Responses of Ghost Crabs and Other Scavengers to Habitat Modification of Urban Sandy Beaches

Responses of ghost crabs and other scavengers to habitat modification of urban sandy beaches

Talia Stelling-Wood B.A.Sc and M.M.S.M

A thesis in fulfilment of the requirements for the degree of Master of Philosophy at The University of New South Wales

School of Biological, Earth and Environmental Sciences Faculty of Science

March 2016

i PLEASE TYPE

THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: STELLING-WOOD

First name: TALIA Other name/s: PETA

Abbreviation for degree as given in the University calendar: M.PHIL

School: BEES Faculty: SCIENCE

Title: Scavenger responses to human impacts on urban sandy beaches

Abstract 350 words maximum: (PLEASE TYPE)

Populations in coastal areas are growing at a rate far exceeding that of other areas, resulting in escalating pressures on coastal areas around the world. Sandy beaches make up over 70% of the world coastlines, comprising some of the most highly sought after areas for human activities. Consequently, sandy beaches are some of the most vulnerable habitats to the effects of urbanisation, with beaches now threatened by a wide range of anthropogenic practices.

Ghost crabs are widely used as ecological indicators of ecosystem health on sandy beaches, but have rarely been studied on beaches within large urban centres. In this thesis, ghost crab burrow densities were used to assess human impacts on beaches in the highly urbanised estuary, Sydney Harbour. Across 38 beaches in the harbour, mechanical beach cleaning frequency was found to be the most influential predictor of ghost crab distributions in the area, with burrow counts reduced on the most frequently cleaned beaches and highest burrow counts found on beaches that were cleaned no more than three times per week. Mechanical cleaning was associated with the volume of wrack accumulated onshore, with frequently cleaned beaches supporting significantly lower volumes of wrack then beaches that were cleaned less frequently. To examine how organic subsidies on sandy beaches affect crab behaviour, I experimentally manipulated the availability of organic marine matter (wrack and carrion). Ghost crabs demonstrated no short term response to either subsidy (measured by the initiation of new burrows near food source), while vertebrate scavengers demonstrated a rapid response to carrion subsidies only. The presence of vertebrate scavengers in treatment plots reduced the fossorial behaviour of ghost crabs, suggesting biological interactions among invertebrates and vertebrates. The arrival of marine subsidies to sandy beaches facilitates the exchange of resources between the marine and terrestrial zone. Given that these systems can rely heavily on this resource exchange any alteration in the natural supply of organic marine matter, e.g. more frequent beach cleaning in response to urbanisation, could have far reaching implications for both zones.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

………………………………………………………… ……………………………………..……………… ……….……………………...…….…

Signature Witness Signature Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………......

Date ……………………………………………......

iii

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………...... Acknowledgements

First of all, I’d like to thank my supervisor Alistair Poore. Thank you for giving me this amazing opportunity and for constantly providing me with the guidance and support I needed to get me here today. Alistair thank you for always calming me when I felt overwhelmed and for letting me interrupt too many lunches with my ‘quick’ questions. Your dedication and passion for science is infectious and you are a true inspiration. I could not have wished for a better supervisor.

My next thanks go to each and every member, past and present, volunteer and intern, of Poore’s lab and Applied Marine and Estuarine Ecology Lab at UNSW that I met whilst completing this Masters. In no particular order, Nigel Coobes, Jamie-Louise Morrison, Melanie Sun, Vivian Sim, Simone Birrer, Kingsley Griffen, Aria Lee, Brendan Lanham, Hannah Sheppard Brennand, Jaz Lawes, Shin Ushiama, Jess Merrett, Nina Schaefer, Natalie Rivero, Luke Hedge, Damon Bolton, Katie Dafforn, Mariana Mayer Pinto, Sally Bracewell, Keryn Bain, Ana Bugnot, Mark Browne and Hannah Ward. You are all amazing, talented people and I am honoured to have met you all. Thanks to Graeme Clark for all his help and statistical advice and for co-authoring my first published manuscript. Special thanks go out to Janine Ledet for her extensive help with fieldwork and her constant motivational talks from across the globe. Another special mention goes out to James Lavender who I absolutely couldn’t have done this without. You were my stats tutor, advisor and just all round go-to-guy for any thesis related question and I thank you for your help and support through it all.

I also have to thank my family for supporting me through it all, and especially to my dad for his ‘sponsorship’ through not only this degree, but my entire life as a student (which has been quite a long time). A special thanks goes out to my sister Arkie for all those 4.30am wake-ups to ‘count crabs’ and her dedication to the 6am club. Thank you to all my friends for their endless supply of crab jokes and for at least feigning interest when I tell them about my research. Finally, I’d like to thank the friendly and helpful staff at the various local councils for their assistance throughout this study.

Abstract

Contents

Acknowledgements iv Abstract v Contents vi List of Figures vii List of Tables ix List of Appendices x List of Publications xi

1. General Introduction 1 Habitat modification of urban sandy beaches ...... 2 Thesis aims ...... 3

2. Responses of ghost crabs to habitat modification on urban sandy beaches 5 Abstract ...... 5 Introduction ...... 6 Materials and methods ...... 8 Results ...... 14 Discussion ...... 21

3. Scavenger responses to enhanced trophic subsidies on a sandy beach 26 Abstract ...... 26 Introduction ...... 27 Materials and methods ...... 29 Results ...... 32 Discussion ...... 35

4. General discussion 39 Habitat modification of sandy beaches ...... 39 Trophic subsidies to sandy beaches ...... 40

References 42

Appendices 49

List of Figures

Fig 1.1. Examples of habitat modification of urban sandy beaches in Sydney. a) coastal armouring, b) mechanical beach cleaning, c) coastal development, and d) tyre tracks from vehicle use...... 3

Fig. 2.1. Map of Sydney Harbour showing location of study sites, with Sydney Harbour indicated on insert of map of Australia (33° 51’ S, 151° 14’ E). The size of symbol at each site corresponds to mean ghost crab burrow abundance/m2...... 9

Fig. 2.2. a) The smooth-handed ghost crab, Ocypode cordimanus, in burrow opening, b) active ghost crab burrow ...... 11

Fig. 2.3. Variation in ghost crab burrow counts with the anthropogenic predictor variables of (a) cleaning regime, (b) seawall extent and (c) level of development surrounding each beach. Plots are GLMM estimates, error bars are 95% confidence intervals and categories that share a letter do not differ in post hoc analysis ...... 16

Fig. 2.4. Variation in ghost crab burrow count with the environmental variables of (a) beach length, (b) beach slope and (c) mean grain size. Each point represents one transect. Points were given opacity, therefore darker areas represent higher densities...... 18

Fig. 2.5. Variation in ghost crab burrow count with the productivity variables of (a) %TOM and (b) wrack volume (cm3). Each point represents one transect. Points were given opacity so that darker areas represent higher densities. (c) Variation in wrack volume with cleaning regime. Plots are GLMM estimates, error bars are 95% CI and categories that share a letter do not differ in post hoc analysis...... 19

Fig. 3.1. Vertebrate scavenger tracks adjacent to experimental plots, a) rodent tracks and b) avian tracks...... 31

vii

Fig. 3.2. Number of new ghost crab burrows formed per plot overnight in response to wrack and carrion subsidies. Data are means ± standard error...... 33

Fig. 3.3. (a) Vertebrate scavenger response to treatments. The metric of vertebrate response was the presence of tracks on treatment plots. (b) Numerical response of ghost crabs to the presence of vertebrate scavengers within treatment plots. The metric of ghost crab response was the number of new burrows formed overnight in a treatment plot ...... 34

viii

List of Tables

Table 2.1. Bivariate GLMMs for ghost crab burrow count and burrow diameter tested against each predictor variable. For burrow count, beach and time were random factors and a negative binominal error distribution was used. For burrow diameter, beach, time and transect were random factors and a negative binominal error distribution was used. Significant predictors are indicated with *...... 15

Table 2.2. Model selection table using GLMMs to predict ghost crab burrow count, with Poisson distribution and beach and time as random effects. Wrack, length and %TOM were log transformed. Best model is shown in bold. ∆AIC indicates the difference in model parsimony as explained by AIC relative to the best model; lower ∆AIC values indicate higher support for the model ...... 20

List of Appendices

Supplementary material published in Stelling-Wood et al. (2016)

Fig. S1. Ranked importance of variables in predicting ghost crab burrow counts according to random forest algorithm ...... 49

Table S1. List of surveyed beaches and corresponding anthropogenic variable category classifications ...... 50

List of Publications

Stelling-Wood, T.P., Clark, G.F., Poore, A.G.B., 2016. Responses of ghost crabs to habitat modification of urban sandy beaches. Marine Environmental Research. 116: 32-40.

Author contributions: TSW designed the study, collected the data, analysed the data and wrote the manuscript; GC analysed data and edited the manuscript; AP designed the study and edited the manuscript.

Chapter 1

General introduction

Coastal zones contain diverse and productive habitats that are important for development and human settlement, with 44 % of the world's population living within 150 kilometres of the coast (UN Atlas of Oceans, 2010). Sandy beaches make up over 70% of the world’s open coastlines (Schlacher et al., 2007a), and comprise some of the most sought after and highly valued areas for human activities. This makes sandy beaches some of the most vulnerable habitats to threats from urbanisation, both presently and in the future with predicted population increases.

Sandy beaches are resilient environments, with the ability to change shape and size naturally in response to variations in wind and wave energy. Human modification of coastlines limits the capacity of beaches to withstand this environmental variability, making them more susceptible to erosion and habitat loss (Nordstrom, 2004). Beaches in urban areas are now trapped between intense pressure from human activities on the terrestrial side and global climate change on the oceanic side (Schlacher et al., 2007a).

Sandy beaches are dynamic environments, governed by physical and biological factors at different temporal and spatial scales, making them difficult environments to study. For example, waves, tides and currents often limit opportunity for spatial and temporal replication in beach environments (Barros, 2001). Consequently management of beaches has traditionally focused on maintaining and restoring physical and geomorphological features important for shoreline protection, with little consideration for the conservation of ecological features and processes (James, 2000).

Beaches are highly permeable boundaries between marine and terrestrial zones, facilitating the exchange of resources between these systems (Polis et al., 1997). They are dynamic environments, inhabited by biotic assemblages that are often endemic to these habitats (McLachlan and Brown, 2006). They provide unique ecological services, such as water filtration and nutrient cycling that are not covered by any other ecosystems (McLachlan and Brown, 2006). Coastal dunes provide essential habitat for plants and invertebrates, as well as feeding and nesting sites for birds and turtles (Schlacher et al., 2011).

Chapter 1

Habitat modification of urban sandy beaches

The ecological value and functioning of sandy beaches, is often considered secondary to their economic value, which combined with surging population growth in coastal areas, has led to wide spread modification of these environments (Defeo et al., 2009). Beaches are now threatened by a range of human activities, including coastal development and armouring practices, beach management practices, and human recreational activities (Defeo et al., 2009). Coastal development involves the construction of human infrastructure in coastal areas, whilst shore armouring is the process of building seawalls to protect this infrastructure from oceanic forces (Lucrezi et al., 2010). Both involve the conversion of natural habitat into hard substrate, with the potential to have severe negative impacts on these environments (Nordstrom, 2004). The increased use of beaches in urban areas has led to the intensification of beach management practices, especially mechanical beach cleaning, to improve the superficial quality of these beaches. This process uses heavy machinery to remove debris from the beach surface. Whilst removing potentially hazardous human waste this process also removes organic material that washes up on shore, which is an important food source for these ecosystems (McLachlan and Brown, 2006). This process is known to reduce the abundance of sandy beach fauna, (Defeo et al., 2009; Dugan and Hubbard, 2010; Gilburn, 2012) but is used anyway as an easy method to maintain the recreational quality of beaches. Human leisure and recreation also have the potential to heavily impact sandy beach ecosystems through the use of recreational off-road vehicles (Lucrezi et al., 2014; Schlacher et al., 2007b) and trampling associated with human activities (Brown and McLachlan, 1990; Reyes-Martínez et al., 2015).

Chapter 1

a) b)

c) d)

Fig 1.1. Examples of habitat modification of urban sandy beaches in Sydney. a) coastal armouring, b) mechanical beach cleaning, c) coastal development, and d) vehicle use on beaches. Photos: T.Stelling- Wood.

