post-nesting flatback turtles on the Australian continental shelf. Journal of Experimental Marine Biology and Ecology. 477, 112–119.

Wilson, B. (2013) The Biogeography of the Australian North West Shelf: Enviornmental Change and Life’s Responses. Elsevier, Perth, Australia.

Witherington, B., Hirama, S., Mosier, A. (2011a) Sea turtle responses to barriers on their nesting beach. Journal of Experimental Marine Biology and Ecology. 401, 1–6.

Wood, D. W., Bjorndal, K. A. (2000) Relation of temperature, moisture, salinity, and slope to nest site selection in loggerhead sea turtles. Copeia. 2000, 119–119.

Woodroffe, C. D. (2002) Beach and barrier coasts. In: Coasts pp. 248–320. Cambridge University Press, Cambridge, UK.

Woolard, J. W., Colby, J. D. (2002) Spatial characterization, resolution, and volumetric change of coastal dunes using airborne LIDAR: Cape Hatteras, North Carolina. Geomorphology. 48. 269–287.

Xiaoye, L. (2008) Airborne LiDAR for DEM generation: some critical issues. Progress in Physical Geography. 32, 31–49.

Yamamoto, K. H., Anderson, S. J., Sutton, P. C. (2015) Measuring the effects of morphological changes to sea turtle nesting beaches over time with LiDAR data. J. Sea Res. doi: 10.1016/j.seares.2015.07.001.

Yamamoto, K. H., Powell, R. L., Anderson, S., Sutton, P. C. (2012) Remote sensing of environment using LiDAR to quantify topographic and bathymetric details for sea turtle nesting beaches in Florida. Remote Sensing of Environment. 125, 125–133.

Zhang, K., Whitman, D., Leatherman, S., Robertson, W. (2005) Quantification of beach changes caused by Hurricane Floyd along Florida’s Atlantic coast using airborne laser surveys. Journal of Coastal Research. 211, 123–134.