Thesis aims

Sandy beaches have been poorly studied, relative to other marine habitats, despite being some of the most vulnerable to threats from urbanisation. One way to address this shortfall in environmental data is to develop rapid assessment methods that monitor the abundance of an indicator species which is known to correlate with certain ecosystem functions or decline with known threats. On sandy beaches, ghost crabs are a widely used indicator species (Schlacher et al., 2016). They play an important ecological role, being both key consumers and important prey on sandy beaches (Schlacher et al., 2013b). They also display fossorial behaviour, making counting burrows an easy and efficient way to rapidly assess population size (Moss and McPhee, 2006).

Chapter 1

The aim of this thesis was to broaden our understanding of how human activities impact sandy beach communities, using ghost crabs as indicators of ecological health in these systems. While ghost crabs are a widely accepted indicator species, much of this reach stems from the ease with which populations size and structure can be easily estimated by examining burrow openings (Schlacher et al., 2016). This has led to an over representation in the literature of simple ‘compare and contrast designs’, with few studies investigating the mechanisms that accounts for these contrasts or lacking the potential to isolate the stressors responsible for these effects, given the multiple nature of human pressures on beaches (Schlacher et al., 2016). Consequently, the aim of this thesis was to fill this gap by quantifying the relationships between multiple possible stressors and their impacts on ghost crab densities.

The survey and experimental work was conducted in Sydney Harbour, located in New South Wales, on the eastern coast of Australia. It is one of Australia’s most urbanised estuaries, containing an extensive range of habitats, including over 50 sandy beaches. A recent review into the state of knowledge about Sydney Harbour identified that sandy beaches within the harbour have been poorly studied, finding only four research papers that investigated infaunal communities in these environments (Johnston et al. 2015). No studies to date have examined the possible impacts of the extensive habitat modification to coastal environments in this important harbour.

The aim of Chapter 2 was to identify which variables relating to physical beach properties, management practices and human induced habitat modification best predicted ghost crab distributions within Sydney Harbour. The aim of Chapter 3 was to examine the relationship between ghost crabs and washed up organic marine matter, by testing the short term response of sandy beach scavengers to enhanced trophic subsidies.

Chapter 2

Responses of ghost crabs to habitat modification of urban sandy beaches

Published as ‘Stelling-Wood, T.P., Clark, G.F., Poore, A.G.B., 2016. Responses of ghost crabs to habitat modification of urban sandy beaches. Marine Environmental Research 116: 32-40’

Abstract

Sandy beaches in highly urbanised areas are subject to a wide range of human impacts including mechanical beach cleaning and habitat modification. Ghost crabs are a commonly used ecological indicator on sandy beaches, as they are key consumers in these systems and display fossorial habits, allowing for rapid assessment of population size by counting burrow openings. This study assessed the pressures of urbanisation on sandy beaches in the highly urbanised estuary of Sydney Harbour. Across 38 beaches, I examined which physical beach properties, management practices and human induced habitat modification best predicted ghost crab distributions. Of all the variables measured, the frequency of mechanical beach cleaning was the most important predictor of crab abundance, with beaches cleaned less frequently (≤ 3 times per week) showing the highest burrow densities. These results indicate that ghost crab populations in Sydney Harbour are more robust to the impacts of urbanisation than previously thought.

Chapter 2

Introduction

The alteration of natural habitats by rapidly expanding human populations has caused impacts to ecosystems around the world (Defeo et al., 2009). Populations in coastal areas are growing significantly faster than anywhere else (Schlacher et al., 2007b), which coupled with increases in leisure time and higher human settlement in coastal areas, have resulted in escalating pressures on coastal areas around the world (Schlacher et al., 2008). Some of the most vulnerable habitats to the threats associated with high populations are sandy beach ecosystems, with high social, recreational and economic values combined with important ecological functioning (Schlacher et al., 2008).

Sandy beaches are dynamic environments, governed by a variety of physical and biological factors at different temporal and spatial scales (Defeo and McLachlan, 2005). Unconstrained, beaches can be resilient, changing shape and size naturally in response to storms and variations in wave action and currents (Schlacher et al., 2007a). Human modifications of sandy beaches interfere with these natural processes, thereby limiting their resilience and typically having negative impacts (Brown and McLachlan, 2002). Beaches in urban areas are now considered to be trapped in a ‘coastal squeeze’ between the impacts from human activities on the terrestrial side and the manifestations of climate change on the ocean side (Schlacher et al., 2007a).

Threats to beaches include a wide range of anthropogenic processes, ranging from habitat loss as a result of development and shore armouring, the over exploitation of resources (e.g., fishing and mining) to recreation activities and coastal management practices (Defeo et al., 2009). Coastal engineering practices such as land reclamation, development and shore armouring often result in the destruction of ecologically important natural habitat in coastal areas. These processes involve the conversion of natural habitat, such as native dune systems, into hard substrates like seawalls, with the potential for strong changes to the ecological values of these environments (Nordstrom, 2004). Recreational use of off-road vehicles on beaches and associated trampling can negatively affect beach habitats (Reyes-Martínez et al., 2015; Schlacher et al., 2007a), but even lower impact activities, such as swimming in the surf zone, can affect the activities of macrofauna and inhibit the feeding of higher order predators in the intertidal (Brown and McLachlan, 1990).

Increased use of beaches for recreational purposes has led to the intensification of beach cleaning practices, especially in urban areas. This process involves the use of heavy machinery to drag a large rake or sieve across the surface of the sand to remove debris from the surface layer. This process is designed to remove human waste and debris, however is non-selective

Chapter 2

and results in the removal of much of the surface layer including washed up organic marine matter known as wrack (Brown and McLachlan, 2002). Sandy beaches often lack attached larger plants and consequently in situ primary production is low (Schlacher et al., 2013a). The exception to this is beaches that support accumulations of surf diatoms, which can result in high levels of primary production (Campbell, 1996). As a result sandy beaches are reliant on this washed up organic material to fuel local food webs (McLachlan and Brown, 2006). The wrack supports a diverse array of organisms, including invertebrate macrofauna that may consume the wrack, meiofauna associated with sediments and predators such as shore birds (Dugan et al., 2003). The removal of wrack from these ecosystems can greatly reduce the abundance of sandy beach fauna (Defeo et al., 2009; Dugan and Hubbard, 2010; Gilburn, 2012). Despite these known impacts, the practice is used frequently by coastal managers around the world as an easy and efficient tool to maintain the aesthetic quality of sandy beaches for human use.

While known to be threatened by a range of human impacts, sandy beaches are poorly studied relative to other marine habitats. One potential solution to address the shortfall in environmental data on sandy beaches is to develop rapid assessment methods to provide measures that act as a surrogate for the health of the ecosystem as a whole. These include monitoring the abundance of indicator species whose abundance is assumed to correlate with known ecosystem functions or decline with known threats. Such methods have widely been used in other ecosystems, but have rarely been used in sandy beaches despite the fact they are some of the most vulnerable to threats from urbanisation (Barros, 2001).

On sandy beaches, ghost crabs (Crustacea: Brachyura: Ocypodidae) are widely used as a surrogate for the health of sandy beach ecosystems (Schlacher et al., 2016), and previous studies have found that populations of ghost crabs are susceptible to human activities on beaches around the world (e.g. Jonah et al., 2015; Schlacher et al., 2011). They are key consumers on sandy beaches (Schlacher et al., 2013b), playing the important ecological role of being apex invertebrate predator in these systems, whilst also being important prey for many higher order vertebrate consumers from nearby terrestrial ecosystems (Lucrezi and Schlacher, 2014). Their position as an apex consumer means that their population structure may reflect that of lower trophic levels (Lucrezi et al., 2009b). Finally, they display fossorial habits, constructing deep and complex burrows for shelter which are clearly visible on the surface of the sand which allow for rapid assessment of population size by counting burrow entrances (Moss and McPhee, 2006).

The aim of this study was to assess beaches in urban areas for the impacts of urbanisation. To do this I examined which variables relating to physical beach properties, management practices

Chapter 2

and human induced habitat modification best predicted ghost crab distributions in the highly urbanised estuary of Sydney Harbour. Sydney Harbour contains an extensive range of habitats, including over 50 sandy beaches, but a recent extensive review of the state of knowledge for the harbour (Johnston et al. 2015) identified only four research papers that have investigated infaunal communities of beach environments within the harbour (Dexter, 1983, 1984; Jones, 2003; Keats, 1997). Three of these studies were primarily concerned with describing the communities on only a few beaches, whilst the fourth study investigated the impact of an accidental oil spill on the amphipod Exeodiceros fossor (Jones, 2003). No studies to date have examined the possible impacts of the extensive habitat modification in this coastal system on sandy beach fauna.

To assess the impacts of urbanisation of ghost crab populations, I modelled burrow densities as a function of variables that described the biotic and abiotic properties of 38 sandy beaches in Sydney Harbour. Predictor variables included those that related to human modification of the beach environments (the presence or absence of a seawall, the level of surrounding development and the mechanical beach cleaning regime), the productivity of individual beaches (estimated using wrack accumulation and organic content of sediment) and beach morphology (length, beach slope and sediment grain size).

Materials and methods

Study sites and species distribution

Sydney is Australia’s largest city with a population of over 4 million people (Hutchings et al., 2013) and likely to rise to 8.5 million by 2061 (Australian Bureau of Statistics, 2013). Sydney Harbour is located on the eastern coast of New South Wales, Australia (33° 51’ S, 151° 14’ E, Fig. 2.1). The harbour is entirely within Sydney’s urban area, with the beaches being exposed to high levels of human activity. Ninety percent of the harbour’s catchment is now urbanised (Hutchings et al., 2013) and more than 50% of the harbour’s shoreline comprises artificial structures (Chapman and Bulleri, 2003). Previous studies involving human impacts on ghost crabs have been concentrated on open ocean beaches, and to date there have been no studies on beaches in estuarine environments within a large urban centre.

Beaches within the harbour experience a range of wave energy depending on their position within the estuary, from those directly exposed to open-ocean swell to those completely protected from all wave inputs, apart from wakes generated by boats (Kennedy, 2010). Consequently, harbour beaches represent a range of morphological types, with large amounts of environmental

Chapter 2

heterogeneity among beaches. Thirty-eight sandy beaches were selected for investigation in this study (Fig. 2.1, Table S1), being located throughout all sections of the harbour, representing a variety of morphological types and experiencing a range of urbanisation pressures (Table S1).

I surveyed burrow densities of the Smooth-handed Ghost Crab Ocypode cordimanus Latreille, 1818 (Fig. 2.2a). This species is found in the Indo-West Pacific and, in Australia, along the northern and eastern coasts, from the north of Western Australia to southern New South Wales (Jones and Morgan, 2002). Ocypode ceratophthalma also occurs along the east coast of Australia as far south as Sydney (Barros, 2001), but only O.cordimanus is known to occur within Sydney Harbour (Hutchings et al., 2013).

Australia

Chapter 2

Classification of beaches by degree of human modification

I identified the level of coastal development and the frequency of mechanical beach cleaning as likely threats to sandy beach communities in Sydney Harbour. Both involve intense modification of natural habitats and have been found to have negative impacts on sandy beach communities worldwide (Dugan and Hubbard, 2006; Gilburn, 2012; González et al., 2014). For each beach, the degree of coastal development was assessed using two measures, the presence of seawalls and the level of urban development on the landward side of the beach.

Beaches were assigned to one of three categories relating to the presence and size of a seawall; none present, partial (i.e., seawall for part or the length of the beach) or complete. Similarly, each beach was categorised into four levels of urban development directly adjacent to the beach (none, low, medium and high). Beaches allocated to the ‘none’ category were surrounded by dense natural vegetation and required visitors to walk a short distance down a bush track to reach the beach. Beaches assigned to the ‘low’ category were backed by vegetation commonly visited by humans, such as a grassy berm or park. Beaches assigned to the ‘medium’ category were adjacent to developed landscapes interspaced with some vegetation. Lastly, beaches assigned to the ‘high’ category had dense commercial or residential development adjacent to the sand, and no vegetation. Google Maps, local municipal councils and the surf lifesaving website were initially used to record beach length and to classify landscapes. These observations were then verified during the first field survey in July 2014.

Each beach was classified according to the technique and frequency with which they were cleaned. Local councils or beach management authorities were contacted and requested to provide the cleaning regimes for all beaches within their jurisdiction. From this information, three categories of cleaning frequency were established; the first category was ‘no mechanical cleaning’, including those beaches that were either not cleaned at all by local councils or those that were only cleaned by hand intermittently. The second category included those beaches that were infrequently cleaned (3 times or less per week) using a mechanical rake. The third category included those beaches that were frequently raked (5-7 times per week) using a mechanical rake. Similar methods of mechanical cleaning were used at all beaches cleaned in this fashion.

Chapter 2

Survey of ghost crab densities

All beaches were surveyed three times: July 2014, October 2014 and January 2015. On each sampling event, 10 quadrats (5 x 1 m) were randomly allocated to positions on the beach and the total number of crab burrows in each quadrat recorded. Quadrats were positioned at the base of either the seawall or at the most landward boundary of the beach, with the upper most edge of the quadrat adjacent to this boundary. They were positioned horizontally so that the long axis was parallel to the shoreline. The lower and middle shore regions were not sampled because a pilot study showed very few burrows lower on the beach.

a) b)

Fig. 2.2. a) The smooth-handed ghost crab, Ocypode cordimanus, in burrow opening, b) active ghost crab burrow. Photos: T. Stelling-Wood.

Counting burrows of crabs is a widely accepted tool for assessing the ecological impacts of human activities on beaches (e.g., Barros, 2001; Lucrezi et al. 2009) and previous studies have found burrow count to correlate with population size (Warren, 1990). However, to increase the accuracy of burrow counts as a measure of population size, only ‘active’ burrow openings were counted, recognised by either the presence of fresh tracks emanating from the opening, or evidence of recent re-working of the burrows visible as small mounds of excavated sediment next to the entrance, or both (Fig. 2.2b) (Pombo and Turra, 2013). To test for possible differences in population structure among beaches in Sydney Harbour, I measured the diameter of all burrows in each transect (to the nearest mm). Burrow diameter is known to correlate with body size for ghost crabs (Lucrezi et al., 2009b; Turra et al., 2005).

Chapter 2

Beach slope and grain size

Given that beach slope varies with sediment grain size and wave energy (Bascom, 1951), I used beach slope and mean grain size as proxy measures for beach exposure. Beach slope was measured during the October survey event using the modified Emery method (Delgado and Lloyd, 2004). This technique uses a stationary, vertical ruler and a 2 m long levelling set square with a fixed arc of 90°. The ruler is driven vertically into the sand and the set square is held perpendicular and lowered until the shortest end rests on the sand. The reading off the ruler indicates the vertical elevation difference across the 2m distance. The ruler is taken out and driven into the previous location of the levelling set square short leg and the procedure is repeated for the width of the beach. Three transects were taken in this way, perpendicular to the shoreline from the most landward boundary of the beach down to the waterline at low tide. These measurements were then averaged for each beach.

The grain size distributions of each beach were determined from 10 sediment samples collected during the October survey event. Five of these samples were taken from the swash zone low on the shore and five were taken from the upper reaches of the beach adjacent to the landward boundary, where most crab burrows were found. For each sample, 30 mL of sand was taken from the surface layer (approx. 100 g). Samples were refrigerated at 2 ᵒC to slow microbial degradation until they were processed. In the laboratory, samples were then split, with a small portion (approx. 10 g) being removed to measure organic content and the remainder of the sample being retained for sediment granulometry. The retained portion of the sample was washed in fresh water to remove salts from the sediment and dried at 80-90 ᵒC for 48 hours. Sediment granulometry was determined by dry sieving through a nested series of eight sieves arranged in decreasing order of mesh aperture size (4,000, 2,000, 1,000, 500, 250, 200, 180, 63 μm). Sediment parameters were calculated according to the Folk and Ward method using the Gradistat software (Blott and Pye, 2001).

Organic inputs to the sandy beaches

Ghost crabs are known to show high trophic plasticity, obtaining food from a wide variety of sources (Lucrezi and Schlacher, 2014). With the productivity on sandy beaches largely dependent on allochthonous organic inputs from other systems (Brown and McLachlan, 1990), I used wrack volume and percent total organic material of sediment (%TOM) as two proxies for potential food supply to each of the beaches. To measure %TOM, small portions of the sand samples collected for grain size analysis (above) were oven dried at 80-90 ᵒC for 24 hrs. Once dry, samples were weighed and then combusted in a muffle furnace at 550 ᵒC for 4 hours to

Chapter 2

remove all organic material from samples. Samples were left to cool in desiccators until room temperature and the weight loss of these samples was taken as the organic content.

The biomass of wrack was visually estimated using a modified photopoint method developed and validated by Duong (2008). I further developed this method to include depth measurements to allow for wrack volume to be estimated from photos. On each sampling event, 20 photos were taken with a GoPro Hero3+ with the field of view set to ‘medium’ to reduce distortion. Photos were taken within a 25 x 25 cm quadrat randomly positioned on each beach along the most recent high tide mark at each beach. Five measurements of wrack depth were taken from within the quadrat, which were then averaged to obtain a mean depth measurement for each quadrat. The percent cover of wrack in each photograph was measured using the image analysis program ImageJ (1.48v; National Institute of Health, Bethesda, MD, USA). The image contrast was adjusted so that the background (sand) was light and the foreground (wrack) was dark, then converted to a binary image and the percent cover of wrack calculated using partial analysis function in Image J. The binary images were compared to the original unprocessed images and, when necessary modifications were carried out by hand to ensure all part of the image that were not wrack were white and only wrack was black. This resulting percent cover measurement was then combined with the mean depth measurement to calculate volume of wrack per quadrat.

Data analyses

I used generalised linear mixed effects models (GLMM) to test for relationships between predictor variables and two response variables: burrow count and burrow diameter. I tested for relationships between both response variables and each predictor separately. Continuous predictor variables were beach slope, beach length, mean grain size, percent total organic material of sediment, volume of accumulated wrack; and categorical predictors were seawall extent, level of development surrounding beach, and mechanical cleaning regime. For grain size and organic content, separate analyses were conducted for samples taken from high and low positions on beaches. Each model included time (date of survey event) and beach as random factors. Beach was included as a random factor to allow me to generalise finding across all beaches in Sydney Harbour, and not just those sampled. The models used a negative binominal error distribution.

I then undertook model selection to determine the best single model for burrow count, using combinations of predictors. As above, the model included beach and time as random factors, and assumed Poisson error distribution. To avoid co-linearity between predictor variables, I

Chapter 2

first generated a covariation matrix, and removed variables until all pairwise correlations were < 0.6. Wrack, length and %TOM were log transformed. Grain size and %TOM samples from different beach heights were pooled, since no significant relationships were found between either variable and burrow count in the bivariate analyses. The initial (full) model was simplified by backwards selection, sequentially removing non-significant variables until all remaining variables were significant. The best model was selected based on AIC value (lowest AIC). Delta AIC was also calculated for each model, which indicates the difference in model parsimony as explained by AIC. Lower delta AIC values indicate higher support for the model.

A similar series of bivariate and backward selection analyses were performed with burrow diameter as the dependant response variable. Beach, time and transect were used as random factors, and models assumed Gaussian error distribution. To test how wrack volume varied with cleaning regime, I performed a GLMM with wrack volume (rounded to the nearest cm³) as the dependent variable, cleaning regime as a fixed factor and beach and time as random factors, again using a negative binominal error distribution. A similar analysis was used to test how wrack volume varied with seawall extent.

GLMMs were performed using the lme4 package (Bates et al., 2015) or the glmmADMB package (Skaug et al., 2012) for negative binominal error distributions in R v3.1.0 (R Core Team, 2014). Residual plots were inspected for model validation (Zuur et al., 2009). Differences among levels of significant categorical variables were analysed with pair-wise comparisons using the ‘relevel’ function.

To determine the robustness of these results I also analysed the data with random forest algorithms (Breiman, 2001). Tree-based algorithms have different assumptions to GLMM, and consistency between predictors of various methods can be used to gauge confidence in results. The random forest analysis was done using the party package (Strobl et al., 2008) in R v3.1.0 (R Core Team, 2014).

Results

Burrows of O. cordimanus were found on all but two Sydney Harbour beaches surveyed (Sirius Cove beach and Bradley Head amphitheatre beach, Fig. 2.1). Burrow densities were highly variable throughout the harbour, showing no clear geographical patterns (Fig. 2.1). The overall mean burrow density across all harbour beaches was 0.27 burrows m-².

Sixty-five percent of the beaches surveyed in Sydney Harbour had some form of shore armouring, with 42% being entirely backed by a seawall. Despite this extensive habitat

Chapter 2

modification, burrow count did not differ significantly with either the presence or type of seawall nor the level of development adjacent to the beach (Fig. 2.3b, c, Table 2.1). Seawall extent was, however, found to correlate with the volume of wrack on beaches (F(2, 4) = 6.32, p = 0.042). Beaches without any form of seawall were found with more variable but higher volumes of wrack, compared to those beaches backed by a partial or complete seawall.

Burrow count Burrow diameter

Variable Z P(> Z) Z P(> Z)

Length 0.68 0.41 0.04 0.85

Slope 0.02 0.89 0.06 0.81

Grain Size (High) 0.38 0.54 0.37 0.55

Grain Size (Low) 0.58 0.45 0.30 0.58

%TOM (High) 1.14 0.29 0.95 0.33

%TOM (Low) 0.54 0.46 1.97 0.16

Wrack 0.14 0.71 0.22 0.65

Seawall 0.92 0.63 0.58 0.75

Development 1.56 0.67 0.65 0.88

Cleaning 10.22 0.01* 1.48 0.48

Chapter 2

Fig. 2.3. Variation in ghost crab burrow counts with the anthropogenic predictor variables of (a) cleaning regime, (b) seawall extent and (c) level of development surrounding each beach. Plots are

Chapter 2

GLMM estimates, error bars are 95% confidence intervals and categories that share a letter do not differ in post hoc analysis.

Burrow count varied significantly among beaches with varying cleaning regime (Fig. 2.3a, Table 2.1), with burrow counts more variable and highest on infrequently cleaned beaches. In comparison, burrow counts were always low on frequently cleaned beaches and on beaches that were not mechanically cleaned at all (Fig. 2.3a).

There was no relationship between burrow count and any of our measures of beach morphology; beach length (Fig. 2.4a, Table 2.1), beach slope (Fig. 2.4b, Table 2.1) or mean grain size (Fig. 2.4c, Table 2.1). Neither of our proxy measures for beach productivity, wrack volume and %TOM, predicted burrow counts (Fig. 2.5a, b, Table 2.1). The volume of wrack did however vary significantly among beaches with differing cleaning regimes (Fig. 2.5c). Frequently cleaned beaches had significantly lower volumes of wrack, whilst infrequently cleaned beaches and beaches that were not mechanically cleaned at all had more variable and higher volumes of wrack.

There was little evidence that body size distributions of the crab populations varied among beaches (Table 2.1). No significant relationships were found between burrow diameter and the length, slope, mean grain size and either %TOM or wrack volume at either level on shore (Table 2.1). There were no significant differences in the burrow diameter across any of the categories of cleaning regime, seawall extent and coastal development (Table 2.1).

Chapter 2

relative to the best model; lower ∆AIC values indicate higher support for the model.

Model Cleaning Seawall Develop %TOM Wrack Length Slope Grainsize AIC ∆AIC

M1 X X 3496.0 0.0

M2 X X X 3496.2 0.2

M3 X X X X 3496.4 0.4

M4 X 3497.4 1.4

M5 X X X X X 3498.2 2.2

M6 X X X X X X 3499.3 3.3

M7 X X X X X X X 3501.4 5.4

M8 X X X X X X X X 3503.4 7.4

Null 3504.5 8.5

Model selection found that burrow count was best predicted by the combination of beach length and cleaning regime (M1: F3,9 = 14.435, p = 0.002). However, models containing the beach productivity variables, %TOM (M2) and wrack and %TOM (M3), as well as length and cleaning regime, were within two AIC of the best model (Table 2.2), and therefore are similar in performance to the best model (Burnham and Anderson, 2004). Despite similar AIC values, no significant differences were found between models containing beach productivity variables and

Chapter 2

the best model, suggesting these variables were not significant in predicting burrow distribution throughout Sydney Harbour. The model containing only cleaning regime (M4) also falls within two AIC values of the best model however, when length was removed from the model the AIC value significantly increased (Table 2.2).

Model selection using burrow diameter as the response variable found no significant difference between the full model and the null model (F(9,9) = 4.884, p = 0.9617), suggesting that none of the measured variables were able to predict burrow diameter. Consistent with the GLMMs, random forest analyses also found cleaning regime to be the strongest predictor of burrow counts (Fig. S1).

Discussion

This study indicates that ghost crab populations in a highly urbanised estuary were persisting despite extensive modification of their habitats and only reduced in the most intensively cleaned beaches. This study was the first to examine the impacts of urbanisation on ghost crab populations on estuarine sandy beaches.

Impact of beach cleaning

The frequency of mechanical beach cleaning was found to be the most influential predictor of ghost crab burrow densities in Sydney Harbour. This is consistent with studies globally, that have identified mechanical beach cleaning as a highly disturbing process on sandy beaches. For example, Dugan et al. (2003) found species richness, abundance and biomass of wrack- associated macrofauna was significantly reduced on beaches that were mechanically cleaned in contrast to those that were not cleaned. Similarly, Gilburn (2012) reported reductions in macroinvertebrate biodiversity on sandy beaches that were mechanically cleaned. In this study, however, I classified beaches by the frequency of mechanical cleaning, and found reduced burrow densities at the most intensively cleaned beaches while beaches that were infrequently cleaned (up to three times per week) supported the highest densities of ghost crab burrows. These results suggest that ghost crab populations within Sydney Harbour are resilient to some level of human disturbances. Similar results were found by Aheto et al. (2011) who found that a moderately disturbed beach had higher ghost crab burrow density compared to a disturbed beach, although this difference was not significant.

The finding of the lowest burrow densities on the most frequently cleaned beaches is consistent with several examples from the literature where intense disturbance has led to decreases in

Chapter 2

ghost crab populations (reviewed by Schlacher et al., 2016). For example, Noriega et al. (2012) found high use beaches that were mechanically cleaned every day to also have low burrow counts compared with lower usage beaches that were mechanically cleaned less frequently. Surprisingly, I also found similarly low burrow densities on unmanaged beaches that were never mechanically cleaned. Our analyses were not able to detect which variables might explain this pattern. However, while beach length did not predict burrow density across all beaches, the unmanaged beaches were often quite short, with most being no longer than 100m. Alternatively, differential predation of ghost crabs among sites within the harbour cannot be ruled out, as higher order vertebrate predators were not considered in this study. Finally, it is possible that lower crab numbers on unmanaged beaches represent the natural state of ghost crab populations in Sydney Harbour and that populations on the infrequently cleaned beaches are benefitting from the human activity on those beaches (e.g., ghost crabs benefitting from anthropogenic food subsidies, Schlacher et al. 2011).

Few studies have looked at the direct impacts of mechanical beach cleaning on ghost crab densities (Schlacher et al., 2016). However, the process of mechanical cleaning has been likened to the use of 4WD or off-road vehicles on sandy beaches in regards to environmental impact (Noriega et al, 2012). The impacts of vehicles on ghost crabs have been well studied. Lucrezi et al. (2014) found that despite heavy restrictions and regulations for vehicular traffic on South African beaches, burrow density and size were still impacted by the activity. They found burrow density at impacted sites to be less than a third of that at non-impacted sites, and found burrow size to be reduced by up to 50% at impacted sites. Similar results were found on North Stradbroke Island in Australia where Schlacher et al. (2007) found fewer crab burrows on sections of the beach which were subject to vehicle use. These reductions in population size are thought to be the result of crushing from the use of heavy vehicles on the sand, either at night when the crabs move down to the intertidal zone to feed (Moss and McPhee, 2006) or whilst they shelter in shallow burrows (Schlacher et al., 2007b).

In addition to the physical disturbance associated with mechanical beach cleaning, this activity can also impact sandy beach fauna by removing a potential food supply. Allochthonous wrack inputs are the dominant source of organic matter on sandy beaches and are highly variable in nature. I found that wrack volume did vary with cleaning regime, with the frequently cleaned beaches maintaining consistently low volumes of wrack, while infrequently cleaned and uncleaned beaches had higher, and more variable, volumes of wrack on our sampling dates. I did not, however, detect any relationship between burrow densities and wrack volume, or sediment organic content (as a potentially less variable proxy for the delivery of organic matter to the beach). This suggests that wrack accumulation may be too variable to accurately predict

Chapter 2

crab densities unless sampled extensively through time, further temporal sampling would be required to better understand the nature of wrack delivery to beaches in Sydney Harbour. I suggest cleaning regime is a better proxy for levels of organic material on each beach than wrack volume at a single sampling time. Alternatively this lack of relationship could suggest that ghost crab populations in Sydney Harbour do not rely on marine sourced trophic subsidies, instead potentially seeking our terrestrial based food sources. This, however, is unlikely given the highly modified nature of Sydney Harbour, where most of the native habitat surrounding beaches has been replaced with urban development.

The finding of highest burrow densities on beaches with high volumes of wrack (infrequently cleaned beaches) suggests that wrack may be an important trophic subsidy for ghost crabs in Sydney Harbour. The macroinvertebrates associated with wrack, or dead animal material deposited with the wrack, are potential food sources for crabs, and many studies have demonstrated that macroinvertebrate abundances are positively correlated with standing stock of wrack on sandy beaches (e.g. Ince et al., 2007 and Dugan et al., 2003). Ghost crabs show a strong positive response to trophic subsidies, such as those associated with wrack, carcasses from fishing activities and food scraps associated with beach side campsites (Schlacher et al. 2011; 2014).These strong responses to pulse food subsidies suggests that ghost crabs might be negatively impacted by the removal of trophic subsidies on sandy beaches through processes such as mechanical beach cleaning.

Ghost crabs are highly opportunistic feeders, known to demonstrate trophic plasticity, which is the ability to switch between scavenging and active predation in response to the availability of resources in their given environment (Morrow et al., 2014). This allows ghost crabs to be highly adaptive when it comes to feeding and could be the reason why their populations are able to persist in highly dynamic and increasingly urbanised environments such as Sydney Harbour.

Modification of beach habitats

The shoreline of Sydney Harbour is highly modified with more than 50% of its length comprised of artificial structures and 65% of the beaches surveyed here having coastal armouring. It involves the destruction of natural habitat on the upper reaches of beaches, areas that are known to be important for ghost crab populations (Noriega et al., 2012). Several studies have demonstrated that habitat loss and modification, as a result of coastal armouring, can result in significant reductions and even local extinctions of plant and animal species that occupy the upper reaches of sandy beaches (Dugan et al., 2008; Hubbard et al., 2014). However despite expectations, I found no association between the presence and extent of seawalls and ghost crab

Chapter 2

burrow densities. This contrasts with the well-known effects of coastal armouring practices on rocky shores, that include reductions in biodiversity and changes in community structure (Chapman and Bulleri, 2003), and the previous studies demonstrating lowered ghost crab abundances where beaches are armoured (Barros, 2001; Lucrezi et al. 2010). Lucrezi et al. (2009b) looked at the impacts of trampling in conjunctions with beach armouring, finding that burrow counts in disturbed areas were half those in undisturbed areas. Studies not involving ghost crabs have found coastal armouring can affect sandy beaches by reducing connectivity between marine and terrestrial environments (Heerhartz et al. 2015) as well as reducing the standing stock of wrack that washes up on beaches (Dugan and Hubbard 2006), consistent with our finding of wrack volumes correlating with seawall extent.

Urban developments on coastal fringes are often hubs of human activity, and human trampling has been identified as impacting ghost crab densities (reviewed by Schlacher et al., 2016). I used the level of surrounding development as a proxy for human visitation at beaches, but did not detect any relationship between this and ghost crab burrow densities. This contrasts with Jonah et al. (2015) classified beaches according to human usage, including a direct measure of human visitors. They found low use beaches had significantly higher numbers of burrows and larger burrow sizes compared to medium use and high use beaches. Reyes-Martinez et al. (2015) similarly found a dramatic reduction in species density and a significant change in the structure of the community from before to after impact (trampling). However this study however did not include ghost crabs. Noriega et al. (2012) found a strong negative correlation between ghost crab burrow density and degree of human use along an urbanisation gradient in the Gold Coast, Australia. However, caution should be used when considering both these results and those of the present study, as both used proxy measures for human visitation to beaches, with no direct measure of likely trampling. Classifying beaches into discrete categories may not reflect likely variation in beach visitation through time in response to weather or local tourist events.

Noriega at al. (2012) concluded intact dune systems with high vegetation cover was the most important habitat type for supporting high ghost crab densities at the low use beaches they studied. Although I found no relationship between development and ghost crab densities across all beaches, one Sydney Harbour beach that does still have an intact sand dune system (Chinamans Beach) had the highest burrow densities of all beaches surveyed in this study. Sand dunes are thought to provide refuge for ghost crabs especially during extreme weather events (Lucrezi et al., 2010) and numerous studies have found highest burrow densities occurring at or just below the dune-beach interface (Barros, 2001; Lucrezi et al., 2009b; Schlacher et al., 2011). Unfortunately, with only one beach with intact sand dunes in Sydney Harbour, I are unable to provide any formal test of the importance of dunes in the ecology of the urban ghost crabs.

Chapter 2

Sandy beaches of Sydney Harbour display high environmental heterogeneity, depending on their location in the estuary, with most being low-energy beaches that are only intermittently reworked by storm events (Kennedy, 2010). Despite this variability, I detected no relationships between ghost crab burrow densities and beach slope, grain size or beach length alone (although beach length in combination with cleaning regime occurred in our best predictive model). Interestingly smaller beaches in Sydney Harbour in general accumulated larger volumes of wrack. This could be attributed to the fact that they were often not mechanically cleaned, or that smaller, wave protected beaches often accumulate larger amounts of wrack (Barreiro et al., 2011). These results are consistent with Quijón et al. (2001) who found the distribution of ghost crabs was unrelated to beach morphodynamics, but contrast with Lucrezi (2015) who showed that burrow counts increased with increasing beach slope and deceased with grain size, and Noriega et al. (2012) who also found slope to be positively correlated with burrow density but no relationship with grain size.

Conclusions

Here, I show that ghost crab populations in Sydney Harbour may be more resilient than originally thought. I found highest burrow densities on those beaches that were still mechanically raked up to three times per week, suggesting ghost crab populations are somewhat robust to moderate human disturbance. These findings give hope to the persistence of similar populations in increasingly modified environments. Our study highlights the need for ecologically informed management strategies that ensure the continued co-existence of humans and nature in highly urbanised environments.

In our study area, Mosman Municipal Council has already recognised the need for such strategies. In 2004, they made the decision to actively manage one of their ‘low-use’ beaches (Chinamans Beach) for the promotion of biodiversity, and that beach supported the highest burrow densities of all Sydney Harbour beaches in our survey. In order to accomplish this, they ceased mechanical cleaning at this beach and now only remove human waste by hand as needed. They provide signage and information at Chinamans beach to educate the public on the importance of wrack to sandy beach ecosystems. The use of only manual hand cleaning, however, is unlikely to be economical for many coastal managers around the world, and a combined beach cleaning approach that incorporates some manual hand cleaning and limits the frequency and number of mechanically cleaned beaches would be desirable to conserve ghost crab populations (Noriega et al., 2012). By implementing these mixed approaches I can allow coastal managers to uphold the important social and economic value of beaches, but not to the detriment of their ecological value.

Chapter 2

Chapter 3

Scavenger responses to enhanced trophic subsidies on a sandy beach

Abstract

Sandy beaches form long, permeable borders between marine and terrestrial zones over which resources are exchanged. Organic marine matter washes onshore providing a critical dietary input not only for sandy beach consumers but also for adjacent terrestrial food webs. This study tested the short term response of sandy beach scavengers to variations in food supply, by experimentally manipulating the availability of two common type of organic subsidy found washed up on beaches, wrack and carrion. Despite ghost crabs being the dominant invertebrate scavengers on sandy beaches they did not demonstrate a short-term shift in burrow distribution in response to either subsidy. Vertebrate scavengers did demonstrate a rapid response to carrion subsidies only. Vertebrate scavengers were also found to modify the behaviour of ghost crabs, with lower numbers of new burrows constructed in plots where vertebrates had been detected. The results of this study suggests behavioural interactions between scavengers are likely, and highlight the need to better understand how alterations to the natural supply of organic inputs to sandy beaches will affect food web dynamics in these ecosystems.

Chapter 3

Introduction

Sandy beaches form a long and highly permeable border between the terrestrial and marine zones (Polis et al., 1997). Their dynamic nature permits the transfer of organic marine matter onshore via wind, waves and currents (Orr et al., 2005). This material often forms distinct accumulations which play important functional and structural roles, including the provision of food, habitat and nutrients to local ecosystems (Ince et al., 2007).

Sandy beaches often lack attached larger plants and consequently in situ primary production is low (Schlacher et al., 2013a), with the exception of the beaches that support high accumulations of surf diatoms (Campbell, 1996). As a result washed up organic marine matter constitutes a critical dietary input for sandy beach consumers, and an important trophic subsidy for marine food webs and even local terrestrial food webs (Polis et al., 1997).

The deposits of organic marine matter on sandy beaches consist primarily of detached marine plants and algae (known as wrack) and animal carcasses (carrion) (Orr et al., 2005). Once deposited onshore this material undergoes fragmentation, decomposition and remineralisation by bacteria, mieofauna and grazers (Colombini et al., 2003). It enters food webs either directly by detritivores which consume the wrack, or indirectly by increasing prey availability (Hagen et al., 2012). Wrack- and carrion-associated species thus occupy a range of trophic levels, including consumers of plant and algal material, scavengers of carrion, and predators (Schlacher et al., 2013b).

Despite the importance of washed up marine matter for sandy beach communities, strandline accumulation are often thought to reduce the aesthetic and recreational quality of these environments. In response to this, beaches in urban areas are often subject to regular and frequent mechanical beach cleaning to remove these shoreline accumulations. This process commonly uses heavy machinery to drag a rake or sieve across the surface of the sand, thereby removing all debris from the surface layer of the beach. The process of mechanical beach cleaning, however, can impact sandy beach fauna in two ways; either through a physical disturbance of the environment (e.g. compaction of sand or destruction of burrows) or by the removal of a potential food supplies. The effects on food supply on sandy beach organisms could be direct (e.g., for consumers of the washed up wrack or carrion) or indirect by reductions in the abundance of prey (e.g., for predators of animals associated with the wrack or carrion). Larger piles of wrack can provide visual cues for shorebirds to assist with foraging, whilst also provide habitat structure and refuges for macroinvertebrates from predation and exposure (Dugan et al., 2003).

Chapter 3

Ghost crabs (Crustacea: Brachyura: Ocypodidae) are abundant on many sandy beaches (Schlacher et al., 2013b), playing the important ecological role of being apex invertebrate predators, while also being important prey for many higher order vertebrate consumers (Lucrezi and Schlacher, 2014). Their diet encompasses a wide range of live animal prey and dead carcasses, which can include both animal tissues and plant material (Lucrezi and Schlacher 2014).

Most predators of ghost crabs have a predominantly terrestrial affinity (e.g. raptors and foxes), using ocean beaches and coastal dunes as foraging sites to various degrees. Thus, predation on ghost crabs may provide an important functional pathway for the trophic coupling of marine and terrestrial ecosystems where predators forage across habitat boundaries (Lucrezi and Schlacher, 2014).

In Chapter 2, I showed that ghost crab abundances in an urbanised estuary (Sydney Harbour) were lowest on the beaches with a high frequency of mechanical beach cleaning (Fig. 2.2a.) (Stelling-Wood et al., 2016). This could result from physical disturbance to the sandy beach habitat or from changes to the availability of trophic subsidies. Previous studies with ghost crabs have shown that they respond to variations in the supply of carrion (Schlacher et al., 2013a; Schlacher et al., 2013b). The carrion supply to sandy beaches however is sporadic, providing an unpredictable, pulsed resource for these environments (Cortés‐Avizanda et al., 2009; Roth, 2003). In comparison, the animals associated with wrack may represent a comparatively more consistent food source (Dugan et al., 2003). To test how ghost crabs respond to variations in the supply of both wrack and carrion to sandy beach ecosystems I experimentally manipulated the availability of these two common types of organic subsidy. I then measured the short-term response of the key scavenger, ghost crabs (O.cordimanus) to these subsidies. Ghost crabs construct burrows which are clearly visible on the surface and known to correlate with body size (Lucrezi et al., 2009; Turra et al., 2005). This allows for the rapid assessment of crab abundance and body size by counting and measuring burrow openings on the beach surface.

I asked the following specific questions: 1.) Does the arrival of algal wrack increase the density of burrows adjacent to subsidies. This would be expected if the presence of wrack is a visual or olfactory cue for crabs to begin foraging on animals associated with the wrack. 2.) Does the arrival of carrion increase the density of crab burrows independently of wrack arrival? This would be expected if crabs responded to visual or olfactory cues of animal food only, and were not using wrack to identify foraging sites. Washed up macroalgae are commonly employed as a visual cue for foraging, as its large size and dark colouring making it easy to distinguish against

Chapter 3

the sand (Pelletier et al., 2011). Carrion arrives on beaches in various shapes and sizes, some highly identifiable along the shoreline (e.g. large whale carcass) whilst others are less so (e.g. fish scraps from fishing refuse or small, dead invertebrates). Carrion thus presents a variable visual cue but does produce strong odours, especially as it decays, providing a strong olfactory cue for feeding on sandy beaches (McLachlan and Brown, 2006).

Materials and methods

Study site and species

The experiments were conducted at Chinamans beach within Sydney Harbour, on the eastern coast of New South Wales, Australia (Fig. 2.1, 33° 51’ S, 151° 14’ E). It is a relatively sheltered beach, with fine grained sand (225.4µ) and a moderate beach slope (6.1ᵒ) (Stelling-Wood et al., 2016). Chinamans beach is the only beach within Sydney Harbour that still has remnants of a natural sand dune system and is backed by native vegetation. In 2004 the local managing council (Mosman Council) ceased mechanically cleaning at Chinamans beach and now preserve the beach for its biodiversity values. In my survey of harbour beaches (Chapter 2), Chinamans beach was found to support the largest populations of ghost crabs, being the only site with burrow densities over 0.9 burrows per square metre (Fig. 2.1, Stelling-Wood et al., 2016).

Two species of ghost crab are known to occur on the east coast of Australia. The Smooth- handed Ghost Crab Ocypode cordimanus Latreille, 1818 is found in the Indo-West Pacific and, in Australia, along the northern and eastern coasts, from the north of Western Australia to southern New South Wales (Jones and Morgan, 2002). Ocypode ceratophthalma (Pallas 1772) also occurs along the east coast of Australia as far south as Sydney (Barros, 2001), but only O.cordimanus is known to occur within Sydney Harbour (Hutchings et al., 2013). Burrow counts are therefore reported as those of the Smooth-handed Ghost Crab Ocypode cordimanus.

Feeding experiment

To test the response of scavengers to trophic subsidies on beaches, I manipulated food supply over 10 nights during December (austral summer) 2015 and measured the numerical response of the key scavenger, ghost crabs. The experiment consisted to two trophic subsidies in a factorial design: a wrack subsidy (macroalgae present or absent) and a carrion subsidy (prawn meat present or absent). This resulted in four different treatments: 1) control (no subsidy present); 2) wrack (macroalgae only present); 3) carrion (prawn meat only present) and 4) wrack and carrion (macroalgae and prawn meat present).

Chapter 3

The macroalgae added in ‘wrack’ treatments was the brown algae Sargassum vestitum (Phaeophyceae: Fucales) as it is commonly found washed up on beaches around Sydney (personal ob.). S.vestitum was collected locally earlier in the day, washed in fresh water to remove epifauna and allowed to air-dry for 3-4 hours. Existing epifauana was removed from the collected wrack to standardise initial conditions and ensure that the only source of animal carrion was the material added in the prawn treatment. For wrack treatments, roughly similar sized individuals (~30cm long) were placed in the centre of each plot and secured with tent pegs. The prawns used in ‘prawn’ treatments were school prawns purchased frozen from a local bait shop and thawed on the day of the experiment. For prawn treatments, five prawns of roughly equal size were cut in half (10 pieces of prawn) and were either placed in a pile in the centre of the plot (‘prawn only’ treatments) or were spread within the wrack (‘wrack + prawn’ treatments). Once treatment plots had been mapped and treatments were in place, a circular area (60cm diameter) surrounding the centre was raked flat and lightly sprayed with water for track identification the next morning.

Each experimental plot measures 2m (wide) x 3m (long), with the longest edge perpendicular to the shoreline, positioned at the landward boundary of the beach. On each of the 10 nights, the experiment consisted of two replicates of each combination of treatments (i.e. 8 plots per night) dispersed over a 230m long stretch of beach. The allocation of treatments to plots and the distance between plots was randomly assigned using random number tables. Distance between plots was restricted to a minimum of 10m. Once treatment plots had been mapped and treatments were in place, a circular area (60 cm in diameter) surrounding the centre was raked flat and lightly sprayed with water for track identification the next morning.

Ghost crabs are known to be primarily nocturnal (Lucrezi and Schlacher, 2014), so treatments were added at sunset to each plot and were checked and removed the following morning at sunrise. Consequently, the number of burrows present in plots was counted: 1.) before the treatment (i.e. initially present in plot), and 2.) after the treatment (i.e. present in plot at sunrise the morning after). At each time burrows were counted, their diameters measured and their locations roughly mapped (to enable identification of new burrows formed overnight). In the morning census, I also measured the presence of vertebrate scavengers or their tracks. Each experiment only lasted from sunset to sunrise the following day as Chinamans Beach is a popular beach in close proximity to Sydney’s CBD, surrounded by residential properties and longer exposures would thus have introduced a greater possibility of human disturbance. Experiments were never undertaken on consecutive nights.

Data analyses

Chapter 3

I used a generalised linear mixed effect models (GLMM) to test for relationships between burrow counts and the two treatments: wrack subsidy and prawn subsidy. Burrow count was the response variable with predictor variables including wrack, carrion and time (before or after the treatment), as well as two-way interactions between all these variables. With the before and after data coming from repeated measures of a single plot (and thus not being independent), I used a random factor of plot (a unique identifier of date, replicate number and treatment). With count data being analysed, I used a Poisson error distribution. The effects of either treatment are tested by the interactions between time (‘before’ and ‘after’) and wrack, and time and carrion. To test for possible changes in the size of crabs using plots, I conducted a similar analysis with burrow diameter as the response variable.

Throughout the study three species of avian scavengers were seen feeding along the strandline at Chinamas beach. These were the Australian Magpie (Cracticus tibicen), Silver Gull (Chroicocephalus novaehollandiae) and Masked Lapwing (Vanellus miles). Rodent tracks were also observed in several treatment plots (Fig. 3.1), with two species of rodent known to occur in the area, the native bogul rat (Rattus fuscipes) and the introduced black rat (Rattus rattus). To test the response of vertebrate scavenger to treatments, and possible effects on ghost crabs, I ran two further analyses. Firstly, I contrasted the occurrence of vertebrate scavengers among treatments using a generalised linear model (GLM). Vertebrate scavenger occurrence was the response variable and the treatments (wrack and carrion subsidies) and their interactions were fixed, predictor variables. I used a binominal error distribution as the vertebrate occurrence was scored as presence or absence. The presence of scavengers in treatment plots was determined from footprints and track markings making identification difficult, as such all vertebrate scavengers were grouped together as ‘vertebrate scavengers’ for analyses.

a) b)

Chapter 3

Fig. 3.1. Vertebrate scavenger tracks adjacent to experimental plots. a) rodent tracks and b) avian tracks. Photo: T.Stelling-Wood.

Secondly, I contrasted the number of new burrows occurring in plots between those plots where there was evidence of vertebrate scavengers, and those with no evidence of vertebrate visitation. For this, I used a GLM with number of new burrows as the response variable and the presence/absence of vertebrates as the predictor variable with a Poisson error distribution. Of the 80 experimental plots undertaken during the study period 78 were still intact the following morning, with only two ‘prawn only’ plot being disturbed by human and consequently removed from all analyses.

All analyses were performed using the lme4 package (Bates et al., 2015) in Rv3.1.0 (R Core Team, 2014). Residual plots were inspected for model validation (Zuur et al., 2009).

Results

New crab burrows were found to have been constructed overnight in 45% of the plots. Overall, there was no effect of either trophic subsidy on the count of burrows in experimental plots after one night (Fig. 3.2).

The GLMM using burrow count as the response variable found that neither interactions between time (before and after the treatment) and either subsidy were significant (Wrack: z = 0.148, p = 0.7; carrion: z = 1.806, p = 0.179), suggesting the number of new burrows formed overnight was not significantly altered by treatments. Similarly, there were no significant interactions between time and either subsidy for burrow diameter (Wrack: z = 0.615, p = 0.433; carrion: z = 0.184, p = 0.668).

Vertebrate scavengers were detected in 21.8% of our experimental plots. Of the 17 instances where vertebrate scavengers were detected six were birds and 11 were rodents. Rodents and birds were never detected in the same plots. Vertebrate scavengers were detected in all plots where prawn piece removal was 100%. The occurrence of vertebrates in treatment plots was found to be significantly higher in treatment plot containing carrion subsidies (Fig. 3.3a, z = 1.587, p = 0.018), compared with plot lacking carrion (Fig. 3.3a, z = 0.00, p = 0.979). The presence of vertebrates in treatment plots was found to be significantly associated with the number of new burrows constructed overnight, with fewer new burrows found in plots where vertebrate tracks were found (Fig. 3.3b, z = -2.87, p = 0.004).

Chapter 3

Fig. 3.2. Number of new ghost crab burrows formed per plot overnight in response to wrack and carrion subsidies. Data are means ± standard error.

Chapter 3

Fig. 3.3. (a) Vertebrate scavenger response to subsidies. The metric of vertebrate response was the presence of tracks on treatment plots. (b) Numerical response of ghost crabs to the presence of vertebrate scavengers within treatment plots. The metric of ghost crab response was the number of new burrows formed overnight in a treatment plot.

Discussion

This study found that despite ghost crabs being the dominant scavengers on sandy beaches they did not demonstrate a short-term shift in burrow density in response to enhanced trophic subsidies. Instead, higher order vertebrate scavengers were found to respond quickly to carrion. Behavioural interactions among scavengers seem likely, with the presence of vertebrates being associated with lower numbers of new burrows constructed in plots where vertebrates had been detected.

Ghost crab response to trophic subsidies

Sandy beaches are oligotrophic environments, relying almost entirely of the input of allochthonous organic material from the marine environment (Schlacher et al., 2013a). The delivery of this material is highly variable in nature and as such the colonisation, processing and associated consumption of this material is expected to be rapid (Colombini et al., 2003; Lastra et al., 2008). The weak response of ghost crabs to enhanced trophic subsidies found in this study was unexpected and contrasts with many studies from the literature which investigate wrack and sandy beach fauna relationships. For example, Ince et al. (2007) found sandy beach fauna and infauna were consistently abundant on high-wrack beaches, but were either absent or only present in low numbers on low wrack beaches. Similarly, Dugan et al. (2003) found two shorebird species, plovers that use visual cues of foraging, were positively correlated with the standing crop of wrack and with the abundance of wrack-associated fauna. Spiller et al. (2010) also found the influence of beach cast wrack to penetrate up into terrestrial food webs, finding lizards and herbivorous arthropod numbers to be higher in plots that had been subsidies with seaweed. All these studies however, investigated macrofauna response over much longer time- scales then was tested in this study, ~2 weeks to 3 months in other studies vs. over-night in our study. With a short term study, I was testing for rapid shifts in foraging behaviour only, and the possibility remains that these trophic subsidies will impact crab populations on longer time scales.

Ghost crabs have some of the most highly evolved eyes among crustaceans, but at best can visually detect only large objects, instead relying on an acute sense of taste and smell (Lucrezi

Chapter 3

and Schlacher, 2014). They use chemosensory hairs to detect odours from potential food, reacting to olfactory cues from both animal and plant material (Rittschof, 1992; Wellins et al., 1989). The weak response of ghost crabs to wrack treatments was not unexpected, given that we employed wrack primarily as a visual feeding cue. The lack of response to carrion however was unexpected, given that ghost crabs are known to respond to odours that signal food sources (Wellins et al., 1989). Two recent studies, Schlacher et al. (2013a) and Schlacher et al. (2013b), demonstrated ghost crabs to show a strong positive response to carrion augmentation on an ocean exposed sandy shore on North Stradbroke Island, in eastern Australia. Both these studies however repeatedly added carrion to plots over an extended period of time (6-8 days), with Schlacher et al. (2013a) only reporting a significant difference between control and treatment plots after day 4. In comparison our experiments only ran overnight (approx. 9 hours), suggesting that ghost crabs may not have been able to locate the enhanced food sources in the short amount of time permitted in our experiment. Alternatively ghost crabs are known to forage variable distances and may have fed in plots with trophic subsidies, but not altered the distribution of their burrows (as would be required for me to detect a treatment effect in these experiments), instead simply returning to existing burrows. I did, however, find new burrows in almost half of the plots, indicating that short term changes to burrow positions were occurring.

Ghost crabs are known to be highly opportunistic feeders, demonstrating strong trophic plasticity. This allows them to be highly adaptive with feeding and may have allowed them continue feeding on alternate food sources. Morrow et al. (2014) demonstrated this using staple isotopic analysis to identify the dominant prey items of ghost crabs Ocypode quadrata. They found distinctly different patterns of prey utilisation between sites, whereby ghost crabs consumed mainly wrack-associated amphipods in areas where swash zone prey was only present in low numbers, but fed primarily on swash zone animals in areas where they were highly abundant. Interestingly the beach at which this experiment was undertaken (Chinamans Beach) is known, at times, to support very large numbers of the wrack-associated amphipod Exoediceroides maculosus (Jones et al 1991).

The sub-littoral macroalgae in temperate Australia is dominated by the kelp Ecklonia radiata (Andrew & Jones 1990) and several species in the genus Sargassum (Order Fucales), including S. vestitum (Steinberg, 1994). Despite it commonly occurring on Sydney beaches, ghost crabs did not respond to S.vestitum, which was used as a wrack subsidy in my experiment, suggesting they may have been able to exploit other resources naturally present on the beach and as such were unaffected by our treatments.

Chapter 3

Vertebrate scavengers

Sandy beaches form a barrier between marine and terrestrial zones, whereby resources can be exchanged (Polis et al., 1997). While ghost crabs are the apex invertebrate predator on many sandy beaches, the abutting terrestrial systems host an array of vertebrate scavengers that use the beach to exploit marine food sources (Polis et al., 1997). I detected vertebrates (via tracks and footprints) in nearly 22% of the treatment plots, indicating that ghost crabs were not the only active scavengers on the study beach. Rodent tracks were most commonly observed, occurring in 65% of plots where vertebrates were detected, while avian tracks were detected in 35% of plots with vertebrate tracks were present. These results are in line with Huijbers et al. (2013) who found that scavenger guilds on urban beaches were dominated by mammals that were mainly active at night. Chinamans beach is located in Sydney Harbour, one of Australia’s most urbanised estuaries. While the beach is maintained for biodiversity by its managing council and as such experiences lower pressure from human activities than other beaches located nearby (Stelling-Wood et al., 2016), Sydney Harbour itself is located within a large urban centre in close proximity to Sydney’s CBD. Therefore it is not surprising then that the scavenger guild found here is similar to those found on other urban beaches.

Vertebrate activity was detected more frequently in plots with carrion subsidies and unrelated to the addition of wrack subsidies. Similar aggregative responses of vertebrate scavengers to carrion have been reported in the literature (Huijbers et al., 2013; Schlacher et al., 2013a; Schlacher at el., 2013b). Cross-habitat boundary foraging behaviour is not unusual in coastal areas, for example, Rose and Polis (1998) observed carnivorous mammals foraging in the intertidal and along the shoreline in coastal areas along the Sea of Cortez and the Pacific Ocean. They also reported coyote populations in coastal areas were larger then at sites further inland, with individuals from coastal populations found to eat a greater variety and larger quantities of food. Similarly enhanced trophic subsidies have been demonstrated to have population-wide benefits for the spider Zygiella x-notata, where the addition of marine macrophytes were found to positively impact population size and fecundity by increasing prey availability. While the duration of this experiment did not allow us to measure the vertebrate population responses to our treatments, it does not seem unreasonable that these scavenger populations could be benefiting from marine trophic subsidies. The effect of carrion, but not wrack subsidies, on vertebrate foraging suggests they are responding primarily to olfactory cues rather than using the presence of wrack as a visual cue to investigate wrack for possible animal resources associated with the wrack. This is not surprising given that rodents were the dominant vertebrate scavenger detected in plots, a group known to rely heavily on olfactory cues. Longer

Chapter 3

term experiments, where wrack is allowed the time to accumulate invertebrate decomposers, are needed to better understand how wrack availability (and its removal at other urban beaches) could affect vertebrate consumers on these beaches.

Although ghost crabs showed no response to my treatments, they were found to build fewer new burrows in plots where vertebrate tracks were present. Schlacher et al. (2013a) and Schlacher et al. (2013b) found a similar response, with ghost crabs modifying their fossorial behaviour in the presence of vertebrates, by reducing burrowing activity. Schlacher et al. (2013a) suggest several reasons for this relationship; 1.) competition for carrion, resulting in reduced food availability as carrion is removed rapidly and completely by the competitively superior vertebrates, before crabs have the chance to locate it, 2.) reduced foraging activity of ghost crabs in the presence of vertebrates, due to higher risk of predation, and/or 3.) predation of crabs that had entered carrion treatment plots. This behavioural shift to reduced burrowing activity in the presence of predators could be the reason no relationship was found between our treatments and ghost crabs burrow density. Morrow et al. (2014) reported similar behaviour modification, finding 50% of ghost crab burrows were located amongst the wrack line at a site with high racoon activity, suggesting that ghost crabs reduced their movement in response to higher risk of predation. If ghost crabs reduce foraging distance in the presence of predators, this could make them less likely to build new burrows, instead favouring returning to their original burrow and could explain the weak relationship we observed between our treatments and burrow density.

Conclusion

This experiment demonstrated the dynamic nature of sandy beaches, as permeable boundaries between the terrestrial and marine zones. While there was no evidence for invertebrate scavengers responding on short time scales to enhanced trophic subsidies, I did show that vertebrate scavengers responded quickly to carrion subsidies and that these vertebrates were likely modifying the behaviour of ghost crabs. All scavengers examined in this study are highly mobile, facilitating the transfer of marine matter across sandy beaches, and linking ocean productivity and terrestrial food webs (Schlacher et al., 2013b). Given that these systems often rely heavily on this resource exchange the intensification of beach cleaning practices, in response to urbanisation, could have substantial ecological effects on these systems by removing washed up organic marine matter (Dugan et al., 2003). As human populations in coastal areas continue to grow, and sandy beaches continue to bear the brunt of associated habitat modification, it is important to consider how alterations to the natural supply of organic marine matter can have far reaching implications for both marine and terrestrial ecosystems.

Chapter 3

Chapter 4

General discussion

This thesis explored how urbanisation is impacting sandy beach ecosystems, using ghost crabs as indicators of ecosystem health and functioning. This link is well established for open ocean beaches, but until now had not been investigated on urban estuarine beaches. Physical measurements, beach management practices and forms of habitat modification were assessed to quantify the relationship between these and ghost crab distributions on beaches. A better understanding of these relationships can assist in the development of more ecologically sound management practices.

Habitat modification of sandy beaches

Human-induced habitat modification is a direct consequence of urbanisation, and comes in many forms, many of which are destructive to natural ecosystems. On sandy beaches, habitat modification by humans range from coastal development and armouring, to beach grooming practices and nourishment. Documenting the biotic responses to these physical modifications of the environment is a critical step in predicting the consequences of human-induced change for beach ecosystems (Defeo et al., 2009). Measuring the densities of ghost crab burrows provided a way to assess urban beaches for the impacts of habitat modification. Examining variables that relate to physical beach properties, management practices and habitat modification provided a means of testing which aspects of habitat modification were the most likely mechanistic links to altered burrow densities.

In contrast to many recent studies, ghost crabs in Sydney Harbour did not appear to be as vulnerable to habitat modification as had been previously suggested (Barros, 2001; Lucrezi et al., 2009; Noriega et al., 2012). Neither level of coastal development or armouring was found to correlate with ghost crab distributions, and burrow densities were only significantly reduced on beaches that were mechanically cleaned more than five times per week. Mechanical beach cleaning can impact sandy beach communities in two ways; 1.) it is a physical disturbance, thereby altering habitat structure and complexity (Brown and McLachlan, 1990), and 2.) it removes a critical food source by raking up accumulations of organic marine matter along the strandline (wrack) (Dugan et al., 2003). Few studies have looked directly at the impacts of mechanical beach cleaning on ghost crab populations and none have explicitly tested the reason behind these negative impacts (Schlacher et al., 2016). In this study, frequent mechanical

Chapter 4 cleaning was associated with lower volumes of wrack accumulating on beaches, suggesting that the process can impact food availability on beaches, however no correlation was found between burrow density and wrack volume. This suggests that either ghost crab populations in Sydney Harbour are not reliant on animals associated with wrack as a food source, or more likely that wrack accumulation was too variable in time to be accurately sampled in my survey. Chapter 3 of this thesis investigated the possible relationship between wrack and ghost crabs, by testing the response of ghost crabs to enhanced organic subsidies.

Chapter 2 fills a gap in the literature, providing key insights into how human induced habitat modification impacts sandy beach communities. Ghost crabs were found to possess some resilience to habitat modification, giving hope to the persistence of similar populations in increasingly urbanised environments.

A useful next step would be to develop ‘dose-response’ style studies to develop better practices for beach cleaning processes that balance reduced ecological impacts with aesthetics. As the number and magnitude of global environmental change increases over time, scaling up the findings of sandy beach research, such as the research undertaken in this thesis, to larger, and longer scales will become more important (Defeo et al., 2009).

Trophic subsidies to sandy beaches

Sandy beaches lack major sources of primary production, making these systems heavily reliant on washed up organic material to fuel local food webs (Polis et al., 1997). Organisms associated with this organic material occupy a range of different trophic levels, including consumers, scavengers and predators (Schlacher et al., 2013b). An experiment that enhanced food supply (Chapter 3) provided an empirical means of testing the importance of allochthonous marine subsidies to sandy beach scavengers. Furthermore, the field experiment employed here examined the scavenger response to different food types, testing the relative importance of visual versus olfactory cues when foraging. Ghost crabs did not demonstrate a rapid response to enhanced subsidies (as measured by shifts in burrow densities), but in line with recent studies, vertebrate scavengers quickly found subsidies that included carrion. The lower number of new crab burrows in plots that had evidence of vertebrate visitation suggests behavioural interactions between invertebrate and vertebrate scavengers on sandy beaches. Population responses to enhanced trophic subsidies was beyond the scope of this study, however population wide benefits from organic marine subsidies is not unreasonable, and has been demonstrated in other coastal vertebrate populations (Rose and Polis, 1998). Future longer- term studies with the capacity to investigate potential population-wide impacts are important

Chapter 4 as human activities continue to alter the availability of organic inputs on sandy beaches through processes such as mechanical beach cleaning.

Understanding the mechanisms of resource usage and exchange on sandy beaches is important, given that these environments are extremely vulnerable to threats from human activities on the terrestrial side (habitat modification) and the global climate change on the oceanic side. The consequence of changes to fundamental trophic processes in coastal food-web structure could be profound, and should be taken into consideration by coastal managers (Huijbers et al., 2013). Furthermore understanding of how these processes vary with different levels of human activities and associated habitat modification is important as a way to develop better and more ecologically sustainable coastal management practices. Future management policies for beaches need to integrate the interacting natural, social and economic values of beaches to protect biodiversity and maintain important ecological function in these environments. To this end, scientists, managers, policy makers and the public will have to work together to devise and implement appropriate and effective coastal management policies.

References

Aheto, D. W., Asare, C., Mensah, E., Aggrey-Fynn, J., 2011. Rapid assessment of anthropogenic impacts of exposed sandy beaches in Ghana using ghost crabs (Ocypode spp.) as ecological indicators. Momona Ethiopian Journal of Science 3, 93-1033.

Andrew, N.L. and Jones, G.P., 1990. Patch formation by herbivorous fish in a temperate Australian kelp forest. Oecologia, 85,.57-68.

Australian Bureau of Statistics, 2013. Population projections, Australia, 2012 (base) to 2101.

Barreiro, F., Gómez, M., Lastra, M., López, J., De la Huz, R., 2011. Annual cycle of wrack supply to sandy beaches: effect of the physical environment. Marine Ecology Progress Series 433, 65– 74.

Barros, F., 2001. Ghost crabs as a tool for rapid assessment of human impacts on exposed sandy beaches. Biological Conservation 97, 399-404.

Bascom, W.N., 1951. The relationship between sand size and beach‐face slope. Eos, Transactions American Geophysical Union 32, 866-874.

Bates, D., Mächler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67, 1-488.

Blott, S.J., Pye, K., 2001. GRADISTAT: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surface Processes and Landforms 26, 1237-1248.

Breiman, L., 2001. Random forests. Machine Learning 45, 5-32.

Brown, A., McLachlan, A., 1990. Ecology of sandy shores. Elsevier, Amsterdam.

Brown, A., McLachlan, A., 2002. Sandy shore ecosystems and the threats facing them: some predictions for the year 2025. Environmental Conservation 29, 62-77.

Burnham, K.P., Anderson, D.R., 2004. Multimodel inference understanding AIC and BIC in model selection. Sociological Methods & Research 33, 261-304.

Campbell, E. E., 1996. The global distribution of surf diatom accumulations.Revista Chilena de Historia Natural 69, 495-501.

Chapman, M., Bulleri, F., 2003. Intertidal seawalls—new features of landscape in intertidal environments. Landscape and Urban Planning 62, 159-172.

Colombini, I., Chelazzi, L., Gibson, R.N. and Atkinson, R.J.A., 2003. Influence of marine allochthonous input on sandy beach communities. Oceanography and Marine Biology: an Annual Review 41, 115-159.

Cortés‐Avizanda, A., Carrete, M., Serrano, D. and Donázar, J.A., 2009. Carcasses increase the probability of predation of ground‐nesting birds: a caveat regarding the conservation value of vulture restaurants. Animal Conservation 12, 85-88.

Defeo, O., McLachlan, A., 2005. Patterns, processes and regulatory mechanisms in sandy beach macrofauna: a multi-scale analysis. Marine Ecology Progress Series 295, 1-20.

Defeo, O., McLachlan, A., Schoeman, D.S., Schlacher, T.A., Dugan, J., Jones, A., Lastra, M., Scapini, F., 2009. Threats to sandy beach ecosystems: a review. Estuarine, Coastal and Shelf Science 81, 1- 12.

Delgado, I., Lloyd, G., 2004. A simple low cost method for one person beach profiling. Journal of Coastal Research, 1246-1252.

Dexter, D.M., 1983. Community structure of intertidal sandy beaches in New South Wales, Australia, Sandy beaches as ecosystems. Springer, pp. 461-472.

Dexter, D.M., 1984. A guide to sandy beach fauna in New South Wales. Wetlands (Australia) 3, 94-104.

Dugan, J.E., Hubbard, D.M., 2006. Ecological responses to coastal armoring on exposed sandy beaches. Shore & Beach 74, 10-16.

Dugan, J.E., Hubbard, D.M., 2010. Loss of coastal strand habitat in southern California: the role of beach grooming. Estuaries and Coasts 33, 67-77.

Dugan, J.E., Hubbard, D.M., McCrary, M.D., Pierson, M.O., 2003. The response of macrofauna communities and shorebirds to macrophyte wrack subsidies on exposed sandy beaches of southern California. Estuarine, Coastal and Shelf Science 58, 25-40.

Dugan, J. E., Hubbard, D. M., Rodil, I. F., Revell, D. L., Schroeter, S., 2008. Ecological effects of coastal armoring on sandy beaches. Marine Ecology, 29(s1), 160-170.

Duong, H.S., 2008. Investigating the ecological implications of wrack removal on South Australian sandy beaches. PhD Thesis, School of Biological Sciences, Flinders University.

Gilburn, A.S., 2012. Mechanical grooming and beach award status are associated with low strandline biodiversity in Scotland. Estuarine, Coastal and Shelf Science 107, 81-88.

González, S.A., Yáñez-Navea, K., Muñoz, M., 2014. Effect of coastal urbanization on sandy beach coleoptera Phaleria maculata (Kulzer, 1959) in northern Chile. Marine Pollution Bulletin.

Hagen, E.M., McCluney, K.E., Wyant, K.A., Soykan, C.U., Keller, A.C., Luttermoser, K.C., Holmes, E.J., Moore, J.C. and Sabo, J.L. 2012. A meta‐analysis of the effects of detritus on primary producers and consumers in marine, freshwater, and terrestrial ecosystems. Oikos 121, 1507-1515.

Hubbard, D. M., Dugan, J. E., Schooler, N. K., Viola, S. M., 2014. Local extirpations and regional declines of endemic upper beach invertebrates in southern California. Estuarine, Coastal and Shelf Science 150, 67-75.

Huijbers, C.M., Schlacher, T.A., Schoeman, D.S., Weston, M.A. and Connolly, R.M. 2013. Urbanisation alters processing of marine carrion on sandy beaches. Landscape and Urban Planning 119, 1-8.

Hutchings, P.A., Ahyong, S.T., Ashcroft, M.B., McGrouther, M.A., Reid, A.L. 2013. Sydney Harbour: its diverse biodiversity. Australian Zoologist 36, 255-320.

Ince, R., Hyndes, G.A., Lavery, P.S., Vanderklift, M.A., 2007. Marine macrophytes directly enhance abundances of sandy beach fauna through provision of food and habitat. Estuarine, Coastal and Shelf Science 74, 77-86.

James, R.J. 2000. From beaches to beach environments: linking the ecology, human-use and management of beaches in Australia. Ocean & Coastal Management 43, 495-514.

Johnston, E. L., Mayer-Pinto, M., Hutchings, P. A., Marzinelli, E. M., Ahyong, S. T., Birch, G., Booth, D. G., Creese, R. G., Doblin, M. A., Figureia, W., Gribben, P. G., Pritchard, T, Roughan, M., Steinberg, P. F. Hedge, L. H. 2015. Sydney Harbour: what we do and do not know about a highly diverse estuary. Marine and Freshwater Research 66, 1073-1087.

Jonah, F. E., Agbo, N. W., Agbeti, W., Adjei-Boateng, D., Shimba, M. J., 2015. The ecological effects of beach sand mining in Ghana using ghost crabs (Ocypode species) as biological indicators. Ocean & Coastal Management, 112, 18-24.

Jones, A., 2003. Ecological recovery of amphipods on sandy beaches following oil pollution: an interim. Assessment Journal of Coastal Research 19, 66-73.

Jones, D.S., Morgan, G.J., 2002. Field guide to crustaceans of Australian waters. Reed New Holland, Sydney.

Jones, A.R., Murray, A. and Marsh, R.E., 1991. Patterns of abundance of exoedicerotid amphipods on sandy beaches near Sydney, Australia. In VIIth International Colloquium on Amphipoda. Springer Netherlands, 119-126.

Keats, M., 1997. A cautionary tale: a study of the macro bivalve and gastropod molluscan fauna of Spectacle Island, Sydney Harbour. Wetlands (Australia) 16, 72-82.

Kennedy, D. M. 2002. Estuarine beach morphology in microtidal Middle Harbour, Sydney. Australian Geographical Studies 2, 231-240.

Lastra, M., Page, H.M., Dugan, J.E., Hubbard, D.M. and Rodil, I.F. 2008. Processing of allochthonous macrophyte subsidies by sandy beach consumers: estimates of feeding rates and impacts on food resources.Marine Biology 154, 163-174.

Lucrezi, S., 2015. Ghost crab populations respond to changing morphodynamic and habitat properties on sandy beaches. Acta Oecologica 62, 18-31.

Lucrezi, S., Saayman, M., van der Merwe, P., 2014. Impact of off-road vehicles (ORVs) on ghost crabs of sandy beaches with traffic restrictions: A case study of Sodwana Bay, South Africa. Environmental Management 53, 520-533.

Lucrezi, S., Schlacher, T.A., Robinson, W., 2009a. Human disturbance as a cause of bias in ecological indicators for sandy beaches: Experimental evidence for the effects of human trampling on ghost crabs (Ocypode spp.). Ecological Indicators 9, 913-921.

Lucrezi, S., Schlacher, T.A., Walker, S., 2009b. Monitoring human impacts on sandy shore ecosystems: a test of ghost crabs (Ocypode spp.) as biological indicators on an urban beach. Environmental Monitoring and Assessment 152, 413-424.

Lucrezi, S., Schlacher, T.A., Robinson, W., 2010. Can storms and shore armouring exert additive effectson sandy-beach habitats and biota? Marine and Freshwater Research 61, 951-962.

Lucrezi, S., Schlacher, T.A., 2014. The ecology of ghost crabs, Oceanography and Marine Biology. CRC Press, 201-256.

McLachlan, A., Brown, A.C., 2006. The ecology of sandy shores (2nd ed.). Amsterdam ; London: Elsevier.

McLachlan, A., Defeo, O., Jaramillo, E., Short, A. D., 2013. Sandy beach conservation and recreation: guidelines for optimising management strategies for multi-purpose use. Ocean & coastal management, 71, 256-268.

Morrow, K., Bell, S.S., Tewfik, A., 2014. Variation in ghost crab trophic links on sandy beaches. Marine Ecology Progress Series 502, 197-206.

Moss, D., McPhee, D., 2006. The impacts of recreational four-wheel driving on the abundance of the ghost crab (Ocypode cordimanus) on a subtropical sandy beach in SE Queensland. Coastal Management 34, 133-140.

Nordstrom, K.F., 2004. Beaches and dunes of developed coasts. Cambridge University Press.

Noriega, R., Schlacher, T., Smeuninx, B., 2012. Reductions in ghost crab populations reflect urbanization of beaches and dunes. Journal of Coastal Research 28, 123-131.

Orr, M., Zimmer, M., Jelinski, D.E. and Mews, M. 2005. Wrack deposition on different beach types: spatial and temporal variation in the pattern of subsidy. Ecology 86, 1496-1507.

Pelletier, A.J., Jelinski, D.E., Treplin, M. and Zimmer, M. 2011. Colonisation of beach-cast macrophyte Wrack patches by Talitrid amphipods: a primer. Estuaries and coasts 34, 863-871.

Polis, G.A., Anderson, W.B. and Holt, R.D. 1997. Toward an integration of landscape and food web ecology: the dynamics of spatially subsidized food webs. Annual review of ecology and systematics 28, 289-316.

Pombo, M., Turra, A., 2013. Issues to be considered in counting burrows as a measure of Atlantic ghost crab populations, an important bioindicator of sandy beaches. PloS one 8, e83792.

Quijón, P., Jaramillo, E., Contreras, H., 2001. Distribution and habitat structure of Ocypode gaudichaudii H. Milne Edwards & Lucas, 1843, in sandy beaches of northern Chile. Crustaceana 74, 91-103.

R Core Team, 2014. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2012. ISBN 3-900051-07-0.

Reyes-Martínez, M.J., Ruíz-Delgado, M.C., Sánchez-Moyano, J.E., García-García, F.J., 2015. Response of intertidal sandy-beach macrofauna to human trampling: An urban vs. natural beach system approach. Marine Environmental Research 103, 36-45.

Rittschof, D. 1992. Chemosensation in the daily life of crabs. American Zoologist 32, 363-369.

Rose, M.D. and Polis, G.A. 1998. The distribution and abundance of coyotes: the effects of allochthonous food subsidies from the sea. Ecology79, 998-1007.

Roth, J.D. 2003. Variability in marine resources affects arctic fox population dynamics. Journal of Animal Ecology, 72, 668-676.

Schlacher, T.A., De Jager, R., Nielsen, T., 2011. Vegetation and ghost crabs in coastal dunes as indicators of putative stressors from tourism. Ecological Indicators 11, 284-294.

Schlacher, T.A., Dugan, J., Schoeman, D.S., Lastra, M., Jones, A., Scapini, F., McLachlan, A., Defeo, O., 2007a. Sandy beaches at the brink. Diversity and Distributions 13, 556-560.

Schlacher, T.A., Lucrezi, S., Connolly, R.M., Peterson, C.H., Gilby, B.L., Maslo, B., Olds, A.D., Walker, S.J., Leon, J.X., Huijbers, C.M., Weston, M.A., 2016. Human threats to sandy beaches: A meta- analysis of ghost crabs illustrates global anthropogenic impacts. Estuarine, Coastal and Shelf Science 169, 56-73.

Schlacher, T.A., Schoeman, D.S., Dugan, J., Lastra, M., Jones, A., Scapini, F., McLachlan, A., 2008. Sandy beach ecosystems: key features, sampling issues, management challenges and climate change impacts. Marine Ecology 29, 70-90.

Schlacher, T.A., Schoeman, D.S., Lastra, M., Jones, A., Dugan, J., Scapini, F. and McLachlan, A. 2006. Neglected ecosystems bear the brunt of change. Ethology Ecology & Evolution 18, 349-351.

Schlacher, T.A., Strydom, S., Connolly, R.M., 2013a. Multiple scavengers respond rapidly to pulsed carrion resources at the land–ocean interface. Acta Oecologica 48, 7-12.

Schlacher, T.A., Strydom, S., Connolly, R.M. and Schoeman, D., 2013b. Donor-control of scavenging food webs at the land-ocean interface. PloS ONE 8, e68221.

Schlacher, T.A., Thompson, L., Price, S., 2007b. Vehicles versus conservation of invertebrates on sandy beaches: mortalities inflicted by off‐road vehicles on ghost crabs. Marine Ecology 28, 354- 367.

Skaug, H., Fournier, D., Nielsen, A., Magnusson, A., Bolker, B., 2012. Generalized linear mixed models using AD model builder. R package version 0.7 2.

Spiller, D.A., Piovia-Scott, J., Wright, A.N., Yang, L.H., Takimoto, G., Schoener, T.W. and Iwata, T. 2010. Marine subsidies have multiple effects on coastal food webs. Ecology 91, 1424-1434.

Steinberg, P. 1994. Lack of short-term induction of phlorotannins in the Australasian brown algae Ecklonia radiata and Sargassum vestitum. Marine Ecology Progress Series 112, 129-133.

Stelling-Wood, T.P., Clark, G.F. and Poore, A.G.B. 2016. Responses of ghost crabs to habitat modification of urban sandy beaches. Marine Environmental Research 116, 32-40.

Strobl, C., Boulesteix, A.-L., Kneib, T., Augustin, T., Zeileis, A., 2008. Conditional variable importance for random forests. BMC Bioinformatics 9, 307.

Turra, A., Gonçalves, M., Denadai, M., 2005. Spatial distribution of the ghost crab Ocypode quadrata in low‐energy tide‐dominated sandy beaches. Journal of Natural History 39, 2163- 2177.

UN Atlas of Oceans. 2010. Human Settlements on the Coast: The ever more popular coasts. Accessed 17 January 2016, < http://www.oceansatlas.org>.

Warren, J., 1990. The use of open burrows to estimate abundances of intertidal estuarine crabs. Australian Journal of Ecology 15, 277-280.

Wellins, C.A., Rittschof, D. and Wachowiak, M. 1989. Location of volatile odor sources by ghost crab Ocypode quadrata (Fabricius). Journal of chemical ecology 15, 1161-1169.

Zuur, A., Ieno, E.N., Walker, N., Saveliev, A.A., Smith, G.M., 2009. Mixed effects models and extensions in ecology with R. Springer Science & Business Media.

Appendices

Fig. S1. Ranked importance of variables in predicting ghost crab burrow counts according to random forest algorithm.

Mean Beach Beach Mean grain wrack %TOM Beach length (m) slope (°) size (µm) volume (m3)

100 8.65 289.26 0.58 0.00013 Athol Bay Beach

860 5.59 306.96 0.69 4.1E-06 Balmoral Beach

100 6.71 328.08 0.56 4.55E-05 Bradleys Head Amph

100 6.08 335.02 1.05 3.04E-05 Bradleys Head Beach

240 6.83 372.20 0.32 1.33E-05 Camp Cove Beach

30 2.33 315.10 0.50 0.00024 Castle Rock Beach

230 6.05 225.42 3.87 1.65E-05 Chinamans Beach

280 7.07 282.48 1.80 1.75E-05 Clifton Gardens

570 4.87 295.59 0.74 4.12E-06 Clontarf Beach

100 5.76 285.01 0.68 4.32E-06 Cobblers Beach

100 3.06 328.64 0.56 6.13E-05 Collins Beach

30 4.82 560.13 0.69 6.29E-05 Delwood Beach

270 7.99 288.12 0.54 5.43E-05 Double Bay Beach

430 5.73 254.23 1.74 7.01E-07 Edwards Beach

50 5.17 606.99 1.03 0.00010 Esplanade Beach

80 4.54 849.91 0.74 5.67E-05 Fairlight Beach

100 4.09 352.43 1.06 0.00013 Forty Baskets Beach

120 6.05 448.22 1.13 2.11E-05 Gibsons Bay Beach

100 2.78 360.96 0.90 0.00013 Hermit Beach

80 6.43 382.80 0.66 2.43E-05 Kutti Beach

140 4.51 420.49 0.66 1.66E-05 Little Manly Cove

240 4.45 303.84 0.45 4.26E-05 Manly Cove East

230 2.47 291.33 0.71 1.36E-05 Manly Cove West

90 5.44 307.18 0.83 0.00033 Obelisk Beach

40 2.79 304.65 1.26 0.00027 Parsley Bay Beach

380 5.18 269.32 0.37 2.75E-05 Percival Park Beach

80 4.93 473.10 1.14 3.27E-05 Reef Beach

230 5.31 269.32 0.48 5.86E-06 Rose Bay Park Beach

150 0.85 258.28 0.96 4.66E-05 Sandy Bay Beach

220 8.77 486.19 0.60 4.94E-06 Shark Beach

100 4.50 294.59 0.59 6.56E-06 Sirius Cove Beach

50 2.60 394.49 0.57 0.00033 Taylors Bay Beach

500 6.27 186.43 1.76 6.7E-06 The Spit Beach

50 7.12 343.25 0.56 5.54E-05 Tingara Reserve

80 2.52 281.59 0.52 0.00012 Vaucluse Bay Beach

200 7.15 356.08 0.51 5.89E-05 Wharf Beach

110 6.82 289.53 0.50 4.45E-05 Wharf Beach South

90 5.67 263.35 0.92 0.000118 Whiting Beach

Table S1. List of surveyed beaches and corresponding anthropogenic variable category classifications.

Beach Beach slope Mean grain Mean wrack %TOM Seawall Cleaning Beach length (m) (°) size (µm) volume (m3) Development type regime

100 8.65 289.26 0.58 0.00013 Athol Bay Beach Absent None No mechanical

860 5.59 306.96 0.69 4.1E-06 Balmoral Beach Complete High Frequently

100 6.71 328.08 0.56 4.55E-05 Bradleys Head Amph Complete Low No mechanical

100 6.08 335.02 1.05 3.04E-05 Bradleys Head Beach Absent Medium No mechanical

240 6.83 372.20 0.32 1.33E-05 Camp Cove Beach Complete High Infrequently

30 2.33 315.10 0.50 0.00024 Castle Rock Beach Absent None No mechanical

230 6.05 225.42 3.87 1.65E-05 Chinamans Beach Absent None No mechanical

Clifton Gardens 280 7.07 282.48 1.80 1.75E-05 Partial Low Frequently Beach

570 4.87 295.59 0.74 4.12E-06 Clontarf Beach Complete Medium Frequently

100 5.76 285.01 0.68 4.32E-06 Cobblers Beach Absent None No mechanical

100 3.06 328.64 0.56 6.13E-05 Collins Beach Absent None No mechanical

30 4.82 560.13 0.69 6.29E-05 Delwood Beach Absent Medium No mechanical

270 7.99 288.12 0.54 5.43E-05 Double Bay Beach Complete High Infrequently

430 5.73 254.23 1.74 7.01E-07 Edwards Beach Complete High Frequently

50 5.17 606.99 1.03 0.00010 Esplanade Beach Absent Medium No mechanical

80 4.54 849.91 0.74 5.67E-05 Fairlight Beach Complete Medium No mechanical

100 4.09 352.43 1.06 0.00013 Forty Baskets Beach Partial Low Infrequently

120 6.05 448.22 1.13 2.11E-05 Gibsons Bay Beach Complete Medium Infrequently

100 2.78 360.96 0.90 0.00013 Hermit Beach Partial Medium Infrequently

80 6.43 382.80 0.66 2.43E-05 Kutti Beach Partial Medium Infrequently

140 4.51 420.49 0.66 1.66E-05 Little Manly Cove Complete Medium Frequently

240 4.45 303.84 0.45 4.26E-05 Manly Cove East Complete High Frequently

230 2.47 291.33 0.71 1.36E-05 Manly Cove West Complete High Frequently

90 5.44 307.18 0.83 0.00033 Obelisk Beach Absent None No mechanical

40 2.79 304.65 1.26 0.00027 Parsley Bay Beach Partial Low Infrequently

380 5.18 269.32 0.37 2.75E-05 Percival Park Beach Complete High Infrequently

80 4.93 473.10 1.14 3.27E-05 Reef Beach Partial Low No mechanical

230 5.31 269.32 0.48 5.86E-06 Rose Bay Park Beach Partial Low Infrequently

150 0.85 258.28 0.96 4.66E-05 Sandy Bay Beach Complete Medium No mechanical

220 8.77 486.19 0.60 4.94E-06 Shark Beach Complete Low No mechanical

100 4.50 294.59 0.59 6.56E-06 Sirius Cove Beach Complete Medium No mechanical

50 2.60 394.49 0.57 0.00033 Taylors Bay Beach Absent None No mechanical

500 6.27 186.43 1.76 6.7E-06 The Spit Beach Partial High Infrequently

50 7.12 343.25 0.56 5.54E-05 Tingara Reserve Absent None No mechanical

80 2.52 281.59 0.52 0.00012 Vaucluse Bay Beach Absent Low Infrequently

200 7.15 356.08 0.51 5.89E-05 Wharf Beach Complete High Infrequently

110 6.82 289.53 0.50 4.45E-05 Wharf Beach South Complete High Infrequently

90 5.67 263.35 0.92 0.000118 Whiting Beach Absent None No mechanical